Lisa Hirschhorn, M.D., M.P.H., Alexis Beattie, Duncan Davidson, Bruce Agins, M.D., M.P.H.

                                   September 9, 2005
Viral Load Report Contents

Panel Members

Executive Summary

Background from review of literature and guidelines

Report from the Expert Meeting
        Morning session
        o Morning overview (questions for the morning)
        o Viral load measurement and interpretation in clinical practice
        o Viral load as a measure of quality care in clinical practice
        o Viral suppression as a measure of quality
        o Use of viral load for patients not receiving antiretroviral therapy
        Afternoon session
        o Afternoon overview (questions for the afternoon)
        o Common themes
                 Defining quality measures using viral load levels in the general population
                 What are potential quality measures for viral load?
                                 a.        Viral suppression
                                 b.        Viral load change
                                 c.        Maintenance of viral load below a specified threshold
                                 d.        Viral load’s limitations as a quality measure
                                 e.        Viral load in the context of other measures
        o Group 1
                 How should viral load measurement be used as a measure of quality in the general population
                    of people with HIV? Can target rates for viral suppression be defined for general and specific
                 Viral suppression as a quality marker in specific populations
        o Group 2
                 What should the performance measure for viral load be in settings of different populations?
                         Naïve vs. experienced?
                         Patients with vs. without documented resistance?
                         Adherence challenged populations?
                         Others (HCV, for example)?
        o Group 3
                 How can we use viral load as a measure of quality in the absence of suppression? What
                    should the performance measure be in the absence of suppression? What other markers
                    would you need to measure quality of care in the absence of suppression?
                 Can viral load be used as a quality measure of populations at the public health level?

Proposed next steps and areas for operational research

                                         Viral Load Report

Since AZT was introduced in the late 1980s, health care providers in New York State, through a
comprehensive statewide HIV care delivery system, have successfully improved access to
antiretroviral treatment. With the evolution of antiretroviral therapy and its increasing
complexity, public health agencies, providers and consumers have struggled to identify
appropriate measures of quality to assess HIV care provided in hospitals and clinics. The
introduction and availability of viral load testing in clinical practice changed how treatment
decisions were made, and prompted development of guidelines to define its role as a marker to
guide both initiation and monitoring of antiretroviral therapy. However, its use as an indicator to
assess quality of care in a population has proven challenging.

With increasing attention to accountability for the use of public health funds and to measuring
the impact and quality of care, public health officials and clinicians have naturally looked for
simple, basic and relevant markers to use for these purposes. Viral load offers a seductive and
potentially simple measure that could easily be tracked and aggregated for assessment of
healthcare quality. Numerous attempts have been made to use viral load measurement in quality
of care monitoring activities in spite of limited published literature addressing the topic.
Proposals to use viral load suppression rates have been introduced by several authorities to
measure quality and provide a process for accountability. With limited experience in the use of
viral load for these purposes, providers and consumers alike have raised numerous concerns
about these suggestions. In response, the New York State Department of Health AIDS Institute
convened a panel to discuss this topic, review the existing literature addressing viral load
outcomes and develop recommendations to the state to guide its use. This report includes a
review of the pertinent literature, a summary of the presentations and discussion of the panel, and
the recommendations of the group.

                                  Viral Load Report

Panel Members
Judith Aberg, M.D. Bellevue Hospital Center
Bruce Agins, M.D. AIDS Institute
Julia Arnsten, M.D. Montefiore Medical Center
Alexis Beattie New York County Health Services Review
Sheldon Brown, M.D. Bronx VA Medical Center
Alvaro Carrascal, M.D. AIDS Institute
Bettina Carroll Medical & Health Research Association of NYC
Moupali Das-Douglas, M.D. NYC Department of Health & Mental Hygiene
Duncan Davidson New York County Health Services Review
Jay Dobkin, M.D. New York Presbyterian Hospital
Wendy Ferguson, RN IPRO
Robert Gass, MA, MPH AIDS Institute
Gregg Gonsalves Gay Men’s Health Crisis
Scott Hammer, M.D. New York Presbyterian Hospital
Steve Hemraj Central Harlem HIV Care Network
Lisa Hirschhorn, M.D. Havard Medical School and JSI Research & Training
Charles Hyman, M.D. Kings County Hospital Center
Vanessa Johnson NYS HIV Quality of Care Consumer Advisory Committee, Chair
Barbara Johnston, M.D. St. Vincent’s Hospital
Jessica Justman, M.D. Mailman School of Public Health, Columbia University
Jason Leider, M.D., Ph.D. North Bronx Health Care Network
Amneris Luque, M.D. University of Rochester
Joseph Masci, M.D. Elmhurst Medical Center
Patrick McGovern Harlem United Community AIDS Center, Inc.
Grace Moon NYC Department of Health & Mental Hygiene
Joanna Omi NYC Health and Hospitals Corporation
Sanjiv S. Shah, M.D. MetroPlus Health Plan
Lou Smith, M.D. NYS Department of Health, Division of Epidemiology
Clemens Steinbock, MBA AIDS Institute
Edward Telzak, M.D. Bronx Lebanon Hospital Center
Rona Vail, M.D. Callen-Lorde Community Health Center
Judith Verdino, MPA Medical & Health Research Association of NYC
Gregg Weinberg Medical & Health Research Association of NYC
Barry Zingman, M.D. Montefiore Medical Center

                                        Viral Load Report

Executive Summary
An expert panel of experienced HIV clinicians, public health officials and consumer
representatives was convened to address the use of viral load measurements to assess the quality
of HIV care and its use as a population-based outcome measure of quality. Background
information was presented based on a review of the literature. Data, including viral load levels,
were presented from the New York State HIV Quality of Care program. The expert panel was
asked to consider the role of viral load as a component of the measurement strategy to assess the
quality of ambulatory HIV care.

Discussion focused on the potential role of viral suppression and other indicators as measures of
quality in the general HIV population, as well as selected subpopulations. The group agreed that
viral suppression was the goal of antiretroviral therapy (ART) for treatment-naïve individuals,
but not a feasible goal for many other patients, including those with significant drug resistance or
poor adherence. The group concluded that absence of suppression is not always a marker for
poor quality of care because of the multiple patient and virologic factors which may render viral
suppression unachievable.

The interpretation of viral load levels in the context of measuring quality of care is complex, and
requires identification of those populations for whom suppression may neither be achievable nor
even the goal of treatment. These include patients with drug resistance, poor adherence and
those who choose not to accept antiretroviral therapy, all factors which would not be reflected
with simple measurement of viral load levels as an indicator of quality. Other key parameters,
such as CD4 count, resistance, treatment history and adherence, would need to be included to
assess quality of care provided to a clinic population. For some populations, the use of other
markers of virologic response, such as the extent of viral load decline or maintenance below a
specified threshold, would be necessary. Without looking at these factors, the use of viral
suppression rates as a marker for quality of care may prove to be a disincentive for clinics to
provide HIV care to patients who represent particular challenges to achievement of this goal,
particularly if funding decisions are based upon success in attaining high suppression rates.
Furthermore, the use of viral load suppression as a marker of quality of care for patients who are

                                          Viral Load Report

eligible but not yet on ART could lead to unintended pressure for clinicians to prescribe ART to
achieve these goals before patients are ready to start treatment.

As a general population quality measure, both viral suppression and viral load levels were
considered difficult to interpret because of the clinical heterogeneity of the population and
because expected rates of suppression have not been determined. Other parameters would need
to be assessed to understand situations where absence of suppression is not due to poor quality of
care. In addition, the inclusion of patients not on ART will further complicate the interpretation
of viral load suppression rates at the population level, since treatment is not recommended until
viral load levels are higher than 100,000 copies/mL. However, viral suppression rates may be
useful as a potential quality of care marker in the specific sub-population of ART-naïve HIV
patients infected with wild-type virus.

The use of viral load levels as population-based measures of quality was highlighted for future
research at both the clinic and community levels. Specific areas identified include defining
appropriate performance measures, additional data elements that would need to be collected, and
the resources that would be required to collect this information.

                                        Viral Load Report

II. Background1
1. Introduction
Since 1995, the use of highly active antiretroviral therapy (HAART) has revolutionized HIV care
and management. HAART has the potential to suppress viral replication, resulting in the
reduction of plasma viremia, often below the level of detection by commercially available tests.
This decline in viral load has been correlated with improvement in immunologic function, as
measured by rise in CD4 count, and decreased frequency of opportunistic infections, progression
to AIDS, and death.1-3 In all patients, higher viral loads are also correlated with increased risk
of HIV disease progression and death.4 The level of HIV in the blood, or viral load, has
therefore become both a marker to guide initiation of HAART and a measure of the efficacy of
HAART treatment in clinical trials and of the effectiveness of HAART in clinical care. The
uniform application of viral load to measure quality of care has not been extensively studied.

In 1990, the Institute of Medicine defined quality of care as the degree to which health services
for individuals and populations increase the likelihood of desired health outcomes and are
consistent with current professional knowledge.5 Evaluating and improving quality of care is a
critical activity for providers and public health agencies to ensure that effective HIV care and
treatment is being delivered to people with HIV. This importance has been highlighted in a
number of papers6, 7 and by leading organizations including the New York State Department of
Health (NYSDOH) AIDS Institute, the Health Resources and Services Administration (HRSA),
and the World Health Organization (WHO) as well as others. A number of indicators have been
proposed to measure the quality of HIV care, including use of HAART for eligible patients, rates
of opportunistic infection prophylaxis, and performance of screening tests for infections and
other diseases. While viral load suppression has also been used as a measure of quality of care in
some populations and studies, the validity of its use as a general measure of quality for all
patients or as a population health outcome measure remains unknown. In particular, given the
multiple factors necessary for suppression of plasma viremia and growing recognition that not all
patients receiving HAART can ultimately achieve viral suppression, the interpretation of the

 The following is a condensed version of the background paper that was distributed before the
expert panel discussion meeting. Hirschhorn L, Beattie A, Davidson D, Agins, B. The role of
viral load as a measure of the quality of care for people with HIV, 4/11/05.

                                         Viral Load Report

range of viral load levels found in a clinical setting in the context of evaluating quality continues
to be a challenging and complex endeavor.

3. Review of HIV viral load in the clinical context

A viral load test measures the amount of virus in plasma and correlates with the level of viral
replication in other parts of the body. Different technologies are used to measure viral load in
clinical practice. While initially, these different methods resulted in significant differences in
results, this variability has decreased over time due to advances in the test.

A degree of variability of plasma viral load levels normally occurs in any individual patient.8 It
is generally believed that this variation can be as much as 3-fold (or 0.5 log10). Therefore, only
a change greater than a 3-fold difference is considered a significant change in a person’s viral
load. In 2005 in the United States, the initiation of effective antiretroviral therapy (ART) is
probably the most common cause of significant change in viral load seen in clinical practice,
with decreases usually occurring within the first weeks of treatment. In the absence of ART, a
number of events can cause viral load levels to increase, although usually these are transient.
The most dramatic increases are seen with new infections, such as opportunistic infections,
tuberculosis, bacterial pneumonia and herpes reactivation. A number of studies have also found
that vaccinations may cause transient and clinically insignificant increases in viral load in some
patients.9 Therefore, viral loads should be interpreted with caution when samples are drawn
within 4-6 weeks of vaccination or during an acute illness.

The term viral suppression is used in clinical practice to signify a level of plasma viremia below
the level of detection of the commercially available test being used. As a result, the definition of
plasma viral suppression has changed over time as technological advances have decreased the
lower limit of detection from under 10,000 copies/mL to less than 20 copies/mL. Tests are now
available in research settings that detect plasma viral levels as low as 2 copies/mL. Currently,
most clinicians use either standard viral load tests, with thresholds of 400-500 copies/mL, or
ultra-sensitive methodology with levels of detection as low as 20-50 copies/mL. However, even
in the setting of viral suppression in plasma to these very low levels, replication in other
compartments, including the genital tract and lymph nodes, may still occur.

                                           Viral Load Report

What is the significance of viral load in the absence of ART?
In addition to CD4 count levels, viral load predicts risk of disease progression.4 Based on these
calculated risks of progression, guidelines in the United States have recommended the use of
viral load levels as one of the markers to guide initiation of ART. However, the viral load
thresholds used to guide initiation of therapy have increased over time and thus become less
important in this role. The most current guidelines from both the NYSDOH and the United State
Public Health Service (USPHS) recommend a level of >100,000 copies/mL as the threshold at
which to initiate ART regardless of CD4 counts.10, 11 Some programs have also used this
threshold to measure access and utilization of ART for eligible patients by identifying patients
with viral load levels above it, but who are not receiving ART.

How is viral load used in the setting of ART?
Along with CD4 counts and clinical response, HIV viral load is the most common indicator used
in the United States to monitor the success or failure of ART. Viral load measurements are used
to determine the risk of disease progression, to decide when to initiate ART, to monitor response
to treatment, and to detect viral breakthrough as a marker of regimen failure. As a result, viral
load response is also used as a surrogate marker for efficacy in ART drug trials and in clinical

Although the ideal outcome of treatment is suppression of viremia below the level of detection,
virologic response to ART varies widely, as do the definitions and patterns of virologic response
and success used in guidelines and studies. The range of possible virologic responses described
in the literature and in practice includes the following categories:2, 10, 12-14
                •failure to ever see a virologic response
                •any decline without achieving suppression
                •decline in viral load greater than a defined amount, such as 0.5 or 1 log10
                •decline followed by rebound
                •ever achieving suppression
                •durable suppression over time

                                         Viral Load Report

               •suppression with rare, very low viral “blips”
               •intermittent suppression with higher level rises in viral load
               •loss of suppression after it has been achieved, with subsequent
                measurements either showing stable level of viremia or rising viral load levels.
The USPHS guidelines define virologic failure as repeated viral load levels greater than 400
copies/mL after 24 weeks or greater than 50 copies/mL after 48 weeks in a treatment-naïve
patient initiating therapy. Table 1 describes the definitions of other virologic responses used in
the USPHS guidelines.11
Table 1. Definitions of Patterns of Viral Responses to ART11

Term                          USPHS Definition
Viral suppression             repeated undetectable viral load levels to the lowest level of detection
                              (<400, 50, or 20 copies/mL depending on sensitivity of viral load test
Virologic failure             repeated HIV RNA >400 copies/mL after 24 weeks of initiating ART

                              >50 copies/mL after 48 weeks of initiating ART in a treatment-naïve
                              repeated HIV RNA levels >400 copies/mL after prior suppression to
                              <400 copies/mL
Viral blips                   a single viral load test of 50-1000 copies/mL when otherwise
Viral                         repeated HIV RNA levels >400 copies/mL after prior suppression to
breakthrough/Virologic        <400 copies/mL;
rebound                       increase of HIV RNA levels following decline of defined magnitude

Partial viral suppression     a decrease of >0.5 log10 in HIV RNA levels from baseline

Can everyone on ART achieve viral suppression?
Many factors determine whether plasma viral load can be suppressed below the level of detection
through ART. These include virologic, patient, pharmacokinetic, provider, and system factors.
The most common reasons for failure to achieve viral suppression are viral resistance and poor
adherence. Other factors, including poor absorption, altered bioavailability, drug-drug
interactions and co-morbid illness, can interfere with suppression of viremia. Use of a regimen
with inadequate potency, errors in dosing, inadequate support to maintain adherence and

                                         Viral Load Report

interrupted access to providers or medications may also be contributing factors. It is important to
note that many of these factors may not be within the control of the clinic or provider and thus
not amenable to change through efforts aimed at improving success rates and quality of care
within the clinical setting.

The highest rates of viral suppression have been described in clinical trials that involve first line
therapy among patients with little or no ART exposure, with an increase from 60% to as high as
90% in recent years as ART regimens have become more potent and tolerable. However, when
considering potentially achievable rates of viral suppression in a general population, it is
important to recognize that the higher suppression rates achievable in these clinical trials are not
likely to be replicated in the more diverse populations of clinical practices or in cohorts which
have been enrolled in natural history studies. One of the main reasons is that these clinical trials
include only ART-naïve patients or those with limited ART exposure (Table 4) who are more
likely to succeed than the mix of patients from a clinical practice that includes those with more
ART exposure and drug resistance. In practice, goals would need to include higher viral load
thresholds, such as those used in ART trials involving more treatment-experienced patients. In
addition, many of the patients who enrolled into these clinical trials were highly motivated, at
lower risk for nonadherence and, therefore, more likely to achieve higher rates of viral
suppression than similarly treatment-experienced patients in general practice. For example,
while suppression rates above 70% have been described in clinical practice among ART-naïve
individuals without primary resistance , they are still lower than those achieved in the clinical
trial setting.15-17

The time period during which a patient started ART is also linked with probability of treatment
success. Earlier in the HIV treatment era, lower rates of viral suppression in clinical practice
occurred. In part, this failure was a result of high rates of drug resistance in patients who had
started treatment with sequential monotherapy or dual therapy in the early 1990’s with less
potent and tolerable regimens. Lohse and Tuboi observed a 30% failure rate among patients at
six months with the year of initiation found to be one of the main predictors for virologic
failure.18, 19 Moore also found significant improvements in virologic as well as immunologic

                                        Viral Load Report

and clinical outcomes in patients started on HAART following development of newer drugs and
improved knowledge of treatment efficacy (Table 2).17

Table 2. Improvement in Virologic, Immunologic, and Clinical Outcomes in Clinical Practice
from 1996 to 2002: First HAART Regimen17

Year started            1996            1997-1998           1999-2000         2001-2002

N                        238               591                365                132
Outcomes (%)
VL Undetectable         45%                54%                68%                73%
at 6 mos
VL Undetectable         43%                47%                60%                68%
at 12 mos
CD4 count                100               120                148                139
increase, 12 mos
AIDS-defining           21%                25%                 9%                10%
illness (1st year)
Death (1st year)        18%                19%                 2%                1%

Viral response and suppression in treatment-naïve patients including those with primary
In general, higher rates of viral suppression are seen in ART-naïve patients than in those with
prior ART-exposure, due largely to the absence of resistance. The exception is individuals who
are infected with virus characterized by primary drug resistance mutations; they may have
suppression rates similar to those who are experienced. The frequency and pattern of primary
drug resistance varies depending on location and population, but is of ongoing concern for
clinicians, patients and the public health community.11

                                        Viral Load Report

Viral response and suppression in treatment-experienced patients
Suppression rates for treatment-experienced patients in clinical practice and clinical trials have
varied widely from 12% to over 80% depending on the definition of suppression and the study
cohort (Table 3). One of the most common reasons for absence of viral suppression in
treatment-experienced patients is the presence of acquired HIV drug resistance as a consequence
of the failure of earlier regimens. The frequency and degree of resistance will vary depending on
the regimens used, the duration of failure, and patterns of non-adherence or other factors
associated with inadequate drug exposure. The impact of resistance mutations on the ability to
achieve suppression will also vary depending on the type and number of mutations present, with
rates ranging from as low as 24% to close to 80% depending on definition, cohort, and

In patients known to be infected with highly drug-resistant virus, response rates are even lower
than in the treatment-experienced population overall. Response rates have varied depending on
the salvage regimen used, timing of viral load measurement, and the degree of resistance when
the new regimen was initiated, although improvements in suppression rates are possible as new
ARTs are being developed.23-25

                                               Viral Load Report

Table 3. Viral Response Rates in ART-Experienced Populations

Type             N     Definition of             Response Rate         Time Period    Comments
                       Response                                        of Study
Prospective,    2674   < 400 copies/mL at        70.3% of those        Sept. 1995 –   HAART study
observational          12 months of              treated with 1 new    Nov. 1998
cohort15               treatment                 drug
                                                 78.7% of those
                                                 treated with 3 new
Multi-center,   679    < 50 copies/mL at         Study 1: 38% (14.7%   July 1995 to   HAART
observational          month 6 of therapy        < 50 copies/mL,       Dec. 1999      initiation study,
cohort26                                         22.8% <400                           majority of
                                                 copies/mL)                           populations is
                                                 Study 2: 32% (12.5%                  ART-
                                                 <50 copies/mL,                       experienced (68-
                                                 19.7% <400                           70%)
Prospective,    1022    1 log10 decline in      37% at three months   Jan. 1996 –
observational          viral load level with     51% at six months     Apr. 1999
cohort26               stable or increased
                       CD4 count
Prospective     3736   <500 copies/mL            29.8% on 1-2 prior    Jan. 1997 –    Study of
cohort27               throughout two year       NRTIs                 Dec. 1998      durability of VL
                       study                     23.7% on 3+ NRTIs                    response
                                                 37.7% ( 1.5 log10

Viral response in the absence of suppression
Because not all patients are able to achieve viral suppression, studies and guidelines have also
used the magnitude of change in viral load or drop to levels below a specified threshold as a goal
of treatment. The majority have used change in viral load compared with pre-intervention levels,
with the most common goals ranging from a decline of 0.5 to 1 log10,2, 12-14 (a selection are
summarized in Table 4). Various thresholds have been proposed for viral response in the
absence of undetectable suppression including viral load values less than 30,000 copies/mL; 28
less than 10,000 copies/mL;29 less than 5000 copies/mL;27 and less than 1000 copies/mL.2 In
general, these definitions of virologic response have been used in populations that are ART-
experienced, and particularly in populations with significant drug resistance. The NYS criteria
define response as a drop by at least one log10 since the last 4-month review period, or an
increase of less than three fold from the lowest viral load value in the last 12 months on that

                                          Viral Load Report

Table 4. Goals in Published Studies for Viral Load Responses in Different Clinical Populations2,
10, 12-14

      Population                 Magnitude of decrease                  Time span
ART Naïve                             0.5-1.0 log10                    24, 48 weeks
ART Experienced                         0.5 log10                      24, 48 weeks
Patients with Known                     0.5 log10                        12 weeks
ART Resistance

4.5 Type of ART prescribed
The relative efficacy of different regimens is a focus of ongoing study. In a meta-analysis of 64
triple therapy studies involving over 10,000 patients and follow-up for 24 weeks, Bartlett found
differences among treatment regimens based on type of regimen.31 Contrary to his earlier
analysis in 2001,32 pill burden was no longer associated with outcome. Other studies have also
found superior response rates among NNRTI-based and boosted-PI based regimens compared
with NRTI or un-boosted PI-based regimens.33

How do we interpret detectable viral load levels in the setting of ART?
Many individuals receiving HAART who achieve viral suppression will subsequently have some
pattern of viral breakthrough. These patterns range from isolated small bursts of virus to frank
breakthrough with a rapidly rising viral load. The interpretation of these patterns is critical to
understanding how they may or may not relate to treatment failure, and whether to classify them
as changes in quality of care.

Viral blips
Viral blips are generally defined as a single viral load test of 50-1000 copies/mL in someone
whose viral load level is otherwise suppressed.11 In a retrospective study, researchers reviewed
viral load data from two HAART treatment trials.34 Transient episodes of viremia or blips were
common, occurring in 29-40% of 241 patients, depending on regimen. Occurrence of viral blips
did not predict treatment failure (9.3% vs. 13.8% failure in patients with and without blips during
15 months of follow-up). Havlir also reported that patients with transient, very small increases
in viral load levels (median viral load levels of 30 copies/mL) were at no greater risk of rebound

                                        Viral Load Report

than those with viral load levels of ~ 3 copies/mL.34 A more recent study of viral blips and drug
resistance found that blips, defined as a single measurement ≥ 50 copies/mL followed by a
measurement < 50 copies/mL, also did not predict the occurrence of failure or resistance. The
authors concluded that blips were the result of random biological and statistical variation.35

Viral breakthrough or rebound
Viral breakthrough or rebound is defined as repeated HIV RNA levels > 400 copies/mL after
prior suppression of < 400 copies/mL.11 Causes of breakthrough or rebound can include
development of ART resistance, non-adherence, treatment interruption, concomitant illness, and
less commonly, lab error. The pattern of viral breakthrough can also vary depending on cause
for regimen failure, whether the regimen is maintained, intrinsic viral replicative capacity, and
pretreatment viral load.

Virologic failure of a regimen
Some studies suggest that in the setting of a failing regimen when patients have limited treatment
options, maintenance of viral load levels below a predetermined threshold or below pre-
treatment values still has clinical benefit over a finite time period. For example, Ledergerber
noted that keeping viral load levels below 10,000 copies/mL or maintaining a viral load decrease
from baseline of more than 1.5 log10 appeared to prevent a significant decrease in CD4 count
over time despite the absence of suppression.29 Similar findings are reported in a retrospective
study carried out by Bahrani evaluating 497 HIV-infected patient beginning HAART from early
1996 through March 1999.36 The study found that patients with suppressed viral load levels that
rebounded to levels above 10,000 copies/mL also had a decline in their CD4 counts related to the
level of breakthrough. However, patients with viral rebound where levels consistently remained
below 10,000 copies/mL experienced no significant change in their CD4 cell counts. These data
suggest that during a period of up to 2 years, patients with viral rebound to levels below 10,000
copies/mL may be able to maintain their CD4 counts despite a failing regimen. Deeks and
colleagues also found that CD4 counts are maintained for prolonged periods following virologic
failure, with duration related to the continuation of the regimen, larger differences between pre-
treatment and post failure viral load, and higher absolute level of viremia during failure.37 These

                                          Viral Load Report

findings have led to growing interest in patients who may have discordant virologic and
immunologic responses to ART (see below).

Other causes of viral rebound or breakthrough
The occurrence of opportunistic or other significant infections can cause a transient rise in viral
load or loss of suppression where both the duration and degree of increase vary. Viral load
levels measured in the setting of acute infection should be repeated after resolution of the
intercurrent illness before considering regimen failure.

Several New York City clinics reported problems with viral load testing for about three years
with an increased rate of low level viremia in previously suppressed patients. A number of these
patients were evaluated for drug resistance, however, the vast majority of the resistance assays
were not amplifiable because of insufficient virus in the specimen. This finding supported the
contention that there were problems with the test resulting in spurious elevations of viral load
levels and causing unnecessary concerns for both clinicians and patients. These viral load
elevations were attributed to the switch from EDTA to PPT tubes. Another potential cause of
errors was the lack of timely transport and centrifugation.

Structured treatment interruptions (STIs) were initially evaluated as a strategy to increase
immunologic response to HIV,38 improve response to salvage therapy for multi-drug resistant
virus,23, 24 or decrease exposure to drugs in order to reduce side effects. In patients undergoing
STIs, viral breakthrough is an expected event, and does not constitute evidence of treatment
failure or lapse in quality or effectiveness of care.

Discordant immunologic and virologic responses
The population of people receiving HAART includes a growing group of treatment-experienced
individuals with multidrug resistant virus in whom treatment with currently available ARTs will
not result in long-term viral suppression. Several studies have shown that continuation of a
regimen in the setting of ongoing viremia provides immunologic and clinical benefits.29, 36, 39, 40
These discordant results represent a critical change in the approach to the management of

                                         Viral Load Report

virologic failure and interpretation of clinical management of patients on ART with ongoing

Use of other clinical factors to assess quality of care in the context of viral suppression
A number of studies, in addition to those noted above, examined the potential use of other
measures to assess ART management outcomes in the setting of ongoing viremia, defining
success or goals of treatment. These goals include the following measures which could be used
with, or instead of, suppression and which are addressed in the sections below:
      Change in viral load level
      Maintenance of viral load level below a specified threshold
      CD4 count response or maintenance, or
      A combination of both viral load and CD4 count
      Clinical response or stability

5. Population characteristics and viral suppression rates

Studies of selected populations such as substance users, homeless persons, and those with
HCV/HIV co-infection have demonstrated decreased rates of viral suppression among patients
prescribed ART. Most of these differences are linked to challenges related to achieving and
maintaining adherence. A complete review of patient-related factors associated with rates of
viral load suppression is beyond the scope of this review, however, a few specific areas are
addressed below.

Maintenance of high levels of adherence is critical to improving the probability of achieving
viral response. Most studies have shown that very high levels of adherence (> 95%) are
necessary to achieve viral suppression and that small differences in adherence rates can translate
into dramatically decreased suppression rates.

Table 5 summarizes key studies that address the level of adherence necessary to achieve viral
suppression. It is notable that even in the setting of very high adherence rates, viral suppression
was still not achieved in 100% of patients. Table 5 details findings from a sample of studies on
adherence rates and risk of virologic failure.

                                            Viral Load Report

Table 5. Adherence Rates and Viral Suppression from Selected Studies
Population                 N            Time Period        Adherence Level       Suppression
                                        and Definition     Reported              Rate
                                        of Suppression
HIV-patients               99           6 month study,            > 95%               78%
seen in a VA and                        Pre-HAART                80-95%               39%
a university                            < 400 copies/mL           <80%                20%
medical HIV
HIV-infected              173           6 month study,            100%                47%
patients with a                         < 400 copies/mL
mean baseline
CD4 count of
142 x 106cells/l                                                 95-99%               41%

                                                                  <80%                 7%
ARV-naïve                 886           19 month study,           >95%                84%
patients starting                       < 500 copies/mL          90-95%               64%
HAART43                                                          90-95%               82%

Virologic response among a sample of reported studies of cohorts of patients with substance use
are summarized in Table 6. Rates ranged from as low as 13% early in the HAART era to 52% in
later studies. Active, but not prior, substance use was associated with lower rates of adherence
and subsequently lower rates of viral suppression.

Table 6. Viral Response Rates in Patients with Active Substance Use

Type                N     Definition of          Response Rate                    Time Period    Comments
                          Response                                                of Study
Prospective,        85    < 500 copies/mL        active cocaine users: 13%        July 1998 –    Heavily
observational                                    active heroin users: 29%         Apr 1999       ART-
cohort44                                                                                         experienced
                                                 those who use alcohol several
                                                 times/wk or daily: 21%
Prospective         695    1 log10 decline in   No substance use to use (52%     Nov. 1998 to   Studied
cohort study45            viral load level or    to 37%)                          May 2001       transition
                          < 400 copies/mL                                                        from
                                                 Substance use to no use (35%                    user to non-
                                                 to 52%)                                         user and vice

                                         Viral Load Report

The homeless population has higher rates of co-morbid conditions, such as active substance use
and mental illness, conditions associated with decreased adherence, which can translate into
decreased rates of viral suppression. However, with extensive supportive services and improved
adherence, treatment success can be achieved in this population.46 Other patient factors
associated with decreased adherence and hence higher risk of non-suppression include certain
active psychiatric diagnoses, in particular depression,47, 48 presence of side effects related to
ART, and decreased belief in the ability of the medications to help.

CD4 count response or maintenance
A number of studies noted above have shown that in the absence of viral suppression,
maintenance of immunologic response as defined by increase or maintenance of CD4 count
levels during treatment remains a valid goal, suggesting its role as a measure to assess quality of
care. A number of factors will influence absolute and incremental response and therefore require
additional data collection, including pretreatment CD4 counts. These studies are supported by
other literature, including NYSDOH AIDS Institute guidelines,10 USPHS guidelines,11 Sabin,49
and Ledergerber.29 The predictive value of CD4 count is particularly relevant in the context of
significant viral resistance which limits the ability to achieve virologic response and further
effective treatment options.50

Combining CD4 response and viral load
The addition of measuring CD4 counts and change in viral load to measurement of viral load
suppression to assess quality of care is also complex, but may add to the ability to develop and
measure treatment goals for ART-experienced populations, particularly those with significant
ART resistance.

In a multi-cohort study of 2,488 ART-experienced patients with three-class virologic failure (VL
>1000 for > 4 months), Ledergerber found that CD4 counts < 50 cells/mm3 compared with >
200 cells/ mm3 confer a significant increased risk of death, while the most recent VL failed to
predict death.29 The authors conclude that CD4 > 200 cells/mm3 should be the goal of treatment
when viral load suppression seems unattainable. The authors also state that maintaining viral
load levels below 10,000 copies/mL, or showing a viral load decrease of greater than 1.5 log10

                                         Viral Load Report

from baseline, appeared to prevent a meaningful decrease in CD4 count. Nasta found that
keeping viral load below 30,000 copies/mL in treatment-experienced patients results in only low
level change in CD4 count.28 Mijch reported that in a multivariate model, the latest CD4 count,
number of different antiretroviral drugs per year and more than one hospitalization, but not viral
load, were found to be predictive of death.51 Table 7 summarizes some proposed definitions of
CD4 count (immunologic) response and of virologic response in the absence of virologic

Table 7. Summary of Potential Definitions of Immunologic Response in the Absence of Viral
Definition of Immunologic          Definition of Virologic
Response                           Response or Proposed Goal in
(cells/mm3)                        Absence of Suppression

Immunologic failure: Return of     N/A
CD4 count to pre-regimen level
after at least 16 weeks of
continuous treatment37
N/A 37                             VL >500
Unstable response:                 Stable: Viral load is undetectable;
CD4 has dropped by 50% since the   or has dropped by at least one log10
last 4-month review period; CD4    since last 4-month review
> 200 10
                                   Unstable: load has increased by
                                   more than 1 log10 and absolute
                                   value is over 1,000
CD4 > 200                          VL 100,000
N/A 28                             VL 30,000
CD4 > 250, 350 49                  VL 10,000
CD4 > 200 29                       VL 1,000; 10,000
N/A 40                             > 1 log10 decrease
N/A 2                              > 1 log10 decrease; >1000

                                        Viral Load Report

Use of viral load levels as a quality measure at the clinic or population level
Population surveys of viral load values have not been routinely published. Similarly, quality of
care performance measurement summary data that includes viral load values have not been
widely available. Three sets of data were reviewed that include samples of viral load data from
large populations and thus provide a glimpse of how viral load values might be distributed both
within a population of people with HIV engaged in ambulatory care and among subsets of this

Population level evaluation of viral load levels in New York State
During the year 2002, 3618 patient records were reviewed in New York State from 60
ambulatory care clinics. These records were selected as part of a random sample drawn from
clinic submitted case lists. Table 8 shows the distribution of viral load levels by sample
subgroups and the proportions of patients with CD4 greater than 200 cells/mm3 and patients
receiving antiretroviral therapy. These data show that even in the group that never had viral load
values lower than 55,000 copies/mL, 58% had a CD4 count above 200 cells/mm3.

                                                      Viral Load Report

Table 8: Patients with Viral Load below and above Specified Thresholds from the New York
State (NYS) HIV Quality of Care Program
                                   Number (%)            Cumulative Number      Number on ART
                                                         Percent    with        in Any Period
Lowest Viral Load Value During 2002 (copies/mL)*
VL undetectable (<50)              1151 (32%)                  32%         78      1086 (94%)
50<VL<1000                         1271 (35%)                  67%        148      1056 (83%)
1000<VL<10,000                      601 (17%)                  84%         99       356 (59%)
10,000<VL<55,000                    338 (9%)                   93%         97       168 (48%)
VL>55,000                           224 (6%)                   99%        131        95 (42%)
No VL value                          27 (1%)                  100%          9         1 (3.7%)

Highest Viral Load Value During 2002 (copies/mL)*
VL undetectable (<50)              280      (8%)         8%           244       260 (93%)
50<VL<1000                         1018 (28%)            36%          875       908 (89%)
1000<VL<10,000                     767     (21%)         57%          635       587 (76%)
10,000<VL<55,000                   617     (17%)         74%          416       402 (65%)
VL>55,000                          909     (25%)         99%          330       604 (66%)
No VL value                        27     (1%)           100%         7         1 (3.7%)

*IPRO, NYS HIV Quality of Care Program (unpublished data).

In the NYS HIV Quality of Care Program, patients who are prescribed antiretroviral therapy are
categorized as “stable” or “unstable” based on the specific definitions that were established by
the state’s clinical advisory panel, the HIV Quality of Care Advisory Committee. These
definitions are based upon changes in viral load levels. Patients classified as “stable” have either
achieved viral load suppression or experienced a decrease of a specified magnitude, whereas
unstable patients fail to meet these criteria (see Appendix B). In 2002, sufficient information
was available to classify 2183 patients as “stable” or “unstable” during all 3 periods of review.
Table 9 reports how many of those remained stable for all 3 four-month periods during the

                                                         Viral Load Report

calendar year, remained stable for 2 of the 3 periods, remained stable for just one period and
those who were not stable during any period.

Table 9. Distribution of ART status and Clinical Stability Classifications as Defined by the New
York State HIV Quality of Care Program*
Total patients                                                                2183
Patients who remain stable for all 3 Periods                                  999               46%
Patients who remain stable for 2 Periods                                      699               32%
Patients who remain stable for 1 Period                                       410               19%
Patients who are not stable during any period                                   75                3%
*IPRO, NYS Department of Health AIDS Institute: Unpublished data. Note that data are limited to reviewable periods in which sufficient
information is available in the medical record.

In addition, of 337 patients who were unstable in one 3-month period, 213 (63%) became stable
in the subsequent period. In Table 10, results from the review of 2520 patient records from
patients who were reviewed during 2002 in all 3 periods are displayed. Finally, for patients who
were considered stable during the entire year 2002, the distribution of the highest and lowest
viral load values are shown in Tables 10 and 11.

Table 10: Distribution of Viral Load Levels among Patients Defined as Clinically Stable by the
New York State HIV Quality of Care Program*
Highest VL (2002)                           Number of              Cumulative Data
(copies/mL)                                   Patients
VL undetectable (<50)                             159 (16%)             159 (16%)
50<VL<1000                                        559 (56%)             718 (72%)
1000<VL<10,000                                    197 (20%)             915 (92%)
10,000<VL<55,000                                   47 (5%)              962 (96%)
VL>55,000                                          37 (4%)              999 (100%)
No VL value                                         0                   999 (100%)
*NYSDOH AIDS Institute, IPRO, unpublished data.

                                                  Viral Load Report

Table 11: Lowest Viral Load Levels in 2002 among 999 Patients Reviewed as Part of the New
York State HIV Quality of Care Program*
Lowest VL (2002)                      Number of          Cumulative
(copies/mL)                            Patients          Number (%)

VL undetectable (<50)                  571 (57%)           571 (57%)
50<VL<1000                             339 (34%)           910 (91%)
1000<VL<10,000                          60 (6%)            970 (97%)
10,000<VL<55,000                        21 (2%)            991 (99%)
VL>55,000                               8 (1%)             999 (100%)
No VL value                             0                  999 (100%)
*NYSDOH AIDS Institute, IPRO, unpublished data.

Medical record reviews also examined the documented reasons for why patients are not
receiving antiretroviral therapy. These included nonadherence, refusal, working on readiness,
pending resistance results, lack of therapeutic options, missed appointments, active substance
use, elevated serum liver enzymes, intolerable side effects, intercurrent illness and limited access
to medications. Other reasons that were not documented but were apparent to the record
abstractors included provider failure to prescribe and overwhelming social factors, such as lack
of housing and family issues. The frequency of these reasons was not quantified. While some of
the reasons were attributed to quality of care, more were related to clinical judgment, patient
readiness and attendance in clinic.

As part of a large national quality improvement collaborative aimed at accelerating the pace of
improvement of HIV care through methods of rapid organizational change, viral load data were
collected as part of a set of indicators designed to assess the impact of the collaborative. Over 80
facilities participated for 12 months during 2000-2001, and 45 participated for 18 months.
10,346 patients were reported in registries from clinics where approximately 70,000 patients
receive care. At each site, monitoring of key measures was conducted and included the number
of patients with viral load below 50 copies/mL and the number with CD4 counts below 200
cells/mm3. During the collaborative, numerous changes were made in the systems of care in
these clinics, including implementation of adherence services, reminder systems and flow sheets,

                                          Viral Load Report

in addition to self-management programs among other strategies. Data were collected monthly
from patients attending clinic during that interval of time and reported continuously over two

Results showed that the proportion of patients with a viral load < 50 copies/mL increased from
38% to 45% during the span of the collaborative. The proportion of patients with CD4 < 200
cells/mm3 increased slightly from 23% to 25%. The patient population included both treatment-
naïve and experienced patients. During this period when resistance testing was not widely
available, known reasons for failure to meet expected higher rates of viral load suppression were
primarily related to patient factors and include clinical status, willingness to start therapy, ability
to adhere to medications, clinic attendance, drug use, mental health conditions and unstable
housing. Conclusions from analysis of these data suggest that clinicians in the collaborative
appropriately relied upon clinical status, CD4 count and viral load to manage patients.
Achievable rates of viral suppression were limited by effectiveness of current care strategies and
tools during that time period, including availability of medications, unavailability of resistance
testing and the state of knowledge about the management of comorbid conditions and
nonadherence to HIV treatment.

Finally, data were shared by the New York City Department of Health and Mental Hygiene
(NYCDOHMH) which analyzed reported viral load levels and reported patient zipcodes (whether
the reported zipcode was of patient residence, provider or lab is unknown).53 These data show
that zipcode areas of patients with higher viral load levels correlate with higher rates of HIV-
related mortality. However, the complexities of interpretation of these data as markers of quality
of care are significant given the absence of other patient level data including ART status and
clinical condition. For example, potential confounding factors include that zipcode regions with
higher mortality and higher viral loads also included a number of AIDS nursing homes in which
patients receive palliative care, and for whom ART is not always clinically indicated. Higher
viral load levels and higher mortality rates would be expected from these zipcodes, but would not
suggest that quality of care issues related to access or provision of care had occurred.

                                         Viral Load Report

Published literature that evaluates the use of viral loads and viral suppression as a quality
measure across clinics or populations remains sparse. The performance of a viral load as a
process measure of quality of care has been used (NYSDOH), however, only limited data have
been published evaluating suppression or response rates in the context of quality. In one study,
viral suppression among patients prescribed HAART or eligible for HAART was used as an
outcome.6 This prospective study included 9,986 patients and compared 44 intervention clinics
to 25 control clinics where the studied intervention was a quality improvement collaborative
from 1999 to 2001. The results showed that in the intervention group, the viral suppression rate
increased from 40.1% to 51.1% with an increase from 43.6% to 48.8% in the control group (not
statistically significant; p = 0.18).

In a US metropolitan community of 150,000, Parry assessed quality of care across the population
using decline in viral load levels, stable or increased CD4 counts and other clinical and
utilization markers including hospitalizations, opportunistic infections, and weight.7 Over the
period of 3 years, 95% of patients in care achieved stable or increased CD4 counts, 90% had
decreased viral loads and 93% had stable or increased weight. No comparison data were
available from other populations or from patients not in active care.

Summary of the Review
Based on this comprehensive review of the literature, it is evident that both virologic and patient
factors, independent of the quality of care provided, may change the probability that an
individual will be able to achieve viral suppression. If viral suppression rates are used to
measure quality of care, then consideration of how to gather data about and adjust for these other
specific factors will be critical. These may include baseline CD4 count and viral load, change in
viral load and timing of that change, planned treatment interruptions, viral load levels following
blips, treatment history, drug resistance, and patient characteristics associated with decreased
adherence and tolerability of the ART regimen.

                                         Viral Load Report

III. Report from the Expert Meeting
In the morning session, panel members were asked to discuss the use and interpretation of viral
load measurements in the context of their personal clinical practice and whether viral load data
were used to assess quality of care in their clinics. In the afternoon session, the panel was
divided into three groups each focusing on different elements of the potential use of viral load as
a quality measure at the programmatic and population level. In addition, the expert panel was
asked to consider the potential role of viral load measurement as part of the measurement
strategy to assess the quality of care at these levels and identify other factors, if any, that would
need to be included.

Morning Session
The first part of the discussion included the entire group and focused on how the participants
assess the effectiveness of antiretroviral therapy (ART) in their clinical practices. Specifically,
the group considered the use of viral load levels and other parameters and measures, how
treatment goals, including viral load levels, are established for different populations, and the
definition of viral suppression. In addition, participants were asked how they manage patients
who have been prescribed ART but fail to achieve viral suppression, and, finally, how these
cases were interpreted in the context of quality of care. Both areas of consensus and contention
emerged regarding the current use of viral load in clinical practice as a clinical tool and its
current use to measure quality at the individual patient and clinic levels.

Viral load measurement and interpretation in clinical practice.
Viral load measurement was widely acknowledged by the group as an extremely valuable tool in
the management of HIV. Although still used in making decisions about initiating ART, its
importance for this purpose has decreased given changes in both guidelines and practice. The
primary use of viral load measurement in current clinical practice is to monitor the effectiveness
of ART in a given patient.

All of the panel members use the same definition of viral load suppression in clinical practice,
specifically defined as the lowest level that can be achieved with commercially available ultra-

                                          Viral Load Report

sensitive technology (now usually fewer than 50 copies/mL). The panel agreed that viral load
suppression is the accepted goal for all patients who are treatment-naïve and in whom primary
resistance is not present before starting HAART. However, even among treatment-naïve
individuals, viral load suppression is not always realistic, primarily because of poor adherence,
although, tolerability and primary resistance were also noted as factors. For this reason, viral
load measurement is generally used in conjunction with other measures of response and
consideration of patient-level factors, including adherence, to determine if a regimen is
appropriate and if it is successful or failing.

In treatment-experienced individuals unable to achieve viral suppression, additional factors were
noted to be at least as important as viral load level for assessing the effectiveness of ART and
guiding appropriate management of these patients. Some of the measures commonly used in
participants’ clinical practices include CD4 count, maintenance of CD4 count above a specified
threshold, viral load decline or maintenance below a specified threshold, clinical response,
presence of drug resistance and adherence rates. Similarly, detectable viral load levels (loss of
suppression) were not always used in practice as the sole indicator for when to switch treatment
regimens, particularly in patients on their second or higher regimen.

Some providers reported use of viral load level as a potential marker for lapses in adherence
among patients who had achieved suppression in the past, but all acknowledged that other factors
such as resistance or drug interactions could cause virologic breakthrough. As in other
situations, viral load was interpreted in the context of other measures and the clinical context.

Viral load as a measure of quality of care in clinical practices
The current use of viral load as a quality measure was discussed in two broad categories: within
the context of individual patient care and at the clinic population level. Because of the highly
individualized treatment history and tolerance for each patient, participants asserted that viral
load measurement alone could not adequately reflect quality and appropriateness of care. Some
clinicians use suppression rates to identify patients who may require further review or
intervention, but with recognition that many of these patients might not be able to achieve
suppression for the various reasons described earlier. In addition, even with patients for whom

                                         Viral Load Report

achieving suppression was biologically feasible, other factors beyond the clinician’s control,
including active substance use, refusal to accept treatment, or severe liver disease precluding
prescription, may still make suppression unachievable or even in some cases not the appropriate
clinical goal. Any current use of viral load in an overall effort to assess quality of care in clinical
practice was therefore deemed appropriate only if it can be interpreted in the context of other
important clinical measures such as CD4 count, viral resistance and clinical outcomes.

Viral suppression as a measure of the quality of individual patient care for treatment-experienced
In participants’ clinical practices, for patients on second- to fourth-line regimens, the desired and
achievable viral load outcome for patients is not always suppression, but rather maintenance
below a threshold or drop in viral load from pre-regimen levels. For example, one common
measure of virologic response for this group is the change in viral load (0.5-1.0 log10), which
while feasible for an individual patient, was reported to pose significant challenges for existing
information systems. Assessment of quality across an entire clinic population would require the
ability to compare viral load levels over time for each patient in care. In addition, the achievable
response rates, given any definition of virologic response, will change as new therapeutic
strategies become available, posing further challenges for longitudinal measurement. In fact, the
appropriate outcome measure for highly treatment-experienced patients with limited options did
not always include viral load. Instead, most clinicians focused on CD4 count as a more
meaningful measure for heavily experienced patients because it is has been consistently shown to
correlate and correspond better with clinical outcomes. Other reasons for viral load elevation
that were not related to quality of care were also noted including structured treatment
interruptions, intercurrent illnesses, or technical problems with the actual test.

Use of viral load for patients who are not receiving antiretroviral therapy
The use of viral load as a sole measure of overall quality in clinical practice was considered
equally challenging for patients not on treatment, even if only patients eligible for ART were
considered. Several clinical scenarios were described in which the appropriate decision would
be to not initiate therapy despite appropriate virologic and immunologic criteria. These included
situations where patients were not ready to initiate treatment or where significant adherence

                                         Viral Load Report

barriers would need to be addressed prior to ART initiation. Panel members expressed concerns
that use of viral suppression as a measure of quality in this population, which includes patients
who were not prescribed ART for appropriate reasons even though they met CD4 count or viral
load eligibility criteria, would underestimate the quality of care and might put pressure on
clinicians to start ART before the patient is ready in order to raise scores.

Afternoon session

The afternoon discussions focused more specifically on the role of viral load measurement and
viral suppression as measures of quality of care. Participants were asked to define appropriate
measures for assessment of clinic-based or population outcomes. Participants were divided into
three groups, each of which addressed one of the following three questions:

1. How should viral suppression be used as a measure of quality in the general population
   of people with HIV? Can target rates for viral suppression be defined for general and
   specific populations?

2. How can we use viral load as a measure of quality in the absence of suppression? What
   should the performance measure be in the absence of suppression? What other
   markers would you need to measure quality of care in the absence of suppression?

3. What should the performance measure for viral load be in settings of different
        Naïve vs. experienced?
        Patients with vs. without documented resistance?
        Adherence challenged populations?
        Others (HCV, for example)?

The discussions focused on the well-documented complexity of management associated with the
multiple interacting factors in clinical practice that influence the probability of an individual’s
achieving viral suppression or other specific virologic response. Below are summarized the
common themes that emerged from the three groups, followed by the issues unique to each

                                        Viral Load Report

Defining quality measures using viral load levels
Potential definitions of quality measures based on viral load measurement for general and
specific sub-populations were discussed. Panel members agreed that a single measure of viral
load response such as suppression would not be meaningful if applied across all populations.
Specific sub-populations in which the measure would necessarily be applied differently include
ART-naïve and experienced patients, those with and without documented resistance, adherence
challenged populations and other groups in whom suppression may not be achievable.

The groups also noted that any viral load measure would require further definition. For example,
even the specifications for a single measure of viral load when viral suppression is feasible as a
quality measure are complex (Table 12). Longitudinal measurement would be more complex
and involve categorizations of “always suppressed”, “ever suppressed”, and “suppression while
on ART”. Cross-sectional analysis of a population would only signify suppression at a specific
point in time. Determination of eligibility for evaluation would require elaboration, such as
clarification about the minimum number of viral load test results that would need to be
documented in the medical record. The definition of “being on ART” would also require more
clarification. This term could mean “only during viral load measurements” or “only if on ART
for a specified amount of time before the viral load is measured”. Other factors which impact
upon suppression rates that would need further definition include the degree of drug resistance,
and selected patient characteristics such as active substance use, and active mental illness.

                                        Viral Load Report

Table 12. Examples of Measures of Quality and Factors Which Will Require Defining
Term             Areas Requiring Further Definition
VL suppression   Does this refer to the patient’s viral load always being suppressed
                 during the time period or ever suppressed?
                 Do we look at a single time point for all patients, or over a longer
                 specified period of time?
                 How should blips or viremia due to planned interruptions be addressed?
VL response      Does this include decline in viral load or maintenance below a specified
                 level or only complete suppression?
                 If decline or maintenance below a threshold other than undetectable is
                 selected, what threshold and criteria are to be used?
ART naïve        Does this include patients who may have primary resistance?
                 What about women who have had short course treatment for PMTCT?
ART              Does this include any ART exposure or only if treatment failure has
experienced      occurred?
                 How should experience and resistance be quantified?
On treatment     Does this refer to a minimum amount of time on treatment or only at the
                 time of viral load measurement?
                 How should structured treatment interruptions be considered?

What are potential quality measures for viral load?
a.     Suppression

As in clinical practice, the general consensus of the panel was that viral suppression to < 50
copies/mL was the ideal goal for a population of ART-naïve individuals without primary
resistance, and was a valid measure of ART effectiveness in this population.

Because all clinics will include a considerable number of patients without viral suppression,
categories of viral load levels were suggested as potential quality measures to identify this group.
Although the proportion of patients in a given clinic with detectable viremia is useful as a
potential marker for problems such as resistance, co-morbidities (mental health, active substance
use), inappropriate management or poor bioavailability, it may not signify gaps in quality.
Therefore, interpretation of rates of detectable viremia would require careful consideration.
Another potential role for viral load suppression as a quality measure at the clinic level may be to
monitor management of the treatment for patients with unsuppressed viral load to determine if
appropriate actions were taken, such as starting or switching ART regimens, resistance testing, or
adherence interventions. However, measurement systems would then necessarily include CD4

                                         Viral Load Report

counts and ART status, at a minimum, and also likely include resistance testing and adherence
service interventions.

b.      Viral load change
The panel agreed that a viral load decline of 0.5 – 1.0 log10 in the setting of ART initiation over
24 to 48 weeks may be considered a positive outcome when viral load is unsuppressed (defined
as > 50 (or 400) copies/mL) depending on the assay available to the clinic. However, this
measure would require the capacity to follow viral load levels longitudinally at the patient level
throughout the clinic or general population as well as linking them temporally to date of ART

c.      Viral load below a specified threshold
Using the detection of patients with a viral load level above a specified maximum threshold as a
measure of quality may identify patients with very high viral load levels who are potentially
most in need of services to access or improve outcomes with ART. This measure could facilitate
service planning and resource allocation for patients with potentially higher risk of mortality and
possible decreased use of HAART. It would require careful consideration of criteria to establish
the threshold with evidence-based or consensus-based guidelines.

d.      Viral load’s limitations as a quality measure

Should viral load be used as a measure for quality?

In order to define the goals, benefits and limitations of using viral load as a quality measure, the
group agreed that the purpose of any quality measure should be to assist in improving care by
helping programs identify potential gaps in service or quality that can be addressed by the
providers, programs or larger community. In this context, the utility of viral load as a quality
measure has not yet been established. Significant reservations about the use of viral load levels
as a marker of quality remain. Specific concerns were raised about the use of unadjusted rates of
viral suppression or other viral load measures for comparison between institutions. The need for
an appropriate case-mix adjustment methodology was expressed. The major factors for
consideration would again include treatment experience as well as patient factors that pose

                                         Viral Load Report

additional challenges to viral suppression, such as homelessness, active substance use and rates
of clinic attendance.

Measuring rates of viral suppression to < 50 copies/mL was not considered useful for assessing
quality of care for the HIV population. Defining the index population would be difficult,
requiring that factors related to ART eligibility be captured, as well as appropriateness of ART
initiation, even in the setting of clinical eligibility. Within a given clinic population, groups of
patients will include those on ART, those eligible and those not fully engaged in care who may
not have viral load tests performed. Should these groups be treated similarly and measured by
the same yardstick? Would the number of clinic visits make any difference? Should all groups
be considered the same for purposes of quality management? Because of these challenges, the
group decided that the use of viral suppression alone as a quality measure should only be
considered after further research on its utility in this setting had been completed.

Further, the level of suppression for specific populations is also difficult to determine. Expected
suppression rates tend to be lower for patients on second- and fourth-line regimens. Complete
suppression for these and more treatment-experienced patients is not necessarily feasible and,
therefore, undetectable viremia is not a useful marker of the quality of care. In addition,
proposing a target rate for this population would also run the risk of unfairly penalizing providers
caring for more challenging patients. Establishing a performance standard for minimum viral
suppression rates within a general population either in care or for those receiving ART would
involve the same limitations encountered when establishing a desired viral load level for a given

How analyses would be performed
Interpretation and the ability to detect potential gaps in quality of care will vary depending on
how the analyses are performed, such as whether all patients eligible for ART (intent-to-treat), or
only those on ART (as treated) are included.

If too much focus is placed on addressing high or unsuppressed viral loads, resources might be
inappropriately re-allocated from programs and populations where they are effectively used to

                                         Viral Load Report

support existing services. This potential resource drain would likely have a negative impact on
public health. The panel, therefore, recommended that before viral load is adopted by a public
health agency to assess quality, an evaluation of its utility and impact upon quality, as well as an
assessment of the human and material resources needed to determine if this information would
be effective in assessing or improving quality should take place.

e.     Viral load in the context of other measures

Can viral load response or suppression be used alone as a measure of quality?
Following consensus that any measure of viral load is not likely to stand alone as a quality
measure in most, if not all, clinical settings, the panel recommended that a matrix that also
includes CD4 counts, ART response over time, and clinical factors would be both a more useful
approach, and one that is more reflective of current management.

Group 1

       Question: How should viral suppression (or viral load measurement) be used as a
       measure of quality in the general population of people with HIV? Can target rates for
       viral suppression be defined for general and specific populations?

The use of viral load measurement or suppression as a measure of quality in the general
population is addressed above as a common theme discussed by all three groups (see pages 32-

Viral suppression as a quality marker in specific populations
Establishing goals for rates of suppression among specific populations, which would be used as
indicators for measuring the quality of care, was acknowledged to pose numerous challenges.
While the group agreed that suppression would be a valid indicator for treatment-naïve patients,
it expressed concern that not all of these patients would be able to attain viral suppression,
necessitating the definition of realistic performance goals and cautious interpretation of results
even in this population with relatively high potential rates of viral suppression.

                                         Viral Load Report

Discussion about achievable rates of suppression for patients on second-line regimens led to
agreement that expected suppression rates were lower for this group than for ART-naïve patients,
rendering its utility as a quality measure questionable for this population. Because complete
suppression cannot be expected for all patients in this treatment-experienced group, it cannot be
considered a fair marker of quality to assess whether appropriate care has been provided. The
group decided that proposing a target rate for viral suppression in experienced populations would
also run the risk of unfairly penalizing providers caring for more challenging patients who would
have lower suppression rates despite the provision of potentially equal or better quality of care.
In fact, lower suppression rates across a population or community might instead flag clinics or
patient populations requiring increased services, rather than reflecting poorer quality of care.

Based on this discussion, the group concluded that establishing a performance standard for
minimum viral suppression rates which could be applied across the general HIV population,
either in care or receiving ART, was also viewed as unrealistic. The group noted that data do not
exist that define achievable rates, or their relationship to quality of care. The group then
concluded that using viral suppression as a quality measure, even in limited well-defined
populations, would require consideration of several other factors that might alter the
interpretation of non-suppression. These include the distribution of patients at risk for decreased
adherence or other causes of non-suppression, ART experience and resistance.

Group 2

Question: What should the performance measure for viral load be in settings of different

      Naïve vs. experienced?
      Patients with vs. without documented resistance?
      Adherence challenged populations?
      Others (HCV, for example)?

This group suggested the approach of categorizing patients into three tiers according to estimates
of achievable rates of viral suppression. Classification of patients into these tiers was based upon

                                         Viral Load Report

treatment history and resistance, and achievable suppression rates were based on clinical trials

        TIER I. Treatment-naïve patients with wild-type virus. Patients with acquired single-
        class drug resistance are included. Treatment-experienced patients with wild-type virus
        where regimens are switched because of side effects are also included.

        TIER II. Treatment-experienced patients who are on a secondary regimen because of
        resistance or treatment-naïve patients with 2-class resistance determined by genotype.

        TIER III. Treatment-experienced patients who have failed their second or third
        regimens. This group may also include patients with multi-class drug resistance.

For each tier, if viral load does not decline and the patient is non-adherent, the patient may not
automatically move into a new tier. A resistance test would need to determine whether in fact
true virologic failure has occurred that qualifies the patient for moving into a higher tier. A
patient, for example, may be prescribed a change in regimen to promote improved adherence,
just as when side effects or toxicities occur, that would not be considered a higher tier regimen
but would still be considered first-line. In this model, only resistance can be used to reclassify
the tier, which would then change the expected quality marker. In the absence of resistance
testing, clinical decision based on treatment history would be required to change tiers.

Tier                Potential Performance Measure
I                   VL<50 copies/mL in 24 or 48 weeks following initiation
II                  VL<50 copies/mL
III                 Hierarchical approach based on feasibility including absolute level of
                    suppression (<50 copies/mL); magnitude of VL decline (>1.0 log10);
                    CD4 count <200 cells/mm3; CD4 count <100 cells/mm3

For Tier I, a simple measure was proposed based on the assumption that viral suppression should
be achievable to levels below 50 copies/mL measured at 24 and 48 weeks following initiation of
therapy. Issues that were raised for this group of patients included the availability of specific
viral load assays that would influence the rate of suppression. For example, if the viral load test
available at a clinic does not measure as low as 50 copies, should the measure be < 400
copies/mL, or, for quality improvement, should ultrasensitive tests capable of measuring to this
50 copies/mL level become the standard of care? Another measurement issue involves transient

                                        Viral Load Report

viremic episodes. The group agreed that blips and other transient viremic episodes, such as those
following vaccination, should not be linked with failure or nonadherence, and should be
excluded from the analysis.

For purposes of quality measurement, patients in Tier II should be able to achieve the same
outcomes as those in Tier I.

Patients who are classified as Tier III are highly treatment-experienced. This group requires
thoughtful clinical judgment, and a hierarchy of potential measures based upon what can
realistically be achieved by ART. Some patients might be able to achieve viral suppression to
rates below < 50 copies/mL. For those who cannot, a decline in viral load levels of > 1.0 log10
would be one potential measure. In conjunction with this measure, or if it cannot be achieved,
the appropriate quality measure should be whether a CD4 count above 200 cells/mm3 is attained
and maintained. Finally, for some patients who have limited therapeutic options, the goal would
be attaining and maintaining a CD4 count above 100 cells/mm3.

The group agreed that any viral load level and CD4 count measured would need to be interpreted
in the context of clinical status. For example, the determination of whether the patient is deemed
clinically “stable” or not might prompt reconsideration of the desired goal and measure for any
given patient. Clinicians would be expected to assess and treat both non-adherence and co-
morbid infections and to evaluate and identify resistance before concluding that treatment failure
has occurred. The potential for co-infection with HCV to influence viral load suppression rates
was raised. The group concluded that evidence did not support any effect on the expected
outcome that would alter the proposed measures, although tolerability of some regimens may be
compromised. Given that significant challenges would exist to implementing this type of
assessment on a large scale, the group also recommended testing these criteria in a cohort
analysis to determine whether the measures were valid and reliable and whether the generated
information would make sense in the context of clinical practice and quality management.

                                         Viral Load Report

Group 3
Question: How can we use viral load as a measure of quality in the absence of suppression?
What should the performance measure be in the absence of suppression? What other markers
would you need to measure quality of care in the absence of suppression?

The group discussed specific populations of patients whose viral load levels are not fully
suppressed and how to assess the quality of their care and management. One population
identified included patients with viral load levels of more than 1,000 copies/mL measured
following 8 weeks of treatment on a given regimen and who were both actively engaged in care
as defined by at least two clinic visits and had a CD4 < 200 cells/mm3. Identification of patients
meeting these criteria would trigger a series of suggested steps that could be useful indicators to
evaluate the quality of care. If the patient is confirmed as still being prescribed ART, then
adherence should be assessed. If non-adherence is identified, then an intervention and follow-up
viral load level should be obtained. As part of this quality measure, assessment would include
whether resistance testing was performed, and, if non-adherence is identified, whether an
adherence intervention results in viral suppression. If the patient has drug-resistant virus, then
the measure would include indication of whether a change in the regimen was made based upon
the results. The group emphasized that this last step would be difficult to standardize given that
clinically appropriate care may involve not changing the regimen even in the setting of resistance
because of limited treatment options, among other factors.

If the patient is not receiving antiretroviral therapy, the record should document why the patient
was not prescribed therapy. Possible reasons which may not indicate a gap in quality would
include participation in a treatment readiness program and refusal, among others. The group
noted that this general approach would be useful in identifying for further evaluation specific
patients with higher risk of potential problems related to the quality of their care. It would not
provide useful clinic rates of non-suppression because this assessment would only examine a
subpopulation of patients, rather than the general population attending the clinic.

                                          Viral Load Report

This approach would require more extensive data collection in order to obtain the elements
necessary to assess whether appropriate care is being provided. In most cases, record review
would be required, unless a specifically constructed database were designed for this purpose.

How to implement viral load as a quality measure at a public health level
Similar to the description of management of individual patients in clinical practice and
measuring quality of care among a clinic population, the group also suggested consideration of
multiple factors which may be useful for interpreting population-wide viral load levels as a
measure of quality or unmet needs. These factors include those patient and virologic factors
which impact upon viral suppression rates, discussed in detail in pervious sections. As in the
other scenarios, appropriate interpretation would require other strategies for using viral load
levels as a measure. Rather than suppression rates to measure quality of care in a population,
other measures such as change in viral load levels over time, suppression to a specified threshold
or CD4 count responses would likely be more useful. They would need to be linked with other
patient-level factors which would change the probability that the virologic goal could
realistically be achieved. However, this approach would require significant data and resources to
collect as well as expertise in interpretation. The potential feasibility and reliability of these
measures, even if these factors would be collected, was a matter of considerable concern to the
panel which repeatedly emphasized the complexity of interpreting the absence of suppression in
the context of quality.

Viral load measures in a population have also been proposed to assess whether patients are fully
connected to care. Using performance of one viral load test to identify patients who have had
contact with the care system, analysts could assess whether subsequent tests are obtained or
whether clinical visits are billed or recorded to determine if the patient has received any follow-
up care and connection with the system of care. Detection of only one viral load test would
suggest that a patient has been lost from the system and might need to be located. Whether this
effort would be feasible and worth expenditure of resources would require evaluation. Clinicians
also wanted to caution public health authorities from using these data to identify patients and
inappropriately attempting to force patients to accept services.

                                         Viral Load Report

Proposed next steps and areas for operational research
Throughout the discussions, multiple areas were highlighted for future work and operational
research, critical for further consideration of the potential role of viral load as a measure of
quality. If viral load is used to measure quality, appropriate clinical performance standards for
practice should be carefully chosen, clearly understood and examined for appropriateness once
implemented. One approach would be to first determine across different clinic settings the
highest possible response rates and ranges in clinical practice with current therapeutic options.
This process would entail the definition of appropriate cohorts, duration of follow up, and which
measures to use. Another approach would be to construct a longitudinal cohort of eligible
treatment-naïve patients in clinics to follow over time to assure uniformity across clinic
populations. Then, after 1-2 years, viral load outcomes and related appropriate quality measures
would be established. However, this longitudinal prospective cohort analysis would require a
long period of follow-up before it would be able to produce the data necessary to guide these

Opportunities to link viral suppression rates with electronic sources of data that include viral load
levels, treatment data and adherence service provision would greatly assist in the interpretation
and evaluation of the role of viral load as a quality measure while also minimizing resource
utilization for data abstraction. These may include Medicaid claims data, pharmacy refill rates,
service utilization data and epidemiologic surveillance data, which in New York State include
viral load and CD4 lab-based reports. It was noted that in New York State, there are ongoing
efforts to link databases including those at NYCDOMMH (surveillance), MHRA (Datalink),
IPRO (AIMS), HHC (Clinical Desktop Application) with the statewide Uniform Reporting
System (URS) which is required for all Ryan White CARE Act-funded providers.

Another example would be to use preliminary data from NYCDOHMH, which links viral load
levels and mortality rates by zipcode. This example illustrates the complexities of using these
types of data for assessing quality. Mortality data were described as useful, but must be
interpreted carefully to account for local factors linked with higher or lower rates of overall or
AIDS-related deaths, rather than quality or access to care

                                         Viral Load Report

The group also questioned whether the measurement of viral load suppression rates would add
any additional value to the existing data elements being measured in the assessment of ART
management undertaken by the New York State HIV Quality of Care Program. If added, they
noted that it would be important to define the purpose of using viral load information in the
context of these ongoing activities and then assess whether it would add value to the existing

Finally, the group raised concerns that clinics with viral suppression rates above the chosen goal
could become complacent which could, in turn, potentially lead to declining quality of care. It
will, therefore, be necessary to pilot any new viral load quality measure in small groups with
appropriate in-depth evaluation to determine if it is helpful or harmful to overall clinical
effectiveness and quality of care. In addition, the need to monitor the potential negative impact
of using viral load as a quality measure for assessment of treatment-challenged populations was

The expert panel developed the following specific recommendations for next steps:

1. Assess the feasibility and resources required to evaluate viral load levels longitudinally and for
cross-sectional analyses, comparing the potential resources and utility of both approaches to
measure quality and stimulate improvement.

2. Evaluate data from existing databases before implementing a widespread approach. One
example might be to include retrospective data from clinics to compare viral suppression rates
and outcomes for 2003 and 2004.

3. Data currently being collected through ongoing quality and monitoring projects at the city and
state levels should be assessed to determine if the addition of data elements to viral load levels
would contribute to their utility as a measure of quality. One example would be to use data for
this assessment from the NYS HIV Quality of Care Program, and the National HIVQUAL
project, which include definitions of stable and unstable patients, and incorporate ART and
resistance data.

                                         Viral Load Report

4. Consider evaluating subgroups that experience higher rates of recent treatment failure, and
assess these subgroups to determine if appropriate and rapid interventions have been provided
that ultimately decrease risk of viral resistance. These groups may include patients with viral
load levels between 50 and 400 copies/mL, or those receiving ART with detectable viral load
levels above a specified higher threshold which would indicate treatment failure, such as above
10,000 copies/mL.

5. Consider developing new models for exploring how viral load and other clinical data might be
used to evaluate and improve care. One specific proposal was to pilot test new measures for
treatment-experienced patients using a hierarchical matrix based on viral load levels, CD4
counts, treatment history and resistance data to determine potential utility.

6. Explore the possibility of using viral load measurement as a process measure to track patients
in and out of care, identify if they are receiving care in different clinics, and identify those who
are potentially out of care and would require assistance to return to care. The absence of follow-
up viral load test results would accordingly suggest a lapse in care. However, a large database
which looked solely at these results and suppression rates in the absence of other data was
uniformly considered to be of little use or value for measuring quality. This process would have
to be compared with visit data or other measures to assess whether patients are connected to care.

Other relevant questions which could not be answered include whether there are regional
differences in viral suppression and whether they signify differences in quality in clinical
practice or in the population? Which viral suppression rates represent desired outcomes for
populations in care? Can we use viral load data to help define results of different approaches
that would inform its use as a marker of quality?

                                         Viral Load Report

The issues involved in consideration of the role of viral load measurement in the assessment of
the quality of HIV care are complex, requiring a careful interpretation of viral load test results in
the broader clinical context in which they are used. The process to implement viral load to
assess quality of care is further complicated by the need to define appropriate performance
measures for application in a clinic or among specific populations.

Currently, the appropriate performance measure for viral load, whether suppression rate,
magnitude or rate of change, maintenance below a threshold or some other measure, remains
unknown. To measure quality at a population level, or even in many subpopulations, viral
suppression rates are not useful as an independent measure. In these settings, other viral load
measures would need to be considered such as change in viral load levels over time. They would
also likely be combined with other measures, such as CD4 count and clinical response. Whether
different measures are required for different populations of patients such as those on treatment or
those not on treatment is similarly unclear. Regardless of which measures are chosen, the
assessment would always need to include other patient-level factors which influence the
probability that suppression or even viral load response can be attained.

Before rates of viral suppression or other viral load measurements are used to assess quality of
care, they should be further evaluated to assess whether they are valid indicators of best practice.
Resources required to collect and report meaningful data need to be assessed. Furthermore, the
impact of using these data to assess quality requires further study to assure that a decline in
appropriate care does not occur as a result of the measures chosen. Concerns were raised by the
panel about the potential misinterpretation of lower suppression rates or other scores which could
unfairly penalize providers of patients with more risks for lower virologic response and which
would ignore the role of patient preference and choice in use of ART.

In conclusion, many unanswered questions remain regarding the potential utility and benefit of
using viral load results to assess quality of care at the population and clinic levels. The panel
reached consensus that even within the context of clinical management of an individual patient,
viral load measurements, including the presence or absence of suppression, when used by

                                         Viral Load Report

themselves have significant limitations as criteria for defining optimal care. Both their
interpretation and appropriate care in the setting of viremia will continue to evolve as our
understanding of optimal ART management expands. Both clinicians and public health officials
must recognize that suppression rates do not always reflect the quality of care provided because
of the multiple patient and virologic factors which may render suppression unachievable. It is
therefore critical that we better understand how to use viral load data and the patterns of viral
response within the context of assessing and improving the quality of care before they are used to
determine if optimal care is being offered.

Public health agencies must define and measure quality with careful attention to desired and
realistically achievable health outcomes. If outcomes are measured that cannot be achieved,
providers may choose not to provide care or prescribe treatment to vulnerable patients who are
less likely to become suppressed. The community of people affected by HIV, providers serving
them, and public health agencies all share the common goal of implementing measurement of
ART therapy in order to improve the quality of care and health outcomes for people living with
HIV. Viral load measurement may be valuable in achieving these goals. As of today, however,
its implementation remains unchartered and its benefit unproven.

                                        Viral Load Report


Appendix A:
Guidelines for use of viral loads and CD4 counts: NYSDOH Adult Guidelines
Viral load recommendations (NYSDOH HIV Adult Guidelines)
      ART therapy should suppress the viral load to the lowest possible level for as long as
      Standard viral load assays with a lower limit of sensitivity of 400 copies/mL are adequate
       for measuring plasma HIV levels for persons who are not receiving ART therapy.
      Ultrasensitive assays that detect as few as 25 to 50 copies/mL should be used to monitor
       patients who have viral loads < 400 copies/mL

CD4 recommendations (NYSDOH HIV Adult Guidelines)
      Clinicians should repeat CD4 or viral load results that are inconsistent with the clinical
       presentation before management decisions are made.
      Historically, quantization of CD4 cells was the first effective predictor of HIV
       progression and still is for persons infected with HIV-2 or HIV-1 variants that cannot be
       accurately quantified using viral load assays.
      Absolute CD4 cell counts are calculated values that may fluctuate widely. The CD4
       count = total white blood cell count (in thousands) x % of total lymphocytes x % of CD4
       lymphocytes. Therefore, any change in one of these three parameters will cause the
       absolute CD4 count to vary. As a result, HIV clinicians should measure and follow the
       CD4 percentage in addition to the absolute count because the CD4 percentage is a direct
       measurement and more reliable.

                                                            Viral Load Report

Appendix B: NYSDOH AIDS Institute HIV Quality of Care Program– Adult and Adolescent Indicators (Ages 13 Years and Above)30

                CATEGORY                                      ELIGIBILITY                                           MEASURE
Antiretroviral (ARV) Therapy Indicators     General Eligibility: Patients who are either
(appropriate management of ARV therapy,     receiving ARV therapy, received ARV in the
treatment adherence, care by an HIV         past, or are eligible for ARV therapy based
specialist)                                 upon current New York State ARV therapy
                                            guidelines. Patients incarcerated and with no
                                            ambulatory clinic visits during any four-month
                                            review period are not eligible for review
                                            during that period.
Appropriate Management for patients         Eligibility: Viral load is undetectable; or has   Measure: The number of patients for whom viral load is
stable on ARV therapy                       dropped by at least one log10 since last 4-       monitored every four months
                                            month review period; or has increased by less
                                            than 3X from the lowest value in last 12
                                            months on that regimen; or there is a note in
                                            record by treating physician that patient is
                                            deemed stable.

                                                                   - 46 -
                                                                   Viral Load Report

  Appropriate management for patients              Eligibility: Viral load has increased by more      Measure: One of the following four management options
       unstable on ARV therapy                     than 1 log10 and absolute value is over 1,000;     is documented in the medical record in every 4-month
                                                   or CD4 has dropped by 50% since the last 4-        period the patient is considered unstable:
(Note: ‘unstable’ is akin to treatment failure.)   month review period; or OI in the last four
                                                   month review period (new or recurrent); or                  Regimen was changed and viral load assay
                                                   patient deemed unstable by physician                   performed within 8 weeks of decision
                                                                                                               Justification provided not to change
                                                                                                          therapy (intercurrent illness, recent vaccination,
                                                                                                          adherence intervention documented, viral load
                                                                                                          reordered, patient prefers not to change
                                                                                                          medication, provider documents that patient is
                                                                                                          clinically/immunologically stable, resistance
                                                                                                          testing ordered, other) and viral load assay
                                                                                                          performed within 8 weeks of decision
                                                                                                               Documentation that patient decides not to
                                                                                                          take medication and viral load assay performed
                                                                                                          within four months
                                                                                                               Decision made to discontinue therapy and
                                                                                                          planned clinical follow-up plan noted in record
                                                                                                          within four months

Appropriate management for end-stage               Eligibility: (1) Patient meets unstable criteria   Measure: The number of patients for whom a follow-up
patients, or patients with no other                outlined above, but clinician documents that       clinic visit is recorded every four months
therapeutic options                                patient has no other therapeutic options
                                                   available, or
                                                   (2) Patient documented to be end stage within
                                                   last twelve months

Treatment Adherence                                Eligibility: All patients prescribed ARV           Measure: Adherence is measured and described
                                                   therapy                                            quantitatively at least once every four months

                                                                           - 47 -
                                                                       Viral Load Report

Other Performance Indicators
CD4 Count Measurement                                Eligibility: All patients, with the exception of Measure: The number of patients for a CD4 count test was
                                                     those incarcerated and with no ambulatory        performed every four months
                                                     clinic visits, during a four-month review period

Viral Load Measurement                               Eligibility: All patients, with the exception of        Measure: The number of patients for a viral load test was
                                                     those incarcerated and with no ambulatory               performed every four months
                                                     clinic visits during a four-month review period

                                                                               - 48 -
                                                                                      Viral Load Report

Appendix C: Summary of selected studies reviewed
           Citation                   Design                   Setting                  N                        Subjects                    (Intervention)     Starting/ Enrolled     Duration/ Follow Up
Abgrall, S., X. Duval, et al.       Prospective        HIV patients receiving         3,736        PI and NNRTI-naïve, not previously          first line PI-   from 1 January 1997     through 31 December
(2003). Clin Infect Dis          cohort study (The     treatment at 68 French                      included in a blinded clinical trial of       HAART                                          1999
37(11): 1517-26.                  French Hospital        university hospitals                      ART, and with VL < 500 on at least 1
                                    Database on                                                     occasion during initial PI-HAART,
                                       HIV)                                                                      aged 15+

Arnsten, J. H., P. A. Demas,       Longitudinal      Montefiore Medical Center          85         HIV-infected current and former drug         HAART            between July 1998       6 months follow-up
et al. (2002). J Gen Intern           study              Substance Abuse                            users recruited from the Bronx HIV                            and April 2000
Med 17(5): 377-81.                                     Treatment Program                           Epidemiologic Research on Outcomes
                                                                                                               (HERO) cohort
Bahrani, A., R. Ramaswamy,        Retrospective       HIV clinic of Eastern            497            “Control group" 75 patients with            (study          from early 1996      through March 1999 (24
et al. (2001). Clin Infect Dis    cohort study       Virginia Medical School,                      consecutive VL < 400 for the duration      coincided with                                month study)
32(8): 1231-2.                                                Norfolk                              of the study, with “time zero” defined    the introduction
                                                                                                      as the time of the first VL < 400.       of HAART)
                                                                                                      “Study group” 134 patients with
                                                                                                    consecutive VL > 400 (after having
                                                                                                    had an initial VL < 400), with “time
                                                                                                   zero” defined as the onset of virologic
Bangsberg, D. R., F. M.          Cross-sectional     Research in Access to Care         34         HIV-infected people with a median of        PI therapy         between January       and July 1998; 6±10
Hecht, et al. (2000). AIDS          analysis of      in the Homeless (REACH)                               12 months of PI therapy                                     1998             weeks follow-up for
14(4): 357-66.                    subjects in an         study, San Francisco                      systematically sampled from free meal                                                  each participant
                                  observational                                                       lines, homeless shelters and low-
                                   prospective                                                         income, single-room occupancy
                                      cohort                                                                     (SRO) hotels
Bangsberg, D. R., E. D.            Prospective          Research on Access to          330           HIV+ urban poor individuals on a         a stable ART         January 1998;         6 month follow-up
Charlebois, et al. (2003).         cohort study          Care in the Homeless                         stable regimen and with VL > 50            regimen           subjects were
AIDS 17(13): 1925-32.            (measuring the           (REACH) cohort, a                                                                                        recruited into
                                 number of new       systematic sample of HIV-                                                                                   REACH between
                                    HIV ART            positive adults recruited                                                                                July 1996 and April
                                    resistance            from San Francisco                                                                                           2000
                                  mutations that        homeless shelters, free
                                 occurred over 6       meal programs and low-
                                     months)             income single-room-
                                                           occupancy hotels
Bangsberg, D., S. Weiser, et       Longitudinal        Subjects were recruited     110 (56 on PI      HIV+ urban poor individuals on           single-PI or        REACH cohort         median of 9.1 months
al. (2005). Abstract # 616,           study           from the REACH Cohort,        and 54 on       single-PI or NNRTI ARV regimens              NNRTI           established in 1997         follow-up
12th Conference on                                        San Francisco; San         NNRTI)          for at least 3 months prior to study
Retroviruses and                                     Francisco General Hospital                                     baseline
Opportunistic Infections,

                                                                                               - 49 -
                                                                                         Viral Load Report

Bartlett, J., M. Fath, et al.     Meta-analysis of              various                 > 10,000                        N/A                        triple ART              ~1999                    ~ 2004
(2005). Abstract #586. 12th         triple ART
Conference on Retroviruses
and Opportunistic Infections,

Blankson, J. N., D. Persaud,        Review paper                  N/A                     N/A                           N/A                           N/A                   N/A                      N/A
et al. (2002 Annu Rev Med
53: 557-93.

Brigido, L., R. Rodrigues, et      Observational         Sao Paulo University,             148                   open clinical cohort            dual and triple     from October 1987         to February 2001
al. (2004). AIDS Patient            longitudinal                Brazil                                                                            antiretroviral
Care STDS 18(4): 189-98.               study                                                                                                         therapy

Deeks, S., G. Beatty, et al.       Retrospective         San Francisco General             108           > 24 weeks of continuous treatment      clinic treatment       from ~ 1995           18-month follow-up
(1998). Abstract #419, 5th          cohort study           Hospital (SFGH)                                   with a PI-regimen and two
Conference on Retroviruses         (chart review)                                                       consecutive VLs > 500 at the last two
and Opportunistic Infections,                                                                                        clinic visits

Deeks, S. G., J. D. Barbour, et     Cohort study        an ongoing observational           291            HIV-infected adults who failed to         protease         initiated a PI-based    27.9 months (median)
al. (2002). AIDS 16(2): 201-                              study of HIV-infected                         maintain durable viral suppression on    inhibitor-based      regimen prior to 1
7.                                                     patients receiving protease                       a protease inhibitor-based regimen         regimen                Jan 1998
                                                         inhibitor therapy at San                         initiated before 1 January 1998 +
                                                            Francisco General                            continuous treatment for at least 16
                                                            Hospital, an urban                                          weeks
                                                       municipallyfunded hospital
Deeks, S. G., R. Hoh, et al.      1. Cross-sectional      San Francisco General       71 patients: 36    HIV patients who continued using          ART clinic       cross-sectional study,   cross-sectional study,
(2002). J Infect Dis 185(3):            design;              Hospital (SFGH)         in the virologic      antiretroviral therapy despite          treatment         based on 18 months      based on 18 months +
315-23.                              2. Treatment                                      failure group     detectable plasma viremia (RNA                             + 24 weeks treatment      24 weeks treatment
                                     interruption                                      and 18 in the            copies >2500 /mL)                                          history                  history
                                        design                                           virologic
                                                                                      success and 17
                                                                                     in the untreated
                                                                                     group (controls)
Deeks, S. G., J. N. Martin, et      Prospective        ongoing prospective cohort           189           (1) “incomplete” virus suppression,         N/A               from ~ 1999          followed up at 4 month
al. (2004). J Infect Dis            cohort study        study of the LT outcomes                        continuous combination ART and VL                                                           intervals
189(2): 312-21.                                         of drug-resistant viremia                        > 50 (2) “complete” VS, continuous
                                                              (Study of the                               combination ART and VL < 50 (3)
                                                           Consequences of the                             ARV untreated, no ART received
                                                          Protease Inhibitor Era                          during the preceding 24 weeks and
                                                                [SCOPE])                                evidence of progressive CD4 cell loss
                                                                                                         (4) LTNPs, self-reported duration of
                                                                                                         HIV infection for at least 10 years +
                                                                                                         median CD4 count 1500, despite no

                                                                                                   - 50 -
                                                                                          Viral Load Report

Demeter, L. M., M. D.               Observational         40 AIDS Clinical Trials           489           pts who had a CD4 count of < 200 at         indinavir,      enrollment began in      40 weeks follow-up
Hughes, et al. (2001). Ann          analysis of one               units                                    week 8 or later and VLs at baseline     zidovudine, and       January 1996
Intern Med 135(11): 954-64.        treatment group                                                                     and week 8                    lamivudine
                                  in a phase III trial
Di Mascio, M., M.                    Longitudinal           clinical trial with 8           123                (HIV clinical trial participants)      HAART              all HAART era          mean of 2.6 years
Markowitz, et al. (2003). J              study                 different drug                                                                                                                  observation (range, 5
Virol 77(22): 12165-72.                                        combinations                                                                                                                    months to 5.3 years)

Durant, J., P. Clevenbergh, et      Prospective,         three hospitals in southern   control, N=43;       HIV-1-infected patients in whom        standard care or    from March, 1997,       3, 6 month follow-up
al. (1999). Lancet                     open,                       France                genotypic            combination therapy was not              treatment       until March, 1998,
353(9171): 2195-9.                  randomized,                                         group, N=65        successful: VL > 10,000 despite at      according to the     114 patients were
                                  controlled study                                                           least 6 months' treatment with            resistance      screened, of whom
                                   (ITT analysis)                                                          nucleoside analogues and at least 3       mutations in     108 were randomized
                                                                                                               months' treatment with a PI           protease and
Dybul, M. (2002). Curr              Review paper                    N/A                    N/A                               N/A                          N/A                 N/A                      N/A
Infect Dis Rep 4(2): 175-180.

Dybul, M., M. Daucher, et al.       Prospective           NIH-approved clinical             12                     adult HIV+ individuals              HAART              from ~ 1998           84 week follow-up
(2003). J Virol 77(5): 3229-        cohort study                protocol                                                                              structured
37.                                                                                                                                                   treatment
Esteve, A., A. Jaén, et al.         Longitudinal            PISCIS cohort study          1,542 (787       All HIV patients newly attending the        HAART,                  2000               3.6 years median
(2005). Abstract #611 12th          cohort study          (Catalonia and Balearic          virally        participating hospitals from January     treated mainly                                    follow-up
Conference on Retroviruses                               islands regional hospitals,     suppressed       1998, independently of the stage of      with PI-based
and Opportunistic Infections,                                      Spain)               throughout)                disease or degree of               regimens
Boston.                                                                                                           immunosuppression.

Fischl, M. A. (1999). AIDS          Review paper                    N/A                    N/A                               N/A                        N/A                   N/A                      N/A
13 Suppl 1: S49-59.

Florence, E., J. Lundgren, et       Longitudinal             EuroSIDA cohort                780             patients initiating HAART with            HAART                progressive           4 year follow-up
al. (2003). HIV Med 4(3):           cohort study                                                           baseline CD4 count < 350 + VLs <                           enrollment start dates
255-62.                                                                                                    500 6 to 12 months after initiating                         from 2 May 1994 to
                                                                                                                         HAART                                           September 2001
Fuller, J. D., D. E. Craven, et     Prospective           Boston Medical Center,            46                        HIV+ adults                     influenza              ~ 1998             4 week follow-up
al. (1999). Clin Infect Dis         cohort study                  MA                                                                                 vaccination
28(3): 541-7.

                                                                                                      - 51 -
                                                                                     Viral Load Report

Gandhi, R., E. Chan, et al.        Randomized        ACTG 384, 58 sites in the         980           treatment-naive patients with VL >          6 treatment       from October 1998 to      96 week follow-up
(2005). Abstract #610, 12th       controlled trial   U.S. and 23 sites in Italy                                     500                             arms:             November 1999
Conference on Retroviruses                                                                                                                     (1) d4T + ddI +
and Opportunistic Infections,                                                                                                                     efavirenz
Boston.                                                                                                                                        (2) d4T + ddI +
                                                                                                                                               (3) AZT + 3TC
                                                                                                                                                 + efavirenz
                                                                                                                                               (4) AZT + 3TC
                                                                                                                                                 + nelfinavir
                                                                                                                                               (5) d4T + ddI +
                                                                                                                                                 efavirenz +
                                                                                                                                               (6) AZT + 3TC
                                                                                                                                                + efavirenz +
Goodkin, K., P. Shapshak, et      Cross-sectional     Depts of Psychiatry and          135            135 older (> 50 years of age) and            HAART           N/A - cross-sectional    N/A - cross-sectional
al. (2004). AIDS 18 Suppl 1:          study           Behavioral Sciences, U                         younger (18-39 years of age) HIV+                                     study             study; Adjusted for
S87-98.                                                  Miami School of                                  individuals (at CDC early                                                               HAART
                                                       Medicine, Miami, FL                               symptomatic stage B or late
                                                                                                      symptomatic stage/AIDS C) were
                                                                                                    enrolled as part of a larger cohort also
                                                                                                      consisting of HIV-1-seronegative
Grabar, S., V. Le Moing, et        Prospective         68 hospitals in France         2,236            protease inhibitor-naive patients         Initiation of      between July 1996        6 months follow-up
al. (2000). Ann Intern Med         cohort study                                                                                                 HAART with           and March 1997
133(6): 401-10.                                                                                                                                  one protease
                                                                                                                                                inhibitor and
                                                                                                                                               two nucleoside
Greub, G., B. Ledergerber, et      Prospective         Frankfurt HIV Clinic        704 out of        HCV HIV co-infected patients from             effective        between 1 January      24 weeks follow-up; the
al. (2000). Lancet                observational      Cohort and the Swiss HIV        1,157           the Frankfurt HIV Clinic Cohort and        antiretroviral         1998 and 31           database used for the
356(9244): 1800-5.                cohort study             Cohort study                                the Swiss HIV Cohort who started        clinic treatment      December 1999             analysis included
                                                                                                     potent antiretroviral therapy between                                                 information recorded up
                                                                                                        June 1, 1996, and May 31, 1999                                                          to April, 2000
Greub, G., A. Cozzi-Lepri, et      Longitudinal      the Frankfurt HIV Clinic       704 out of       Patients with low-level viral rebound        HAART               samples taken          median follow-up of
al. (2002). AIDS 16(14):           cohort study      Cohort and the Swiss HIV     2,055 patients      on effective antiretroviral treatment                         between 1 January            17.7 months
1967-9.                                                    Cohort Study                             from the Frankfurt HIV Clinic Cohort                               1998 and 31
                                                                                                           and the Swiss HIV Cohort                                  December 1999
Gross, R., W. B. Bilker, et al.   Observational      General Clinical Research         41            PI-naive subjects with VLs > 10 000           HAART            between February       and November 1999; 4
(2001). AIDS 15(16): 2109-        cohort study        Center at University of                             referred from HIV clinics in         including initial          1998               months follow-up
17.                                                    Pennsylvania Medical                                       Philadelphia                    nelfinavir
                                                      Center Philadelphia, PA

                                                                                                - 52 -
                                                                                      Viral Load Report

Grossman, H., G. Frechette,         Longitudinal             not available              34             individuals with a mean of 9 drugs            received        not available          not available
et al. (1999). Abstract #23,        cohort study                                                                prior experience                     abacavir,
2nd International Workshop                                                                                                                       efavirenz and
on Salvage Therapy for HIV                                                                                                                           adefovir
Infection, Toronto, Canada.                                                                                                                           through
                                                                                                                                                 programs plus
                                                                                                                                                      two PIs
                                                                                                                                                       vir) +
                                                                                                                                                  AZT/3TC or
Gulick, R. M., X. J. Hu, et al.     Randomized,        multi-center (42 units of       277           (HIV)-infected patients >16 years old      saquinavir with    between September     24 week follow-up
(2000). J Infect Dis 182(5):      partially double-       the adult ACTG)                            naive to nonnucleoside analogues who          ritonavir or     1997 and October
1375-84.                          blinded, factorial                                                  had taken indinavir >6 months with            nelfinavir            1998
                                   study of 6 oral                                                             VL 2000–200,000                   together with
                                    antiretroviral                                                                                                 delavirdine
                                      "salvage"                                                                                                 and/or adefovir
Hammer, S. M., F. Vaida, et          Multicenter,       Thirty-one participating       481              HIV-infected persons with prior            Selectively      between October    and April 2000, with a
al. (2002). JAMA 288(2):           randomized, 4-           AIDS (acquired                            exposure to a maximum of 3 PIs and          randomized             1998             24-week primary
169-80.                           arm trial, double-      immunodeficiency                                         VL > 1000                    assignment (per                        analysis with extension
                                      blind and        syndrome) Clinical Trials                                                                     prior PI                               to 48 weeks
                                       placebo-        Units in the United States                                                                 exposure) to
                                    controlled for                                                                                              saquinavir (n =
                                      second PI                                                                                                 116); indinavir
                                                                                                                                                    (n = 69);
                                                                                                                                                 nelfinavir (n =
                                                                                                                                                     139); or
                                                                                                                                                 placebo twice
                                                                                                                                                  per day (n =
                                                                                                                                                     157); in
                                                                                                                                                 efavirenz, and
Harrigan, P. R., R. S. Hogg,       Retrospective          HOMER (HAART              1,191 (298         antiretroviral-naive adults initiating       HAART             from ~ 2001       30 month follow-up
et al. (2005). J Infect Dis        longitudinal         Observational Medical        resistant)              HAART with VL >1000
191(3): 339-47.                        study           Evaluation and Research)
                                                        cohort in B.C., Canada

                                                                                                  - 53 -
                                                                                      Viral Load Report

Haubrich, R. H., S. J. Little,      Prospective      Five university-affiliated         173           HIV-infected patients with a mean    Individualized,         ~ 1997          2, 6, months follow-up
et al. (1999). AIDS 13(9):         observational     HIV clinics in California,                      baseline CD4 count of 142 (range 3-     unrestricted
1099-107.                             study of                 USA                                  515) of whom 164 and 119 completed      antiretroviral
                                   adherence to                                                      adherence questionnaires at 2 and 6       therapy
                                 therapy nested in                                                           months, respectively
                                   a randomized
                                 comparative trial
                                     of frequent
                                 versus infrequent
                                   monitoring of
                                    plasma HIV
Havlir, D. V., R. Bassett, et      Retrospective     well-characterized cohorts,        241         ACTG 343 patients receiving triple-    various ART in      August 1997 to        84 week median
al. (2001). JAMA 286(2):            cohort study      the AIDS Clinical Trials                      drug therapy + patients from Merck       ACTG and          November 1998       observation follow-up
171-9.                                               Group (ACTG) 343 trial of                      035 with VL < 200 after 6 months of    Merck 035 drug     (ACTG 343) and
                                                       induction-maintenance                          indinavir-zidovudine-lamivudine           trials        since March 1995
                                                     therapy and the Merck 035                                                                                (Merck 035 trial)
Hicks, C., M. S. King, et al.    Long-term, open-           multicenter            100 (16 vs. 51     antiretroviral-naive HIV-infected     Three blinded        from ~ 1999       204 weeks follow-up
(2004). AIDS 18(5): 775-9.        label follow-up                                    vs. 33; 72           patients with VL > 5,000         doses of LPV/r
                                   of a phase II,                                  followed up at                                           with stavudine
                                    prospective,                                     week 204)                                                40 mg and
                                 randomized, trial                                                                                         lamivudine 150
                                                                                                                                             mg every 12
                                                                                                                                           hours. After 48
                                                                                                                                             weeks, LPV/r
                                                                                                                                              was dosed
                                                                                                                                             open-label at
                                                                                                                                             400/100 mg
                                                                                                                                            every 12 hours
                                                                                                                                            with stavudine
                                                                                                                                           and lamivudine
Hirsch, M. S., B. Conway, et           N/A            Recommendation of an             N/A                          N/A                          N/A                N/A                    N/A
al. (1998). JAMA 279(24):                              International AIDS
1984-91.                                               society-USA panel

Hirsch, M., R. Steigbigel, et     Randomized,                 21 sites                  320          HIV+ patients > 18 years old, with       indinavir,     between August 1995    24 week follow-up
al. (1999). J Infect Dis          double-blind,                                                      CD4 count < 50 and extensive prior    zidovudine, and      and April 1996
180(3): 659-65.                  multicenter study                                                         zidovudine therapy                lamivudine

Hirsch, M. S., F. Brun-                N/A            Recommendation of an             N/A                          N/A                         N/A                 N/A                    N/A
Vezinet, et al. (2000).                                International AIDS
JAMA 283(18): 2417-26.                                 society-USA panel

Holodniy, M., L. Mole, et al.     Longitudinal        AIDS Research Center,              8                     HIV+ veterans                    N/A                ~ 1999                 1 month
(1994). J Acquir Immune           observational       Department of Veterans
Defic Syndr 7(4): 363-8.              study           Affairs Medical Center,
                                                       Palo Alto, CA 94304

                                                                                                - 54 -
                                                                                        Viral Load Report

Hunt, P. W., S. G. Deeks, et        Cohort study     four clinic-based cohorts in        423           HIV-infected patients who initiated      initial HAART      initiated HAART      for up to 4 years or until
al. (2003). AIDS 17(13):                              San Francisco, Cleveland,                       HAART prior to 1998 and achieved a                             prior to 1998      plasma HIV RNA levels
1907-15.                                                      and Seattle                              viral load </= 1000 copies/ml by 48                                                increased to > 1000
                                                                                                                      weeks                                                                     copies/ml
Jacobson, L. P., R. Li, et al.      Longitudinal     Multicenter AIDS Cohort             679            HIV-infected HAART-naïve men               HAART           since January 1993      3.5 year follow-up
(2002). Am J Epidemiol                 study                  Study                                                                                                   (pre-HAART)
155(8): 760-70.

Jensen-Fangel, S., H. F.             Prospective,    Dept Infectious Diseases,            56          HAART experienced (2 NRTIs + 1 PI             nevirapine     between October        24 week follow-up
Thomsen, et al. (2001). J             open-label,     Marselisborg/Aarhus                              [saquinavir-hgc/ritonavir/indinavir]),    combined with     1998 and October
Acquir Immune Defic Syndr         randomized study     University Hospital,                           age > 18 years, nelfinavir and NNRTI        nelfinavir and         1999
27(2): 124-9.                       (ITT analysis)           Denmark                                      naive, indication for changing PI      two nucleoside
                                                                                                         because virologic or immunologic              reverse
                                                                                                            failure, and adverse effects or        transcriptase
                                                                                                       replacement of saquinavir-hgc with a          inhibitors
                                                                                                                    more potent PI                    (NRTIs)
Katlama, C., S. Dominiguez,          Open label       Hôpital Pitié-Salpêtrière,      68 patients     CD4 cells < 200 and VL > 50 000, on           immediate           ~ 2000          12, 20, 24, 32, 48 week
et al. (2002). Abstract #5887,    randomised trial     Hôpital Bichat, Paris,       (N=34 in each     a currently failing ART treatment (> 3    regimen change                                 follow-up
XIV International AIDS                                         France                    arm)         drugs), experienced to RTIs, NNRTIs       to GigaHAART
Conference, Barcelona,                                                                                and PIs. Exclusion criteria: acute OIs     (with > 8 ART
Spain.                                                                                                     or intolerance to any drug class        drugs) vs. 8-
                                                                                                                                                   followed by
Kaufmann, G., I. Perrin, et al.    Longitudinal       Swiss HIV cohort study            6,509              patients initiating HAART             initial HAART          ~ 1997            4, 7 years follow-up
(2005). Abstract #612, 12th         cohort study                                                                                                 (> ART drugs)
Conference on Retroviruses         (retrospective
and Opportunistic Infections,         analysis)

Kempf, D. J., R. A. Rode, et             Two         3 Randomized Controlled        (1) 29; (2) 102    (1) patients receiving ritonavir (PI)       initial PI-          ~ 1995            400 days/52 weeks
al. (1998). AIDS 12(5): F9-         retrospective     Drug Trials (M93-112 -                          whose VL rebounded from viral nadir           therapy
14.                                 longitudinal     M94-169, M94-245, M96-                               with resistance to ritonavir; (2)
                                      analyses                 462)                                     patients who initially responded to
                                                                                                         randomized treatment with either
                                                                                                        ritonavir or ritonavir + zidovudine
Kieffer, T., M. Finucane, et      Cross-sectional    Johns Hopkins University             286         patients with VL < 50 for the previous          N/A             from ~ 2003         N/A (cross-sectional
al. (2004). Abstract #650,        nested RT-PCR        School of Medicine,          independent RT                   15 months                                                                  study)
11th Conference on                 amplification          Baltimore, MD               or protease
Retroviruses and                      strategy                                        clones were
Opportunistic Infections, San                                                       obtained from 8
Francisco.                                                                          HIV-1-infected

                                                                                                 - 55 -
                                                                                     Viral Load Report

Knobel, H., A. Guelar, et al.      Prospective        Department Infectious           842           28 HIV patients aged 60 years or older        first-line      initiated treatment    24-month follow-up
(2001). AIDS 15(12): 1591-         cohort study       Diseases, Hospital del                         vs. 671 patients aged 40 years or less       HAART           between 1997 and
3.                                                     Mar, Autonomous                                                                        containing two              1998
                                                     University of Barcelona,                                                                    nucleoside
                                                        Barcelona, Spain                                                                      analogues and a
Kousignian, I., S. Abgrall, et      Prospective       HIV patients receiving         3,736           PI and NNRTI-naïve patients with an        first line PI-    from 1 January 1997   To 31 December 1999
al. (2003). J Acquir Immune      cohort study (The    treatment at 68 French                            undetectable viral load on a first        HAART          through 31 December    (12 month follow-up)
Defic Syndr 34(1): 50-7.          French Hospital       university hospitals                         course of highly active antiretroviral                               1998
                                    Database on                                                          therapy (HAART), aged 15+
Landon, B. E., I. B. Wilson,      Controlled pre-    44 intervention clinics and     9,986                                                       Quality                1999                     2001
et al. (2004). Ann Intern             and post-           25 control clinics                                                                   Improvement
Med 140(11): 887-96.                intervention                                                                                               Collaborative
Law, W. P., C. J. Duncombe,        Retrospective         8 ART randomized             692            Thai HIV-infected patients including        initial          patients initiated    -4 (screening), baseline,
et al. (2004). AIDS 18(8):          longitudinal      controlled trials at HIV-                      with HCV and/or HBV co-infection          combination        therapy between        4, 8, 12, 24, 36 and 48
1169-77.                          cohort analysis       NAT, HIV outpatient                                                                       ART            December 1996 and           week follow-up
                                                      clinic of Chulalongkorn                                                                                       March 2001
                                                         Hospital, Bangkok,
Lawrence, J., D. L. Mayers, et    Randomized           16 units of the CPCRA       138 vs. 132         Patients (> 13 years old) with          STI followed      Enrollment opened in     20 month follow-up;
al. (2003). N Engl J Med           controlled           consortium (conducts         control          multidrug-resistant HIV and VL >        by a change in          June 2000         analyses censoring date:
349(9): 837-46.                   nonblinded,         clinical research on HIV                                      5,000                      ART regimen                                 May 9, 2002; study
                                 multicenter trial   and AIDS in U.S. primary                                                                  vs. immediate                             closed prematurely on
                                 (CPCRA 064)                care settings)                                                                    change in ART                                  June 26, 2002
Ledergerber, B., M. Egger, et    Prospective data    Swiss HIV Cohort Study          2,674           adult outpatients from the Swiss HIV     clinic treatment   patients who started     30 month follow-up
al. (1999). Lancet                   analysis                                                                    Cohort study                                    HAART in 1995-98
353(9156): 863-8.

Ledergerber, B., J. D.             Longitudinal        13 HIV cohorts from           2,488               HIV patients with three-class         single or dual         1989-1999           5,015 person years
Lundgren, et al. (2004).           cohort study       Europe, North America,                         virological failure (viral load >1000    ART as part of                                  follow-up
Lancet 364(9428): 51-62.                              and Australia (PLATO)                             copies per mL for >4 months)          clinic treatment

Lincoln, D., K. Petoumenos,        Longitudinal           Australian HIV             2,086           AHOD patients including with HBV         initial HAART       by September 2002     up to 8 years follow-up
et al. (2003). HIV Med 4(3):       cohort study       Observational Database,                             and HCV coinfection
241-9.                                                from 26 sites (hospitals,
                                                      sexual health clinics and
                                                       GPs) around Australia
Lohse, N., N. Obel, et al.       Population-based     The Danish HIV Cohort          2,722            Danish patients initiating HAART        initial HAART            ~ 1997              7 year follow-up
(2005). Abstract #594, 12th       observational                Study
Conference on Retroviruses         cohort study
and Opportunistic Infections,

                                                                                                 - 56 -
                                                                                      Viral Load Report

Low-Beer, S., B. Yip, et al.       Longitudinal       British Columbia Centre     886 (766 men)      patients naive to triple therapy with    two NRTIs +          from ~ 1997          median of 19 month
(2000). J Acquir Immune               study          for Excellence HIV/AIDS                        CD4 count consistently < 500 or VL >       one PI or a                                  follow-up
Defic Syndr 23(4): 360-1.                            Drug Treatment Program                                          5000                       NNRTI

Lucas, G. M., R. E. Chaisson,     Retrospective      Johns Hopkins HIV Clinic          273                       PI-naive pts                   PI regimen         HAART was           14 months follow-up
et al. (1999). Ann Intern         cohort study           in Baltimore, MD                                                                      containing at     initiated between
Med 131(2): 81-7.                                                                                                                             least one other    March 1996 and
                                                                                                                                               ART drug to        February 1998
                                                                                                                                               which the pts
                                                                                                                                                  had not
                                                                                                                                             previously been
Lucas, G. M., K. A. Gebo, et       Prospective        The Johns Hopkins HIV            695          protease inhibitor–naive HIV patients       PI-HAART        between March 1996     14 months follow-up
al. (2002). AIDS 16(5): 767-       cohort study               Clinic                                  who completed two or more semi-                            and February 1998
74.                                                                                                 annual standardized surveys initiated
                                                                                                                 on HAART
Maggiolo, F., L. Ravasio, et     Prospective study     an infectious diseases          543            on steady HAART (previous > 6             HAART              from ~ 2003          6 month follow-up
al. (2005). Clin Infect Dis                             outpatient clinic in                                months) with VL< 50
40(1): 158-63.                                        northern Italy: Ospedali
                                                         Riuniti, Bergamo
Manfredi, R. and F. Chiodo         Case–control      Department of Clinical and      21 vs.84         21 consecutive patients 55 years of       first-line          from ~1999          12-month follow-up
(2000). AIDS 14(10): 1475-            study           Experimental Medicine,                        age or older retrospectively compared       HAART
7.                                                     Division of Infectious                        with 84 HIV+ individuals randomly       containing two
                                                      Diseases, University of                       selected from nearly 450 patients aged     nucleoside
                                                        Bologna, S. Orsola                             35 years or less who had started a    analogues and a
                                                      Hospital, Bologna, Italy                       similar HAART regimen since 1997        potent protease
Manfredi, R., L. Calza, et al.    Cross-sectional     HIV outpatient center in    67 (out of 930)              > 55 years old                     ART           from 1 January 2002    all on HAART > 12
(2003). J Acquir Immune               study           Bologna, northern Italy                                                                                                         months; to December 31
Defic Syndr 33(1): 112-4.                                                                                                                                                                      2002

Mellors, J. W., C. R. Rinaldo,     Longitudinal      The Pittsburgh portion of         180               HIV+ gay or bisexual men             N/A (natural      between April 1984           10 years
Jr., et al. (1996). Science           study           the multicenter AIDS                                                                    history study)      and march 1985
272(5265): 1167-70.                                    cohort study MACS

Mezzaroma, I., M. Carlesimo,          Open           the Department of Allergy          77          multi-drug experienced AIDS patients       PI-HAART         April and September         24 months
et al. (1999). Clin Infect Dis     longitudinal       and Clinical Immunology                             with CD4 counts < 50 in 2                                     1996
29(6): 1423-30.                       study          at the University of Rome                      determinations at least 1 month apart
                                                           “La Sapienza”                             and intolerant of or unresponsive to
                                                                                                            NRTI-based therapies
Mijch, A. M., J. Hoy, et al.      Retrospective         The Alfred Hospital,           555            All individuals treated at between     clinic treatment   between January and   outcome was assessed at
(2001). J Clin Virol 22(3):          cohort             Melbourne, Victoria                          January and June 1997 who had had                               June 1997             October 2000
271-8.                                                                                                  plasma HIV RNA measured.

                                                                                                - 57 -
                                                                                      Viral Load Report

Miller, V., A. N. Phillips, et     Longitudinal      The EuroSIDA study ( >            1,106           patients > 16 years old with CD4         clinic treatment     progressive cohort     follow-up to summer
al. (2002). J Infect Dis           cohort study        8,500 patients from 64                                     count < 500                                      enrollment start dates           2000
186(2): 189-97.                                      centers across Europe and                                                                                      from 2 May 1994 to
                                                         Israel; information                                                                                         Sept 2001; Median
                                                     collected from case notes                                                                                     HAART start date in
                                                      at baseline and every 6                                                                                               1997
                                                         months thereafter)
Mocroft, A., L. Ruiz, et al.      Observational           patients from the            2,444         Patients were followed from their first    clinic treatment        progressive           2 years follow-up
(2003). AIDS 17(12): 1741-        cohort study       EuroSIDA study (70 HIV                           viral load under 400 copies/ml to the                        enrollment start dates
51.                                                    centers across Europe)                          first of two consecutive viral loads                         from 2 May 1994 to
                                                                                                               above 400 copies/ml                                    September 2001
Montaner, J. S., P. R.            Observational      a single, university-based         106             Patients who failed several prior        most common       between August 1997        ITT VL analysis:
Harrigan, et al. (2001).             study             tertiary referral clinic                       regimens started MDRT (multi-drug             starting           and June 1998        between weeks 47 and
AIDS 15(1): 61-9.                                                                                    rescue therapy ) including at least five    regimen: three                                57 of follow-up;
                                                                                                            antiretroviral (ART) drugs          NRTIs + 2 PIs,                               Pre/post VL change
                                                                                                                                                   which was                                time-span: 15 months
                                                                                                                                                prescribed to 45                                   median
Moore, A. L., O. Kirk, et al.    Prospective study   63 centers across Europe       2,547 (511        patients naive to protease inhibitors         HAART            progressive cohort      72 month follow-up
(2003). J Acquir Immune                               (subjects were from the         female)              and nonnucleoside reverse                               enrollment start dates   since starting HAART
Defic Syndr 32(4): 452-61.                               EuroSIDA cohort)                             transcriptase inhibitors, and had at                         from May 1994-Sept
                                                                                                          least one viral load and CD4                                 2001; Median
                                                                                                     measurement prior to starting HAART                             HAART start date
                                                                                                                                                                        was in 1997
Moore, R., J. Keruly, et al.       Longitudinal      Johns Hopkins University,         1,326           cohort of HIV-infected patients in       initial HAART          1996 to 2002         6, 12 month follow-up
(2004). Abstract #558, 11th        cohort study           Baltimore, MD                               longitudinal HIV care in Maryland
Conference on Retroviruses
and Opportunistic Infections,
San Francisco.

Moore, R., J. Keruly, et al.       Longitudinal      Johns Hopkins, Baltimore,     256 patients >        patients who started their first          first line         all HAART era              1,200 days
(2004). 11th Conference on         cohort study                MD                   50 years vs.      HAART regimen at age ≥50 years               HAART
Retroviruses and                                                                   1,322 patients     were compared with adult patients
Opportunistic Infections, San                                                     35-50 years and       <50 years in an cohort of HIV-
Francisco.                                                                        895 patients 20-   infected patients in longitudinal HIV
                                                                                    34 years old                      care

                                                                                                 - 58 -
                                                                                         Viral Load Report

Murphy, D. A., M. Sarr, et al.         Structured        REACH (Reaching for              114              human immunodeficiency virus-            HAART              from ~ 1998         N/A - questionnaire
(2003). Arch Pediatr                interviews were     Excellence in Adolescent                        infected adolescents (12-19 years old)                                                    study
Adolesc Med 157(3): 249-55.           conducted to      Care and Health) Project                          recruited from 13 US cities in the
                                     determine the                                                       REACH cohort, prescribed HAART
                                    associated with
                                   component factor
                                      analysis was
                                     performed on
                                    scores of the 19
                                    barrier variables
Murri, R., A. C. Lepri, et al.       Observational         67 Italian infectious      2323 men and             previously ART naive                initial ART        started in 1997    > 28 weeks follow; cut-
(2003). J Acquir Immune               cohort study       disease wards; part of a     1335 women                                                                                           off date for reported
Defic Syndr 34(2): 184-90.                               multicenter cohort study                                                                                                         data to the data-base:
                                                            (I.Co.N.A. Study)                                                                                                              September 24, 2002
Nasta, P., F. Castelnuovo, et        Open-label,        HIV clinics in hospitals in    102 vs. 99        Patients with ≥ 2 previous HAART              current           ~ 2003          24, (48) week follow-up
al. (2005). Abstract #605,           randomized,         northern Italy (Master-                         failed regimens, with 1,000 < VL <        treatment vs.
12th Conference on                 prospective trial      IM.P.R.O.V.E study                            20,000 for at least 6 months, and CD4       switch to an
Retroviruses and                                                  group)                                   count ≥ 200 twice consecutively           optimized
Opportunistic Infections,                                                                                                                        lopinavir/ritona
Boston.                                                                                                                                           vir-containing
Nettles, R. E., T. L. Kieffer,      Active patients      the Moore Clinic at the           10           HIV+ asymptomatic adults with VL <            a stable      from June 19, 2003     to February 9, 2004
et al. (2005). JAMA 293(7):           underwent          Johns Hopkins Hospital                               50 while receiving a stable         antiretroviral
817-29.                            sampling (every                                                      antiretroviral regimen for 6 months or        regimen
                                   2-3 days) over 3                                                                      longer
                                    to 4 months to
                                       define the
                                   magnitude, and
                                   duration of blips
                                       and their
                                   association with
                                    drug levels and
                                     other clinical
Ostrowski, S. R., T. L.               Prospective          Dept of Infectious             101           patients with reproducible VL > 200          various         September 1997–       24 month follow-up
Katzenstein, et al. (2005). J        cohort study       Diseases, Rigshospitalet,                                                                   HAART              August 1998
Infect Dis 191(3): 348-57.                               Copenhagen, Denmark                                                                        regimens

Palella Jr, F. J., J. S. Chmiel,      Prospective       The HIV Outpatient Study      1,022 (out of      patients with CD4 count ever < 500         HAART           from January 1996     to April 1999 (at least
et al. (2002). AIDS 16(12):          observational      (HOPS) cohort from eight         1,769)                                                                                           12 months follow-up)
1617-26.                                cohort             clinics in the USA

Palepu, A., M. Tyndall, et al.      Retrospective           Population-based              578                   HIV+, naive to ART                    ART            between August 1       until March 2002
(2003). J Acquir Immune              clinic study           HIV/AIDS Drug                                                                                           1996 and December

                                                                                                    - 59 -
                                                                                         Viral Load Report

Defic Syndr 32(5): 522-6.                                 Treatment Program in                                                                                        31 2000
                                                       British Columbia, Canada
Palmer, N. B., J. Salcedo, et       Questionnaire         HIV/AIDS Treatment               107           triply diagnosed individuals (HIV,     methadone          all HAART era       3 months retrospective
al. (2003). AIDS Patient               study                Adherence Health                             mental health disorders, substance     treatment +                                questionnaire
Care STDS 17(12): 635-44.                              Outcomes and Cost Study,                          use) on methadone treatment, who         HIV care
                                                       from the Bronx arm of the                              volunteered for the study
                                                       study: Dept Psychiatry and
                                                          Behavioral Sciences,
                                                           Montefiore Medical
                                                         Center/Albert Einstein
                                                          College of Medicine,
                                                               Bronx, NY
Parry, M. F., J. Stewart, et al.    Prospective and         US metropolitan                                                                      Quality                1998                    2001
(2004). AIDS Care 16(6):             retrospective       community of 150,000                                                                  Improvement
690-9.                             chart data review             people                                                                        Collaborative

Paterson, D. L., S. Swindells,      Prospective,        HIV clinics in a Veterans          99            HIV patients prescribed a PI who         PI-ART         from August 1997 to     median of 6 months
et al. (2000). Ann Intern           observational      Affairs medical center and                       neither used a medication organizer                          March 1999         follow-up (range, 3 to
Med 133(1): 21-30.                      study          a university medical center                      nor received their medications in an                                                 15 months)
                                                                                                         observed setting (such as a jail or
                                                                                                                   nursing home)
Phillips, A. N., S. Staszewski,        Inception          Three cohort (Swiss,            3,226          ART-naive HIV patients initiating     ART initiation        from 1996            until 2000 (median
et al. (2001). JAMA                  longitudinal        Frankfurt, EuroSIDA)                             ART (> 3 drugs, 1996 onwards)                                                follow-up of 119 weeks)
286(20): 2560-7.                     cohort study       studies of patients cared
                                                          for in HIV clinics in
Piketty, C., E. Race, et al.        Open trial of a      HIV hospitals in Paris,           32           SQV and NNRTI naïve patients with        RTV/SQV +             ~ 1999          24, 48 week follow-up
(2000). AIDS 14(5): 626-8.            five-drug                  France                                   advanced HIV with VL > 5,000          efavirenz + 2
                                     combination                                                                                               recycled NRTIs
Piketty, C., L. Weiss, et al.        Prospective       HIV outpatient clinics of     42 discordants     PI-naïve, HAART-naïve HIV patients      first-line PI-      from April to            30 months
(2001). J Infect Dis 183(9):            cohort         Hospital Broussais, Paris,    identified at 12                                             HAART             October 1996
1328-35.                                                        France                months, from
Pilcher, C. D., W. C. Miller,       Retrospective      UNC Hospitals Infectious            168            patients VL < 400 on stable ART       stable ART       from September 1995    13 month follow-up
et al. (1999). AIDS 13(11):         cohort study        Diseases Clinic, Chapel                                                                                                         (29,576 person-days
1337-42.                                                       Hill, NC                                                                                                                  overall) between 1
                                                                                                                                                                                        October 1996 and 31
                                                                                                                                                                                          November 1997
Polis, M. A., I. A. Sidorov, et      Prospective            (Laboratory of                 124                      PI-naive pts                  initial PI         from ~ 2000         12 week follow-up
al. (2001). Lancet                   cohort study         Immunoregulation,
358(9295): 1760-5.                                       National Institute of
                                                        Allergy and Infectious
                                                       Diseases, NIH, Bethesda,

                                                                                                   - 60 -
                                                                                     Viral Load Report

Powderly, W. G., M. S. Saag,     Retrospective       multicenter, randomized,          196          Patients treated with nelfinavir; All     multicenter,       from ~ 1996       48 week follow-up
et al. (1999). AIDS 13(14):        analysis           double-blind Phase III                         subjects participated voluntarily        randomized,
1873-80.                                            study of combination anti-                                                                double-blind
                                                         HIV therapy with                                                                   Phase III study
                                                    nelfinavir, zidovudine, and                                                             of combination
                                                            lamivudine                                                                           anti-HIV
                                                                                                                                              therapy with
                                                                                                                                            zidovudine, and
Price, D. A., G. Scullard, et     Prospective,       All patients were drawn           53          subjects with chronic HIV-1 infection    clinic treatment     from ~ 1999      monitored intensively
al. (2003). J Virol 77(10):     single-blinded,       from the Jefferiss Wing                      who had been treated with continuous                                            over a period of 19
6041-9.                         nonrandomized            Clinic at St Mary’s                                       ART                                                                   months
                                     cohort            Hospital, London, UK
Raboud, J. M., J. S.             Randomized          the INCAS international           104         patients with CD4+ T cell counts of         randomized        July 20 1994      52 weeks follow-up
Montaner, et al. (1998).          control trial      multi-center randomized                       between 200 and 600 x 10(6) cells/l           trial with                       (INCAS study ended
AIDS 12(13): 1619-24.                                        trial study                             who were naive to antiretroviral       combinations of                           July 31 1996)
                                                                                                   therapy and AIDS-free at enrolment          zidovudine,
                                                                                                                                             nevirapine and
Raguin, G., G. Chene, et al.     Prospective,         multicentre Puzzle 1-         40 patients     patients with CD4 < 500 and VL >          First 2 weeks:       ~ 2000        2, 6, 26 week follow-up
(2002). Abstract #H-1078,        randomized,           ANRS104 study                    were         10,000 after at least 2 PIs and 1       randomized to
42nd Interscience Conference    open-label, trial                                 randomized, 37                  NNRTI                        either LPV/r
on Antimicrobial Agents and                                                            started                                                  (gr1), APV
Chemotherapy, San Diego.                                                             treatment                                                (1200mg/d) +
                                                                                                                                            RTV (200mg/d)
                                                                                                                                             (gr2), LPV/r +
                                                                                                                                            RTV (200mg/d)
                                                                                                                                                (gr3), APV
                                                                                                                                              (1200mg/d) +
                                                                                                                                            RTV (400mg/d)
                                                                                                                                               (gr4). From
                                                                                                                                             weeks 2 to 26,
                                                                                                                                                all patients
                                                                                                                                              received APV
                                                                                                                                               and LPV/r +
                                                                                                                                               200mg/d of
                                                                                                                                            RTV for gr 3, 4
Reijers, M. H., G. J.             Open-label        Dept Internal Medicine,            62          HIV patients with CD4 count > 200,        initial HAART     from March 1997     36 week follow-up;
Weverling, et al. (1998).        randomized         National AIDS Therapy                               VL > 1000, ART naïve                                                       ended April 6 1998
Ned Tijdschr Geneeskd           controlled trial     Evaluation Centre, U.                                                                                                            prematurely
142(40): 2230.                                      Amsterdam, Netherlands

                                                                                              - 61 -
                                                                                         Viral Load Report

Roge, B. T., T. S. Barfod, et         Retrospective       Danish multi-center drug        56        primary virological failures (viral load   Randomized to      from 1998 to January     median of 90 weeks
al. (2004). HIV Med 5(5):              (post-hoc)           trial study (n=293)                               > 400 copies/mL)                 two nucleoside            2001               follow-up, until
344-51.                                                                                                                                             reverse                                 December 2001
                                                                                                                                                 (n=115), two
                                                                                                                                                  (n=118), or
                                                                                                                                                 e (ASD-arm)
Sabin, C. A., F. C. Lampe, et             Audit          large treatment clinic in the   1,264                     clinic patients             clinic treatment     all patients under         N/A: audit
al. (2003). HIV Med 4(2):                                UK: Royal Free Centre for                                                                                 follow-up between
87-93.                                                     HIV Medicine, London,                                                                                  1st January 2000 and
                                                                     UK                                                                                             1st January 2001
                                                                                                                                                                      were included
Schrooten, W., E. Florence, et        Longitudinal         HIV outpatient clinic,        225         Initiating HAART in a clinical setting     HAART in a                 1997             5 years follow-up
al. (2004). Int J STD AIDS               study              Institute of Tropical                    (with an interim follow-up at 2 years)    clinical setting
15(8): 523-8.                                               Medicine, Antwerp

Sension, M. G., C. Farthing,           Open-label,        Two centers in the USA.         42         HIV+ antiretroviral-naive individuals       treated with                             48 weeks intervention
et al. (2001). AIDS Patient         single-arm, phase       Primary center: HIV                       with VL > 20,000 and CD4 count             SQV-SGC,
Care STDS 15(3): 129-36.             3b study (part of    Clinical Research, North                                  >100                       1200 mg three
                                          phase 3        Broward Hospital District,                                                            times per day;
                                     development) to        Fort Lauderdale, FL                                                                ZDV, 300 mg;
                                       evaluate the                                                                                            and 3TC, 150
                                       efficacy and                                                                                            mg each twice
                                         safety of                                                                                                  per day
                                      + zidovudine
                                         (ZDV) +
                                    lamivudine (3TC)
Servais, J., J. C. Schmit, et al.           N/A          Luxembourg HIV cohort,           73              PI-naïve, previously treated and       PI-HAART                                36 months of treatment
(2000). HIV Clin Trials                                   retrospectively analysis                         untreated patients, initiated on
1(2): 17-24.                                                                                                          HAART

Shapiro, M. F., S. C. Morton,       Cohort study that      HIV Cost and Services         2,864          Multistage probability sample          clinic treatment    baseline: between     first follow-up: between
et al. (1999). JAMA                  consisted of 3        Utilization Consortium                   representing the 231,400 persons > 18                          January 1996 and         December 1996 and
281(24): 2305-15.                      interviews                                                   years old, with HIV receiving medical                             April 1997           July 1997; the second
                                                                                                    care in continental US in early 1996 in                                                 follow-up: between
                                                                                                    facilities other than EDs, the military,                                                 August 1997 and
                                                                                                                   or prisons                                                                   January 1998

                                                                                                 - 62 -
                                                                                        Viral Load Report

Sklar, P. A., D. J. Ward, et al.   Non-randomized        9 HIV clinics in eight           448           Ambulatory HIV patients with > 2        clinic treatment    between 1 January        and 30 June 2000
(2002). AIDS 16(15): 2035-         dynamic cohort        cities, USA (The HIV                               consecutive VLs < 50                   (HAART)                1997               (median 485 days
41.                                     study           Outpatient Study (HOPS))                                                                                                                follow-up)

Smith, C. J., S. Staszewski, et      Longitudinal           the J.W. Goethe               656               antiretroviral-naive patients          starting             from 1999           24 weeks follow-up
al. (2004). J Acquir Immune          cohort study        University, Chelsea and                                                                   HAART
Defic Syndr 37(1): 1155-                                 Westminster, and Royal
1159.                                                        Free Hospitals

Sterling, T. R., C. M. Lyles,        Nested case-        Division of Infectious          Rapid           272 HIV-1 seroconverters among                             between February       between March 1989
et al. (1999). J Infect Dis          control study      Diseases, Johns Hopkins      progressors to    female and male HIV+ IDU patients                             1988 and March           and July 1996
180(3): 666-72.                                           University Schoolof           AIDS (24        age >18 years, injection of drugs at                              1989
                                                         Medicine, Baltimore,         patients) and    least once after 1977, without AIDS
                                                         Maryland 21287, USA         nonprogressors
                                                                                      (47 controls)
Sufka, S. A., G. Ferrari, et al.     Longitudinal       Duke University Medical            30         > 18 years old; HIV+; current therapy        HAART            from ~1999 (all              > 2 years
(2003). J Infect Dis 187(7):         cohort study       Center Infectious Diseases                    with PI-based HAART; documented                                 HAART-era)
1027-37.                                                          Clinic                              VL > 50,000 and/or CD4 nadir < 50,
                                                                                                        while not on antiretroviral therapy
Sulkowski, M. S. and D. L.          Review paper                  N/A                     N/A                          N/A                           N/A                   N/A                     N/A
Thomas (2003). Ann Intern
Med 138(3): 197-207.

Torti, C., E. Quiros-Roldan,       Prospective study        Multicenter, Italy            159             heavily pretreated HIV-positive          HAART                from 1999           48 weeks follow-up
et al. (2003). HIV Med 4(3):           of real- vs.                                                                   patients
263-70.                            virtual-phenotype

Tuboi, S., E. Sprinz, et al.         Longitudinal        Hospital de Clinicas de          454                Treatment-naïve patients           initial HAART       between January         6 month follow-up
(2005). Abstract #600, "             cohort study        Porto Alegre, Universite                                                                                  1996 and November
12th Conference on                                      Federal do Rio Grande do                                                                                          2003
Retroviruses and                                                Sul, Brazil
Opportunistic Infections,

Valdez, H., M. M. Lederman,         Retrospective          University-affiliated          310                      clinic patients               regimens that     patients treated from    to December 1997
et al. (1999). Arch Intern          observational       human immunodeficiency                                                                       included           June 1995
Med 159(15): 1771-6.                    study              virus (HIV) clinic in                                                                     protease
                                                             Cleveland, Ohio                                                                        inhibitors
Walmsley, S., B. Bernstein, et       Randomized         93 centers in 13 countries        653         HIV patients (> 12 years old) with VL         lopinavir-     assigned to treatment   24, 48 week follow-up
al. (2002). N Engl J Med           double-blind trial    in North America, South                            > 400 who had not received           ritonavir with     between March 30,
346(26): 2039-46.                                       America, Europe, Africa,                      antiretroviral therapy for more than 14       nelfinavir     1999, and September
                                                               and Australia                                            days                       placebo or            30, 1999
                                                                                                                                                 nelfinavir with
                                                                                                                                                placebo + open-
                                                                                                                                                label stavudine
                                                                                                                                                and lamivudine
                                                                                                                                                  (all patients)

                                                                                                 - 63 -
                                                                                 Viral Load Report

Yamashita, T. E., J. P. Phair,    Cohort study      the Multicenter AIDS          397                            men                       HAART        initiated HAART      3-33 months
et al. (2001). AIDS 15(6):                          Cohort Study (MACS)                                                                                 between October
735-46.                                                                                                                                                  1995 and March
Yazdanpanah, Y., D. Sissoko,      Review and              drug trials         nearly 6,000       Participants had no previous exposure     PI-based            1996             2000
et al. (2004). BMJ                meta-analysis                                                    to protease inhibitors or NNRTIs      regimens vs.
328(7434): 249.                                                                                                                          NNRTI-based
Yeni, P. G., S. M. Hammer, et         N/A           Recommendations of the        N/A                            N/A                         N/A              N/A               N/A
al. (2004). JAMA 292(2):                              International AIDS
251-65.                                               Society-USA Panel

Yun, L. W., M. Maravi, et al.        Pre/post        urban HIV health care   data from chart            1,713 (57% depressed)             clinic ART    patients seen from    until 2001
(2005). J Acquir Immune            intervention             setting            review and                                                                      1997
Defic Syndr 38(4): 432-8.        study (data from                            administrative
                                 chart review and                             and pharmacy
                                  administrative                                   files
                                  and pharmacy

                                                                                             - 64 -
                                 Viral Load Report

1.    Mezzaroma I, Carlesimo M, Pinter E, et al. Clinical and immunologic response
      without decrease in virus load in patients with AIDS after 24 months of highly
      active antiretroviral therapy. Clin Infect Dis. Dec 1999;29(6):1423-1430.
2.    Torti C, Quiros-Roldan E, Scudeller L, et al. Characterization of viro-
      immunological responses in a closely followed cohort of heavily pretreated
      patients: evidence from the GenPheRex Study. HIV Med. Jul 2003;4(3):263-270.
3.    Sufka SA, Ferrari G, Gryszowka VE, et al. Prolonged CD4+ cell/virus load
      discordance during treatment with protease inhibitor-based highly active
      antiretroviral therapy: immune response and viral control. J Infect Dis. Apr 1
4.    Mellors JW, Rinaldo CR, Jr., Gupta P, White RM, Todd JA, Kingsley LA.
      Prognosis in HIV-1 infection predicted by the quantity of virus in plasma.
      Science. May 24 1996;272(5265):1167-1170.
5.    Institute of Medicine. Medicare: A strategy for quality assurance. Vol 1.
      Washington, D.C.: National Academy Press; 1990.
6.    Landon BE, Wilson IB, McInnes K, et al. Effects of a quality improvement
      collaborative on the outcome of care of patients with HIV infection: the EQHIV
      study. Ann Intern Med. Jun 1 2004;140(11):887-896.
7.    Parry MF, Stewart J, Wright P, McLeod GX. Collaborative management of HIV
      infection in the community: an effort to improve the quality of HIV care. AIDS
      Care. Aug 2004;16(6):690-699.
8.    Holodniy M, Mole L, Winters M, Merigan TC. Diurnal and short-term stability of
      HIV virus load as measured by gene amplification. J Acquir Immune Defic Syndr.
      Apr 1994;7(4):363-368.
9.    Fuller JD, Craven DE, Steger KA, Cox N, Heeren TC, Chernoff D. Influenza
      vaccination of human immunodeficiency virus (HIV)-infected adults: impact on
      plasma levels of HIV type 1 RNA and determinants of antibody response. Clin
      Infect Dis. Mar 1999;28(3):541-547.
10.   New York State Department of Health AIDS Institute. Criteria for the medical
      care of adults with HIV infection: Antiretroviral therapy January 2005 2005.
11.   United States Department of Health and Human Services. Guidelines for the use
      of antiretroviral agents in HIV-1 infected adults and adolescents: Panel on
      Clinical Practices for Treatment of HIV Infection Panel; October 29, 2004 2004.
12.   Durant J, Clevenbergh P, Halfon P, et al. Drug-resistance genotyping in HIV-1
      therapy: the VIRADAPT randomised controlled trial. Lancet. Jun 26
13.   Dybul M, Daucher M, Jensen MA, et al. Genetic characterization of rebounding
      human immunodeficiency virus type 1 in plasma during multiple interruptions of
      highly active antiretroviral therapy. J Virol. Mar 2003;77(5):3229-3237.
14.   Grabar S, Le Moing V, Goujard C, et al. Clinical outcome of patients with HIV-1
      infection according to immunologic and virologic response after 6 months of
      highly active antiretroviral therapy. Ann Intern Med. Sep 19 2000;133(6):401-

                                       - 65 -
                                  Viral Load Report

15.   Ledergerber B, Egger M, Opravil M, et al. Clinical progression and virological
      failure on highly active antiretroviral therapy in HIV-1 patients: a prospective
      cohort study. Swiss HIV Cohort Study. Lancet. Mar 13 1999;353(9156):863-868.
16.   Greub G, Cozzi-Lepri A, Ledergerber B, et al. Intermittent and sustained low-
      level HIV viral rebound in patients receiving potent antiretroviral therapy. Aids.
      Sep 27 2002;16(14):1967-1969.
17.   Moore R, Keruly J, Gebo K, Lucas G. Abstract #558, Improvement in virologic,
      immunologic, and clinical outcomes in clinical practice from 1996 to 2002. 11th
      Conference on Retroviruses and Opportunistic Infections, San Francisco. 2004.
18.   Lohse N, Obel N, Kronborg G, et al. Paper #594, Risk and predictors of triple-
      class antiretroviral drug failure in Danish HIV-infected individuals. 12th
      Conference on Retroviruses and Opportunistic Infections, Boston. 2005.
19.   Tuboi S, Sprinz E, Harrison L, Albernaz R, Schechter M. Abstract #600,
      Predictors of virological failure at 6 months of therapy in HIV-1-infected patients
      starting HAART in Porto Alegre, Brazil. 12th Conference on Retroviruses and
      Opportunistic Infections, Boston. 2005.
20.   Hirsch M, Steigbigel R, Staszewski S, et al. A randomized, controlled trial of
      indinavir, zidovudine, and lamivudine in adults with advanced human
      immunodeficiency virus type 1 infection and prior antiretroviral therapy. J Infect
      Dis. Sep 1999;180(3):659-665.
21.   Hirsch MS, Brun-Vezinet F, D'Aquila RT, et al. Antiretroviral drug resistance
      testing in adult HIV-1 infection: recommendations of an International AIDS
      Society-USA Panel. Jama. May 10 2000;283(18):2417-2426.
22.   Yeni PG, Hammer SM, Hirsch MS, et al. Treatment for adult HIV infection: 2004
      recommendations of the International AIDS Society-USA Panel. Jama. Jul 14
23.   Lawrence J, Mayers DL, Hullsiek KH, et al. Structured treatment interruption in
      patients with multidrug-resistant human immunodeficiency virus. N Engl J Med.
      Aug 28 2003;349(9):837-846.
24.   Miller V, Phillips AN, Clotet B, et al. Association of virus load, CD4 cell count,
      and treatment with clinical progression in human immunodeficiency virus-
      infected patients with very low CD4 cell counts. J Infect Dis. Jul 15
25.   Katlama C, Dominiguez S, Duvivier C. Abstract #5887, Benefits of treatment
      interruption in patients with multiple therapy failures, CD4 cells <200/mm3 and
      HIV RNA > 50,000 cp/mL (GIGHAART ANRS 097). XIV International AIDS
      Conference, Barcelona, Spain. 2002.
26.   Jacobson LP, Li R, Phair J, et al. Evaluation of the effectiveness of highly active
      antiretroviral therapy in persons with human immunodeficiency virus using
      biomarker-based equivalence of disease progression. Am J Epidemiol. Apr 15
27.   Abgrall S, Duval X, Joly V, Descamps D, Matheron S, Costagliola D. Clinical
      and immunologic outcome in patients with human immunodeficiency virus
      infection, according to virologic efficacy in the year after virus undetectability,
      during antiretroviral therapy. Clin Infect Dis. Dec 1 2003;37(11):1517-1526.

                                         - 66 -
                                  Viral Load Report

28.   Nasta P, Castelnuovo F, Paraninfo G, et al. Abstract #605, To maintain or to
      switch HAART in heavily pre-treated patients with low-level viremia. 12th
      Conference on Retroviruses and Opportunistic Infections, Boston. 2005.
29.   Ledergerber B, Lundgren JD, Walker AS, et al. Predictors of trend in CD4-
      positive T-cell count and mortality among HIV-1-infected individuals with
      virological failure to all three antiretroviral-drug classes. Lancet. Jul 3-9
30.   New York State Department of Health AIDS Institute. Quality of care indicators:
      Adult indicators.
      care/qoc-ny-inidicator/a-indicators.htm. Accessed 9/2/2005.
31.   Bartlett J, Fath M, DeMasi R, Quinn J, Hermes A, Rousseau F. Abstract #586, An
      updated meta-analysis of triple combination therapy in antiretroviral-naive HIV-
      infected adults. 12th Conference on Retroviruses and Opportunistic Infections,
      Boston. 2005.
32.   Bartlett JA, DeMasi R, Quinn J, Moxham C, Rousseau F. Overview of the
      effectiveness of triple combination therapy in antiretroviral-naive HIV-1 infected
      adults. Aids. Jul 27 2001;15(11):1369-1377.
33.   Yazdanpanah Y, Sissoko D, Egger M, Mouton Y, Zwahlen M, Chene G. Clinical
      efficacy of antiretroviral combination therapy based on protease inhibitors or non-
      nucleoside analogue reverse transcriptase inhibitors: indirect comparison of
      controlled trials. Bmj. Jan 31 2004;328(7434):249.
34.   Havlir DV, Bassett R, Levitan D, et al. Prevalence and predictive value of
      intermittent viremia with combination hiv therapy. Jama. Jul 11 2001;286(2):171-
35.   Nettles RE, Kieffer TL, Kwon P, et al. Intermittent HIV-1 viremia (Blips) and
      drug resistance in patients receiving HAART. Jama. Feb 16 2005;293(7):817-
36.   Bahrani A, Ramaswamy R, Oldfield EC, 3rd. Effects of virologic rebound on
      CD4 cell counts. Clin Infect Dis. Apr 15 2001;32(8):1231-1232.
37.   Deeks SG, Barbour JD, Grant RM, Martin JN. Duration and predictors of CD4 T-
      cell gains in patients who continue combination therapy despite detectable plasma
      viremia. Aids. Jan 25 2002;16(2):201-207.
38.   Dybul M. Structured Treatment Interruption: Approaches and Risks. Curr Infect
      Dis Rep. Apr 2002;4(2):175-180.
39.   Deeks SG, Hoh R, Grant RM, et al. CD4+ T cell kinetics and activation in human
      immunodeficiency virus-infected patients who remain viremic despite long-term
      treatment with protease inhibitor-based therapy. J Infect Dis. Feb 1
40.   Piketty C, Weiss L, Thomas F, Mohamed AS, Belec L, Kazatchkine MD. Long-
      term clinical outcome of human immunodeficiency virus-infected patients with
      discordant immunologic and virologic responses to a protease inhibitor-containing
      regimen. J Infect Dis. May 1 2001;183(9):1328-1335.
41.   Paterson DL, Swindells S, Mohr J, et al. Adherence to protease inhibitor therapy
      and outcomes in patients with HIV infection. Ann Intern Med. Jul 4

                                        - 67 -
                                  Viral Load Report

42.   Haubrich RH, Little SJ, Currier JS, et al. The value of patient-reported adherence
      to antiretroviral therapy in predicting virologic and immunologic response.
      California Collaborative Treatment Group. Aids. Jun 18 1999;13(9):1099-1107.
43.   Low-Beer S, Yip B, O'Shaughnessy MV, Hogg RS, Montaner JS. Adherence to
      triple therapy and viral load response. J Acquir Immune Defic Syndr. Apr 1
44.   Arnsten JH, Demas PA, Grant RW, et al. Impact of active drug use on
      antiretroviral therapy adherence and viral suppression in HIV-infected drug users.
      J Gen Intern Med. May 2002;17(5):377-381.
45.   Lucas GM, Gebo KA, Chaisson RE, Moore RD. Longitudinal assessment of the
      effects of drug and alcohol abuse on HIV-1 treatment outcomes in an urban clinic.
      Aids. Mar 29 2002;16(5):767-774.
46.   Bangsberg DR, Hecht FM, Charlebois ED, et al. Adherence to protease inhibitors,
      HIV-1 viral load, and development of drug resistance in an indigent population.
      Aids. Mar 10 2000;14(4):357-366.
47.   Yun LW, Maravi M, Kobayashi JS, Barton PL, Davidson AJ. Antidepressant
      treatment improves adherence to antiretroviral therapy among depressed HIV-
      infected patients. J Acquir Immune Defic Syndr. Apr 1 2005;38(4):432-438.
48.   Palmer NB, Salcedo J, Miller AL, Winiarski M, Arno P. Psychiatric and social
      barriers to HIV medication adherence in a triply diagnosed methadone population.
      AIDS Patient Care STDS. Dec 2003;17(12):635-644.
49.   Sabin CA, Lampe FC, Chaloner C, et al. An audit of antiretroviral treatment use
      in HIV-infected patients in a London clinic: the limitations of observational
      databases when auditing antiretroviral treatment use. HIV Med. Apr 2003;4(2):87-
50.   Deeks SG, Martin JN, Sinclair E, et al. Strong cell-mediated immune responses
      are associated with the maintenance of low-level viremia in antiretroviral-treated
      individuals with drug-resistant human immunodeficiency virus type 1. J Infect
      Dis. Jan 15 2004;189(2):312-321.
51.   Mijch AM, Hoy J, Watson K, Dunne A, Crowe S, Wesselingh SL. Does plasma
      HIV RNA predict outcome in a cohort of treated HIV-infected individuals
      followed over 3 years? J Clin Virol. Oct 2001;22(3):271-278.
52.   Sherer R.
53.   Das-Douglas Mea. shared with permission; 2005.

                                        - 68 -

To top