THERE are at least two potential pathways through which perioperative events may increase the risk of intermediate (i.e. , 1 yr or less after surgery) and long-term (i.e. , more than 1 yr after surgery) cardiovascular outcomes. First, perioperative events (e.g. , myocardial ischemia) may result in unstable coronary artery plaques that are prone to fissure and cause acute thrombosis weeks to months later. Second, perioperative myocardial infarction (MI) may result in myocardial scaring that may lead months to years later to a major cardiovascular outcome (e.g. , heart failure, cardiac arrest, cardiovascular death).
Unfortunately, no direct imaging or molecular studies are available for evaluating whether perioperative events result in unstable coronary artery plaques. Less direct evidence that supports the first hypothesized pathway comes from three small prospective studies.1–3Wallace et al. undertook a nested cohort study within a 200-patient perioperative β-blocker trial.1This study demonstrated that perioperative myocardial ischemia (detected on Holter electrocardiography) on postoperative days 0–2 was a univariate predictor of 2-yr mortality (36 patients died; relative risk 2.06; 95% confidence interval [CI] 1.04–4.06).
Pasternack et al. undertook a prospective cohort study of 385 patients.2Logistic regression demonstrated that only total perioperative percentage time ischemic of 1% or more (based on continual electrocardiography monitoring for an average of 31 h after surgery) and age were statistically significant independent predictors of cardiovascular outcomes (44 patients died, and 17 suffered MI during 2-yr follow-up; estimates of association were not reported).2
Mangano et al. undertook a prospective cohort study of 444 consecutive patients with or at high-risk of coronary artery disease who were discharged home after surgery.3During the 2-yr follow-up, 47 patients suffered cardiac complications as defined by a broad composite that included cardiac death and nonfatal MI. Multivariable analysis demonstrated that postoperative myocardial ischemia (detected on Holter electrocardiography) was an independent predictor of long-term cardiac complications (hazard ratio 2.2; 95% CI 1.1–4.3).3
Several studies support the second hypothesized pathway that perioperative MI may lead to a major cardiovascular outcome months to years later. Five small studies (total of 753 patients) all demonstrated that an elevated troponin measurement after surgery was a statistically significant independent predictor of mortality (total of 98 deaths) within 1 yr of surgery.4–8Two small studies (total of 840 patients) both demonstrated that an elevated troponin measurement after surgery was a statistically significant independent predictor of mortality (total of 162 deaths) up to 4 yr after surgery.9,10The prospective cohort study by Mangano et al. also demonstrated that a perioperative MI was an independent predictor of a cardiac complication (hazard ratio, 20.0, 95% CI 7.5–53.0) at 2-yr follow-up.3Finally, a large (105,951 patients) Veterans Affairs study that used prospective and administrative data demonstrated that a perioperative MI was an independent predictor of 8-yr mortality (37,743 deaths, hazard ratio 1.5, 95% CI 1.4–1.6).11
There are at least two potential explanations for these study results. First, the groups of patients (e.g. , patients with and without perioperative myocardial ischemia) in each study had a similar extent of cardiovascular disease and a similar risk of subsequent events before surgery, and the occurrence of the perioperative event changed the patients’ long-term prognosis. A second potential explanation is that perioperative myocardial ischemia and infarction are markers of more severe underlying cardiovascular disease and thus a worse prognosis.
If the first explanation is correct, then preventing perioperative myocardial ischemia or MI may prevent intermediate or long-term cardiovascular outcomes; if the second explanation solely accounts for the demonstrated associations, then preventing perioperative myocardial ischemia or MI is unlikely to affect distant events.
Although it is not possible to draw firm conclusions on the basis of the current evidence, the consistency of the signal in the adjusted analyses across the perioperative studies and the strong evidence that myocardial ischemia and MI alter intermediate and long-term prognosis in the nonoperative setting suggests that perioperative events independently alter intermediate and long-term cardiovascular prognosis. Therefore, exploring whether interventions that prevent perioperative myocardial ischemia or MI result in a decrease in intermediate or long-term cardiovascular complications is warranted.
Can Perioperative β-blockers Prevent Intermediate or Long-term Cardiovascular Outcomes
Before considering the intermediate and long-term impact of administering a β-blocker around the time of noncardiac surgery, it is relevant to determine whether a β-blocker can attenuate the perioperative stress response. A meta-analysis of high quality β-blocker randomized controlled trials (RCTs) among patients undergoing noncardiac surgery demonstrated at 30-day follow-up a lower rate of myocardial ischemia among patients assigned a β-blocker (43 of 1,059 patients) compared to control (76 of 1,059 patients, odds ratio 0.42, 95% CI 0.27–0.65, I220%), and a lower rate of nonfatal MI among patients assigned a β-blocker (174 of 5,610 patients) compared to control (240 of 5,426 patients, odds ratio 0.72, 95% CI 0.59-0.87, I20%).12
Five trials have reported whether the favorable perioperative effects of a β-blocker translate into intermediate or long-term cardiovascular benefits (table 1).13–17Only one small trial (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography Study [DECREASE]) with few events (112 patients, with 41 patients experiencing the primary outcome) and methodological limitations (unblinded and recruitment was stopped early for an unexpected very large treatment effect)18demonstrated a statistically significant long-term benefit with β-blocker therapy.14Unlike all the other trials, patients in this trial continued the study drug during the long-term follow-up14; therefore, DECREASE addresses a different question than the other trials that evaluated the intermediate or long-term effects of a β-blocker only given around the time of surgery.
The largest trial (Diabetic Postoperative Mortality and Morbidity [DIPOM], 921 patients with 192 patients experiencing the primary outcome) demonstrated no effect on major cardiovascular outcomes at 18 months of follow-up. Although some authors have suggested that the difference in results of the β-blocker trials is the result of variations in the dosing (i.e. , high dose vs. low dose), the current evidence does not support this perspective (table 1); it is more likely that the differences relate to chance and methodological quality (i.e. , the high-quality trials demonstrate a consistent signal).
Although the current trials do not provide encouraging evidence that a perioperative β-blocker affects intermediate or long-term cardiovascular outcomes, there is still a limited amount of data. The PeriOperative ISchemic Evaluation (POISE) Trial (perioperative extended-release metoprolol succinate with a target dose of 50% of the maximum daily therapeutic dose vs. placebo) included 8,351 patients and will report the 1-yr follow-up data next year; 22 countries have completed their direct patient follow-up, and Canada will complete its 1-yr follow-up through its national databases in 2010.19If POISE demonstrates a benefit from a perioperative β-blocker at 1 yr, clinicians and patients will have to balance this benefit against the 30-day excess of death and stroke with a β-blocker, as demonstrated in POISE and the high-quality RCTs.12,20
The prior perioperative β-blocker discussion, except for a few patients in the trial by Mangano et al. , relates to patients who were not taking chronic β-blocker therapy before surgery. Therefore, these trials do not inform the intermediate or long-term effects of continuing, withholding, or titrating β-blockers around the time of noncardiac surgery among patients who have a history of taking a β-blocker chronically. Potentially relevant issues to the short, intermediate, and long-term effects include the following: the potential exacerbation of cardiac ischemia that may occur from stopping a β-blocker acutely before a patient undergoes surgery, and the β-blocker dose that is safe in the nonoperative setting may still exacerbate clinically significant hypotension after surgery and result in the negative consequences demonstrated in POISE.20Until a large high-quality trial is undertaken to directly inform this issue, physicians will have to use indirect evidence to individualize the perioperative management of each patient who is chronically on a β-blocker.
Can Perioperative α2Agonists Prevent Intermediate and Long-term Cardiovascular Outcomes
RCT evidence suggests that α2agonists can attenuate the perioperative stress response (e.g. , reduce perioperative myocardial ischemia).21,22Wallace et al. undertook an RCT evaluating the effect of 4 days of perioperative clonidine in patients undergoing noncardiac surgery.23Clonidine demonstrated an absolute risk reduction of 5.4% for mortality at 30 days (total of 5 deaths, P = 0.048) and an absolute risk reduction of 14% for mortality at 2 yr (total of 38 deaths, P = 0.035). These encouraging but limited data (Wallace is the only clonidine trial that followed patients beyond 30 days) highlight the need for further RCTs to examine whether perioperative clonidine reduces long-term mortality.
Conclusion
Perioperative cardiovascular events appear to affect intermediate and long-term cardiovascular outcomes. The current β-blocker evidence is not encouraging, but we will have more data in 2010. Although the clonidine evidence is encouraging, there is a need for confirmatory trials. Considering that globally 200 million adults undergo noncardiac surgery annually highlights why there is an urgent need for large high-quality RCTs to establish ways to ensure that patients obtain the benefits of their noncardiac surgery without suffering a major cardiovascular outcome that compromises their quality or duration of life in the short, intermediate, or long-term.
Departments of Clinical Epidemiology and Biostatistics and Medicine, Michael G. DeGroote School of Medicine, McMaster University, Hamilton, Ontario, Canada. philipj@mcmaster.ca
