Introduction
Perioperative beta-blockade (PBB) has been advocated for the reduction of cardiac risk for noncardiac surgery. The initial interest was based on nearly 50 years of research in the cardiology literature documenting the cardioprotective effects of beta-blockers.
The primary role for beta-blockers in the perioperative setting is for the prevention of major adverse cardiac events (MACE). These adverse cardiac events account for up to 40% of all perioperative mortality. Additionally, perioperative myocardial ischemia has been associated with a significant increased risk of nonfatal myocardial infarction (MI), as well as cardiovascular death for up to 6 months after surgery.
To understand the role of PBBs in preventing MACE, it is important to understand the etiology of perioperative myocardial infarction (PMI). PMI is often preceded by prolonged tachycardia with ST depression–type ischemia and generally develops into non–Q-wave infarction with the resting electrocardiogram subsequently returning to baseline. Thus PMI has traditionally been ascribed mostly to prolonged stress-induced ischemia in the setting of fixed coronary stenosis; only a small percentage has related to acute plaque rupture. Given these assumptions, the natural role of beta-blockers in preventing PMI has been seen as improving myocardial oxygen balance by slowing the heart rate, reducing contractility, and improving diastolic coronary filling, thereby decreasing myocardial oxygen consumption.
Simply improving the balance of oxygen supply and demand is not the only benefit to PBBs, as the attenuation of perioperative hemodynamic stress can help prevent rupture or fissuring on the intimal surface of a vulnerable plaque. There are several other pathologic effects of perioperative stress and inflammation that PBBs cannot readily modify. The perioperative milieu can promote thrombosis by increasing platelet activity and decreasing fibrinolysis, as well as cause endothelial coronary vasoconstriction and further plaque destabilization. This more complex nature of PMI provides a rationale for why perioperative ischemia does not consistently lead to PMI and why beta-blockers may not affect the incidence of PMI or perioperative mortality in some patients, despite reduction of perioperative ischemia.
Options
Beta-adrenergic receptor antagonists have been the best studied medical therapy during the perioperative period. Beta-blockers can be both short- and long-acting (12- or 24-hour dosage scheduling), can be nonselective or beta 1 selective, and can be administered intravenously or orally. The potency of the agents varies greatly, and some demonstrate weak stimulatory properties.
There are several paradigms with respect to modes of delivery of these agents perioperatively. Patients may be taking beta-blockers on a long-term basis. Alternatively, beta-blocker therapy could be initiated several days to weeks before surgery with titration of effect, administered the day of surgery, or begun intraoperatively.
Evidence
Early Studies
The first randomized trial of perioperative beta-blockers came from the Multicenter Study of Perioperative Ischemia (McSPI) study group. The study consisted of 200 Veterans Affairs patients with or at risk of coronary artery disease (CAD) undergoing noncardiac surgery. ( Table 39-1 ) Patients were randomly assigned to receive either placebo or atenolol (50 or 100 mg) 30 minutes before surgery and continued for 7 days. A 50% reduction in postoperative ischemia (from 34% to 17% in days 0 to 2, p = 0.008) based on Holter monitoring was detected. Although beta-blockers did not affect perioperative MACE, during follow-up, the incidence of postoperative cardiac events and overall mortality were both shown to be significantly lower in the atenolol group at 6 months (0% versus 8%; p < 0.001) and remained significant throughout the 2-year study period (10% versus 21%; p = 0.019).
| Risk Stratification | Procedure Type |
|---|---|
| Major vascular surgery (reported cardiac risk generally >5%) |
|
| Intermediate-risk surgery (reported cardiac risk generally 1%-5%) |
|
| Low-risk surgery (reported risk generally <1%) |
|
There were several important limitations to this trial. Patients were not excluded if they were already taking a beta-blocker; thus some patients randomly assigned into the placebo arm could have had effects from abrupt cessation of therapy. Beta-blocker withdrawal can lead to increases in heart rate and myocardial oxygen demand and predispose to myocardial ischemia. Additionally, only patients who survived to hospital discharge were examined because it was not an intention-to-treat analysis. If all in-hospital mortalities were included, the actual 2-year mortality rate would not have been significantly different ( p = 0.1).
In contrast to the McSPI study in which patients just at “risk” of CAD were included, the first DECREASE (Dutch Echocardiographic Cardiac Risk Evaluation Applying Stress Echocardiography) trial enrolled only patients with positive results on preoperative dobutamine stress echocardiography before major vascular surgery. Patients were randomly assigned to either titrated bisoprolol therapy or standard perioperative care. Patients taking preoperative beta-blockers or those with extensive wall motion abnormalities were excluded. Bisoprolol (5 to 10 mg) was started at least 1 week before surgery (average, 37 days prior) and then continued for 30 days postoperatively; the target heart rate was 51 to 79 beats/min.
The results showed a significant reduction in the primary endpoint of composite death from cardiac causes or nonfatal MI within 30 days postoperatively (34% versus 3.4%, p < 0.001). The trial was stopped early despite only enrolling 112 patients (20 cardiac events). This study was not double-blind, the degree of risk reduction was larger than many authors deemed reasonable, and the event rate in the placebo arm was also greater than expected.
The Perioperative Beta-Blockade (POBBLE) study found no difference in cardiovascular outcome in 97 vascular surgical patients randomly assigned to perioperative metoprolol versus placebo who underwent screening to ensure that CAD was not present. Similarly, both the Diabetic Postoperative Mortality and Morbidity (DIPOM; 921 patients) and Metoprolol after Vascular Surgery (MaVS; 496 patients) studies found no benefit in short- and long-term (6- to 18-month) cardiac outcomes with the administration of perioperative metoprolol initiated immediately prior or very close to surgery. A statistically significant increase in perioperative bradycardia and hypotension in the treatment arm was also shown.
It is important to note that the study populations in POBBLE, DIPOM, and MaVS represented a lower risk cohort than the DECREASE I trial did. Although DIPOM studied diabetic patients undergoing major surgery, major was defined as any procedure lasting greater than 1 hour and had somewhat vague exclusion criteria for patients with significant cardiac disease. The MaVS study specifically studied major vascular surgical patients but also had a low incidence of patients with known CAD and excluded those with significant comorbidities.
A large retrospective database cohort study of PBB by Lindenauer and colleagues used propensity score matching to adjust for differences in patients. They found that the administration of any beta-blocker perioperatively to patients not already taking them was associated with no benefit and possible harm in patients with a Revised Cardiac Risk Index (RCRI) of 0 to 1 (1 point each for the following: high-risk surgery, serum creatinine >2 mg/dL, with diabetes and taking insulin, or history of CAD, congestive heart failure [CHF], or cerebrovascular disease) ( Box 39-1 ). However, in patients with an RCRI of 2 or greater, perioperative beta-blockade was associated with a decreased risk of death.
History of ischemic heart disease
History of compensated or prior heart failure
History of cerebrovascular disease
Diabetes mellitus
Renal insufficiency (serum creatinine, >2 mg/dL)
Although the results of the POBBLE, DIPOM, and MaVS studies question the utility of PBB in mostly intermediate risk patients, the DECREASE IV study showed more positive results for a similar cohort. The study enrolled 1066 patients considered to have a 1% to 6% perioperative cardiovascular risk who were randomly assigned to receive bisoprolol or placebo as well as fluvastatin or placebo. The study design for bisoprolol administration was identical to that of the DECREASE I trial. A significant reduction in MACE within 30 days was shown for the 533 patients who received bisoprolol (2.1% versus 6.0%, p = 0.002). Similar to the original DECREASE trial, this trial also was not double-blind and was terminated early.
Perioperative Ischemic Evaluation Trial and Subsequent Studies
The Perioperative Ischemic Evaluation (POISE) trial included 8351 patients with or at risk of atherosclerotic disease who were randomly assigned to placebo or controlled release (CR) metoprolol. Initial therapy consisted of 100 mg orally 2 to 4 hours before surgery. Dosing was increased to 200 mg a day postoperatively and continued for 30 days. For those unable to take oral medications, 15 mg of intravenous metoprolol every 6 hours was given until oral therapy could be restarted. The study drug was held if heart rates were less than 50 beats/min or if systolic blood pressure (SBP) was less than 100 mm Hg; adjustments were then made in subsequent dosing.
Metoprolol CR therapy reduced the incidence of the primary endpoint of composite cardiovascular death, nonfatal MI, and nonfatal cardiac arrest within 30 days postoperatively (5.8% versus 6.9%, p = 0.04), and the reduction was driven by a decrease in PMI (4.2% versus 5.7%, p = 0.0017). Despite this decrease in MACE, there was a significant increase in the overall mortality in the study group (3.1% versus 2.3%, p = 0.0317) as well as a doubling in the stroke incidence (1% versus 0.5%, p = 0.0053). The increase in the all-cause mortality rate occurred mostly because of an increase in what was described as mortality related to infectious complications.
As stated by the POISE authors, for every 1000 patients treated with PBB, metoprolol CR would prevent 15 PMIs and seven cases of new onset atrial fibrillation while resulting in an excess of eight deaths and five strokes. The negative effect of stroke becomes even more significant when one considers that in the POISE cohort, most of those who sustained PMI had limited continued symptoms and only 6% to 9% required revascularization. In contrast, only 15% to 21% of the nonfatal stroke patients had complete recovery, whereas a significant number required functional support for daily activities.
Evaluation of the cause of this increase in morbidity and mortality showed the dominant factor to be perioperative hypotension. Significant hypotension (SBP <90 mm HG requiring intervention) was common in POISE, developing in 15% of the study group. The risk of death increased by fivefold, and the incidence of stroke associated with hypotension doubled. In addition to hypotension, intraoperative bleeding was also a predictor of stroke.
Subsequent to the release of POISE, the 2008 meta-analysis of Bengalore and colleagues demonstrated a 116% increase in stroke risk with PBB, which significantly offset their cardiovascular benefit (35% reduction in nonfatal PMI). A similar relationship with beta-blockers in the cardiology literature has recently been shown. Although beta-blockers still remain one of the mainstays of pharmacologic management for ischemic heart disease, this is not necessarily true for beta-blocker use in primary prevention, management of hypertension, and, now, even stable angina without prior MI in contemporary practice. In a study of 19,257 hypertensive patients with three or more coronary risk factors but without overt CAD, there was a 14% higher incidence of coronary events and a 23% higher incidence of stroke when an atenolol-based regimen was used compared with an amlodipine regimen.
Recent meta-analysis of beta-blocker use in the management of chronic hypertension has shown no benefit in all-cause mortality, cardiovascular mortality, and MI rates when they were compared with other antihypertensive agents. These results were seen even when beta-blockers were compared with placebo. To make matters worse, the incidence of stroke has consistently been shown to be higher with beta-blockers when compared with other therapies, ranging from 16% to 30% higher. Inferior control of blood pressure was also found with beta-blockers. This inefficiency in lowering pressure was even greater in the more important control of central aortic pressure.
Conclusions regarding the association of PBB with sepsis or other infectious causes of death are less clear. Perioperative hypotension was likely a marker for varying degrees of shock in many patients, potentially affecting the maintenance of gut integrity. The early use of intravenous beta-blockade after MI was studied in the Clopidogrel and Metoprolol in Myocardial Infarction (COMMIT) trial and was associated with a 30% increase in cardiogenic shock. The prevention of responsive tachycardia with the use of PBB in POISE could also have potentially delayed the recognition and treatment of sepsis. The effect of increased insulin resistance with PBB has not been studied in perioperative patients.
However, not all the recent evidence has been negative. In a large cohort study, Wallace and colleagues reported on the experience at the San Francisco Veterans Administration Medical Center implementing a perioperative cardiac risk reduction protocol. The addition of beta-blockade was associated with a reduction in 30-day (odds ratio [OR], 0.52; 95% confidence interval [CI], 0.33 to 0.83; p = 0.006) and 1-year mortality (OR, 0.64; 95% CI, 0.51 to 0.79; p < 0.0001). Flu and colleagues questioned the association of PBB use and stroke risk by demonstrating that beta-blocker treatment initiated more than 1 week before surgery was associated with improved outcomes compared with treatment initiated less than 1 week preoperatively, which was associated with an increased risk of stroke.
Continuation of Long-Term Therapy
Only a few studies and some nonsurgical data have implicated the withdrawal of beta-blockers in cardiac morbidity and mortality after surgery. Shammash and colleagues showed a significant increase in mortality (from 1.5% to 50%) and MI (from 5.3% to 50%) in those undergoing vascular surgery who had beta-blockers discontinued. It is important to note that this study was retrospective, only eight patients had beta-blockers discontinued, and the 50% mortality rate seems excessive for withdrawal. In a prospective survey study by Hoeks and colleagues, those receiving beta-blockers the day of surgery who then had them discontinued after vascular surgery had a significant increase in mortality rate compared with those who either were never given a beta-blocker, those who newly started them, or those who continued their beta-blocker perioperatively. The group who had beta-blockers stopped comprised only 21 of the 711 patients, and no information was available about why the drug was discontinued. The increased short-term and 1-year mortality rates were, however, significant.
Wallace and colleagues also studied perioperative beta-blocker withdrawal as part of their perioperative cardiac risk reduction protocol. A significant increase in 30-day and 1-year mortality rates were found with withdrawal and were more predictive of mortality than the presence of CAD or peripheral vascular disease (PVD). Similar to groups in previous studies, the withdrawal group represented only a small percentage of patients (4.6%), and no information was provided about why beta-blockers were discontinued. Because the University of California San Francisco has long employed an aggressive protocol for prophylactic PBB, the early withdrawal of beta-blockers in some patients was likely in response to other perioperative complications, potentially skewing their evidence.
Areas of Uncertainty
Although the findings of the POISE trial should cause clinicians to proceed with significant caution regarding PBB, one must consider several important issues before making conclusions based on this trial. First, would the cardiovascular benefits of PBB have been fully offset by increases in stroke and overall mortality rates if the cohort from POISE had had more baseline cardiovascular disease? In contrast to the DECEASE cohort in which provocative ischemia in vascular surgical patients was the inclusion criteria, only approximately 40% of patients in the POISE trial had either known CAD or underwent vascular surgery. The incidence of previous CHF was just 6%.
The outcome of PBB in the patients with higher cardiovascular risk from the POISE cohort is not known. The POISE authors merely state that this relationship did not reach statistical significance. It is not unreasonable to expect that the cardiovascular benefits would have been better in the higher risk patients compared with the only modest benefit (5.8% versus 6.9%) for MACE shown for the POISE trial cohort overall. It is unclear what degree of MACE risk reduction would have been needed to offset the complications recorded.
The second big issue regarding the POISE trial rests with the dosage of metoprolol used. As already discussed, therapy was initiated at 100 mg of metoprolol CR daily with a target daily dosing of 200 mg a day (the target dose could be reached as early as postoperative day 0). A potential starting dose of 200 mg/day is twice the maximum recommended starting dose. It is also 50% of the maximum recommended daily dose. In addition to the aggressive metoprolol dosage used in POISE, the threshold for allowing hypotension before withholding metoprolol (SBP >100) was also considered quite high. The POISE study also did not account for relative hypotension based on the percentage drop from baseline, although neither did any other study.
Despite the higher incidence of hypotension in study patients (15% versus 9.7% in control patients) and the significant association between hypotension and the risk of death and stroke, the POISE authors have taken issue with the notion that their negative results were related to excessive metoprolol dosing. They state that the dosing was not significantly different than in the DECREASE trial, where the starting 5-mg dose of bisoprolol was at 100% of the maximum recommended starting dose and that the target dose of 5 to 10 mg daily was up to 50% of the maximum recommended dose. Additionally, the target dosing of atenolol in the McSPI trial was also 50% of the maximum recommended dose.
One potential explanation is that metoprolol may have led to worse outcomes with POISE. Wallace and colleagues have published a comparison of outcomes depending on the choice of beta-blocker used for prophylaxis in surgical patients from the San Francisco Veterans Administration Medical Center. Thirty-day and 1-year mortality rates were significantly lower in the 1011 patients who received atenolol versus the 2776 patients who were given metoprolol. Possible mechanisms for this finding include the longer action of atenolol, which may allow for greater cardioprotection and improved cardioselectivity with atenolol over metoprolol. Additionally, there may be greater variability in the metabolism of metoprolol. A similar discrepancy has also recently been shown in the cardiology literature regarding outcomes after long-term use of atenolol versus metoprolol.
Although the choice of beta-blocker prophylaxis may be relevant, the most important difference between DECREASE and POISE is probably not the drug or its starting dose but rather the timing of the initiation of therapy in relation to the surgery. It certainly seems plausible that titration with bisoprolol therapy (average, 37 days) in the DECREASE I and IV trials could have allowed for a greater safety margin in the avoidance of hypotension and, potentially, subsequent related complications. This schedule may also allow for the early recognition of variability in drug response, given the relatively recent acknowledgment of significant interpatient genetic differences in both the clinical response to beta-blockers and metabolism of the drug. Some hypotension seen in the POISE trial may have been related to undiagnosed latent left ventricular dysfunction, which could also potentially be recognized with early initiation. Additionally, the pleotropic effects of beta-blockers on plaque stability and inflammation will take days to develop.
The Erasmus group published the pooled results of DECREASE I, II, and IV trials in terms of stroke risk in patients receiving titrated bisoprolol therapy greater than 30 days before major noncardiac surgery. The incidence of stroke was only 0.46% in the 3884 studied patients. A similar but even lower risk of stroke was also shown by the same group in patients undergoing noncardiac surgery who were receiving long-term beta-blocker therapy. In the more than 186,000 patients examined, only 34 suffered from a stroke (0.02%).
In contrast to the titration schedule in the DECREASE trials (average, 30 days) or to patients receiving long-term therapy, there was no preoperative titration in POISE, and maximum dosing occurred as early as day 0. The recommended titration schedule for metoprolol CR is on a weekly basis. The POISE authors have refuted the importance of the lack of titration by noting that 10% of the patients receiving placebo in the POISE study also developed perioperative hypotension.
An important question is which patients should avoid PBB. Several of the major randomized trials have excluded patients with asthma, whereas others have generally been more liberal in their inclusion. Fortunately, the use of beta 1 selective beta-blockers has been shown to have minimal effect on bronchial tone in cardiovascular patients. A depressed left ventricular ejection fraction or a previous history of CHF has rarely been an exclusion criterion in most of the major trials, but only a small percentage of patients enrolled have actually had this history. The incidences of CHF or previous CHF in the McSPI, DECREASE, CARP (Coronary Artery Revascularization Prophylaxis), and POISE trials were only 8%, 12.5%, 10%, and 6%, respectively. Given the association of presumed hypoperfusion and increased mortality rate from the POISE trial, short-term beta-blockade, particularly large fixed doses, in those with depressed myocardial function should be used with great caution until better safety data are available.
An additional group of patients in whom initiation of PBBs should be carefully considered is those with known cerebrovascular disease. In the POISE cohort, a history of stroke or transient ischemic attack was a significantly better predictor of postoperative stroke (population attributable risk [PAR], 30.5%) than either perioperative hypotension (PAR, 14.6%), atrial fibrillation (PAR, 6.9%), or intraoperative bleeding (PAR, 10.1%). Limburg and colleagues also showed a profound relationship: in their study, previous cerebrovascular disease led to a more than twelvefold increase in the risk of postoperative stroke. These findings highlight the risk associated with beta-blocker–induced hypotension or hypoperfusion in those with a compromised cerebrovascular tree. This risk may be exacerbated in patients with concomitant ventricular dysfunction or anemia.
Related posts:
Is a Preoperative Screening Clinic Cost-Effective?
Which Patient Should Have a Preoperative Cardiac Evaluation (Stress Test)?
What Is the Optimal Airway Management in Patients Undergoing Gastrointestinal Endoscopy?
What Works in a Patient with Acute Respiratory Distress Syndrome?
Which Are the Best Techniques for Reducing the Incidence of Postoperative Deep Vein Thrombosis?
Is Propofol Safe If Given by Nonanesthesia Providers?
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
