Contra-Beta bloqueador

Con: Beta-Blockers Are Indicated for All Adults at Increased Risk Undergoing Noncardiac Surgery
[Cardiovascular Anesthesia: Editorial]
London, Martin J. MD

From the Department of Clinical Anesthesia, University of California, San Francisco, San Francisco, California.
Accepted for publication May 23, 2006.
Address correspondence and reprint requests to Martin J. London, MD, Anesthesia (129), Veterans Affairs Medical Center, 4150 Clement St., San Francisco, CA 94121. Address e-mail to londonm@anesthesia.ucsf.edu.
Although anesthesiologists have long recognized the value of using [beta]-adrenergic receptor blocking drugs perioperatively to attenuate adrenergic “stressors,” it is only recently that other specialties have rallied around “perioperative [beta]-blockade” (PBB) (1). This enthusiasm is linked to the publication of two seminal but controversial reports in the New England Journal of Medicine (2–3) and the tentative recommendation for PBB by the American College of Physicians in 1997 (4). Interest in PBB has grown to a “fevered pitch” with its designation as a top tier “safety practice” by the Agency for Healthcare Research and Quality’s report (5) and has become “highly desirable” in the eyes of clinicians interested in optimizing patient outcome, hospital administrators eager to enhance their hospitals status as a provider of “safe care,” and, more recently, by administrative organizations developing performance measures for benchmarking care and reducing costs. From the onset, however, there was skepticism about PBB by clinicians and researchers trained in classical epidemiologic techniques for evaluating efficacy (e.g., results in a highly controlled setting such as the randomized clinical trial with strict inclusion and exclusion criteria) and effectiveness (e.g., results in the larger universe of clinical practice) (6).

Although efficacy (either perioperative or long-term) has been challenged by a few outspoken critics (supported in part by 2 meta-analyses), (7–10) this debate focuses on the evidence that PBB should be routinely administered to all “at-risk patients” and excludes patients already receiving [beta]-blockers or those with clear-cut indications for this therapy regardless of surgery (11). “At-risk” patients are usually considered to belong to either of 2 categories: 1) those undergoing high-risk vascular surgery with no evidence of coronary artery disease (CAD) or with stable CAD but without easily inducible ischemia, and 2) those undergoing nonvascular procedures with comorbidities predictive of CAD identified with traditional risk factors (advanced age, high total and HDL cholesterol, elevated blood pressure, cigarette smoking, family history of premature CAD, and diabetes mellitus). Although vascular surgery is recognized as producing the greatest percentage of perioperative adverse cardiac events, the latter group of patients undergoing noncardiac surgery is numerically much larger and thus, in many respects, of greatest interest.

Central to this debate are several linked questions: Just how large a problem is cardiac morbidity and mortality in patients without overt CAD (and in what types of surgery)? Do risk factors (or even overt CAD) influence short-term or longer-term (e.g., 1–2 yr) “intermediate” outcome after surgery? What does the current literature of PBB report?

With regards to the magnitude of the problem, the literature on the epidemiology of perioperative myocardial infarction (PMI) is derived primarily from investigations of patients with known prior MI and those undergoing vascular surgery (12). Both groups are recognized to have substantially higher risk over the general surgical population. Definitive reports of increased risk for PMI in patients with risk factors alone undergoing nonvascular surgery are lacking. Although accumulating evidence suggests that even low-grade postoperative “troponin leakage” has adverse implications for outcome for up to a year after surgery, even less data are available with regard to patients with CAD risk factors only (13).

Using classic Framingham predictors for CAD to risk stratify patients for PMI is problematic. Although predictive for CAD events over a timeframe measured in decades, these clinical markers are not intended for risk prediction over a period of weeks to months (14). To further confuse matters, the National Cholesterol Education Program-III considers diabetes or peripheral vascular disease as CAD equivalents (based on a 10-yr risk of a CAD event of >=20%) (14). These epidemiologic complexities have contributed to the favored use by consultants of the “revised Cardiac Risk Index”(RCRI), which has identified stronger risk factors such as overt CAD, congestive heart failure, and highest risk surgery as most predictive of adverse perioperative outcomes (15). Although the RCRI has significant limitations, a recent report suggests its predictive ability can be enhanced by incorporating age and additional surgical details (16).

This discussion of the potential efficacy of PBB is complicated not only by the issue of risk factors alone versus overt CAD but also by purported effects of PBB on longer-term outcome after surgery (e.g., 1 to 2 yr). Although anesthesiologists have traditionally focused on perioperative outcomes, extending attention to long-term events entails the probability that one is evaluating the “natural history” of the patient’s surgical indication (e.g., malignancy) or comorbidities (peripheral vascular disease, renal disease, diabetes) along with the impact of other important unmeasured factors (e.g., surgeon skill and outpatient medical care). There is currently no well-defined hypothesis as to why a short course of PBB might influence long-term outcomes leading to poorly substantiated speculation regarding perioperative inflammation and plaque stability as potential mechanisms (17).

Regardless, the existing literature of PBB centers primarily on two well-publicized studies. Mangano et al. (2) evaluated a short perioperative course (immediately before induction to up to 7 days after surgery or the time of hospital discharge) of atenolol versus placebo titrated to heart rate in 200 male veterans selected based on a history of known CAD or the presence of CAD risk factors. Although perioperative outcomes were not different between treatment groups, risk for adverse cardiac events was reduced approximately 65% the first year after surgery. In the other study, Poldermans et al. (3) reported a striking reduction (90%) in perioperative risk in a small study of 112 high-risk patients (easily inducible ischemia on preoperative dobutamine stress echo), recommending a prolonged period of preoperative and postoperative PBB.

It is important to consider the direct precursor for the atenolol trial of Mangano et al. (2), a National Institutes of Health-funded observational study of the predictors of perioperative cardiac morbidity. In this study, 454 male veterans (of which approximately 40% underwent vascular surgery and 50% had known CAD) were evaluated using perioperative Holter monitoring (2 days preoperatively, intraoperatively, and 2 days postoperatively). Postoperative myocardial ischemia occurred in 40% of patients and it imparted a ninefold increased risk of combined cardiac death, nonfatal MI, or unstable angina (18). The latter end-points occurred in only 3.2% of patients. In a subsequent 2-yr follow-up report of this cohort, 11% of patients developed major cardiovascular complications (19). Independent predictors of longer-term adverse cardiac events were known vascular disease, history of congestive heart failure (CHF), known CAD, and perioperative cardiac events that included PMI, unstable stable angina, and Holter-detected myocardial ischemia (hazards ratio, 2.2; P = 0.03) (19). Curiously, the hypotheses and sample size estimations for the subsequent atenolol study (performed at the same center) were presented as dual goals to simultaneously evaluate reduction of in-hospital “surrogate” events (hemodynamic changes, dysrhythmias, and Holter-detected myocardial ischemia) and longer-term outcome, rather than the 3.2% perioperative cardiac event rate (fatal/nonfatal MI or unstable angina) of more interest to clinicians and with the most direct physiologic rationale. However, a properly performed power analysis suggests that the latter hypothesis would require 6,000–10,000 patients. Thus, the rationale for PBB in at-risk patients (particularly the large group of nonvascular surgery patients) was never really supported by existing data, which suggested that known CAD, CHF, and vascular surgery were the major risk factors.

Regardless, in the multivariate analysis of the atenolol trial, diabetes was identified as the major risk factor for adverse long-term outcome (hazard ratio, 2.8; P = 0.01), and atenolol use was actually not a significant protective factor in this model (with a 95% confidence interval of 0.2–1.1; P value of 0.06). This finding is of interest given the preliminary report of a large (more than 900 patients) randomized trial of diabetic patients (DIPOM) that reported that PBB did not influence either perioperative or intermediate adverse cardiac outcomes in this group of patients (20). Furthermore, although Holter-detected myocardial ischemia was reduced by approximately 50% by atenolol in the trial of Mangano et al. (2), it is unclear why this reduction did not influence perioperative outcome given its role as a major prognostic factor in the original “predictors” study (21). This failure is most consistent with the well accepted truism that myocardial ischemia alone is a relatively nonspecific “surrogate outcome.”

The recent large-scale retrospective observational analysis of Lindenauer et al (22) of in-hospital mortality in over 780,000 patients at 329 United States hospitals (predominantly nonteaching facilities) in 2000 and 2001, using data obtained from a large proprietary administrative database, has generated considerable attention with regard to the findings of neutral or even adverse associations of PBB in low and at-risk only patients (22). Of the 85% of patients without contraindications to [beta]-blockers, 18% received them (tracked only during the first 2 hospital days), increasing from 14% in those with no RCRI risk factors (50% of patients) to 44% in those with >=4 risk factors (which notably were present in only <1% of patients). PBB was associated with lower mortality only in patients with 3 or more risk factors (3% of the total cohort). The most controversial findings were that in the lowest-risk patients PBB actually increased mortality. As speculated by the authors, the increased mortality may reflect the use of [beta]-blockers to treat complications, rather than as prophylaxis. Despite its large size, this study has numerous important caveats and limitations, the most important of which is its nonrandomized design. Other preliminary data by Yang et al. (23) in more than 400 patients (MAVS trial) and a recently reported peer-reviewed small (103 patients) randomized controlled trial (POBBLE) (24) found no differences in perioperative outcomes in lower-risk vascular patients with PBB. The peer reviewed results of the MAVS trial are eagerly awaited. A large ongoing multinational randomized trial (POISE) is likely to provide the most definitive data within the next few years regarding the benefits versus risk of PBB, particularly for low-risk patients (10).

To summarize, there is little evidence demonstrating that patients not already identified as having CAD or CHF and not undergoing major vascular surgery (particularly aortic or lower extremity revascularization) are at substantially increased risk for PMI. Further, there has been little effort to document a rational mechanism for the purported long-term protection afforded by a short-term course of PBB. The focus on long-term outcomes in vascular patients is driven by the observation that, in the absence of medical therapy and lifestyle modifications, these patients have a high mortality over time as a result of their underlying cardiovascular disease (25,26). In contrast to the documented benefits of [beta]-adrenergic blockers on secondary prevention in post-MI patients (e.g., prevention of a subsequent recurrent MI with enhanced long-term survival) (27) and their long-term benefits in patients with CHF (28), there is minimal, if any, evidence for primary preventive effects of [beta]-blockers alone, either on development of overt CAD or MI, in patients with risk factors only (particularly in large cohorts of patients treated for hypertension). These medical observations may be analogous to the perioperative setting. There is no debate that [beta]-blockers are a great option that can be offered to any at-risk patient undergoing major surgery. As Devereaux and Yusuf (29) emphasize, evidence-based decision-making should equally use research evidence, the clinical state, the patient’s preference, and the clinician’s expertise. Some patients will clearly be interested in therapy and others will refuse it. The skillful perioperative use of [beta]-adrenergic blockers to control hemodynamic stress is rapidly approaching, if not already established as, standard of care for all patients. Their mandatory use, especially when used as a measure of quality of care, is still a hypothesis awaiting adequate supporting data.