Preoperative Characteristics Predicting Intraoperative Hypotension

Document Sample
Preoperative Characteristics Predicting Intraoperative Hypotension Powered By Docstoc
					Preoperative Characteristics Predicting
Intraoperative Hypotension and Hypertension
Among Hypertensives and Diabetics
Undergoing Noncardiac Surgery


We prospectively studied patients with hypertension and diabetes                    From the Clinical Epidemiology Unit, the Departments of
undergoing elective noncardiac surgery with general anesthesia                    Medicine, Anesthesiology, and Surgery, Cornell University
to test the hypothesis that patients at high risk for prognostically                                  Medical College, New York, New York
significant intraoperative hemodynamic instability could be
identified by their preoperative characteristics. Specifically we
hypothesized that patients with a low functional capacity, de-
creased plasma volume, or significant cardiac comorbidity would                cardiac complications after operation. Patients with cardiac dis-
be at high risk for intraoperative hypotension and those with a                ease, especially diabetics, and those with negative fluid balances
history of severe hypertension would be at risk for intraoperative             also had increased complications. Preoperative characteristics
hypertension. Patients who had a preoperative mean arterial                    influence the susceptibility to intraoperative hypotension and
pressure (MAP) .110, a walking distance of less than 400 m,                    hypertension, which are related to postoperative complications.
or a plasma volume less than 3000 cc were at increased risk of
intraoperative hypotension (i.e., more than 1 hour of .20 mmHg
decreases in the MAP). Hypotension was also more common                             x r HILE MOST STUDIES have demonstrated that
among patients having intra-abdominal or vascular surgery, and                              intraoperative hypotension increases the risk
among those who had operations longer than 2 hours. Patients                                of postoperative cardiacl'5 and renal6'7 com-
older than 70 years or with a decreased plasma volume were at                  plications, there is less agreement about the impact of
increased risk of having more than 15 minutes of intraoperative
elevations of .20 mmHg over the preoperative MAP in com-                       other intraoperative hemodynamic patterns (e.g., hyper-
bination with intraoperative hypotension; this was also more                   tension alone or hypertension and hypotension).8'10 We
common when surgery lasted more than 2 hours. Patients who                     have shown that patients with hypertension or diabetes
had intraoperative hypotension tended to have an immediate de-                 whose intraoperative mean arterial pressure (MAP) de-
crease in MAP at the onset of anesthesia and were often pur-                   creased .20 mmHg below their preoperative MAP for
posefully maintained at MAPs less than their usual level during
surgery with fentanyl and neuromuscular blocking agents. Pa-                   more than 1 hour had increased cardiac and renal com-
tients who had intraoperative hyper/hypotension tended to have                 plications.' ''- In addition patients who experienced both
repeated elevations in MAP above their preoperative levels dur-                15 minutes or more of elevated MAP (.20 mm above
ing the course of surgery, and such elevations precipitated in-                their usual MAP) and less than 1 hour of .20 mmHg
terventions with neuromuscular blocking agents and/or fentanyl.                decreases in their MAP hour had increased complica-
Neither pattern was more common among patients who developed
net intraoperative negative fluid balances. Both hypotension and               tions. "'-4 Together these two patterns, which will be called
hyper/hypotension were associated with increased renal and                     intraoperative hypotension and intraoperative hyper/hy-
                                                                               potension, respectively, had a 46% sensitivity rate and a
                                                                               70% specificity rate as predictors of postoperative com-
   Dr. Charlson was a Henry J. Kaiser Family Foundation Faculty Scholar        plications.
in General Internal Medicine. Dr. Ales was a George Morris Piersol
American College of Physicians Teaching and Research Scholar.                     There has been no reliable method of identifying pa-
   Supported by a grant from the National Heart, Lung and Blood In-            tients at risk for intraoperative hypotension, hypertension,
stitute, R23HL276 13, in part by the Robert Wood Johnson Foundation            or both. Furthermore it has not been possible to distin-
Program to Improve Patient Functional Status, and PHS Research Grant           guish patients specifically at risk for intraoperative hy-
RR00047 from the Division of Research Resources, National Institute            potension as opposed to hypertension, which clearly
of Health.                                                                     would have importance for patient management, because
   Address reprint requests to Dr. Mary E. Charlson, Cornell Medical
College, 515 E. 71st Street, New York, NY 10021.                               strategies to prevent intraoperative hypotension may differ
   Accepted for publication October 30, 1989.                                  substantially from those designed to prevent hypertension.

   The explicit prior hypothesis of this study was that pa-     to the Rose criteria.'5 Patients with nocturnal or rest an-
tients at high risk for significant episodes ofintraoperative   gina, or with a crescendo pattern, were designated as hav-
hypotension or hypertension could be identified by their        ing unstable angina. '5 Myocardial infarction included pa-
preoperative characteristics. Specifically patients with a      tients who had been hospitalized for chest pain and de-
low functional capacity, decreased plasma volume, or sig-       veloped either new Q waves in at least two leads that were
nificant cardiac comorbidity were hypothesized to be at         0.04 seconds in duration and 1 mm in depth, new ST
high risk for intraoperative hypotension and those with a       segment depression of 1 mm or more, or new T wave
history of severe hypertension were hypothesized to be at       inversion that persisted for 7 days, with elevation of cre-
risk for intraoperative hypertension. The purpose of this       atine kinase (CK) or CK-MB isoenzyme. Congestive heart
analysis was to test this hypothesis. The intraoperative        failure included patients with a definite history of pul-
management of patients with hypotension, hypertension/          monary edema, paroxysmal nocturnal dyspnea, or dys-
hypotension, or those without either of these high-risk         pnea on exertion, '5 not in the setting of an acute myo-
patterns was also examined to develop a qualitative con-        cardial infarction, arrhythmia, sepsis, and so on who re-
text for the evaluation of these intraoperative hemody-         sponded to treatment and who required continued
namic patterns.                                                 pharmacologic therapy. Patients with documented val-
                                                                vular disease of any type (aortic, mitral, prosthetic valve,
                         Methods                                and so on), except mitral valve prolapse, were defined as
Assembly of Population                                          having valvular heart disease. Aortic valvular disease in-
                                                                cluded two patients with previously documented hemo-
   All patients who had essential hypertension or diabetes
who were undergoing elective general surgery were eligible      dynamically insignificant aortic stenosis, 14 patients with
                                                                grade ii/vi or louder systolic ejection murmur radiating
for enrollment. The criteria for hypertension were (1) for      to the carotids, and two patients with aortic insufficiency.
patients younger than 30 years, systolic .150 mmHg or           Mitral valvular disease included two patients with pre-
diastolic .90 mmHg; (2) for patients 30 years or older,         viously documented mitral regurgitation and seven pa-
systolic .160 mmHg or diastolic .95 mmHg; or (3)                tients with grade ii/vi or louder holosystolic murmur ra-
treatment with any medication explicitly used to reduce         diating to the axilla. Patients were also classified according
blood pressure (i.e., patients treated with diuretics for       to Goldman's cardiac risk classification.'6 Patients were
congestive heart failure alone were not considered hyper-       classified as having severe hypertension if they had a his-
tensive). The criteria for diabetes were (1) treatment with     tory of diastolic blood pressure more than 120 mmHg or
insulin or oral hypoglycemic agents, including treatment        had been hospitalized for control of blood pressure.
started during hospitalization, and (2) elevated fasting        Chronic pulmonary disease included patients with chronic
glucose on more than one occasion (plasma .140 mg/              bronchitis'5 and patients with a definite history of asthma
dL; whole blood .120 mg/dL).                                    that occurred without upper respiratory tract infections
   Between July 1982 and September 1985, 278 patients           and that required pharmacologic treatment.
were enrolled in our study. To document the proportion             The patient's right and left arm sitting blood pressure
of eligible patients who were entered into the study, the       and pulse, as well as the recumbent and standing pulse
charts of all 1398 patients undergoing several types of         and blood pressure, were obtained. Blood pressure was
operation (vascular surgery or cholecystectomy) during          recorded using a standard mercury manometer17 and Ko-
the same secular interval were reviewed; these operations       rotkoff phase V was recorded as the diastolic pressure.
accounted for 40% of the surgery performed in the pro-          Blood pressures were obtained every 4 hours before sur-
spective study. In these groups 26% of eligible hyperten-       gery by the nursing staff. These systolic and diastolic blood
sives and 34% of eligible diabetics had been entered into       pressures were recorded to delineate the patient's usual
the prospective study.'3                                        preoperative MAP. A 12-minute walking distance test was
Preoperative Evaluation
                                                                performed with encouragement whenever possible before
                                                                operation.'8-20 Basic preoperative laboratory values were
   The protocol was reviewed and approved by the Insti-         recorded, including hemoglobin, electrolytes, urinanalysis,
tutional Human Rights Committee. Informed consent was           creatinine, creatine kinase, electrocardiogram, and chest
obtained on all patients. Before operation basic demo-          x-ray. Cardiomegaly was present when the heart was la-
graphic and clinical data were recorded and a physical          beled by the radiologist as definitely enlarged. An abnor-
examination focused on abnormalities in the pulmonary,          mal electrocardiogram included bundle branch block, def-
cardiac, neurologic, and vascular systems. The history of       inite or probable evidence of an old infarction, or definite
comorbid conditions was obtained using standardized             left ventricular hypertrophy. Plasma volume was deter-
questions and criteria.                                         mined whenever possible using Evans blue2' 12 hours be-
   Angina was defined as definite or probable according         fore operation; results were available for 50 patients.
68                                               CHARLSON AND OTHERS                                        Ann. Surg. * July 1990

Intraoperative Monitoring                                       course of any patient who developed dyspnea, palpita-
                                                                tions, chest pain, signs of congestive failure, electrocar-
   On the day of surgery, a research assistant, blinded to      diographic changes, or increased serum creatinine level
the specific hypotheses of the study, recorded pulse and        was reviewed by two physicians who were blinded to the
blood pressure during induction and intubation and every        patient's preoperative status and intraoperative course.
5 minutes thereafter, using either an intra-arterial catheter   These physicians determined whether the patients who
(at the discretion of the anesthesiologist), a mercury ma-      were suspected of having myocardial ischemia, congestive
nometer, or an indirect automatic device. In 52 patients        heart failure, or postoperative renal impairment met the
intra-arterial measurements were made using a 20-gauge          diagnostic criteria.23
teflon catheter in the radial artery and connected to a
Marquette 7000 dual pressure monitor (Marquette Elec-           Definition of Outcomes
tronics, Milwaukee, WI) calibrated before each use (ac-            Life-threatening events. A total of eight patients had
curacy of ± 1 mmHg). In 76 patients the Omega 1000              cardiac death, acute pulmonary edema, cardiac arrest, or
(Vivo Research Labs, Tulsa, OK), an indirect oscillometric      myocardial infarction. There were two cardiac deaths, of
blood pressure recorder was used (accuracy of ±2 mmHg);         which one occurred after intraoperative ventricular
the large adult cuff (15 X 33 cm) was used if the arm           tachycardia, and one after an anginal episode. One cardiac
circumference was 30 cm or larger. The oscillometric de-        arrest that did not lead to cardiac death occurred in the
vice has been shown to produce reproducible and valid           setting of pulmonary edema (pulmonary capillary wedge
measurements of blood pressure while used in the oper-          pressure more than 20 cm and a typical roentgenologic
ating room. 22 In 150 patients the anesthesiologist used a      picture) and symptomatic myocardial ischemia. Two ar-
standard mercury manometer instead of the Omega 1000.           rests occurred in the setting of a myocardial infarction:
In these cases the research assistant obtained and recorded     one with ST-T depression .48 hours with 12% MB (def-
the blood pressures according to the procedures described       inite) and the other with ST elevation less than 48 hours
previously. To standardize the definition of the MAP with       with 6% MB (possible). In addition there was one definite
the different techniques, MAPs were calculated as the di-       myocardial infarction without an arrest or death and two
astolic pressure + 1/3 (systolic - diastolic.) The primary      patients with probable myocardial infarction: one patient
analysis of intraoperative blood pressures was based on         had new T inversion of more than 48 hours with 25%
the differences between the intraoperative MAPs and the         MB and another had new T inversions of more than 48
preoperative MAPs, after adjusting for differences between      hours with 8% MB.
arms.                                                               Other cardiac complications. There was one patient with
   The research assistant recorded all pharmacologic            a possible myocardial infarction and 18 patients with def-
agents used during each 5-minute interval of the opera-         inite ischemia with ECG changes. Twelve of these patients
tion. The use of anesthetic agents and adjuvants was re-        had elevated CK-MB, and the seven patients without el-
corded. Major intraoperative events, such as skin incision,     evated CK-MB had ST-T changes that persisted for more
reversal of anesthesia, aortic cross-clamping, drainage of      than 48 hours, In addition one patient had unstable angina
ascites, and changes in position, were also recorded.           without ECG changes and two patients had definite angina
   In 24 (8.6%) of the patients, the research assistant was
not present in the operating room to record events and
                                                                by the Rose criteria. Seven patients had postoperative
blood pressures. Because the purpose of this analysis is to
                                                                congestive heart failure with at least two of the following:
                                                                S3 gallop, typical roentgenologic picture, bibasilar rales
assess preoperative predictors of high-risk intraoperative      involving more than one third of the lung fields, or an
hemodynamic patterns, we have omitted these 24 patients         elevated central venous pressure (more than 18 cm of
(seven of whom had complications) because the intra-            water).
operative data was not recorded by an independent ob-               Renal complications. Fourteen patients had postoper-
server. Of these 254 patients, 47 patients had either local/    ative renal dysfunction, e.g., an increase in the serum cre-
regional or spinal/epidural anesthesia; 11 of these patients    atinine of .20% that persisted for 48 hours or more and
had complications. The rest of the paper will focus on the      that began in the first 3 days after surgery. In identifying
207 patients who had general anesthesia.                         patients whose serum creatinines remained persistently
Postoperative Follow-up                                          elevated at discharge from the hospital, an increase in the
  All patients were followed daily for 7 days after oper-        serum creatinine of .20% persisting for more than 48
ation or until death, discharge, or reoperation with clinical    hours had a true-positive rate of 93%.24
examinations, serum creatinine tests, electrocardiograms,           There were 48 complications in 44 patients; 4 patients
and creatine kinase isoenzyme tests. Patients who under-         had both postoperative renal dysfunction and congestive
went reoperation were considered censored in terms of            failure, and 4 had myocardial ischemia and congestive
follow-up at the time of reoperation. The postoperative          failure.
Vol. 212 * No. I     PREOPERATIVE PREDICTORS OF INTRAOPERATIVE HYPOTENSION AND HYPERTENSION                                                       69
Data Analysis                                                                 ening events (arrest, myocardial infarction, or pulmonary
                                                                              edema), 7 congestive failure, 17 myocardial ischemia, and
   Logistic regression was performed with the PROC                            9 renal dysfunction. Thirty-four per cent of patients with
LOGIST program in SAS25 to determine differences be-                          intraoperative hypotension, 23% of patients with intra-
tween those patients with and without a given high-risk                       operative hyper/hypotension, and 16% of those without
pattern according to the preoperative characteristics. An                     either high-risk pattern had complications.'1-3 According
effort was made to constrain the number of variables en-                      to previous papers, patients with cardiac disease, partic-
tered into the model so that one covariate was entered                        ularly diabetics with cardiac disease, patients who had
for each 5 to 10 outcome events.26 In the analysis of walk-                   either intraoperative hypotension or hyper/hypotension,
ing distance, patients who were unable to complete the                        and patients who received lower net intraoperative fluids
test were combined with those who were able to walk                           had increased rates of postoperative complications. The
more than 640 m because there were no significant dif-                        predictors of each specific complication varied.' 1-13 How-
ferences between the groups. The logistic regression re-                      ever, for the purposes ofthis paper, the complication rates
sults, while reported in different sections of the paper,                     are shown in Table 1 according to the presence or absence
were obtained from a single equation, with six variables                      ofcardiac disease and diabetes, intraoperative hypotension
entered for hypotension, and five for hyper/hypotension.                      or hyper/hypotension, and according to whether the pa-
Because plasma volume was assessed in only 50 of the                          tient had 1 or more hours during operation with a net
207 patients, the logistic regression was performed using                     negative fluid balance. Forty-nine per cent of the 19 di-
only those covariates that were significant predictors of                     abetics with cardiac disease had complications; in contrast
hypotension or hyper/hypotension in the 207 patients as                       the 52 diabetics without cardiac disease had only 18%
a whole.                                                                      complications. Patients with a high-risk pattern and at
                           Results                                            least 1 hour with a negative fluid balance had especially
                                                                              high complication rates.
   Of the total 207 patients, 41 (20%) had intraoperative
hypotension: a .20 mmHg decrease from their usual pre-
operative MAP that lasted for 1 hour or more. An addi-                         Preoperative Predictors of Patients with High-risk He-
tional 39 (18%) patients had intraoperative hyper/hypo-                        modynamic Patterns
tension: . 15 minutes of hypertension with a .20 mmHg
increase above their usual MAP in combination with a                             The purpose of the initial analysis was to test whether
.20 mmHg decrease in their MAP lasting less than 1                            a history of severe hypertension, decreased functional ca-
hour. The remaining 127 (62%) patients had neither high-                      pacity, decreased plasma volume, or the presence of car-
risk pattern.                                                                 diac disease predicted high-risk intraoperative hemody-
   As mentioned previously, intraoperative hypotension                        namic patterns. Other preoperative characteristics, such
and hyper/hypotension had been identified as predictors                       as age and diabetes, were also examined to determine
of specific postoperative complications: renal dysfunction,                   whether patients with either high-risk pattern differed from
myocardial ischemia, or infarction and congestive failure.                    those without that pattern. In Figure 1 the percentage of
Of the 207 patients, 44 had complications: 8 life-threat-                     patients who had intraoperative hypotension according

            TABLE 1. Rates of Postoperative Complications: Congestive Failure, Myocardial Infarction and Ischemia, and Renal Dysfunction
                                                        According to the Significant Predictors
                                             Neither High-risk Pattern                         Intraoperative Hypotension or Hyper/hypotension
                                        Intraoperative Hourly Fluid Balance                           Intraoperative Hourly Fluid Balance
                                 Positive               Negative              Total           Positive             Negative                    Total

No cardiac disease
  No diabetes                    10% (50)               15% (13)          11% (63)            10% (30)              53% (13)                  23% (43)
  Diabetes                       19% (32)               25% (4)           19% (36)            26% (15)               0% (1)                   25% (16)
Total                            13% (82)               18% (17)          14% (99)            16% (45)              50% (14)                  24% (59)
Cardiac disease
  No diabetes                     7% (14)                0% (3)            6% (17)            40% (10)              67% (3)                   46% (13)
  Diabetes                       55% (11)                 -               55% (11)            43% (7)              100% (1)                   38% (8)
Total                            28% (25)                0% (3)           25% (28)            41% (17)              75% (4)                   47% (21)
    Model X2 = 16.2 p < .0003; R = 0.196                                                High-risk pattern    0.71 ± 0.36       p   =   0.04
            Cardiac disease     0.82 ± 0.38         p   =   0.03                        Negative balance     0.92 ± 0.43       p   =   0.03
            Diabetes            0.82 ± 0.37         p   =   0.03
70                                                               CHARLSON AND OTHERS                                                        Ann.   Surg. * July   1990





             Not Yes
            No                             <100    100- >110        Not >400 <400           Not <3000 3000 >3500
             Severe Severe                         110              done                    done         3500
 n=    47     125 35                         81 51 36               106 68        33        157     19 14         17
A History of Hypertension                    Pre-operative MAP    Walking distance (meters)    Plasma Volume (oc)


                                                                                                                                 FIGS. lA-C. Percentage of
                                                                                                                                 patients with intraoperative
 40                                                                                                                              hypotension, hyper/hypo-
                                                                                                                                 tension, or neither high-risk
 30                                                                                                                              pattern according to preop-
                                                                                                                                 erative characteristics.
 Ba -n1aM


           NoYes          No      <   :6     >6      No    Yes Severe     No Yes            No Yes            No     Yes
  n   =
           176~           185N        7      15      186    9    12       187 20            158 49            186     21
      B    Angina                 ml                       CHF            Valvular         Any Cardiac         0    wave






 26   so            Ye           oY

           No       Yes      No            Yes      No Ye           No Yes           <70     >70          I                III

  n=        172 35           187 20                 167 40          136 71            141     66         124 61 22
      Chronic Pulmonary          Renal              Cardiomgaly     Diabetes           Age               Cardiac risk class
to their preoperative characteristics is shaded black, the    respectively.) Patients with cardiac disease had slightly
percentage with hyper/hypotension is cross-hatched, and       but not significantly increased rates. None of the other
the percentage with neither pattern is white. First we will   variables shown in Figure 1 were significant predictors.
discuss the predictors of intraoperative hypotension, and     When logistic regression was repeated in the subset of 50
second, intraoperative hyper/hypotension.                     patients for whom plasma volume was available, using
   Predictors of intraoperative hypotension. As shown in      preoperative MAP and a walking distance of less than
Figure 1A, patients with a history of severe hypertension     400 m as covariates, decreased plasma volume achieved
and those with a preoperative MAP >110 mmHg had               significance as a predictor of intraoperative hypotension
increased intraoperative hypotension. When these two          (p = 0.01).
characteristics were evaluated simultaneously, the actual        Predictors of intraoperative hyper/hypotension. As
preoperative MAP played the important role. For ex-           shown in Figure 1 A, intraoperative hyper/hypotension
ample, among patients who had a history of severe hy-         was not more common among patients with a history of
pertension but whose preoperative MAP was less than           severe hypertension or among patients with high preop-
100 mmHg, only 8% developed prolonged intraoperative          erative MAPs. Patients with walking distances of less than
hypotension; however 62% of those whose preoperative          400 m had an increased risk of intraoperative hyper/hy-
MAP was .1 10 mmHg had intraoperative hypotension,            potension: 30% of patients with a walking distance of less
regardless of the previous severity. Thirty per cent of pa-   than 400 m had hyper/hypotension as opposed to 16% of
tients with a 12-minute walking distance of less than 400     those without. In addition intraoperative hyper/hypoten-
m had intraoperative hypotension versus 21% of patients       sion occurred in 42% of the 19 patients with a plasma
with walking distances .400 m and 18% of patients who         volume less than 3000 cc, in 29% of those with a plasma
did not have a walking distance test performed. Only 6        volume .3000 and less than 3500 cc, and in only 6% of
of the 33 patients with a walking distance of less than 400   patients with a plasma volume .3500 cc.
m had any cardiac disease, so that the relationship between      Intraoperative hyper/hypotension was slightly increased
decreased walking distance and intraoperative hypoten-        with the presence of cardiac disease (Figure 1 B). Patients
sion was independent of their history of cardiac comor-       older than 70 years were more likely to have hyper/hy-
bidity. Plasma volumes were measured in 50 patients;          potension (Figure 1C). The presence of diabetes, pul-
26% of the 19 patients with a plasma volume less than         monary disease, renal disease, or cardiomegaly did not
3000 cc had intraoperative hypotension, 29% of the 14         increase the risk of hyper/hypotension.
patients with a plasma volume >3000 and less than 3500           Multivariate analysis revealed that age of more than 70
cc, but only 12% of the 30 patients with a plasma volume      years (p < 0.006) was the most important predictor of
.3500 cc.                                                     hyper/hypotension; walking distance less than 400 m did
   As shown in Figure 1B, patients with cardiac comor-        not achieve significance (p = 0.08). None of the other
bidity other than angina had slightly increased rates of      variables in Figure 1 were significant predictors. When
intraoperative hypotension. In total 27% of patients with     the logistic regression was repeated in the 50 patients for
congestive failure, previous myocardial infarction, or val-   whom plasma volume was available, using age and a
vular heart disease had intraoperative hypotension versus     walking distance of less than 400 m as covariates, de-
18% of those without. Once other cardiac comorbidity          creased plasma volume again achieved significance as a
was taken into account, a Q wave on ECG was not an            predictor of intraoperative hypotension (p = 0.03).
independent predictor; only 13% of patients with a Q wave        Preoperative predictors of high-risk patterns. Patients
but without a history of other cardiac comorbidity had        with decreased functional capacity (as measured by a 12-
intraoperative hypotension.                                   minute walking distance of less than 400 m were at in-
   There were no differences according to age or sex. Di-     creased risk for intraoperative hypotension and hyper/
abetics were slightly but not significantly more likely to    hypotension. In the 50 patients in whom it was measured,
have hypotension. Hypotension rates did not differ ac-        decreased plasma volume was associated with increased
cording to cardiac risk class or for patients with chronic    intraoperative hypotension and hyper/hypotension.
pulmonary disease. Patients with chronic renal disease        Therefore these two hypotheses were confirmed. Cardiac
and those with cardiomegaly did have higher hypotension       comorbidity did not attain significance in the multivari-
rates; however, once the preoperative MAP was taken into      able models. It was initially hypothesized that intraoper-
account, neither of these characteristics independently       ative hypertension would be most common among pa-
increased the risk of hypotension.                            tients with severe hypertension; however this was not true.
   Multivariate analysis confirmed that the preoperative      Instead patients with MAPs more than 110 had a signif-
MAP, especially when .110 mmHg, and a walking dis-            icantly increased rate of hypotension. In addition age of
tance of less than 400 m, were significant predictors of      more than 70 years was the most important predictor of
intraoperative hypotension (p < 0.0001 and p = 0.044,         hyper/hypotension.
72                                                         CHARLSON AND OTHERS                                               Ann. Surg. * July 1990

TABLE 2. Percentage of Patients with Intraoperative High-risk Patterns   pattern as compared to 42% of those having abdominal
 (Hypotension and Hyper/hypotension) According to Plasma Volume          surgery and 32% of those having vascular surgery. Only
                      and the Use of Diuretics
                                                                         18% of those having other types ofsurgery had either high-
                           Plasma Volume                                 risk pattern.
Diuretics    <3000 cc     .3000-<3500 cc        .3500 cc       Total         Furthermore the occurrence of either high-risk patterns
None         75% (8)          33% (6)           25% (12)     42% (26)    was directly related to the duration of surgery, with such
Yes          63% (11)         71% (7)           16% (6)      54% (24)    hemodynamic changes being infrequent when surgery
Total        68% (19)         46% (13)          22% (18)     46% (50)    lasted less than 2 hours, and more common when surgery
                                                                         lasted more than 4 hours. However high-risk patterns were
                                                                         no more frequent if the duration of the operation exceeded
   Once these characteristics were considered, preoperative              the expected duration.
administration ofdifferent pharmacologic agents (i.e., di-                   Multivariate analysis showed that, in addition to the
uretics, beta adrenergic blocking agents, calcium channel                previously identified preoperative characteristics, a du-
blockers, other antihypertensive drugs, digitalis) did not               ration of surgery more than 2 hours (p < 0.0001) or AAA
affect the rates of occurrence of either high-risk pattern.              repair (p = 0.03) were significant predictors of intraoper-
Because diuretics decrease plasma volume, it was some-                   ative hypotension, while a surgical procedure other than
what surprising that the use of diuretics was not associated             abdominal or vascular surgery was inversely related to
with a significant increase in high-risk patterns. Table 2               intraoperative hypotension (p = 0.02). Similarly, in ad-
shows the rates of intraoperative high-risk patterns among               dition to the preoperative characteristics, a duration of
the 50 patients who had plasma volume measured ac-                       surgery more than 2 hours (p = 0.02) was a significant
cording to the plasma volume and the use of diuretics.                   predictor of intraoperative hyper/hypotension, while a
Fifty-eight per cent (11 of 19) ofthe patients with a plasma             surgical procedure other than vascular, abdominal or aor-
volume less than 3000 cc, 54% (7 of 13) of those with a                  tic surgery (p = 0.03) was inversely related to the risk.
plasma volume between 3000 and 3500 cc, but only 33%                         The distribution of adverse intraoperative hemody-
(6 of 18) of those with a plasma volume .3500 cc were                    namic patterns according to the significant predictors (age
on diuretics. Thus the use ofdiuretics was correlated with               more than 70 years, decreased functional capacity, and
a lower plasma volume. However, as shown in Table 2,
                                                                         the type and duration of operation) are shown in Figure
it was the plasma volume that was the principal predictor                 3. Plasma volume was not included because data was only
of high-risk patterns. Patients on diuretics with a higher               available for 50 patients. In this figure patients whose
plasma volume did not experience the pattern; patients                   walking distance was less than 400 m, regardless of age,
not on diuretics with a lower plasma volume did.                         are grouped together. Patients whose surgery lasted less
   Relationship ofpreoperative characteristics that predict              than 2 hours and those undergoing surgery other than
high-risk intraoperative patterns to postoperative compli-                vascular, aneursym, or abdominal surgery were combined
cations. Table 3 shows that the preoperative characteristics              into a lower risk group (Figure 3B); patients who had
that predicted intraoperative hypotension or hyper/hy-                    more than 2 hours of abdominal, vascular, or aortic
potension did not directly predict cardiac and renal com-
plication rates. Complications occurred at increased rates                  TABLE 3. Percentage of Patients with Postoperative Cardiac and
among patients who had either high-risk pattern. De-                          Renal Complications According to Preoperative Predictors
creased functional capacity (according to a walking dis-                                 of High-risk Intraoperative Patterns
tance of less than 400 m), decreased plasma volume, pre-                                                    Intraoperative Pattern
operative MAP, and age define the characteristics of pa-
                                                                            Preoperative                       Hypotension or
tients who are susceptible to particular high-risk                         Characteristics       Neither        Hypertension             Total
intraoperative patterns, but patients with those preoper-
ative characteristics who do not have intraoperative hy-                 Age
                                                                           <70                  12% (92)          28% (49)            18% (131)
potension or hyper/hypotension are not at increased risk                   270                  26% (35)          29% (31)            27% (66)
of postoperative complications.                                          Walking distance
                                                                           2400                 14% (42)          23% (26)            17% (68)
    Clearly, in addition to these preoperative characteristics,            <400                  8% (13)          30% (20)            21% (33)
the type of operation and the expected duration of surgery                 Not done             19% (72)          32% (34)            24% (106)
must be considered when evaluating the issue of high-risk                Preoperative MAP
                                                                           <100                 18% (72)          32% (37)            25% (109)
intraoperative hemodynamic patterns.                                       .100-<1 10           16% (36)          23% (26)            19% (62)
 The Type and Duration of Surgery and the Risk of Hy-                      .110                 11% (19)          29% (17)            19% (36)
                                                                         Plasma Volume
potension and Hyper/hypotension                                            Not done             16% (101)         32% (56)            22% (157)
    As shown in Figure 2, 76% of patients undergoing ab-                   <3500                16% (12)          20% (20)            19% (32)
                                                                           >3500                21% (14)          25% (4)             22% (18)
 dominal aortic aneursym (AAA) repair had a high-risk
Vol. 212 * No. I     PREOPERATIVE PREDICTORS OF INTRAOPERATIVE HYPOTENSION AND HYPERTENSION                                                       73


FIG. 2. Percentage of patients      68
with intraoperative hypoten-
sion, hyper/hypotension, or
neither high-risk pattern ac-       45
cording to type and duration        30
of operation.

                                                 Other       Abdominal     Vascular     AAA*                  < 2 hrs       2-4 hrs     > 4 hrs
                                         n=       49              85         50          25                     49            97           61
                                                         Type of Surgery                                             Duration of Operation
                                             Abdominal Aortic Aneurysm

aneursym surgery were the high-risk group (Figure 3A).                            Induction and Intubation
High-risk patterns were more frequent when surgery was                               The vast majority of patients received thiopental (n
more than 2 hours and the operation was intra-abdominal,                          = 186), innovar (n = 172), and nitrous oxide (n = 200)
vascular, or aneursym repair. Within that group, however,                         for induction of anesthesia. There were no differences in
the patients who had decreased walking distances or pre-                          hemodynamic patterns according to the agents used for
operative MAPs          .110     had increased rates of high-risk                 induction; however, because most patients received sim-
patterns.                                                                         ilar combinations of agents, differences would not be an-
   Having identified the preoperative characteristics of the                      ticipated.
patients who were at risk for one of the high-risk patterns,                         A number of different neuromuscular blocking agents
we then examined the intraoperative course of those with                          were used before intubation. Pancuronium alone (n = 16),
the high-risk patterns versus those without to understand                         pancuronium and succinylcholine (n = 29), pancuronium
the qualitative context in which these patterns occurred.                         and D-turbocurarine (n = 4), D-turbocurarine alone (n
We divided the analysis into two periods: from the onset                          = 11), D-turbocurarine and succinylcholine (n = 28),

of anesthesia to the beginning of surgery (including in-                          atracurium alone (n = 13), and atracurium and succi-
duction and intubation) and the period of surgery itself.                         nylcholine (n = 5) were all used. Succinylcholine alone
                                                                                  was used in 72 patients and 29 patients received no neu-
Intraoperative Course at the Onset ofAnesthesia: Patients                         romuscular blocking agent. There were no differences in
with Hypotension or Hyper/hypotension Versus Those                                the hemodynamic patterns according to the neuromus-
Without                                                                           cular blocking agents used for induction.
   In this section we evaluated how intraoperative man-
agement differed between those with and without high-                             Inhalation Anesthetic Agent
risk patterns according to premedications, agents used for                           There were no significance differences according to the
induction and intubation, and according to the hemo-                              inhalation anesthetic in the occurrence of high-risk pat-
dynamic patterns at the beginning of anesthesia.                                  terns; however the majority of patients received either
   Premedications. Eighty per cent of patients received                           enflurane or isoflurane. The seven patients who received
atropine and nembutal, while only 5% received neither.                            halothane were slightly more likely to have hypotension
There were no differences in the occurrence of either high-                       or hyper/hypotension, but this was not significant. Only
risk pattern according to the use of one or more of these                         one patient received methoxyflurane and 10 patients re-
premedications. Meperidine was given to 21 patients, who                          ceived nitrous oxide and fentanyl (none of these patients
did not have an excess of either pattern. On the other                            had either high-risk pattern).
hand, intravenous valium was given to 16 patients, 31 %
(5 of 16) of whom developed prolonged hypotension ver-                            Hemodynamic Pattern during Induction and Intubation
sus 19% (36 of 191) without, but this did not achieve                               Shown in Table 4A are the differences between the pre-
statistical significance.                                                         operative and intraoperative MAPs observed in the dif-
74                                                                               CHARLSON AND OTHERS                                                    Ann. Surg. - July 1990

                       Operation > 2 hours and Intra-abdominal, vascular or aneurysm surgery

                  I        -        Hypoten. Ion                   M        H-p.r & Hypotnesion                       He ith.z       I
188           I                 I    I             9         I      I                           5       a

 9 -
 78       -

 6 -
 so _
 30    -

 in    _

                  Age<70             Aoe>70                Walk <400                           Aeo<70         Agq>70            Walk <400
      n=              56                 34                    21                                13               6              4

A                 Pre-operative MAP                    <   110 mm Hg                           Pre-operative MAP       >   1 10 mm Hg       FIGS. 3A and B. Percentage
                                                                                                                                            of patients with intraopera-
                                                           Operation    <   2 hours or other surgery                                        tive hypotension, hyper/hy-
                                                                                                                                            potension, or neither high-
                                                                                                                                            risk pattern according to pre-
                                                                                                                                            operative MAP, type, and
iSO                                                                                                                                         duration of operation.






                      Aoe<70                  Age>70             Walk <400                                  Age<70              Age>70
          na=              38                 14                     8                                       11                  2

          Pre-operative MAP                    <       110 mm Hg                                Pre-operative MAP          >   1 10 mm Hg

 **Other surgery indudes operations other than intra-abdominal vascular or aneurysm

B No patients with walking distance <400 in this group
ferent hemodynamic groups immediately before induc-                                             of anesthesia. In contrast patients who had hyper/hypo-
tion, after induction, after intubation, and after 5 and 10                                     tension tended to be hypertensive when they arrived in
minutes ofanesthesia. Patients without either ofthe high-                                       the operating room, became markedly hypertensive after
risk patterns had a slight increase in MAP before induc-                                        intubation, and only experienced a decrease in MAP after
tion, returned to preoperative levels after induction, had                                      10 minutes of anesthesia. Patients who developed intra-
a modest increase in MAP after intubation, and then ex-                                         operative hypotension tended to have a decrease in MAP
perienced a 10 mmHg decrease in MAP after 10 minutes                                            after induction, had MAPs that remained low after in-
Vol. 212 * No. I      PREOPERATIVE PREDICTORS OF INTRAOPERATIVE HYPOTENSION AND HYPERTENSION                                                     75
   TABLE 4.        Changes in MAP During Induction, Intubation, and at the Onset ofAnesthesia in Relation to Subsequent Hemodynamic Pattern
                                                                                           Intraoperative Hemodynamic Pattern
                                                                                    Intraoperative                        Intraoperative
                                                   Neither                        Hyper/hypotension                        Hypotension
                                                  (n = 127)                            (n = 39)                              (n = 41)
                                     A. Within patient differences in MAP (Intra-operative minus preoperatives)
             Before induction                      9.3 ± 13.2                          16.7 ± 12.9                           3.6 ± 17.9
             After induction                       2.6 ± 16.9                           5.4 ± 15.7                         -9.4 ± 19.0
             After intubation                      8.7 ± 22.2                          24.5 ± 20.6                         -1.5 ± 21.6
             After inhalation
                 5 minutes                        -6.0 ± 17.8                           3.2 ± 23.0                        -15.6 ± 15.8
                10 minutes                       -10.0 ± 16.4                        -12.8 ± 11.9                         -23.8 ± 14.8

                                           B. Percentage of Patients with Specific Hemodynamic Patterns
MAP During Induction, Intubation,                   Neither                Hyper/hypotension                Hypotension                      Total
 and First 5 Minutes of Anesthesia                 (n = 127)                    (n = 39)                      (n = 41)                     (n = 207)
. 10 min .20 mm increase in MAP                     32% (10)                    56% (18)                      12% (4)                  100% (32)
.10 min .20 mm decrease in MAP                      62% (18)                                                  38% (11)                 100% (29)
Neither                                             68% (99)                    14% (21)                      17% (26)                 100% (146)

tubation, and more of them developed a pronounced de-                       hypotension as the predominant hemodynamic pattern
crease after the beginning of anesthesia.                                   revealed that only 21% of the patients experienced a 5-
   In Table 4B, patients were stratified according to                       to 10-minute episode of an increase in MAP above pre-
changes in their MAPs during the sequence of induction                      operative levels at any time during operation, usually with
and intubation; patients are divided into three groups:                     incision or manipulation.
those who had .10 minutes of .20 mmHg increases in                             As expect hypotension was higher when the inhalation
their MAPs, those who had .10 minutes of .20 mmHg                           anesthetic was increased (Figure 4). Vasoactive drugs were
decreases in their MAPs, and those who had neither. Of                      used infrequently and generally only when the absolute
the 32 patients with .10 minutes of increased MAP, 56%                      MAP decreased to less than 70 mmHg (Fig. 4A). For ex-
developed the pattern of intraoperative hyper/hypotension                   ample pressors were given to only four of the patients in
versus 14% of patients without (X2 = 35; p < 0.001); the                    the hypotension group, while hypotensive drugs were
sensitivity for subsequent hyper/hypotension was 27%,                       given to only five of the patients in this group.
the specificity was 89%. Of the 29 patients with 10 min-                       Six of the 41 patients in the hypotension group expe-
utes of decreased MAP, 38% developed prolonged hy-                          rienced a decrease in MAP at the onset of anesthesia and
potension in contrast to 17% of those without initial hy-                   continued to have low MAP throughout the operation.
potension. (X2 = 6.9; p < 0.01); the sensitivity for subse-                 In the other 35 patients, when their MAPs increased to
quent intraoperative hypotension was 62% and the                            values similar to their preoperative MAPs, they were given
specificity was 88%. One patient had hypertension only                      either fentanyl alone, fentanyl and D-turbocurarine, fen-
during induction/intubation and two patients had hy-                        tanyl and pancuronium, pancuronium alone, D-turbo-
potension only during induction/intubation. Excluding                       curarine, or all three. Consequently each of these agents
these three patients from the group defined as having ad-                   were associated with intraoperative hypotension. For ex-
verse intraoperative hemodynamic fluctuations does not                      ample 31% of the patients who received pancuronium
change the results. Specifically the fluctuations in MAP                    versus 16% without pancuronium had intraoperative hy-
that occurred during surgery were often heralded by the                     potension. Twenty-five per cent of those who received D-
events accompanying the sequence of induction, intu-                        turbocurarine versus 17% without this agent had hypo-
bation, and the onset of anesthesia; this was especially                    tension. Thirty-seven per cent of those who received more
true if the patient was immediately hypotensive. Having                     than 2 cc of fentanyl (excluding those receiving nitrous
evaluated whether the patterns differed during the period                   oxide-fentanyl for anesthesia) had hypotension versus 12%
from the beginning of anesthesia to surgery, we analyzed                    of those without it. Figure 4B shows the rates of hypoten-
what happened during surgery itself.                                        sion among patients receiving combinations of these
                                                                            agents. Of the 47 patients who received fentanyl and a
Hemodynamic Patterns During Surgery                                         neuromuscular blocking agent, 49% had hypotension ver-
   Course ofpatients with intraoperative hypotension. De-                   sus 11% of the 160 patients without agents. The use of
tailed evaluation of the 41 patients with intraoperative                    fentanyl and neuromuscular blocking agents was partic-
76                                                      CHARLSON AND OTHERS                                                      Ann. Surg. * July 1990

               Nitrous Halo- Enflur- Iso- Methoxy-             No      Yes            No     Yes                  No     Press- Hypo-        Both
                and    thane ane flurane flurane                                                                          ors    tensave
              fentanyl                                        Increase inhalaton      Pentothal                                  agents
                10      7     87      102     1                 172     35            175    32                 169       13      20          5

                                                                                                                  Vasoactive agents

         68 _
         4    -

         40   _
         1    -

          a    -

                   None   C)TC    Fentanyl    DTC & Pan-            DTC & Fentanyl All 3                     No        Yes
                                             Fentanyl curonium Pan-       Pan-                                Atracurium
                                                                 curonium curonium
   B          n=    73     20        44         16       13         10      28     5                          178       29
FIGS. 4A and B. Percentage of patients with intraoperative hypotension, hyper/hypotension, or neither    high-risk pattern according to inhalation
anesthetic, the use of neuromuscular blocking agents, narcotics, and vasoactive agents during surgery.

ularly correlated with prolonged hypotension; transient                      derstood when it is remembered that many of these pa-
increases in the pulse of more than 20 beats sometimes                       tients had preoperative MAPs 21 10.
accompanied the subsequent decrease in MAP. In contrast
prolonged hypotension did not occur more frequently                          Course ofPatients with Intraoperative Hyper/hypotension
among patients who received atracurium.
   The anesthesiologist was apparently actively intervening                     Detailed evaluation of patients who had intraoperative
when the patient's intraoperative MAP value increased                        hyper/hypotension revealed that nearly one half of the
and approached their preoperative MAP value. The pa-                         patients had elevated MAPs during induction and intu-
tients were being actively maintained at levels 20 mmHg                      bation, and then the MAPs decreased. However, as shown
below their preoperative MAP. This is more readily un-                       in Table 5, the changes in MAP were not limited to those
                  TABLE 5. The Patients with Intraoperative Hyper/hypotension: Changes in MAP of .20 mmHg that Occurred
                                               After the Onset of Surgery (e.g., After the Incision)
                                                                   Duration of Decreased Intraoperative MAP
       Duration of Increased in
         Intraoperative MAP              None              5-14 min.                15-29 min.                >30 min.             Total
       None                                0                   1                          0                      0                    1
       5-14 minutes                         1                  3                          1                      2                   7
       15-29 minutes                        1                  9                         10                      1                  21
       >30 minutes                         0                  10                          1                      0                  11
       Total                               2                  23                         12                      3                  39

occurring during the first minutes of anesthesia. Most of              balance, the number of hours in which a patient had a
the patients had repeated episodes of intraoperative hy-               negative fluid balance was also evaluated. For example a
pertension. Only those patients with .20 mmHg increases                patient may have had a positive fluid balance at the end
and decreases in MAP during surgery are shown in Table                 of the operation but may have had 1 or more hours of a
5. After their intraoperative MAPs increased during sur-               negative fluid balance during the operation. When this
gery, many of these patients were given fentanyl and/or                was examined, 47% of the 38 patients who had a negative
a neuromuscular blocking agent. For example 27% of the                 fluid balance for 1 hour or more had a high-risk pattern,
patients who received pancuronium versus 16% without                   in contrast to 36% of those without any hourly negative
had intraoperative hyper/hypotension. Twenty-eight per                 balances (p > 0.05).
cent of those who received D-turbocurarine versus 16%
without had hyper/hypotension. Twenty-nine per cent of                                                 Discussion
those who received more than 2 cc of fentanyl (excluding
those receiving nitrous oxide-fentanyl for anesthesia) had                Underlying the study are two basic assumptions: first
hyper/hypotension versus 17% of those without this agent.              that preoperative characteristics influence the suscepti-
   After a neuromuscular blocking drug and fentanyl, the               bility to different intraoperative hemodynamic patterns;
MAP often decreased. In total 44% of the 47 patients who               and second that intraoperative hemodynamic course re-
were given fentanyl and a neuromuscular blocking agent                 lates to postoperative complications. Obviously the ability
had the pattern of hyper/hypotension (versus 11% who                   to identify patients who are likely to develop hypotension
did not receive both of these agents). This pattern did not            or hyper/hypotension forms the basis for preoperative and
occur among patients who received atracurium. Not sur-                 intraoperative interventions designed to prevent or ame-
prisingly, given the changes in MAP, vasoactive drugs were             liorate such trends.
used most commonly in this group: 5 patients received
pressors, 11 hypotensive agents, and 5 received both.                  Are Changes of 20 mmHg Too Small to Be Important?
Thirty-one per cent of these patients also received thio-                 A basic issue is how to characterize the intraoperative
pental, presumably to assess whether the hypertension was              hemodynamic pattern. Studies have used a wide variety
the result of light anesthesia.                                        of strategies to define intraoperative blood pressure and
   The group without either high-risk pattern had either               pulse. In previous studies the hemodynamic patterns have
hypertension and hypotension of a shorter duration, but                been characterized as a 30% to 50% decrease or increase
not both. This group tended to receive either fentanyl or
neuromuscular blocking drugs for elevated MAPs but not
                                                                       in systolic pressure for 10 minutes or more,'   4'5'827 as a
                                                                       20% increase or decrease in MAP,9 or as a 20 to 50 mmHg
both. Only nine received vasoactive agents. Having eval-               absolute decrease in MAP.23"0 Others have evaluated
uated the anesthetic agents that were given during surgery,
we also analyzed the intraoperative course in relation to
fluid management.                                                                 TABLE 6. Intraoperative Hemodynamic Pattern and
                                                                                          Intraoperative Fluid Management
Fluid Management                                                                                                 Intraoperative    Intraoperative
                                                                                                  Neither      Hyper/hypotension    Hypotension
   There were no important differences in the hourly net                       MAP               (n= 127)           (n =39)          (n = 41)
fluids (total input - total output), in the hourly net saline-
                                                                       Total net hourly fluids   904   1015          831   440       801 ± 458
containing fluids, or in the hourly blood loss between the             Net hourly saline
groups (Table 6). Patients with hypotension and hyper/                   containing fluids       770 ± 1039          689 ± 433      666 ± 410
hypotension had higher mean central venous pressures                   Hourly blood loss         117 289             217 254        207 ± 211
                                                                       CVP                        7.8 4.4             9.5 4.4       11.3 ± 6.6
than those without. Apart from the mean overall fluid
78                                                 CHARLSON AND OTHERS                                          Ann. Surg. * July 1990

whether the systolic pressure increased to more than 180          Preoperative Predictors of Intraoperative Hemodynamics
mmHg or decreased to less than 80 mmHg; while others                 In this study preoperative functional capacity was a
have evaluated whether the MAP decreased to less than             significant predictor of intraoperative hypotension. Pa-
70 mmHg.8'28 Before this study we had selected intra-             tients who were unable to walk .400 m during 12 min-
operative changes of .20 mmHg in MAP in relation to               utes, most of whom did not have cardiac comorbidity,
the patient's usual MAP as potentially important.                 had increased rates of intraoperative hypotension. Initially
   This definition was chosen because of studies of auto-         functional capacity was assessed because measures of car-
regulation of cerebral and renal blood flow. Normally             diac performance at rest are poor predictors of circulatory
vessels in key organs, the brain, kidneys, heart, and some        performance during stress, such as surgery. Most studies
viscera, vasodilate in response to decreased blood pressure       to date have evaluated either performance on treadmill
and constrict in response to increased blood pressure.29'30       tests39'40 or radionucleotide cineangiograms41"42 as predic-
This autoregulation serves to maintain a relatively con-          tors of postoperative myocardial infarction or ischemia.
stant blood flow despite swings in systemic pressure; for         One study found that 22% of patients who were able to
example, in the normotensive, cerebral perfusion will re-         achieve less-than five METS on a treadmill test had post-
main normal despite fluctuations in MAP between 60                operative ischemic complications versus 11% without; in
and 120 mmHg.2930 When MAP decreases below the                    addition, five of the six patients with major complications
lower limit of autoregulation, hypoperfusion and ischemia         had a low exercise capacity.39 Another study found a re-
occur.3'133 When MAP increases over the upper limit of            lationship between ejection fraction and ischemic com-
autoregulation, blood flow suddenly increases, resulting          plications.4' None ofthese studies evaluated intraoperative
in local hyperemia.33'34                                          course in relation to preoperative exercise capacity. The
   In hypertensives the autoregulatory range is higher than       present study does not provide data about the patient's
in normotensives; in uncontrolled hypertensives, it may           ability to increment their ejection fraction in response to
be 1 10 to 180 mmHg and in controlled hypertensives 90            stress; however it is logical that patients who are unable
to 150 mmHg.3' 35'36 Thus in the hypertensive, the upper          or unwilling to walk .400 m in 12 minutes may be com-
and lower limits for autoregulation are generally raised          promised in this regard.
by 20 to 30 mmHg.37 Because autoregulation in regional               Patients with a decreased plasma volume were also at
circulations is normally preserved within 20 to 30 mmHg           increased risk of hypotension. Previous studies of trauma
ofthe usual MAP, we selected 20 mmHg above and below              and shock patients have demonstrated the relationship
the usual MAP as the definition of important changes in           between plasma volume and blood pressure."3 Surgery
intraoperative MAP. If these limits are important, then           results in losses of extracellular fluid."" Furthermore pa-
patients with >20 mm changes of longer intraoperative             tients who were volume depleted at the time of surgery
duration would have increased evidence of renal and car-          have been shown to have decreased renal plasma flow and
diac complications after operation. In fact we have dem-          glomerular filtration rate." If extracellular fluids were im-
onstrated in previous papers that intraoperative hypoten-         mediately replaced, renal function was restored, but be-
sion of 1 hour or more, or more than 15 minutes of hy-            yond a critical interval, the ability of the kidney to respond
pertension, in combination with lesser episodes of                was lost and only circulatory overload occurred with in-
hypertension, are significantly related to postoperative          creased fluids.45 Therefore it is not surprising that patients
renal dysfunction, myocardial ischemia, and congestive            with a decreased plasma volume were at increased risk
failure.' 1-13                                                    for intraoperative hypotension. It should be noted also
   It is critical to note that these definitions of hypotension
and hypertension are not used in standard practice. More          that patients older than 70 years and hypertensive patients
                                                                  may have decreased plasma volumes."6"'7
often absolute systolic (i.e., less than 90 mmHg or more
than 180 mmHg) or percentage changes in MAPs are                     For both plasma volume and functional capacity there
used.8'38 In a previous paper we evaluated the sensitivity        was missing data; it would have been preferable to have
and specificity of different definitions of intraoperative        complete data on the whole population. Nonetheless the
hypotension and hypertension and found that the defi-             relationship between these characteristics and intraoper-
nitions used in this paper had the highest true-positive          ative hemodynamics was sufficiently strong to be statis-
rate coupled with the lowest false-positive rate in pre-          tically significant.
dicting postoperative complications.'4 Because changes               Postoperative complications were related to the occur-
of this magnitude are not customarily used to define hy-          rence of the intraoperative hypotension and hyper/hy-
potension and hypertension, it is important to note that          potension, to the presence of cardiac disease and diabetes,
the patients in this study were managed according to ac-          were increased if net negative fluid balances occurred, and
cepted standard practice.                                         were higher among patients with cardiac disease, especially
Vol. 212 * No. I   PREOPERATIVE PREDICTORS OF INTRAOPERATIVE HYPOTENSION AND HYPERTENSION                                             79
diabetics with cardiac disease. Diabetes and cardiac disease      Changes in the intraoperative MAP can result from
have been identified as risk factors for postoperative car-    surgical stimulation, ventilatory changes, or anesthetic
diac complications.8'48 One study found that intraopera-       agents. Traction,68 declamping the aorta,69 or hypo-
tive hypotension in hypertensive patients correlated with      capnia" may produce hypotension, while incision and
postoperative complications but intraoperative hyperten-       clamping the aorta may produce hypertension.69 Hyper-
sion did not.49 In this study intraoperative hypertension      tensives may have exaggerated responses to noxious stim-
correlated with postoperative complications when it was        uli,7"72 resulting in greater fluctuations in MAP than in
accompanied by intraoperative hypotension, rather than         normotensives. While each of these factors may have
when it occurred alone.                                        played a role in the patterns seen in this study, there were
                                                               several other trends.
                                                                  In the patients who had intraoperative hypotension, as
Hemodynamic Response to Anesthesia and Surgery                 defined in this study, and who tended to have MAPs > 1 10
                                                               before operation, the anesthesiologist often intervened
   As found in this study, induction of anesthesia with        with either fentanyl, D-turbocurarine, or pancuronium
barbiturates usually results in peripheral vasodilatation,     when the MAP started to approach (but not exceed) pre-
producing a 10 to 20 mmHg decrease in MAP; the de-             operative levels. D-turbocurarine causes reduction in
crease may be greater in volume-depleted patients or in        MAP by reducing systemic vascular resistance,73 while
untreated hypertensive patients with a markedly increased      pancuronium may produced an increase in MAP by sym-
systemic vascular resistance.50'5' While fentanyl alone does   pathetic stimulation and selective cardiac vagal block-
not result in a decrease in pressure,52 the combination of     ade.73 Thus a different target MAP was maintained during
fentanyl and nitrous oxide decreases cardiac output and        operation. For this reason the relationship between the
pressure.53 During intubation, normotensives have in-          preoperative MAP and intraoperative hypotension may
creases of 20 to 25 mmHg in MAP54 and hypertensives            be related primarily to physician management, rather than
may have even greater increases, possibly because they         to inherent patient characteristics. Among the patients
have higher norepinephrine responses to this noxious           with the hyper/hypotension pattern, they began surgery
stimulus,55 or because of the increased vascular reactivity    with high MAPs and tended to have recurring substantial
to pressor stimuli.56'57 These patterns were seen in this      elevations of MAP, which often resulted in the adminis-
study with a decrease in MAP after induction and an im-        tration of either fentanyl, D-turbocurarine, or pancuron-
mediate increase after intubation.                             ium, and was followed by a substantial decrease in MAP.
   After intubation all but 10 of these patients were given    These patients had an intraoperative course that was ex-
an inhalation anesthetic for maintenance of anesthesia         tremely labile.
(10 were maintained on a combination of nitrous-oxide             Patients who had a preoperative mean arterial pressure
and fentanyl) and the MAP decreased from postintubation        (MAP) .1 10, a walking distance less than 400 m, or a
levels. Halothane, enflurane, and isoflurane produce a         plasma volume less than 3000 cc were at increased risk
dose-related depression of MAP,50'58'59 the first two from     of intraoperative hypotension (i.e., more than 1 hour of
a decrease in myocardial contractility and the latter from     .20 mmHg decreases in the MAP). Hypotension was also
a decrease in peripheral resistance. Myocardial blood flow     more common among patients having intra-abdominal
decreases with halothane and enflurane; however myo-           or vascular surgery, and among those who had operations
cardial oxygen consumption also decreases.60                   lasting longer than 2 hours. Patients older 70 years or
   Inhalation agents alter regional autoregulation so that     with a decreased plasma volume were at increased risk of
blood flow is more pressure dependent during anesthesia.       having more than 15 minutes of intraoperative elevations
With isoflurane, coronary blood flow is maintained, de-        of .20 mmHg over the preoperative MAP in combination
spite the decrease in systemic pressure.50 In dogs all three   with intraoperative hypotension; this was also more com-
inhalation anesthetics impair coronary autoregulation and      mon when surgery lasted more than 2 hours.
make flow more dependent on pressure.6' Halothane re-             Identifying patients who are susceptible to intraoper-
duces cerebral metabolism and produces cerebral vaso-          ative fluctuations in MAP is only a first step. However
dilation at constant pCO2,62 while isoflurane maintains        the data may help to tailor interventions designed to pre-
cerebral autoregulation.63 Renal autoregulation in the in-     vent or ameliorate these changes or their impact.
tact human may be abolished by inhalation anesthesia.30
With these alterations in autoregulatory function, it is not                                 References
surprising that changes in pressure were shown to corre-        1. Mauney FM, Ebert PA, Sabiston DC. Post-operative myocardial
late with postoperative cardiac and renal complica-                  infarction: a study of predisposing factors, diagnosis, and mortality
                                                                     in a high risk group of surgical patients. Ann Surg 1970; 172:
tions.27,64-67                                                        497-502.
80                                                           CHARLSON AND OTHERS                                                       Ann. Surg. * July 1990

  2. Nachlas MM, Abrams SJ, Golberg MM. The influence of arterio-             26. Harrell FE, Califf RM, Pryor DB, et al. Regression modelling strat-
        sclerotic heart disease on surgical risk. Am J Surg 1961; 10 1:447-          egies for improved prognostic prediction. Stat Med 1984; 3:143-
        455.                                                                         152.
 3. Chamberlain DA, Edmonds-Seal J. Effects of surgery under general          27. Coriat P, Harari A, Daloz M, et al. Clinical predictors of intraoper-
       anesthesia on the electrocardiogram in ischaemic heart disease                ative myocardial ischemia in patients with coronary artery disease
       and hypertension. Br Med J. 1964; 2:784-787.                                  undergoing noncardiac surgery. Acta Anaesthesiol Scand 1982;
 4. Steen PA, Tinker JH, Tarhan S. Myocardial reinfarction after anes-               26:287-290.
       thesia and surgery. JAMA 1978; 239:2566-2570.                          28. Slogoff S, Keats AS. Does perioperative myocardial ischemic lead
 5. Knorring JV. Postoperative myocardial infarction: a prospective                  to postoperative myocardial infarction. Anesthesia 1985; 62:107-
       study in a risk group of surgical patients. Surgery 1981; 90:55-                114.
        60.                                                                   29. Smith AL, Wollman H. Cerebral blood flow and metabolism: effects
 6. Barry KG, Mazze RI, Schwartz FD, et al. Mannitol infusion: II.                    of an/esthetic drugs and techniques. Anesthesiology 1972;36:378-
        Prevention of acute functional renal failure during resection of              400.
        aneursym of abdominal aorta. New Eng J Med 1961; 264:967-             30. Larson CP, Mazze RI, Cooperman LH, Wollman H. Effects of an-
        971.                                                                          esthetics on cerebral, renal and splanchnic circulations. Anesthesia
  7. Mazze RI, Barry KG. Prevention of functional renal failure during                 1974; 41:169-181.
       anesthesia and surgery by sustained hydration and mannitol in-         31. Farhat SM, Schneider RC. Observations on the effect of systemic
        fusion. Anesth Analg 1967; 46:61-68.                                          blood pressure on intracranial circulation in patients with cere-
  8. Goldman L, Caldera DL, Southwick FS, et al. Cardiac risk factors                 brovascular insufficiency. J Neurosurg 1967; 27:441-445.
        in noncardiac surgery. Medicine 1978; 57:357-370.                     32. Strangaard S. Autoregulation of cerebral blood flow in hypertensive
  9. Rao TKL, Jacobs KH, El-Etr AA. Reinfarction following anesthesia                 patients. Circulation 1976; 53:720-727.
        in patients with myocardial infarction. Anesthesiology 1983; 9:       33. Wright FS. Intrarenal regulation of glomerular filtration rate. N Engl
        499-505.                                                                      J Med 1974; 48:91-96.
10. Sechzer PH, Wachtel J, Keats AS. Circulatory lability and myocardial      34. Skinhoj E, Strangaard S. Pathogenesis of hypertensive encephalop-
        stress in patients undergoing cardiovascular surgery. Am J Surg               athy. Lancet 1973; i:461-462.
        1968; 116:686-692.                                                    35. Johansson B, Strangaard S, Lassen NA. On the pathogenesis of hy-
11. Charlson ME, MacKenzie CR, Gold JP, et al. Post-operative renal                   pertensive encephalopathy. Circ Res 1974; 34 (Suppl): 1-167-
        dysfunction can be predicted. Surg Gynecol Obstet 1989; 169:                   174.
        303-309.                                                              36. Strangaard S, Olsen J, Skinhoj E, et al. Autoregulation of brain cir-
12. Charlson ME, MacKenzie CR, Gold JP, et al. Risk for post-operative                culation in severe arterial hypertension. Br Med J 1973; 1:507-
        congestive heart failure (Submitted for publication).                         509.
13. Charlson ME, MacKenzie CR, Gold JP, et al. The pre-operative              37. Langfitt TR. Cerebral circulation and metabolism. J Neurosurg 1971;
        and qintra-operative hemodynamic predictors of post-operative                 40:561-576.
        myocardial infarction of ischemia in patients undergoing non-         38. Slogoff S, Keats AS. Randomized trial of primary anesthetic agents
        cardiac surgery. Ann Surg 1989; 5:637-648.                                    on outcome of coronary artery bypass operations. Anaesthesia
14. Charlson ME, MacKenzie CR, Gold JP, et al. The assessment of                       1989; 70:179-188.
        intra-operative blood pressure and pulse: how should change be        39. Carliner NH, Fisher ML, Plotnick, GD, et al. Routine preoperative
        measured and what changes are clinically important? Ann Surg                  exercise testing in patients undergoing major noncardiac surgery.
        (In press).                                                                   Am J Cardiol 1985; 56:51-58.
 15. Rose GA, Blackburn H. Cardiovascular Survey Methods. World               40. Leppo J, Plaja J, Gionet M, et al. Noninvasive evaluation of cardiac
         Health Organization Monograph 1968, Belgium, pp. 172-177.                    risk before elective vascular surgery. J Am Coll Cardiol 1987; 9:
 16. Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index                  269-76.
         of cardiac risk in noncardiac surgical procedures. N Engl J Med      41. Pasternack PF, Imparato AM, Riles TS, et al. The value of radio-
          1977; 297:845-50.                                                           nuclide angiogram in the prediction of perioperative mycoardial
 17. Kirkendall WM, Feinlieb M, Freis ED, Mark AL. American Heart                     infarction in patients undergoing lower extremity revasculariza-
         Association recommendations for human blood pressure deter-                  tion procedures. Circulation 1985; 72(Suppl II): 13-17.
         minations by sphygmomanometer. Hypertension 1981; 2:509-             42. Gibbons RJ. Predicting the presence of severe coronary artery disease
         519A.                                                                        with exercise radionuclide angiography. Cardiol Board Rev 1988;
 18. McGavin CR, Gupta SP, McHardy GJR. 12 minute walking distance                    5:28-32.
         for assessing disability in chronic bronchitis. Br Med J 1976; 1:    43. Noble RP, Gregersen MI. Blood volume in clinical shock II The
         822-823.                                                                     extent and cause of blood volume reduction in traumatic, hem-
 19. McGavin CR, Artvin L, Naoeh, et al. Dyspnea, disability and distance             orrhagic and burn shock. J Clin Invest 1946; 25:172-183.
         walked. A comparison of estimates of exercise performance in         44. Shires GT, Williams J, Brown F. Acute change in extracellular fluids
         respiratory disease. Br Med J 1978; 2:241-243.                               associated with major surgical procedures. Ann Surg 1961; 154:
20. Guyatt GHQ, Sullivan MJ, Thompson PJ, et al. The 6 minute walk:                    803-810.
         a new measure of exercise capacity in patients with chronic heart    45. Barry KG, Mazze RI, Schwartz FD. Prevention of surgical oliguria
         failure. Can Med Assoc J 1985; 132:919-923.                                  by sustained hydration. N Engl J Med 1964; 270:1371-1377.
21. Gregerson MI. A practical method for the determination of blood           46. Gregersen MI, Rawson RA. Blood volume. Physio Rev 1959; 39:
         volumes with the dye T-1824. J Lab Clin Med 1944; 29:1266-                   307-342.
          1286.                                                                47. Tarazi RC, Frohlich ED, Dustan HP. Plasma volume in men with
22. Charlson ME, Gold JP, MacKenzie CR, Shires GT. Blood pressure                      essential hypertension. N Engl J Med 1968; 278:762-765.
         monitoring during anesthesia and surgery: evaluation of an au-        48. Eagle KA, Singer DE, Brewster DC, et al. Dipyridamole-thallium
         tomatic recorder. Infections in Surgery 1988; 7:497-505.
23. Charlson ME, MacKenzie CR, Ales KL, et al. The post-operative                      scanning in patients undergoing vascular surgery. JAMA 1987;
         electrocardiogram and creatine kinase: implications for diagnosis             257:2185-2189.
         of myocardial infarction after non-cardiac surgery. J Clin Epi-       49. Goldman L, Caldera DL. Risks of general anesthesia and the elec-
         demiol 1989; 42:25-34.                                                        tivew operation in the hypertensive patient. Anesth 1979; 39:
24. Charlson ME, MacKenzie CR, Gold JP, Shires GT. Post-operative                      285-292.
         changes in serum creatinine: when do they occur and how much          50. Filner BF, Karliner JS. Alterations of normal left ventricular per-
         is important? Ann Surg 1989; 209:328-333.                                     formance by general anesthesia. Anesth 1976; 45:610-20.
25. Harrell F. PROC LOGIST Supplemental Users Guide. Cary, NC:                 51. Prys-Roberts C, Greene LT, Meloche R, Foex P. Studies ofanesthesia
         SAS Institute, 1983, pp. 180-202.                                             in relation to hypertension II. Hemodynamic consequences of
      induction and endotracheal intubation. Br J Anaesth 1971; 43:         63. Madsen JB, Cold GE, et al. Cerebral blood flow and metabolism
       531-546.                                                                    during isoflurane induced hypotension in patients subjected to
52. Roscow CE, Mos J, Philbin DM, Savarese JJ. Histamine release                    surgery for cerebral aneurysms. Br J Anesth 1987; 59:1204.
      during morphine and fentanyl anesthesia. Anaesthesia 1982; 56:        64. Roy WL, Edelist G, Gilbert B. Myocardial ischemia during non-
       93-96.                                                                      cardiac surgical procedures in patients with coronary artery dis-
53. Ngai SH. Effects of anesthetics on various organs. N Engl J Med                ease. Anesth 1979; 51:393-397.
        1980; 302:564-566.                                                  65. Godet G, Coriat P, Baron JF, et al. Prevention of intraoperative
54. Forbes AM, Daly FG. Acute hypertension during induction of anes-               myocardial ischemia during noncardiac surgery with intravenous
       thesia and endotracheal intubation in normotensive man. Br J                diltiazem: a randomized trial vs. placebo. Anesth 1987; 66:24 1-
       Anaesth 1970; 42:618-624.                                                   245.
55. Low JM, Harvey JT, Prys-Roberts C, Dagnino J. Studies ofanesthesia      66. Thompson IR, Mutch WA, Culligan JD. Failure of intravenous ni-
       in relation to hypertension VII. Adrenergic responses to laryn-             troglycerin to prevent intraoperative myocardial ischemia during
       goscopy. BrJ Anaesth 1986; 58:471-477.                                      fentanyl-pancuronium anesthesia. Anesth 1984; 61:385-393.
56. Folkow B. The hemodynamic consequences of adaptive structural           67. Hayes MA. The influence of shock with clinical renal failure on
       changes of the resistance vessels in hypertension. Clin Sci 1971;           renal function. Ann Surg 1957; 146:523-529.
       41:1-9.                                                              68. Vandam LD, Schweizer H, Kubota Y. Circulatory response to intra-
57. Bohr DR, Berecek KH. Relevance of vascular structural and smooth               abdominal manipulation during ether anesthesia in man. Circ
       muscle sensitivity change in hypertension. Aust NZ J Med 1976;              Res 1962; 11:287-295.
       6 (Suppl 2):26-32.                                                   69. Kalman PG, Wellwood MR, Weisel RD, et al. Cardiac dysfunction
58. Calverly RK, Smith NT, Prys-Roberts C, et al. Cardiovascular effects           during abdominal aortic operation: the limitations of pulmonary
       of enflurane anesthesia during controlled ventilation in man.               wedge pressures. J Vasc Surg 1986; 3:773-781.
       Anesth Analg 1978; 57:619-628.                                       70. Prys-Roberts C, Foex P, Greene LT, et al. Studies of anesthesia in
59. Stevens WC, Cromwell TH, Halsey MJ, et al. The cardiovascular                  relation to hypertension IV The effects of artificial ventilation on
       effects of a new inhalation anesthetic Forane in human volunteers           the circulation and pulmonary gas exchanges. Br J Anaesth 1972;
       at constant arterial carbon dioxidc tension. Anesth 1971; 35:8-             44:335-348.
60. Moffit EA, Barker RA, et al. Myocardial metabolism and hemo-            71. Shapiro AP, Moutsos Se, Krifcher E. Patterns of pressor response
       dynamic responses with isoflurane anesthesia for coronary arterial          to noxious stimuli in normal hypertensive and diabetic subjects.
       surgery. Anesth Analg 1986; 65:53-61.                                       J Clin Invest 1963; 42:1890-1898.
61. Hickey RF, Stiebert PE, et al. Effects of halothane, enflurane and      72. Lorimer AR, MacFarlane PW, Provan G, et al. Blood pressure and
       isoflurane on coronary blood flow autoregulation and coronary               catecholamine response to stress in normotensive and hyperten-
       vascular reserve in the canine heart. Anesth 1988; 68:21-30.                sive subjects. Cardiovasc Res 1971; 5:169-173.
62. Albrecht RG, Miletich DJ, et al. Cerebral blood flow and metabolic      73. Stoelting RK. The hemodynamic effects of pancuronium and d-
       changes from induction to onset of anesthesia with halothane                turbocurarine in anesthetized patients. Anesth 1972; 136:612-
       and pentobarbitol. Anesth 1977; 47:252-256.                                 615.

Shared By: