Patients with coronary artery disease undergoing noncardiac surgery are at increased risk of adverse events, including myocardial infarction (MI), congestive heart failure, and death (1). Guidelines for the management of these patients emphasize preoperative clinical risk assessment coupled with targeted surgical or medical interventions (2). However, there are few strategies whose efficacy in reducing perioperative cardiac risk is well established. Risk stratification indices may be highly predictive; however, it is not clear that their use is associated with improved outcomes. The value of a preventive strategy involving percutaneous or surgical myocardial revascularization has not been tested in a prospective, controlled fashion. However, clinical trials indicate that drugs may favorably influence perioperative cardiac outcomes (3). These drugs directly or indirectly modify the relation between oxygen supply and demand in the myocardium and have established efficacy in the medical management of MI, angina, and hypertension.
Data from perioperative cardioprotection trials were examined recently in two systematic reviews. Drawing from five randomized trials in patients undergoing noncardiac surgery, Auerbach and Goldman (4) noted a benefit of β-blockade in preventing perioperative cardiac morbidity; however, several potentially relevant clinical trials involving β-blockers were omitted in this review. Nishina et al. (5), assessing the trials on perioperative clonidine, concluded that myocardial ischemia was reduced in patients undergoing both cardiac (five studies) and noncardiac surgery (two studies); of note, several pertinent clonidine trials and large-scale trials of other α2-agonists such as mivazerol were not evaluated.
Our aim in conducting this systematic review was as follows:
- To identify all randomized-controlled trials of pharmacologic cardioprotection in patients undergoing noncardiac surgery, including β-blockers, α2-agonists, calcium-channel blockers, and nitrates
- To quantitatively estimate efficacy and risk of these therapies in preventing perioperative ischemia, MI, and cardiac death
- To provide a framework for comparing the efficacy and risk of these classes of drugs.
We sought randomized trials comparing any cardioprotective drugs (experimental intervention) with an inactive control (placebo or no treatment) or another cardioprotective drug in patients undergoing noncardiac surgery. To be included, studies had to report on at least one of three end-points: perioperative myocardial ischemia, perioperative nonfatal MI, or cardiac mortality. We did not assume that observer bias was an important factor with such end-points. Whether or not a study was blinded did not constitute a criterion for inclusion or exclusion. Our interest was major noncardiac surgery. Studies dealing with cardiac surgery, percutaneous myocardial revascularization, dental procedures, electroconvulsive therapy, and diagnostic or endoscopic procedures were excluded. Further criteria for exclusion were length of observation (less than at least the entire intraoperative period), study size (<10 patients per group), treatment regimens (combinations of different drugs or treatment started more than 24 h after the end of surgery), and trials centered on cardiovascular responses during the induction of anesthesia.
We searched the Medline, Embase, and Cochrane Controlled Trials Register databases using the following terms and combinations thereof: surgery, anaesthesia or anesthesia, perioperative, postoperative, myocardial infarction, myocardial ischemia or ischaemia, β-blocker, β adrenergic blocker, α adrenergic agonist, calcium-channel blocker, angiotensin converting enzyme inhibitor, nitrate, and randomised or randomized. Generic names of individual drugs were included in the searches. The last electronic search was on November 26, 2002. References from retrieved reports and from relevant review articles and editorials were checked. Language of publication did not influence article selection. Authors of papers were contacted when results were unclear or when relevant data were not reported.
One author (RS) screened all retrieved articles, rejecting those that did not clearly meet inclusion criteria. Each potentially relevant report was then independently assessed by the other authors. Papers were given a numerical score using the three-item 5-point Oxford validity scale that evaluates adequacy of randomization (0–2 points), blinding (0–2 points), and follow-up of study participants (0–1 point) (6). The minimum score for a randomized study was 1, and the maximum score was 5. The authors convened to compare assessments and make a definitive selection of papers; differences were resolved by discussion.
Information about clinical setting, patient characteristics, treatment regimens, and length of postoperative follow-up was noted. The reported prevalence of previous MI in patients entering the study was used as an estimate of the baseline risk. Definitions of ischemia, nonfatal MI, and adverse effects were taken as reported in the original trials. When death was not reported or not characterized as cardiac, we wrote to the principal authors to clarify. Because all but two studies (7,8) assessed outcomes within 1 mo of surgery, we did not analyze data beyond that time. All data were extracted by one author (RS) and checked by the two others independently. Authors agreed on extracted data by discussion.
Heterogeneity was tested using a standard χ2 test. Because we combined only clinically homogeneous data (same clinical setting, same active drug versus same comparator, same end-points, and similar observation periods) and because, for most comparisons, there was evidence of statistically homogeneity (P > 0.1), we used a fixed effect model throughout. Because many trials reported on zero events, we chose the Peto odds ratio (OR) as an estimate of efficacy and harm; this model excludes trials that do not report on any event (9). We calculated 95% confidence intervals (CI). Numbers-needed-to-treat and harm (NNT/H), the reciprocal of the absolute risk reduction/increase, were estimated using the weighted means of the combined experimental and control event rates. Data from individual trials were displayed using both forest plots (Review Manager [RevMan] software, version 4.1, Oxford, England: The Cochrane Collaboration, 2000) to show relative efficacy of treatment effects and event rate scatters (10) to illustrate variability in incidence of outcomes. On event rates scatters, data points above the line of equality would suggest efficacy with the active intervention.
The different search strategies yielded 1093 titles, 58 of which were potentially relevant and singled out for detailed review. Thirty-seven of these were subsequently rejected: 24 did not report on relevant end-points, 4 were on diagnostic procedures (11–14) (2 were on the same cohort (13,14)), 2 were not randomized (15,16), 1 compared two nitroglycerin regimens (17), 1 had an insufficient follow-up period (18), 1 was too small (19), and 1 (20) contained the same data as another full report that had been published in the same year (21) (only one publication (20) was analyzed by us).
Two studies reported on perioperative outcomes of cohorts (22,23) that were subsequently followed up for 2 yr (7,8); because all other trials concerned events in the immediate perioperative period, the first reports only were included in our analysis (22,23). Finally, one study reported on intraoperative data (24) of patients who were subsequently followed for 10 days (25); the second report only was selected by us (25). We eventually analyzed data from 21 randomized-controlled trials (20,22,23,25–42) (Table 1).
The studies were published between 1980 and 2000; one in French (27), and all others in English. The median Oxford score was 3 (range, 1 to 5); 2 scored 1, 7 scored 2, and 4 each scored 3, 4, and 5. Six trials had a no treatment control group (21,22,30,32,41,42), and all others were placebo-controlled. No head-to-head comparisons were retrieved.
There were analyzable data on 3646 patients. In the largest study, the European Mivazerol Trial (37), 2854 patients were randomized, but the published analysis was restricted to 1897 patients with known coronary artery disease, presumably to maximize the likelihood of a treatment effect. Twelve of 21 trials (57%) involved vascular surgery. α2-adrenergic agonists (clon-idine and mivazerol) were tested in 6 trials with 2614 patients; median trial size was 179 patients (range, 29 to 1897 patients). β-blockers (atenolol, bisoprolol, esmolol, labetalol, metoprolol, oxprenolol, and propranolol) were tested in 11 trials with 866 patients; median trial size was 59 patients (range, 26 to 200 patients). Calcium-channel blockers (diltiazem and verapamil) were tested in 3 trials with 121 patients; median trial size was 30 patients (range, 25 to 66 patients). Finally, nitroglycerin was tested in one trial with 45 patients. No studies of perioperative angiotensin converting enzyme inhibitors, platelet antagonists, or HMG-CoA Reductase inhibitors that met inclusion criteria were retrieved.
Metoprolol (25,28,35) and clonidine (31,32,38,40) were the most frequently studied drugs; they were tested respectively in three and four trials but each time using different regimens. Most other drugs were tested in one trial each. Study drugs were given by various routes and at several time points (Table 1). In six trials, drugs were given as a single-dose premedication regimen (20,28,29,32,38,40). In 11 trials, duration of treatment was 2 days or longer. In 3 trials, more than one regimen of an intervention was tested (20,36,42); data from different regimens of the same drug class were pooled.
Six β-blocker and four α2-agonist trials reported on the number of patients with a history of MI before entering the study (which we used as an estimate of preoperative cardiac risk) and on the number of patients who suffered a nonfatal perioperative MI (22,23,25,26,31,36,38,40–42).
In the β-blocker trials, the prestudy prevalence of MI ranged from 0% to 52%. Graphical display suggested a correlation between the prestudy prevalence of MI and the control event rate (i.e., the incidence of nonfatal MI in controls); the prestudy prevalence of MI seemed also to correlate with the magnitude of risk reduction in patients receiving β-blockers (Fig. 1). Two β-blocker trials included high proportions of patients with preexisting MI (i.e., 32%(42) and 52%(22)); these two trials had the largest control event rates (15.8% and 15.1%), the smallest ORs (0.04 and 0.1), and largest absolute risk reductions (15.8% and 15.1%; corresponding to NNTs of 6.3 and 6.6) for nonfatal MI. In the α2-agonist trials, the prevalence of prestudy infarction ranged from 19%(40) to 24%(36) only, and a correlation between preoperative risk, perioperative MI, and cardioprotective efficacy was not observed (data not shown).
The definitions of ischemia included ST segment depressions of more than 1 mm or 1.5 mm on a continuous electrocardiographic (EKG) recording (Holter) (Table 2). Data from trials that reported on the time point ischemia had occurred (intra- or postoperative) were further analyzed.
Ten trials reported on intraoperative ischemia (20,23,28,31–33,35,36,38,40) (Fig. 2). With β-blockers, the risk decreased from 20.2% to 7.6% (OR 0.32 [95% CI, 0.17–0.58], NNT 8). With α2-agonists, it decreased from 32.8% to 19.4% (OR 0.47 [95% CI, 0.33–0.68], NNT 7). In a small trial with a 60% prestudy MI prevalence and a very frequent incidence of ischemia in controls (73.3%), the beneficial effect of diltiazem almost reached statistical significance (33).
Five trials reported on postoperative ischemia (22,23,31,34,36) (Fig. 2). With β-blockers, the risk decreased from 27.9% to 15.2% (OR 0.46 [95% CI, 0.26–0.81], NNT 8). With α2-agonists, there was no benefit. The largest treatment effect, although not statistically significant, was reported in one small trial that tested verapamil (34).
Perioperative nonfatal MI was reported in 16 studies (22,23,25,27,29–31,33,36–42) (Fig. 3), of which, 11 included a definition referring to EKG or serologic markers (Table 2). The time at which MI occurred after surgery was not clearly reported in the majority of studies. β-blockers reduced the risk of MI from 5.2% to 0.9% (OR 0.19 [95% CI, 0.08–0.48], NNT 23). When in a sensitivity analysis the two trials with the most frequent prestudy MI prevalence and extremely frequent control event rates were excluded (Fig. 1) (22,42), the incidence of nonfatal MI in controls (i.e., the control event rate) decreased to 2.4%, and the effect of β-blockers was no longer significant (OR 0.51 [95% CI, 0.14–1.89]) (Fig. 3). Five trials testing α2-agonists showed no significant effect on the rate of nonfatal MI (OR 0.85 [95% CI, 0.62–1.14]). Two small trials testing diltiazem (27) and nitroglycerin (30) reported on large but statistically not significant treatment effects.
Cardiac mortality was not clearly defined in any of the studies. To obtain additional information on cardiac deaths, we contacted authors of the 11 studies that reported that no cardiac deaths had occurred or which did not report on mortality at all; eight answered (20,25,26,29,32,38,39,42). In one of these, one cardiac death had occurred in the control group but was not reported in the original paper (26).
Overall, 7 trials (2966 patients), all testing β-blockers or α2-agonists, reported cardiac deaths (22,23,26,31,36,37,40). Seven trials (342 patients) reported that no cardiac deaths had occurred (20,25,29,32,38,39,42). Seven trials (338 patients) did not provide any information on cardiac deaths (27,28,30,33–35,41). There were 18 cardiac deaths among 1743 patients (1.0%) receiving a β-blocker or an α2-agonist compared with 41 deaths in 1565 (2.6%) controls (OR 0.42 [95% CI, 0.25–0.71]).
Eight β-blocker studies (694 patients) reported on the presence or absence of cardiac deaths (Fig. 4). There were 15 deaths in 3 trials (22,23,26) and none in 5. β-blockers significantly decreased the risk of cardiac death from 3.9% to 0.8% (OR 0.25 [95% CI, 0.09–0.73], NNT 32). One trial included a population with a particularly high baseline risk (Fig. 2) and reported on an particularly large death rate in controls (17%) (22); in this trial, the NNT to prevent one cardiac death with bisoprolol compared with no cardioprotection was 7. When this trial was not considered, the pooled effect of β blockade on cardiac mortality was no longer significant (OR 0.37 [95% CI, 0.05–2.66]).
Six α2-agonist trials (2614 patients) reported on the presence or absence of cardiac deaths (Fig. 4). There were 44 deaths in 4 trials (31,36,37,40) and none in 2. α2-agonists significantly decreased the risk of cardiac death from 2.3% to 1.1% (OR 0.50 [95% CI, 0.28–0.91], NNT 83). In the largest trial (1897 patients), the benefit with mivazerol was borderline statistically significant; death rate in controls was 2.7%, and the NNT to prevent one cardiac death was 77 (37).
Sensitivity analysis revealed that efficacy of mortality reduction was not dramatically influenced by inclusion or exclusion of studies based on the degree of information provided on cardiac death. When data from all 21 trials were included (with the assumption that no reporting of deaths meant no deaths), the NNT was 74. When only the trials that did provide information on cardiac deaths were considered, the NNT was 67.
Adverse drug reactions were reported in 19 of 21 trials and defined variably (Table 2). With β-blockers, bradycardia was defined as a heart rate <60 bpm (26,29), <50 bpm (23), <45 bpm (20), or <40 bpm (35,42); it was the most common adverse effect, occurring in 24.5% of treated patients compared with 9.1% of controls (OR 3.76 [95% CI, 2.45–5.77], NNH 6.5) (Table 3). β-blockers were also associated with an increased risk of pulmonary edema and atrioventricular block, but these results were not statistically significant. With α2-agonists, bradycardia (4.8% compared with 3.6% with controls) and hypotension (16.4% compared with 10% with controls) happened more often, although not statistically significantly. Calcium-channel blockers increased the risk of bradycardia (31.9% compared with 16.3% with control); this result was borderline significant, and the NNH was 6.4. For nitroglycerin there were not enough data to draw meaningful conclusions.
This systematic review indicates that β-blockers and α2-agonists offer significant protection against cardiac morbidity in patients undergoing major noncardiac surgery. In quantitative terms, for every 100 patients receiving a β-blocker, approximately 13 (NNT 8) will be prevented from having intra- or postoperative ischemia, approximately 4 (NNT 23) will not have a MI, and approximately 3 deaths will be prevented (NNT 32). With α2-agonists, a little less than one and a half deaths will be prevented (NNT 73). Data on β-blocker trials that we did not analyze here indicate that this survival benefit may extend months and even years beyond the perioperative period (7,8), which is consistent with extensive data on the outcome-modifying effect of β-blockers in ischemic heart disease, congestive heart failure, and hypertension.
A recent review that analyzed data from individual trials suggested that β-blockers decrease mortality in high-risk patients undergoing major noncardiac surgery (NNT 2.5 to 8.3) (4). The results presented in this study, which reflect a broader range of randomized-controlled trials, are a more conservative estimate of mortality reduction (NNT 32), possibly because (a) the pooled risk reduction reflects trials in which baseline risk was small and thus treatment effects limited, or (b) the types of β-blockers, doses, or timing of administration were less effective. Another recent systematic review of 7 randomized trials in both cardiac (5 trials) and noncardiac (2 trials) surgery concluded that clonidine reduces perioperative ischemic episodes (5). Our study assessed 6 randomized trials in which α2-agonists (clonidine in 4 studies and mivazerol in 2) were used in noncardiac surgery, demonstrating not only an effect on intraoperative ischemia, but also a diminution in cardiac mortality.
Whereas perioperative cardioprotection seems to be feasible and effective, the question arises as to how best to apply it in clinical practice. In particular, we may ask: (a) Who should be targeted? (b) What is the most effective drug? (c) Over what time period should it be given? (d) What is the optimal dose? (e) What is the risk-benefit profile of these therapies?
Taken together, our data on β-blockers, as those of others (4), indicate that outcome is most likely to be improved in higher-risk patients undergoing major surgical procedures. Patients may be selected with the help of validated risk stratification indices (43,44). Yet, even with pharmacologic intervention, a subset of patients will nevertheless sustain cardiac complications, and there are few data indicating how to differentiate those high-risk surgical patients who may be adequately managed with cardioprotective drugs alone from those in whom myocardial revascularization should be considered first. One study suggests that this type of stratification may be aided by additional diagnostic testing such as dobutamine stress echocardiography (45). Although targeting high-risk patients undergoing elective surgery seems appropriate, the usefulness of β blockade or α2-adrenergic therapy in critically ill or in patients requiring emergent surgery is unclear.
Because direct (head-to-head) comparative trials are not yet available, indirect comparisons between placebo-controlled trials must be used. Ideally, indirect comparisons should be made between populations with similar characteristics and baseline risks. β-blockers were shown to decrease the incidence of MI, but this result was dependent on the inclusion of trials with high-risk patients (22,42). α2-agonists were not tested in comparably high-risk populations, and combined data showed no significant impact on the rate of MI (Fig. 3). Based on these indirect comparisons, we cannot infer that β-blockers are a better choice. However, many patients with hypertension, coronary artery disease, or congestive heart failure are already on chronic β-blocker therapy, and continuing with the same class of drug in the perioperative period represents a pragmatic approach. α2-agonists may be of interest as an alternative therapy in patients with a contraindication to β-blocker use (for instance, documented intolerance in patients with asthma or decompensated systolic heart failure), and some anesthesiologists may prefer α2-agonists for their antinociceptive (46) and sedative (47) properties. Results with calcium-channel blockers and nitroglycerin for the prevention of ischemia and of nonfatal MI were promising; however, the numbers were too small to make meaningful conclusions.
In the analyzed trials, schedules of drug administration were highly variable. Two β-blocker studies randomized patients to receive treatment two to five weeks before surgery (22,35). Therapy was continued for a week or more after surgery in four studies (22,23,25,34). In some trials, a single dose was given immediately before surgery; in others, the drug was administered during or after surgery. One study compared pre- and postoperative drug dosing with intraoperative dosing, yet outcomes were not different between these groups (42). Such variability precluded useful inferences as to the most effective schedule. For β-blockers, it has been recommended to begin therapy at least one week before and to continue for at least two weeks after surgery (22). However, this may be impractical because patients often present for evaluation on the day before surgery. No study has assessed the optimal run-in period for these drugs and for how long they should be continued after surgery. Until more clinical data are available, prolonging therapy beyond the immediate perioperative period seems to be reasonable and is consistent with the natural history of the surgical stress response that may take days to weeks to resolve after major surgery.
We cannot make firm conclusions regarding the question of what is the most effective dose of cardioprotective drugs. Two studies randomized patients to different drug doses without demonstrating any significant difference in efficacy or harm (28,36). The relevant question is perhaps not so much the drug dose but whether or not the desired end-point is achieved. Whereas the ultimate end-point is prevention of death and MI, a practical strategy, illustrated in several of the β-blocker trials analyzed here, is to titrate therapy to a surrogate goal such as a heart rate of 50 to 60 bpm. Perioperatively, this may be achieved with parenteral forms of esmolol, metoprolol, or atenolol.
Analysis of adverse drug effects was limited by inconsistent reporting and a tendency not to explain how drug effects might have been differentiated from other sources of complications. With β-blockers, bradycardia was the only reported adverse drug reaction that occurred significantly more often than with controls. Bradycardia appeared more frequently in treatment groups with α2-agonists and with calcium-channel blockers, although these results did not reach statistical significance. Instances of other complications (hypotension, pulmonary edema, or atrioventricular block) were reported, but none was significantly associated with any drug, perhaps because of small numbers. In practice, careful titration may allow practitioners to maximize efficacy while limiting adverse effects.
Several inherent limitations of this study should be noted. The inclusion of older studies might introduce bias in particular with reference to (a) the evolution of surgical and anesthesiological techniques and (b) refinements that have been made in the detection and diagnosis of myocardial injury (such as measurement of Troponin levels).
Another limitation is the use of previous MI as a marker of preoperative or baseline cardiac risk. There was considerable heterogeneity in the way the studies enrolled patients with, and reported on, cardiovascular risk factors and on indicators of coronary artery disease. Ischemic heart disease was an unequivocal inclusion criteria in only one study (22). In our assessment, previous MI was the most consistently reported and least ambiguous indicator available. A relationship between previous MI and the rate of perioperative MI (Fig. 1) suggested that this indicator had some validity (although this association did not hold for α2-agonists). However, history of MI may not be an accurate predictor of cardiac risk because (a) a significant proportion of patients with coronary artery disease never have an MI, (b) individuals who have sustained a MI may not have any further myocardial territory at risk, and (c) previous non-Q wave MIs may be undetectable with standard screening methods (history and EKG).
A further limitation is the potential heterogeneity in the definition and reporting of efficacy end-points. Myocardial ischemia was reported in 13 of 21 trials, with definitions that were generally consistent (Table 2). Perioperative MI was reported in 16 trials and defined in 11, with no 2 definitions that were exactly the same. However, new Q waves were part of the definition in 10 of the 11 studies. Detection of myocardial injury is complicated because most perioperative events are clinically silent and because the sensitivity and specificity of objective tests, such as Holter monitoring and serologic markers in the perioperative period, are not well established. Such variability may undermine our ability to compare the studies and pool their results, yet tests of heterogeneity indicated an acceptable degree of consistency between trials (Figs. 2–4).
Finally, mortality data were difficult to interpret because many trials did not report on deaths. In at least one trial, a death occurred but was not reported in the original publication (26). Sensitivity analysis indicated that depending on whether trials with no mortality data were included or not, the NNT to prevent one cardiac death with a β-blocker or an α2-agonist was between 60 and 75. Taken individually, only the bisoprolol study by Poldermans et al. (22), which used stringent criteria to include high-risk patients, and the mivazerol study by Oliver et al. (37), which was the largest trial analyzed here, were able to show a decrease in perioperative mortality. Thus, it could be argued that the pooled results on mortality were skewed for β-blockers by a trial that studied uncommonly high-risk patients with a dramatic treatment effect and for α2-agonists by a large study that was powered to show an effect that had statistical significance but whose clinical implication is less clear.
Results of this systematic review suggest several areas for further investigation. Trials are required to directly compare therapies, for instance, a β-blocker versus an α2-agonist, or to evaluate combinations of drugs. Studies are required to determine the optimal run-in period for these drugs and the duration of required postoperative treatment. Randomized studies designed to identify patients who would benefit from a preventive strategy of pharmacologic cardioprotection as opposed to revascularization are still lacking. Finally, novel, mechanistically distinct therapeutic strategies need to be investigated. Drugs such as those studied here are directed to the adrenergic-driven myocardial oxygen supply and demand imbalance. However, the pathophysiology of perioperative myocardial ischemia may also involve inflammation, hypercoagulability, and endothelial dysfunction. Drugs such as cyclooxygenase inhibitors or statins that target these processes could be useful candidates for clinical evaluation.
We thank Daniel Haake from the Medical libraries of the Centre Medical Universitaire, Geneva University, Geneva, for his help in searching electronic databases. We are grateful to Dr. Bayliff, Professor Bonnet, Dr. Davies, Professor Foëx, Dr. Inculet, Dr. Jacobson, Dr. Quintin, Dr. Raby, and Dr. Zaugg who responded to our enquiry.
1. Mangano DT, Browner WS, Hollenberg M, et al. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery: The Study of Perioperative Ischemia Research Group. N Engl J Med 1990; 323: 1781–8.
2. Eagle KA, Berger PB, Calkins H, et al. ACC/AHA guideline update for perioperative cardiovascular evaluation for noncardiac surgery: executive summary—a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee to Update the 1996 Guidelines on Perioperative Cardiovascular Evaluation for Noncardiac Surgery). Anesth Analg 2002; 94: 1052–64.
3. Fleisher LA, Eagle KA. Clinical practice: lowering cardiac risk in noncardiac surgery. N Engl J Med 2001; 345: 1677–82.
4. Auerbach AD, Goldman L. Beta-blockers and reduction of cardiac events in noncardiac surgery: clinical applications. JAMA 2002; 287: 1445–7.
5. Nishina K, Mikawa K, Uesugi T, et al. Efficacy of clonidine for prevention of perioperative myocardial ischemia: a critical appraisal and meta-analysis of the literature. Anesthesiology 2002; 96: 323–9.
6. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996; 17: 1–12.
7. Mangano DT, Layug EL, Wallace A, et al. Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery: Multicenter Study of Perioperative Ischemia Research Group. N Engl J Med 1996; 335: 1713–20.
8. Poldermans D, Boersma E, Bax JJ, et al. Bisoprolol reduces cardiac death and myocardial infarction in high-risk patients as long as 2 years after successful major vascular surgery. Eur Heart J 2001; 22: 1353–8.
9. Yusuf S, Peto R, Lewis J, et al. Beta blockade during and after myocardial infarction: an overview of the randomized trials. Prog Cardiovasc Dis 1985; 27: 335–71.
10. L’Abbé K, Detsky AS, O’Rourke K. Meta-analysis in clinical research. Ann Intern Med 1987; 107: 224–33.
11. Korpinen R, Simola M, Saarnivaara L. Effect of esmolol on the hemodynamic and electrocardiographic changes during laryngomicroscopy under propofol-alfentanil anesthesia. Acta Anaesthesiol Belg 1998; 49: 123–32.
12. Matot I, Sichel JY, Yofe V, et al. The effect of clonidine premedication on hemodynamic responses to microlaryngoscopy and rigid bronchoscopy. Anesth Analg 2000; 91: 828–33.
13. Rosenberg J, Overgaard H, Andersen M, et al. Double blind randomised controlled trial of effect of metoprolol on myocardial ischaemia during endoscopic cholangiopancreatography. BMJ 1996; 313: 258–61.
14. Rosenberg J, Overgaard H, Anderson MM, et al. Metoprolol prevents myocardial ischemia during endoscopic retrograde cholangiopancreatography. Ugeskr Laeger 1997; 159: 5514–8.
15. Neil-Dwyer G, Walter P, Cruickshank JM. Beta-blockade benefits patients following a subarachnoid haemorrhage. Eur J Clin Pharmacol 1985; 28: 25–9.
16. Prys-Roberts C, Foëx P, Biro GP, et al. Studies of anaesthesia in relation to hypertension. V. Adrenergic beta-receptor blockade. Br J Anaesth 1973; 45: 671–81.
17. Coriat P, Baron JF, Natali J, et al. Incidence of myocardial ischemia during carotid endarterectomy. Int Angiol 1986; 5: 203–6.
18. Oxorn D, Knox JW, Hill J. Bolus doses of esmolol for the prevention of perioperative hypertension and tachycardia. Can J Anaesth 1990; 37: 206–9.
19. Talke P, Li J, Jain U, et al. Effects of perioperative dexmedetomidine infusion in patients undergoing vascular surgery. Anesthesiology 1995; 82: 620–33.
20. Stone JG, Foëx P, Sear JW, et al. Risk of myocardial ischaemia during anaesthesia in treated and untreated hypertensive patients. Br J Anaesth 1988; 61: 675–9.
21. Stone JG, Foëx P, Sear JW, et al. Myocardial ischemia in untreated hypertensive patients: effect of a single small oral dose of a beta-adrenergic blocking agent. Anesthesiology 1988; 68: 495–500.
22. Poldermans D, Boersma E, Bax JJ, et al. The effect of bisoprolol on perioperative mortality and myocardial infarction in high-risk patients undergoing vascular surgery. N Engl J Med 1999; 341: 1789–94.
23. Wallace A, Layug B, Tateo I, et al. Prophylactic atenolol reduces postoperative myocardial ischemia. Anesthesiology 1998; 88: 7–17.
24. Jakobsen CJ, Bille S, Ahlburg P, et al. Preoperative metoprolol improves cardiovascular stability and reduces oxygen consumption after thoracotomy. Acta Anaesthesiol Scand 1997; 41: 1324–30.
25. Jakobsen CJ, Bille S, Ahlburg P, et al. Perioperative metoprolol reduces the frequency of atrial fibrillation after thoracotomy for lung resection. J Cardiothorac Vasc Anesth 1997; 11: 746–51.
26. Bayliff CD, Massel DR, Inculet RI, et al. Propranolol for the prevention of postoperative arrhythmias in general thoracic surgery. Ann Thorac Surg 1999; 67: 182–6.
27. Caramella JP, Goursot G, Carcone B, et al. [Prevention of per- and postoperative myocardial ischemia in non-cardiac surgery by intravenous diltiazem]. Ann Fr Anesth Reanim 1988; 7: 245–50.
28. Coleman AJ, Jordan C. Cardiovascular responses to anaesthesia: influence of beta-adrenoreceptor blockade with metoprolol. Anaesthesia 1980; 35: 972–8.
29. Davies MJ, Dysart RH, Silbert BS, et al. Prevention of tachycardia with atenolol pretreatment for carotid endarterectomy under cervical plexus blockade. Anaesth Intensive Care 1992; 20: 161–4.
30. Dodds TM, Stone JG, Coromilas J, et al. Prophylactic nitroglycerin infusion during noncardiac surgery does not reduce perioperative ischemia. Anesth Analg 1993; 76: 705–13.
31. Ellis JE, Drijvers G, Pedlow S, et al. Premedication with oral and transdermal clonidine provides safe and efficacious postoperative sympatholysis. Anesth Analg 1994; 79: 1133–40.
32. Ghignone M, Calvillo O, Quintin L. Anesthesia and hypertension: the effect of clonidine on perioperative hemodynamics and isoflurane requirements. Anesthesiology 1987; 67: 3–10.
33. Godet G, Coriat P, Baron JF, et al. Prevention of intraoperative myocardial ischemia during noncardiac surgery with intravenous diltiazem: a randomized trial versus placebo. Anesthesiology 1987; 66: 241–5.
34. Lindgren L, Lepantalo M, von Knorring J, et al. Effect of verapamil on right ventricular pressure and atrial tachyarrhythmia after thoracotomy. Br J Anaesth 1991; 66: 205–11.
35. Magnusson J, Thulin T, Werner O, et al. Haemodynamic effects of pretreatment with metoprolol in hypertensive patients undergoing surgery. Br J Anaesth 1986; 58: 251–60.
36. Perioperative sympatholysis: beneficial effects of the alpha 2-adrenoceptor agonist mivazerol on hemodynamic stability and myocardial ischemia—McSPI-Eure Research Group. Anesthesiology 1997; 86: 346–63.
37. Oliver MF, Goldman L, Julian DG, et al. Effect of mivazerol on perioperative cardiac complications during non-cardiac surgery in patients with coronary heart disease: the Eur Mivazerol Trial (EMIT). Anesthesiology 1999; 91: 951–61.
38. Pluskwa F, Bonnet F, Saada M, et al. Effects of clonidine on variation of arterial blood pressure and heart rate during carotid artery surgery. J Cardiothorac Vasc Anesth 1991; 5: 431–6.
39. Raby KE, Brull SJ, Timimi F, et al. The effect of heart rate control on myocardial ischemia among high-risk patients after vascular surgery. Anesth Analg 1999; 88: 477–82.
40. Stühmeier KD, Mainzer B, Cierpka J, et al. Small, oral dose of clonidine reduces the incidence of intraoperative myocardial ischemia in patients having vascular surgery. Anesthesiology 1996; 85: 706–12.
41. Urban MK, Markowitz SM, Gordon MA, et al. Postoperative prophylactic administration of beta-adrenergic blockers in patients at risk for myocardial ischemia. Anesth Analg 2000; 90: 1257–61.
42. Zaugg M, Tagliente T, Lucchinetti E, et al. Beneficial effects from beta-adrenergic blockade in elderly patients undergoing noncardiac surgery. Anesthesiology 1999; 91: 1674–86.
43. Goldman L, Caldera DL, Nussbaum SR, et al. Multifactorial index of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977; 297: 845–50.
44. Lee TH, Marcantonio ER, Mangione CM, et al. Derivation and prospective validation of a simple index for prediction of cardiac risk of major noncardiac surgery. Circulation 1999; 100: 1043–9.
45. Boersma E, Poldermans D, Bax JJ, et al. Predictors of cardiac events after major vascular surgery: role of clinical characteristics, dobutamine echocardiography, and beta-blocker therapy. JAMA 2001; 285: 1865–73.
46. De Kock MF, Pichon G, Scholtes JL. Intraoperative clonidine enhances postoperative morphine patient-controlled analgesia. Can J Anaesth 1992; 39: 537–44.
47. El-Kerdawy HM, Zalingen EE, Bovill JG. The influence of the alpha2-adrenoceptor agonist, clonidine, on the EEG and on the MAC of isoflurane. Eur J Anaesthesiol 2000; 17: 105–10.