Perioperative Beta-Blocker Therapy and Mortality
http://www.100md.com
《新英格兰医药杂志》
To the Editor: The large observational study by Lindenauer et al. (July 28 issue)1 seems to confirm that the administration of beta-blockers improves the postoperative outcome of very high-risk patients undergoing major surgery. Owing to the nonrandomized design of the study, the authors cannot exclude the possibility that other cardioprotective strategies — including prolonged monitoring, aggressive pain control and rehabilitation, intensive insulin therapy, and continuation of lipid-lowering agents — could have been used more frequently or adequately for the patients at highest risk.2,3
On the other hand, the potential harmful effects of beta-blockers raise several questions, including why almost 20 percent of patients at "no risk" or low risk received such medications perioperatively. Since no data on the cardiac or noncardiac causes of in-hospital mortality are available, we should consider that inappropriate use of beta-blockers can be associated with serious adverse effects, including reduced delivery of oxygen to tissue, which could lead to postoperative organ failure and death. Such a hypothesis is supported by the results of numerous randomized trials indicating that perioperative optimization of oxygen delivery significantly reduces mortality and morbidity among high-risk surgical patients.4,5
Benoit Vandenbunder, M.D.
Alexandre Mignon, M.D., Ph.D.
H?pital Cochin, Assistance Publique–H?pitaux de Paris
75014 Paris, France
alexandre.mignon@cch.aphp.fr
References
Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med 2005;353:349-361.
Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345:1359-67.
Lindenauer PK, Pekow P, Wang K, Gutierrez B, Benjamin EM. Lipid-lowering therapy and in-hospital mortality following major noncardiac surgery. JAMA 2004;291:2092-2099.
Boyd O, Grounds RM, Bennett ED. A randomized clinical trial of the effect of deliberate perioperative increase of oxygen delivery on mortality in high-risk surgical patients. JAMA 1993;270:2699-2707.
Davies SJ, Wilson RJT. Preoperative optimization of the high-risk surgical patient. Br J Anaesth 2004;93:121-128.
To the Editor: In the study by Lindenauer et al., the most striking finding was the paradox that the excess mortality linked to prophylactic beta-blockade (an increase of 43 percent) in low-risk patients (with a Revised Cardiac Risk Index of 0 or 1) was as great as the survival benefit in high-risk patients (with an RCRI of 2 or more). Two key methodologic flaws cast doubt on the premise that these associations are causal: the lack of any data on preoperative long-term beta-blockade and the arbitrary distinction between beta-blockade administered during the first two hospital days (prophylactic group) and beta-blockade administered later (treatment or control group).
Patients who receive long-term beta-blocker therapy are likely to be sicker than patients who do not receive this therapy. In addition, even if beta-blocker therapy is continued in the hospital, patients who receive long-term therapy may have complications of acute withdrawal related to operations associated with large fluid shifts and blood loss or as a result of reduced gastrointestinal absorption.1 Thus, worse survival in the low-risk group may reflect long-term, rather than short-term, use of beta-blockers perioperatively.
Patients who receive beta-blockers for tachycardia or ischemia on the third hospital day or later are sicker by definition than are patients who do not receive such treatment. The inclusion of patients receiving extended beta-blockade in the control group does not necessarily imply that perioperative beta-blockade improved outcome.
Peter F. Dunn, M.D.
Giora Landesberg, M.D., D.Sc.
Massachusetts General Hospital
Boston, MA 02114
pdunn@partners.org
References
Giles JW, Sear JW, Foex P. Effect of chronic beta-blockade on peri-operative outcome in patients undergoing non-cardiac surgery: an analysis of observational and case control studies. Anaesthesia 2004;59:574-583.
To the Editor: Lindenauer et al. used the RCRI to stratify patients who received perioperative beta-blocker therapy according to risk. They concluded that beta-blocker therapy was of no benefit or even detrimental in low-risk patients.
We disagree with the exclusion of patients with congestive heart failure from this study. The authors stated that the basis for this decision was that beta-blocker therapy could be considered contraindicated in patients with congestive heart failure. Since beta-blocker use in patients with stable congestive heart failure is a class I indication (according to the criteria of the American College of Cardiology–American Heart Association), presumably many of these patients would have been receiving beta-blockers before hospitalization. Furthermore, congestive heart failure is one of the six criteria that constitute the RCRI. The exclusion of patients with congestive heart failure resulted in risk stratification that was not truly consistent with the RCRI. Also, since these patients benefit from beta-blockade, their exclusion from the study may have dampened the treatment effect in low-risk patients, leading to a type II error.
Henry A. Tran, M.D.
New York University
New York, NY 10016
tranh01@med.nyu.edu
Paul H. Tran, M.D.
Eastern Virginia Medical School
Norfolk, VA 23507
The authors reply: A primary aim of our study was to assess the quality of care by evaluating the use of beta-blockers among ideal patients. We excluded those with heart failure because we did not know whether patients had tolerated beta-blockers in the outpatient setting and did not expect that on the day of surgery physicians would initiate treatment in patients who may not have received previous treatment. The perioperative administration of beta-blockers may benefit patients with heart failure, but the results of our analysis cannot be applied to this population. Although this fact limits the generalizability of our results, we would not consider this an issue of power or a type II error, as Drs. Tran and Tran suggest.
Drs. Dunn and Landesberg express concern that our lack of information about long-term beta-blocker use may have led us to conclude erroneously that beta-blockers were harmful in low-risk patients. Although we were careful to address this limitation in our report, it should be noted that among the low-risk patients who were most likely to have been long-term users of beta-blockers (i.e., those with documented hypertension), the perioperative administration of beta-blockers did not appear harmful. The higher mortality we reported in the group of low-risk patients without hypertension who received beta-blockers may have been attributable to differential underdocumention of risk factors (including hypertension), which would not have been incorporated into our models. Furthermore, among the patients who did not appear to be candidates for prophylactic therapy, beta-blockers may have been prescribed as treatment for an early complication. We chose to use the second hospital day as a cutoff when defining prophylaxis because administrative data sources count days at midnight, and the first hospital "day" can be as short as one hour if a patient is admitted at 11 p.m.
We agree with Drs. Vandenbunder and Mignon that we could not exclude the possibility of differences in care between the two groups of patients. Although we lacked information about the quality of perioperative monitoring and pain control, we included lipid-lowering therapy, insulin and oral hypoglycemic medications, antibiotic prophylaxis, and measures to prevent venous thromboembolism in our models. As described above, there are several potential explanations for the increased mortality we observed among the low-risk patients without hypertension. The possibility of harm stemming from beta-blocker use — and potential explanations for this harm, as highlighted by Drs. Vandenbunder and Mignon — must also be considered.
Peter K. Lindenauer, M.D.
Penelope Pekow, Ph.D.
Evan M. Benjamin, M.D.
Baystate Medical Center
Springfield, MA 01199
peter.lindenauer@bhs.org
On the other hand, the potential harmful effects of beta-blockers raise several questions, including why almost 20 percent of patients at "no risk" or low risk received such medications perioperatively. Since no data on the cardiac or noncardiac causes of in-hospital mortality are available, we should consider that inappropriate use of beta-blockers can be associated with serious adverse effects, including reduced delivery of oxygen to tissue, which could lead to postoperative organ failure and death. Such a hypothesis is supported by the results of numerous randomized trials indicating that perioperative optimization of oxygen delivery significantly reduces mortality and morbidity among high-risk surgical patients.4,5
Benoit Vandenbunder, M.D.
Alexandre Mignon, M.D., Ph.D.
H?pital Cochin, Assistance Publique–H?pitaux de Paris
75014 Paris, France
alexandre.mignon@cch.aphp.fr
References
Lindenauer PK, Pekow P, Wang K, Mamidi DK, Gutierrez B, Benjamin EM. Perioperative beta-blocker therapy and mortality after major noncardiac surgery. N Engl J Med 2005;353:349-361.
Van den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345:1359-67.
Lindenauer PK, Pekow P, Wang K, Gutierrez B, Benjamin EM. Lipid-lowering therapy and in-hospital mortality following major noncardiac surgery. JAMA 2004;291:2092-2099.
Boyd O, Grounds RM, Bennett ED. A randomized clinical trial of the effect of deliberate perioperative increase of oxygen delivery on mortality in high-risk surgical patients. JAMA 1993;270:2699-2707.
Davies SJ, Wilson RJT. Preoperative optimization of the high-risk surgical patient. Br J Anaesth 2004;93:121-128.
To the Editor: In the study by Lindenauer et al., the most striking finding was the paradox that the excess mortality linked to prophylactic beta-blockade (an increase of 43 percent) in low-risk patients (with a Revised Cardiac Risk Index of 0 or 1) was as great as the survival benefit in high-risk patients (with an RCRI of 2 or more). Two key methodologic flaws cast doubt on the premise that these associations are causal: the lack of any data on preoperative long-term beta-blockade and the arbitrary distinction between beta-blockade administered during the first two hospital days (prophylactic group) and beta-blockade administered later (treatment or control group).
Patients who receive long-term beta-blocker therapy are likely to be sicker than patients who do not receive this therapy. In addition, even if beta-blocker therapy is continued in the hospital, patients who receive long-term therapy may have complications of acute withdrawal related to operations associated with large fluid shifts and blood loss or as a result of reduced gastrointestinal absorption.1 Thus, worse survival in the low-risk group may reflect long-term, rather than short-term, use of beta-blockers perioperatively.
Patients who receive beta-blockers for tachycardia or ischemia on the third hospital day or later are sicker by definition than are patients who do not receive such treatment. The inclusion of patients receiving extended beta-blockade in the control group does not necessarily imply that perioperative beta-blockade improved outcome.
Peter F. Dunn, M.D.
Giora Landesberg, M.D., D.Sc.
Massachusetts General Hospital
Boston, MA 02114
pdunn@partners.org
References
Giles JW, Sear JW, Foex P. Effect of chronic beta-blockade on peri-operative outcome in patients undergoing non-cardiac surgery: an analysis of observational and case control studies. Anaesthesia 2004;59:574-583.
To the Editor: Lindenauer et al. used the RCRI to stratify patients who received perioperative beta-blocker therapy according to risk. They concluded that beta-blocker therapy was of no benefit or even detrimental in low-risk patients.
We disagree with the exclusion of patients with congestive heart failure from this study. The authors stated that the basis for this decision was that beta-blocker therapy could be considered contraindicated in patients with congestive heart failure. Since beta-blocker use in patients with stable congestive heart failure is a class I indication (according to the criteria of the American College of Cardiology–American Heart Association), presumably many of these patients would have been receiving beta-blockers before hospitalization. Furthermore, congestive heart failure is one of the six criteria that constitute the RCRI. The exclusion of patients with congestive heart failure resulted in risk stratification that was not truly consistent with the RCRI. Also, since these patients benefit from beta-blockade, their exclusion from the study may have dampened the treatment effect in low-risk patients, leading to a type II error.
Henry A. Tran, M.D.
New York University
New York, NY 10016
tranh01@med.nyu.edu
Paul H. Tran, M.D.
Eastern Virginia Medical School
Norfolk, VA 23507
The authors reply: A primary aim of our study was to assess the quality of care by evaluating the use of beta-blockers among ideal patients. We excluded those with heart failure because we did not know whether patients had tolerated beta-blockers in the outpatient setting and did not expect that on the day of surgery physicians would initiate treatment in patients who may not have received previous treatment. The perioperative administration of beta-blockers may benefit patients with heart failure, but the results of our analysis cannot be applied to this population. Although this fact limits the generalizability of our results, we would not consider this an issue of power or a type II error, as Drs. Tran and Tran suggest.
Drs. Dunn and Landesberg express concern that our lack of information about long-term beta-blocker use may have led us to conclude erroneously that beta-blockers were harmful in low-risk patients. Although we were careful to address this limitation in our report, it should be noted that among the low-risk patients who were most likely to have been long-term users of beta-blockers (i.e., those with documented hypertension), the perioperative administration of beta-blockers did not appear harmful. The higher mortality we reported in the group of low-risk patients without hypertension who received beta-blockers may have been attributable to differential underdocumention of risk factors (including hypertension), which would not have been incorporated into our models. Furthermore, among the patients who did not appear to be candidates for prophylactic therapy, beta-blockers may have been prescribed as treatment for an early complication. We chose to use the second hospital day as a cutoff when defining prophylaxis because administrative data sources count days at midnight, and the first hospital "day" can be as short as one hour if a patient is admitted at 11 p.m.
We agree with Drs. Vandenbunder and Mignon that we could not exclude the possibility of differences in care between the two groups of patients. Although we lacked information about the quality of perioperative monitoring and pain control, we included lipid-lowering therapy, insulin and oral hypoglycemic medications, antibiotic prophylaxis, and measures to prevent venous thromboembolism in our models. As described above, there are several potential explanations for the increased mortality we observed among the low-risk patients without hypertension. The possibility of harm stemming from beta-blocker use — and potential explanations for this harm, as highlighted by Drs. Vandenbunder and Mignon — must also be considered.
Peter K. Lindenauer, M.D.
Penelope Pekow, Ph.D.
Evan M. Benjamin, M.D.
Baystate Medical Center
Springfield, MA 01199
peter.lindenauer@bhs.org