120596 effect of atenolol on mortality and cardiovascular morbidity
Journal of Medicine
V O L U M E 3 3 5
N U M B E R 2 3
EFFECT OF ATENOLOL ON MORTALITY AND CARDIOVASCULAR MORBIDITY
AFTER NONCARDIAC SURGERY
DENNIS T. MANGANO, PH.D., M.D., ELIZABETH L. LAYUG, M.D., ARTHUR WALLACE, PH.D., M.D., AND IDA TATEO, M.S.,
FOR THE MULTICENTER STUDY OF PERIOPERATIVE ISCHEMIA RESEARCH GROUP*
ORTALITY and morbidity due to car-
Perioperative myocardial ischemia is
diovascular disease are prevalent and
the single most important potentially reversible risk
costly for the 30 million patients who
factor for mortality and cardiovascular complications
undergo noncardiac surgery annually
after noncardiac surgery. Although more than 1 mil-
in the United States, affecting more than 1 million
lion patients have such complications annually, there
of them.1,2 In the subgroup of 3 million patients
is no effective preventive therapy.
who require noncardiac surgery who have or are at
We performed a randomized, double-
risk for coronary artery disease, the most significant
blind, placebo-controlled trial to compare the effect of
risk factors for mortality and cardiovascular morbid-
atenolol with that of a placebo on overall survival and
ity are myocardial ischemia and nonfatal myocardial
cardiovascular morbidity in patients with or at risk for
infarction during the first week after surgery; these
coronary artery disease who were undergoing non-cardiac surgery. Atenolol was given intravenously be-
factors increase the risk of serious cardiovascular
fore and immediately after surgery and orally there-
outcomes by a factor of 2 to 20 over the two years
after for the duration of hospitalization. Patients were
after surgery.3-5 Postoperative ischemic events appear
followed over the subsequent two years.
to be related to the persistently exaggerated sympa-
A total of 200 patients were enrolled. Nine-
thetic response that is associated with substantial in-
ty-nine were assigned to the atenolol group, and 101
creases in the heart rate throughout the hospitaliza-
to the placebo group. One hundred ninety-four pa-
tion.6-10 Several small clinical trials have investigated
tients survived to be discharged from the hospital,
the effect of the preoperative or intraoperative use of
and 192 of these were followed for two years. Over-
nitrates,11,12 beta-blockers,13-15 calcium-channel block-
all mortality after discharge from the hospital was
ers,16,17 or a
-agonists18,19 on hemodynamics and meas-
significantly lower among the atenolol-treated pa-
tients than among those who were given placebo
ures of myocardial ischemia. Although several of the
over the six months following hospital discharge
preliminary findings were encouraging, previous tri-
(0 vs. 8 percent, P0.001), over the first year (3 per-
als have not investigated the effects of such therapies
cent vs. 14 percent, P0.005), and over two years
when administered over the entire postoperative hos-
(10 percent vs. 21 percent, P0.019). The principal
pital stay. Even more important, the effects of such
effect was a reduction in deaths from cardiac caus-
intensive therapy on long-term mortality (i.e., mor-
es during the first six to eight months. Combined
tality after discharge from the hospital) and the in-
cardiovascular outcomes were similarly reduced
cidence of cardiovascular events remained to be de-
among the atenolol-treated patients; event-free sur-
vival throughout the two-year study period was 68
We hypothesized that in patients who had or were
percent in the placebo group and 83 percent in theatenolol group (P0.008).
In patients who have or are at risk for
coronary artery disease who must undergo noncardi-ac surgery, treatment with atenolol during hospital-ization can reduce mortality and the incidence of car-
From the San Francisco Veterans Affairs Medical Center and University
of California, San Francisco (D.T.M., A.W.), and the Ischemia Research and
diovascular complications for as long as two years
Education Foundation, San Francisco (E.L.L., I.T.). Address reprint re-
after surgery. (N Engl J Med 1996;335:1713-20.)
quests to Dr. Mangano at the San Francisco Veterans Affairs Medical Cen-
1996, Massachusetts Medical Society.
ter, 4150 Clement St., 129, San Francisco, CA 94121.
*Members of the Multicenter Study of Perioperative Ischemia Research
Group are listed in the Appendix.
at risk for coronary artery disease, the intensive ad-
100 mm Hg, no atenolol (or placebo) was given. No treating cli-
ministration of beta-blockers before and after sur-
nician was allowed to observe administration of the study drug
gery, continuing throughout the period of hospital-
either before or after surgery.
ization, would decrease mortality and the incidence
of serious cardiovascular events during the two years
All patients underwent general anesthesia with endotracheal in-
tubation; preoperative medications were continued until the timeof surgery, with beta-blockers replaced by the study drug on the
morning of surgery. There were no other protocol-based restric-
tions of anesthetic or surgical technique, and clinical decisionswere not controlled by the study protocol. Perioperative informa-
Eligible patients were those with or at risk for coronary artery
tion, which was recorded and analyzed for possible confounding
disease who were scheduled for elective noncardiac surgery re-
effects, included the type and duration of surgery, the anesthetic
quiring general anesthesia at the San Francisco Veterans Affairs
techniques used, details of fluid and blood loss and replacement,
Medical Center. The presence of coronary artery disease was in-
cardiovascular medications, hemodynamic variables, electrocar-
dicated by a previous myocardial infarction, typical angina, or
diographic data, and adverse events.
atypical angina with a positive stress test; a patient was consideredat risk for coronary artery disease when he or she had at least two
Follow-up and Outcome Measures
of the following cardiac risk factors: age 65 years, hypertension,current smoking, a serum cholesterol concentration 240 mg per
Of the 200 patients enrolled, 194 were discharged after surgery
deciliter (6.2 mmol per liter), and diabetes mellitus.3,4 No more
and 6 died during hospitalization. Three deaths were secondary to
than 1 patient per day was enrolled, and of the 204 patients who
myocardial infarction (two in the placebo group and one in the
agreed to participate in the study, 200 were enrolled and under-
atenolol group). Three deaths had noncardiac causes; two were
went randomization (1 withdrew and 3 did not have surgery); 99
secondary to metastatic cancer (both in the atenolol group), and
were assigned to the atenolol group and 101 to the placebo
one was caused by pulmonary failure after an infusion of 23 liters
of crystalloid, colloid, and blood over a period of 24 hours for fluidloss (in the atenolol group). Among the 194 patients discharged,
Administration of the Study Drugs
outcome data were collected for 192 (99 percent); 1 was lost tofollow-up and 1 was not followed because surgery was not per-
Patients were randomly assigned to receive either atenolol or
formed after he received the study drug. Six months, one year, and
placebo before the induction of anesthesia, immediately after sur-
two years after surgery, study physicians conducted scheduled eval-
gery, and daily throughout their hospital stay (up to seven days).
uations that were independent of the patients' usual clinical care.
All clinical and study personnel were blinded to the study-group
Death was considered due to cardiac causes if the patient died
assignments throughout all phases of this trial. Intravenous and
of a myocardial infarction, dysrhythmia, or congestive heart fail-
oral preparations of the active drug (atenolol) and placebo were
ure caused primarily by a cardiac condition. The diagnosis of my-
prepared by the hospital pharmacy according to a computer-gen-
ocardial infarction required at least one of the following: devel-
erated, randomized list that was retained only by the pharmacy
opment of new Q waves (as defined by Minnesota Code 1-1-1 or
and that remained confidential until formal unblinding after the
1-2-7); new persistent ST-segment or T-wave changes (Minnesota
data base was closed. The intravenous preparation consisted of
Code 4-1, 4-2, 5-1, or 5-2)3 associated at the time of hospitaliza-
two 10-ml syringes, each containing 5 mg of atenolol or placebo;
tion with an elevation of total creatine kinase and creatine kinase
the oral preparation consisted of two 50-mg tablets of atenolol
MB isoenzyme values; evidence at autopsy of acute myocardial in-
or two placebo tablets. Approximately one hour before surgery,
farction; or documentation of myocardial infarction in the hospi-
patients entered the preoperative area, their blood pressure was
tal records.3 Unstable angina was defined as severe precordial
recorded with an automated cuff, and a five-lead continuous elec-
chest pain that lasted at least 30 minutes, was unresponsive to
trocardiogram was recorded. Thirty minutes before entry into the
standard therapeutic maneuvers, and was associated with transient
operating room, the intravenous administration of the study drug
ST-segment or T-wave changes without the development of
Q waves or diagnostic enzyme abnormalities. The diagnosis of con-
Administration of the study drug at each time point required
gestive heart failure was made when a patient had symptoms or
that the heart rate be 55 beats per minute, that the systolic
signs of pulmonary congestion (shortness of breath and rales),
blood pressure be 100 mm Hg, and that there be no evidence
signs of new left or right ventricular failure (cardiomegaly, an S
of congestive heart failure, third-degree heart block, or broncho-
sound, jugular venous distention, and peripheral edema), abnor-
spasm (as defined in the first International Study of Infarct
mal results on chest radiography (vascular redistribution, intersti-
Survival20). If these criteria were met, the first syringe of the study
tial edema, and alveolar edema), and a change in medication in-
drug was infused over a period of five minutes, the patient was
volving at least the institution of treatment with diuretic agents.3
observed for an additional five minutes, and if the criteria con-
The outcome measures were prescribed by the study protocol.
tinued to be met, the second syringe was infused over a period
The primary outcome was mortality from all causes during the
of five minutes. Immediately after surgery, the study drug was
two years after discharge from the hospital. The secondary out-
again given, in the same way.
come variable combined myocardial infarction, unstable angina
Starting on the morning of the first postoperative day, and each
or congestive heart failure requiring hospital admission and clin-
day thereafter until the patient was discharged from the hospital
ical diagnosis and treatment, myocardial revascularization (coro-
(up to a maximum of seven days), patients received the study
nary-artery bypass graft surgery or percutaneous transluminal cor-
drug in the manner described for intravenous infusion (every 12
onary angioplasty), and death. Autopsy data, if available for patients
hours) or once a day orally (if possible), at which time, if the
who died during the two-year period, were reviewed at the cen-
above criteria were met, atenolol (50 or 100 mg) or placebo was
tral laboratory (at the Ischemia Research and Education Founda-
given daily. If the heart rate was above 65 beats per minute and
tion) by a pathologist who was blinded to the patients' treatment
the systolic blood pressure was 100 mm Hg or higher, two tablets
of atenolol (total dose, 100 mg) or two tablets of placebo weregiven orally; if the heart rate was 55 or higher but no higher than
65 and the systolic blood pressure was 100 mm Hg or higher, onetablet of atenolol or one tablet of placebo was administered; if the
We designed the study to permit the assessment of both in-hos-
heart rate was below 55 or the systolic blood pressure was below
pital events (such as hemodynamic changes, dysrhythmias, and
December 5, 1996
ischemia21) and adverse cardiovascular outcomes during the twoyears after discharge. Using the log-rank survival test for the esti-
CHARACTERISTICS OF THE PATIENTS, ACCORDING TO
mation of the sample size (BMDP statistical software), we calcu-
lated that 198 patients would be necessary for the assessment ofmortality and 158 for the assessment of the combined outcomevariable; using the z statistic, we calculated that 170 patients
would be required for the assessment of in-hospital events. The
risk of death in different categories (death from all causes, from
Definite coronary artery disease (%)
cardiac causes, and from noncardiac causes, all at six months, oneyear, and two years) was compared between the groups by Kap-
At risk for coronary artery
lan–Meier methods, as was event-free survival after discharge.
Univariable predictors of two-year mortality were identified with
History of cardiac disease (%)
Cox proportional-hazards regression techniques,22 after we first
Coronary bypass surgery
verified that the assumption of the hazards model was valid.23
Predictors with two-tailed P values below 0.10 were entered into
the multivariable models, and a series of models was constructed
by adding variables, as long as the resulting multivariable model
had a lower P value (by chi-square analysis) than competing mod-
Congestive heart failure
els. Analyses were performed with use of Statistical Analysis Sys-
Cardiac risk factors (%)
tem software (SAS Institute, Cary, N.C.).
Cholesterol 240 mg/dl
The 200 study patients were middle-aged or eld-
erly persons who smoked and had a history of hy-
Preoperative medications (%)
pertension and chronic medical problems (Table 1).
There were no significant differences between the
groups, except that a higher proportion of the aten-
olol group was receiving treatment for hypertension.
Overall Mortality and Mortality from Cardiac Causes
Thirty patients (15.6 percent of the 192 who were
followed after hospital discharge) died during the
Duration of anesthesia
two-year follow-up period. Twenty-one of these
Type of surgery (%)
deaths (12 of which were from cardiac causes) oc-
curred in the placebo group, and 9 (4 of which were
from cardiac causes) in the atenolol group; thus,
overall mortality was 55 percent lower in the ateno-
lol group (P0.019), and mortality from cardiac
*Plus–minus values are means SD. Ranges are shown in parentheses.
causes was 65 percent lower (P0.033). The prin-
†Patients considered at risk for coronary artery disease were those with
cipal effect of atenolol therapy was on cardiac out-
two or more of the following: age 65 years, hypertension, current smok-
comes occurring during the first six to eight months
ing, serum cholesterol concentration 240 mg per deciliter (6.2 mmol per
(1 death from noncardiac causes in the atenolol
liter), and diabetes mellitus.
group vs. 10 in the placebo group, 7 of which were
‡Other surgery included intrathoracic procedures (in 1 patient in the
atenolol group and 3 in the placebo group), other general or plastic sur-
from cardiac causes; P0.001); the length of time
gery (in 15 in the atenolol group and 11 in the placebo group), and head
to the first death was 19 days in the placebo group
and neck surgery (in 2 in the atenolol group).
and 237 days in the atenolol group. After eightmonths, there was no substantial difference betweenthe groups; however, the early difference in survivalbetween the groups was preserved at one year
decrease of 67 percent from the rate in the placebo
(3 deaths in the atenolol group vs. 14 in the placebo
group within one year (7 events vs. 22 events, P
group, P0.005) and at two years (9 vs. 21, P
0.003), and a decrease of 48 percent in the two
0.019); the survival rate was significantly higher in
years after surgery (16 events vs. 32 events, P
the atenolol group at all times (Fig. 1).
0.008). The principal effect of atenolol treatmentwas evident over the first 6 to 8 months after sur-
Combined Cardiac Outcomes
gery; the time to the first adverse event in each
Atenolol-treated patients who survived to hospital
group was 6 days for the placebo group, as com-
discharge had a significant decrease in the rate of
pared with 158 days for the atenolol group. There-
cardiac events, as compared with the rate in the pla-
after, there was no substantial difference between
cebo group, within six months after surgery (there
the groups; however, the early difference in event-
were no such events in the atenolol group, as com-
free survival was preserved over the two years after
pared with 12 in the placebo group; P0.001), a
surgery (Fig. 2).
PREDICTORS OF DEATH AMONG PATIENTS UNDERGOING
HAZARD RATIO (95% CI)
Oral hypoglycemic treatment
Ischemia on Holter monitoring on
postoperative days 0–2
Days after Surgery
Overall Survival in the Two Years after Noncardiac
Surgery among 192 Patients in the Atenolol and Placebo
Groups Who Survived to Hospital Discharge.
*The patients included in these models were the 192 of the original ran-
The rate of survival at 6 months (180 days) was 100 percent in
domized group of 200 who survived to hospital discharge and were fol-
the atenolol group and 92 percent in the placebo group
lowed for two years after discharge; 30 of these 192 patients (15.6 percent)
(P0.001); at 1 year (360 days), the rates were 97 percent and
died during the two years of follow-up. CI denotes confidence interval.
86 percent, respectively (P0.005); and at 2 years (720 days),90 percent and 79 percent (P0.019).
(hazard ratio, 1.2; P0.76), whereas in patients giv-
en placebo, the presence of diabetes was associated
with a quadrupling of the risk (hazard ratio, 4.0;
P0.003). No other perioperative variables were
associated with outcome. Several medications were
used more frequently in one group than in another
(Table 3), but there was no independent association
between the use of these medications and outcome,
as shown by the estimated odds ratios and the cor-
Event-free Survival (%)
responding P values (Table 3).
During the two years after discharge from the
Days after Surgery
hospital, there was no difference between the groups
Event-free Survival in the Two Years after Noncardiac
in the use of any cardiovascular medication (Table
Surgery among 192 Patients in the Atenolol and Placebo
3); therefore, the use of such medications did not
Groups Who Survived to Hospital Discharge.
confound the observed effect of atenolol on two-
The outcome measure combined the following events: myocar-
year mortality. These data also indicate that cardio-
dial infarction, unstable angina, the need for coronary-artery
vascular medications administered before admission
bypass surgery, and congestive heart failure. The rate of event-
to the hospital were still given at hospital discharge
free survival at 6 months (180 days) was 100 percent in the aten-olol group and 88 percent in the placebo group (P0.001); at
in most patients, but not all, and that after discharge
1 year (360 days), the rates were 92 percent and 78 percent, re-
the patients in the placebo group continued to use
spectively (P0.003); and at 2 years (720 days), 83 percent and
cardiovascular medications at least as often as the pa-
68 percent (P0.008).
tients in the atenolol group.
Tolerance and Adverse Reactions
Other Indicators of Treatment Effect
More than 85 percent of the patients tolerated the
During treatment, the average heart rate was sig-
intravenous administration of atenolol before and
nificantly lower in the atenolol group (75 beats per
immediately after surgery and its oral administration
minute, vs. 87 in the placebo group; P0.001), as
during the postoperative period; more than 60 per-
was the maximal heart rate (113 vs. 130 beats per
cent were able to receive the full daily dose of aten-
minute, P0.001). Multivariable correlates associat-
olol (10 mg intravenously or 100 mg orally) (Table
ed with survival at two years (shown in Table 2) were
4). In approximately 10 percent of the patients, the
a history of diabetes mellitus and atenolol therapy,
intravenous administration of atenolol before or af-
with atenolol improving two-year survival in patients
ter surgery was associated with a decrease of 20 per-
with diabetes by approximately 75 percent (hazard
cent or more in the systolic blood pressure or heart
ratio for death, 0.25; P0.03). Similarly, in ateno-
rate (Table 4); however, no patient had a systolic
lol-treated patients, the presence of diabetes was not
blood pressure below 90 mm Hg or a heart rate be-
associated with a significantly increased risk of death
low 40 beats per minute, and none required therapy.
December 5, 1996
USE OF CARDIOVASCULAR MEDICATIONS BEFORE AND AFTER SURGERY, ACCORDING TO STUDY GROUP.*
NO. WITH DATA
PLACEBO VALUE ATENOLOL
*Chi-square statistics were used to compare the two groups. ACE denotes angiotensin-converting enzyme.
†Odds ratio for mortality at two years associated with beta-blocker use before admission0.80 (P0.73).
‡Odds ratio for mortality at two years associated with use of calcium-channel blockers before admission1.06 (P0.90).
§Odds ratio for mortality at two years associated with ACE-inhibitor use before admission1.45 (P0.50).
¶One patient of the 95 in the atenolol group was not included in these calculations because surgery was delayed for several days after the study drug
Odds ratio for mortality at two years associated with beta-blocker use at discharge0.61 (P0.52).
**Odds ratio for mortality at two years associated with use of calcium-channel blockers at discharge0.85 (P0.74).
††Odds ratio for mortality at two years associated with nitrate use at discharge1.32 (P0.64).
‡‡Odds ratio for mortality at two years associated with ACE-inhibitor use at discharge1.17 (P0.79).
§§Odds ratio for mortality at two years associated with use of calcium-channel blockers for six months1.05 (P0.92).
The oral administration of atenolol was not associ-
discharge, with the odds of such outcomes 28 times
ated with an increased incidence of hypotension,
higher in patients with postoperative ischemia, as
bradycardia, or other events.
compared with those without ischemia, by six monthsafter surgery, 20 times higher at one year, and 14
times higher at two years.3-5,24-26 In addition, studies
The results of this trial demonstrate that, in pa-
have demonstrated an association between postop-
tients who have or are at risk for coronary artery dis-
erative ischemia and an elevated heart rate and have
ease and who are undergoing noncardiac surgery,
suggested that mitigation of this heart-rate response
mortality and cardiovascular events after discharge
may reduce the incidence or severity of ischemia (or
from the hospital can be substantially reduced by the
both).6-10,21 Thus, we concluded that intensive peri-
administration of atenolol throughout hospitaliza-
operative beta-blockade, if it could attenuate the
tion for surgery. The length of time to the first ad-
heart-rate response and limit the development of is-
verse event, survival, and event-free survival were all
chemia, might substantially reduce longer-term car-
significantly improved with atenolol, particularly
during the first six to eight months after surgery,
The large treatment effect that we observed —
and the effects on survival persisted for at least two
namely, an absolute increase of 15 percentage points
years. Among the atenolol-treated patients who sur-
in event-free survival after hospital discharge (from
vived to discharge from the hospital, survival was 90
68 percent to 83 percent) in the atenolol group as
percent two years after surgery, as compared with 79
compared with the placebo group — was unexpect-
percent in the placebo group, and event-free survival
ed. Several smaller trials, however, have reported
was 83 percent, as compared with 68 percent. More-
sizable effects of beta-blockers on perioperative
over, perioperative beta-blockade appeared to be well
ischemia,13-15,21 and observational studies have dem-
tolerated by these patients, despite the high preva-
onstrated an 18 percent difference in event-free sur-
lence of cardiac and pulmonary disease.
vival at two years between patients who had post-
What is the rationale for using perioperative beta-
operative ischemia and those who did not.4
blockade for the prevention of long-term adverse
Furthermore, we found that the principal effect of
outcomes? Studies conducted over the past decade
beta-blockade was evident within the first six to
have established the association between postopera-
eight months after surgery; this finding is consistent
tive myocardial ischemia and adverse outcomes after
with the temporal profile of the association between
DAILY DOSE AND SIDE EFFECTS OF ATENOLOL.*
DAY OF SURGERY
percentage of patients
Systolic BP 90 mm Hg
20% decrease in systolic BP
Heart rate 40 bpm
20% decrease in heart rate
Bradycardia and hypotension
Systolic BP 90 mm Hg and
heart rate 40 bpm
20% decrease in heart rate
Congestive heart failure
*There were 99 patients in the atenolol group and 101 in the placebo group. BP denotes blood
pressure, and bpm beats per minute.
†Effects include hypotension, bradycardia, congestive heart failure, or bronchospasm at any time
on days 1 through 7, as reported by the clinical staff.
‡The full dose was 10 mg for intravenous administration, and 100 mg for oral administration. A
half-dose was 5 mg intravenously and 50 mg orally. Patients not treated received no study drug be-cause the criteria for administration were not met.
§Two patients in the atenolol group whose condition was stable after intravenous drug adminis-
tration received treatment after intubation for bradycardia 30 to 75 minutes after the intravenousadministration of atenolol.
¶Two patients had bronchospasm after intubation, 1 to 3 hours after intravenous atenolol admin-
istration; one patient had bronchospasm after extubation, 8.5 hours after intravenous atenolol ad-ministration.
perioperative myocardial ischemia and outcomes at
calcium-channel blockers, nitrates, angiotensin-con-
six months (short-term results) and one year (inter-
verting–enzyme inhibitors, or aspirin) before or after
mediate results).4 Although our trial was small, the
surgery or at the time of discharge. Most variables
observed rates of events in the placebo group were
were distributed evenly in the two groups, and the
similar to those reported in observational studies of
variables that may not have been similar in the two
similar patients. In addition, as in nonsurgical pa-
groups, such as treatment for heart failure or diabetes,
tients,27-31 beta-blockade also had effects on nonfatal
were shown not to affect the results of the trial.
cardiac outcomes, such as myocardial infarction,
Assessing the effect of the long-term use of cardio-
congestive heart failure, and unstable angina requir-
vascular medications over the two years of the study
is critical to the analysis of the results of this trial, be-
The treatment effect in this trial cannot be attrib-
cause one interpretation might be that the patients
uted to important differences between the two study
treated with atenolol received more intensive cardio-
groups at base line; in fact, a larger proportion of the
vascular therapy than the patients given placebo,
atenolol-treated patients had cardiovascular disease
thereby confounding our findings. However, this did
before surgery, and this group had a greater number
not occur. First, the use of beta-blockers, calcium-
of risk factors known to affect the incidence of cardio-
channel blockers, nitrates, angiotensin-converting–
vascular complications after surgery.3,32-34 Our results
enzyme inhibitors, and aspirin did not differ signifi-
also cannot be explained by differences in surgical
cantly between groups 6, 12, or 24 months after
technique, details of the hospital stay, or the use of
surgery (Table 4). Nor was the use of any of these
cardiovascular medications (specifically, beta-blockers,
medications associated with any study outcome be-
December 5, 1996
fore or after surgery or at discharge. This finding is
duction), then intensive perioperative beta-blockade
not surprising, given the results of our previous two-
might give 60,000 U.S. patients each year at least an
year observational study of 474 patients with similar
additional two years of life, or save 120,000 life-years
risk profiles, in which no association was demonstrat-
(3 million surgical patients 2 percent reduction in
ed between the routine use of cardiovascular medica-
mortality 2 additional years per patient) at a cost
tions and long-term outcome.3-5 Finally, as Table 4
of less than $100 per patient (a conservative esti-
suggests, the patients in the placebo group continued
mate for one week of atenolol therapy). For the
to use these cardiovascular medications at least as of-
3 million patients at risk, the overall cost, based on
ten as the patients in the atenolol group during the
the conservative assumption, would equal $2,500
two years after hospital discharge. These results con-
per life-year saved.
firm that the observed effect of atenolol on mortality
The results of this trial thus indicate that in pa-
was not confounded by the use of these medications
tients who have or are at risk for coronary artery dis-
during the two years after discharge.
ease and who must undergo major noncardiac sur-gery, mortality and the incidence of cardiovascular
events after hospital discharge can be reduced by the
In patients at risk for coronary artery disease who
use of beta-adrenergic blockade throughout the hos-
are about to undergo major surgery, the standard
pital stay. Intensive perioperative beta-blockade ap-
practice is to control the heart rate before surgery,
pears to be safe and well tolerated, and given the
to continue beta-blocking medication up to the
availability of a generic beta-blocking agent, the es-
time of surgery, and to modulate the heart-rate re-
timated savings in lives more than outweighs the
sponse during surgery by means of anesthetic tech-
cost of therapy.
niques. After surgery, however, the heart rate is notwell controlled and rises above preoperative levels by
Supported by the Ischemia Research and Education Foundation.
30 percent or more throughout the extended post-operative period.4,9,10,24,26 Furthermore, even brief
periods of tachycardia during the postoperative pe-
The Multicenter Study of Perioperative Ischemia Research
riod may precipitate ischemia in this group of pa-
Group is a consortium of investigators from approximately 150
tients, who also are subject to alterations in perfu-
medical centers worldwide; its focus is the problems of perioper-ative myocardial infarction, stroke, and renal dysfunction (as well
sion, oxygenation, and coagulation as well as other
as other organ dysfunction) and the implications of such diseases
types of stress imposed by the exaggerated sympa-
for health economics. The Ischemia Research and Education
thetic and inflammatory responses to surgery.
Foundation is a nonprofit foundation that supports multicenter
Despite the recognition of the general problem of
research in these areas and is closely affiliated with the study in-
perioperative infarction, as well as the potentially del-
vestigators and their institutions. The coordinating analysis groupconsisted of the following: director
— D.T. Mangano; data collec-
eterious effect of an unchecked postoperative sympa-
— E. Layug, J. Li, C. Dietzel, S. Kaileh, and D.T. Mangano;
thetic response, and despite their awareness of the
— I. Tateo, E. Layug, A. Wallace, and D.T. Manga-
efficacy of beta-blockade in ambulatory patients with
no; editorial administrative assistants
— D. Beatty, B. Xavier,
coronary artery disease, clinicians have been reluc-
M. Riddle, and W. von Ehrenburg; and consultants
— S. Zhou,A. Herskowitz, W. Browner, M. Hollenberg, and K. Ziola.
tant to prescribe beta-blockers after surgery, even forpatients who were maintained on beta-blockers be-
fore their admission for surgery. Such reluctance ap-pears to be based on several areas of concern, includ-
Mangano DT. Perioperative cardiac morbidity. Anesthesiology 1990;72:
ing safety (the fear of precipitating postoperative
Mangano DT, Goldman L. Preoperative assessment of patients with
heart failure, hypotension, and bronchospasm), the
known or suspected coronary disease. N Engl J Med 1995;333:1750-6.3.
Mangano DT, Browner WS, Hollenberg M, et al. Association of peri-
efficacy of these drugs (which is unproved for surgi-
operative myocardial ischemia with cardiac morbidity and mortality in men
cal patients), and cost. Our study addressed the first
undergoing noncardiac surgery. N Engl J Med 1990;323:1781-8.
two of these issues, and our findings demonstrate
Mangano DT, Browner WS, Hollenberg M, Li J, Tateo IM. Long-term
cardiac prognosis following noncardiac surgery. JAMA 1992;268:233-9.
the efficacy and safety of perioperative beta-block-
Browner WS, Li J, Mangano DT. In-hospital and long-term mortality
ade, even for patients with a history of heart failure
in male veterans following noncardiac surgery. JAMA 1992;268:228-32.
and pulmonary disease. Regarding cost, we chose to
Rao TLK, Jacobs KH, El-Etr AA. Reinfarction following anesthesia in
patients with myocardial infarction. Anesthesiology 1983;59:499-505.
evaluate a therapy that is available in generic-drug
Gottlieb SO, Weisfeldt ML, Ouyang P, Mellits ED, Gerstenblith G. Si-
form. By conservative estimates, our study popula-
lent ischemia predicts infarction and death during 2 year follow-up of un-stable angina. J Am Coll Cardiol 1987;10:756-60.
tion represents approximately 10 percent of the 30
Siliciano D, Mangano DT. Postoperative myocardial ischemia: mecha-
million patients who undergo noncardiac surgery
nisms and therapies. In: Estafanous FG, ed. Opioids in anesthesia II. Bos-
each year (or 3 million patients). Even assuming that
ton: Butterworth–Heinemann, 1991:164-77.9.
Mangano DT, Hollenberg M, Fegert G, et al. Perioperative myocardial
atenolol has an effect only one fifth as strong as the
ischemia in patients undergoing noncardiac surgery. I. Incidence and se-
11 percent absolute reduction (from 21 percent to
verity during the 4 day perioperative period. J Am Coll Cardiol 1991;17:
10 percent) in overall long-term mortality found in
Mangano DT, Wong MG, London MJ, Tubau JF, Rapp JA, Study of
our trial (or approximately a 2 percent absolute re-
Perioperative Ischemia (SPI) Research Group. Perioperative myocardial is-
chemia in patients undergoing noncardiac surgery. II. Incidence and sever-
Modelling survival data. In: Collett D. Modelling survival data in med-
ity during the 1st week after surgery. J Am Coll Cardiol 1991;17:851-7.
ical research. London: Chapman & Hall, 1994:53-106.
Coriat P, Daloz M, Bousseau D, Fusciardi J, Echter E, Viars P. Preven-
Kaplan EL, Meier P. Nonparametric estimation from incomplete ob-
tion of intraoperative myocardial ischemia during noncardiac surgery with
servations. J Am Stat Assoc 1958;53:457-81.
intravenous nitroglycerin. Anesthesiology 1984;61:193-6.
Raby KE, Goldman L, Creager MA, et al. Correlation between preop-
Gallagher JD, Moore RA, Jose AB, Botros SB, Clark DL. Prophylactic
erative ischemia and major cardiac events after peripheral vascular surgery.
nitroglycerin infusions during coronary artery bypass surgery. Anesthesiol-
N Engl J Med 1989;321:1296-300.
Slogoff S, Keats AS. Does perioperative myocardial ischemia lead to
Stone JG, Foëx P, Sear JW, Johnson LL, Khambatta HJ, Triner L. My-
postoperative myocardial infarction? Anesthesiology 1985;62:107-14.
ocardial ischemia in untreated hypertensive patients: effect of a single small
Eisenberg MJ, London MJ, Leung JM, et al. Monitoring for myocar-
oral dose of a beta-adrenergic blocking agent. Anesthesiology 1988;68:495-
dial ischemia during noncardiac surgery: a technology assessment of trans-
esophageal echocardiography and 12-lead electrocardiography. JAMA
Magnusson J, Thulin T, Werner O, Jarhult J, Thomson D. Haemody-
namic effects of pretreatment with metoprolol in hypertensive patients un-
Yusuf S, Peto R, Lewis J, Collins R, Sleight P. Beta blockade during
dergoing surgery. Br J Anaesth 1986;58:251-60.
and after myocardial infarction: an overview of the randomized trials. Prog
Cucchiara RF, Benefiel DJ, Matteo RS, DeWood M, Albin MS. Eval-
Cardiovasc Dis 1985;27:335-71.
uation of esmolol in controlling increases in heart rate and blood pressure
The MIAMI Trial Research Group. Metoprolol in acute myocardial in-
during endotracheal intubation in patients undergoing carotid endarterec-
farction: patients and methods. Am J Cardiol 1985;56:3G-9G.
tomy. Anesthesiology 1986;65:528-31.
Pepine CJ, Cohn PF, Deedwania PC, et al. Effects of treatment
Chung F, Houston PL, Cheng DC, et al. Calcium channel blockade
on outcome in mildly symptomatic patients with ischemia during daily
does not offer adequate protection from perioperative myocardial ischemia.
life: the Atenolol Silent Ischemia Study (ASIST). Circulation 1994;90:762-
Merin RG. Calcium channel blocking drugs and anesthetics: is the
Lau J, Antman EM, Jimenez-Silva J, Kupelnick B, Mosteller F, Chalm-
drug interaction beneficial or detrimental? Anesthesiology 1987;66:111-3.
ers TC. Cumulative meta-analysis of therapeutic trials for myocardial in-
Ghignone M, Calvillo O, Quintin L. Anesthesia and hypertension: the
farction. N Engl J Med 1992;327:248-54.
effect of clonidine on perioperative hemodynamics and isoflurane require-
May GS, Eberlein KA, Furberg CD, Passamani ER, DeMets DL. Sec-
ments. Anesthesiology 1987;67:3-10.
ondary prevention after myocardial infarction: a review of long-term trials.
Talke P, Li J, Jain U, et al. Effects of perioperative dexmedetomidine in-
Prog Cardiovasc Dis 1982;24:331-52.
fusion in patients undergoing vascular surgery. Anesthesiology 1995;82:620-
Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index of
cardiac risk in noncardiac surgical procedures. N Engl J Med 1977;297:
ISIS-I (First International Study of Infarct Survival) Collaborative
Group. Randomised trial of intravenous atenolol among 16 027 cases of
Detsky AS, Abrams HB, Forbath N, Scott JG, Hilliard JR. Cardiac as-
suspected acute myocardial infarction: ISIS-I. Lancet 1986;2:57-66.
sessment for patients undergoing noncardiac surgery: a multifactorial clin-
Wallace A, Layug E, Browner W, et al. Randomized, double blinded,
ical risk index. Arch Intern Med 1986;146:2131-4.
placebo controlled trial of atenolol for the prevention of perioperative my-
Hollenberg M, Mangano DT, Browner WS, London MJ, Tubau JF,
ocardial ischemia in high risk patients scheduled for non cardiac surgery.
Tateo IM. Predictors of postoperative myocardial ischemia in patients un-
Anesthesiology 1994;81:Suppl:A99. abstract.
dergoing noncardiac surgery. JAMA 1992;268:205-9.
December 5, 1996
Biomaterials 25 (2004) 5375–5385 Stimulation of porcine bone marrow stromal cells by hyaluronan, dexamethasone and rhBMP-2 Xuenong Zoua,b,*, Haisheng Lia, Li Chenc, Anette Baatrupa, Cody B .ungera, Martin Linda a Orthopaedic Research Laboratory, Spine Section/Department of Orthopaedics, Center of Nanoscience and Biocompitability, University of Aarhus, Nørrebrogade 44, Building 1A, 8000 Aarhus C, Denmark
Yorkshire Centre for Health Informatics HL7 Summer School Capturing clinical requirements: Owen Johnson, Senior Fellow, YCHI, Leeds University Design Process Capture clinical Delegate Capturing Clinical Requirements A UML Design Process Learning Objectives • To introduce a light-weight, agile approach to UML design