Correspondence
Effect of Atenolol on Mortality and Cardiovascular Morbidity after Noncardiac Surgery
N Engl J Med 1997; 336:1452-1454May 15, 1997
- Article
To the Editor:
Dr. Mangano and other investigators in the Multicenter Study of Perioperative Ischemia Research Group (Dec. 5 issue)1 have failed to report directly the outcome of the eight patients in the placebo group who were taking beta-blockers preoperatively. There were 12 deaths from cardiac causes in the placebo group. If all eight patients who were taking beta-blockers preoperatively died from cardiac events, then what the authors have really found is that the sudden perioperative withdrawal of beta-blockers is deleterious. This is quite possible, given the well-supported position, which the authors promote, that the effects of tachycardia and hypertension are deleterious. The fact that many of the patients in the atenolol group were also receiving beta-blockers does not compensate for this type of bias, since those patients never underwent short-term perioperative withdrawal of the beta-blockers but, rather, were maintained on this therapy throughout hospitalization by virtue of receiving the drug rather than the placebo.
1 ReferencesJohn A. Petros, M.D.
Emory University School of Medicine, Atlanta, GA 303221
Mangano DT ,Layug EL ,Wallace A ,Tateo I . Effect of atenolol on mortality and cardiovascular morbidity after noncardiac surgery. N Engl J Med 1996;335:1713-1720
Full Text | Web of Science | Medline
To the Editor:
Mangano and colleagues claim that “in patients who have or are at risk for coronary artery disease who must undergo noncardiac surgery, treatment with atenolol during hospitalization can reduce mortality and the incidence of cardiovascular complications for as long as two years after surgery.” That the rate of use of angiotensin-converting–enzyme (ACE) inhibitors was three times as high in the atenolol group as in the placebo group before admission (23.7 percent vs. 8.2 percent) and at hospital discharge (20.4 percent vs. 6.1 percent) may suggest another explanation for the observed decrease in postoperative cardiovascular complications. Although the authors state that the odds ratios for death at two years associated with the use of ACE inhibitors before admission and after discharge were not significant, it would be more helpful to calculate odds ratios for the six-to-eight-month postoperative period, when the greatest difference in mortality due to cardiovascular causes occurred.
Timothy Ho, M.R.C.P.
David Taussig, M.R.C.P.
Gregory McAnulty, F.R.C.A.
St. George's Hospital, London SW17 0QT, United KingdomTo the Editor:
Dr. Mangano and colleagues should be commended for taking research in perioperative medicine beyond the focus of risk assessment. In the 20 years since the first multifactorial index was developed to assess cardiac risk associated with noncardiac surgery,1 researchers have thoroughly studied perioperative risk assessment, but there has been a paucity of studies on risk modification.
Mangano et al. have appropriately used a randomized, controlled design to assess the value of beta-blockers given perioperatively. Despite this strong study design, we are concerned about some possible areas of bias. Randomization yielded potentially important differences between the treatment and placebo groups, perhaps because the authors did not stratify patients according to whether they had definite coronary artery disease or were at risk for coronary artery disease (both were inclusion criteria). The authors state that “there were no significant differences between the groups” except for the rate of treatment for hypertension. However, they are referring to statistical significance, which is not the pertinent issue, since we already know that all differences between the groups arose by chance in the randomization process. The more important issue is the magnitude of the differences between the treatment groups and the clinical significance of these differences.2 There was a 6 percent absolute difference in the proportions with a history of definite coronary artery disease and an 8 percent absolute difference in the rate of prior myocardial infarction. Both these differences favored the treatment group, and may have been clinically important. With only 100 patients in each group, the authors should not have relied on statistical significance alone.
Likewise, in the multivariate analysis, Mangano et al. used the statistical significance of univariate associations to guide the selection of variables. By using a criterion of P<0.10 for the selection of variables, the authors excluded a large number of potentially important confounders,3 especially such variables as prior myocardial infarction and known coronary artery disease.
This study is certainly an important contribution, and its results are encouraging, but we want to caution readers about the issues just discussed.
3 ReferencesP.J. Devereaux, M.D.
William A. Ghali, M.D., M.P.H.
University of Calgary, Calgary, AB T2N 4N1, Canada1
Goldman L ,Caldera DL ,Nussbaum SR , et al. Multifactorial index of cardiac risk in noncardiac surgical procedures. N Engl J Med 1977;297:845-850
Full Text | Web of Science | Medline2
Guyatt GH ,Sackett DL ,Cook DJ . Users' guides to the medical literature. II. How to use an article about therapy or prevention. A. Are the results of the study valid? JAMA 1993;270:2598-2601
CrossRef | Web of Science | Medline3
Rothman KJ. Modern epidemiology. Boston: Little, Brown, 1986:302.
To the Editor:
Mangano and colleagues concluded that in patients who have or are at risk for coronary artery disease and who are undergoing noncardiac surgery, mortality and cardiovascular events after discharge can be substantially reduced by in-hospital atenolol therapy. However, it is unclear whether this finding is generalizable to the entire population at risk. Specifically, the authors do not report the proportion of patients who were women, although they imply that the study population included both men and women at their Veterans Affairs hospital. In addition, they do not comment on the sex-specific incidence of events or the influence of sex on the effects of atenolol therapy. Indeed, if adequate numbers of women were enrolled in their study, the authors may make an important contribution to the identification of sex-specific prognostic variables and preventive strategies for high-risk patients undergoing noncardiac surgery.
Steven E. Reis, M.D.
Adam H. Feldman, M.D.
University of Pittsburgh Medical Center, Pittsburgh, PA 15213Author/Editor Response
Dr. Mangano replies:
To the Editor: Our study identified an important therapy for reducing mortality after noncardiac surgery — namely, the perioperative use of beta-adrenergic blockade with atenolol.
Dr. Petros addresses the important issue of the discontinuation of beta-blockers before surgery.1 In our study of 200 randomized patients, 26 (13 percent) received beta-blockers preoperatively, and these medications were continued until the morning of surgery, then replaced by study drug (placebo or atenolol). In the placebo group, 12 patients died from cardiac causes over two years, with 1 death (12 percent) among the 8 patients taking beta-blockers preoperatively and 11 deaths (12 percent) among the 91 patients not taking beta-blockers. Likewise, in the atenolol group, 1 patient died of cardiac causes (6 percent) among the 18 patients receiving beta-blockers, as compared with 3 of 75 patients (4 percent) not receiving beta-blockers. These findings are consistent with previous results,2-4 indicating that our findings were not explained by the discontinuation of beta-blockers.
Ho et al. raise similar questions regarding the use of ACE inhibitors. Although our results suggest differences between the study groups with respect to the use of these drugs before surgery and at discharge, there was no suggestion of an association between such use and mortality at two years (Table 3 of our article). Regarding our six-month data, the estimated odds ratios for the association between six-month mortality and the use of ACE inhibitors preoperatively or at discharge are 0.29 (P = 0.21) and 0.35 (P = 0.25), respectively. The low odds ratios suggest a protective effect; however, given that there were only eight deaths at six months (all in the placebo group), no conclusion should be drawn regarding the association between the use of ACE inhibitors and six-month mortality.
In response to Drs. Devereaux and Ghali: neither a history of definite coronary artery disease nor prior myocardial infarction had a significant confounding effect on the relation between the study drug and mortality at six months, one year, or two years — findings consistent with those of studies in similar cohorts.2-4 Regarding our use of the P<0.10 criterion for multivariable modeling, we agree that using a less stringent criterion, such as P <0.20, would have allowed the inclusion of a greater number of potential predictors (in our case, two additional univariable predictors). However, when we used this less stringent criterion, our final multivariable model and associated conclusions did not change.
Drs. Reis and Feldman question the generalizability of our findings to other groups. Formally, such questions require prospective assessment in randomized, controlled trials. However, even if such trials were conducted, their results would not be available for two to four years, and given our compelling findings, the extensive experience with the use of beta-blockers in women, and recent results in surgical patients,5 we believe that there is rationale for aggressive control of the perioperative heart rate for all high-risk patients (women and men) and that such an approach should be implemented sooner, rather than two to four years later.
5 ReferencesDennis T. Mangano, Ph.D., M.D.
San Francisco Veterans Affairs Medical Center, San Francisco, CA 941211
Mangano DT . Perioperative cardiac morbidity. Anesthesiology 1990;72:153-184
CrossRef | Web of Science | Medline2
Mangano DT ,Browner WS ,Hollenberg M , et al. Association of perioperative myocardial ischemia with cardiac morbidity and mortality in men undergoing noncardiac surgery. N Engl J Med 1990;323:1781-1788
Full Text | Web of Science | Medline3
Mangano DT ,Browner WS ,Hollenberg M ,Li J ,Tateo IM . Long-term cardiac prognosis following noncardiac surgery. JAMA 1992;268:233-239
CrossRef | Web of Science | Medline4
Browner WS ,Li J ,Mangano DT . In-hospital and long-term mortality in male veterans following noncardiac surgery. JAMA 1992;268:228-232
CrossRef | Web of Science | Medline5
McSPI-Europe Research Group
CrossRef | Web of Science | Medline
- Citing Articles (2)
Citing Articles
1
Steven L. Cohn, Andrew D. Auerbach. (2007) Preoperative cardiac risk stratification 2007: Evolving evidence, evolving strategies. Journal of Hospital Medicine 2:3, 174-180
CrossRef2
Elizabeth A. Martinez, Peter Pronovost. (2002) Perioperative β-blockers in high-risk patients. Journal of Critical Care 17:2, 105-113
CrossRef
