Docstoc

Executive Summary

Document Sample
Executive Summary Powered By Docstoc
					                                 Executive Summary

Telmisartan is an angiotensin receptor blocker (ARB) currently approved in the United
States (since 10-Nov-1998) for the treatment of hypertension.
With this efficacy supplement, the sponsor seeks the approval of telmisartan for the
indication for telmisartan to reduce risk of myocardial infarction (MI), stroke,
cardiovascular (CV) deaths or hospitalization for congestive heart failure (CHF) in
patients 55 years or older at high risk of developing major CV events.
This submission consists of one pivotal clinical trial, ONTARGET, and two supportive
trials, TRANSCEND and PRoFESS.

ONTARGET
ONTARGET was a large multicenter, double-blind, parallel group, active-controlled,
double-dummy trial.
ONTARGET randomized 25,620 subjects over age 55 and at high risk for CV events to
telmisartan (T) 80 mg (n=8,542), ramipril (R) 10 mg (n=8,576), or their combination
(T/R) (n=8502).
The objective of ONTARGET is to show that the T/R combination is superior to R alone
or that T is non-inferior to R to reduce the 4-fold composite endpoint of MI, stroke, CV
deaths or hospitalization for CHF. Reference is also made here to a “3-fold” composite,
which omits hospitalization for heart failure, because this was the primary end point in
the study that established the effectiveness of R.
The mean follow up was 54 months.
Baseline therapy included anti-platelet drugs (95%), lipid lowering drugs (65%),
β-blockers (57%), calcium channel blockers (34%), nitrates (29%) and diuretics (28%).

TRANSCEND
TRANSCEND, similar to ONTARGET in design, was conducted contemporaneously
with ONTARGET, the only differences being that the subjects were angiotensin
converting enzyme inhibitor (ACE-I)-intolerant, and patients with clinically significant
proteinuria were excluded.
TRANSCEND randomized 5,926 subjects to T (n=2,954) or placebo (PBO) (n=2,972).
The objective was to compare effect of T vs PBO on the same 4-fold composite
endpoint as used in ONTARGET.
The mean duration of follow up was 56 months.
All patients received concomitant medications as needed.
PRoFESS
PRoFESS was a multicenter, double-blind, parallel-group, active- and placebo-
controlled, double-dummy, factorial design trial in subjects who had had an ischemic
stroke within 120 days of study entry.
PRoFESS randomized 20,332 subjects to clopidogrel or a fixed-dose combination of
aspirin plus dipyridamole (ASA+DP), and to T or placebo (PBO).
The mean duration of follow up was 30 months.
The primary endpoint was time to first recurrent stroke.
For the T vs PBO comparison, the secondary endpoints included the same 4-fold
composite endpoint used in ONTARGET.

EFFICACY REVIEW FINDINGS
ONTARGET
ONTARGET failed to demonstrate superiority of the T/R combination over R for the
primary efficacy endpoint (HR 0.99, 95% CI 0.92, 1.07, p=0.8462).
The sponsor’s NI margin was based upon a 3-component end point (omitting heart
failure hospitalizations) in the HOPE study (ramipril vs. placebo in a similar high-risk
population).
The sponsor successfully demonstrated non-inferiority of T to R for the Full Analysis Set
(FAS), the Per-Protocol Set (PPS) and for subjects who took the full 10 mg dose of
ramipril (FAS-10) using the pre-specified NI margin of 1.13 (Table 1).

Table 1 Non-Inferiority analyses of ONTARGET primary efficacy endpoint (T vs R)
                                           T                 R            HR (97.5% CI)
Randomized, n (%)                     8,542 (100)        8,576 (100)           ---
                                     FAS Population
Primary endpoint, n (%)               1,423 (16.7)       1,412 (16.5)
                                                                          1.01 (0.93, 1.10)
Events per 100 patient-years              3.87               3.82
                                     PPS Population
Primary endpoint, n (%)               1,214 (14.2)       1,164 (13.6)
                                                                          1.02 (0.93, 1.12)
Events per 100 patient-years              3.66               3.58
                                    FAS-10 Population
Primary endpoint, n (%)               1,237 (14.5)       1,183 (13.8)
                                                                          1.03 (0.94, 1.13)
Events per 100 patient-years              3.83               3.72

The reviewers calculated alternative NI margins from a meta-analysis of the triad of
placebo-controlled ACE-I trials—HOPE (1993-1999), EUROPA (perindopril; 1997-
2003), and PEACE (trandolapril; 1996-2003).




                                         Page 2 of 5
                             Table 2 Alternative NI margins
                           CI          90     95       95   99
                           Preserve 50% 20% 50% 50%
                           HOPE        1.09 1.13 1.08 1.07
                           3 studies 1.06 --           1.04 --


Depending on the confidence interval (CI) used, the amount of the reference effect to be
preserved, and whether only data from HOPE trial alone or integrated data from HOPE,
EUROPA and PEACE trials are used, the NI margin varies from 1.04 to 1.13.
The reviewers cite other reasons to warrant a conservative approach. With one study,
one cannot estimate between-trial variability. The effect size seen in HOPE (2000) was
larger than has been seen in subsequent studies EUROPA (2003) and PEACE (2004).
This calls into question the constancy assumption.




TRANSCEND and PRoFESS
The supportive trials, TRANSCEND and PRoFESS, do not show superiority of T over
PBO for the protocol-specified primary endpoints (Table 3).
The failure of T to win over PBO in TRANSCEND trial, which contemporaneously
enrolled patients with similar CV risk as in ONTARGET (albeit ACE-I intolerant), does
not support interchangeability of T for R, and further suggests that changes in the
standard of care have altered the clinical event rates in this patient population so that
the constancy assumptions do not hold.
The sponsor’s post-hoc analysis of pooled PRoFESS and TRANSCEND data showed a
nominally significant reduction in the relative risk of the 3-fold composite (HOPE) or
4-fold composite (ONTARGET) endpoints (Table 3). After adjusting for multiple
comparisons, however, no post hoc result is statistically significant.




                                         Page 3 of 5
   Table 3 Analysis of the 3- and 4-component composite endpoints in HOPE, 

                ONTARGET, TRANSCEND and PRoFESS trials 

                        HOPE (superiority of ramipril vs placebo)
                                        R                 PBO          HR (95% CI)
Randomized, n (%)                  4,645 (100)         4,652 (100)          ---
4-component endpoint, n (%)         726 (15.6)          919 (19.8) 0.775 (0.703, 0.854)
3-component endpoint, n (%)         651 (14.0)          826 (17.8)   0.78 (0.70, 0.86)
                ONTARGET (non-inferiority of telmisartan vs ramipril)
                                         T                  R         HR (97.5% CI)
Randomized, n (%)                  8,542 (100)         8,576 (100)          ---
4-component endpoint, n (%)       1,423 (16.7)        1,412 (16.5)   1.01 (0.93, 1.10)
3-component endpoint, n (%)       1,190 (13.9)        1,210 (14.1)   0.99 (0.90, 1.08)
                         TRANSCEND (telmisartan vs placebo)
                                         T                PBO          HR (95% CI)
Randomized, n (%)                  2,954 (100)         2,972 (100)          ---
4-component endpoint, n (%)         465 (15.7)          504 (17.0)   0.92 (0.81, 1.05)
3-component endpoint, n (%)         384(13.0)           440 (14.8)   0.87 (0.76, 1.00)
                           PRoFESS (telmisartan vs placebo)
                                         T                PBO          HR (95% CI)
Randomized, n (%)                 10,146 (100)        10,186 (100)          ---
4-component endpoint, n (%)       1,367 (13.5)        1,463 (14.4)   0.94 (0.87, 1.01)
3-component endpoint, n (%)       1,289 (12.7)        1,377 (13.5)   0.94 (0.87, 1.02)
     TRANSCEND & PRoFESS (telmisartan vs placebo) Total/NoACE-I population
                                         T                PBO          HR (95% CI)
Randomized, n (%)                  8,587 (100)        8,290 (100)           ---
4-component endpoint, n (%)       1,120 (13.0)        1,196 (14.4)   0.90 (0.83, 0.98)
3-component endpoint, n (%)       1,015 (11.8)        1,106 (13.3)   0.88 (0.81, 0.96)


SAFETY REVIEW FINDINGS
For safety, the reviewers evaluated the safety data from the ONTARGET, TRANSCEND
and PRoFESS trial populations of a total of 58,553 patient years’ exposure on 17,085
patients treated with T (average observation time = 1,251 days) in ONTARGET,
TRANSCEND and PRoFESS trials, compared to 35,207 patient-years exposure on
8,576 patients treated with R (av. observation time = 1,499 days) in ONTARGET, and
22,929 patient-years exposure on 8,249 patients who received placebo (av. observation
time = 1,015 days) in TRANSCEND and PRoFESS.
The frequencies of deaths and SAEs while on treatment were similar between treatment
groups in ONTARGET.
The T/R combination, compared to R or T group, experienced more deaths associated
with renal failure {T/R group (0.19%) vs T (0.09%) or R (0.13%)}, more SAEs of renal
failure, renal and urinary disorders, GI disorders, hypotension and hyperkalemia, and
more frequent AEs leading to permanent discontinuations (namely, hypotension and
cough).
Safety signals were found for the following, which are of uncertain clinical significance:


                                        Page 4 of 5
•	 An increased risk of sepsis was found (not statistically significant) in T group
   compared to R or PBO in all 3 trials. In TRANSCEND and PRoFESS, more deaths
   associated with sepsis were associated with T compared to PBO. No explanation
   for this finding is known. A review of case report forms suggests that these patients
   had multiple co-morbid illnesses (e.g., decubitus ulcers, urosepsis, gall stones, heart
   valve replacements with predisposition to infective endocarditis, COPD and
   pneumonia, different types of cancers requiring chemotherapy with consequent
   neutropenia, surgery with post-surgical infections, diabetes with predisposition to
   skin infections, etc.,) or confounding conditions (e.g., burns, motor vehicle accidents,
   cat scratch, falls, etc.) associated with sepsis.
•	 An increased risk of malignancies was found in ONTARGET for T/R group vs R for
   all patients (HR 1.14, 95% CI 1.03, 1.26, p = 0.0089) and for patients without cancer
   at baseline (HR 1.12, 95% CI 1.01, 1.25, P = 0.036). In TRANSCEND, too, a higher
   rate of malignancies was found in patients treated with T vs PBO. There is no
   evidence to date that ARBs are associated with a higher risk of malignancies.




                                        Page 5 of 5
 COMBINED CLINICAL & STATISTICAL REVIEW 


           Application TypeNDA (New Efficacy
                           Supplement)
    Application Number(s) 20-850
      Priority or Standard SE 1, Sequence 025
            Submit Date(s)        2008-12-16
          Received Date(s)        2008-12-16
         PDUFA Goal Date          2009-10-15
           Division / Office      DCaRP/ODE 1/OND
     Reviewer Name(s) Khin Maung U, M.D. 

                       Jialu Zhang, Ph.D.
Review Completion Date 2009-06-29
      Established Name            Telmisartan
 (Proposed) Trade Name            Micardis®
       Therapeutic Class          Angiotensin Receptor Blocker
               Applicant          Boehringer Ingelheim
           Formulation(s) Tablets
         Dosing Regimen 80 mg once daily, orally
             Indication(s) Reduction of risk of myocardial
                           infarction, stroke, death from
                           cardiovascular causes or
                           hospitalization for congestive
                           heart failure
   Intended Population(s) Patients 55 year or older at
                           high risk of developing major
                           cardiovascular events
Template Version: March 6, 2009
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

                                                Table of Contents

1     RECOMMENDATIONS/RISK BENEFIT ASSESSMENT ....................................... 10

    1.1     Recommendation on Regulatory Action ........................................................... 11

    1.2     Risk Benefit Assessment.................................................................................. 11

    1.3     Recommendations for Postmarket Risk Evaluation and Mitigation Strategies . 12

    1.4     Recommendations for Postmarket Requirements and Commitments .............. 12

2     INTRODUCTION AND REGULATORY BACKGROUND ...................................... 13

    2.1     Product Information .......................................................................................... 13

    2.2     Tables of Currently Available Treatments for Proposed Indications ................. 13

    2.3     Availability of Proposed Active Ingredient in the United States ........................ 13

    2.4     Important Safety Issues With Consideration to Related Drugs......................... 13

    2.5     Summary of Presubmission Regulatory Activity Related to Submission .......... 14

    2.6     Other Relevant Background Information .......................................................... 15

3     ETHICS AND GOOD CLINICAL PRACTICES....................................................... 16

    3.1     Submission Quality and Integrity ...................................................................... 16

    3.2     Compliance with Good Clinical Practices ......................................................... 16

    3.3     Financial Disclosures........................................................................................ 17

4	 SIGNIFICANT EFFICACY/SAFETY ISSUES RELATED TO OTHER REVIEW

   DISCIPLINES ......................................................................................................... 18

    4.1     Chemistry Manufacturing and Controls ............................................................ 18

    4.2     Clinical Microbiology......................................................................................... 18

    4.3     Preclinical Pharmacology/Toxicology ............................................................... 18

    4.4     Clinical Pharmacology ...................................................................................... 18

5     SOURCES OF CLINICAL DATA............................................................................ 19

    5.1 Tables of Studies/Clinical Trials ....................................................................... 19

    5.2 Review Strategy ............................................................................................... 19

    5.3 Discussion of Individual Studies/Clinical Trials ................................................. 22

      5.3.1 Pivotal Clinical Trial.................................................................................... 22

      5.3.2 Supportive clinical trials.............................................................................. 26

      5.3.3 Statistical Methods ..................................................................................... 28

6     REVIEW OF EFFICACY......................................................................................... 32

    Efficacy Summary...................................................................................................... 32

    6.1 Indication .......................................................................................................... 35

      6.1.1 Methods ..................................................................................................... 35

      6.1.2 Demographics ............................................................................................ 38

      6.1.3 Subject Disposition..................................................................................... 40

      6.1.4 Analysis of Primary Endpoint(s) ................................................................. 42

      6.1.5 Analysis of Secondary Endpoints(s) .......................................................... 52



                                                               2

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

       6.1.6	    Statistical considerations of Primary and Secondary Efficacy Endpoints ... 60

       6.1.7	    Other Endpoints ......................................................................................... 90

       6.1.8	    Subpopulations .......................................................................................... 91

       6.1.9	    Analysis of Clinical Information Relevant to Dosing Recommendations .... 93

       6.1.10    Discussion of Persistence of Efficacy and/or Tolerance Effects................. 94

       6.1.11    Additional Efficacy Issues/Analyses ........................................................... 96

7     REVIEW OF SAFETY........................................................................................... 102

    Safety Summary ...................................................................................................... 102

    7.1 Methods.......................................................................................................... 103

      7.1.1	 Studies/Clinical Trials Used to Evaluate Safety ....................................... 104

      7.1.2	 Categorization of Adverse Events............................................................ 105

      7.1.3	 Pooling of Data across Studies/Clinical Trials to Estimate and Compare 

             Incidence.................................................................................................. 105

    7.2 Adequacy of Safety Assessments .................................................................. 105

      7.2.1	 Overall Exposure at Appropriate Doses/Durations and Demographics of 

             Target Populations ................................................................................... 105

    7.3 Major Safety Results ...................................................................................... 108

      7.3.1	 Deaths...................................................................................................... 108

      7.3.2	 Nonfatal Serious Adverse Events ............................................................ 111

      7.3.3	 Dropouts and/or Discontinuations ............................................................ 112

      7.3.4	 Significant Adverse Events ...................................................................... 117

    7.4 Supportive Safety Results .............................................................................. 118

      7.4.2	 Laboratory Findings ................................................................................. 118

      7.4.3	 Vital Signs ................................................................................................ 120

      7.4.4	 Electrocardiograms (ECGs) ..................................................................... 120

    7.5 Other Safety Explorations............................................................................... 121

      7.5.1	 Dose Dependency for Adverse Events .................................................... 121

      7.5.3	 Drug-Demographic Interactions ............................................................... 121

      7.5.4	 Drug-Disease Interactions........................................................................ 121

      7.5.5	 Drug-Drug Interactions............................................................................. 121

    7.6 Additional Safety Evaluations ......................................................................... 122

      7.6.2	 Human Reproduction and Pregnancy Data.............................................. 122

      7.6.4	 Overdose, Drug Abuse Potential, Withdrawal and Rebound.................... 122

    7.7 Additional Submissions / Safety Issues .......................................................... 122

      7.7.1	 Sepsis and fatal sepsis as a new potential safety signal in patients treated 

             with telmisartan ........................................................................................ 122

      7.7.2	 Malignancies as a new potential safety signal in patients treated with 

             telmisartan ............................................................................................... 124

8     POSTMARKETING EXPERIENCE ...................................................................... 125


9     APPENDICES ...................................................................................................... 125

    9.1     Literature Review/References ........................................................................ 125



                                                              3

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

    9.1.1	 Effect of ACE-Is on improving survival in patients with chronic coronary 

            artery disease .......................................................................................... 125

    9.1.2 	Dose considerations for Angiotensin Converting Enzyme Inhibitors (ACE-Is)

            ................................................................................................................. 128

    9.1.3 Dose considerations for Angiotensin Receptor Blockers (ARBs) ............... 128

    9.1.4 ARBs as substitutes for ACE-Is in treatment of ACE-I intolerant patients .. 129

    9.1.5 The issue of ARB-MI paradox .................................................................... 130

    9.1.6 Are the effects of ARBs additive on top of ACE-inhibitors? ........................ 132

  9.2 Labeling Recommendations ........................................................................... 133

  9.3 Advisory Committee Meeting.......................................................................... 133

  9.4 Non-essential tables of analyses of clinical trial data ..................................... 133

    9.4.1	 Non-essential tables of analyses of follow-up clinical outcome data at 6 

            months and cumulatively at Year 1, 2, 3, 4, and 5 ................................... 133





                                                              4

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

                                                   Table of Tables
Table 1 Sites requested for GCP inspection by DSI..................................................... 17

Table 2: Tables of pivotal and supportive clinical trials for NDA 20-850........................ 19

Table 3 Treatment groups in HOPE, ONTARGET, TRANSCEND & PRoFESS trials .. 21

Table 4 2x2 factorial design of PRoFESS trial ............................................................. 27

Table 5 Statistical analyses and data sets used........................................................... 30

Table 6 Number of patients included in sensitivity analyses (ONTARGET) ................. 31

Table 7 Composite 4-fold and 3-fold primary endpoint data in HOPE .......................... 36

Table 8 Baseline characteristics and primary efficacy endpoint results in HOPE, 

          EUROPA and PEACE trials of ACE inhibitors ............................................... 36

Table 9 Baseline characteristics in HOPE, ONTARGET and TRANSCEND trials ....... 37

Table 10 Demographics and baseline characteristics by treatment (FAS) ................... 39

Table 11 Disposition of patients in ONTARGET, TRANSCEND & PRoFESS (FAS).... 40

Table 12 Analysis of ONTARGET primary endpoint for superiority (T/R vs R)............. 42

Table 13 Analysis of the primary renal endpoint in ONTARGET / FAS (DN) ............... 43

Table 14 Non-Inferiority analyses of ONTARGET composite primary efficacy endpoint 

          (T vs R) .......................................................................................................... 44

Table 15 Sensitivity analyses for non-inferiority of T vs R in ONTARGET trial............. 46

Table 16 Analysis of the primary composite endpoint in TRANSCEND (FAS)............. 47

Table 17 Cox proportional analysis of recurrent stroke in PRoFESS (RAN set) .......... 48

Table 18 Time-dependent analysis of recurrent stroke in PRoFESS (RAN set)........... 49

Table 19 The primary endpoint in pooled TRANSCEND and PRoFESS (FAS – 

          Total/NoACE-I) population ............................................................................. 50

Table 20 Time-dependent analysis of the primary composite endpoint in the pooled 

          TRANSCEND and PRoFESS populations (FAS – Total/NoACE-I)................ 51

Table 21 Non-inferiority analyses of T vs R for 3-fold composite endpoint................... 53

Table 22 Individual components of the primary composite endpoint / FAS (ONTARGET 

          trial)................................................................................................................ 54

Table 23 Incidence of MI subcategories** (FAS).......................................................... 54

Table 24 Composite 3-fold endpoint in TRANSCEND (FAS) compared to HOPE ....... 56

Table 25 Analysis of composite 3-fold (HOPE) endpoint in pooled TRANSCEND & 

          PRoFESS patient population (FAS – Total/NOACE-I) ................................... 58

Table 26 Time-dependent analysis of the composite 3-fold (HOPE) endpoint in the 

          pooled TRANSCEND and PRoFESS populations (FAS – Total/NoACE-I) .... 59

Table 27 NI Margins derived based on HOPE Trial alone............................................. 61

Table 28 Summary of three ACE-I trials in patients with CV risk factors, listed in 

          chronological order ........................................................................................ 62

Table 29 Summary of total number of events in PEACE trial ........................................ 63

Table 30 NI Margins based on HOPE, EUROPA and PEACE trials ............................. 63

Table 31 Non-Inferiority analyses of ONTARGET composite primary efficacy endpoint 

          (T vs R) .......................................................................................................... 66

Table 32 Sensitivity analyses for non-inferiority of T vs R in ONTARGET trial............. 67

Table 33 Summary of Analyses on HOPE and ONTARGET......................................... 68



                                                                5

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 34 Analyses of HOPE trial by adjusting covariates............................................. 72

Table 35 Analyses on EUROPA Trial by Adjusting Covariates (HOPE endpoint) ......... 76

Table 36 Analyses on HOPE and EUROPA Trials by Adjusting Covariates ................. 79

Table 37 Analyses on PEACE Trial by Adjusting Covariates (reconstructed 3-fold 

         endpoint)........................................................................................................ 82

Table 38 Analyses of ONTARGET Trial by Adjusting Covariates (HOPE endpoint) ..... 85

Table 39 Analysis of ONTARGET primary endpoint for superiority (T/R vs R)............. 87

Table 40 Analysis of the primary composite endpoint in TRANSCEND (FAS)............. 88

Table 41 Cox proportional analysis of recurrent stroke in PRoFESS (RAN set) .......... 88

Table 42 Summary of Analyses in TRANSCEND and PRoFESS ................................ 89

Table 43 Frequency of composite primary endpoint by year (FAS) ............................. 94

Table 44 ACE inhibitor vs Placebo trials in Post-MI Left Ventricular dysfunction ......... 97

Table 45 ACE inhibitor trials in patients with cardiovascular risk factors ...................... 97

Table 46 Meta-analysis of mortality data in HOPE, EUROPA and PEACE trials ......... 98

Table 47 Placebo outcomes in HOPE, EUROPA, and PEACE trials. .......................... 98

Table 48 Annualized event rates in HOPE, ONTARGET and TRANSCEND trials ...... 99

Table 49 Baseline characteristics and the primary efficacy endpoint results in 

         ONTARGET and TRANSCEND compared to HOPE, EUROPA and PEACE
         trials ............................................................................................................. 101

Table 50 Numbers of patients included in different safety sets for safety analyses ... 104

Table 51: Observation time on treatment in randomized periods of clinical trials........ 106

Table 52 Exposure by duration (days) of treatment ................................................... 107

Table 53 Deaths in ONTARGET trial during randomized period (safety evaluation) .. 108

Table 54 Deaths in the subgroup of patients with diabetic nephropathy ONTARGET 

         trial during randomized period (safety evaluation) ....................................... 109

Table 55 Deaths in TRANSCEND trial – randomized period (safety evaluation) ....... 110

Table 56 Deaths in PRoFESS trial – randomized period (safety evaluation) ............. 110

Table 57 Nonfatal SAEs in ONTARGET trial – randomized period ............................ 111

Table 58 Nonfatal SAEs in TRANSCEND trial – randomized period.......................... 112

Table 59 All SAEs in PRoFESS trial – randomized period ......................................... 112

Table 60 Permanent discontinuation of study drugs in ONTARGET trial ................... 113

Table 61 Cumulative permanent discontinuation of study drug in ONTARGET trial ... 114

Table 62 Permanent discontinuation of active study drugs in ONTARGET trial.......... 114

Table 63 Cumulative permanent discontinuation of active drug in ONTARGET trial... 115

Table 64 Permanent discontinuations in patients with diabetic nephropathy in 

         ONTARGET trial .......................................................................................... 115

Table 65 Permanent treatment discontinuations in TRANSCEND trial ...................... 116

Table 66 Permanent treatment discontinuations in PRoFESS trial ............................ 116

Table 67 AEs of special interest in the ONTARGET trial............................................ 118

Table 68 Number (%) of patients with sepsis and fatal sepsis in large clinical outcome 

         trials of telmisartan....................................................................................... 123

Table 69 Postmarketing exposure for the indication: treatment of hypertension ........ 125

Table 70 Summary of findings in the medical literature regarding issues related to the 

         effect of ACE-Is in patients with heart disease............................................. 127



                                                               6

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 71 Role of ARBs in patients with CAD and/or high risk of CVD........................ 130

Table 72 Hazard ratios (95% CI) of cumulative outcomes in patients who had a non-

         fatal MI: comparison of T/R vs R.................................................................. 134

Table 73 Hazard ratios (95% CI) of cumulative outcomes in patients who had a non-

         fatal MI: comparison of T vs R ..................................................................... 134

Table 74 Hazard ratios (95% CI) of cumulative outcomes in patients who had a non-

         fatal stroke: comparison of T/R vs R ............................................................ 134

Table 75 Hazard ratios (95% CI) of cumulative outcomes in patients who had a non-

         fatal stroke: comparison of T vs R................................................................ 135

Table 76 Hazard ratios (95% CI) of cumulative outcomes in patients who were 

         hospitalized due to CHF: comparison of T/R vs R ....................................... 135

Table 77 Hazard ratios (95% CI) of cumulative outcomes in patients who were 

         hospitalized due to CHF: comparison of T vs R ........................................... 135





                                                       7

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

                                                  Table of Figures
Figure 1 ONTARGET: study flow chart ........................................................................ 22

Figure 2 ONTARGET – Run-in Phase flow chart ......................................................... 24

Figure 3 ONTARGET – Randomization Phase flow chart ............................................ 24

Figure 4 Kaplan-Meier plot for time-to-recurrent stroke in PRoFESS – T vs PBO (RAN 

          set)................................................................................................................. 48

Figure 5 Kaplan-Meier plot for the primary composite endpoint in the pooled 

          TRANSCEND and PRoFESS populations – FAS (Total/NoACE-I)................ 50

Figure 6 Kaplan-Meier plot for the composite 3-fold (HOPE) endpoint in the pooled 

          TRANSCEND and PRoFESS populations – FAS (Total/NoACE-I)................ 58

Figure 7 Summary of HOPE, EUROPA and PEACE trials using the 3-fold composite 

          (HOPE) endpoint............................................................................................ 62

Figure 8 NI Margins based on Different Discounting ..................................................... 65

Figure 9 Hazard Ratio and 95% Confidence Interval in Three Trials ............................ 69

Figure 10 Comparison of placebo event rate in HOPE, TRANSCEND and PEACE ..... 70

Figure 11 Comparison of event rate in ramipril in HOPE and ONTARGET................... 70

Figure 12 Comparison of event rate in telmisartan in TRANSCEND and ONTARGET . 71

Figure 13 Hazard ratio in HOPE Trial by Covariates ..................................................... 74

Figure 14 Hazard ratio of ramipril over placebo adjusted for covariates in HOPE........ 75

Figure 15 NI margins based on hazard ratio adjusted for covariates in HOPE ............ 75

Figure 16 Hazard Ratio in EUROPA Trial by Covariates (HOPE endpoint) .................. 78

Figure 17 Hazard ratio adjusted for each covariate in EUROPA ................................... 78

Figure 18 NI margins based on hazard ratio adjusted for covariates in EUROPA ........ 79

Figure 19 Hazard ratios adjusted by covariates in integrated HOPE and EUROPA trials

          ....................................................................................................................... 81

Figure 20 NI margins based on hazard ratio adjusted for covariates in integrated 

          EUROPA and HOPE trials ............................................................................. 82

Figure 21 Hazard ratio of perindopril over placebo adjusted for each covariate in 

          PEACE........................................................................................................... 84

Figure 22 Comparison of T vs R in subgroups (ONTARGET) / FAS ............................ 92

Figure 23 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint over 

          time (FAS)...................................................................................................... 94

Figure 24 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint over 

          time (PPS) ..................................................................................................... 95

Figure 25 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint over 

          time (FAS–10)................................................................................................ 95

Figure 26 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint over 

          time (PPS–without ACE-Is/ARBs).................................................................. 96

Figure 27 Unadjusted Kaplan–Meier curves according to ACEI prescribed within 30 

          days of discharge. (Based on data from Ann Intern Med 2004; 141: 102-12)

          ..................................................................................................................... 126

Figure 28 Survival of patients on ramipril therapy, other ACEI therapy, and no ACEI 

          therapy (Kaplan- Meier curves).................................................................... 126



                                                                 8

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 29 ARBs do not increase risk of MI overall (Forest plot of MI analyses in 25 

         clinical trials of ARBs) .................................................................................. 131

Figure 30 ARBs increase risk of MI overall (Forest plot of MI analyses in 12 clinical 

         trials of ARBs) .............................................................................................. 132





                                                            9

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


1 Recommendations/Risk Benefit Assessment
Background/Introduction
Telmisartan is currently approved in the United States (since 10-Nov-1998) for the
treatment of hypertension. This submission consists of one pivotal clinical trial {ONgoing
Telmisartan Alone and in combination with Ramipril Global Endpoint Trial,
(ONTARGET)} and two supportive trials {Telmisartan Randomized AssessmeNt Study
in ACE iNtolerant subjects with cardiovascular Disease, (TRANSCEND) and The
Prevention Regimen For Effectively avoiding Second Strokes (PRoFESS)}. This efficacy
supplement seeks the approval of telmisartan for the indication of use of 80 mg
telmisartan once a day to reduce risk of myocardial infarction (MI), stroke,
cardiovascular (CV) deaths or hospitalization for congestive heart failure (CHF) in
patients 55 year or older at high risk of developing major CV events.
ONTARGET was a large multicenter, double-blind, parallel group, active-controlled,
double-dummy trial which randomized 25,620 patients (1:1:1 ratio) to one of three
treatment groups: telmisartan (T) 80 mg (n=8,542), ramipril (R) 10 mg (n=8,576), or
combination (T/R) of telmisartan 80 mg and ramipril 10 mg (n=8502). The objective of
ONTARGET is to show that the T/R combination is superior to R alone and/or that T is
non-inferior to R to reduce the 4-fold composite primary efficacy endpoint of MI, stroke,
CV deaths or hospitalization for CHF in patients ≥ 55 years old at high risk of developing
major CV events. The mean follow up was 54 months. Baseline therapy included anti-
platelet drugs (95% of patients), lipid lowering drugs (65%), β-blockers (57%), calcium
channel blockers (34%), nitrates (29%) and diuretics (28%).
TRANSCEND, similar to ONTARGET in design, methods and endpoints, was
conducted contemporaneously with ONTARGET, the only differences being that the
patients were angiotensin converting enzyme inhibitor (ACE-I) intolerant, and patients
with clinically significant proteinuria were excluded. TRANSCEND randomized (1:1
ratio) 5,926 patients to T (n=2,954) or placebo (PBO) (n=2,972). The objective was to
compare effect of T vs PBO on the 4-fold composite endpoint in high CV risk patients
intolerant to ACE-Is. The mean duration of follow up was 56 months. All patients
received concomitant medications as needed.
PRoFESS was a large, multicenter, double-blind, parallel-group, active- and placebo-
controlled, double-dummy, factorial design trial of Aggrenox® (fixed-dose combination
product of aspirin plus extended-release dipyridamole, ASA+ER-DP) vs. clopidogrel,
with and without telmisartan to compare Aggrenox® to clopidogrel, and T to PBO in the
prevention of recurrent stroke in patients who had had an ischemic stroke within 120
days of study entry. PRoFESS randomized 20,332 patients (stratified by baseline usage
of ACE-Is) using a 2x2 factorial design to platelet therapy with clopidogrel or a fixed-
dose combination of ASA+ER-DP, and at the same time to T or PBO. The mean
duration of follow up was 30 months. The primary endpoint was time to first recurrent



                                           10 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

stroke. For the T vs PBO comparison, the secondary endpoints included the 4-fold
composite primary efficacy endpoint of MI, stroke, CV deaths or CHF hospitalizations.

1.1 Recommendation on Regulatory Action
Will defer to comment until after the AC meeting.

1.2 Risk Benefit Assessment
In the efficacy data, ONTARGET fails to demonstrate superiority of the T/R combination
over R for the primary efficacy endpoint (HR 0.99, 95% CI 0.92, 1.07, p=0.8462).
Non-inferiority of T to R is not found for any of the data sets {namely, the Full Analysis
Set (FAS), the Per-Protocol Set (PPS) and the FAS-10 (patients who took the full 10 mg
dose of ramipril)} using the reviewers’ NI margin of 1.08 (including the sensitivity
analyses), but the upper confidence bounds are within the relatively wider and more
liberal NI margin of 1.13 used by the sponsor. However, the sponsor’s NI margin is
based on (i) 20% two-sided confidence interval, (ii) the single HOPE trial so that
between-trial variability cannot be estimated, and (iii) the HOPE trial population that
received the prevailing treatment 10 years ago which is different from the current clinical
practice.
The annualized event rates in the placebo group are lower in TRANSCEND compared
to HOPE for the 4-fold composite endpoint (3.72% vs 5.09%, respectively) and the 3­
fold composite endpoints (3.34% vs 4.53%). Similarly, the annualized event rates in the
ramipril group are lower in ONTARGET compared to HOPE for the 4-fold (3.67% vs
3.95%, respectively) and the 3-fold (3.23% vs 3.51%) composite endpoints. These
findings suggest that the constancy assumption which the non-inferiority analysis relies
upon may no longer hold.
Regarding treatment-by-covariate interactions, most of the interaction terms of
treatment and clinical covariates are not significant, and do not seem to affect the
treatment effect in EUROPA, HOPE or ONTARGET trials with the possible exception of
use of anti-platelets in HOPE trial and use of β-blockers in ONTARGET.
The supportive trials, TRANSCEND and PRoFESS, do not show superiority of T over
PBO for the protocol-specified primary endpoints. The failure of T to win over PBO in
TRANSCEND trial, which contemporaneously enrolled patients with similar CV risk as in
ONTARGET (albeit ACE-I intolerant), does not support interchangeability of T for R, and
further suggests that changes in the standard of care have altered the clinical event
rates in this patient population so that the constancy assumptions do not hold any more.
From the safety perspective, the frequencies of deaths and SAEs while on treatment
were comparable between treatment groups in ONTARGET. The T/R combination,
compared to R or T group, experienced more deaths associated with renal failure {T/R
group (0.19%) vs T (0.09%) or R (0.13%)}, more SAEs of renal failure, renal and urinary


                                            11 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

disorders, GI disorders, hypotension and hyperkalemia, and more frequent AEs leading
to permanent discontinuations (namely, hypotension and cough).
Safety signals were found for the following, which are of uncertain clinical significance:
•	 An increased risk of sepsis was found (not statistically significant) in T group
   compared to R or PBO in all 3 trials. In TRANSCEND and PRoFESS, more deaths
   associated with sepsis were associated with T compared to PBO. No explanation
   for this finding is known. A review of case report forms suggests that these patients
   had multiple co-morbid illnesses or confounding conditions associated with sepsis.
•	 An increased risk of malignancies was found in ONTARGET for T/R group vs R for
   all patients (HR 1.14, 95% CI 1.03, 1.26, p = 0.0089) and for patients without cancer
   at baseline (HR 1.12, 95% CI 1.01, 1.25, P = 0.036). In TRANSCEND, too, a higher
   rate of malignancies was found in patients treated with T vs PBO. There is no
   evidence to date that ARBs are associated with a higher risk of malignancies.
While no major safety concerns are identified for this product, there was an overall lack
of a clinically meaningful beneficial effect.

1.3 	 Recommendations for Postmarket Risk Evaluation and Mitigation
      Strategies
Will defer to comment until after the AC meeting.

1.4 	 Recommendations for Postmarket Requirements and Commitments
Will defer to comment until after the AC meeting.




                                            12 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


2 Introduction and Regulatory Background

2.1 Product Information
Telmisartan is an orally-active, non-peptide, angiotensin II receptor blocker (ARB); it is
chemically described as 4'-[(1,4'dimethyl-2'-propyl [2,6'-bi-1H-benzimidazol]-1'­
yl)methyl]-[1,1'-biphenyl]-2-carboxylic acid. Its empirical formula is C33H30N4O2
(molecular weight: 514.63), and its structural formula is:




NDA 20-850 for telmisartan, marketed as MICARDIS®, was approved in the United
States November 10, 1998 for the treatment of hypertension. Other trade names are
Pritor® and Kinzalmono®.


2.2 Tables of Currently Available Treatments for Proposed Indications
For the proposed indication of the reduction in risk of myocardial infarction, stroke,
death from cardiovascular causes, or hospitalization for congestive heart failure, in
patients 55 years or older at high risk of developing a major cardiovascular event, there
are no other currently approved drugs in the US.

2.3 Availability of Proposed Active Ingredient in the United States
MICARDIS® tablets for oral administration are available in the US in three dose
strengths (20 mg, 40 mg, and 80 mg).

2.4 Important Safety Issues With Consideration to Related Drugs
As an angiotensin receptor blocker drug that acts directly on the renin-angiotensin
system, patients:
•	 who are pregnant may have fetal abnormalities (skull hypoplasia, craniofacial
   deformation, fetal limb contractures, hypoplastic lung development, anuria, renal
   failure, prematurity, intrauterine growth retardation, patent ductus arteriosus),
   oligohydramnios and neonatal morbidity and neonatal death;




                                            13 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

•	 with an activated renin-angiotensin system such as volume and/or salt depleted
   patients may develop symptomatic hypotension.
•	 whose renal function depend on the activity of the renin-angiotensin-aldosterone
   system (e.g., patients with severe congestive heart failure) may experience oliguria
   and/or progressive azotemia and acute renal failure;
•	 with unilateral or bilateral renal artery stenosis may experience increases in serum
   creatinine or blood urea nitrogen.

2.5 Summary of Presubmission Regulatory Activity Related to Submission
One pre-submission regulatory issue was that an “agreement was not reached between
the sponsor and the Division on defining the margin for the non-inferiority arm of the
ONTARGET trial.” The sponsor’s proposed non-inferiority (NI) margin was 1.13 (which
was calculated using the Hasselblad/Kong approach). The Division’s position during the
meeting between BI and FDA on 10-Apr-2001 was that “it would be acceptable to
exclude one half of the excess risk associated with placebo to give an excess risk of
8.5% (1.085 margin, which was 50% of the lower bound of the 95% C.I.) that can be
allowed for telmisartan compared to ramipril.” This criterion assures preservation of at
least 50% of the conservatively estimated ramipril effect, which the Division defines as
non-inferior effectiveness.
The sponsor’s argued that the upper bound of the 95% at 1.129 (<1.13) for the hazard
ratio of telmisartan compared to ramipril, is still 2.7 standard errors (SEs) of the control
effect from the HOPE trial. The NI margin of 1.085 would force telmisartan to maintain
3.5 SEs of effect above HOPE trial, which was too conservative.
The sample size calculated for the 1.085 NI margin was approximately 52,000 patients
(≈18,000 patients per treatment arm). The ONTARGET trial was sized at 25,620 (< half
of the calculated sample size).
BI and the Division did not come to an agreement on a statistical plan for the non-
inferiority analysis to determine efficacy; this issue will need to be discussed before a
Cardio-Renal Advisory Committee meeting.
Another regulatory issue was related to accounting for a potential effect of blood
pressure (BP) lowering on the primary outcome. While all patients received telmisartan
and/or ramipril (both effective antihypertensive drugs), the protocol did not pre-specify
target BP goals. Based on investigator judgment, additional anti-hypertensive agents
(excluding ARBs and ACE-Is) were permitted. To demonstrate that the effect of the
combination treatment on the primary endpoint is present independent of its blood
pressure (BP) lowering effect, the sponsor proposed to use the Cox proportional
hazards model with systolic blood pressure (SBP) measured at the last visit before the
primary outcome events occurred to be included as a time-dependent covariate. The
sponsor submitted that this model would quantify the effect of SBP on outcome in terms
of hazard ratio per mmHg increase in SBP, and would give hazard ratios between
treatment groups that were adjusted for in-trial BP measurements.


                                             14 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The Division pointed out that it would be very difficult for such a time-dependent
covariate adjustment to answer the question of whether the drug effect was attributed to
BP lowering, and did not agree that such a Cox regression analysis could show that the
effect on the primary endpoint was not attributable to the BP lowering effect. Additional
important considerations would be (i) the dose of ramipril (whether a sub-maximal or
maximal dose was used), (ii) whether a random measurement of BP or the full 24-hour
(or peak and trough) BP effects was used, (iii) whether historical constancy in the
treatment effect of ramipril (in the HOPE trial) was maintained in the ONTARGET trial,
and (iv) whether there were differences in use and dose of concomitant medications
(e.g., β-blockers, calcium channel blockers, diuretics, etc.) at baseline and during the
course of the study.
At the pre-NDA meeting on 25-Oct-2007, the limitations of the Cox proportional hazards
model in terms of the inherent difficulties to fully interpret the results were discussed
and mutually acknowledged. No alternative analyses (to the Cox proportional hazards
model) was offered and agreed upon. There was no final agreement as to what would
constitute a “persuasive result” which would remain a review issue.

2.6 Other Relevant Background Information
Not applicable.




                                           15 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


3 Ethics and Good Clinical Practices

3.1 Submission Quality and Integrity
The sponsor submitted that the clinical studies at non-US sites were not conducted
under an IND, but these studies were conducted in accordance with the ethical
principles stated in the Declaration of Helsinki issued by the World Medical Association
as well as applicable local regulations. While maintaining that these studies were
conducted according to ICH good clinical practice (GCP), the sponsor requested a
waiver from the requirements outlined in 21 CFR 312.120 related to foreign clinical
studies not conducted under an IND because the non-IND studies were conducted prior
to the effective date (27-Oct-2008) of the revised 21 CFR 312.120, and it is possible
that not all technical requirements of the new rule – as outlined in 21 CRF 312.120(b)
related to foreign clinical studies not conducted under an IND – may have been met.

3.2 Compliance with Good Clinical Practices
29,019 patients were enrolled, and 25,620 patients were randomized during December
2001 to July 2003 at 732 centers. Funnel plots {hazard ratios (on the log-scale) plotted
against the standard errors} of the four-fold composite primary efficacy endpoint events
in ONTARGET trial were made by site, by country and by region. These funnel plots
(pages 1-9 of Attachment 1 in Sponsor’s response submitted 14-Apr-2009) show that
data by region and by country are evenly distributed in a consistent manner, and that
data from nine sites which enrolled and randomized large number of patients (167 to
298 enrolled / 147 to 271 randomized, per site) are located within the expected range of
variation, with no site showing as an outlier.
Four sites that participated in the ONTARGET trial were identified for GCP inspection
(Table 1). These four sites are essential for approval because in addition to enrolling
and randomizing relatively large number of patients, (i) sites in USA and Slovakia have
relatively low hazard ratios (<1.0), and patients with CHF hospitalization (a “soft”
endpoint) comprised 50% or more of total endpoint events, and (ii) sites in the UK and
Australia have very low hazard ratios (0.326 to 0.369) suggesting large effect size at
these sites.
On 12- and 18-Jun-2009, the Division of Scientific Investigations (DSI) submitted a
preliminary communication regarding the GCP Inspections of these four sites that there
were no major issues with protocol compliance or reporting of AEs.




                                           16 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 1 Sites requested for GCP inspection by DSI
                                                  # subjects     # with total    Hazard Ratio
Center #   Site (Name and Address)                 enrolled/   endpoint / CHF    (Telmisartan
                                                 randomized    hospitalization     /Ramipril)
           Nauman Qureshi, MD
 0256      Athens Internal Medicine
                                                 267 / 234        60 / 30           0.749
           1005 West Market St., Suite 16
           Athens, AL 35611, USA
           Dr. Jon Waites
 0544      Coffs Harbour Cardiovascular
                                                 162 / 139         12 / 2           0.326
           Clinic
           Mackays Road
           Coffs Harbour NSW 2450,
           AUSTRALIA
           Dr. A. Cowie
           BARONET (Bath Area Research
                                                 133 / 114         14 / 2           0.369
 1706      Organisation Network), The Porch
           Surgery
           Beechfield Road, Corsham
           Wiltshire SN13 9DL, UNITED
           KINGDOM
           Andrej Dukat, MD
 1535      2nd Dept. of Internal Medicine,
                                                  109 / 93        22 / 13           0.999
           Faculty Hospital Comenius
           University, Mickiewiczova 13, 913
           69 Bratislava, SR, SLOVAKIA



3.3 Financial Disclosures
The sponsor submitted a list of 21 clinical investigators (18 in US, 2 in Canada, 5 in
Italy, 4 in Germany, 3 in Spain, 2 in the UK and 1 in Sweden) who hold financial
interests requiring disclosure, together with Forms FDA 3455.
A review of the Form FDA 3455 and financial disclosure statement made by each
showed that
(1) the clinical investigators who enrolled large numbers and showed positive results
    were not among the 21 investigators, and
(2) these 21 investigators did not enroll large numbers of patients.
Comment: It is unlikely that the number of patients enrolled at these 21 sites where the
investigators had significant equity interest in this drug product could have contributed
to the overall outcome of the study in a significant manner.




                                               17 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

4 Significant Efficacy/Safety Issues Related to Other Review
  Disciplines

4.1 Chemistry Manufacturing and Controls
There was no new chemistry, manufacturing and controls data submitted for this NDA
supplement.

4.2 Clinical Microbiology
There was no new clinical microbiology study submitted for this NDA supplement.

4.3 Preclinical Pharmacology/Toxicology
There was no new pre-clinical pharmacology/toxicology study submitted for this NDA
supplement.

4.4 Clinical Pharmacology
No new clinical pharmacology, pharmacodynamic or pharmacokinetic studies of
telmisartan was submitted in support of the new indication for this NDA supplement.
The sponsor performed an open-label, randomized, single-dose, 6-sequence, 3-way
cross-over design Phase 1 trial (U08-1096-01) in 84 healthy volunteers, comparing the
80 mg telmisartan / 10 mg ramipril fixed-dose combination with the free combination of
telmisartan and ramipril given as either Micardis® + Altace® (ramipril) capsules (Mic+Alt)
or Micardis® + Delix® (ramipril) tablets (Mic+Del) at a dose of telmisartan 80 mg and
ramipril 10 mg.
This trial showed bioequivalence of an 80 mg telmisartan/ 10 mg ramipril fixed-dose
combination compared with the individual components, telmisartan and ramipril, either
as Micardis® 80 mg plus 10 mg Delix® or as Micardis® 80 mg plus 10 mg Altace®, given
concurrently in healthy volunteers. The trial also showed bioequivalence between the
two ramipril formulations, Delix® and Altace® at a dose of 10 mg on the Cmax and AUCs
of ramiprilat (the active moiety) and the parent compound, ramipril.




                                            18 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


5 Sources of Clinical Data

5.1 Tables of Studies/Clinical Trials
Data in support of the NDA Supplement for the new indication come from the following
pivotal and supportive clinical trials (Table 2):
(i) 	 ONgoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial
       (ONTARGET) trial. This was the pivotal trial.
(ii) 	 Telmisartan Randomized AssessmeNt Study in ACE iNtolerant subjects with 

       cardiovascular Disease (TRANSCEND) trial. This was a supportive trial. 

(iii) The Prevention Regimen For Effectively avoiding Second Strokes (PRoFESS) trial:
       This study was in a population of stroke patients, the data being submitted for
       safety analyses.

Table 2: Tables of pivotal and supportive clinical trials for NDA 20-850
                          Pivotal Trial                 Supportive Trials
                           ONTARGET Trial               TRANSCEND Trial                PRoFESS Trial
Number of centers                  732                           650                         695
Total enrolled                   29,019                         6,665                       20,403
Number of patients       Total = 25,620;              Total = 5,926                Total = 20,332
randomized:              Telmisartan = 8542,          Telmisartan = 2,954,         Telmisartan =10,146;
                         Ramipril = 8576              Placebo = 2,972              Placebo = 10,186
                         Telmi+Ramipril = 8502
Population studied       Patients at high risk of     Patients at high risk of a   Patients who had
                         a major CV event             major CV event who are       suffered an ischemic
                                                      intolerant to ACE-Is         stroke within 120 days
Mean follow up (month)            54                              56                           30
Primary efficacy         CV death, MI, stroke,        CV death, MI, stroke, &      Time to first recurrent
endpoint                 & CHF hospitalization        CHF hospitalization          stroke



5.2 Review Strategy

FDA reviewer’s explanation:
While we try to adhere to the format of the new OND Clinical Review Template, due to
extensive statistical evaluations relevant to clinical decision making, we will present our
consolidated and detailed statistical analyses and discussion under Section 6.1.6
Statistical considerations of Primary and Secondary Efficacy Endpoints(authored by the
statistical reviewer), and present only our summary of methods in Section 6.1.1
Methods, and our summarized review of the primary and secondary efficacy
endpoints, respectively, in Section 6.1.4 Analysis of Primary Endpoint(s), and Section
6.1.5 Analysis of Secondary Endpoints(s).


                                                    19 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

For the primary efficacy endpoint review, our plan was to first evaluate the superiority of
the T/R combination over R for the primary efficacy endpoint in ONTARGET. When we
found that ONTARGET failed to demonstrate superiority of the T/R over R for the
primary efficacy endpoint, we reviewed the renal endpoint only in an exploratory
manner, in keeping with the protocol-specified statistical analysis plan that testing of the
renal primary endpoint in the subgroup of diabetic patients with proteinuria would be
performed only if both superiority of T/R vs R and non-inferiority of T vs. R regarding
the primary CV composite were found.
Next, we planned to review the non-inferiority of T vs R for the primary efficacy endpoint
in ONTARGET. For the purpose of this review, the implied “statistical” understanding of
“non-inferiority” is that:
(i) the test treatment is better than placebo, and
(ii) the confidence interval of the difference between the test treatment and the control
      treatment is within a clinically relevant non-inferiority (NI) margin (i.e., therapeutic
      interchangeability).
Other statistical considerations include the evaluation of constancy assumption and 

adjustments for clinically relevant covariates to investigate whether the changes in 

standard of care over time have altered the effect size and/or altered the disease status 

or CV risk in the patient population studied. 

We evaluated ONTARGET to determine whether T is non-inferior to R using the 

following steps: 


Step 1: Calculation of NI margin(s)

First, we calculated the NI margin based on the historical HOPE trial1,2 data for the 

original 3-fold composite (HOPE) endpoint. We calculated the NI margin for the 4-fold 

composite (ONTARGET) endpoint in HOPE trial as well. 

We also calculated NI margins from a meta-analysis of HOPE and two other ACE-I trials
(EUROPA3 and PEACE4) trials in patients with CV risk factors.
Thus, we have NI margins from HOPE trial with 3-fold and 4-fold composite endpoints,
and from a meta-analysis of the 3-fold composite endpoint in the 3 ACE-I trials.
For the purpose of our NI analyses, we planned to use a conservative 50% discount of
the lower confidence bounds, and also to study how changing the discount would
influence the NI margin.

Step 2: Non-inferiority testing of the primary efficacy endpoint data in ONTARGET
We evaluated ONTARGET for both the 4-fold and 3-fold composite efficacy endpoints
using the FAS, PPS and FAS-10 data sets (Section 5.3.3 Statistical Methods).
For sensitivity analyses, we evaluated these data adjusting for (i) CHF hospitalizations
confirmed by chest X-ray (similar to HOPE protocol), (ii) systolic BP at end of the trial




                                             20 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

and (iii) systolic BP at the visit just before the primary event occurred, and (iv) use of
open-label ACE-I and/or ARBs.

Step 3: Evaluation of constancy assumptions and adjustments for covariates
We planned to investigate the validity of our constancy assumptions by examining
hazard ratios and placebo event rates in multiple trials.
To determine if a clinical covariate which occurred with different frequencies in the
HOPE and ONTARGET trials was influencing the treatment-effect, we calculated
treatment-by-covariate interactions in HOPE, EUROPA and ONTARGET data for the
following clinically important covariates:
• Female gender
• Previous coronary artery disease
• Previous MI
• Previous stroke/TIA
• History of diabetes mellitus
• History of hypertension
• Concomitant use of antiplatelet drugs
• Concomitant use of β-blockers
• Previous coronary revascularization procedures.

Step 4: Review of supportive trials
In addition to determining whether T is non-inferior to R based upon a NI margin that is
acceptable to FDA, we investigated whether the effect of R (interchangeable with T)
was still maintained in the current patient population by evaluating superiority of
telmisartan over placebo in the two large telmisartan-vs-placebo clinical outcome trials:
(i)    T
       	 RANSCEND, which contemporaneously enrolled patients with the same high
       CV risk as in ONTARGET (albeit ACE-I intolerant), and
(ii)   PRoFESS (and/or pooled data from the TRANSCEND/PRoFESS trials).
       	
The overlapping treatment groups in HOPE, ONTARGET, TRANSCEND and PRoFESS
are shown in Table 3. We think that statistical confirmation of interchangeability of effect
of T with R would be shown if T showed superiority over placebo (PBO) in a convincing
manner (i.e., very small p values).


Table 3 Treatment groups in HOPE, ONTARGET, TRANSCEND & PRoFESS trials
                        Placebo Ramipril Telmisartan Telmisartan+Ramipril
       HOPE                 X        X
       ONTARGET                      X            X            X
       TRANSCEND            X                     X
       PRoFESS‡             X                     X
       ‡
        telmisartan vs. placebo comparison in PRoFESS trial




                                             21
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The following is the layout of this combined clinical and statistics review keeping to the
four steps outlined above:
•	 Step I is summarized in Section 6.1.1 Methods,
•	 Steps 2(a) and 2(b) are summarized in Sections 6.1.4         Analysis of Primary
   Endpoint(s)and 6.1.5 Analysis of Secondary Endpoints(s), and
•	 Steps 3(a) and (b) and Step 4, together with details of Step 1 and 2, are all
   presented in Section 6.1.6 Statistical considerations of Primary and Secondary
   Efficacy Endpoints.


5.3 Discussion of Individual Studies/Clinical Trials

5.3.1    Pivotal Clinical Trial
The ONgoing Telmisartan Alone and in combination with Ramipril Global Endpoint Trial
(ONTARGET) is the pivotal study for this NDA. ONTARGET is a multi-center, multi­
national, double-blind, parallel-group, active-controlled, double-dummy trial in which
patients with high risk of CV disease were randomized (1:1:1 ratio) to receive one of the
3 treatments: combination therapy of 80 mg telmisartan and 10 mg ramipril (T/R),
monotherapy with 80 mg telmisartan (T), or monotherapy with 10 mg ramipril (R). The
objective of the ONTARGET trial was to determine if
(i) T/R is superior to ramipril 10 mg R alone and,
(ii) T alone is not inferior to R alone,
in reducing the four-component composite endpoint of CV death, MI, stroke, or
hospitalization for CHF.
The ONTARGET study flow chart is shown in Figure 1.

Figure 1 ONTARGET: study flow chart




        Source: Figure 1.3:1 on page 20 of Summary of Clinical Efficacy.


                                                  22 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The mean follow up was 54 months. The Coordinating Investigator of ONTARGET and
TRANSCEND trials was Prof Salim Yusuf (Canada), and worldwide conduct of the trials
was coordinated by Dr. Koon Teo (Canada), with a National Coordinator or National
Leader in each participating country.
Enrollment criteria: ONTARGET enrolled male or female patients, 55 years of age or
older, and at high risk of CV events indicated by one of the following:
1. Coronary artery disease (CAD) defined as:
      a) Previous MI (>2 days post uncomplicated MI, prior to informed consent), or
      b) Stable or unstable angina (>30 days prior to informed consent) each with
         documented multivessel coronary disease, >50% stenosis in at least 2 major
         coronary arteries on coronary angiography, or a positive stress test (ST
         depression ≥2 mm or a positive nuclear perfusion scintigram), or
      c) Multi-vessel percutaneous transluminal coronary angioplasty (PTCA) (>30
         days prior to informed consent), or
      d) Previous multi-vessel coronary artery bypass graft (CABG) without angina (>4
         years prior to informed consent) or with recurrent angina after surgery.
2. Peripheral arterial disease (PAD) defined as:
      a) Previous limb bypass surgery or percutaneous transluminal angioplasty, or
      b) Previous limb or foot amputation, or
      c) History of intermittent claudication, with an ankle/arm blood pressure (BP)
         ratio ≤0.80 on at least one side, or
      d) Significant peripheral artery stenosis (>50%) documented by angiography or
         non-invasive testing.
3. Previous stroke (included definite or presumed cerebral infarction, intracerebral
   hemorrhage, stroke of uncertain subtype, but not subarachnoid hemorrhage).
4. Transient ischemic attack (TIA) >7 days and <1 year prior to informed consent (TIA
   was defined as acute loss of focal cerebral or monocular function with symptoms
   lasting <24 h, and thought to be due to inadequate cerebral or ocular blood supply
   as a result of arterial thrombosis or embolism).
5. High-risk diabetes (insulin-dependent or non-insulin dependent) with evidence of
   end-organ damage (retinopathy, left ventricular hypertrophy, micro- or macro-
   albuminuria, or any evidence of previous cardiac or vascular disease).
Exclusion criteria: Patients who had CHF, uncontrolled hypertension (BP>160/100
mmHg), significant renal artery disease or hepatic dysfunction at study entry were not
enrolled.
Doses:
•	 Run-in period: R2.5 and T placebo daily (3 days) followed by T40/R2.5 daily (7 days)
   followed by T40/R5 daily (11 to 18 days) (see Figure 2)



                                           23 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

• After randomization: T80/R5 or T80 or R5 daily for 2 weeks (see Figure 3)
• Maintenance period: T80/R10 or T80 or R10 daily over 3.3 – 5.5 years (see Figure 3)
Duration of treatment:
• Run-in period: 3 weeks (+ 1 optional week) (see Figure 2)
• Further up-titration (after randomization): 2 weeks (see Figure 3)
• Maintenance period: planned: 3.5 to 5.5 years

Figure 2 ONTARGET – Run-in Phase flow chart




Figure 3 ONTARGET – Randomization Phase flow chart




Primary endpoints:
•	 4-fold composite endpoint of time to CV death, non-fatal MI, non-fatal stroke, or
   hospitalization for CHF
•	 Composite of doubling of serum creatinine, progression to ESRD (as defined by
   initiation of dialysis, need for renal transplantation, or eGFR <15 mL/min/1.73 m2),
   and all-cause mortality in the subgroup of diabetic nephropathy patients (i.e. diabetic
   patients with macro-albuminuria assessed as a UACR ≥300 mg/g Crea at baseline)
Secondary endpoints: 

An additional objective, using an exploratory analysis, was to investigate whether T/R is 

more effective than R concerning these endpoints: 

•	 3-fold composite endpoint of CV death, non-fatal MI, and non-fatal stroke (HOPE
   trial endpoint)


                                           24 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

•   The individual components of the primary endpoints
•   Occurrence of nephropathy analyzed based on the following 6 subcategories:
    o	 Doubling of serum creatinine
    o	 Progression to ESRD (i.e. initiation of dialysis, need for renal transplantation, or
       eGFR <15 mL/min/1.73 m2)
    o	 New microalbuminuria (UACR ≥30 mg/g Crea in patients with a UACR <30 mg/g
       Crea at baseline)
    o	 New macroalbuminuria (UACR ≥300 mg/g Crea in patients with a UACR <300
       mg/g Crea at baseline)
    o	 Combined endpoint of doubling of serum creatinine, progression to ESRD, new
       microalbuminuria, or new macroalbuminuria
    o	 Normalization from micro- or macroalbuminuria to normoalbuminuria (UACR <30
       mg/g Crea in patients with a baseline UACR ≥30 mg/g Crea)
•   Newly diagnosed CHF
•   CV revascularization procedures
•   Newly diagnosed diabetes
•   Cognitive impairment and cognitive decline
•   New onset of atrial fibrillation
Other endpoints:
• All-cause mortality (non-CV mortality + CV mortality)
• Non-CV death
• Angina (unstable, new, and worsening)
• Transient ischemic attack (TIA)
• Malignancy (fatal and non-fatal)
• Laser therapy for diabetic retinopathy
Additionally, changes from baseline in BP, ankle/arm BP ratio, Mini Mental State
Examination, eGFR, serum creatinine, UACR, urinary potassium creatinine ratio, and
urinary sodium-creatinine ratio were determined.
A total of 29,019 patients were enrolled by 732 centers in 40 countries worldwide. After
a 3-week single-blind run-in period, 25,620 patients were randomized to one of the 3
treatments: T/R (8,502 patients), T (8,542 patients), or R (8,576 patients) for a treatment
period of 3.5 - 5.5 years; the actual median observation time was 4.6 to 4.7 years. All
patients received concomitant medications according to their medical needs.
The most frequent reasons for non-randomization were the patients' request to stop
treatment (5.5%), insufficient compliance with study medication (4.6%), persistent
symptomatic hypotension (1.7%), and cough (1.6%).
A large number of patients in ONTARGET received open label ACE-Is (5.2% of patients
randomized to telmisartan and 5.7% of patients randomized to ramipril) and open label
ARBs (5.0% of patients randomized to telmisartan and 5.5% of patients randomized to


                                            25 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

ramipril). Non-inferiority analyses are also performed excluding these patients for the
composite 4-fold primary efficacy endpoint.

5.3.2   Supportive clinical trials
The following trials are submitted as supportive trials to provide evidence based on
comparison to telmisartan to placebo.
TRANSCEND trial
The TRANSCEND (Telmisartan Randomized AssessmeNt Study in ACE iNtolerant
subjects with cardiovascular Disease) trial was a large, simple, randomized, double-
blind, multicenter, international trial to compare the effects of telmisartan with placebo
on outcomes in patients at high risk for cardiovascular events and intolerant of ACE-I.
The study design, enrollment criteria, methodology, efficacy and safety endpoints, and
statistical methods were the same as for the ONTARGET trial, with two exceptions:
(i) only patients who were ACE-I intolerant were enrolled, and
(ii) patients with clinically significant proteinuria were excluded (because of the ethical
     consideration that if randomized to placebo, they would be without treatment with a
     RAAS-blocking agent).
The total study duration was 4 years 10 months, and the mean duration of follow-up
was 56 months. The trial was conducted at 650 centers world wide. The Coordinating
Investigator was Prof Salim Yusuf (Canada).
The TRANSCEND trial enrolled 6,665 patients at 650 centers worldwide into the run-in
period of 3 to 4 weeks (starting with placebo (PBO) for 1 week, continuing with
telmisartan 80 mg (T) for 2 to 3 weeks), and randomized (1:1 ratio) a total of 5,926
patients to T (2,954 patients) or PBO (2,972 patients) who were followed up to 1,766
days (T) to 1,772 days (PBO). All patients received concomitant medication according
to their medical needs. The most frequent reasons for non-randomization were the
patients' request to stop treatment (6.1%), insufficient compliance with study medication
(5.1%), and the occurrence of AEs (2.6%).

PRoFESS trial
The PRoFESS (Prevention Regimen For Effectively avoiding Second Strokes) trial was
prospective, randomized, multi-national, double-blind, double-dummy, active- and
placebo-controlled, parallel-group, 2x2 factorial design study of Aggrenox® (fixed-dose
combination product aspirin plus extended-release dipyridamole (Aggrenox®, ASA+ER­
DP) vs. clopidogrel, with and without Micardis® to compare
(i) the efficacy and safety of Aggrenox® with that of clopidogrel, and
(ii) the efficacy and safety of telmisartan with that of placebo,
in the prevention of recurrent stroke in patients who had had an ischemic stroke within
120 days of study entry.



                                            26 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 4 2x2 factorial design of PRoFESS trial




The trial design assumed that the effects of antiplatelet treatments and telmisartan were
independent of each other (treatment-by-treatment interaction was absent by statistical
analysis). The primary endpoint for both comparisons was the time to recurrent stroke.
For the telmisartan comparison, secondary endpoints included the 4-fold composite of
recurrent stroke, MI, death due to vascular causes, or new or worsening CHF (the
primary endpoint in ONTARGET and TRANSCEND), and a 3-fold composite of stroke,
MI, or death due to vascular causes (HOPE endpoint, defined post-hoc).
The primary endpoint of PRoFESS trial was the time to first recurrent stroke.
The secondary endpoints were:
(a) For the telmisartan vs. placebo comparison:
    (i) Composite outcome defined as time to the first of: recurrent stroke, MI, new or
        worsening CHF, or CV death, and
    (ii) Time to new onset of diabetes mellitus.
(b) 	For the antiplatelet comparison: Composite outcome defined as time to the first
     occurrence of recurrent stroke, MI, or death due to vascular causes.
The primary efficacy outcome event of strokes as well the outcome events of MI, major
hemorrhagic events, new or worsening CHF, retinal arterial occlusion, intraocular
hemorrhage and causes of deaths in PRoFESS trial were centrally adjudicated by an
independent Adjudication and Assessment Committee (AAC). The statistical analyses
of outcome events were based on the final assessments of the AAC.
The 3-fold composite point of MI, stroke and CV death (the HOPE trial endpoint) was
analyzed post hoc to put the PRoFESS trial results within the context of those of HOPE,
ONTARGET and TRANSCEND trials where this endpoint was pre-specified.
The enrolment period was 2 years 10 months, total study duration was 4 years 5
months, but the mean duration of follow-up was only 30 months. The trial was
conducted at 695 centers in 35 countries world wide. The 3 Principal Investigators of the
PRoFESS trial were Prof Hans-Christoph Diener (Germany), Prof Ralph L. Sacco (USA)
and Prof Salim Yusuf (Canada).
A total of 20,403 patients were enrolled in PRoFESS. They were slightly younger (≥50
years of age), had suffered a previous ischemic stroke within 120 days of enrolment,



                                           27 

Clinical Review 

Khin Maung U, M.D. & Jialu Zhang, Ph.D. 

NDA 20-850 Efficacy Supplement SE1-025 

Micardis® (telmisartan) tablets 


and were neurologically and clinically stable at enrollment. For patients aged 50 – 54
years and/or patients whose qualifying stroke had occurred 91 to 120 days prior to
admission, at least 2 of the following additional risk factors were required: diabetes
mellitus, hypertension (SBP ≥140 mmHg, or DBP ≥90 mmHg), smoker at the time of
qualifying stroke, obesity (BMI ≥30 kg/m2), previous vascular disease (stroke prior to the
qualifying stroke, MI, or peripheral artery disease), end-organ damage (retinopathy, left
ventricular hypertrophy, or micro-albuminuria), or lipidemia.
20,332 patients were randomized using a 2 x 2 factorial design to platelet therapy with
either a fixed-dose combination of ASA plus ER-DP or clopidogrel, and at the same time
to either T or PBO. Randomization was stratified by baseline usage of ACE-Is.
Although the original treatment regimen for patients randomized to the clopidogrel arm
consisted of a combination of clopidogrel (75 mg QD) and aspirin (75 mg QD), after
about 1,920 patients were enrolled, the aspirin component was removed {based on the
findings of the MATCH (Management of ATherothombosis with Clopidogrel in High-risk
patients with recent transient ischemic attack or ischemic stroke) trial5, which showed an
increased incidence of life-threatening and major bleeding events in clopidogrel plus
aspirin compared to clopidogrel alone, with no significant therapeutic benefit of adding
aspirin to clopidogrel in the prevention of stroke, MI, CV death or rehospitalization due
to an ischemic event.}

5.3.3    Statistical Methods
The following is a summary of the sponsor’s statistical plan and methods.
The primary outcome events and deaths in ONTARGET and TRANSCEND trials were 

centrally adjudicated by the independent Events Adjudication Committee (EAC). The 

statistical analyses were based on these adjudicated outcome events. Doubling of 

serum creatinine and end-stage renal disease (ESRD) were not centrally adjudicated. 

For the confirmatory evidence of efficacy, the intent-to-treat principle was followed in all
3 trials. To determine consistency of the results and the robustness of statistical
analysis, evaluation of the primary efficacy endpoint parameters in ONTARGET and
TRANSCEND were performed using the following analysis sets:
(i) 	 The run-in set (RIS) which included all patients who were enrolled into the run-in
      period irrespective of whether they were randomized or not; (the RIS was used for
      the description of the run-in period of ONTARGET and TRANSCEND trials for safety
      analyses.)
(ii) The full analysis set (FAS) which included all patients who were randomized
      (intent-to-treat principle),
(iii) The per-protocol set (PPS) which included all patients of the full analysis set (FAS)
      who were taking study medication according to the trial protocol for the majority
      (≥50%) of their time in the study,
(iv) The set of patients in FAS who stayed on the full 10-mg dose of ramipril or ramipril



                                             28 

Clinical Review 

Khin Maung U, M.D. & Jialu Zhang, Ph.D. 

NDA 20-850 Efficacy Supplement SE1-025 

Micardis® (telmisartan) tablets 


    placebo (FAS-10). These patients were censored at the date of visit they started to
    reduce their ramipril (in T/R and R group) or ramipril placebo (in T group) dose to 5
    mg or discontinued ramipril or ramipril placebo, and for sensitivity analyses.
ONTARGET primary efficacy data were analyzed based on these data sets. The FAS

was used for (i) non-inferiority comparison of T vs. R, (ii) superiority comparison of T/R 

vs R for the primary efficacy endpoint data, and (iii) for safety analyses. The PPS and 

FAS-10 data were used for sensitivity analysis of the FAS analysis results. 

For PRoFESS, the sponsor used the following:
(i) 	 the randomized set (RAN) which consisted of patients who signed informed consent
       and were randomized (similar to FAS in ONTARGET and TRANSCEND),
(ii) 	 PPS as in ONTARGET and TRANSCEND, subdivided into PPS-A and PPS-B (A=
       antiplatelet comparison and B= telmisartan comparison),
(iii) Treated sets (TS-A and TS-B) which comprised patients who were dispensed and
       documented to have taken at least one dose of telmisartan or placebo (for safety
       analysis).
Table 5 summarizes the endpoint analyses performed, the analysis data set used and 

the statistical test applied per sponsor’s statistical analyses. 

For sensitivity analyses of non-inferiority comparisons, the following data sets were used:
ONTARGET
(i) 	 excluding patients who took an open-label ACE-I in the T group or an open-label
       ARB in the R group [NoARB for R+ NoACE-I for T (ONTARGET)]
(ii) 	 excluding patients who took an open-label ACE-I from all treatment groups
       [NoACE-I for all treatments (ONTARGET)]
(iii)	 excluding patients who took an open-label ARB from all treatment groups [NoARB
       for all treatments (ONTARGET)]
(iv) 	 excluding patients who took an open-label ARB or ACE-I from all treatment groups
       [NoARB+noACE-I for all treatments (ONTARGET)]
(v) 	 excluding patients who took an open-label ACE-I from the T group [NoACE-I for T
       (ONTARGET)]
(vi) 	 excluding patients who took an open-label ARB from the R group [NoARB for R
       (ONTARGET)]
 TRANSCEND and PRoFESS pooled population
(i) 	 excluding patients who took an ACE-I at baseline from both treatment groups
       [Total/NoACE-I at baseline]
(ii) 	 excluding patients who took an ACE-I or an ARB from both treatment groups
       [Total/NoACE-I + NoARB]
(iii)  including all randomized patients [Total]
       	



                                             29 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 5 Statistical analyses and data sets used
                                                         Analysis set            Model, Test
                                              Primary endpoint
  Non-inferiority: T vs R                                FAS (ONTARGET)        Cox regression
  Superiority:     T vs PBO                              FAS (Total/NoACE-I)   Cox regression
                   T vs R                                FAS (ONTARGET)        Cox regression
                                      Key Secondary (HOPE) endpoint
    Non-inferiority: T vs R                              FAS (ONTARGET)        Cox regression
    Superiority:     T vs PBO                            FAS (Total/NoACE-I)   Cox regression
                                             All-cause mortality
    Non-inferiority: T vs R                              FAS (ONTARGET)        Cox regression
    Superiority:     T vs PBO                            FAS (Total/NoACE-I)   Cox regression
          Sensitivity analyses for the primary endpoint and key secondary (HOPE) endpoint
Per-protocol analysis                                    PPS (ONTARGET)        Cox regression
Persistence of effect after correction for SBP           FAS (ONTARGET)        Cox regression,
                                                         FAS (Total/NoACE-I)   time-dependent SBP
Excluding patients in ONTARGET who took open             NoARB for R+NoACE-I   Cox regression
label ACE-I in T group or open-label ARB in R group      for T (ONTARGET)
Excluding patients who took an open-label ACE-I from NoACE-I for all           Cox regression
all treatment groups in ONTARGET                         treatments (ONTARGET)
Excluding patients who took an open-label ARB from       NoARB for all         Cox regression
all treatment groups in ONTARGET                         treatments (ONTARGET)
Excluding patients who took an open-label ARB or         NoARB+noACE-I for all Cox regression
ACE-I from all treatment groups in ONTARGET              treatments (ONTARGET)
Excluding patients who took an open-label ACE-I from NoACE-I for T             Cox regression
the T group in ONTARGET                                  (ONTARGET)
Excluding patients who took an open-label ARB from       NoARB for R           Cox regression
the R group in ONTARGET                                  (ONTARGET)
Excluding patients of the pooled patient population of   Total/NoACE-I at      Cox regression
TRANSCEND and PRoFESS who took an ACE-I at               baseline
baseline from both treatment groups
Excluding patients of the pooled patient population of   Total/NoACE-I + NoARB Cox regression
TRANSCEND and PRoFESS who took an ACE-I or
an ARB from both treatment groups
Including all patients randomized in TRANSCEND           Total                 Cox regression
and PRoFESS
Comparison with R 10 mg                                  FAS-10 (ONTARGET)     Cox regression
Measurements over time
Sitting SBP, DBP, change from                            FAS (ONTARGET) FAS    Descriptive
baseline and BP categories                               (Total/NoACE-I)

For supporting evidence, the pooled analysis of patients in TRANSCEND and
PRoFESS was based on FAS (Total/NoACE-I) which included all randomized patients
who did not take ACE-I at any time during the trial. 16,877 patients randomized in
TRANSCEND (5,866 patients) and PRoFESS (11,011 patients) did not take an ACE-I
during the trials. Thus, the FAS (Total/NoACE-I) comprised 16,877 patients; 8,587 of
these patients were randomized to T, and 8,290 patients were randomized to PBO.
All analyses of secondary and other endpoints, and subgroup analyses, were based on



                                              30 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

the FAS (ONTARGET) and FAS (Total/ACE-I). No per-protocol analysis was performed
for the pooled patient population who did not take ACE-I during the TRANSCEND and
PRoFESS trials.
The number of patients randomized and included in the above sets of patients are
shown in Table 6. We noted that the number of patients in the datasets of 'PPS without
ACE-I,' 'PPS without ARB' and 'PPS without ACE-I and ARB,' appeared to be more
than the number of patients in PPS from which they were taken.
The reason given by the sponsor was that in the PPS and FAS-10, ‘… patients who
deviated from the protocol (mainly due to non-adherence to study medication and
addition of non-study medication) were censored at the time they started to deviate,’
whereas for sensitivity analyses which ‘… excluded patients taking open-label ACE-Is
and/or ARBs,’ these groups were ‘…defined post-hoc as a result of the unexpected
findings in the combination arm…’ and ‘… were based on the set of all randomized
patients (FAS) without any premature censoring due to protocol deviations.’

Table 6 Number of patients included in sensitivity analyses (ONTARGET)
                                                                     ONTARGET
                                                       T/R: n (%)      T: n (%)       R: n (%)
Randomized (FAS)                                      8502 (100.0)   8542 (100.0)   8576 (100.0)
Per-protocol set (PPS)                                 6531 (76.8)    7142 (83.6)    6988 (81.5)
FAS-10                                                 5802 (68.2)    6585 (77.1)    6484 (75.6)
No ARB for R + No ACE-I for T (ONTARGET)              8502 (100.0)    7502 (87.8)    8107 (94.5)
PPS without ACE-I                                      7414 (87.2)    7502 (87.8)    7522 (87.7)
PPS without ARB                                        8058 (94.8)    8118 (95.0)    8107 (94.5)
PPS without ACE-I and ARB                              7154 (84.1)    7263 (85.0)    7240 (84.4)
Source: Table 3.1.1:2 on page 64 of Summary of Clinical Efficacy

The primary and secondary outcome events were tested by time-to-event analysis
(Kaplan Meier) and hazard ratios based on the intention-to-treat principle and using the
Cox proportional hazards model (Cox regression).
 Χ2-test was used to compare for cognitive impairment and cognitive decline.
Other statistical analyses include:
•	 Exploratory analysis for the comparison of T/R vs. T
•	 Sensitivity analyses using Cox regression including time-dependent covariates and
   Χ2-test
•	 Analysis of covariance (ANCOVA) and
•	 Descriptive statistics




                                                   31 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

6 Review of Efficacy
Efficacy Summary
Telmisartan is currently approved in the United States (since 10-Nov-1998) for the
treatment of hypertension. The submission consists of one pivotal clinical trial
(ONTARGET) and two supportive trials (TRANSCEND and PRoFESS). This
supplement seeks the indication of use of 80 mg telmisartan once a day to reduce the
risk of MI, stroke, CV deaths or hospitalization for CHF in patients 55 year or older at
high risk of developing major CV events.
ONTARGET was a large multicenter, double-blind, parallel group, active-controlled,
double-dummy trial which randomized (1:1:1 ratio) 25,620 patients to one of three
treatment groups: telmisartan (T) 80 mg (n=8,542), ramipril (R) 10 mg (n=8,576), or
combination (T/R) of telmisartan 80 mg and ramipril 10 mg (n=8502). The objective of
ONTARGET is to show that the T/R combination is superior to R alone and/or that T is
non-inferior to R to reduce the 4-fold composite primary efficacy endpoint of MI, stroke,
CV deaths or hospitalization for CHF in patients ≥ 55 years old at high risk of developing
major CV events. The mean follow up was 54 months. Baseline therapy included anti-
platelet drugs (95% of patients), lipid lowering drugs (65%), β-blockers (57%), calcium
channel blockers (34%), nitrates (29%) and diuretics (28%).
TRANSCEND, similar to ONTARGET in design, methods and endpoints, was
conducted contemporaneously with ONTARGET, the only differences being that the
patients were ACE-I intolerant, and patients with clinically significant proteinuria were
excluded. TRANSCEND randomized (1:1 ratio) 5,926 patients to T (n=2,954) or placebo
(PBO) (n=2,972). The objective was to compare effect of T vs PBO on the 4-fold
composite endpoint in high CV risk patients intolerant to ACE-Is. The mean duration of
follow up was 56 months. All patients received concomitant medications as needed.
PRoFESS, was a large, multicenter, double-blind, parallel-group, active- and placebo-
controlled, double-dummy, factorial design trial of Aggrenox® (fixed-dose combination
product of aspirin plus extended-release dipyridamole, ASA+ER-DP) vs. clopidogrel,
with and without telmisartan to compare Aggrenox® to clopidogrel, and T to PBO in the
prevention of recurrent stroke in patients who had had an ischemic stroke within 120
days of study entry. PRoFESS randomized 20,332 patients (stratified by baseline usage
of ACE-Is) using a 2x2 factorial design to platelet therapy with clopidogrel or a fixed-
dose combination of ASA+ER-DP, and at the same time to T or PBO. The primary
endpoint was time to first recurrent stroke. For the T vs PBO comparison, the secondary
endpoints included the 4-fold composite primary efficacy endpoint of MI, stroke, CV
deaths or CHF hospitalizations. The mean duration of follow up was 30 months.
ONTARGET fails to demonstrate superiority of the T/R combination over R for the
primary efficacy endpoint (HR 0.99, 95% CI 0.92, 1.07, p=0.8462).
ONTARGET is evaluated to determine if T is non-inferior to R using the following steps:




                                            32 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Step I: Calculation of the NI margin
The sponsor claimed that telmisartan monotherapy is non-inferior to ramipril based on
their NI margin of 1.13. This margin of 1.13 seems to be too liberal.
The NI margin calculated by the reviewers from the HOPE trial data for the 3-fold
composite (HOPE) endpoint (using a 50% discount of the lower confidence bound of
1.17) is 1.08. Other NI margins calculated from a meta-analysis of the triad of placebo-
controlled ACE-I trials (HOPE, EUROPA and PEACE) are 1.04, using the fixed effects
model and 1.03, using the random effects model. Depending on the confidence interval
(CI) used (90% to 99%), the degree of discounting, and whether only data from HOPE
trial alone or integrated data from HOPE, EUROPA and PEACE trials are used in the
reviewers’ calculation, the NI margin varies from 1.03 to 1.09.
The NI margins generated by the reviewers are more conservative than the NI margin of
1.13 used by the sponsor. The sponsor’s NI margin is based on (i) 20% two-sided
confidence interval, (ii) single HOPE trial so that between-trial variability cannot be
estimated, and (iii) the HOPE trial population which received the prevailing treatment 10
years ago different from current clinical practice.

Step 2: Non-inferiority testing of primary efficacy data in ONTARGET 

Using the NI margin of 1.08 generated from the HOPE trial, the primary endpoint of 

ONTARGET is evaluated for the Full Analysis Set (FAS), the Per-Protocol Set (PPS) 

and the FAS-10 (patients who took the full 10 mg dose of ramipril) data sets. 

Non-inferiority of T to R is not found for any of the data sets using the reviewers’ NI
margin of 1.08 (including the sensitivity analyses), but the upper confidence bounds are
within the relatively wider NI margin of 1.13 used by the sponsor.
Sensitivity analyses were performed adjusting for systolic blood pressure (SBP) at
baseline and SBP change from baseline or SBP change before an event, and excluding
patients who took an open-label ACE-I in T group (1,040 patients) or an open-label ARB
in the R group (469 patients) or both during the time they belonged to the PPS (i.e.,
before censoring). These analyses do not change the initial evaluation except that
statistically significant (p<0.0001) differences between T and R are found for analyses
which exclude (i) patients who took open label ACE-I in T group, and (ii) patients who
took open label ACE-I in T group or open-label ARB in R group. No clinically significant
conclusions can be made on these findings which are the result of post hoc multiple
statistical testing.

Step 3: Evaluation of constancy assumptions and adjustments for covariates
The annualized event rates in the placebo group are lower in TRANSCEND compared
to HOPE for the 4-fold composite endpoint (3.72% in TRANSCEND vs 5.09% in HOPE)
and for the 3-fold composite endpoint (3.34% in TRANSCEND vs 4.53% in HOPE).
The annualized event rates in the ramipril group are lower in ONTARGET compared to
HOPE for the 4-fold composite endpoint (3.67% in ONTARGET vs 3.95% in HOPE) and



                                           33 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

for the 3-fold composite endpoint (3.23% in ONTARGET vs 3.51% in HOPE).
These findings suggest that the constancy assumption which the non-inferiority test
relies upon may no longer hold.
Regarding treatment-by-covariate interactions, most of the interaction terms of
treatment and clinical covariates are not significant at all, and do not seem to affect the
treatment effect in EUROPA, HOPE or ONTARGET trials with the exception of use of
anti-platelets in HOPE trial and use of β-blockers in ONTARGET.
After adjusting for covariates (with the exception of the covariate anti-platelet), the NI
margin, based on HOPE trial alone, varies from 1.02 to 1.10. Including the covariate
anti-platelet into the model seems to favor ramipril and increases the NI margin to 1.15.
The EUROPA trial shows a smaller treatment effect when adjusting different covariates.
The NI margin based on EUROPA trial alone appears to vary from 1.02 to 1.06
depending on the covariate included in the model. In most cases, the NI margins
adjusted for covariates are below 1.10.
In ONTARGET, the upper limit of the confidence interval of the point estimate varies
from 1.07 to 1.20 if adjusted for each covariate.
Thus, in almost all analyses with or without covariate adjustment, the NI margin derived
by discounting 50% on the worst limit of 95% confidence interval for the effect of ramipril
over placebo cannot be ruled out by the non-inferiority analysis in ONTARGET.

Step 4: Evaluation of supportive trials
The supportive trials, TRANSCEND and PRoFESS, do not show superiority of T over
PBO for the protocol-specified primary endpoints.
In TRANSCEND, T did not win over PBO for the 4-fold composite primary efficacy
endpoint. TRANSCEND shows a nominal significance for the 3-fold composite (HOPE)
endpoint (HR 0.87; 95% CI 0.76, 1.00; p = 0.0483).
In PRoFESS, T did not win over PBO for the primary efficacy endpoint (time to the first
recurrent stroke) or for the secondary 4-fold composite (ONTARGET) endpoint (HR
0.94; 95% CI 0.87, 1.01; p = 0.1073) or post-hoc 3-fold composite (HOPE) endpoint (HR
0.94; 95% CI 0.87, 1.02; p = 0.1260).
The sponsor’s post-hoc analysis of pooled PRoFESS and TRANSCEND data showed a
nominally significant reduction in the relative risk of 4-fold composite (ONTARGET)
endpoint (HR 0.90, CI 0.83, 0.98, p=0.0107) and the 3-fold composite (HOPE) endpoint
(HR 0.88, CI 0.81, 0.96, p=0.0029). The post hoc analyses did not take multiplicity
adjustment into consideration. Also, the analyses did not account for all of the possible
ways to win.
The failure of T to win over PBO in TRANSCEND trial, which contemporaneously
enrolled patients with similar CV risk as in ONTARGET, does not support
interchangeability of T for R, and further suggests that changes in the standard of care



                                             34 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

have altered the clinical event rates in this patient population so that the constancy
assumptions do not hold any more.
Will defer to comment until after the AC meeting.

6.1 Indication
Reduction of risk of myocardial infarction, stroke, death from cardiovascular causes or
hospitalization for congestive heart failure in patients 55 year or older at high risk of
developing major cardiovascular events

6.1.1   Methods
FDA reviewers’ statistical methods to evaluate the data in ONTARGET
{Note: We will present only our summary of methods here. We will present all of
our consolidated, detailed statistical analyses and discussion in Section 6.1.6
Statistical considerations of Primary and Secondary Efficacy Endpoints.}
Following a review of the demographics and patient disposition in the pivotal and
supportive trials, we first reviewed the primary efficacy endpoint results. When
superiority of T/R over R was not found, we proceeded with non-inferiority analysis of T
vs R for the primary efficacy endpoint following the Steps described earlier in this review
(please see Section 5.2       Review Strategy).
For Step I, we calculated the NI margin using the HOPE trial data for both the 3- fold
(original) composite endpoint and 4-fold composite endpoint. Detailed statistical
determination of the NI margin, the confidence intervals used, and the degree of
discounting necessary are presented in Section 6.1.6 Statistical considerations of
Primary and Secondary Efficacy Endpoints. The following summarizes the reviewers’
rationale in the context of clinical trials of ACE-Is for their calculation of the NI margin.
To determine the NI margin for the 4-fold composite (ONTARGET) endpoint based on
the HOPE trial data, we retrieved the dates of CHF hospitalizations in the HOPE trial,
and obtain the hazard ratio (and the confidence interval) in the HOPE data for the 4-fold
composite primary endpoint (Table 7) similar to that in ONTARGET.
We used the NI margin(s) thus generated to compare the data for the 4-fold primary and
3-fold secondary composite endpoints in ONTARGET to the corresponding 4-fold and
3-fold composite endpoints in HOPE.




                                              35 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 7 Composite 4-fold and 3-fold primary endpoint data in HOPE
                           HOPE (ramipril vs placebo)
                                  R           PBO         PBO vs R: HR (95% CI)
Randomized, n (%)                         4,645 (100)              4,652 (100)                          ---
4-fold Primary endpoint, n (%)             726 (15.6)               919 (19.8)                 0.775 (0.703, 0.854)
3-fold Primary endpoint, n (%)             651 (14.0)               826 (17.8)                   0.78 (0.70, 0.86)
Source: FDA analyses of HOPE data in NDA


A 50% discount of the lower confidence bound in the HOPE trial gives a NI margin of
1.08 (using the fixed effects model), and a 75% discount gives a NI margin of 1.04. The
NI margin changes depending upon the amount discounted. We believe that a clinical
decision is needed to determine how much to discount and obtain a clinically relevant NI
margin around which we can evaluate the non-inferiority of T vs. R in ONTARGET. For
our review, we will use 50% discount on the lower limit of the confidence interval which
gives a conservative NI margin of 1.08 and preserves at least 50% of the effect size.
We calculated other NI margins from a meta-analysis of HOPE, EUROPA and PEACE
(Table 8) trials which were placebo-controlled trials of ACE-Is (ramipril, perindopril and
trandolapril, respectively) conducted in patients with similar CV risk.

Table 8 Baseline characteristics and primary efficacy endpoint results in HOPE,
EUROPA and PEACE trials of ACE inhibitors
                                    HOPE (2000)                     EUROPA (2003)              PEACE (2004)
Study drug                           Ramipril                         Perindopril                   Trandolapril
Population studied            History of CAD, stroke,         History CAD (documented MI       History of CAD (documented
                              PVD or diabetes plus at         >3 months before), per­          MI >3 months before), CABG
                              least one other CV risk         cutaneous or surg coronary       or PTCA ≥ 3 months before,
                              factor (hypertension,           revascularization >6 mths        ≥50% obstruction of luminal
                              elevated total cholesterol,     before or angiographic           diameter of at least one native
                              Low HDL cholesterol             evidence of 70% narrowing of     vessel (coronary angiography),
                              levels, cigarette smoking or    ≥1 major coronary arteries) or   and LVEF >40% on contrast or
                              documented                      chest pain with positive ECG,    radionuclide ventriculography/
                              microalbuminuria)               echo or nuclear stress test.     echocardiography, normal LV
                                                                                               wall motion echocardiography
N (ACEI/ARB : Placebo)         9,297 (4,645 : 4,652)             12,218 (6,110 : 6,108)          9,290 (4,158 : 4,132)
Composite primary efficacy    MI, stroke or death from        CV death, non-fatal MI and       CV death or non-fatal MI
endpoint                      CV causes                       cardiac arrest with
                                                              successful resuscitation
Placebo group: # of events              826                               603                             929
                 Event rate            17.8%                             9.9%                            22.5%
ACEI/ARB:      # of events              651                               488                             909
                 Event rate            14.0%                             8.0%                            21.9%
HR (95% CI); p value          0.78 (0.70, 0.86);<0.001        0.80 (0.71, 0.91);p=0.0003        0.96 (0.88, 1.04);p=0.43
HR (95% CI); p value (for     0.78 (0.70, 0.86);<0.001         0.83 (0.74, 0.93);p=0.001        0.93 (0.81, 1.07);p=0.32
HOPE endpoint)
Source: FDA analyses of HOPE and EUROPA data in NDAs and PEACE data from literature.

The sponsor argued that the Breslow-Day test showed lack of homogeneity when data
from all 3 trials are used, but shows homogeneity when data from HOPE and EUROPA



                                                             36
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

only are used. The NI margin calculated from the combined data in the 3 ACE-I trials is
1.06 using the fixed effects model and 1.08 when using only HOPE and EUROPA trials.
For Step 2, we evaluated the full analysis set (FAS), the “per-protocol set” (PPS) and
the set of patients who took the full 10 mg/day dose of ramipril (FAS-10) in ONTARGET
for the 4-fold and 3-fold composite primary efficacy endpoints. (Please see Section
6.1.4 Analysis of Primary Endpoint(s)).
(Note that the NI margins calculated by the reviewers do not agree with the NI margin of
1.13 that was proposed by the sponsor.)
For Step 3, we found that there are differences in patient characteristics and in the
concomitant medications taken by patients enrolled in HOPE vs. ONTARGET and
TRANSCEND trials (Table 9) which are clinically relevant covariates that may influence
the primary efficacy events. So we performed exploratory analyses by adjusting those
covariates in each trial.

Table 9 Baseline characteristics in HOPE, ONTARGET and TRANSCEND trials
                                   HOPE (2000)                     ONTARGET (2008)               TRANSCEND (2008)
Study drug                           Ramipril                     Telmisartan vs Ramipril           Telmisartan
Population studied           History of CAD, stroke, PVD      Patients at high risk of CV        Same patients as in
                             or diabetes plus at least one    events (Previous MI >2 d,          ONTARGET who are
                             other CV risk factor             angina with multivessel CAD,       ACE-I intolerant (by
                             (hypertension, elevated          stenosis >50% in 2 major artery,   history, or – following
                             total cholesterol, Low HDL       positive stress test), PAD         ramipril administration –
                             cholesterol levels, cigarette    (previous bypass/ angioplasty,     symptoms of cough,
                             smoking or documented            limb/foot amputation, peripheral   hypotension or renal
                             microalbuminuria)                artery stenosis >50%), prev.       dysfunction likely to be
                                                              stroke/TIA, high risk diabetes+    associated with ACE-I
                                                              organ damage.                      intolerance)
N (ACEI/ARB : Placebo)        9,297 (4,645 : 4,652)             17,118 (T 8,54:R 8,576)        5,926 (2,954 : 2,972)
Average Age (years)                      66                               66                            67
Female gender                           27%                              27%                           43%
Known CAD                               80%                              66%                           67%
Previous MI                             53%                              49%                           46%
Stroke/TIA                              11%                              21%                           22%
Diabetes mellitus                       38%                              37%                           36%
Hypertension                            47%                              69%                           77%
Hypercholesterolemia                    66%                               ---                           ---
Antiplatelet drugs                      76%                              76%*                           ---
β-blockers                              39%                              57%                           70%
Lipid lowering drugs                    29%                              65%                           62%
Coronary revascularization              40%                              51%                           45%
Composite primary            MI, stroke or death from         CV death, non-fatal MI, stroke or CHF hospitalization
efficacy endpoint            CV causes
Source: FDA reviewers’ analyses of HOPE, ONTARGET and TRANSCEND data in NDAs. * antiplatelet
use as aspirin only


We investigated the validity of constancy assumption by calculating the annualized
event rates in the placebo groups in HOPE and TRANSCEND trials and the annualized
rates in the ramipril groups in HOPE and ONTARGET trials to determine the effect of
changes in the standard care over the past years (e.g., greater use of medications such


                                                             37
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

as statins, β-blockers and diuretics which are known to reduce CV risk) on the effect
size of the primary efficacy clinical outcome parameters in the placebo group and the
comparator (ramipril) group in patients with similar CV risk.
We also determined treatment-by-covariate interactions in HOPE and ONTARGET data
for the following clinically important covariates at baseline:
• Female gender
• Previous coronary artery disease
• Previous MI
• Previous stroke/TIA
• History of diabetes mellitus
• History of hypertension
• Concomitant use of antiplatelet drugs
• Concomitant use of β-blockers
• Previous coronary revascularization procedures.
Lastly, in Step 4, we reviewed superiority test of the combination of telmisartan and
ramipril versus ramipril, and also the two supportive trials (TRANSCEND and
PRoFESS) to determine whether telmisartan (T) is significantly superior to placebo
(PBO) in the patient population receiving the current standard care {i.e. to determine the
interchangeability of T (in ONTARGET) with R (in HOPE)}.
The above Steps I, 2, 3 and 4 are presented in detail in Section 6.1.6 Statistical
considerations of Primary and Secondary Efficacy Endpoints.


6.1.2   Demographics
The demographic characteristics of patients by treatment groups in ONTARGET,
TRANSCEND and PRoFESS trials are shown in Table 10.

ONTARGET trial:
As shown in Table 10, demographics and baseline characteristics of the randomized
patients in ONTARGET are similar in the 3 treatment groups. CAD was the primary risk
factor for inclusion in the trial for the majority (66.1%) of patients.
At baseline, 82.8% of patients had hypertension (68.7% by medical history and 59.8%
by baseline BP measurement), 40.9% had diabetes (37.5% by medical history and
27.6% by their baseline plasma glucose levels).
The mean SBP/DBP at baseline was 141.8/82.1 mmHg, and it did not differ between
treatment groups. The mean SBP/DBP at randomization was also similar in the 3
treatment groups (T/R: 134.5/77.1 mmHg; T: 134.3/77.1 mmHg; R: 134.2/77.0 mmHg).
At baseline, patients were receiving aspirin (75.7%), statins (61.6%, increasing to 70.6%
at study end), or β-blockers (56.9%).



                                            38 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 10 Demographics and baseline characteristics by treatment (FAS)




1
  Patients in TRANSCEND and PRoFESS who did not take an ACE-I during the trials

2
  As given by the patient; the question to the patient referred to ethnic origin, not to nationality or race. 

3
  For 7 patients in ONTARGET (T/R: 2 patients; T: 3 patients, R: 2 patients), 0 patients in 

TRANSCEND/NoACE-I, and 0 patients in PRoFESS/noACE-I, ethnicity was not recorded. 

4
  Obesity was defined as BMI ≥30 kg/m2 in non-Asian patients and BMI ≥27 kg/m2 in Asian patients. 

  Source: Table 3.1.3.1:1 on page 75 of Summary of Clinical Efficacy


TRANSCEND trial:
As shown in Table 10, demographics and baseline characteristics of the randomized
patients in TRANSCEND are similar in the 2 treatment groups. For the majority of
patients, CAD was the primary risk factor for inclusion in the trial (66.8%). At baseline,
hypertension was present in 86.0% of patients and diabetes mellitus in 41.6%.
The majority of patients received ASA (74.7% at baseline, 71.2% at study end), β­
blockers (58.3% at baseline, 57.8% at study end), and statins (55.2% at baseline,
increasing to 63.5% at study end). After randomization, differences between treatment
groups in the use of concomitant medication were seen for diuretics (33.0% at baseline,
33.7% at study end in the T group, 40.0% at study end in the PBO group), calcium
channel blockers excluding diltiazem/verapamil (31.1% at baseline, 30.8% vs. 39.2%,
respectively, at study end), and β-blockers (58.3% at baseline, 56.6% vs. 59.0%,
respectively, at study end).


                                                       39 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

PRoFESS trial:
As shown in Table 10, demographics and baseline characteristics of the randomized
patients in PRoFESS are similar in the 2 treatment groups. Overall 39.8% of patients
were randomized within 10 days after the qualifying stroke; this stroke was attributed to
small-artery occlusion in 52.0% patients, and to large-artery atherosclerosis in 28.6% of
patients. At baseline, mean SBP/DBP was 144.1/83.8 mmHg, and 74.0% of patients
were hypertensive.
The most frequent concomitant medication classes at baseline were anti-hypertensives
(67.3%), anti-thrombotics (65.8%) and lipid-modifying medications (48.5%). The
treatment groups (both ASA+ER-DP vs. clopidogrel and telmisartan vs. placebo) were
matched for all baseline characteristics.
In the pooled patient population of TRANSCEND and PRoFESS who did not take an
ACE-I during the trials, too, demographic characteristics of patients randomized to T
and those randomized to PBO were similar.


6.1.3    Subject Disposition
ONTARGET trial:
The majority (99.8%) of randomized patients completed the trial (defined as final visit
performed or vital status confirmed including death) (Table 11).

Table 11 Disposition of patients in ONTARGET, TRANSCEND & PRoFESS (FAS)




Source: Table 3.1.2.2:1 on page 69 of Summary of clinical efficacy




                                                   40 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

A total of 18.2% of all randomized patients permanently discontinued study medication
(T/R: 20.3%; T: 17.0%; R: 17.1%). Of those who permanently discontinued study
medication, 53.8% stopped due to AEs (T/R: 60.9%; T: 48.1%; R: 51.0%). The risk for
permanent treatment discontinuation of both study medications was significantly greater
with T/R than with T or R (risk ratio vs. T: 1.20; risk ratio vs. R: 1.19; p<0.0001 for each
comparison); no difference was seen between T and R (risk ratio: 0.99; p=0.8678).
Study drug compliance (defined as ≥75% of study medication taken based on blister
card count at each visit) was slightly worse in the T/R group than in the T or R group.

TRANSCEND trial:
Of 6,665 patients enrolled into the run-in period, the most frequent reasons for non-
randomization were the patients' request to stop treatment (6.1%), insufficient
compliance with study medication (5.1%), and the occurrence of AEs (2.6%).
18.5% of all randomized patients permanently discontinued study medication (T: 17.7%;
PBO: 19.4%) with a trend for fewer discontinuations in the telmisartan group (risk ratio
for T vs. PBO: 0.91; 95% CI 0.82, 1.02; p=0.0969). Of those who permanently
discontinued study medication, 42.9% stopped due to AEs (T: 47.2%; PBO: 39.1%).
77.1% (PBO) to 79.2% (T) patients were compliant with study medication. The weighted
mean change in SBP/DBP from baseline post-randomization was −6.5/−4.9 mmHg with
telmisartan and −2.3/−2.6 mmHg with placebo; this treatment difference in BP was
generally maintained throughout the study.

PRoFESS trial:
A total of 20,332 patients were randomized (T: 10,146; PBO: 10,186). Follow-up visits
were performed at 1 month, 3 months, and 6 months after randomization, and every 6
months until the end of the study, similar to ONTARGET trial. Although the overall trial
duration (first-patient-in to last-patient-out) was 4.5 years, the median observation time
for telmisartan was only 2.44 years (892 days).




                                            41 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


6.1.4    Analysis of Primary Endpoint(s)
{Note: We will present only our summarized review of the primary efficacy
endpoints here. We will present all of our consolidated, detailed statistical
analyses and discussion in Section 6.1.6 Statistical considerations of Primary
and Secondary Efficacy Endpoints.}


ONTARGET trial: For the analysis of primary efficacy endpoints, we will follow the
steps outlined in Section 5.2  Review Strategy.
As mentioned in section 5.3.3 Statistical Methods, the primary efficacy endpoints were 

analyzed on three types of patient data sets (FAS, PPS and FAS-10) to determine 

consistency of the results. 

Superiority of T/R over R: 

Analysis of the FAS shows that the primary outcome occurred in 1,386 patients (16.3%) 

in the T/R group, 1,423 patients (16.7%) in the T group, and 1,412 patients (16.5%) in 

the R group (Table 12). 


Table 12 Analysis of ONTARGET primary endpoint for superiority (T/R vs R)
                                 T/R           R        HR (95% CI) / P value
Randomized, n (%)            8,502 (100) 8,576 (100)               ---
                                 FAS Population
Primary endpoint, n (%)      1,386 (16.3) 1,412 (16.5) 0.99 (0.92, 1.07) / 0.8462
Events per 100 patient-years     3.79         3.82
                                 PPS Population
Primary endpoint, n (%)      1,095 (12.9) 1,164 (13.6) 1.00 (0.93, 1.09) / 0.9093
Events per 100 patient-years     3.60         3.58
                                FAS-10 Population
Primary endpoint, n (%)               1,094 (12.9)      1,183 (13.8)
                                                                            1.00 (0.92, 1.09) / 0.9926
Events per 100 patient-years              3.72              3.72
 Sensitivity analysis accounting for CHF-hospitalizations confirmed by chest X-ray
Primary endpoint, n (%)        1,323 (15.6)  1,359 (15.8)
                                                            0.98 (0.91, 1.06) / 0.6738
Events per 100 patient-years       3.61          3.67
Sensitivity analysis accounting for SBP at BL, & SBP change from BL, time-dependent
Primary endpoint, n (%)        1,386 (16.3) 1,412 (16.5)    1.00 (0.93, 1.08) / 0.9553
Events per 100 patient-years       3.79          3.82
Sensitivity analysis accounting for SBP at BL and SBP change for BL before event
Primary endpoint, n (%)      1,386 (16.3) 1,412 (16.5)   0.99 (0.92, 1.07) / 0.8111
Events per 100 patient-years     3.79         3.82
Source: Sponsor’s Tables 15.2.1.1:1, 15.2.1.1:3, 15.2.1.1:5, 15.2.1.1:7 & 15.2.1.1:8 on pages 541, 544,
547, 550 - 552 of ONTARGET Clinical Trial Report; BL = baseline




                                                   42
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

In the FAS, the event rate per 100 patient years (PY) is 3.79 for T/R, 3.87 for T, and
3.82 for R. The hazard ratio of T/R vs. R is 0.99 (95% CI 0.92, 1.07; p=0.8462). Thus,
the ONTARGET trial did not demonstrate superiority of T/R vs. R.
For the PPS (where adherence to study medication is assumed), the incidences
observed across treatment groups are lower than the FAS and replicate the results of
the FAS analysis (Table 12). This confirms the lack of superiority of T/R over R.
Analysis of the primary endpoint on the FAS-10 population also shows a similar trend
(Table 12) and further confirms the lack of superiority of T/R over R.

Renal endpoint (in the subgroup of patients with diabetic nephropathy):
ONTARGET also has 3-fold composite primary endpoint of doubling of serum
creatinine, progression to ESRD (defined by initiation of dialysis, need for renal
transplantation or eGFR <15 mL/min/1.73m2) and all-cause mortality in a subgroup of
diabetic nephropathy patients (diabetic patients with macro-albuminuria UACR≥300
mg/g Crea at baseline). The statistical analysis plan pre-specifies that confirmatory
testing of this renal primary endpoint will be performed only if both, superiority of T/R
vs. R and non-inferiority of T vs. R regarding the primary cardiovascular composite
endpoint are proven.
As superiority is not demonstrated for the primary CV endpoint, the analysis of the renal
3-fold endpoint is done only as an exploratory analysis and for completeness of review
of the primary efficacy endpoint data.

Table 13 Analysis of the primary renal endpoint in ONTARGET / FAS (DN)




Source: Table 11.4.1.1.2:1 on page 205 of ONTARGET Clinical Study Report.




                                                43 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

For the 3-fold renal primary efficacy endpoint (Table 13), there is no benefit of either T/R
or T over R, and no evidence of non-inferiority of T vs R. Death was the main
component contributing to the primary renal endpoint (25.4% in T/R vs. 29.8% in R).
Analysis of the ‘per protocol set of patients with diabetic nephropathy’ – PPS (DN) – and
‘the set of patients with diabetic nephropathy who stayed on the full dose of ramipril or
ramipril placebo’ – FAS-10 (DN) – shows the same results.

Non-inferiority of T vs. R:
The largest portion of our review will be focused on the evaluation of non-inferiority of T
vs. R, following the steps outlined in Section 5.2.
For Step 2 – We evaluated non-inferiority analysis of the primary efficacy endpoint
(please see Section 6.1.1 Methods and Section 6.1.6 Statistical considerations of
Primary and Secondary Efficacy Endpoints).
Non-inferiority analyses using the data sets for the FAS, PPS and FAS-10 populations
(Table 14) show that the upper limits of the 97.5% CIs are below the sponsor’s pre­
defined NI margin of 1.13 but all exceed the NI margin of 1.08 calculated by the
reviewers.

Table 14 Non-Inferiority analyses of ONTARGET composite primary efficacy
endpoint (T vs R)
                                   T            R       HR (97.5% CI) / P value*
Randomized, n (%)             8,542 (100)  8,576 (100)              ---
                                  FAS Population
Primary endpoint, n (%)       1,423 (16.7) 1,412 (16.5) 1.01 (0.93, 1.10) / 0.0019
Events per 100 patient-years      3.87         3.821




                                  PPS Population
Primary endpoint, n (%)       1,214 (14.2) 1,164 (13.6)
                                                        1.02 (0.93, 1.12) / 0.0078
Events per 100 patient-years      3.66         3.58
                                 FAS-10 Population
Primary endpoint, n (%)       1,237 (14.5) 1,183 (13.8) 1.03 (0.94, 1.13) / 0.0113
Events per 100 patient-years      3.83         3.72
Source: Tables 15.2.1.1:1, 15.2.1.1:3, 15.2.1.1:5, on pages 541, 544 and 547 of ONTARGET Clinical
Trial Report; *Sponsor’s analyses based on NI margin of 1.13.



FDA Reviewers’ comments: An important issue is whether the NI margin needs to be
closer to 1.08 (the reviewers’ calculation) or to 1.13 (the sponsor’s proposed NI margin).
We think that the NI margin should be closer to 1.08, because the effect size has
decreased with time {HOPE in 2000 showed a relative risk reduction (RRR) of 22%,
EUROPA in 2003 showed a RRR of 20%, and PEACE in 2004 showed a RRR of 4%.
The NI margin calculated from a meta-analysis of the 3 ACE-I trials is 1.06.


                                                 44 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The current placebo-controlled trial of telmisartan – TRANSCEND in 2008 –showed a
RRR of 8% for the 4-fold composite endpoint and a RRR of 13% for the 3-fold
composite endpoint}.
Please see Section 6.1.6 Statistical considerations of Primary and Secondary Efficacy
Endpoints for more detailed discussion.


Sensitivity Analyses
For sensitivity analyses (Table 15), we evaluated these data which are adjusted for (i)
systolic BP at end of the trial and (ii) systolic BP at the visit just before the primary event
occurred, and (iii) use of open-label ACE-I and/or ARBs.
We think that the CHF-hospitalization component of the four-fold composite primary
efficacy endpoint was a relatively “soft” endpoint, particularly because following Protocol
Amendment 3.1, evidence of heart failure from chest X-rays was not mandatory. To be
consistent with the HOPE trial where CHF-hospitalization was analyzed (as a secondary
endpoint) based only on cases confirmed by chest X-ray, the sponsor performed a post-
hoc sensitivity analysis including only CHF-hospitalizations where pulmonary
congestion was confirmed by chest X-ray. The hazard ratio of T vs. R was 1.00 (97.5%
CI 0.92, 1.09, with a one-sided p-value for non-inferiority: 0.0007, Table 15).
The mean change from baseline in sitting SBP was -6.7 mmHg in T group and -5.6
mmHg in R group. Throughout the study, BP in the T/R and T group was lower than that
in the R group; the weighted mean change from baseline post-randomization was -8.3/­
5.9 mmHg, -6.7/-5.1 mmHg, and -5.6/-4.4 mmHg, respectively for T/R, T and R
treatment groups. Thus, it is necessary to determine if there is an effect of BP reduction
on the primary endpoint. The sensitivity analysis (Cox proportional hazards model) with
baseline SBP and SBP-change as covariates, and another analysis with baseline line
SBP and SBP-change before event as covariates (Table 15) showed that adjusting for
change in SBP did not alter the hazard ratio or upper bounds of the confidence intervals
significantly.
Some subjects in the ONTARGET trial received open label ACE-Is (5.2% of patients
randomized to telmisartan and 5.7% of patients randomized to ramipril) and open label
ARBs (5.0% of patients randomized to telmisartan and 5.5% of patients randomized to
ramipril). Sensitivity analyses excluding patients who took an open-label ACE-I in T
group (1,040 patients) or an open-label ARB in the R group (469 patients) or both
during the time they belonged to the PPS (i.e., before censoring) for the primary efficacy
endpoint are shown in Table 15. These analyses showed that excluding patients who
took open-label ACE-Is and/or ARBs did not alter the primary endpoint findings except
when excluding patients who took open-label ACE-I for T group (HR 0.88, CI 0.81, 0.97,
p<0.0001) and when excluding patients who took open-label ACE-I for T group or open-
label ARB in R group (HR 0.90, CI 0.82, 0.99, p<0.0001).




                                              45 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 15 Sensitivity analyses for non-inferiority of T vs R in ONTARGET trial
                                     T                R         HR (97.5% CI)
       Accounting for CHF-hospitalizations confirmed by chest X-ray
Randomized, n (%)               8,542 (100)     8,576 (100)
Primary endpoint, n (%)          1,353 (15.8)  1,359 (15.8)  1.00 (0.92, 1.09)
Events per 100 patient-years         3.66          3.67
       Accounting for SBP at BL, & SBP change from BL, time-dependent
Randomized, n (%)               8,542 (100)  8,576 (100)
Primary endpoint, n (%)        1,423 (16.7)  1,412 (16.5)     1.02 (0.94, 1.11)
Events per 100 patient-years       3.87           3.82
        Accounting for SBP at BL and SBP change for BL before event
Randomized, n (%)               8,542 (100)  8,576 (100)
Primary endpoint, n (%)        1,423 (16.7)  1,412 (16.5)     1.02 (0.93, 1.10)
Events per 100 patient-years       3.87           3.82
    Excluding patients who took open-label ACE-I for all treatment groups
Randomized, n (%)               7,502 (100)  7,522 (100)
Primary endpoint, n (%)                 1,100 (14.7)       1,086 (14.4)      1.02 (0.92, 1.12)
Events per 100 patient-years                3.38               3.32
        Excluding patients who took open-label ACE-I for T group
Randomized, n (%)            7,502 (100)    8,576 (100)
Primary endpoint, n (%)                 1,100 (14.7)       1,412 (16.5)     0.88 (0.81, 0.97)*
Events per 100 patient-years                3.38               3.82
   Excluding patients who took open-label ARB for all treatment groups
Randomized, n (%)             8,118 (100)   8,107 (100)
Primary endpoint, n (%)                 1,315 (16.2)       1,308 (16.1)      1.00 (0.92, 1.10)
Events per 100 patient-years                3.76               3.74
        Excluding patients who took open-label ARB for R group
Randomized, n (%)            8,542 (100)   8,107 (100)
Primary endpoint, n (%)                 1,423 (16.7)       1,308 (16.1)      1.03 (0.95, 1.13)
Events per 100 patient-years                3.87               3.74
 Excluding patients who took open-label ACE-I and/or ARB for all treatment
                                  groups
Randomized, n (%)              7,263 (100)   7,240 (100)
Primary endpoint, n (%)                 1,051 (14.5)       1,035 (14.3)      1.01 (0.92, 1.12)
Events per 100 patient-years                3.33               3.29
Excluding patients who took open-label ACE-I in T group or open-label ARB
                in R group (No ARB for R + No ACE-I for T)
Randomized, n (%)              7,502 (100)   8,107 (100)
Primary endpoint, n (%)                 1,100 (14.7)       1,308 (16.1)     0.90 (0.82, 0.99)**
Events per 100 patient-years                3.38               3.74
Source: Compiled using data from Tables 15.2.1.1:7 & 15.2.1.1:8 on p. 550 – 552 of ONTARGET Clinical
Trial Report and Tables A 2.1.1:1 to A 2.1.1:6 on p. 294 – 304 of Integrated Summary of Efficacy; BL =
baseline; *P=0.0021, **P=0.0118, for 2-sided test for superiority.



                                                 46
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

(Of 1040 patients who took open-label ACE-I in the T group, 446 patients (42.9%) took
open-label ACE-I concomitantly with study medication, 222 patients (21.3%) after
discontinuation of study medication, and for 372 patients (35.8%) on and post study
medication. Of patients who took open-label ARB in the R group, 132 patients (28.1%)
took open-label ARB concomitantly with study medication, 118 patients (25.2%) after
treatment discontinuation, and for 219 patients (46.7%) on and post study medication.)

FDA Reviewers’ comments: Most of the sensitivity analyses show consistent results,
but do not meet the NI margin of 1.08. The sensitivity analyses probably are not of any
clinical significance because they are found post hoc without adjusting for multiple
statistical comparisons.

TRANSCEND trial – Primary Efficacy Endpoint
In TRANSCEND trial, the primary 4-fold composite endpoint occurred in 465 patients
(15.7%) in the T group and 504 patients (17.0%) in the PBO group (Table 16). The trial
did not demonstrate superiority of T over PBO in the prevention of CV death, non-fatal
MI, non-fatal stroke, and hospitalization for CHF. Baseline SBP and SBP changes over
time did not substantially influence the results for the primary endpoint; the SBP-
adjusted hazard ratio of T vs. PBO was 0.94 (95% CI 0.83, 1.07; p=0.3638).

Table 16 Analysis of the primary composite endpoint in TRANSCEND (FAS)




      Source: Table 15.2.1:1 on page 403 of TRANSCEND Clinical Trial Report

For non-fatal stroke and non-fatal MI, incidences were lower in the T group than in the
PBO group (Table 16); no differences were seen for hospitalization for CHF and CV
death. No meaningful subgroup-by-treatment interaction was observed for any of the 30
subgroups analyzed.
While the TRANSCEND trial was initially planned and powered (at 94% to show
superiority of T over PBO) based on the event rates observed in HOPE, the event rates
in TRANSCEND were less than expected. Post hoc recalculation of power for the
TRANSCEND trial using the observed event rates showed that it has only a 25%
chance of demonstrating superiority of T over PBO.


                                              47 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


PRoFESS trial – Primary Efficacy Endpoint
For this review, only the T vs PBO comparison portion in the PRoFESS trial for the
primary efficacy endpoint (time to the first recurrent stroke (ischemic, hemorrhagic or
uncertain cause, fatal or non-fatal) will be discussed. The statistical analyses assume
that there was no interaction between ARBs and anti-platelet agents.
1,814 (8.9%) patients experienced the primary endpoint in the PRoFESS trial. For the T
vs PBO comparison, the incidences of recurrent stroke were 8.7% for T and 9.2% for
PBO (HR 0.95; 95% CI 0.86, 1.04; p = 0.2312; Table 17); thus the PRoFESS trial failed
to confirm the superiority of telmisartan over placebo for the primary endpoint.

Table 17 Cox proportional analysis of recurrent stroke in PRoFESS (RAN set)




      Source: Table 15.2.2.1: 2 on page 514 of PRoFESS Clinical Study Report.

The lines on Kaplan-Meier plot of the primary endpoint for the RAN set cross at 9
months (Figure 4).

Figure 4 Kaplan-Meier plot for time-to-recurrent stroke in PRoFESS – T vs PBO
(RAN set)




      Source: Figure 15.2.2.1:2 on page 513 of PRoFESS Clinical Study Report.



                                               48 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


A post-hoc time-dependent analysis of the primary endpoint (Table 18) showed a time-
by treatment interaction with differential results for the time period ≤6 months after
randomization (T: 3.4%, PBO: 3.2%; HR 1.07; 95% CI 0.92, 1.25) compared with the
time period >6 months after randomization (T: 5.3%, PBO 6.0%; HR 0.88; 95% CI 0.78,
0.99, p= 0.042).

Table 18 Time-dependent analysis of recurrent stroke in PRoFESS (RAN set)




       Source: Table 15.2.2.1:8 on page 524 of PRoFESS Clinical Study Report


Post hoc analyses of the primary endpoint by time from qualifying stroke to day of
randomization using (i) a cut off point of 10 days, or (ii) 4 groups of 0-7, 8-14, 15-30 and
>30 days, showed no difference in the primary endpoint. This suggests that treatment
with T very early after the qualifying stroke did NOT affect the frequency of a recurrent
stroke in the first month or at any time during the trial.
In general, the results were consistent across the 25 subgroups tested with no
significant sub-group-by-treatment interactions (all p-values >0.01). However, the
subgroup analysis according to the proportion of visits with BP control (i.e., BP control in
<25%, 25-50%, 50-75% and ≥75% of visits) revealed a significant interaction (p=0.02).
For those patients with BP control at <25% of visits, the HR was 1.07 (95% CI 0.90,
1.25) in favor of PBO, whereas for those with BP control at ≥75% of visits, the HR was
0.81 (95% CI 0.68, 0.97) in favor of T.

TRANSCEND & PRoFESS (Total/NoACE-I population) – Primary Efficacy Endpoint
Neither TRANSCEND nor PRoFESS trial on its own showed superiority of T over
PBO for the protocol-specified primary endpoint.
The sponsor made a post hoc analysis of a pooled populations of patients in
TRANSCEND and PRoFESS who did not take any ACE-I during the trials


                                               49 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

(Total/NoACE-I). In this post-hoc pooled patient population, there was a nominally
significant (p=0.0107) RRR of the primary endpoint by 10% (HR 0.90, 95% CI 0.83,
0.98) in patients randomized to T vs PBO (Table 19). No study-by-treatment interaction
was observed (p=0.4122).
In the pooled analyses, non-fatal stroke made the highest contribution to the composite
endpoint (Table 19) which is different from that in ONTARGET where CV death was the
largest contributor to the primary composite endpoint (Table 22).

Table 19 The primary endpoint in pooled TRANSCEND and PRoFESS (FAS –
Total/NoACE-I) population




      Source: Table 3.2.1.3: 1 on page 119 of Summary of clinical Efficacy

Figure 5 Kaplan-Meier plot for the primary composite endpoint in the pooled
TRANSCEND and PRoFESS populations – FAS (Total/NoACE-I)




              Source: Table 3.2.1.3:1 on page 120 of Summary of Clinical Efficacy



                                                50 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The lines on Kaplan-Meier plot of the primary endpoint for the pooled Total/NOACE-I
population (Figure 5) separate at about 6 months after which the difference between T
and PBO increased over time in favor of T.
In this pooled TRANSCEND and PRoFESS data, 549 patients (49.0% of those who had
a primary event) treated with T had experienced a stroke as primary outcome event
(Table 19) whereas in ONTARGET, 364 patients (only 25.6% of those who had a
primary event) treated with T experienced a non-fatal stroke (Table 22). The risk profile
of the pooled population of patients in TRANSCEND and PRoFESS is not similar to that
of patients in ONTARGET. This difference is probably attributable to patients enrolled in
PRoFESS who had just experienced a stroke (median time from stroke to enrollment
was 15 days in PRoFESS) and were therefore at a very high risk of a second stroke.
A post-hoc time-dependent analysis of the primary endpoint for the pooled patient
population of TRANSCEND and PRoFESS who did not take any ACE-I during the trials
(Total/ NoACE-I) revealed a statistically significant time-by treatment interaction
(p=0.0151, Cox regression model, Table 20). In the period ≤6 months after
randomization, the incidences of the primary endpoint were similar in the 2 treatment
groups (T: 3.5%; PBO: 3.3%). In the period >6 months, significantly fewer patients in the
T group than in the PBO group experienced a primary event (9.9% vs. 11.6%; HR: 0.85,
95% CI 0.78, 0.94; p=0.0009), suggesting a time-dependent effect of telmisartan.

Table 20 Time-dependent analysis of the primary composite endpoint in the
pooled TRANSCEND and PRoFESS populations (FAS – Total/NoACE-I)




      Source: Table 3.2.1.4:2 on page 122 of Summary of Clinical Efficacy



                                                51 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets



6.1.5   Analysis of Secondary Endpoints(s)
{Note: We will present only our summarized review of the secondary efficacy
endpoints here. We will present all of our consolidated, detailed statistical
analyses and discussion in Section 6.1.6 Statistical considerations of Primary
and Secondary Efficacy Endpoints.}

ONTARGET trial
Superiority of T/R vs R for the composite primary endpoint was not demonstrated;
further analyses of secondary endpoints for superiority of T/R vs R will not be made.
The acceptability of the sponsor’s non-inferiority analyses of T vs R for the 4-fold
composite primary endpoint still remains in question because of the disagreement
between the Division and the sponsor regarding the choice of the NI margin. This issue
will need to be discussed at the Cardiovascular and Renal Drugs Advisory Committee
meeting in July, 2009. We will continue to review the non-inferiority analyses of the
secondary endpoints for ONTARGET.
Key secondary 3-fold composite endpoint of CV death, non-fatal MI, and non-fatal
stroke (HOPE primary endpoint):
Table 21 shows the non-inferiority analyses of T vs R for the protocol-specified 3-fold
composite secondary endpoint for the FAS, PPS and FAS-10 populations. The hazard
ratios were similar, and these results persisted when adjustments were made for
baseline SBP and changes in SBP over time or the most recent SBP before an event.
For the composite 3-fold (HOPE) endpoint, the results barely made it for non-inferiority
at the NI margin of 1.08 for (i) the FAS set, and for some of the sensitivity analyses such
as (ii) accounting for SBP change before event, (iii) excluding patients who took open-
label ACEI for T group, (iv) excluding patients open-label ARB for all treatment groups,
and (v) excluding those who took open-label ACE-I in T group or open-label ARB in R
group (Table 21). Non-inferiority of T to R was not found for the PPS and FAS-10 data
sets and other sensitivity analyses (Table 21).
The individual components of the CV primary endpoint (CV death, MI, stroke and CHF
hospitalization), which are protocol-specified secondary endpoints, are further analyzed
(Table 22) as:
(i) first event of each component endpoint, and
(ii) all events of each component endpoint (to determine the total number of patients
      who had the endpoint during the study).
In either case, there are no meaningful differences between the T and R treatment
groups.




                                            52 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 21 Non-inferiority analyses of T vs R for 3-fold composite endpoint
                                           T             R                        HR (97.5% CI)
Randomized, n (%)                     8,542 (100)   8,576 (100)                        ---
                                           FAS Population
Secondary endpoint, n (%)             1,190 (13.9)  1,210 (14.1)
                                                                                0.99 (0.90, 1.08)
Events per 100 patient-years              3.18          3.23
                                           PPS Population
Secondary endpoint, n (%)             1,014 (11.9)   989 (11.5)
                                                                                1.00 (0.91, 1.11)
Events per 100 patient-years              3.02          3.01
                                         FAS-10 Population
Secondary endpoint, n (%)             1,016 (11.9)  1,003 (11.7)
                                                                      0.99 (0.90, 1.10)
Events per 100 patient-years         3.10           3.12
  Sensitivity analysis accounting for SBP at BL, & SBP change from BL, time-dependent
Secondary endpoint, n (%)        1,190 (13.9)   1,210 (14.1)
                                                                      0.99 (0.90, 1.09)
Events per 100 patient-years         3.18           3.23
    Sensitivity analysis accounting for SBP at BL and SBP change for BL before event
Secondary endpoint, n (%)        1,190 (13.9)   1,210 (14.1)
                                                                      0.99 (0.90, 1.08)
Events per 100 patient-years         3.18           3.23
           Excluding patients who took open-label ACE-I for all treatment groups
Randomized, n (%)                7,502 (100)    7,522 (100)
Secondary endpoint, n (%)         961 (12.8)     953 (12.7)           1.01 (0.91, 1.12)
Events per 100 patient-years         2.92           2.88
         Excluding patients who took open-label ACE-I for T group (No ACEI for T)
Randomized, n (%)                7,502 (100)    8,576 (100)
Secondary endpoint, n (%)         961 (12.8)    1,210 (14.1)          0.90 (0.82, 1.00)
Events per 100 patient-years         2.92           3.23
           Excluding patients who took open-label ARB for all treatment groups
Randomized, n (%)                8,118 (100)    8,107 (100)
Secondary endpoint, n (%)        1,115 (13.7)   1,134 (14.0)          1.98 (0.89, 1.08)
Events per 100 patient-years         3.14           3.20
         Excluding patients who took open-label ARB for R group (No-ARB for R)
Randomized, n (%)                8,542 (100)    8,107 (100)
Secondary endpoint, n (%)        1,190 (13.9)   1,134 (14.0)          0.99 (0.91, 1.09)
Events per 100 patient-years         3.18           3.20
    Excluding patients who took open-label ACE-I and/or ARB in all treatment groups
Randomized, n (%)                7,263 (100)    7,240 (100)
Secondary endpoint, n (%)         924 (12.7)     917 (12.7)           1.00 (0.90, 1.11)
Events per 100 patient-years         2.90           2.88
  Excluding patients who took open-label ACE-I in T group or open-label ARB in R group
Randomized, n (%)                7,502 (100)    8,107 (100)
Secondary endpoint, n (%)         961 (12.8)    1,134 (14.0)          0.91 (0.83, 1.00)
Events per 100 patient-years         2.92           3.20
Source: Compiled using data from Tables 15.2.1.1:4 & 15.2.1.1:8 on p. 550 – 552 of ONTARGET Clinical 

Trial Report; Tables A 2.1.2:1 to A 2.1.2:6 on p 307–317 of Integrated Summary of Efficacy; BL= baseline 





                                                  53
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 22 Individual components of the primary composite endpoint / FAS
(ONTARGET trial)
                                          T              R        HR (95% CI) / p value
Randomized, n (%)                    8,542 (100)     8,576 (100)             ---
                                          CV death
First event, n (%)                    598 (7.0)       603 (7.0)
                                                                 1.00 (0.89, 1.12) / 0.9421
First events per 100 patient-years      1.54            1.54
                                         Non-fatal MI
First event, n (%)                    431 (5.0)       400 (4.7)
                                                                 1.08 (0.94, 1.24) / 0.2534
First events per 100 patient-years      1.13            1.05
                                            All MIs
All events, n (%)                     438 (5.1)       409 (4.8)
                                                                 1.08 (0.94, 1.23) / 0.2874
All events per 100 patient-years        1.15            1.07
                                       Non-fatal Stroke
First event, n (%)                    364 (4.3)       402 (4.7)
                                                                 0.91 (0.79, 1.05) / 0.1829
First events per 100 patient-years      0.95            1.05
                                         All Strokes
All events, n (%)                     369 (4.3)       405 (4.7)
                                                                 0.91 (0.79, 1.05) / 0.2102
All events per 100 patient-years        0.97            1.06
                                    CHF hospitalization
First event, n (%)                    394 (4.6)       354 (4.1)
                                                                 1.12 (0.97, 1.29) / 0.1203
First events per 100 patient-years      1.03            0.92
         CHF hospitalization (confirmation of pulmonary congestion by chest X-ray)
First event, n (%)                    285 (3.3)       269 (3.1)
                                                                 1.06 (0.90, 1.26) / 0.4601
First events per 100 patient-years      0.74            0.70
Source: Compiled from data in Tables 15.2.2.1:1 to 15.2.21:14 on p. 569 – 590 of Integrated Summary of
Efficacy.

There appears to be some variability in the number of patients for the sub-categories of
MI (Table 23), particularly with regard to total MI and Q-wave or ST elevation MI.

Table 23 Incidence of MI subcategories** (FAS)




Source: Table 15.2.2.1:5 on page 5769 of Integrated Summary of Efficacy.


                                                 54 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

FDA reviewers’ comments: With regard to the sub-categories of Non Q-wave MI or
Non ST-elevation MI, and MI without significant ECG changes (Table 23), we do not
think that any clinical inference can be drawn.
The sponsor’s response to our queries regarding the degree of peak biomarker
(troponin, CK-MB, LDH, SGPT) re-elevation which was used for adjudicating the
endpoint of MI at the time these peak biomarker elevations occurred was not
informative. It appears that the clinical laboratory parameters of peak CK, peak CK-MB,
peak LDH, peak AST, and troponin T or I were to be measured in case of a suspected
myocardial infarction, and the data had to be entered into Case Report Form 40 (CRF
40). In their response, the sponsor submitted that:
(i) 	 their CRF 40 did not provide a field to enter the time when these parameters were
      assessed, and
(ii) the CRF did not provide instructions on when to assess these parameters.
The sponsor assumed that the investigators entered the highest value observed during
the clinical observation of a patient diagnosed with a myocardial infarction; thus "peak
CK" may well have been the highest CK value observed without any temporal
relationship to the MI.
We note also that in cases where the clinical picture was sufficiently clear to diagnose a
myocardial infarction, e.g. Q-wave MI, the laboratory parameters might not have been
entered on the CRF.
This inaccurate case recording to determine the endpoint of MI is troubling because it
may be concealing an excess of MI known to be associated with ARBs (ARB-MI
paradox). This unexpected effect of ARBs arose from the finding of a 19% relative
increase in MI with valsartan vs amlodipine in the 15,245 patient VALUE (Valsartan
Antihypertensive Long-term Use Evaluation) trial6, and an 11% increase in non-fatal MI
compared to placebo in Candesartan in Heart failure Assessment of Reduction in
Mortality and morbidity Alternative (CHARM-Alt) trial7.
A meta-analysis of 25 clinical trials of ARBs in 68,711 patients at risk for MI and >4000
events showed that the ARBs do not increase the risk of MI with a pooled OR of 1.03
(95% CI 0.93, 1.13, p=0.59)8.
On the other hand, a meta-analysis of 11 randomized clinical trials of ARBs with at least
100 patients in each group treated for at least 6 months involving 55,050 patients
showed that MI was significantly increased by 8% (pooled OR 1.08, 95% CI 1.01, 1.16,
p=0.03)9, with the VALUE and CHARM-Alt trials driving the pooled results.
While there is no consensus to date on whether ARBs increase MI, the data from
ONTARGET related to MI are not useful to help in this controversy.
(Please see FDA reviewers’ discussion and forest plots of the above meta-analyses in
Section 9.1 Literature Review/References).




                                            55 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

TRANSCEND trial
There was a nominally significant lower incidence of the key secondary 3-fold
composite endpoint of CV death, non-fatal MI, and non-fatal stroke (HOPE primary
endpoint, Table 24), in the T group (13.0%) than in the PBO group (14.8%), with hazard
ratio 0.87 (95% CI 0.76, 1.00; p=0.0483). This difference was due to the lower
incidences in T vs PBO (Table 24) for non-fatal MI (3.8% vs. 4.6%) and non-fatal stroke
(3.8% vs. 4.6%); there was no difference between treatment groups in the incidence of
CV death (5.7% in both treatment groups) in TRANSCEND compared to HOPE where a
reduction in CV death was found (R: 6.1% vs. PBO: 8.1%).

Table 24 Composite 3-fold endpoint in TRANSCEND (FAS) compared to HOPE




Source: Table 11.4.1.2.1:1 on page 184 of TRANSCEND Clinical Study Report.

The HOPE and TRANSCEND trials differed in some aspects. In HOPE, only patients
without previous heart failure and with preserved LV function were enrolled; in
TRANSCEND, patients with low LVEF (<40%) were not excluded from participation,
although patients could not have symptomatic heart failure. Patients with severe hepatic
dysfunction were excluded from the TRANSCEND trial but were allowed to participate in
the HOPE trial. Patients with intolerance to ACE-Is were excluded from participation in
HOPE, whereas TRANSCEND enrolled patients with ACE-I intolerance.
In contrast to the HOPE study, the patient population in TRANSCEND included (Table
49) a substantial proportion (21% ~ 21.7%) of Asian patients, patients with TIA as
qualifying diagnosis, a higher proportion of female patients (43.0% vs. 26.7%), a greater
percentage of patients with a medical history of hypertension (76.4% vs. 46.8%) and
stroke (22.0% vs. 10.9%), and a lower percentage of patients with known coronary
artery disease (67% vs. 80%) including previous MI (46.3% vs. 52.6%), stable angina
(37.1% vs. 55.5%), and unstable angina (15.3% vs. 25.5%). Due to changes in medical
practice, the use of β-blockers (70% vs. 39%), lipid lowering drugs (62% vs. 29%), and



                                               56 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

anti-platelet drugs (93% vs. 76%) at baseline was more frequent in TRANSCEND than
in HOPE, while patients in TRANSCEND took less frequently CCBs (including diltiazem/
verapamil) than patients in HOPE (40.2% vs. 47.1%). The median observation time in
TRANSCEND was 58 months (approximately 4.8 years). The HOPE trial was planned
to last 5 years, but trial was stopped early (after 4.5 years) by the Data Monitoring
Committee due to overwhelming evidence of benefit in the ramipril group.

PRoFESS trial
Two secondary endpoints were pre-defined for the telmisartan vs. placebo comparison
in PRoFESS:
(i) 	 the 4-fold composite of either recurrent stroke, MI, death due to vascular causes, or
       new or worsening of CHF (ONTARGET endpoint), and
(ii) 	 new onset of diabetes.
These endpoints were to be tested sequentially for superiority of T over PBO only if the
primary endpoint had shown statistical superiority. Since T not found to be superior to
PBO on the primary endpoint of time to first recurrent stroke, the secondary endpoints
results are not informative, and at best, considered exploratory.
The frequencies of the 4-fold composite secondary endpoint of CV death, MI, stroke,
and new or worsening CHF were 13.5% for T vs. 14.4% for PBO (hazard ratio: 0.94;
95% CI 0.87, 1.01; p=0.1073). Incidences of the post-hoc defined 3-fold composite
secondary endpoint of CV death, MI, and stroke (HOPE primary endpoint) were 12.7%
(T) vs. 13.5% (PBO) resulting in a hazard ratio: 0.94; 95% CI 0.87, 1.02; p=0.1260.

TRANSCEND & PRoFESS (Total/NoACE-I population) – Composite 3-fold (HOPE)
Endpoint
The sponsor analyzed the 3-fold composite (HOPE) endpoint of CV death, non-fatal MI,
and non-fatal stroke in a post-hoc pooled population of TRANSCEND and PRoFESS
who did not take an ACE-I during the trials (Total/NoACE-I population) to compare of
the results of TRANSCEND and PRoFESS with those of the HOPE study.
Table 25 shows that the incidence of the HOPE endpoint as well as the event rate per
100 patient years were lower with T than with PBO (11.8% vs. 13.3%; 3.83 vs. 4.33,
HR: 0.88, 95% CI 0.81, 0.96; p = 0.0029). No study-by-treatment interaction was
observed (p = 0.9328). For each of the individual components, the incidence was lower
in the T than in the PBO group. As observed for the primary endpoint, non-fatal stroke
contributed most to the composite endpoint.
Also, as with the primary endpoint, the lines for T and PBO in the Kaplan-Meier plot
separate increasingly from 6 months after randomization till the end of the trial (Figure
6).




                                            57 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 25 Analysis of composite 3-fold (HOPE) endpoint in pooled TRANSCEND &
PRoFESS patient population (FAS – Total/NOACE-I)




      Source: Table 3.2.2.3:1 on page 128 of Summary of Clinical Efficacy



Figure 6 Kaplan-Meier plot for the composite 3-fold (HOPE) endpoint in the
pooled TRANSCEND and PRoFESS populations – FAS (Total/NoACE-I)




      Source: Figure 3.2.2.3: 1 on page 129 of Summary of Clinical Efficacy

An exploratory time-dependent analysis of the composite 3-fold (HOPE) endpoint for the
pooled patient population of TRANSCEND and PRoFESS who did not take any ACE-I
during the trials (Total/ NoACE-I) suggested that the treatment effect of T was time
dependent (Table 26). The time-by treatment interaction was statistically significant
(p=0.0178, Cox proportional hazards model), with fewer patients in the T group than in
the PBO group experienced a primary event in the time period > 6 months (HR: 0.83,
95% CI 0.75, 0.92; p=0.0002).


                                                58 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 26 Time-dependent analysis of the composite 3-fold (HOPE) endpoint in
the pooled TRANSCEND and PRoFESS populations (FAS – Total/NoACE-I)




      Source: Table 3.2.2.4:1 on page 131 of Summary of Clinical Efficacy




                                                59 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


6.1.6   Statistical considerations of Primary and Secondary Efficacy Endpoints

In this section authored by the statistical reviewer, we present our consolidated,
detailed statistical review and analyses related to non-inferiority of telmisartan to
ramipril in ONTARGET including review and meta-analysis of historical ACE-I
trials (HOPE, EUROPA and PEACE), and supportive data from TRANSCEND and
PRoFESS trials.

ONTARGET
The FDA reviewers performed extensive non-inferiority analyses for telmisartan (T) vs.
ramipril (R) in the following steps (as outlined in Section 5.2 Review Strategy and
Section 6.1.1 Methods).

Step 1: We examined the non-inferiority margin
The sponsor and the Division did not agree on the non-inferiority margin (NI) when the
application was still in IND stage. (Please see Section 2.5  Summary of
Presubmission Regulatory Activity Related to Submission and Section 5.3.3 Statistical
Methods). A detailed discussion can be found in the minutes of the meeting between
the sponsor and FDA on April 10, 2001.
The sponsor derived the NI margin of 1.13 based on HOPE trial alone. The 13% excess
risk for telmisartan to rule out in the non-inferiority analysis based is equal to about half
of the 40th percentile of the excess risk of placebo compared to ramipril in HOPE trial.
The calculation is as follows: The hazard ratio of placebo relative to ramipril is 1.29. The
40th percentile of the standard normal distribution is z0.40 = -0.25. The standard error of
log hazard ratio is 0.0524. The 40th percentile of the excess risk of placebo compared to
ramipril can be computed as
exp(log(1.29)+z0.4*0.0524)=1.27
13% is roughly half of the excess risk of 27%. In essence, this calculation is equivalent
to using the lower limit of the 20% two-sided confidence interval of the effect of placebo
relative to ramipril in HOPE trial and then discounting it by a half.
The Division thought that it would be acceptable to exclude half of the excess risk
associated with placebo. However, the Division suggested use of the lower limit of at
least 95% confidence interval of the effect of placebo relative to ramipril plus a 50%
preservation. However, if only HOPE is used for the computation, it is not clear whether
use of 95% confidence interval is conservative enough, because there is no way to
estimate inter-trial variability that needs to be incorporated in deriving a non-inferiority
margin. Should a 99% confidence interval, instead of 95% confidence interval, be used?
If 95% confidence interval is used, the margin is calculated by discounting 50% of the
lower bound of the 95% confidence interval of log hazard ratio of placebo relative to
ramipril.


                                             60 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

exp[(log(1.29)-1.96*0.0524)*0.5]=1.08 

This gives a non-inferiority margin of 1.08. Thus, in determining non-inferiority margin, 

the essential difference between the sponsor’s calculation and the calculation that the 

Division is willing to articulate is between 20% and at least 95% confidence interval. 

The reviewers performed further exploratory calculations: 

1) We computed the NI margin based on the original 3-fold endpoint in HOPE trial. 

2) We computed the NI margin based on the 4-fold endpoint in HOPE trial. This 

   includes hospitalizations into the composite endpoint and is the same endpoint used
   as the primary endpoint in ONTARGET.
3) We computed the NI margin from a meta-analysis of HOPE and two other ACE-I
   trials (EUROPA and PEACE trials) in patients with CV risk factors. We used both the
   fixed effects model and the random effects model.
All of the calculations above take 50% discount on the lower bound of confidence
intervals of the log hazard ratio to obtain the NI margin. As a general principle, the NI
margin is derived based on a 95% confidence interval. However, if only HOPE trial is
used in deriving the NI margin, the inter-study variability cannot be evaluated. In this
case, it may be necessary to consider using at least a 99% confidence interval instead.
The sponsor stated in section 1.2 of the Advisory Committee briefing package that the
“non-inferiority margin for ONTARGET was selected that was identical to the margin
selected for VALIANT15 (1.13), which resulted in FDA approval for valsartan to reduce
CV mortality in post-MI patients with left ventricular dysfunction”. However, we found
that the non-inferiority margin for VALIANT was 1.09.
Also in the Advisory Committee briefing package, the sponsor performed a 'putative
placebo’ comparison (Hasselblad et al 2001) based on the historical data of the HOPE
trial and came to the conclusion that “telmisartan preserved about 95% (95% CI, 66,
124) of the benefits of ramipril over placebo with respect to the 4-fold endpoint, and
preserved 105% (95% CI, 74 to137) of the benefits with respect to the 3-fold outcome”.
It is well known in the literature that this ‘putative placebo’ method is very sensitive to
the constancy assumption. According to our review, the event rate in placebo appeared
to decrease as the standard of care in clinical practice changed during the past 10
years. There was a trend that the effects of ACE inhibitors are decreasing over time
(Figure 7, and Figure 9). So the constancy assumption may not be valid anymore.

Table 27 NI Margins derived based on HOPE Trial alone
                    HR        90% CI         NI         95% CI         NI         99% CI         NI
                 estimate                   margin                    margin                    margin
                   (P/R)
 HOPE 3 fold
                  1.290    (1.183, 1.406)   1.087    (1.164, 1.429)   1.079    (1.127, 1.476)   1.062
 endpoint
 HOPE 4 fold
                  1.290    (1.189, 1.400)   1.090    (1.171, 1.422)   1.082    (1.135, 1.467)   1.066
 endpoint
Source: FDA analyses of HOPE data in NDA.


                                              61
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets



Depending on the confidence interval used in the computation, the NI margin varies
from 1.06 to 1.09 (Table 27). The triple endpoint and the quadruple endpoint did not
seem to yield a big difference in the NI margin.
We also calculated the NI margin from a meta-analysis of HOPE1, EUROPA3 and
PEACE4 trials (Table 28 and Figure 7) which were placebo-controlled trials of ACE-Is
(ramipril, perindopril and trandolapril, respectively) conducted in patients with similar CV
risk. The constancy assumption may not hold so the validity of pooling the three studies
remains questionable. This meta-analysis is solely for exploratory purposes.


Table 28 Summary of three ACE-I trials in patients with CV risk factors, listed in
chronological order
                                    ACE-I          Treatment                Placebo         HR (T/P)       p-
Trial      Endpoint              treatment     events N      rate     events N      rate    with 95%       value
                                                             (%)                    (%)        CI

HOPE       CV death,                                                                            0.78
                                  ramipril      651     4645   14.0    826    4652   17.8                  <0.001
(2000)     nonfatal MI, stroke                                                              (0.70, 0.85)
EUROPA     CV death, MI,                                                                        0.81
                                 perindopril    488     6110   8.0     603    6108   9.9                   <0.001
(2003)     cardiac arrest                                                                   (0.71, 0.91)
           CV death, MI,                                                                        0.83
                                 perindopril    552     6110   9.0     663    6108   10.9                  0.001
           stroke                                                                           (0.74, 0.93)
PEACE      CV death,                                                                            0.96
                               trandolapril 909   4158 21.9   929      4132 22.5                            0.43
(2004)     nonfatal MI,                                                                     (0.88, 1.06)
           CABG, PCI
           CV death,                                                                            0.93
                               trandolapril 396   4158 9.5    420      4132 10.2                            0.32
           nonfatal MI, stroke                                                              (0.81, 1.07)
Source: FDA analyses of HOPE and EUROPA data in NDAs and PEACE data from literature.

Figure 7 Summary of HOPE, EUROPA and PEACE trials using the 3-fold
composite (HOPE) endpoint




Source: FDA analyses of HOPE and EUROPA data in NDAs and PEACE data from literature.


                                                 62 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Data from EUROPA and HOPE trials were submitted to the Agency before. The Agency
obtained the PEACE trial data from the principal investigator. However, the stroke
information from PEACE trial data obtained by the Agency contains only non-fatal
strokes. So the number of events in the composite 3-fold endpoint (CV death, MI and
stroke) is slightly less than what was reported in the New England Journal of Medicine,
which appears to include the fatal strokes (
Table 29).

Table 29 Summary of total number of events in PEACE trial
                            Trandolapril          Placebo
                              (rate %)            (rate %)
 HOPE endpoint               396 (9.5)           420 (10.2)         from publication
                                                              from publication, verified
  CV death, non-fatal MI   344 (8.3)          352 (8.5)       by the reviewer
  CV death, non-fatal
  MI, non-fatal stroke     392 (9.4)         417 (10.1)       calculated by the reviewer
Source: FDA analyses of PEACE data provided by the principal investigator (Dr. Marc Pfeffer) and data
from literature.

For computational purpose, the reviewer used the same composite endpoint for all three
trials in calculation of NI margins. The original endpoints for PEACE and EUROPA were
not the same composite endpoint as HOPE so the reconstruction of the HOPE endpoint
from those two trials is purely post-hoc. The hazard ratio point estimate and confidence
interval in PEACE trial are taken from the publication. However, in covariate-adjusted
analyses performed later on, the re-constructed 3-fold endpoint (CV death, non-fatal MI
and non-fatal stroke) is used since patient-level data are needed for the analyses.


Table 30 NI Margins based on HOPE, EUROPA and PEACE trials
                                              HR                        NI                   NI
                                           estimate        90% CI     margin       95% CI   margin
 HOPE, EUROPA and PEACE (fixed
 effects model, normal distribution)    1.21  (1.14, 1.28) 1.07    (1.13, 1.29)              1.06
 HOPE, EUROPA and PEACE (fixed
 effects model, t-distribution)         1.21  (1.12, 1.31) 1.06    (1.09, 1.35)              1.04
 HOPE, EUROPA and PEACE
 (random effects model, t-distribution) 1.21  (1.12,1.31)  1.06    (1.08, 1.35)              1.04
 HOPE and EUROPA (fixed effects
 model, normal distribution)            1.25  (1.18, 1.33) 1.08    (1.16, 1.35)              1.08
 HOPE and EUROPA (fixed effects
 model, t-distribution)                 1.25  (1.12, 1.40) 1.06    (1.06, 1.48)              1.03
 HOPE and EUROPA (random effects
 model, t-distribution)                 1.25  (1.12, 1.40) 1.06    (1.06, 1.48)              1.03
Source: FDA analyses of HOPE and EUROPA data in NDAs and PEACE data from literature.



                                                      63
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


The meta-analysis is based on the HOPE endpoint (3-fold endpoint, CV death, MI and
stroke). If the three trials can be integrated in a meta-analysis, the estimate of hazard
ratio based on the fixed effects model is 1.21 with 95% CI (1.13, 1.29). The NI margin
with 50% discount on the lower limit of the 95% confidence interval is 1.06 (Table 30).
Chronologically, the effects of ACE-Is appeared to decrease (the point estimate was
getting larger), if the three trials can be considered jointly to yield a meaningful effect for
ACE-I and better represent the effect of ramipril. This requires clinical judgment. This
troubling trend makes the validity of the constancy assumption even more doubtful.
The estimate of the between-study variability is 0.004, based on the DerSimonian and
Laird’s random effects method. The close-to-zero estimate of between-study variability
may be due to a small number of studies used in the meta-analysis. Since only three
trials are analyzed, the normal distribution used to construct the confidence interval for
the overall estimate of the hazard ratio may arguably be inappropriate too. So we also
used the t-distribution to compute the confidence interval; this results in a slightly more
conservative confidence interval (please refer to Follmann and Proschan (1999)). Again,
the NI margin is calculated by taking a 50% discount on the lower confidence intervals.
If only HOPE and EUROPA trials are used, the estimate of the between-study variability
is 0. The random effects model therefore seems to reduce to the fixed effects model.
We also looked at how the NI margin changes as a degree of discounting. Figure 8
shows the relationship between the NI margin and different degrees of discounting. 0%
discount makes the lower bound of the confidence interval become the NI margin. 100%
discount makes the NI margin to be 1, so the non-inferiority testing becomes a
superiority testing.
The NI margin proposed by the sponsor is equivalent to taking approximately a 20%
discount on the lower bound of the 95% confidence interval from the HOPE trial.
A decision is needed to determine how much to discount so as to derive a NI margin
around which we can, at least, assess whether telmisartan is effective through the non-
inferiority testing in ONTARGET.




                                              64 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

       Figure 8 NI Margins based on Different Discounting




Source: FDA analyses of HOPE and EUROPA data in NDAs and PEACE data from literature. The three
trials discounting curves are based on fixed effects model assuming normal distribution.



Step 2: Non-inferiority testing of the primary efficacy endpoint data in ONTARGET
The sponsor’s non-inferiority analyses of the data (Table 31) were based on their
proposed NI margin of 1.13, and the sponsor concluded that telmisartan is non-inferior
to ramipril. However, if the NI margin 1.08 is used, then most of the non-inferiority tests
(including sensitivity analyses using PPS population and FAS-10 population) do not
conclude that telmisartan is non-inferior to ramipril and thus cannot conclude that
telmisartan is effective.




                                              65 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 31 Non-Inferiority analyses of ONTARGET composite primary efficacy
endpoint (T vs R)
                                          T            R        HR (97.5% CI)
      Randomized, n (%)              8,542 (100)  8,576 (100)         ---
                                    FAS Population
      Primary endpoint, n (%)        1,423 (16.7) 1,412 (16.5) 1.01 (0.93, 1.10)
      Events per 100 patient-years       3.87     1
                                                      3.82
                                    PPS Population
      Primary endpoint, n (%)        1,214 (14.2) 1,164 (13.6) 1.02 (0.93, 1.12)
      Events per 100 patient-years       3.66         3.58
                                   FAS-10 Population
      Primary endpoint, n (%)        1,237 (14.5) 1,183 (13.8) 1.03 (0.94, 1.13)
      Events per 100 patient-years       3.83         3.72
      Source: Tables 15.2.1.1:1, 15.2.1.1:3, 15.2.1.1:5, on pages 541, 544 and 547 of ONTARGET
      Clinical Trial Report

Most of the additional sensitivity analyses (Table 32) seem to show consistent results,
but did not meet the NI margin of 1.08.
Table 33 summarizes the results of 3-fold and 4-fold composite endpoints in both HOPE
and ONTARGET. The 3-fold endpoint is the composite endpoint of CV death, MI and
stroke, the primary endpoint in HOPE. The 4-fold endpoint is the composite endpoint of
CV death, MI, stroke and CHF hospitalization, the primary endpoint used in
ONTARGET. The sponsor’s NI margin was derived based upon the 3-fold endpoint in
HOPE. So we examined the 4-fold endpoint in HOPE (using CHF hospitalization data
retrieved from the HOPE trial datasets) as well as the 3-fold endpoint in ONTARGET (a
pre-specified secondary endpoint). Please refer to Table 30 for the NI margins derived
based on 3-fold and 4-fold endpoints in HOPE trial.




                                              66 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 32 Sensitivity analyses for non-inferiority of T vs R in ONTARGET trial
                                             T                 R         HR (97.5% CI)
                 Accounting for CHF-hospitalizations confirmed by chest X-ray
      Randomized, n (%)                 8,542 (100)       8,576 (100)
      Primary endpoint, n (%)           1,353 (15.8)    1,359 (15.8)     1.00 (0.92, 1.09)
      Events per 100 patient-years          3.66            3.67
                 Accounting for SBP at BL, & SBP change from BL, time-dependent
      Randomized, n (%)                8,542 (100)     8,576 (100)
      Primary endpoint, n (%)          1,423 (16.7)    1,412 (16.5)       1.02 (0.94, 1.11)
      Events per 100 patient-years         3.87            3.82
                  Accounting for SBP at BL and SBP change for BL before event
      Randomized, n (%)                8,542 (100)     8,576 (100)
      Primary endpoint, n (%)          1,423 (16.7)    1,412 (16.5)       1.02 (0.93, 1.10)
      Events per 100 patient-years         3.87            3.82
              Excluding patients who took open-label ACE-I for all treatment groups
      Randomized, n (%)                7,502 (100)     7,522 (100)
      Primary endpoint, n (%)                   1,100 (14.7)         1,086 (14.4)          1.02 (0.92, 1.12)
      Events per 100 patient-years                  3.38                 3.32
                  Excluding patients who took open-label ACE-I for T group
      Randomized, n (%)              7,502 (100)     8,576 (100)
      Primary endpoint, n (%)                   1,100 (14.7)         1,412 (16.5)          0.88 (0.81, 0.97)
      Events per 100 patient-years                  3.38                 3.82
            Excluding patients who took open-label ARB for all treatment groups
      Randomized, n (%)              8,118 (100)    8,107 (100)
      Primary endpoint, n (%)                   1,315 (16.2)         1,308 (16.1)          1.00 (0.92, 1.10)
      Events per 100 patient-years                  3.76                 3.74
                   Excluding patients who took open-label ARB for R group
      Randomized, n (%)               8,542 (100)     8,107 (100)
      Primary endpoint, n (%)                   1,423 (16.7)         1,308 (16.1)          1.03 (0.95, 1.13)
      Events per 100 patient-years                  3.87                 3.74
       Excluding patients who took open-label ACE-I and/or ARB for all treatment groups
      Randomized, n (%)               7,263 (100)     7,240 (100)
      Primary endpoint, n (%)                   1,051 (14.5)         1,035 (14.3)          1.01 (0.92, 1.12)
      Events per 100 patient-years                  3.33                 3.29
       Excluding patients who took open-label ACE-I in T group or open-label ARB in R
                           group (No ARB for R + No ACE-I for T)
      Randomized, n (%)              7,502 (100)      8,107 (100)
      Primary endpoint, n (%)                   1,100 (14.7)         1,308 (16.1)          0.90 (0.82, 0.99)
      Events per 100 patient-years                  3.38                 3.74
      Source: Compiled using data from Tables 15.2.1.1:7 & 15.2.1.1:8 on p. 550 – 552 of ONTARGET
      Clinical Trial Report and Tables A 2.1.1:1 to A 2.1.1:6 on p. 294 – 304 of Integrated Summary of
      Efficacy; BL = baseline; (The results above are verified by the reviewer)




                                                67
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 33 Summary of Analyses on HOPE and ONTARGET
                   HOPE (superiority of ramipril vs placebo)
                                R            PBO          HR (95% CI) / P value
Randomized, n (%)                4,645 (100)       4,652 (100)                 ---
4-fold Primary endpoint, n (%)    726 (15.6)        919 (19.8)    0.775 (0.703, 0.854)/<0.001
3-fold Primary endpoint, n (%)    651 (14.0)        826 (17.8)      0.78 (0.70, 0.86)/<0.001
                 ONTARGET (non-inferiority of telmisartan vs ramipril)
                                 T                R             HR (97.5% CI)
Randomized, n (%)                8,542 (100)       8,576 (100)                ---
4-fold Primary endpoint, n (%)   1,423 (16.7)      1,412 (16.5)        1.01 (0.93, 1.10)
3-fold Primary endpoint, n (%)   1,190 (13.9)      1,210 (14.1)        0.99 (0.90, 1.08)

In section 5.5.1 of the Advisory Committee briefing package, the sponsor stated that the
upper bounds of the 97.5% confidence intervals (adjustment for multiple testing) for
ONTARGET were 1.10 and 1.08 for the 4-fold and 3-fold endpoints, respectively. The
adjustment for multiplicity mentioned by the sponsor was for testing both superiority of
the combination over ramipril and non-inferiority of telmisartan over ramipril for the
primary endpoint. The total alpha of 0.05 was split into 0.025 for superiority test and
0.025 for non-inferiority test. However, the sponsor listed the 4-fold endpoint and 3-fold
endpoint simultaneously without any multiple comparison adjustment. In order to control
the study-wise type I error, the secondary endpoints are tested only when the primary
endpoint wins. The 4-fold endpoint in ONTARGET trial is the primary endpoint and the
3-fold endpoint is the key secondary endpoint. If margin 1.13 is valid, then the 4-fold
endpoint wins the non-inferiority test. The 3-fold endpoint can be tested as a key
secondary endpoint. However, if a tighter margin, e.g., 1.08, was selected, the 4-fold
primary endpoint fails to show statistical significance in the non-inferiority test. The 3­
fold endpoint therefore cannot be tested anymore since the primary endpoint did not
win. Hence, there is an additional multiplicity issue because of testing the two endpoints
(4-fold and 3-fold endpoints)

Step 3: Evaluation of constancy assumptions and adjustments for covariates
The sponsor contended that the results of the EUROPA and PEACE trials became
available when ONTARGET was already running so that these studies could not have
been used to pre-specify calculations of the NI margin for ONTARGET. The sponsor
maintained that their formal meta-analysis (using the Cochran-Mantel-Haenszel test) of
the three studies showed that they are heterogeneous (p=0.0037 in the Breslow-Day
test for homogeneity of odds ratios), but a homogeneous effect of ACE-Is is found if the
PEACE trial is excluded {p=0.5715 in the Breslow-Day test for homogeneity of odds
ratios, with the pooled estimate of the risk ratio being 0.798 (0.742 – 0.858)}.
Table 30 shows the meta-analysis of three trials as well as the meta-analysis of
EUROPA and HOPE. The lower limit of the confidence intervals only increases slightly.
So the NI margins based on 50% of the lower limit of the confidence intervals do not
differ much whether PEACE trial is included or not. How much discount should be taken


                                            68 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

on the confidence intervals makes a much larger impact on the NI margin.
Because of the changes in standard of care that have evolved over the past several
years, it is important to ascertain that the constancy assumption holds and that the
effect size remains reasonably unchanged. It is also necessary to determine whether or
not there is a significant placebo effect creep or change in efficacy endpoint events in
the placebo group due to changes in standard of care.
If we look at the treatment effect in the three ACE-I trials, the hazard ratio point estimate
appears to yield a trend of going upwards (Figure 9 and Table 28). This suggests that
the effect size is decreasing. So can the NI margin derived based on HOPE trial alone
be applied to ONTARGET? The NI margin based on HOPE alone may overestimate
the ramipril effect in ONTARGET. A more conservative margin may be required.


Figure 9 Hazard Ratio and 95% Confidence Interval in Three Trials




We compared the placebo event rates in HOPE and TRANSCEND trials. We also
compared the event rates in ramipril arm and in telmisartan arm in TRANSCEND,
ONTARGET and HOPE trials (Figure 10, Figure 11 and Figure 12). The placebo event
rate decreases in TRANSCEND trial considerably when compared to HOPE trial. The
event rate in ramipril arm appears to be slightly higher in HOPE than in ONTARGET if
we look only at the 3-fold endpoint.



                                             69 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Figure 10 Comparison of placebo event rate in HOPE, TRANSCEND and PEACE




Figure 11 Comparison of event rate in ramipril in HOPE and ONTARGET




                                      70 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Figure 12 Comparison of event rate in telmisartan in TRANSCEND and
ONTARGET




We then investigated the interaction between treatment and a number of covariates in
HOPE, including gender, age, history of MI, history of stroke, history of CABG, use of β­
blockers, use of lipid lowering agents, use of antiplatelet agents, diabetes, hypertension,
high cholesterol, history of CAD, and history of TIA, which appeared to be important
covariates in predicting CV risk. Note that all of these analyses are exploratory and did
not take multiple comparison into consideration. Also, we used the HOPE endpoint (3­
fold composite of CV death, MI and stroke) in all of the covariate analyses below to
enhance the comparability.
Table 34 shows the analyses in the HOPE trial by adjusting these covariates. Each Cox
proportional hazard model only includes one covariate other than treatment. We
examined both models with treatment-covariate interaction and models without
treatment-covariate interaction. The interaction between treatment and use of anti-
platelets is the only one which shows a nominally significant p-value of 0.016. So the
hazard ratio of ramipril over placebo is similar across most of the covariates except
possibly use of anti-platelet agents. By including all of the covariates and two way
treatment*covariate interactions into the Cox proportional hazard model, the hazard
ratio of ramipril over placebo is 0.625 with 95% CI (0.407, 0.959), compared to the
hazard ratio of 0.78 with 95% CI (0.70, 0.85) in unadjusted analysis.




                                            71 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

In order to make comparisons in HOPE, EUROPA, PEACE and ONTARGET, the
reviewer redefine coronary artery disease (CAD) as:
1. Previous MI or
2. Multi-vessel percutaneous transluminal coronary angioplasty (PTCA) , or
3. Previous multi-vessel coronary artery bypass graft (CABG).
The CAD here does not include stable or unstable angina.


Table 34 Analyses of HOPE trial by adjusting covariates

                                                                   Hazard     Lower     Upper
Variable          Estimate    StdErr   ChiSq       ProbChiSq       ratio      bound     bound
ramipril              -0.25     0.05      22.97      <0.0001           0.78      0.70      0.86
Female                -0.26     0.06      17.32      <0.0001
ramipril              -0.24     0.06      16.84          <0.0001      0.78       0.70      0.88
Female                -0.25     0.08       8.51           0.0035
ram*female            -0.03     0.13       0.07           0.7897
ramipril              -0.26     0.05      24.62          <0.0001      0.77       0.70      0.85
age60                  0.40     0.07      30.87          <0.0001
ramipril              -0.31     0.14       5.34           0.0208      0.73       0.56      0.95
age60                  0.38     0.09      15.79           0.0001
ram*age60              0.06     0.15       0.19           0.6620
ramipril              -0.25     0.05      22.76          <0.0001      0.78       0.70      0.86
MI                     0.43     0.05      62.89          <0.0001
ramipril              -0.26     0.09       8.98           0.0027      0.77       0.65      0.92
MI                     0.42     0.07      34.30          <0.0001
ram*MI                 0.01     0.11       0.01           0.9255
ramipril              -0.25     0.05      23.60          <0.0001      0.78       0.70      0.86
stroke                 0.43     0.09      25.25          <0.0001
ramipril              -0.24     0.06      19.40          <0.0001      0.78       0.70      0.87
stroke                 0.48     0.11      18.13          <0.0001
ram*stroke            -0.11     0.18       0.41           0.5195
ramipril              -0.25     0.05      23.43          <0.0001      0.78       0.70      0.86
CABG                   0.15     0.06       7.17           0.0074
ramipril              -0.31     0.06      24.03          <0.0001      0.74       0.65      0.83
CABG                   0.07     0.08       0.90           0.3429
ram*CABG               0.18     0.12       2.40           0.1217
ramipril              -0.25     0.05      23.56          <0.0001      0.78       0.70      0.86
beta blocker           0.02     0.05       0.11           0.7442
ramipril              -0.25     0.07      13.42           0.0002      0.78       0.68      0.89
beta blocker           0.02     0.07       0.11           0.7373
ram*betablocker       -0.01     0.11       0.02           0.8909
ramipril              -0.26     0.05      23.88          <0.0001      0.77       0.70      0.86
lipid                 -0.39     0.06      38.87          <0.0001



                                                  72 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

ramipril             -0.24    0.06     16.94          <0.0001   0.78   0.70     0.88
lipid                -0.37    0.08     19.56          <0.0001
ram*lipid            -0.05    0.13      0.19           0.6663
ramipril             -0.25    0.05     23.26          <0.0001   0.78   0.70     0.86
antiplatelet          0.10    0.06      2.29           0.1304
ramipril             -0.49    0.11     18.96          <0.0001   0.61   0.49     0.76
antiplatelet         -0.04    0.08      0.22           0.6372
ram*antiplatelet      0.31    0.13      5.80           0.0160
ramipril             -0.26    0.05     24.02          <0.0001   0.77   0.70     0.86
diabetes              0.20    0.05     15.15           0.0001
ramipril             -0.24    0.07     11.86           0.0006   0.79   0.69     0.90
diabetes              0.22    0.07     10.19           0.0014
ram*diabetes         -0.04    0.11      0.18           0.6748
ramipril             -0.26    0.05     24.06          <0.0001   0.77   0.70     0.86
hypertension          0.17    0.05     10.22           0.0014
ramipril             -0.22    0.07      8.61           0.0033   0.80   0.70     0.93
hypertension          0.20    0.07      8.31           0.0039
ram*hypertension     -0.08    0.10      0.55           0.4585
ramipril             -0.26    0.05     24.34          <0.0001   0.77   0.70     0.86
cholesterol          -0.34    0.05     41.43          <0.0001
ramipril             -0.19    0.08      5.58           0.0181   0.83   0.70     0.97
cholesterol          -0.29    0.07     16.54          <0.0001
ram*cholesterol      -0.11    0.11      1.15           0.2828
ramipril             -0.25    0.05     23.38          <0.0001   0.78   0.70     0.86
CAD                   0.24    0.06     17.13          <0.0001
ramipril             -0.32    0.10     10.07           0.0015   0.73   0.60     0.89
CAD                   0.20    0.08      6.79           0.0092
ram*CAD               0.09    0.12      0.58           0.4466
ramipril             -0.25    0.05     23.34          <0.0001   0.78   0.70     0.86
TIA                   0.43    0.10     18.74          <0.0001
ramipril             -0.25    0.05     21.32          <0.0001   0.78   0.70     0.87
TIA                   0.44    0.13     11.27           0.0008
ram*TIA              -0.02    0.20      0.01           0.9107



In Figure 13, the hazard ratio within each covariate stratum is plotted. A pair of the
hazard ratio with 95% confidence interval (CI) is produced for each covariate. For
example, the left plot in the covariate “history of MI” is the hazard ratio with 95% CI for
patients without history of MI enrolled in the study. The right one is the hazard ratio with
95% CI for patients with history of MI. The hazard ratio and CI plots in patients using
anti-platelet and in patients not using anti-platelet appear to differ considerably.




                                               73 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 13 Hazard ratio in HOPE Trial by Covariates




Figure 14 shows hazard ratios after adjusting for individual covariates (this is done by
including treatment-covariate interaction in the model).
For exploratory purposes, we also computed the NI margin based on the hazard ratio
adjusted for each covariate (Figure 15), by discounting 50% on the worst limit of the
95% CI interval. With the exception of anti-platelet, NI margins adjusting for other
covariates vary from 1.02 to 1.10.




                                            74 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 14 Hazard ratio of ramipril over placebo adjusted for covariates in HOPE




Figure 15 NI margins based on hazard ratio adjusted for covariates in HOPE




                                       75 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Similar exploratory analyses were made for the EUROPA trial (Table 35, Figure 16).
None of the interaction term is significant. The estimate of hazard ratio of placebo over
perindopril is 0.833 if all the covariates and two way treatment*covariate interactions are
included into the Cox proportional hazard model. The corresponding 95% confidence
interval is (0.482, 1.437).

Table 35 Analyses on EUROPA Trial by Adjusting Covariates (HOPE endpoint)

                                                         Hazard   Lower   Upper
      Variable     Estimate   StdErr   ChiSq   p-value    Ratio   bound   bound
perindopril         -0.19      0.06    10.90   0.0010    0.827    0.738   0.926
female              -0.12      0.08     2.15   0.1424
perindopril         -0.19      0.06     9.06   0.0026    0.830    0.735   0.937
female              -0.11      0.11     0.93   0.3336
peri*female         -0.03      0.17     0.03   0.8522
perindopril         -0.19      0.06    10.96   0.0009    0.826    0.738   0.925
age60                0.38      0.06    38.98   <0.0001
perindopril         -0.31      0.10    10.27   0.0014    0.730    0.603   0.885
age60                0.29      0.08    13.08   0.0003
peri*age60           0.19      0.12     2.44   0.1179
perindopril         -0.19      0.06    11.08   0.0009    0.825    0.737   0.924
MI                   0.36      0.06    31.64   <0.0001
perindopril         -0.09      0.11     0.71   0.3984    0.912    0.737   1.129
MI                   0.42      0.09    23.36   <0.0001
peri*MI             -0.14      0.13     1.17   0.2798
perindopril         -0.19      0.06    10.87   0.0010    0.827    0.739   0.926
stroke               0.78      0.15    26.49   <0.0001
perindopril         -0.21      0.06    12.26   0.0005    0.814    0.725   0.913
stroke               0.57      0.23     6.23   0.0125
peri*stroke          0.42      0.31     1.92   0.1660
perindopril         -0.19      0.06    10.88   0.0010    0.827    0.739   0.926
CABG                -0.11      0.06     2.71   0.0995
perindopril         -0.19      0.07     7.87   0.0050    0.827    0.725   0.944
CABG                -0.11      0.09     1.46   0.2265
peri*CABG            0.00      0.13     0.00   0.9910
perindopril         -0.19      0.06    10.79   0.0010    0.828    0.739   0.927
beta blocker        -0.02      0.06     0.17   0.6810
perindopril         -0.08      0.09     0.80   0.3701    0.920    0.766   1.104
beta blocker         0.05      0.08     0.45   0.5014
peri*betablocker    -0.17      0.12     2.09   0.1478
perindopril         -0.19      0.06    10.82   0.0010    0.827    0.739   0.926
lipid               -0.33      0.06    32.08   <0.0001
perindopril         -0.22      0.08     7.77   0.0053    0.800    0.683   0.936
lipid               -0.36      0.08    21.11   <0.0001


                                               76 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

peri*lipid           0.07   0.12     0.37    0.5408
perindopril         -0.20   0.06    11.49   0.0007     0.823   0.735    0.921
antiplatelet        -0.46   0.09    25.93   <0.0001
perindopril         -0.42   0.17     6.05   0.0139     0.657   0.470    0.918
antiplatelet        -0.57   0.12    23.14   <0.0001
peri*antiplatelet    0.25   0.18     1.96    0.1618
perindopril         -0.18   0.06    10.04   0.0015     0.833   0.744    0.933
diabetes             0.56   0.07    59.24   <0.0001
perindopril         -0.19   0.06     8.84   0.0029     0.827   0.729    0.937
diabetes             0.54   0.10    30.78   <0.0001
peri*diabetes        0.04   0.15     0.08    0.7775
perindopril         -0.19   0.06    10.74   0.0010     0.828   0.740    0.927
hypertension         0.34   0.06    30.62   <0.0001
perindopril         -0.19   0.07     7.27   0.0070     0.826   0.719    0.949
hypertension         0.33   0.08    16.54   <0.0001
peri*hypertension    0.01   0.12     0.00    0.9671
perindopril         -0.19   0.06    10.96   0.0009     0.826   0.738    0.925
cholesterol         -0.31   0.06    28.41   <0.0001
perindopril         -0.26   0.09     8.59   0.0034     0.774   0.652    0.919
cholesterol         -0.36   0.08    21.38   <0.0001
peri*cholesterol     0.12   0.12     0.99    0.3195
perindopril         -0.19   0.06    10.89   0.0010     0.827   0.739    0.926
CAD                  0.08   0.10     0.60    0.4394
perindopril          0.10   0.19     0.30    0.5838    1.107   0.770    1.591
CAD                  0.23   0.14     2.73    0.0986
peri*CAD            -0.32   0.19     2.75    0.0974
perindopril         -0.19   0.06    11.26   0.0008     0.824   0.736    0.923
TIA                  0.79   0.16    25.50   <0.0001
perindopril         -0.19   0.06    10.52   0.0012     0.827   0.737    0.927
TIA                  0.83   0.22    14.81    0.0001
peri*TIA            -0.09   0.31     0.08    0.7759



We also examined the original primary endpoint in EUROPA. The interaction between
treatment and use of β-blocker has a relatively small p-value (0.068). The hazard ratio is
0.959 with 95% CI (0.538, 1.711) if all of the covariates and two way
treatment*covariate interactions are included into the Cox proportional hazard model.
Figure 17 shows hazard ratio adjusted for individual covariates (including treatment­
covariate interaction in the model). For exploratory purpose, we also computed the NI
margin based on the hazard ratio adjusted for each covariate. Some confidence
intervals of hazard ratio estimates include 1, so there is no NI margin in those cases.
We imputed the margin to be 1 for graphical display purpose (Figure 18).




                                            77 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 16 Hazard Ratio in EUROPA Trial by Covariates (HOPE endpoint)




Figure 17 Hazard ratio adjusted for each covariate in EUROPA




                                      78 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 18 NI margins based on hazard ratio adjusted for covariates in EUROPA




We also examined the integrated HOPE and EUROPA trials. The same covariates used
in the above analyses are also used in this integrated analysis. In addition to the
covariates and treatment interactions, this integrated study is also used as a
stratification factor in the model. Table 36 shows that the interaction of
treatment*antiplatelets is the only interaction term with a nominally significant p-value.

Table 36 Analyses on HOPE and EUROPA Trials by Adjusting Covariates
                                                           Hazard      Lower      Upper
Variable        Estimate    StdErr    ChiSq      p-value   Ratio       bound      bound
trt                 -0.22      0.04     33.40    <0.0001       0.799      0.741      0.862
female              -0.21      0.05     17.98    <0.0001
trt                 -0.22      0.04     25.48    <0.0001      0.806       0.741      0.876
female              -0.20      0.07      8.29     0.0040
trt*female          -0.04      0.10      0.19     0.6665
trt                 -0.23      0.04     34.78    <0.0001      0.796       0.737      0.858
age60                0.39      0.05     69.92    <0.0001
trt                 -0.31      0.08     15.59     0.0001      0.730       0.625      0.854
age60                0.34      0.06     31.37    <0.0001
trt*age60            0.11      0.09      1.52     0.2172
trt                 -0.22      0.04     33.38    <0.0001      0.799       0.741      0.862
MI                   0.40      0.04     93.79    <0.0001



                                                79 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

trt                -0.19   0.07    8.35     0.0039   0.824   0.722   0.939
MI                  0.42   0.06   57.07    <0.0001
trt*MI             -0.04   0.08    0.29     0.5892
trt                -0.23   0.04   33.80    <0.0001   0.798   0.740   0.861
stroke              0.51   0.08   45.59    <0.0001
trt                -0.23   0.04   31.44    <0.0001   0.798   0.737   0.863
stroke              0.51   0.10   25.27    <0.0001
trt*stroke          0.00   0.15    0.00     0.9814
trt                -0.23   0.04   33.76    <0.0001   0.798   0.740   0.861
CABG                0.03   0.04    0.67     0.4145
trt                -0.25   0.05   30.44    <0.0001   0.777   0.710   0.850
CABG               -0.01   0.06    0.02     0.8793
trt*CABG            0.10   0.09    1.28     0.2580
trt                -0.23   0.04   33.80    <0.0001   0.798   0.740   0.861
beta blocker        0.00   0.04    0.00     0.9817
trt                -0.19   0.05   12.25     0.0005   0.825   0.741   0.919
beta blocker        0.03   0.05    0.30     0.5817
trt*betablocker    -0.07   0.08    0.75     0.3879
trt                -0.23   0.04   33.96    <0.0001   0.798   0.739   0.861
lipid              -0.36   0.04   71.01    <0.0001
trt                -0.24   0.05   24.62    <0.0001   0.789   0.719   0.867
lipid              -0.37   0.06   43.72    <0.0001
trt*lipid           0.03   0.08    0.15     0.6952
trt                -0.23   0.04   34.06    <0.0001   0.797   0.739   0.860
antiplatelet       -0.06   0.05    1.47     0.2259
trt                -0.47   0.09   24.63    <0.0001   0.626   0.520   0.753
antiplatelet       -0.19   0.07    8.04     0.0046
trt*antiplatelet    0.29   0.10    7.97     0.0048
trt                -0.23   0.04   33.88    <0.0001   0.798   0.740   0.861
diabetes            0.32   0.04   53.92    <0.0001
trt                -0.21   0.05   20.56    <0.0001   0.808   0.737   0.886
diabetes            0.34   0.06   35.24    <0.0001
trt*diabetes       -0.04   0.08    0.24     0.6271
trt                -0.23   0.04   34.29     0.0000   0.797   0.739   0.860
hypertension        0.24   0.04   35.50     0.0000
trt                -0.20   0.05   15.86     0.0001   0.815   0.737   0.901
hypertension        0.26   0.05   24.35    <0.0001
trt*hypertension   -0.05   0.08    0.47     0.4946
trt                -0.23   0.04   34.54    <0.0001   0.796   0.738   0.859
cholesterol        -0.33   0.04   69.57    <0.0001
trt                -0.22   0.06   13.87     0.0002   0.801   0.713   0.900
cholesterol        -0.32   0.05   37.26    <0.0001
trt*cholesterol    -0.01   0.08    0.02     0.8872
trt                -0.23   0.04   33.70    <0.0001   0.798   0.740   0.862
CAD                 0.20   0.05   15.72     0.0001



                                          80 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

trt                 -0.22     0.09     6.55     0.0105    0.799     0.672      0.949
CAD                  0.20     0.07     8.92     0.0028
trt*CAD              0.00     0.10     0.00     0.9989
trt                 -0.23     0.04    33.78    <0.0001    0.798     0.740      0.861
TIA                  0.52     0.08    38.35    <0.0001
trt                 -0.22     0.04    31.26    <0.0001    0.800     0.740      0.865
TIA                  0.54     0.11    23.10    <0.0001
trt*TIA             -0.04     0.17     0.05     0.8189

With the exception of anti-platelet, NI margins adjusting for other covariates vary from
1.03 to 1.08.


Figure 19 Hazard ratios adjusted by covariates in integrated HOPE and EUROPA
trials




                                              81 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 20 NI margins based on hazard ratio adjusted for covariates in integrated
EUROPA and HOPE trials




The same covariate-adjusted analyses were done for the PEACE trial. In order to
perform the analyses, the reviewer re-constructed the 3-fold composite (HOPE)
endpoint. Due to lack of fatal stroke data, the 3-fold endpoint is not exactly the 3-fold
endpoint used in HOPE trial. The hazard ratio point estimate of trandolapril over
placebo is 0.93 with 95% CI (0.81, 1.07) without any covariate adjustment. None of the
interaction term is significant (Table 37).
The estimate of hazard ratio of trandolapril over placebo is 0.71 if all the covariates and
two way treatment*covariate interactions are included into the Cox proportional hazard
model. The corresponding 95% confidence interval is (0.26, 1.92). There will not be any
NI margins based on PEACE trial alone since the trial failed to show superiority to
placebo.


Table 37 Analyses on PEACE Trial by Adjusting Covariates (reconstructed 3-fold
endpoint)
                                                             Hazard      Lower       Upper
Variable          Estimate    StdErr   ChiSq     p-value     Ratio       limit       limit
trandolapril          -0.07     0.07      0.98      0.3221       0.933       0.812       1.071
Female                -0.11     0.09      1.27      0.2606
trandolapril          -0.05     0.08      0.34      0.5591      0.956       0.822       1.112
Female                -0.03     0.13      0.06      0.8065
tran*female            0.15     0.19      0.61      0.4341



                                                 82 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

trandolapril        -0.08   0.07    1.21     0.2707   0.925   0.806   1.062
age60                0.53   0.08   42.15    <0.0001
trandolapril        -0.17   0.14    1.44     0.2305   0.844   0.639   1.114
age60                0.47   0.11   17.96    <0.0001
tran*age60           0.12   0.16    0.57     0.4520
trandolapril        -0.07   0.07    0.95     0.3296   0.934   0.813   1.072
MI                  -0.16   0.07    4.76     0.0291
trandolapril        -0.08   0.21    0.13     0.7143   0.925   0.608   1.406
MI                  -0.16   0.10    2.54     0.1112
tran*MI              0.01   0.14    0.00     0.9621
trandolapril        -0.07   0.07    1.10     0.2951   0.929   0.809   1.066
Stroke              -0.34   0.15    5.04     0.0247
trandolapril        -0.04   0.59    0.00     0.9486   0.963   0.304   3.049
Stroke              -0.33   0.21    2.34     0.1259
tran*stroke         -0.02   0.30    0.00     0.9512
trandolapril        -0.07   0.07    0.92     0.3373   0.935   0.814   1.073
CABG                -0.16   0.07    5.28     0.0215
trandolapril         0.07   0.24    0.09     0.7588   1.076   0.676   1.712
CABG                -0.12   0.10    1.50     0.2208
tran*CABG           -0.09   0.14    0.38     0.5356
trandolapril        -0.07   0.07    1.07     0.3017   0.930   0.810   1.067
beta blocker        -0.08   0.07    1.09     0.2964
trandolapril         0.12   0.21    0.32     0.5728   1.127   0.744   1.709
beta blocker        -0.01   0.10    0.01     0.9268
tran*betablocker    -0.14   0.14    0.92     0.3369
trandolapril        -0.07   0.07    1.10     0.2942   0.929   0.809   1.066
Lipid                0.39   0.07   28.75    <0.0001
trandolapril        -0.16   0.21    0.58     0.4465   0.851   0.561   1.290
Lipid                0.36   0.10   12.42     0.0004
tran*lipid           0.06   0.14    0.19     0.6617
trandolapril        -0.08   0.07    1.17     0.2787   0.927   0.807   1.064
antiplatelet         0.34   0.11   10.30     0.0013
trandolapril         0.00   0.25    0.00     0.9845   1.005   0.616   1.639
antiplatelet         0.38   0.15    6.38     0.0115
tran*antiplatelet   -0.07   0.21    0.11     0.7349
trandolapril        -0.08   0.07    1.30     0.2542   0.923   0.804   1.059
Diabetes            -0.56   0.08   48.15    <0.0001
trandolapril         0.09   0.29    0.09     0.7619   1.093   0.615   1.941
Diabetes            -0.52   0.12   19.76    <0.0001
tran*diabetes       -0.10   0.16    0.35     0.5525
trandolapril        -0.07   0.07    1.12     0.2896   0.928   0.809   1.065
hypertension        -0.37   0.07   27.07    <0.0001
trandolapril        -0.24   0.22    1.20     0.2731   0.788   0.514   1.207
hypertension        -0.42   0.10   18.35    <0.0001
tran*hypertension    0.11   0.14    0.63     0.4257



                                           83 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

trandolapril       -0.07   0.07    1.05     0.3054   0.930   0.811     1.068
cholesterol         0.01   0.07    0.03     0.8652
trandolapril       -0.04   0.09    0.20     0.6583   0.961   0.808     1.145
cholesterol         0.05   0.10    0.29     0.5895
tran*cholesterol   -0.09   0.15    0.36     0.5479
trandolapril       -0.07   0.07    1.09     0.2960   0.929   0.809     1.066
CAD                -0.12   0.11    1.19     0.2751
trandolapril        0.04   0.21    0.04     0.8327   1.045   0.692     1.579
CAD                -0.05   0.16    0.10     0.7464
tran*CAD           -0.13   0.22    0.35     0.5515
trandolapril       -0.08   0.07    1.19     0.2747   0.926   0.807     1.063
TIA                -0.69   0.14   22.71    <0.0001
trandolapril        0.64   0.58    1.23     0.2665   1.901   0.612     5.905
TIA                -0.48   0.23    4.41     0.0358
tran*TIA           -0.37   0.30    1.57     0.2097



Figure 21 Hazard ratio of perindopril over placebo adjusted for each covariate in
PEACE




                                          84 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Table 38 shows the covariate analysis in ONTARGET. The hazard ratios, after adjusting
for covariates and interactions, do not differ much. Particularly, most of the upper
bounds of the 95% confidence intervals are above 1.08.


Table 38 Analyses of ONTARGET Trial by Adjusting Covariates (HOPE endpoint)

                                                              Hazard       Lower       Upper
Variable           Estimate    StdErr   ChiSq     p-value     ratio        limit       limit
telmisartan            -0.01     0.04      0.13      0.7235        0.986       0.910       1.068
female                 -0.10     0.05      4.13      0.0420
telmisartan             0.00     0.05      0.01      0.9253       0.996       0.908       1.092
female                 -0.08     0.07      1.35      0.2460
telm*female            -0.04     0.09      0.18      0.6710
telmisartan            -0.02     0.04      0.16      0.6868       0.984       0.908       1.066
age60                   0.44     0.05     68.00    <0.0001
telmisartan            -0.04     0.10      0.21      0.6452       0.956       0.790       1.157
age60                   0.43     0.07     32.56    <0.0001
telm*age60              0.03     0.11      0.10      0.7482
telmisartan            -0.02     0.04      0.14      0.7098       0.985       0.909       1.067
MI                      0.13     0.04     10.85      0.0010
telmisartan            -0.02     0.06      0.07      0.7915       0.985       0.877       1.105
MI                      0.13     0.06      5.44      0.0197
telm*MI                 0.00     0.08      0.00      0.9926
telmisartan            -0.01     0.04      0.07      0.7952       0.989       0.913       1.072
stroke/TIA              0.44     0.05     94.84    <0.0001
telmisartan            -0.01     0.05      0.07      0.7881       0.987       0.898       1.085
stroke/TIA              0.44     0.06     47.25    <0.0001
telm*strokeTIA          0.01     0.09      0.01      0.9269
telmisartan            -0.01     0.04      0.13      0.7181       0.985       0.910       1.067
CABG                    0.12     0.05      6.09      0.0136
telmisartan            -0.02     0.05      0.23      0.6316       0.978       0.892       1.072
CABG                    0.10     0.07      2.24      0.1348
telm*CABG               0.03     0.10      0.11      0.7370
telmisartan            -0.01     0.04      0.11      0.7366       0.986       0.911       1.069
beta blocker           -0.06     0.04      2.01      0.1566
telmisartan            -0.08     0.06      1.80      0.1801       0.921       0.817       1.039
beta blocker           -0.12     0.06      4.25      0.0394
telm*betablocker        0.12     0.08      2.25      0.1340
telmisartan            -0.01     0.04      0.07      0.7848       0.989       0.913       1.071
lipid                  -0.26     0.04     39.47    <0.0001
telmisartan             0.04     0.06      0.40      0.5268       1.041       0.919       1.179
lipid                  -0.22     0.06     13.78      0.0002
telm*lipid             -0.09     0.08      1.11      0.2924


                                                  85 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

    telmisartan                    -0.01   0.04    0.11      0.7392   0.987   0.911   1.069
    Antiplatelet†                  -0.15   0.05   10.62      0.0011
    Telmisartan                    -0.03   0.08    0.12      0.7241   0.973   0.834   1.134
    Antiplatelet†                  -0.16   0.06    6.10      0.0136
    telm*antiplatelet†              0.02   0.09    0.04      0.8332
    telmisartan                    -0.02   0.04    0.21      0.6504   0.982   0.906   1.063
    diabetes                        0.35   0.04   74.57      0.0000
    telmisartan                     0.00   0.06    0.00      0.9456   1.004   0.901   1.118
    diabetes                        0.38   0.06   42.98     <0.0001
    telm*diabetes                  -0.05   0.08    0.36      0.5476
    telmisartan                    -0.01   0.04    0.09      0.7676   0.988   0.912   1.070
    hypertension                    0.35   0.05   54.38     <0.0001
    telmisartan                     0.02   0.08    0.04      0.8333   1.017   0.866   1.195
    hypertension                    0.37   0.07   30.09     <0.0001
    telm*hypertension              -0.04   0.09    0.17      0.6803
    telmisartan                    -0.01   0.04    0.13      0.7222   0.986   0.910   1.068
    cholesterol                     0.23   0.04   31.94     <0.0001
    telmisartan                    -0.01   0.05    0.08      0.7822   0.985   0.887   1.095
    cholesterol                     0.23   0.06   16.03      0.0001
    telm*cholesterol                0.00   0.08    0.00      0.9928
    telmisartan                    -0.01   0.04    0.11      0.7369   0.986   0.911   1.069
    CAD                             0.02   0.04    0.15      0.6958
    telmisartan                    -0.02   0.07    0.05      0.8161   0.984   0.855   1.132
    CAD                             0.01   0.06    0.06      0.8085
    telm*CAD                        0.00   0.09    0.00      0.9602
†
antiplatelet use as aspirin only




We then examined also the superiority test of the combination of telmisartan plus
ramipril versus ramipril monotherapy in ONTARGET. In essence, the combination of
telmisartan and ramipril performs similarly as ramipril alone (HR 0.99; 95% CI 0.92,
1.07; p=0.8462), which is also consistently shown in the sensitivity analyses (Table
12/Table 39).




                                                          86 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 39 Analysis of ONTARGET primary endpoint for superiority (T/R vs R)
                                 T/R           R        HR (95% CI) / P value
Randomized, n (%)            8,502 (100) 8,576 (100)               ---
                                 FAS Population
Primary endpoint, n (%)      1,386 (16.3) 1,412 (16.5) 0.99 (0.92, 1.07) / 0.8462
Events per 100 patient-years     3.79         3.82
                                 PPS Population
Primary endpoint, n (%)      1,095 (12.9) 1,164 (13.6) 1.00 (0.93, 1.09) / 0.9093
Events per 100 patient-years     3.60         3.58
                                FAS-10 Population
Primary endpoint, n (%)               1,094 (12.9)        1,183 (13.8)
                                                                            1.00 (0.92, 1.09) / 0.9926
Events per 100 patient-years              3.72                3.72
 Sensitivity analysis accounting for CHF-hospitalizations confirmed by chest X-ray
Primary endpoint, n (%)        1,323 (15.6)  1,359 (15.8)
                                                            0.98 (0.91, 1.06) / 0.6738
Events per 100 patient-years       3.61          3.67
Sensitivity analysis accounting for SBP at BL, & SBP change from BL, time-dependent
Primary endpoint, n (%)        1,386 (16.3) 1,412 (16.5)    1.00 (0.93, 1.08) / 0.9553
Events per 100 patient-years       3.79          3.82
Sensitivity analysis accounting for SBP at BL and SBP change for BL before event
Primary endpoint, n (%)      1,386 (16.3) 1,412 (16.5)   0.99 (0.92, 1.07) / 0.8111
Events per 100 patient-years     3.79         3.82
Source: Sponsor’s Tables 15.2.1.1:1, 15.2.1.1:3, 15.2.1.1:5, 15.2.1.1:7 & 15.2.1.1:8 on pages 541, 544,
547, 550 - 552 of ONTARGET Clinical Trial Report; BL = baseline



The sponsor used 95% confidence interval in the clinical study report. This did not take
into account the multiple comparisons used in the trial. Even if we consider all other
analyses (different populations, with various covariates) as sensitivity analyses to show
the robustness of the primary analysis, there are at least two important comparisons for
the primary analysis: the superiority test of combination versus ramipril, and the non-
inferiority test of telmisartan versus ramipril. One simple way of multiple comparison
adjustment is to split the alpha of 0.05 for the two tests. So, strictly speaking, a 97.5%
confidence interval should be used. The sponsor did report the non-inferiority results
using the 97.5% confidence intervals.


Step 4: Evaluation of supportive trials (TRANSCEND and PRoFESS)
The sponsor’s analyses on TRANSCEND and PRoFESS showed that neither
TRANSCEND nor PRoFESS trial on its own demonstrated superiority of T over PBO for
the protocol-specified primary endpoint (Table 16/Table 40 and Table 17/Table 41).




                                                   87 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 40 Analysis of the primary composite endpoint in TRANSCEND (FAS)




Source: Table 15.2.1:1 on page 403 of TRANSCEND Clinical Trial Report (Verified by the reviewer)


Table 41 Cox proportional analysis of recurrent stroke in PRoFESS (RAN set)




Source: Table 15.2.2.1: 2 on page 514 of PRoFESS Clinical Study Report. (Verified by reviewer)

The sponsor pooled data from the two placebo-controlled TRANSCEND and PRoFESS
trials to obtain a population of 16,877 patients who did not use an ACE-I at any time
during the trials (the total/NoACE-I population), with 8,587 patients who received
telmisartan and 8,290 patients who received placebo. This pooled population of patients
in the two supportive trials appears adequate to evaluate telmisartan compared to
placebo for its efficacy to reduce the risk of CV events, but we have to remember that
this is post-hoc.
The sponsor’s analysis of this pooled Total/NoACE-I population showed a nominally
significant (p=0.0107) reduction in the relative risk of the primary endpoint by 10% (HR
0.90, 95% CI 0.83, 0.98) in patients randomized to T vs PBO (Table 42).
In the pooled analyses, non-fatal stroke was the outcome that had the highest
contribution to the composite endpoint which is different from that in ONTARGET where
CV death was the largest contributor to the primary composite endpoint. However, this
is post-hoc analysis.


                                                  88 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 42 Summary of Analyses in TRANSCEND and PRoFESS
                    TRANSCEND (telmisartan vs placebo)
                                T        PBO          HR (95% CI) / P value
Randomized, n (%)                     2,954 (100)       2,972 (100)                    ---
4-fold Primary endpoint, n (%)         465 (15.7)        504 (17.0)        0.92 (0.81, 1.05) / 0.2192
3-fold Primary endpoint, n (%)         384(13.0)         440 (14.8)        0.87 (0.76, 1.00) / 0.0483
                              PRoFESS (telmisartan vs placebo)
                                      T            PBO         HR (95% CI) / P value
Randomized, n (%)                    10,146 (100)       10,186 (100)                   ---
4-fold Primary endpoint, n (%)       1,367 (13.5)       1,463 (14.4)       0.94 (0.87, 1.01) / 0.1073
3-fold Primary endpoint, n (%)       1,289 (12.7)       1,377 (13.5)       0.94 (0.87, 1.02) / 0.1260
    TRANSCEND & PRoFESS (telmisartan vs placebo) Total/NoACE-I‡ population
                             T             PBO         HR (95% CI) / P value
Randomized, n (%)                    8,587 (100)        8,290 (100)                    ---
4-fold Primary endpoint, n (%)       1,120 (13.0)       1,196 (14.4)       0.90 (0.83, 0.98) / 0.0107
3-fold Primary endpoint, n (%)       1,015 (11.8)       1,106 (13.3)       0.88 (0.81, 0.96) / 0.0029
‡
 Patients in TRANSCEND and PRoFESS trials who did not take an ACE-I during the trials. 

Source: Table 11.4.1.2.1: 1 on page 184 of TRANSCEND Clinical Trial Report, Table 11.B4.1.2.1: 1 on 

page 205 and Table 11.B4.1.2.2: 1 on page 208 of PRoFESS Clinical Trial Report. (Results above are 

verified by the reviewer) 


Also, there is a multiple comparison issue. The analysis of pooled TRANSCEND and
PRoFESS data is in addition to the primary analyses of the individual trials, both of
which failed. The post hoc analyses did not take multiplicity adjustment into
consideration. Also, the analyses did not account for all possible ways to win.
The failure of T to win over PBO in the TRANSCEND trial – which contemporaneously
enrolled patients with similar CV risk as in ONTARGET – does not suggest
interchangeability of T for R to support the non-inferiority of T to R. This further confirms
our assertion that changes in the standard of care have altered clinical event rates in
this patient population so that the constancy assumptions do not hold any more.



Summary of statistical considerations
ONTARGET did not show that the combination of telmisartan and ramipril is superior to
ramipril in treating patients with CV risks. In fact, the effects of the combination (T/R)
and ramipril (R) alone are almost identical. Although the sponsor claimed that
telmisartan alone is non-inferior to ramipril, their non-inferiority test is based on the NI
margin proposed by the sponsor (1.13).
This margin of 1.13 seems too liberal. First, the rationale and the clinical meaning of the
NI margin derived based on 20% confidence interval are unclear. Second, the NI margin
is based on a single (HOPE) trial, so between-trial variability cannot be estimated. Third,
the HOPE trial was conducted 10 years ago when clinical practice was different from


                                                 89 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

current clinical practice. This is evident from the placebo event rate which is significantly
lower in TRANSCEND trial compared to HOPE trial. The constancy assumption that the
non-inferiority analysis relies upon may no longer hold.
We explored a number of clinical covariates. Most of the interaction terms of treatment
and clinical covariates are not significant. In HOPE trial, the interaction of
treatment*anti-platelet appears to be significant. The treatment*anti-platelet appears
also to be significant in pooled EUROPA and HOPE trials.
The NI margin, after adjusting the covariates, varies from 1.02 to 1.10 based on HOPE
trial alone with the exception of the covariate anti-platelet. Including covariate anti-
platelet into the model seems to favor ramipril more and the NI margin might be 1.15 as
a result. Similar results are observed in pooled EUROPA and HOPE trials.
The EUROPA trial shows smaller treatment effect when adjusting different covariates.
By including some covariates, e.g., history of MI, history of CAD, and use of β-blockers,
the confidence interval of the adjusted hazard ratio estimate covers 1. The treatment no
longer has a significant effect over placebo in these cases; therefore a NI margin cannot
be defined any more. Nevertheless, if the NI margin exists, it varies from 1.02 to 1.06
depending on the covariate included in the model. In most cases, the NI margins
adjusted for covariates are below 1.10.
In ONTARGET, the upper limit of the confidence interval of the hazard ratio point
estimate varies from 1.07 to 1.20 if adjusting for each covariate (with exception of use of
β- blockers). Thus, in almost all analyses with or without covariate adjustment, the NI
margin derived by discounting 50% on the worst limit of 95% confidence interval for the
effect of ramipril over placebo cannot be ruled out by the non-inferiority analysis in
ONTARGET.
In essence, various covariates do not seem to affect much the treatment effect in
EUROPA, HOPE or ONTARGET trials with the exception of use of anti-platelets in
HOPE trial.
Neither TRANSCEND nor PRoFESS trial on its own shows superiority of T over PBO
for the protocol-specified primary endpoint. TRANSCEND appears to show nominal
significance for the 3-fold composite (HOPE) endpoint. The sponsor also performed a
post-hoc pooled analysis of PRoFESS and TRANSCEND and showed a nominally
significant reduction in the relative risk of the 3-fold composite (HOPE) or 4-fold
composite (ONTARGET) endpoints. However, the multiple comparisons which resulted
from such post hoc analyses were never adjusted. Also, the analyses did not account
for all possible ways to win.
Will defer to comment until after the AC meeting.




                                             90 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


6.1.7   Other Endpoints
The following endpoints of clinical interest will be mentioned briefly.

ONTARGET trial
All-cause mortality: The incidence of all-cause mortality was 12.5% in the T/R group,
11.6% in the T group, and 11.8% in the R group. The hazard ratio for T/R vs. R was
1.07 (95% CI 0.98, 1.16; p=0.1453); and for T vs. R, it was 0.98 (95% CI 0.90, 1.07;
p=0.6378). The between-group difference was not statistically significant. The
predominant reasons for all-cause death were CV death and malignancies.
Renal endpoints: Treatment with T/R compared to R significantly reduced the risk for (i)
new microalbuminuria, (ii) new macroalbuminuria, and (iii) the composite of doubling of
serum creatinine, progression to ESRD, new microalbuminuria, or new
macroalbuminuria.
Also, significantly more patients with micro- or macroalbuminuria who were treated with
T/R compared to R, improved to normo-albuminuria. In these patients with micro- or
microalbuminuria, the hazard ratio of T/R vs. R was 0.88 for microalbuminuria (95% CI
0.80, 0.97), 0.81 for macroalbuminuria (95% CI 0.67, 0.97), 0.90 for the composite
endpoint (95% CI 0.83, 0.98), and 1.19 for normoalbuminuria (95% CI 1.05, 1.35).
TRANSCEND trial
The incidence of all-cause mortality was 12.3% in the telmisartan group and 11.7% in
the placebo group. The hazard ratio of telmisartan vs. placebo was 1.05 (95% CI 0.91,
1.22). The predominant reasons for death were CV death and malignancies.
PRoFESS trial
During the study, 1.7% (telmisartan) and 2.1% (placebo) of patients without diabetes at
baseline were newly diagnosed with diabetes mellitus (HR 0.82, 95% CI 0.65, 1.04).


6.1.8   Subpopulations
ONTARGET trial
30 subgroup analyses of T vs R were performed (Figure 22).




                                             91 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 22 Comparison of T vs R in subgroups (ONTARGET) / FAS




1
 β-blockers; p = 0.0139; Proportion of visits with BP < 140/90 mmHg: p = 0.0018.
Source: Figure 11.4.1.1.1:2 on page 202 of ONTARGET Clinical Study Report



                                                  92 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The subgroup analysis included core subgroups such as age, sex, and ethnicity, for
intrinsic risk factors including diabetes, for extrinsic risk factors, for medical history such
as MI and stroke, and for concomitant medication use. There was no subgroup-by­
treatment interaction observed for the subgroups with the exception of patients who
were taking β-blockers.


FDA reviewers’ comments: About 57% of patients in each of the randomization
groups were taking β-blockers (Table 49). In patients who were NOT taking β-blockers,
T has a beneficial effect; in contrast, in patients taking β-blockers, a beneficial effect
was seen only with ramipril (Figure 22).
The lack of effect of ARBs in patients who were taking β-blockers was also found in
Valsartan Heart Failure Trial (ValHeFT)10: in the subgroup of patients an ACE-I or β­
blockers, valsartan treatment had a modest benefit on CV morbidity; in patients taking
both β-blockers and ACE-I, valsartan had an adverse effect on morbidity and mortality,
whereas valsartan conferred the greatest benefit for patients who were not taking an
ACE-I or a β-blocker, a finding similar to that in ONTARGET. It is possible that the
potential benefits of combination T/R were blunted by concomitant β-blocker therapy11,
leading to lack of superiority of the T/R combination over R alone.


6.1.9   Analysis of Clinical Information Relevant to Dosing Recommendations
In the ONTARGET, TRANSCEND and PRoFESS trials, only a single dose of
telmisartan (80 mg/day) and a singled dose of ramipril (10 mg/day) were used. The
relationship of drug vs dose or drug concentration vs response was not investigated.
FDA reviewers’ comments: At this time, uncertainties regarding use of the optimal
dose of ACE-Is and ARBs in the treatment to reduce CV events (as perceived by
general practitioners and cardiologists) remain an unresolved issue in clinical practice.
The general consensus of opinion is that for ACE-Is to achieve a reduction in
cardiovascular mortality and morbidity, larger than standard doses used for the control
of hypertension12, and probably at the full doses that were shown in clinical trials (e.g.,
HOPE, EUROPA, etc.), must be used. {Please see Section 9.1 Literature
Review/References, 9.1.2 Dose considerations for Angiotensin Converting Enzyme
Inhibitors (ACE-Is)}.
For ARBs, too, a survival benefit is found only when higher doses than that for the
treatment of hypertension are used. It appears that an insufficient dose of ARBs used in
ELITE II13, OPTIMAAL14, and VALIANT15 trials led to lack of beneficial effect, whereas
use of relatively larger doses of ARBs led to a significant survival benefit in high risk
patients in LIFE16 and RENAAL17 trials. {Please see Section 9.1 Literature
Review/References, 9.1.3 Dose considerations for Angiotensin Receptor Blockers
(ARBs)}.



                                              93 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

6.1.10 Discussion of Persistence of Efficacy and/or Tolerance Effects
Table 43 shows that the treatment effect persisted over time, based on the number of
patients who were at risk, had an event or were censored by the year of the study. This
was supported by the Kaplan-Meier curves which were linear until patient numbers
decreased at about 5 years of follow-up.

Table 43 Frequency of composite primary endpoint by year (FAS)




1
  Number of patients entering the respective time interval 

2
  All effort was made to collect information about the primary outcome events for patients lost to follow-up. 

In case of no contact, the patient was censored on the last day of available contact during the study. 

3
  The yearly interval was defined as 365 days in this CTR. The dates of telephone contacts (which may 

have differed from the visit dates) were also considered in the calculation of the date when a patient was 

still seen alive, which may be different from the publication which considered only the visit dates. 


To determine persistency of efficacy (non-inferiority of T vs R) of the 4-fold composite
primary endpoint, the hazards ratios with CI over time (using the fixed effects model) for
the FAS, PPS, FAS-10 and PPS without ACE-Is/ARBs population sets are show in
Figure 23, Figure 24, Figure 25, and Figure 26.

Figure 23 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint
over time (FAS)




                                                                       NI Margin 1.13

                                                                        NI Margin 1.08




                                                     94 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Figure 24 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint
over time (PPS)




                                                       NI Margin 1.13


                                                       NI Margin 1.08




Figure 25 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint
over time (FAS–10)




                                                      NI Margin 1.13


                                                      NI Margin 1.08




                                       95 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 26 Hazard ratios (T vs R) and CIs for 4-fold composite primary endpoint
over time (PPS–without ACE-Is/ARBs)




                                                           NI Margin 1.13

                                                           NI Margin 1.08




FDA reviewers’ comments: The hazard ratios and CIs of the cumulative data of the 4­
fold composite primary endpoint for the above data sets show that at 5 years and at end
of study time-points, non-inferiority is found if the sponsor’s NI margin of 1.13 is used,
but non-inferiority is not demonstrated if the reviewers’ NI margin of 1.08 is used.

To determine patient outcome at follow up (for patients who had MI or Stroke or CHF
hospitalization as a primary outcome event), we evaluated the hazard ratios (95% CIs)
of the follow-up clinical outcome data at 6 months and cumulatively at Year 1, 2, 3, 4,
and 5 for the comparisons of T/R vs R and T vs R, respectively, of patients (FAS) who
were adjudicated as having reached the primary efficacy endpoint of:
(i) non-fatal MI (Table 72 and Table 73),
(ii) non-fatal stroke (Table 74 and Table 75), or
(iii) CHF hospitalization (Table 76 and Table 77), which are presented later in this
      review in “Section 9.4.1: Non-essential tables of analyses of follow-up clinical
      outcome data at 6 months and cumulatively at Year 1, 2, 3, 4, and 5.”


6.1.11 Additional Efficacy Issues/Analyses

The sponsor alluded to a meta-analysis of clinical trials other ACE-inhibitors (SOLVD18,
SAVE19, AIRE20 and TRACE21 trials), from which they derived their NI margin based on
the draft CPMP guidance22 and a ‘putative placebo approach’ (abstracted from
ONTARGET Protocol and Appendix C: Statistical Considerations).


                                           96 

    Clinical Review
    Khin Maung U, M.D. & Jialu Zhang, Ph.D.
    NDA 20-850 Efficacy Supplement SE1-025
    Micardis® (telmisartan) tablets

    Our literature review reveals that these were clinical trials (Table 44) of ACE inhibitors in
    patients (i) in the post MI period (3 – 16 days), (ii) with some degree of heart failure and
    reduced LVEF, (iii) with large variations in the duration of treatment (15 to 50 months),
    and (iv) had different endpoints than those studied in the ONTARGET trial. We do not
    agree that these trials can be used in a meta-analysis to determine the NI margin for the
    ONTARGET trial.


    Table 44 ACE inhibitor vs Placebo trials in Post-MI Left Ventricular dysfunction
   Study (yr      N      Study Population          ACE inhibitor /      Concomitant      Duration      Mortality / end point
  published)                                          Placebo            medication                            findings
  SOLVD18        2,569   LVEF ≤35%, NYHA           Enalapril(1285)                       41.1 mo     ↓ 16% all cause death
  (1992)                 II & III heart failure,   Placebo (1284)                                    (P=0.0036), ↓ 26% CHF
                         asymptomatic                                                                hospitalization (P<0.0001)
  SAVE19         2,231   Post MI 3-16 d,           Captopril            β-blockers,      4 yr        ↓ 19% all cause death
  (1992)                 LVEF ≤40%,                (1115) Placebo       thrombolytics,               (P<0.019), ↓ 21%CV
                         asymptomatic              (1116)               aspirin                      death (P<0.014)
      20
  AIRE           2,006   NYHA II-III, post MI      Ramipril (1014)                       15 mo       ↓ 27% all cause death
  (1993)                 3-10 d, overt signs       Placebo (992)                                     (P<0.02), ↓ 19% death/HF
                         of heart failure                                                            /MI/stroke (P<0.008)
           21
  TRACE          1,749   Post MI 3-7 d, Echo       Trandolapril                          24 – 50     ↓ 22% all cause death
  (1995)                 LVEF ≤35%,                (876)                                 mo          (P<0.001), ↓ 25% CV
                         Danish                    Placebo (873)                                     death, (P=0.03), ↓ 24%
                                                                                                     sudden death (P=0.003)



    The characteristics of patients in the early ACE-I trials are shown in Table 45.


    Table 45 ACE inhibitor trials in patients with cardiovascular risk factors
 Study (yr        N       Study Population              Agents           Design /     Primary         End point findings
published)                                                              Duration      Endpoint
  HOPE          9,297    ≥55 yr, history of CAD,      Ramipril          R, DB,       Combined       22%↓ CV death, MI &
  (2000)                 stroke, PVD or               10mg/d            PC, 2X2      MI, stroke     stroke (P<0.001); 16%↓
                         diabetes plus ≥ 1 other      vs.               factorial    and CV         all-cause death, 23%↓
                         CV risk factor               placebo           4.5 yr       mortality      CHF, 16%↓ DM
                                                                                                    complications; 32%↓
                                                                                                    stroke, 61%↓ fatal stroke
EUROPA          12,218   69±9 yr; 64% previous        Perindopri        R, DB,       Combined       20% ↓ CV death, MI or
  (2003)                 MI, 61% CAD by               l 8 mg/d          PC, MC       CV death,      cardiac arrest (P=0.0003)
                         angiography 55%              vs.               4.2 yr       MI or
                         coronary                     placebo                        cardiac
                         revascularization; 5%
                         +ve stress test
                                                                                     arrest
PEACE           8,290    64±8 yr, stable CAD          Trandolapril      R, DB,       CV death/      No difference in primary
  (2004)                 (72% coronary                4mg/d vs.         PC           MI/coronary    efficacy endpoint events
                         revascularization,70%        placebo           Median       revasculari­
                         on lipid lowering drugs)                       4.8 yr       zation




                                                                 97 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 46 Meta-analysis of mortality data in HOPE, EUROPA and PEACE trials23




        Source: Yusuf S, Pogue J. New Engl J Med (Letter) 2005; 352: 937-8.


For the most important single endpoint of CV death, the treatment effect in the HOPE
trial is the largest and statistically significant. In EUROPA and PEACE, CV deaths are
not statistically significant (Table 46).
Even in the earlier HOPE, EUROPA and PEACE trials, a substantial disparity in
treatment effect in the placebo group is found (Table 47). After a mean follow-up of 4.5
years, HOPE reported a total placebo group mortality of 12.2%, CV mortality of 8.1%
and Q-wave MI of 3.2%, against lower event rates of 7.4%, 4.4% and 2.1%,
respectively, in the EUROPA trial (placebo outcomes are standardized for 4.5 years
follow-up), and relatively lower event rates of 8.1%, 3.7% and 5.3%, respectively, in the
PEACE trial. Thus, the major event rates in HOPE were 50% to 80% more than in
EUROPA or PEACE.
The most recent of the three clinical outcome trials of ACE-Is, PEACE, was negative for
CV benefit. This was explained partly by the fact that 70% of patients were taking lipid
lowering drugs, >90% were taking aspirin, and 72% had undergone prior coronary
revascularization (Table 49).


Table 47 Placebo outcomes in HOPE, EUROPA, and PEACE trials.
                                         HOPE           EUROPA           PEACE
Study drug                              Ramipril       Perindopril    Trandolapril
Total randomized, N                       9,297          12,218           9,290
(ACE-I :Placebo)                     (4,645 : 4,652) (6,110 : 6,108) (4,158 : 4,132)
Total mortality in placebo group (%)       12.2            7.4             8.1
CV mortality in placebo group (%)          8.1             4.4             3.7
Q wave MI in placebo group (%)             3.2             2.1             5.3
Source: FDA reviewers’ analysis of data in NDAs and literature.



The mean duration of follow-up in ONTARGET is 4.6 years, which is comparable to that
of 4.5 years in HOPE.


                                                   98 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

For the HOPE, ONTARGET and TRANSCEND trials, we found the annualized event
rates of the primary 4-fold endpoint (using the definition of CHF hospitalization
confirmed by chest X-ray to have a more valid comparison to HOPE), its single
components, the 3-fold HOPE primary endpoint and overall death (Table 48) as follows:
•	 For the primary 4-fold endpoint: the event rates appear to have decreased over time
    in the placebo group (3.72% in TRANSCEND vs. 5.09% in HOPE), and in the
    ramipril group (3.67% in ONTARGET vs. 3.95% in HOPE),
•	 For the primary 3-fold endpoint: the event rates also appear to have decreased over
    time in the placebo group (3.34% in TRANSCEND vs. 4.53% in HOPE), and in the
    ramipril group (3.23% in ONTARGET vs. 3.51% in HOPE).


Table 48 Annualized event rates in HOPE, ONTARGET and TRANSCEND trials
                            HOPE                   ONTARGET                     TRANSCEND
    Endpoint         Placebo     Ramipril    Ramipril     Telmisartan     Placebo      Telmisartan
Primary 4-fold         5.09       3.95        3.82           3.87           3.87          3.58
Primary 4-fold*        5.09       3.95        3.67           3.66           3.72          3.34
Hope (3-fold)          4.53       3.51        3.23           3.18           3.34          2.92
All-cause death        2.90       2.40        2.60           2.54           2.53          2.66
CV death               1.95       1.45        1.54           1.54           1.61          1.66
MI                     3.06       2.44        1.07           1.15           1.09          0.85
Stroke                 1.19       0.81        1.06           0.97           1.01          0.83
CHF Hosp.*             0.84       0.73        0.70           0.74           0.70          0.67
*required confirmation by chest X-ray; Source: sponsor’s response to FDA, submitted 14-Apr-2009



There appears to be a relatively large reduction in the event rates of MI, compared to
other components of the composite primary endpoint events in HOPE. In HOPE, the
placebo group had a relatively higher annualized event rate 3.06% for MIs). In
ONTARGET and TRANSCEND, the event rates for MIs have decreased >60% to about
1% (Table 48). Note also that the event rate for MI in the ramipril groups, too, has
decreased >50% from 2.44 in HOPE to 1.07 in ONTARGET (Table 48).
We believe that this reduction in the event rates of MI can be attributed in part to
improvement of medical care and increasing use of concomitant effective medications
that have improved CV outcomes during the more recent ONTARGET and
TRANSCEND trials compared to the earlier HOPE trial such as (Table 49).
We found that there has been an increased use of the following therapies in
ONTARGET and TRANSCEND compared to HOPE:
•	 Antiplatelet agents (HOPE: 76% vs. TRANSCEND 93% and ONTARGET: 95%)




                                                 99 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

•	 β-blockers (HOPE: 39% vs. TRANSCEND: 70% and ONTARGET: 57%) {Please
   also see Section 6.1.8 Subpopulations, and our discussion of ARBs showing no
   benefit when patients are taking β-blockers and ACE-Is.}
•	 Lipid lowering agents (HOPE: 29% vs TRANSCEND: 62% and ONTARGET: 65%)
Based on the above findings, we believe that the constancy assumptions do not hold
true. Therefore, for the non-inferiority analysis of the primary efficacy endpoint in
ONTARGET to be based on the HOPE trial with or without the other ACE-I trials
(EUROPA and PEACE) adjusting for the changes in effect size due to improved
standard of care need to be made. A clinical decision is necessary to determine the
acceptable level of effect size from HOPE needed to determine the NI margin to
evaluate non-inferiority of telmisartan vs ramipril. Our proposal to discount 50% of the
lower confidence bound of 1.17 (i.e., NI margin of 1.085) is conservative but probably
justified to ensure that telmisartan will produce at least 50% of the beneficial CV effect
seen with ramipril in the HOPE trial.
We note also that the population of patients in HOPE has difference characteristics from
those in ONTARGET and TRANSCEND (Table 49) including the following:
•	 TIA as qualifying diagnosis in ONTARGET and TRANSCEND,
•	 a greater percentage of patients with a medical history of hypertension ONTARGET
   and TRANSCEND than in HOPE (69% and 77% vs. 47% respectively),
•	 a greater percentage of patients with a medical history of stroke/TIA in ONTARGET
   and TRANSCEND than in HOPE (21% and 22% vs. 11%, respectively), and
•	 fewer patients with a history of chronic coronary artery disease (ONTARGET: 66%
   and TRANSCEND 67% vs. HOPE: 80%).
The above differences in the patient populations, too, may have contributed in part to
the changes in event rates seen in the placebo and ramipril groups (Table 48)
discussed above.




                                            100 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 49 Baseline characteristics and the primary efficacy endpoint results in ONTARGET and TRANSCEND
compared to HOPE, EUROPA and PEACE trials
                                        HOPE (2000)                  EUROPA (2003)                     PEACE (2004)               TRANSCEND (2008)             ONTARGET (2008)
Study drug                                Ramipril                     Perindopril                      Trandolapril                 Telmisartan              Telmisartan vs Ramipril
Population studied               History of CAD, stroke, PVD     History CAD (documented        History of CAD (documented        Same patients as in         Patients at high risk of CV
                                 or diabetes plus at least one   MI >3 months before), per­     MI >3 months before), CABG        ONTARGET who are            events (Previous MI >2 d,
                                 other CV risk factor            cutaneous or surg coronary     or PTCA ≥ 3 months before,        ACE-I intolerant (by        angina with multivessel
                                 (hypertension, elevated         revascularization >6 mths      ≥50% obstruction of luminal       history, or – following     CAD, stenosis >50% in 2
                                 total cholesterol, Low HDL      before or angiographic         diameter of at least one native   ramipril administration –   major artery, positive
                                 cholesterol levels, cigarette   evidence of 70% narrowing      vessel (coronary angiography),    symptoms of cough,          stress test), PAD (previous
                                 smoking or documented           of ≥1 major coronary           and LVEF >40% on contrast or      hypotension or renal        bypass/angioplasty, limb/
                                 microalbuminuria)               arteries) or chest pain with   radionuclide ventriculography/    dysfunction likely to be    foot amputation, peripheral
                                                                 positive ECG, echo or          echocardiography, normal LV       associated with ACE-I       artery stenosis >50%),
                                                                 nuclear stress test.           wall motion echocardiography      intolerance)                prev. stroke/TIA, high risk
                                                                                                                                                              diabetes+ organ damage.
N (ACEI/ARB : Placebo)            9,297 (4,645 : 4,652)           12,218 (6,110 : 6,108)          9,290 (4,158 : 4,132)            5,926 (2,954 : 2,972)      17,118 (T 8,54:R 8,576)
Average Age (years)                          66                              60                             65                                67                          66
Female gender                               27%                            15%                            18%                               43%                          27%
Known CAD                                   80%                           100%                            72%                               67%                          66%
Previous MI                                 53%                            65%                            55%                               46%                         49%
Stroke/TIA                                  11%                             3%                             7%                               22%                          21%
Diabetes mellitus                           38%                            12%                            17%                               36%                         37%
Hypertension                                47%                            27%                            46%                               77%                          69%
Hypercholesterolemia                        66%                            63%                              ---                               ---                         ---
Antiplatelet drugs                          76%                            92%                            91%                                 ---                       *76%
β-blockers                                  39%                            62%                            60%                               70%                         57%
Lipid lowering drugs                        29%                            56%                            70%                               62%                         65%
Coronary revascularization                  40%                            54%                            72%                               45%                         51%
Composite primary                MI, stroke or death from        CV death, non-fatal MI         CV death or non-fatal MI          CV death, non-fatal         CV death, non-fatal MI,
efficacy endpoint                CV causes                       and cardiac arrest with                                          MI, stroke or CHF           stroke or CHF
                                                                 successful resuscitation                                         hospitalization             hospitalization
Placebo group: # of events                 826                              603                            929                               504               Ramipril events: 1412
                 Event rate               17.8%                            9.9%                           22.5%                            17.0%                       16.5%
ACEI/ARB:       # of events                651                              488                            909                               465                 Telmi events: 1423
                 Event rate               14.0%                            8.0%                           21.9%                            15.7%                       16.7%
Relative risk ↓ (RRR)                21.1% (p < 0.001)              19.1% (p=0.0003)                  2.8% (p = 0.43)                  8% (p=0.22)                        NA
Absolute risk ↓ (ARR)                     3.8%                             1.9%                           0.6%                              1.3%                          NA
N needed to treat (NNT)                     26                               53                            167                                77                          NA
ACEI= angiotensin-converting enzyme inhibitor, ARB = Angiotensin Receptor Blocker; RRR = relative risk reduction; ARR = Absolute risk reduction; NTT = number needed to treat.
*Aspirin use only.




                                                                                       101
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


7 Review of Safety
Safety Summary
For safety review, we evaluated the safety data from the ONTARGET, TRANSCEND
and PRoFESS trial populations. The latter two trials provide comparison of T to PBO.
Safety data for the submission was on a total of 58,553 patient years’ exposure on
17,085 patients treated with T (average observation time = 1,251 days) in ONTARGET,
TRANSCEND and PRoFESS trials, compared to 35,207 patient-years exposure on
8,576 patients treated with R (av. observation time = 1,499 days) in ONTARGET, and
22,929 patient-years exposure on 8,249 patients who received placebo (av. observation
time = 1,015 days) in TRANSCEND and PRoFESS.
The safety information of the T/R combination group in ONTARGET is not presented
because the sponsor does not intend to recombine use of T and R together.
Differences in recording and processing of safety data do not permit a simple pooling of
safety date from these 3 trials.
The reporting of safety data in ONTARGET, TRANSCEND and PRoFESS trials during
the randomized periods is limited to AEs leading to treatment discontinuation, and to
SAEs, and, during the run-in periods, to AEs considered ‘related to study medication’.
Deaths are reported as outcome events except deaths >14 days after final visit which
are reported as SAEs. The frequencies of deaths and SAEs while on treatment were
comparable between treatment groups in ONTARGET, except deaths associated with
renal failure which were reported more frequently in T/R group (0.19%) vs T (0.09%) or
R (0.13%) group. The subgroup of patients with diabetic nephropathy had more deaths
compared to the overall ONTARGET population, most frequent causes being MI, stroke,
ventricular arrhythmias, worsening heart failure, cancer and amputation related causes.
In TRANSCEND and PRoFESS, deaths were comparable between T and PBO groups
in the randomized periods.
SAEs were experienced by comparable proportions of patients in the treatment arms in
each of these telmisartan trials. The SAEs of renal failure, renal and urinary disorders,
GI disorders, hypotension and hyperkalemia occur more frequently in the T/R group in
ONTARGET.
The most frequent AEs leading to permanent discontinuations were hypotension and
cough, being most frequent for the T/R combination group compared to R or T. The risk
of discontinuation of study medication was significantly greater for T/R vs monotherapy
with T (HR 1.20, CI 1.12, 1.27, P<0.0001) or R (HR 1.19, CI 1.12, 1.27, P<0.0001), and
the risk of study drug discontinuation due to AEs was also significantly greater for T/R
vs monotherapy with T (HR 1.51, CI 1.38, 1.65, P<0.0001) or R (HR 1.42, CI 1.30, 1.55,
P<0.0001).




                                           102 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

In ONTARGET, the AEs of special interest which occurred significantly more frequently
in the R group are cough (P<0.0001) and angioedema (P<0.0067), and in the T group is
hypotensive symptoms (P<0.0003). In TRANSCEND and PRoFESS, patients in the T
group had significantly higher frequencies of hypotensive symptoms and syncope
compared to PBO group.
It appeared that AEs related to BP-lowering effect (hypotension, syncope, etc.) and
renal effects (renal failure, hyperkalemia) of T and R caused more patients to drop out
from the T/R combination group. Increased adverse effects without additive beneficial
effect of the T/R combination was found.
Safety signals were found for the following, which are of uncertain clinical significance:
•	 Using the definition of ‘sepsis’ by ACCP/SCCM, an increased risk of sepsis was
   found (not statistically significant) in T group compared to R or PBO in all 3 trials. In
   TRANSCEND and PRoFESS, more deaths were associated with T compared to
   PBO.
•	 An increased risk of malignancies was also found in ONTARGET for T/R group vs R
   for all patients (HR 1.14, 95% CI 1.03, 1.26, p = 0.0089) and for patients without
   cancer at baseline (HR 1.12, 95% CI 1.01, 1.25, P = 0.036). In TRANSCEND, too, a
   higher rate of malignancies was found in patients treated with T vs PBO.

7.1 Methods
The reporting of safety data in the ONTARGET, TRANSCEND and PRoFESS trials was
modified from the conventional method of safety data reporting, as follows:
•	 In ONTARGET, TRANSCEND and PRoFESS trials, the sponsor’s reporting AEs is
   limited to (i) AEs leading to treatment discontinuation (using an internal coding
   system), and (ii) SAEs (coded using Medical Dictionary for Regulatory Activities
   [MedDRA]), for which the sponsor’s reason was that the safety profiles of telmisartan
   and ramipril were relatively well-characterized.
•	 During the run-in periods of the trials, only AEs considered ‘related to study
   medication’ were recorded.
•	 For the ONTARGET and TRANSCEND trials, outcome events were analyzed both
   as efficacy and as safety data. Outcome events for efficacy were adjudicated by the
   event adjudication committee, whereas for safety evaluation, these outcome events
   were included as SAEs based on the information provided by the investigators.
   Thus, outcome events may be coded differently in efficacy and safety analyses.
•	 For the PRoFESS trial, only patients who did not receive ACE-Is during the study
   were evaluated for clinical safety (with the explanation that the dual RAAS blockade
   might interfere with the safety profile of patients at high CV risk).
Safety data in the Phase I bioequivalence trial were collected in the conventional way;
i.e., data on both non-serious AEs and SAEs were recorded, and a broader range of


                                            103 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

laboratory parameters were analyzed.
A limited set of laboratory parameters (e.g., serum creatinine, potassium) was analyzed
at baseline and at selected time points during the ONTARGET trial.
We found that AEs from TRANSCEND and PRoFESS trials are not reported in labeling.

7.1.1     Studies/Clinical Trials Used to Evaluate Safety
For this review, we evaluated the safety data from the ONTARGET, TRANSCEND and
PRoFESS trial populations. The latter two trials provide comparison of T to PBO. There
is also a phase 1 bioequivalence trial which was a cross-over study with the T/R
combination conducted in healthy volunteers, a different population.
The sponsor contends that because they do not recommend concomitant use of T and
R, the safety information of the T/R group in ONTARGET trial is of limited relevance to
the safety review.
The number of patients in the three large clinical outcome trials are shown in Table 50.

Table 50 Numbers of patients included in different safety sets for safety analyses

                                            Meta-analysis of
                                            mortality data in
                                            HOPE, EUROPA
                                              and PEACE
                                                trials23




(Source: Sponsor’s Table 1.1.4:1, on page 35 of Summary of Clinical Safety)
RIS: run-in set, comprising all patients enrolled; FAS: full analysis set, comprising all patients randomized
irrespective of treatment; TS: treated set, comprising all patients who received at least one dose of
telmisartan/matching placebo
1
 ACE-I use recorded at baseline or at any time during the PRoFESS trial. This set of patients who did not
receive ACE-Is and the set of patients who received ACE-Is are mutually exclusive.
2
 The bioequivalence study 1236.5 was a cross-over study with T/R combination treatment only.

Patients in each treatment arm of the ONTARGET trial received open-label ACE-Is or
ARBs (e.g., 1040 patients in telmisartan arm received open-label ACE-Is). Similarly, in
the TRANSCEND trial, 60 patients (who are allegedly ACE-I intolerant) in the
telmisartan arm received open-label ACE-Is, and in the PRoFESS trial, 4,430 patients in
the telmisartan arm and 4,776 patients in the placebo arm received open-label ACE-I.




                                                    104 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

7.1.2   Categorization of Adverse Events
For the ONTARGET and TRANSCEND trials, outcome events and SAEs were
categorized as follows:
    • CC0:              outcome events and hospitalizations
    • CC1:              serious, unexpected, study drug-related events
    • CC2:              serious, expected, study-drug related events
    • Exempted SAEs: serious, expected, not study drug-related.
For these trials, analyses SAEs are based on the information reported by investigators,
on death report forms, visit CRFs and outcome event forms.
For the PRoFESS trial, the outcome events are analyzed only as efficacy endpoints and
are not included in the AE reporting. For the PRoFESS trial, SAEs are categorized in a
different way as follows:
• ICH category 0: SAEs not considered study drug-related, excluding outcome events
• ICH category 1: unexpected SAEs considered study drug-related
• ICH category 2: expected SAEs considered study drug-related

7.1.3 Pooling of Data across Studies/Clinical Trials to Estimate and Compare
Incidence
The differences in the recording and processing of safety data do not permit a simple
pooling of safety data from the three large outcome trials.
Since the ONTARGET trial provides the pivotal efficacy data for this submission, its
safety data are the most relevant for review. Safety data are reviewed for each study
separately with emphasis on the ONTARGET trial.
The safety data of patients who were not receiving dual RAAS blockade (i.e., not
receiving ACE-Is) in the TRANSCEND and PRoFESS trials provide a comparison of
telmisartan and placebo for safety evaluation.

7.2 Adequacy of Safety Assessments

7.2.1 Overall Exposure at Appropriate Doses/Durations and Demographics of
Target Populations
Exposure during the run-in period:
The PRoFESS trial has no run-in period.
The run-in period in ONTARGET trial was used to identify patients who were ACE-I
intolerant (to R) or a combination of T and R.




                                          105 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Only the run-in period of the TRANSCEND trial contains data related to 80 mg/day T;
this run-in period consisted of 1 week of PBO and 2 – 3 weeks treatment with T 80
mg/day. In this TRANSCEND trial run-in period, 6,665 patients were enrolled with the
mean observation time of 25.2 days, to make a total exposure of 460 patient years.
Exposure during the randomized period: The exposure in each of the trials (Table 51) is
reported as ‘observation time on treatment’ which is defined as ‘the time difference
between the date of the visit at which the permanent stop of the study medication was
recorded and the date of randomization + 1 day.’
During the randomized period, the overall mean observation time in ONTARGET trial
was about 4.5 years (1507 days in the T group, 1499 days in the R group, and 1444
days in the T/R group) and it was comparable between treatment groups. For safety
analyses, all between-group comparisons of SAEs and AEs were done after adjustment
for observation time.

Table 51: Observation time on treatment in randomized periods of clinical trials




1
The observation time on treatment was determined as the time difference between the date of the visit at
which the permanent stop of the study medications was recorded and the date of randomization+1 day.
        Source: Sponsor’s table 2.5.5.2.2:1 in page 33 of Clinical Overview.

For patients enrolled in ONTARGET, TRANSCEND and PRoFESS trials, the combined
exposure for treatment durations, different doses, demographics (age-group, gender
and ethnic origin), and special population (renal impairment) are shown in Table 52.
FDA Reviewers’ comment: The postmarketing exposure in patients treated for
hypertension is much larger (Table 69 in Section 8        Postmarketing Experience). At
this point in time, no safety signal has been identified in the postmarketing data, but no
specific investigation for any safety signal has been made in the postmarketing data.



                                                 106 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 52 Exposure by duration (days) of treatment
                             Exposure by duration
         Duration of exposure     Persons       Person time (patient
                                                       years)
              1 – 365 days          2572                  764
             366 – 730 days         2694                 4456
            731 – 1095 days         3971                 9752
           1096 – 1460 days         2083                 7082
           1461 – 1825 days         7218                30389
           1826 – 2190 days         3063                14515
               ≥2191 days             37                  176
                  Total            21638                67134
            Exposure by dose (only ONTARGET and TRANSCEND)
           Dose of exposure       Persons       Person time (patient
                                                       years)
                 40 mg               104            Not applicable
                 80 mg             11392            Not applicable
                       Exposure by Age Group and Gender
           Dose of exposure       Persons       Person time (patient
                                                       years)
                                  M       F        M             F
                <65 years        6664 2729      21311          8314
            ≥65 - <75 years      5698 2930      18640           8998
                ≥75 years        2130 1487        5930         3940
                  Total         14492 7146      45881          21253
                           Exposure by Ethnic Origin
           Dose of exposure       Persons       Person time (patient
                                                       years)
               Caucasian           13988                45308
                  Black              673                 1732
                  Asian             5162                13889
                  Other             1812                 6200
                 Missing              3                    6
                  Total            21638                67134
                     Exposure by Special Populations (renal impairment)
             Dose of exposure          Persons         Person time (patient years)
              Renal impairment
                                          8173                   21878
          (eGFR <60 ml/min/1.73m2)
             No renal impairment
                                         13466                   45256
          (eGFR ≥60 ml/min/1.73m2)
            Source: Sponsor’s table1.2.1.1 (c) on page 13 of Risk Management Plan




                                            107
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

7.3 Major Safety Results
In the randomized period,
•	 deaths and SAEs occurring on a date (+1 day) before the date of the visit at which
    permanent discontinuation was recorded were considered 'on treatment',
•	 deaths and SAEs occurring on a date (+1 day) after the reporting of permanent
    treatment discontinuation were considered to have occurred 'post-treatment'.
SAEs reported after the final visit +1 day during the 30-day follow-up period or later
('post final visit') are described as well.
SAEs were analyzed on a by-patient basis, and rates are reported as frequencies of
patients per event. As the observation times differed between treatment groups, event
rates per 100 patient-years were calculated for comparison.

7.3.1    Deaths
Deaths were reported as outcome events, except for deaths >14 days after final patient
contact which were reported as SAEs.
For deaths in safety evaluation, two analyses were performed:
•	 The primary cause of death as recorded by the investigator in CRF (page 99), and
•	 All SAEs reported in association with death.

ONTARGET trial: 

During the run-in period, there were 29 deaths (of which 9 were due to MI, and 7 to 

sudden cardiac death). 


Table 53 Deaths in ONTARGET trial during randomized period (safety evaluation)
                                                   T/R           T            R
Randomized, n (%)                              8502 (100.0) 8542 (100.0) 8576 (100.0)
All deaths, n (%)                               1065 (12.5)  989 (11.6)  1014 (11.8)
Deaths on treatment, n (%)                       780 (9.2)   750 (8.8)    759 (8.9)
    Non-CV deaths, n (%)                         294 (3.5)    267 (3.1)    274 (3.2)
Deaths post-treatment, n (%)                     285 (3.4)    239 (2.8)    255 (3.0)
    Non-CV deaths, n (%)                         151 (1.8)    124 (1.5)    137 (1.6)
                               Data from efficacy analyses
    CV deaths*, n (%)                            620 (7.3)    598 (7.0)    603 (7.0)
    Non-CV deaths*, n (%)                        445 (5.2)   391 (4.6)    411 (4.8)
CV deaths as primary efficacy endpoint, n (%)    402 (4.7)    384 (4.5)    393 (4.6)
*Adjudicated CV and non-CV deaths for efficacy analyses (shown for comparison).
Source: Sponsor’s Table 12.2.1: 2 on page 267 of ONTARGET Study Report.

During the randomized period, 3,068 (12%) patients died overall (Table 53), with more


                                                108
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

deaths in T/R group compared to monotherapy with T or R for (i) all deaths, (ii) deaths
on treatment and (iii) deaths post-treatment.
The most frequently reported SAEs recorded while on treatment in association with
death were ‘sudden cardiac death (excluding MI, stroke and ventricular tachyarrhythmia
specified in CRF)’ (T/R: 1.5%; T: 1.8%; R: 1.6%), cancer (T/R: 1.8%; T: 1.5%; R: 1.6%),
and MI (T/R: 1.1%; T: 1.2%; R: 1.1%).
The frequencies of primary causes of death and of SAEs reported in association with
death while on treatment were comparable between treatment groups, except that renal
failure was reported in association with death more frequently in the T/R group (0.19%)
than in the T (0.09%) or R (0.13%) group. These causes of deaths were also the most
frequently reported post-treatment deaths.
The subgroup of patients with diabetic nephropathy showed more deaths (Table 54)
while on treatment as well as post-treatment compared to the overall ONTARGET
population. The most frequent causes were MI, stroke, ventricular arrhythmias,
worsening heart failure, cancer and amputation related deaths.

Table 54 Deaths in the subgroup of patients with diabetic nephropathy
ONTARGET trial during randomized period (safety evaluation)
Patients with diabetic nephropathy     T/R            T           R
Enrolled, n (%)                    248 (100.0) 288 (100.0) 238 (100.0)
Deaths on treatment, n (%)          54 (21.8)     62 (21.5)   55 (23.1)
Enrolled, n (%)                         69           84          70
Deaths post-treatment, n (%)        21 (30.4)     30 (35.7)   28 (40.0)
Source: Sponsor’s Table 15.3.2.3:1 on page 1198 of ONTARGET Study Report.

The analysis of deaths for subgroups of age group, sex, ethnicity, obesity, diabetes,
hypertension, CAD, PAD, previous stroke or TIA or concomitant use of certain
medications (statins, dihydropyridines, β-blockers, α-blockers, diltiazem/verapamil,
diuretics, antihypertensives, digitalis) or renal impairment (high risk diabetes,
albuminuria, impaired renal function) did not reveal differences from the overall study
population results, with the exception that for all three treatment groups more deaths
were observed in patients:
•	 with diabetes, high risk diabetes, PAD, prior stroke/TIA, albuminuria (UACR 30 to
    300 mg/g Crea, UACAR >300 mg/g Crea), hypertension and impaired renal function,
•	 in Latin America (13.4%) than patients in North America (7.8%), Europe including
    South Africa (9.0%), Asia including Middle East (9.0%), and Australia (5.6%),
•	 not receiving statins, being inversely related to increasing doses of simvastatin, and
•	 receiving digitalis.
On the other hand, fewer deaths were found in patients who received dihydropyridines,
diltiazem/verapamil, β-blockers, α-blockers and any hypertensive medication.




                                              109 

Clinical Review 

Khin Maung U, M.D. & Jialu Zhang, Ph.D. 

NDA 20-850 Efficacy Supplement SE1-025 

Micardis® (telmisartan) tablets 


TRANSCEND trial: 

During the run-in period, there were 3 deaths (one each for MI, other sudden cardiac 

death and other CV cause). 

During the randomized period, there were 713 (12.0%) deaths. The mortality rates were
similar overall as well as on-treatment between telmisartan and placebo groups (Table
55). On the preferred term level, MI and sudden cardiac death were the most frequently
reported SAEs associated with death. Heart failure and ventricular arrhythmias in
association with death were reported more frequently in telmisartan group (0.9% and
0.2%, respectively) than in placebo group (0.6% and 0.03%, respectively).

Table 55 Deaths in TRANSCEND trial – randomized period (safety evaluation)
                                        T          PBO
      Randomized, n (%)            2954 (100.0) 2972 (100.0)
      All deaths, n (%)             364 (12.3)   349 (11.7)
      Deaths on treatment, n (%)    284 (9.6)    266 (9.0)
          Non-CV deaths, n (%)       94 (3.2)     94 (3.2)
      Deaths post-treatment, n (%)   80 (2.7)     83 (2.8)
          Non-CV deaths, n (%)       43 (1.5)     32 (1.1)
      Source: Sponsor’s Table 2.5.5.4.2:1 on page 39 of Clinical Overview.


PRoFESS trial: 

During the first month after randomization, there were few deaths (T: 22 of 56 (0.49%) 

patients; R 12 of 52 (.23%) patients). 


Table 56 Deaths in PRoFESS trial – randomized period (safety evaluation)
                                         T         PBO
      Randomized                  5589 (100.0) 5277 (100.0)
      All deaths                    417 (7.5)    389 (7.4)
                      1
      Incidence % (PY) fatal SAEs    3.6 (1.8)   3.2 (1.6)
      1
       Patients per 100 patient years (adjusted for observation time) 

      Source: Sponsor’s Table 2.1.2.1.2:1 on page 98 of Summary of Clinical Safety. 


During the randomized period of the PRoFESS trial, deaths were not reported as SAEs,
but the incidence of death was analyzed as an efficacy parameter. Overall mortality and
fatal SAEs were comparable in the telmisartan and placebo groups (Table 56). The
most frequently reported SAEs associated with death were general disorders, cancer
and cardiac disorders.
The most frequent causes of death in both T and PBO groups were 'non-vascular
causes' (3.0% vs. 2.8%) and 'other vascular causes' (other than stroke, MI, or
hemorrhage; 3.0% vs. 3.1%).




                                               110 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

7.3.2    Nonfatal Serious Adverse Events
ONTARGET trial:
During the run-in period, 688 (2.4%) of enrolled patients experienced SAEs (including
death), with SAEs leading to permanent discontinuation in 62 (0.2%) patients. Non-fatal
SAEs occurred in 2.3% of patients during the run-in period; these were mainly cardiac
disorders.
During the randomized period, 16,404 (64.0%) patients experienced non-fatal SAEs
(Table 57). The frequencies of patients with non-fatal SAEs were comparable in all
treatment groups (T/R: 59.4%; T: 61.8%; R: 60.1%) during the randomized period.
Renal failure was reported as a SAE on treatment more frequently in the T/R group
(1.65%) than in the T (1.35%) or R (1.25%) group, as were renal and urinary disorders,
GI disorders, hypotension and hyperkalemia.
On the other hand, diabetes mellitus, new onset diabetes and cardiac failure were
reported in fewer patients in T/R group than in T or R group.
In all treatment groups, the frequencies of SAEs remained relatively constant (around
24% to 28% of the at risk population in each treatment group) each year over time for
the 5 year period per Table 12.3.2.2.1:3 on page 283 of ONTARGET clinical trial report.

Table 57 Nonfatal SAEs in ONTARGET trial – randomized period
                                            T/R           T                                   R
Randomized, n (%)                       8502 (100.0) 8542 (100.0)                        8576 (100.0)
All SAEs, n (%)                         5451 (64.1)   5537 (64.8)                        5416 (63.2)
*
  SAEs on treatment, n (%)              5051 (59.4)  5282 (61.8)                         5158 (60.1)
   ‡
     SAEs excluding outcome events and    145 (1.7)    111 (1.3)                          106 (1.2)
hospitalizations, n (%)
*
  SAEs post-treatment, n (%)             868 (10.2)    666 (7.8)                            653 (7.6)
    ‡
      SAEs excluding outcome events and    7 (0.1)     2 (0.02)                              2 (0.02)
hospitalizations, n (%)
*
 SAEs with an onset date before the date of the visit at which permanent discontinuation was recorded +1
day were considered to have occurred 'on treatment', SAEs with an onset date thereafter until the final
visit +1 day were considered 'post-treatment'.
‡
 SAEs based on information provided by the investigators on the SAE report forms CRF 65B, excluding 

outcome events and hospitalizations. 

Source: Sponsor’s Table 12.2.1: 2 on page 267 of ONTARGET Study Report. 



TRANSCEND trial: 

During the run-in period, 125 (1.9%) of enrolled patients experienced SAEs (including 

death), with the most frequent SAEs being cardiac disorders (0.8%) including angina 

pectoris (0.4%), congestive cardiac failure (0.2%) and MI (0.1%). 

During the randomized period, 3,627 (61.2%) patients experienced non-fatal SAEs
(Table 58). The percentages of SAEs were comparable in the two treatment groups.


                                                  111 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


        Table 58 Nonfatal SAEs in TRANSCEND trial – randomized period
                                            T            PBO
          Randomized, n (%)            2954 (100.0) 2972 (100.0)
          All SAEs, n (%)‡             1802 (61.0) 1825 (61.4)
          *
            SAEs on treatment, n (%)   1691 (57.2) 1739 (58.5)
          *
            SAEs post-treatment, n (%)  237 (8.0)      243 (8.2)
        ‡
          Excluding deaths. Non-fatal SAEs that occurred in patients who later died are included. All 

        information is based on the information provided by the investigators. 

        *
         SAEs with an onset date before the date of the visit at which permanent discontinuation was 

        recorded +1 day were considered to have occurred 'on treatment,' SAEs with an onset date 

        thereafter until the final visit +1 day were considered 'post-treatment.' 

        Source: Sponsor’s Table 2.1.22 on page 91 of Summary of Clinical Safety. 



PRoFESS trial: 

During the first month after randomization, there were fewer SAEs in T compared to R 

treatment group (T: 195 of 5,589 (3.5%) patients; R 136 of 5,277 (2.6%) patients). 

The number (%) of patients who experienced all SAEs during the randomized period of
the PRoFESS trial, are shown in Table 59.
The percentages of SAEs were comparable in the two treatment groups. The most
frequent SAEs in both groups by SOC were infections and infestations (4.8% vs. 4.4%),
cardiac disorders (4.4% vs. 4.0%), and nervous system disorders (4.3% vs. 3.9%).
In general, more patients in the T group than the PBO group experienced SAEs related
to hyperkalemia (0.1% vs. 0.0%), hypotensive symptoms and syncope (1.6% vs. 1.3%),
and renal failure (0.9% vs. 0.6%).

        Table 59 All SAEs in PRoFESS trial – randomized period
                                          T           PBO
        Randomized                   5589 (100.0) 5277 (100.0)
        *
          SAEs on treatment, n (%)    1277 (22.9) 1107 (21.0)
        *
          SAEs post-treatment, n (%)   376 (6.7)    316 (6.0)
        *
        Patients were considered on treatment from randomization up to permanent treatment 

        discontinuation +28 days and post- treatment from permanent treatment discontinuation +28 days 

        up to the last patient contact +28 days. 

        Source: Sponsor’s Table 2.1.2: 3 on page 92 of Summary of Clinical Safety. 


7.3.3       Dropouts and/or Discontinuations
ONTARGET trial: 

Permanent discontinuation of study medication was defined as discontinuation of both 

study medications, i.e., the active treatment and the double-dummy placebo treatment 

in the monotherapy arms and the 2 active treatments in the combination therapy arm. 

During the run-in period, AEs leading to permanent treatment discontinuation were


                                                   112 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

reported in 3.7% of patients. The most frequent AEs leading to permanent treatment
discontinuation were cough (1.1%) and AEs related to low BP (e.g. dizziness,
hypotension). SAEs leading to permanent treatment discontinuation during the run-in
period were reported in 0.2% of patients, the most frequent being hypotension (n=13),
syncope (n=10), and angioedema (n=9 patients).
Table 60 shows the number of patients who discontinued treatment during the trial due
to SAEs and AEs. The most frequently reported AEs leading to permanent treatment
discontinuation were hypotension (T/R: 2.8%; T: 1.5%; R: 1.1%) and cough (T/R: 2.1%;
T: 0.6%; R: 1.7%). Hypotension as well as syncope, renal diseases, hyperkalemia, and
diarrhea were reported more frequently in the T/R group than in the T or R group; cough
and angioedema were reported at lower frequencies in the T group than in the T/R or R
group. A few patients discontinued trial medication because of non-fatal SAEs (T/R:
0.7%; T: 0.5%; R: 0.5%).

Table 60 Permanent discontinuation of study drugs in ONTARGET trial
                                                 T/R           T            R
Randomized, n (%)                           8502 (100.0) 8542 (100.0) 8576 (100.0)
Permanent discontinuation of study drugs* 1730 (20.3) 1454 (17.0) 1468 (17.1)
†
  SAEs leading to treatment discontinuation    64 (0.8)     44 (0.5)     40 (0.5)
†
  AEs leading to treatment discontinuation   1053 (12.4)   700 (8.2)    749 (8.7)
Other reasons                                850 (10.0)   880 (10.3)   873 (10.2)
‡
  Patient’s request                            76 (0.9)     92 (1.1)     77 (0.9)
*
 Multiple entries possible; †Reasons for permanent treatment discontinuation during the randomized
period as derived from the free text field on the CRF. These reasons were coded internally by PHRI. In
this table, only the reasons considered as AEs are presented.
‡
 These patients did not refuse to be followed-up but did not want to take study medication.
Source: Sponsor’s Table 10.1: 4 on page 143 of ONTARGET Study Report.

Over the entire course of treatment up to 5 years, the relative risk for permanent
discontinuation was consistently higher for T/R compared to R, and consistently lower
for T vs R, being significantly so for cumulative discontinuations in the first 2 years
(Table 61). At completion of the trial, the risk of discontinuation of study medication was
significantly greater for T/R than for monotherapy with either R (HR 1.19, CI 1.12, 1.27;
P<0.0001) or T (HR 1.20, CI 1.12, 1.27; P<0.0001). In a similar manner, the risk of
study drug discontinuation due to AEs was significantly greater for T/R than for
monotherapy with either R (HR 1.42, CI 1.30, 1.55; P<0.0001) or T (HR 1.51, CI 1.38,
1.65; P<0.0001).


FDA Reviewers’ Comment: In comparison to the HOPE trial, the number of patients
discontinuing study medication was much lower in ONTARGET. For example, in HOPE
trail 1343 (28.8%) patients permanently discontinued in the ramipril group, compared to
17.1%, 17.0% and 20.3% patients who discontinued R, T and T/R, respectively (Table
60), in ONTARGET trial.


                                                  113 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Table 61 Cumulative permanent discontinuation of study drug in ONTARGET trial
             Relative risk for discontinuation T/R vs R      Relative risk for discontinuation T vs R
                Permanent            Discontinuation           Permanent            Discontinuation
              discontinuation          due to AEs            discontinuation          due to AEs
6 months     1.37 (1.18 – 1.59)      1.66 (1.38 – 2.00)      0.81 (0.68 – 0.96)     0.81 (0.65 – 1.01)
1 year       1.40 (1.25 – 1.57)      1.75 (1.49 – 2.00)      0.84 (0.74 – 0.96)     0.84 (0.70 – 0.99)
2 years      1.27 (1.17 – 1.39)      1.53 (1.36 – 1.71)      0.89 (0.81 – 0.98)     0.80 (0.70 – 0.92)
3 years      1.24 (1.15 – 1.34)      1.49 (1.35 – 1.65)      0.95 (0.88 – 1.03)     0.86 (0.77 – 0.97)
4 years      1.21 (1.13 – 1.29)      1.45 (1.32 – 1.59)      0.98 (0.91 – 1.05)     0.92 (0.82 – 1.02)
5 years      1.19 (1.12 – 1.27)      1.41 (1.29 – 1.54)      0.99 (0.93 – 1.06)     0.94 (0.85 – 1.03)
Study end    1.19 (1.12 – 1.27)      1.42 (1.30 – 1.55)      0.99 (0.93 – 1.06)     0.94 (0.85 – 1.04)
Source: Data tables on pages 1 – 26 of Attachment 5 in Sponsor’s response submitted 14-Apr-2009.

Discontinuation of active treatment, i.e. telmisartan in the T group, ramipril in the R
group, and the 2 active treatments in the T/R group, irrespective of discontinuation of
matching placebo, is shown in Table 62.

Table 62 Permanent discontinuation of active study drugs in ONTARGET trial
                                T/R            T             R
Randomized, n (%)           8502 (100.0) 8542 (100.0) 8576 (100.0)
Permanent discontinuation   1730 (20.3)   1512 (17.7)   1689 (19.7)
of active study drug*
†
  Discontinuation of active 1053 (12.4)    743 (8.7)     942 (11.0)
study drug due to AEs
*
 Multiple entries possible
†
 Reasons for permanent treatment discontinuation during the randomized period as derived from the free

text field on the CRF. These reasons were coded internally by PHRI. In this table, only the reasons 

considered as AEs are presented. 

Source: Sponsor’s Table 10.1: 4 on page 143 of ONTARGET Study Report. 


The risk for permanent discontinuation of active drug was significantly greater in the T/R
group than in the T group (HR 1.15, CI 1.08, 1.22; P<0.0001), but not the R group (HR
1.03, CI 0.97, 1.10; P=0.2858).
Over the course of the trial, the risk of discontinuation of active drugs due to AEs was
significantly higher in the T/R group than the R group (Table 63). On the other hand,
the risk of permanent discontinuation of telmisartan in the T group was significantly
lower than of ramipril in the R group over the course of the trial (Table 63). The same
trend was seen for permanent discontinuation of active drug due to AEs. The hazard
ratio of T vs. R for permanent discontinuations of active treatment overall was 0.90 (CI
0.84, 0.96; p=0.0008) and for permanent discontinuations of active treatment due to
AEs was 0.79 (CI 0.72, 0.87; p <0.0001).


                                                 114 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


Table 63 Cumulative permanent discontinuation of active drug in ONTARGET trial
              Relative risk for discontinuation T/R vs R      Relative risk for discontinuation T vs R
                Permanent             Discontinuation           Permanent             Discontinuation
              discontinuation           due to AEs            discontinuation           due to AEs
6 months      1.12 (0.98 – 1.29)     1.25 (1.06 – 1.49)       0.69 (0.59 – 0.81)     0.65 (0.53 – 0.79)
1 year        1.15 (1.03 – 1.28)     1.29 (1.13 – 1.48)       0.72 (0.64 – 0.82)     0.67 (0.57 – 0.78)
2 years       1.06 (0.98 – 1.15)     1.16 (1.04 – 1.28)       0.77 (0.70 – 0.84)     0.64 (0.57 – 0.73)
3 years       1.06 (0.98 – 1.13)     1.15 (1.06 – 1.27)       0.84 (0.78 – 0.91)     0.71 (0.64 – 0.79)
4 years       1.04 (0.97 – 1.11)     1.14 (1.04 – 1.24)       0.88 (0.82 – 0.94)     0.77 (0.69 – 0.84)
5 years       1.03 (0.97 – 1.10)     1.12 (1.03 – 1.22)       0.90 (0.84 – 0.95)     0.79 (0.72 – 0.87)
Study end     1.03 (0.97 – 1.10)     1.13 (1.04 – 1.22)       0.90 (0.84 – 0.96)     0.79 (0.72 – 0.87)
Source: Data tables on pages 27 – 52 of Attachment 5 in Sponsor’s response submitted 14-Apr-2009.


Patients with diabetic nephropathy
Table 64 Permanent discontinuations in patients with diabetic nephropathy in
ONTARGET trial
                                               T/R           T           R
Randomized, n (%)                           248 (100.0) 288 (100.0) 238 (100.0)
Permanent discontinuation of study drugs*    69 (27.8)   84 (29.2)   70 (29.4)
†
  SAEs leading to treatment discontinuation   7 (2.8)     4 (1.4)     4 (1.7)
†
  AEs leading to treatment discontinuation   45 (18.1)   46 (16.0)   41 (17.2)
Other reasons                                29 (11.7)   46 (16.0)   37 (15.5)
‡
  Patient’s request                           5 (2.0)     4 (1.4)     4 (1.7)
*
 Multiple entries possible; †Reasons for permanent treatment discontinuation during the randomized
period as derived from the free text field on the CRF. These reasons were coded internally by PHRI. In
this table, only the reasons considered as AEs are presented.
‡
 These patients did not refuse to be followed-up but did not want to take study medication.
Source: Sponsor’s Table 10.1: 7 on page 146 of ONTARGET Study Report.

Of the 732 recruiting study centers, 350 centers enrolled 776 patients identified as 

having diabetic nephropathy; 774 were randomized. More diabetic nephropathy patients 

were randomized to T (288 patients, 37.2%) than to T/R (248 patients, 32.0%) or R (238 

patients, 30.7%). 


TRANSCEND trial: 

Of 6,665 patients enrolled during the run-in period, 175 (2.6%) patients discontinued 

treatment permanently, of which 169 (2.5%) did so for AEs, the most frequent being 

dizziness/lightheadedness/vertigo (45 (0.7%) patients). 

In contrast to the ONTARGET trial where 17-20 of patients discontinued treatment



                                                  115 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

permanently (Table 60), with 8-13% of randomized patients having discontinued due to
SAEs/AEs (Table 60), in the TRANSCEND trial, there were relatively fewer patients who
discontinued treatment permanently due to SAEs/AEs(Table 65).

      Table 65 Permanent treatment discontinuations in TRANSCEND trial
                                                        T          PBO
      Randomized, n (%)                           2954 (100.0) 2972 (100.0)
      †
        SAEs leading to treatment discontinuation    10 (0.3)     7 (0.2)
      †
        AEs leading to treatment discontinuation    247 (8.4)    225 (7.6)
      †
       Reasons for permanent treatment discontinuation during the randomized period as derived from 

      the free text field on the CRF. These reasons were coded internally by PHRI. 

      Source: Sponsor’s Table 2.1.2: 2 on page 91 of Summary of Clinical Safety. 



PRoFESS trial: 

During the first month after randomization, 66 (1.2%) patients in T group and 50 (1.0%) 

patients in PBO group discontinued due to AEs before or at the first visit after 

randomization. 

In the PRoFESS trial, outcome events were not reported as AEs, and the AEs leading
to permanent discontinuation could be serious or non-serious because no information
was collected in this regard. The frequency of permanent discontinuations in the
PRoFESS trial (Table 66) are more than that in the TRANSCEND trial (Table 65) but
less than that found in the ONTARGET trial (Table 60).

      Table 66 Permanent treatment discontinuations in PRoFESS trial
                                                      T          PBO
      Randomized                                 5589 (100.0) 5277 (100.0)
      †
        AEs leading to treatment discontinuation 797 (14.3)    567 (10.7)
      †
       AEs leading to permanent treatment discontinuation could have been serious or non-serious (no 

      information collected on seriousness). 

      Source: Sponsor’s Table 2.1.2: 3 on page 92 of Summary of Clinical Safety. 


More patients treated with T (14.3%) compared to PBO (10.7%) were reported with AEs
that led to permanent discontinuation (excluding outcome events).
This difference was mainly due to higher incidences of vascular disorders (3.0% vs.
1.2%; especially hypotension with 2.5% vs. 0.9%), dizziness (1.3% vs. 0.9%), and
gastrointestinal disorders (3.1% vs. 2.5%) in the T group than in the PBO group. The
incidences of headache leading to discontinuation were comparable between both
groups (2.3% vs. 2.0%).
More patients of the T group than of the PBO group experienced AEs associated with
hyperkalemia (0.2% vs. 0.1%), hypotensive symptoms and syncope (4.1% vs. 1.9%),
and renal failure (0.3% vs. 0.2%) that led to discontinuation. When hypotensive
symptoms and syncope were analyzed post hoc by baseline blood pressure category,


                                               116 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

the majority of patients with these symptoms were found to have lower baseline blood
pressures (<140/90 mmHg) which could have made them more susceptible to
hypotensive symptoms or syncope.
There was no difference between treatments in major hemorrhagic events (T: 3.8% vs.
PBO: 3.9%) and all hemorrhagic events (minor and major, 5.1% in each treatment
group). The incidences of intracranial hemorrhages were 1.1% in the T group and 1.4%
in the PBO group.

7.3.4   Significant Adverse Events
In ONTARGET and TRANSCEND trials, the sponsor submitted post-hoc analyses
performed using the χ2-test on AEs related to inhibition of the renin angiotensin
aldosterone system (RAAS) called ‘AEs of special interest’ {cough, angioedema,
hypotensive symptoms, syncope, renal failure, signs of renal dysfunction (based on
doubling of serum creatinine, serum potassium >5.5 mmol/L, eGFR <15 ml/min/1.73m2)
and diarrhea}, derived from SAEs and AEs leading to temporary/permanent treatment
discontinuation although frequencies were below 1% in any treatment group. (The
sponsor submitted MedDRA PTs, PHRI codes for these AEs of special interest in Table
1.1.3.1.3:1 on Page 27 of module 2.7.4 Summary of Clinical Safety.)

AEs of special interest in the ONTARGET trial
Of the ‘AEs of special interest’ in the ONTARGET trial (Table 67), cough and
angioedema were significantly more frequent in the R group, and hypotensive
symptoms were more frequent in the T group.




                                         117 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 67 AEs of special interest in the ONTARGET trial




AEs of special interest in the TRANSCEND trial
In the TRANSCEND trial, patients in the T group had higher frequencies of hypotensive
symptoms (RR 1.50, CI 1.08, 2.09, P = 0.0157), syncope (RR 2.06, CI 1.23, 3.46, P =
0.0053), signs of renal dysfunction (RR 1.41, CI 1.18, 1.69, P = 0.0001) and diarrhea
(RR 2.85, CI 1.22, 6.79, P = 0.0116).

AEs of special interest in the PRoFESS trial
Similarly, In the PRoFESS trial, patients in the T group had higher frequencies of
hypotensive symptoms (RR 1.57, CI 1.08, 2.30, P = 0.0185), syncope (RR 1.89, CI
1.15, 3.11, P = 0.0111), and hyperkalemia (RR not determinable, P = 0.0060).


7.4 Supportive Safety Results

7.4.2   Laboratory Findings
Because of the known side effects of telmisartan and ramipril, serum creatinine and
potassium values were of special interest. Note that clinical laboratory analyses were
performed locally, therefore no standardized evaluations can be made across centers or
across regions. Possibly Clinically Significant Abnormalities (PCSAs) and deviations
from the reference range were based on local laboratory standards. The rule for PCSA


                                         118 

Clinical Review 

Khin Maung U, M.D. & Jialu Zhang, Ph.D. 

NDA 20-850 Efficacy Supplement SE1-025 

Micardis® (telmisartan) tablets 


for creatinine was ≥1.5 mg/dL, and for serum potassium was <3 or >5.8 mmol/L. 

These laboratory tests were done at run-in visit 1, prior to visit 2, visit 3 (6 weeks), visit 7 

(2 years) and final visit at end of study for ONTARGET and TRANSCEND trials, and 

visit 1 (baseline, at randomization) and visit 2 (1 month ± 7 days) in PRoFESS trial. 

During the run-in period: 

21.8% of patients in ONTARGET trial had baseline creatinine above upper limit of 

reference range (≥1.2 mg/dL), and 8.1% (2,074) patients had increases in creatinine 

level ≥1.5 mg/dL (defined as PCSA). 

Similarly, 20.8% patients in TRANSCEND trial had baseline creatinine above upper limit
of reference range (≥1.2 mg/dL), and 7.0% (415) patients had increases in creatinine
level ≥1.5 mg/dL (defined as PCSA).
During the randomized period:
Patients in both ONTARGET and TRANSCEND trials had small changes from baseline
in mean potassium and mean creatinine. Mean baseline values for total cholesterol,
HDL cholesterol, LDL cholesterol and triglycerides were about the same in all treatment
groups. Small increases in glucose were observed over the randomized period for both
T and R treatment groups. Analysis of laboratory parameters and change from baseline
values by subgroups of age, sex, ethnicity, obesity, diabetes, hypertension and impaired
renal function did not reveal any unexpected or clinically important findings.
A similar trend was found with laboratory data in the PRoFESS trial.
Serum creatinine: In ONTARGET, the mean increases in serum creatinine from
baseline to the end of the study were higher in the T/R group (0.12 mg/dL) than in the T
(0.09 mg/dL) or R group (0.06 mg/dL). The proportions of patients with transitions from
low or normal to high creatinine values (T/R: 22.6%; T: 18.4%; R: 16.1%) and with
possibly clinically significant increases (≥1.5 mg/dL) in creatinine (T/R: 18.5%; T: 15.7%;
R: 13.0%) were highest in T/R group, lower in T group, and lowest in R group.
In the TRANSCEND trial, too, PCSAs in creatinine were more frequently observed in
the T group than in the PBO group at any time during the randomized period (T: 13.8%
and PBO 11.2%).
The PRoFESS trial showed the same trend as for creatinine in the ONTARGET and
TRANSCEND trials: 3.8% of patients of the T group were reported with an increase to
>1.5 mg/dL at 1 month vs. 3.1% of patients in the PBO group. There was no difference
regarding doubling of creatinine levels at 1 month between the treatment groups.
Serum potassium: PCSAs in potassium were comparable between the T group (1.2%)
and the R group (1.1%) in the ONTARGET trial, but were more frequent in the T group
(1.8%) than the PBO group (0.5%) in the TRANSCEND trial.
In the PRoFESS trial, the frequencies of patients with clinically significant decreased or
increased serum potassium were comparable between treatment groups. In the T



                                              119 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

group, 1.5% of patients were reported with an increase in potassium levels to >5.5
mmol/L at 1 month (vs. 0.8% of patients in the PBO group).
Other laboratory test values: PCSAs in glucose, total cholesterol, low density
lipoprotein, high density lipoprotein, and triglycerides were comparable between
treatment groups in the ONTARGET as well as the TRANSCEND trial.
The analyses of AEs and clinical laboratory parameters in subgroups in the
ONTARGET and TRANSCEND trials did not reveal unexpected differences from the
overall study population.

7.4.3   Vital Signs
Because of the anti-hypertensive effect of telmisartan and ramipril, low BP and
hypotension were sought for as a possible side effect of these treatments. Markedly low
BP was defined as sitting SBP <100 mmHg, sitting DBP <60 mmHg or both sitting SBP
<100 and DBP <60 mmHg.
Markedly low BP was observed with more frequently in the T/R group, but with low
frequency in similar proportions of patients in the T and the R groups in ONTARGET
trial (0.5 – 1.2% and 0.5 – 0.9%, respectively), and in the T and PBO groups in
TRANSCEND trial (0.1% ~ 0.3%)
However, in the PRoFESS trial, markedly low BP was observed with higher frequency
and in relatively larger proportion of patients in the T group (1.1 ~ 1.2%) compared to
the PBO group (0.1 to 0.2%).

7.4.4   Electrocardiograms (ECGs)
At baseline, abnormal ECGs were found in almost equal proportions of patients in
ONTARGET (T: 48.1%, R: 48.2%), TRANSCEND (T: 46.7%, R: 45.5%) and PRoFESS
(T: 5.9%, R: 5.4%) trials.
Over the randomized period till the end of the trials, the proportion of patients with
abnormal ECGs remained similar to baseline in ONTARGET and TRANSCEND trials,
with no clinically significant differences in ECG characteristics between treatment
groups. In the PRoFESS trial, there were too many missing ECGs to be able to analyze.
The sponsor sought information regarding regression of left ventricular hypertrophy
(LVH): no differences between treatment groups were found in ONTARGET and
TRANSCEND trials (6.3% in the latter). The proportion of patients who new developed
LVH during the randomized period were:
 •	 similar between treatment groups in ONTARGET,
 •	 relatively smaller in the T group (4.3%) compared to the PBO group (6.9%) in the 

    TRANSCEND trial, and 





                                          120 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

•	 there were more patients who had regression of LVH in the T group (8.3%)
   compared to the PBO group (6.7%), and fewer patients developed new LVH in the
   T group (2.9%) compared to the PBO group (3.5%), in the PRoFESS trial.



7.5 Other Safety Explorations

7.5.1   Dose Dependency for Adverse Events
This was not studied because only one dose of each drug (telmisartan and/or ramipril)
was used in these three large clinical outcome trials. In general, it is known that there
are no dose-dependent side effects with ARBs or ACE-Is.

7.5.3   Drug-Demographic Interactions
In all three clinical outcome trials, there was no interaction observed for:
•	 Drug and age,
•	 Drug and sex,
•	 Drug and ethnicity (most patients were white or Asians, with very low numbers of
    black patients) or
•	 Drug and obesity.

7.5.4   Drug-Disease Interactions
In all three clinical outcome trials, there was no interaction observed for:
•	 Drug and presence of hypertension, incidences of SAEs being similar in patients
    with and without hypertension;
•	 Drug and presence of diabetes, the incidences of SAEs being – as expected –
    higher in patients with diabetes than in patients without diabetes, with no differences
    between treatment groups in the incidences of SAEs for diabetic and non-diabetics;
•	 Drug and impaired renal function, the incidences of SAEs being – as expected –
    higher in patients with renal impairment than in patients without renal impairment,
    with no differences between treatment groups in the incidences of SAEs in patients
    with or without renal impairment.

7.5.5   Drug-Drug Interactions
Subgroup analyses of drug interactions with use of statins (and by dose of simvastatin),
β-blockers, α-blockers, diuretics, dihydropyridines (excluding diltiazem/verapamil),
diltiazem/verapamil, antihypertensives and digitalis in the ONTARGET and
TRANSCEND trials showed no differences in these subgroups compared with the
overall population of patients, and no differences between the treatment groups.




                                           121 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

7.6 Additional Safety Evaluations

7.6.2   Human Reproduction and Pregnancy Data
Patients randomized in ONTARGET and TRANSCEND trials were ≥ 55 years of age,
and those in the PRoFESS trial ≥ 50 years of age. They were unlikely to experience
pregnancies, and pregnancies were not reported.

7.6.4   Overdose, Drug Abuse Potential, Withdrawal and Rebound
The Patient Terms ‘overdose,’ ‘accidental overdose,’ or ‘multiple drug overdose’ were
reported in 23 patients in the 3 trials (ONTARGET n=15, TRANSCEND n=1, PRoFESS
n=7). All but 1 case required hospitalization, 2 cases were considered life-threatening,
and none of the cases were reported as fatal. All but 1 case reported overdoses of other
medications taken concomitantly. Only in 1 case (ONTARGET: R group), the reported
diagnosis (‘medication overdose’) did not explicitly specify other medication than the
study medications.
Telmisartan and ramipril are not considered to possess abuse potential. Three patients
were reported with possible abuse of study drugs. The diagnoses reported were:
•	 polysubstance abuse (ONTARGET: T/R group),
•	 psychiatric substance abuse (ONTARGET: R group), both resulting in hospitalization
•	 and multiple substance abuse (PRoFESS: T group), resulting in persistent or
   significant disability/incapacity.
The above diagnoses suggest that either other drugs or multiple drugs (probably
including telmisartan) were taken for drug abuse and therefore do not appear to indicate
any abuse potential of telmisartan.
WITHDRAWAL AND REBOUND: In ONTARGET and TRANSCEND trials, study drug
discontinuations of less than 5 days were not recorded; in the PRoFESS study
discontinuations of less than 28 days were not recorded. The 6-monthly visit frequency
did not allow an assessment of withdrawal and rebound phenomena.

7.7 Additional Submissions / Safety Issues

7.7.1 Sepsis and fatal sepsis as a new potential safety signal in patients treated
with telmisartan
On 21-Nov-2008, the sponsor submitted integrated safety data from three large
outcome studies (ONTARGET, TRANSCEND and PRoFESS) in which sepsis was
identified as a new potential safety signal in patients treated with telmisartan (SN 0643;
supporting document # 717 to IND 42,450). The sponsor proposed to add to the
ADVERSE REACTIONS section of the prescribing information the following:


                                           122 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

      Cardiovascular Risk Reduction Trials: 

      In clinical studies with patients at high risk of developing major cardiovascular 

      events, cases of sepsis, including some with fatal outcomes, have been reported. 


We summarize the findings related to sepsis and fatal sepsis in these three large clinical
outcome trials in Table 68:


Table 68 Number (%) of patients with sepsis and fatal sepsis in large clinical
outcome trials of telmisartan
  Clinical trial Telmisartan Ramipril       Placebo          RR (CI)       p-value
                         Number (%) of patients with sepsis
  ONTARGET        70 (0.82%) 60 (0.70%)         --      1.17 (0.83 – 1.65) 0.3665
  TRANSCEND 25 (0.85%)            --       17 (0.57%) 1.48 (0.80 – 2.73) 0.2082
  PRoFESS         70 (0.70%)      --       49 (0.49%) 1.43 (1.00 – 2.06) 0.0513
                      Number (%) of patients with fatal sepsis
  ONTARGET        26 (0.30%) 27 (0.31%)         --      0.97 (0.56 – 1.66) 0.9020
  TRANSCEND 15 (0.51%)            --        9 (0.30%) 1.68 (0.73 – 3.83) 0.2142
  PRoFESS         33 (0.33%)      --       16 (0.16%) 2.07 (1.14 – 3.76) 0.0145


FDA reviewers’ comments: The definition of “sepsis” that was applied consists of the
following definition by ACCP/SCCM (American College of Chest Physicians/Society of
Critical Care Medicine) consensus statement sepsis which defines sepsis as present if a
patient has >1 of the following in the presence of positive evidence of bacteria in the
circulation: (i) body temperature >38°C or <36°C, (ii) heart rate >90 bpm, (iii)
hyperventilation evidenced by respiration rate >20/min or PaCO2 <32mmHg, or (iv)
WBC count >12,000 cells/µL or <4,000 cells/µL or >10% bands or immature
neutrophils).
The ONTARGET trial did not find an increased risk of sepsis with telmisartan compared
to ramipril. One caveat of this study report was that Asian patients on telmisartan had a
higher proportion of sepsis (telmisartan = 24.3% vs ramipril = 15%).
In both the TRANSCEND trial and the PRoFESS trial, there were more deaths
associated with sepsis in patients on telmisartan compared to placebo.
No explanation for the above finding is known. Both telmisartan and ramipril act on the
Renin-Angiotensin pathway and this may be a possible reason for finding no difference
in the incidence of sepsis between telmisartan and ramipril in the ONTARGET trial.
At the reviewers’ request, the sponsor submitted case report forms together with some
case narratives of these patients who experienced sepsis and fatal sepsis in the three
large outcome trials.




                                           123 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

The narratives have very limited information. A review of the CRFs suggests that these
patients had co-morbid illnesses (e.g., decubitus ulcers, urosepsis, gall stones, cardiac
heart replacements with predisposition to infective endocarditis, COPD and pneumonia,
different types of cancers requiring chemotherapy with consequent neutropenia or
surgery with post-surgical infections, diabetes with predisposition to skin infections,
etc.,) or confounding conditions (e.g., burns, motor vehicle accidents, cat scratch, falls,
etc. associated with sepsis).
Thus, we do not believe that there is no evidence of a “real” safety signal of sepsis
and/or fatal sepsis in patients treated with telmisartan. We did not find a signal of sepsis
and/or fatal sepsis in postmarketing data of over 18 million patient-years of treatment of
patients with hypertension.
The sponsor submitted a Risk Evaluation and Management Plan (REMS).

7.7.2 Malignancies as a new potential safety signal in patients treated with
telmisartan
In ONTARGET and TRANSCEND trials, malignancies (fatal and non-fatal) were
evaluated under efficacy as ‘other endpoints.’
In the ONTARGET trial, the hazard ratio for malignancies for T/R vs. R was 1.14 (95%
CI 1.03, 1.26; p=0.0089) in all randomized patients and 1.12 (95% CI 1.01, 1.25;
p=0.0366) in patients without cancer at baseline. There was no difference between T
and R regarding malignancies.
The sponsor attributed this finding of T/R being associated with a higher risk of cancer
compared to monotherapy with R to a chance finding due to multiple testing in this trial.
In the TRANSCEND trial, too, a higher rate of malignancies (overall and in patients
without cancer at baseline) was observed in patients treated with telmisartan than in
those treated with placebo. No differences between treatment groups were seen for
fatal malignancies.
So far, there is no evidence that ARBs are associated with a higher risk of
malignancies. However, we think that other large clinical outcome trials of ARBs and
ACE-Is should be evaluated to investigate if (i) there are similar findings to suggest a
new potential safety signal, and (i) a real risk for higher incidence of malignancies exist
when patients are treated with ARBs or ACE-Is at high (CV risk prevention) doses over
long periods of time.




                                            124 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

8 Postmarketing Experience
Table 69 Postmarketing exposure for the indication: treatment of hypertension
            Post marketing exposure by gender and age group
              Age Group            Exposure (patient years)
                                    M                  F
              0 – 18 years        0.03%             0.08%
             >18 – 40 years       2.26%             2.04%
             41 - <65 years      21.89%            22.37%
               ≥65 years         21.84%            29.48%
                  Total            46%               54%
               Post marketing exposure by dose strengths
                  Dose        Total tablets     Patient years
           Telmisartan 20 mg 504,951.084          1,382,481
           Telmisartan 40 mg 3,410,015,044        9,336.112
           Telmisartan 80 mg 3,004,11,882         8,224,947
                  Total       6,919,128,010      18,943,540

There is no post-marketing data available for telmisartan for the proposed indication.
However, from the Risk Management Plan submitted, I have compiled the post-
marketing (non-study) exposure data for the indication of hypertension in Table 69.


9 Appendices

9.1 Literature Review/References

9.1.1 Effect of ACE-Is on improving survival in patients with chronic coronary
artery disease

It is possible that quantitative differences exist among ACE-Is and that not all ACE-Is
are equally effective for all indications24. In a prospective, randomized study involving
patients who had had a myocardial infarction25, quinapril was significantly more effective
than enalapril (at doses previously shown to be effective in reducing the rate of CV
events among patients with LV systolic dysfunction) in reducing the concentration of C-
reactive protein26. One retrospective study demonstrated lower 1-year mortality after
acute MI in 7,512 Canadian patients treated with either ramipril or perindopril (Figure
27), compared with other ACE inhibitors27.




                                           125 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 27 Unadjusted Kaplan–Meier curves according to ACEI prescribed within
30 days of discharge. (Based on data from Ann Intern Med 2004; 141: 102-12)




                    P < 0.001 for the log-rank test.


In a cohort of 6,381 patients hospitalized with ST-segment elevation MI28, therapy with
ramipril versus other ACE inhibitors was independently associated with a lower 21-day
mortality rate [HR 0.54; 95% CI 0.32, 0.90] (Figure 28) and a lower rate of nonfatal
major adverse coronary and cerebrovascular events [HR 0.79; 95% CI 0.50, 1.27].


Figure 28 Survival of patients on ramipril therapy, other ACEI therapy, and no
ACEI therapy (Kaplan- Meier curves).




                                                       126 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


This differential efficacy among specific ACE inhibitors appears to extend to CAD
patients without a history of MI, as observed for ramipril in the HOPE study and for
perindopril in the EUROPA study. On the other hand, trandolapril in the PEACE trial
demonstrated no reduction in the primary composite endpoint of CV death, MI or
coronary revascularization [HR 0.96; 95% CI 0.88, 1.06, p = 0.43].
These findings suggest that a “class effect” may not exist among ACE inhibitors for the
treatment of stable CAD, and that, treatment with either 10 mg/day of ramipril or 8
mg/day of perindopril remains the only evidence-based strategy to reduce CHD events
in this population of patients with high CV risks.
The results of ACE-I trials in the medical literature pertaining to patients with high CV
risk, CHF and MI are summarized in Table 70.


Table 70 Summary of findings in the medical literature regarding issues related
to the effect of ACE-Is in patients with heart disease
                       Issue                                       Evidence from Clinical Trials*
                                                         Chronic CAD                CHF         Post-MI
Are ACE-Is useful to reduce CV events (in placebo-      HOPE, EUROPA,       ?SOLVD          SAVE., AIRE,
controlled trials)?                           Yes       PROGRESS                            SMILE, TRACE
                                                 No     PEACE
Are ACE-Is useful to reduce CV events (in      Yes
active comparator-controlled trials)?                   ALLHAT, CAPPP
                                                No      ANBP2, STOP-2,
                                                        CAMELOT
To reduce CV events, is the dose of ACE-I      Yes                          ATLAS
important?                                                                  NETWORK,
                                                No                          HEDS,CHARM
                                    Dose was not        HOPE, ANBP2,                        SAVE, AIRE,
                                       addressed        EUROPA, PEACE                       SMILE, TRACE
Are there qualitative differences between ACE-Is?       HOPE, EUROPA                        Canadian study,
                                                                                            STEMI-study
                                               Yes
                                                No
Is there intolerance to ACE-Is?                Yes                          SPICE,SOLVD
                                                 No
    *Clinical trial acronym; CAD = coronary artery disease 

    Source: FDA reviewers’ compilation of ACE-I trials from medical literature. 





                                                     127 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

9.1.2 Dose considerations for Angiotensin Converting Enzyme Inhibitors (ACE-
Is)

There are three clinical trials which examined high vs. low dose ACE-Is in patients with
heart failure showing conflicting results.
The ATLAS (Assessment of Treatment with Lisinopril And Survival)29 study evaluated
the effect of high dose lisinopril (32.5 to 35 mg/day, n= 1,568) vs. low dose lisinopril (2.5
to 5 mg/day, n = 1,596) in the treatment of 3,164 patients with CHF (NYHA class III and
LVEF ≤ 0.30) with a 39 –58 months follow-up time. High dose lisinopril produced
significant reductions in the relative risk of the composite endpoint of death or
hospitalization for any reason (HR 0.88, 95% CI 0.82, 0.96; P=0.002), and the
composite endpoint of all-cause deaths or CHF hospitalizations (HR 0.85, 95% CI 0.78,
0.93; P<0.001), compared with the low-dose lisinopril regimen.
In contrast, the NETWORK (Clinical Outcome with Enalapril in Symptomatic Chronic
Heart Failure) trial 30 found no differences between high-dose and low-dose treatment
groups for any of the endpoints measured among 1,532 patients with NYHA class II
(65% of patients) to class III/IV (35% of patients) heart failure randomized to receive
enalapril 2.5 mg b.i.d., 5 mg b.i.d. or 10 mg b.i.d., followed up for 24 weeks.
The High Enalapril Dose Study Group trial31 also found no differences in deaths [HR
0.998; 95% CI 0.56, 1.79, p=0.995] or a composite endpoint of death and hospital
admission when comparing the treatment of 248 patients with advanced CHF treated
with enalapril at a dose of 20 mg/day (Group 1: mean dose achieved 17.9 ± 4.3 mg/
day, n=122) vs. 60 mg/day (Group 2: mean dose achieved 42 ± 19.3 mg/day, n=126). It
is possible that even maximally recommended doses of ACE-Is do not completely
prevent ACE-mediated formation of angiotensin II in the heart in patients with CHF32.
Thus, the general consensus is that for ACE-Is to achieve a reduction in CV events,
larger than standard doses used for the control of hypertension, and probably at the full
doses shown to be effective in large clinical outcome trials such as HOPE, EUROPA,
etc., must be used.


9.1.3 Dose considerations for Angiotensin Receptor Blockers (ARBs)
In two clinical trials of losartan16,17 in which the dose was increased gradually to 100 mg
per day in asymptomatic patients with hypertension and ECG evidence of left ventricular
hypertrophy, a significant survival benefit among high-risk patients was observed.
In the LIFE (Losartan Intervention For Endpoint reduction in hypertension)16 trial, 9,193
participants 55-80 years old with essential hypertension and left ventricular hypertrophy
documented by ECG, were randomly assigned to receive losartan (titrated to 100 mg)
or atenolol (titrated to 100 mg) once daily. A significant reduction (by 15%, P = 0.009) in
the primary composite endpoint of CV mortality, stroke and MI (the HOPE endpoint)



                                            128 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

was found in the subjects treated with losartan.
In the RENAAL (Reduction of Endpoints in NIDDM with the Angiotensin II Antagonist
Losartan)17 trial, 1,513 patients with type II diabetes and nephropathy were randomized
to losartan (50-100 mg once daily) or placebo, and received conventional
antihypertensive treatment, for a mean of 3.4 years. Losartan significantly reduced the
primary endpoint (composite of a doubling of the base-line serum creatinine
concentration, end-stage renal disease, or death: HR 0.84; 95% CI 0.72, 0.98; P=0.02),
the incidence of doubling of serum creatinine concentration (HR 0.75; 95% CI 0.61,
0.92; P=0.006), and end-stage renal failure (HR 0.72, 95% CI 0.58, 0.89; P=0.002).
In clinical trials where lower doses of ARBs were used, a survival benefit was not found.
In VALIANT (VALsartan In Acute Myocardial INfarction Trial)15, 14,808 patients were
randomized (1:1:1 ratio) to receive valsartan (titrated to 160 mg b.i.d.), captopril (titrated
to 50 mg t.i.d.) or the combination of valsartan (titrated to 80 mg b.i.d.) plus captopril
(titrated to 50 mg t.i.d.), beginning 12 hours to 10 days after a myocardial infarction, and
followed up to a median of 24.7 months. This study was designed to assess non-
inferiority of valsartan relative to captopril. All-cause mortality was 19.9% in the
valsartan group, 19.5% in the captopril group and 19.3% in the combination (valsartan­
and-captopril) group. The hazard ratio for death in valsartan group vs. captopril group
was 1.00 (97.5% CI: 0.90, 1.11, P=0.98), and the hazard ratio for death in valsartan plus
captopril group vs. captopril group was 0.98 (97.5% CI: 0.89, 1.09, P=0.73).
The ELITE II (Evaluation of Losartan In The Elderly)13 trial showed no survival
advantage of losartan over captopril; the dose of losartan used was thought to be
insufficient to substantially block the AT1 receptor, and was the reason, at least in part,
for this lack of effect.
The OPTIMAAL (OPtimal Trial In Myocardial Infarction with the Angiotensin II
Antagonist Losartan) trial14 compared losartan (50 mg/day) vs. captopril (150 mg/day) in
high-risk patients with acute MI. The results were in favor of captopril for all-cause
mortality (not significant, P = 0.069) and for CV mortality (P=0.032). An insufficient dose
of losartan was a reason for the failure to show superiority of losartan over captopril.
Thus, it appears that:
•	 an insufficient dose of an ARB used in ELITE II, OPTIMAAL, and VALIANT trials
    was the reason for an observed lack of beneficial effect, and
•	 a significant survival benefit in high risk patients was observed when relatively larger
    doses of ARBs were used as in LIFE and RENAAL trials.



9.1.4 ARBs as substitutes for ACE-Is in treatment of ACE-I intolerant patients
The Study of Patients Intolerant to Converting Enzyme Inhibitors (SPICE)33 showed that
84% of patients with chronic heart failure and LVEF <35% who are intolerant to ACE-Is
were able to tolerate candesartan (4 mg once/day, titrated to 16 mg once/day);


                                            129
Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

however, the mortality and all-cause hospitalization were not significantly different
between candesartan and the placebo groups in this small study of 270 patients.
In Table 71, we have summarized the results of ARB trials in the medical literature
pertaining to patients with chronic CAD.

Table 71 Role of ARBs in patients with CAD and/or high risk of CVD
                               Issue                                   Clinical trials* in patients with CAD
                                                                                   or heart failure
Are ARBs better than placebo?                                  Yes    SCOPE, CHARM-Alt, RENAAL, IDNT
                                                               No     TRANSCEND, PRoFESS
Is dose of ARB important for the treatment of chronic          Yes    High: RENAAL, LIFE
CAD?                                                                  Low: ELITE II, OPTIMAAL, VALIANT
                                                               No     ---
Are ARBs as useful as ACE-Is in ACE-intolerant patients? Yes          CHARM-Alternative§
                                                               No     SPICE§
Are ARBs as useful as ACE-Is in chronic CAD?                   Yes    LIFE, RENAAL
                                                               No     ---
Are ARBs harmful in causing increased MI events?              Yes     VALUE, OPTIMAAL, DETAIL,
                                                                      CHARM-Alt
                                                               No     ---
Are ARBs additive over ACE-Is for survival in heart failure? Yes      CHARM-Added§
                                                               No     VALIANT§
Are ARBs additive over ACE-Is for survival in chronic CAD? Yes        ?RENAAL
                                                               No     ONTARGET
* Clinical trial acronym; CAD = coronary artery disease; §Patients with heart failure
     Source: FDA reviewers’ compilation of ACE-I trials from medical literature.




9.1.5 The issue of ARB-MI paradox
The major ARB trials in high risk patients show a lack of reduction in MI and mortality
despite significant reductions in BP, with paradoxical increased rates of MI observed in
Valsartan Antihypertensive Long-term Use Evaluation (VALUE)6 trial (19% increase in
relative risk of MI with valsartan vs amlodipine in 15,245 patients) and Candesartan in
Heart failure Assessment of Reduction in Mortality and morbidity Alternative (CHARM-
Alt) trial (11% increase in non-fatal MI compared to placebo)7.
A meta-analysis of 25 clinical trials of ARBs in 68,711 patients at risk for MI and >4000
events (Figure 29) showed that the ARBs do not increase the risk of MI with a pooled
OR of 1.03 (95% CI 0.93, 1.13, p=0.59)8.



                                                    130 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 29 ARBs do not increase risk of MI overall (Forest plot of MI analyses in
25 clinical trials of ARBs)




Source: Tsuyuki RT and McDonald, MA. Angiotensin Receptor Blockers do not increase risk of
myocardial infarction.Circulation 2006; 114: 855-60.



On the other hand, a meta-analysis of 11 randomized clinical trials of ARBs with at least
100 patients in each group treated for at least 6 months involving 55,050 patients
(Figure 30) showed that MI was significantly increased by 8% (pooled OR 1.08, 95% CI
1.01, 1.16, p=0.03)9, with the VALUE and CHARM-Alt trials driving the pooled results.
This issue is not helped by the ONTARGET, TRANSCEND and PRoFESS trials where
MIs were not accurately evaluated by the clinical trial sites or by adequate adjudication.
(Please see Section 6.1.5 Analysis of secondary endpoints.)




                                               131 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Figure 30 ARBs increase risk of MI overall (Forest plot of MI analyses in 12
clinical trials of ARBs)




Source: Strauss MH and Hall AS. Angiotensin Receptor Blockers may increase risk of myocardial
infarction: Unraveling the ARB-MI Paradox. Circulation 2006; 114: 838-54.



9.1.6 Are the effects of ARBs additive on top of ACE-inhibitors?
(That is, can incremental survival benefits be achieved in heart failure by using two
inhibitors (ACE-Is and ARBs) of the renin-angiotensin system?)
Theoretically, ACE-Is only block 13% of the total production of angiotensin II in the
human heart34. One or more ACE-independent pathways for the synthesis of
angiotensin II, including the “chymase pathway,” can produce about 90% of angiotensin
II in the heart35. Thus, local production of angiotensin II can occur despite use of an
ACE-I. ARBs, by inhibiting angiotensin II at the AT1-receptor level, may exert a more
complete inhibition of the local effects of angiotensin II. Thus, ACE-Is and ARBs may
exert different effects at the cardiac and vascular levels, which, in theory, may be
complementary in the treatment of CAD.
In the CHF population (of patients with ejection fraction ≤40%), the CHARM-Added
trial36 showed an additive effect of candesartan and ACE-I to reduce CV deaths and
heart failure hospitalizations (Please refer to FDA clinical review of CHARM-Added trial
and minutes of the Cardio-Renal Advisory Committee Meeting in February 2005).
We have not yet found in the medical literature a study where an ACE inhibitor and an
ARB are used together in patients who are at high CV risk but have no apparent
structural abnormality of the heart (i.e., no studies of use of ACE an inhibitor and an
ARB together among patients with hypertension and/or diabetes mellitus, and/or
dyslipidemia without an apparent structural abnormality of the heart).


                                                132 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

ONTARGET and TRANSCEND trials contain a number of patients who received
concomitant open-label ACE-I and/or ARB, but these patients were censored for data
analyses.
The only data of the effect of an ACE-I and an ARB together come from the T/R
combination arm of ONTARGET, which did not show an added benefit but showed
increased adverse events (hypotensive symptoms, treatment discontinuations).



9.2 Labeling Recommendations
Will defer to comment until after the AC meeting.

9.3 Advisory Committee Meeting
At this point in time, there is yet no consensus regarding the Non-Inferiority margin, the
amount to discount, the clinical outcome endpoint(s) (i.e., 4-fold or 3-fold composite
endpoint) to use in this patient population, and the number of clinical trials that should
be meta-analyzed (HOPE alone or HOPE plus EUROPA plus/not PEACE). These
issues will be addressed at a Cardio-renal Advisory Committee Meeting scheduled on
29-Jul-2009 for public discussion.

9.4 Non-essential tables of analyses of clinical trial data

9.4.1 Non-essential tables of analyses of follow-up clinical outcome data at 6
months and cumulatively at Year 1, 2, 3, 4, and 5
These analyses were performed according to the intent-to-treat principle. Thus,
discontinuations of study medication and/or addition of open label ACE-Is or ARBs, both
interventions which are likely to happen after an initial event, are not taken into account.
All but one of the confidence intervals for the hazard ratios between T and T (for all
endpoints and at all time points) includes unity (Table 77, for stroke in patients with
hospitalization for CHF at 5 years).




                                            133 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 72 Hazard ratios (95% CI) of cumulative outcomes in patients who had a
non-fatal MI: comparison of T/R vs R
                                      Relative Risk (95% CI) for T/R vs R
                   Death             Recurrent MI              Stroke          CHF hospitalization
6 months     1.18 (0.80 – 1.74)    0.93 (0.54 – 1.60)    0.54 (0.16 – 1.85)      1.23 (0.68 – 2.23)
1 year       1.18 (0.82 – 1.70)    1.05 (0.67 – 1.67)    0.76 (0.30 – 1.93)      1.20 (0.71 – 2.02)
2 years      1.15 (0.82 – 1.61)    1.16 (0.77 – 1.75)    0.89 (0.43 – 1.85)      1.17 (0.74 – 1.87)
3 years      1.08 (0.79 – 1.48)    1.22 (0.82 – 1.80)    1.02 (0.51 – 2.07)      1.14 (0.74 – 1.77)
4 years      1.13 (0.83 – 1.55)    1.26 (0.86 – 1.85)    1.09 (0.54 – 2.18)      1.16 (0.76 – 1.77)
5 years      1.13 (0.83 – 1.55)    1.21 (0.83 – 1.77)    1.22 (0.62 – 2.39)      1.16 (0.76 – 1.77)
Source: Data tables on pages 1 – 6 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.


Table 73 Hazard ratios (95% CI) of cumulative outcomes in patients who had a
non-fatal MI: comparison of T vs R
                                       Relative Risk (95% CI) for T vs R
                   Death             Recurrent MI              Stroke          CHF hospitalization
6 months     1.25 (0.85 – 1.82)    0.72 (0.41 – 1.28)    1.61 (0.63 – 4.00)      1.17 (0.64 – 2.13)
1 year       1.36 (0.96 – 1.94)    0.80 (0.49 – 1.30)    1.32 (0.59 – 2.97)      1.04 (0.60 – 1.78)
2 years      1.29 (0.93 – 1.79)    0.78 (0.50 – 1.23)    1.01 (0.50 – 2.04)      1.01 (0.63 – 1.64)
3 years      1.10 (0.81 – 1.51)    0.84 (0.55 – 1.28)    1.07 (0.53 – 2.14)      0.90 (0.57 – 1.42)
4 years      1.16 (0.85 – 1.58)    0.84 (0.55 – 1.28)    1.07 (0.53 – 2.14)      0.90 (0.58 – 1.41)
5 years      1.14 (0.84 – 1.56)    0.83 (0.55 – 1.25)    1.08 (0.54 – 2.15)      0.93 (0.59 – 1.44)
Source: Data tables on pages 1 – 6 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.


Table 74 Hazard ratios (95% CI) of cumulative outcomes in patients who had a
non-fatal stroke: comparison of T/R vs R
                                      Relative Risk (95% CI) for T/R vs R
                   Death                   MI            Recurrent stroke      CHF hospitalization
6 months     1.17 (0.83 – 1.66)    0.69 (0.16 – 2.89)    1.04 (0.44 – 2.44)      1.43 (0.39 – 5.34)
1 year       1.17 (0.83 – 1.65)    1.38 (0.42 – 4.53)    0.78 (0.39 – 1.59)      1.38 (0.42 – 4.52)
2 years      1.15 (0.84 – 1.57)    1.16 (0.43 – 3.09)    0.86 (0.49 – 1.51)      1.00 (0.36 – 2.77)
3 years      1.11 (0.82 – 1.50)    1.40 (0.60 – 3.24)    0.83 (0.48 – 1.41)      1.28 (0.52 – 3.15)
4 years      1.12 (0.84 – 1.51)    1.17 (0.53 – 2.61)    0.88 (0.52 – 1.47)      1.05 (0.45 – 2.47)
5 years      1.12 (0.84 – 1.50)    1.27 (0.58 – 2.78)    0.93 (0.56 – 1.54)      1.06 (0.45 – 2.47)
Source: Data tables on pages 7 – 12 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.




                                                134 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

Table 75 Hazard ratios (95% CI) of cumulative outcomes in patients who had a
non-fatal stroke: comparison of T vs R
                                       Relative Risk (95% CI) for T vs R
                   Death                   MI            Recurrent stroke      CHF hospitalization
6 months     0.97 (0.67 – 1.40)    0.89 (0.24 – 3.33)    1.16 (0.50 – 2.68)      1.44 (0.39 – 5.36)
1 year       1.04 (0.73 – 1.49)    1.57 (0.50 – 4.93)    0.98 (0.51 – 1.91)      1.61 (0.51 – 5.08)
2 years      1.02 (0.74 – 1.40)    1.69 (0.69 – 4.13)    0.93 (0.53 – 1.62)      1.58 (0.63 – 3.92)
3 years      1.00 (0.74 – 1.36)    1.35 (0.59 – 3.13)    1.08 (0.65 – 1.78)      1.92 (0.84 – 4.38)
4 years      1.02 (0.75 – 1.37)    1.41 (0.66 – 3.02)    1.05 (0.64 – 1.73)      1.78 (0.83 – 3.81)
5 years      1.01 (0.75 – 1.36)    1.41 (0.66 – 3.02)    1.03 (0.63 – 1.69)      1.78 (0.83 – 3.81)
Source: Data tables on pages 7 – 12 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.
Table 76 Hazard ratios (95% CI) of cumulative outcomes in patients who were
hospitalized due to CHF: comparison of T/R vs R
                                      Relative Risk (95% CI) for T/R vs R
                                                                                  Recurrent CHF
                   Death                   MI                  Stroke
                                                                                  hospitalization
6 months     1.21 (0.79 – 1.86)    1.44 (0.63 – 3.29)    0.55 (0.14 – 2.21)      0.85 (0.56 – 1.28)
1 year       1.18 (0.82 – 1.71)    1.57 (0.75 – 3.28)    1.10 (0.39 – 3.15)      0.75 (0.53 – 1.06)
2 years      1.17 (0.84 – 1.62)    1.37 (0.72 – 2.59)    1.10 (0.44 – 2.78)      0.83 (0.62 – 1.12)
3 years      1.18 (0.87 – 1.59)    1.28 (0.70 – 2.33)    1.01 (0.43 – 2.37)      0.83 (0.62 – 1.12)
4 years      1.21 (0.91 – 1.60)    1.21 (0.69 – 2.13)    1.20 (0.55 – 2.63)      0.82 (0.61 – 1.10)
5 years      1.20 (0.91 – 1.60)    1.21 (0.69 – 2.13)    1.19 (0.54 – 2.61)      0.83 (0.62 – 1.11)
Source: Data tables on pages 13 – 18 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.
Table 77 Hazard ratios (95% CI) of cumulative outcomes in patients who were
hospitalized due to CHF: comparison of T vs R
                                       Relative Risk (95% CI) for T vs R
                                                                                  Recurrent CHF
                   Death                   MI                  Stroke
                                                                                  hospitalization
6 months     1.02 (0.67 – 1.56)    0.61 (0.23 – 1.61)    1.39 (0.50 – 3.91)      0.76 (0.51 – 1.14)
1 year       1.02 (0.71 – 1.47)    1.25 (0.60 – 2.61)    1.74 (0.69 – 4.35)      0.74 (0.53 – 1.02)
2 years      1.13 (0.82 – 1.55)    1.09 (0.57 – 2.06)    2.08 (0.95 – 4.56)      0.83 (0.62 – 1.12)
3 years      1.05 (0.79 – 1.41)    1.05 (0.58 – 1.91)    1.95 (0.95 – 4.00)      0.85 (0.64 – 1.13)
4 years      1.00 (0.75 – 1.32)    1.03 (0.59 – 1.80)    1.78 (0.89 – 3.59)      0.86 (0.65 – 1.13)
5 years      1.03 (0.78 – 1.35)    1.03 (0.59 – 1.80)    2.09 (1.06 – 4.12)      0.86 (0.65 – 1.13)
Source: Data tables on pages 13 – 18 of Attachment 4 in Sponsor’s response submitted 14-Apr-2009.




                                                135 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets

                                                    REFERENCES
1
    The Heart Outcomes Prevention Evaluation (HOPE) Study Investigators. Effects of an angiotensin-converting­
     enzyme inhibitor, ramipril, on cardiovascular events in high-risk patients. N Engl J Med 2000; 342: 145-153.
2
    The Heart Outcomes Prevention Evaluation (HOPE) Study Investigators. Effects of ramipril on cardiovascular and
     microvascular outcomes in patients with diabetes mellitus: results of the HOPE study and MICRO-HOPE sub-
     study. Lancet 2000; 355: 253-259 [Erratum, Lancet 2000; 356: 860-861].
3
    The European Trial on Reduction of Cardiac Events with Perindopril in Stable Coronary Artery Disease
     Investigators. Efficacy of perindopril in reduction of cardiovascular events among patients with stable coronary
     artery disease: randomized, double-blind, placebo-controlled, multicenter trial (the EUROPA study). Lancet 2003;
     362:782-788.
4
  The PEACE Trial Investigators. Angiotensin-Converting–Enzyme Inhibition in Stable Coronary Artery Disease. N
   Engl J Med 2004;351:2058-68.
5
  Diener H-C, Bogousslavsky J, Brass LM, Cimminiello C, Csiba L, Kaste M, Leys D, Matias-Guiu J, Rupprecht H-J,
   on behalf of MATCH investigators. Management of Atherothrombosis with Clopidogrel in High-Risk Patients with
   Recent Transient Ischaemic Attack or Ischaemic Stroke (MATCH): Study Design and Baseline Data.
   Cerebrovascular Dis 2004; 17: 253-261.
6
  Julius S, Kjeldsen SE, Weber M, Brunner HR, Ekman S, Hansson L, Hua T, Laragh J, McInnes GT, Mitchell L, Plat
   F, Schork A, smith B, Zanchetti A. Outocmes in hypertensive patients at high cardiovascular risk treated with
   regimens based on valsartan or amlodipine: the VALUE randomized trial. Lancet 2004; 363: 2022-31.
7
  Granger CB, McMurray JJ, Yusuf S, Held P, Michelson EL, Olofsson B, Ostergren J, Pfeffer MA, Swedberg K.
   Effects of candesartan in patients with chronic heart failure and reduced left-ventricular systolic function intolerant
   toangiotensin-converting-enzyme inhibitors: the CHARM-Alternative trial. Lancet. 2003;362:772–776.
8
  Tsuyuki RT and McDonald, MA. Angiotensin receptor blockers do not increase risk of myocardial infarction.
   Circulation 2006; 114:855-60.
9
  Strauss MH and Hall AS. Angiotensin receptor blockers may increase risk of myocardial infarction: unraveling the
   ARB-MI paradox. Circulation 2006; 114: 838-54.
10
   Cohn JN, Tognoni G for the Valsartan Heart Failure Trial (ValHeFT) Investigators. A randomized trial of the
   angiotensin-receptor blocker valsartan in chronic heart failure. N Engl J Med 2001; 345: 1667-75.
11
   Gutrie R. Recent advances in cardiovascular risk reduction: implications of ONTARGET. Clinical Cornerstone
   2009; 9[Suppl 3]:S18-S26.
12
   Delahaye F, de Gevigney G. Is the optimal dose of angiotensin-converting enzyme inhibitors in patients with
   congestive heart failure definitely established? J Am Coll Cardiol 2000; 36: 1096-97.
13
     Pitt B, Poole-Wilson PA, Segal R, Martinez FA, Dickstein K, Camm AJ, Konstam MA, Riegger G, Klinger GH,
     Neaton J, Sharma D, Thiyagarajan B: Effect of losartan compared with captopril on mortality in patients with
     symptomatic heart failure: randomized trial – the Losartan Heart Failure Survival Study (ELITE II). Lancet 2000;
     355:1582-1587.
14
     Dickstein K, Kjekshus J, and the OPTIMAAL Steering Committee for the OPTIMAAL Study Group. Effects of
     losartan and captopril on mortality and morbidity in high-risk patients after acute myocardial infarction: the
     OPTIMAAL randomized trial. Optimal Trial in Myocardial Infarction with Angiotensin II Antagonist Losartan. Lancet
     2002; 360: 752-760.
15
     Pfeffer MA, McMurray JJV, Velazquez EJ, Rouleau J-L, et al for Valsartan in Acute Myocardial Infarction Trial
     Investigators (VALIANT). Valsartan, captopril or both in myocardial infarction complicated by heart failure, left
     ventricular dysfunction or both. N Engl J Med 2003; 349: 1893-1906.
16
     Dahlof B, Devereux RB, Kjeldsen SE, et al. Cardiovascular morbidity and mortality in the Losartan Intervention For
     Endpoint reduction in hypertension study (LIFE): a randomized trial against atenolol. Lancet 2002;359:995-1003.
17
   Brenner BM, Cooper ME, de Zeeuw D, et al: Effects of losartan on renal and cardiovascular outcomes in patients
   with type 2 diabetes and nephropathy. N Engl J Med 2001;345:861-869.
18
   The SOLVD Investigators. Effects of enalapril on mortality and the development of heart failure in asymptomatic
   patients with reduced left ventricular ejection fractions. N Engl J Med 1992; 327: 685– 91.
19
   Pfeffer MA, Braunwald E, Moye LA, Basta L, Brown EJ, Cuddy TE, Davis BR, Geltman EM, Goldman S, Flaker
   GC, on behalf of the SAVE investigators. Effect of captopril on mortality and morbidity in patients with left



                                                            136 

Clinical Review
Khin Maung U, M.D. & Jialu Zhang, Ph.D.
NDA 20-850 Efficacy Supplement SE1-025
Micardis® (telmisartan) tablets


   ventricular dysfunction after myocardial infarction: results of the Survival and Ventricular Enlargement (SAVE)
   Trial. N Engl J Med 1992; 327: 669-677.
20
   The Acute Infarction Ramipril Efficacy (AIRE) Study Investigators. Effect of ramipril on mortality and morbidity of
    survivors of acute myocardial infarction with clinical evidence of heart failure. Lancet 1993;342:821-8.
21
   Køber L, Torp-Pedersen C, Carlsen JE, Bagger H, Eliasen P, Lyngborg K, Videbaek J, Cole DS, Auclert L, Pauly
   NC, for the Trandolapril Cardiac Evaluation (TRACE) Study Group. A clinical trial of the angiotensin-converting
   enzyme inhibitor trandolapril in patients with left ventricular dysfunction after myocardial infarction. N Engl J Med
   1995; 333: 1670-1676.
22
   CPMP Committee for proprietary medicinal products (EMEA). Note for guidance on clinical investigation of
   medicinal products for the treatment of cardiac failure. CPMP/EWP, 235/95, Rev 1.
23
   Yusuf S Pogue J. Ace Inhibition In Stable Coronary Artery Disease (Let). N Engl J Med 2005;352:937-8.
24
   Furberg CD, Pitt B. Are all angiotensin-converting enzyme inhibitors interchangeable? J Am Coll Cardiol 2001; 37:
   1456-60.
25
   Tsikouris JP, Suarez JA, Simoni JS, Ziska M, Meyerrose GE. Exploring the effects of ACE inhibitor tissue
   penetration on vascular inflammation following acute myocardial infarction. Coron Artery Dis 2004; 15: 211-7.
26
   Ridker PM. Clinical application of C-reactive protein for cardiovascular disease detection and prevention.
   Circulation 2003; 107: 363-9.
27
   Pilote L, Abrahamowicz M, Rodrigues E, Eisenberg MJ, Rahme E. Mortality rates in elderly patients who take
   different angiotensin-converting enzyme inhibitors after myocardial infarction: a class effect? Ann Intern Med 2004;
   141: 102-12
28
   Wienbergen H, Schiele R, Gitt A, et al. Impact of ramipril versus other angiotensin-converting enzyme inhibitors on
   outcome of unselected patients with ST-elevation acute myocardial infarction. Am J Cardiol 2002; 90: 1045-9.
29
   Packer M, Poole-Wilson PA, Armstrong PW, Cleland JG, Horowitz JD, Massie BM, Ryden L, Thygesen K, on
   behalf of the ATLAS Study Group. Comparative effects of low and high doses of the angiotensin-converting
   enzyme inhibitor, lisinopril, on morbidity and mortality in chronic heart failure. Circulation 1999; 100: 2312-8.
30
   NETWORK Investigators: Clinical outcome with enalapril in symptomatic chronic heart failure: a dose comparison.
   Eur Heart J 1998;19:481-9.
31
   Nanas JN, Alexopoulos G, Anastasiou-Nana MI, et al. Outcome of patients with congestive heart failure treated
   with standard versus high doses of enalapril: a multicenter study. J Am Coll Cardiol 2000; 36: 2090-2095.
32
   Jorde UP, Ennezat PV, Lisker J, et al. Maximally recommended doses of angiotensin-converting enzyme (ACE)
   inhibitors do not completely prevent ACE-mediated formation of angiotensin II in chronic heart failure. Circulation
   2000; 101:844-846.
33
   Granger CB, Ertl G, Kuch J, Maggioni AP, McMurray J, Rouleau J-L, Stevenson LW, Swedberg K, Young J, Yusuf
   S, Califf RM, Bart BA, Held P, Michelson EL, Sellers MA, Ohlin G, Sparapani R, Pfeffer, MA for the Study of
   Patients Intolerant to Converting Enzyme Inhibitors (SPICE) Investigators. Randomized trial of candesartan
   cilexetil in the treatment of patients with congestive heart failure and a history of intolerance to angiotensin-
   converting enzyme inhibitors. Am Heart J 2000; 139; 609-17.
34
   Urata H, Healy B, Stewart RW, Bumpus M, Hussain A. Angiotensin II-forming pathways in normal and failing
   human hearts. Circ Res 1990; 66: 883-90.
35
   Urata H, Kinoshita A, Misono KS, Bumpus FM, Husain A. Identification of a highly specific chymase as the major
   angiotensin II- forming enzyme in the human heart. J Biol Chem 1990; 265(36): 22348-57.
36
   Solomon, SD, Wang D, Finn P, Skali H, Zornoff L, McMurray JJV, Swedberg K, Yusuf S, Granger CB, Michelson
   EL, Pocock S, Pfeffer MA. Effect on candesartan on cause-specific mortality in heart failure patients. The
   Candesartan in Heart failure Assessment and Reduction in Mortality and morbidity (CHARM) Program. Circulation
   2004; 110: 2180-3.




                                                         137 


				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:6
posted:9/10/2011
language:English
pages:142