The Potential Preventability of Postoperative Myocardial Infarction: Underuse of Perioperative β-Adrenergic Blockade

Background  Among selected patients undergoing major noncardiac surgery, β-adrenergic blockade has been shown to reduce the risk for postoperative cardiac complications and mortality. We sought to determine how often postoperative MI might be considered preventable through appropriate use of these medications.

Methods  We reviewed the medical records of patients who developed a postoperative MI between January 1, 1998, and October 31, 2001, at Baystate Medical Center, a 570-bed community-based teaching hospital in Springfield, Mass. We calculated a Revised Cardiac Risk Index score and used criteria from previous randomized trials to determine whether patients would have been candidates for perioperative β-adrenergic blockade. Postoperative MI was considered potentially preventable if the patient appeared to have been an ideal candidate for β-blocker therapy but did not receive it before the infarction. We compared the mortality of ideal candidates who did and did not receive β-blockers before their infarction using multivariable logistic regression.

Results  Seventy (97%) of the 72 patients who developed postoperative MI could have been identified as being at increased risk for cardiac complications, and 58 (81%) appeared to be ideal perioperative β-blocker candidates. Thirty ideal candidates (52%) were treated with β-blockers before the development of the infarction. Among ideal candidates, treatment with a β-blocker before infarction was associated with an odds ratio of in-hospital mortality of 0.19 (95% confidence interval, 0.04-0.87).

Conclusions  A large percentage of the postoperative MIs at our institution might have been prevented if a β-blocker had been administered to all ideal candidates around the time of surgery. Use of β-blockers before infarction may reduces overall mortality, even among patients who go on to develop this complication.

Each year, as many as 1 million Americans who undergo major noncardiac surgery suffer a perioperative cardiac complication.^1,2 Among these complications, the development of postoperative myocardial infarction (MI) is of particular concern, because it is associated with a markedly increased risk for death, prolonged hospitals stays, and higher costs.^3-6 Epidemiological studies carried out during the past 3 decades have identified numerous factors associated with cardiovascular complications after surgery and led to the development of a series of clinical prediction tools.^5,7-10 Despite these advances, few strategies have been shown to reduce the occurrence of complications and, until recently, most prevention efforts have focused on the selective use of coronary revascularization in patients identified as being at the highest risk. Several randomized trials have demonstrated that the perioperative administration of β-adrenergic blockers to selected patients can reduce the incidence of postoperative myocardial ischemia, MI, and mortality.^11-15 On the basis of these studies, recent reviews^2,16-18 have suggested that β-blockers be used routinely in all but the lowest-risk individuals, and β-blockade is recommended in perioperative management guidelines produced by both the American College of Physicians and the American College of Cardiology.^19,20 The Agency for Healthcare Research and Quality report, Making Health Care Safer: A Critical Analysis of Patient Safety Practices,²¹ placed perioperative β-blockade near the top of its short list of interventions labeled "clear opportunities for safety improvement."

Relatively little is known about the use of β-blockers among high-risk patients undergoing major noncardiac surgery, and available data suggest that β-blockers may be underused.^22,23 In addition, although perioperative β-blockade is now widely advocated, it is unclear how many cases of postoperative MI might actually be prevented through encouraging their use. For example, if only a small percentage of the patients who develop postoperative MI would have been labeled as high risk at the time of surgery, then it is unlikely that a physician would have considered prescribing a β-blocker as a risk reduction strategy. Similarly, it is conceivable that most or all of the patients who develop postoperative MI are already being treated with β-blockers. If this is the case, alternative means of prevention will need to be discovered. For those intent on improving the safety of noncardiac surgery, it is important to have a clear understanding of the benefit that might accrue from placing greater emphasis on perioperative β-blockade. With this in mind, we sought to determine the percentage of cases of postoperative MI that could be considered potentially preventable through appropriate use of perioperative β-blockers.

We conducted a retrospective review of medical records at Baystate Medical Center, a 570-bed community-based teaching hospital in Springfield, Mass, that serves as the western campus of the Tufts University School of Medicine, Boston, Mass. All patients were treated by attending surgeons with the assistance of house staff. Permission to carry out the study was obtained from the Institutional Review Board at Baystate Medical Center.

We identified consecutive patients who underwent major noncardiac surgery and who had a secondary diagnosis of acute MI (International Classification of Diseases, Ninth Revision, Clinical Modification codes 410.01, 410.11, 410.21, 410.31, 410.41, 410.51, 410.61, 410.71, 410.81, and 410.91)²⁴ between January 1, 1998, and October 31, 2001. Major surgery was defined as surgery requiring inpatient admission. The coding of acute MI was considered correct if the patient had a new elevation of creatine kinase isoenzyme or troponin I level that met our laboratory's diagnostic criteria for MI during the study period and if this was accompanied by written documentation in the medical record attesting to the MI by the attending surgeon, house staff, or medical consultant. The MI was designated postoperative if the first evidence of elevation of enzyme level and documentation of infarction occurred after surgery.

Information collected from the medical record included demographic information about the patient, the date and type of surgery, the date the MI was documented by the physician, the creatine kinase isoenzyme or troponin I level, comorbidities, the patient's previous use of β-blockers, β-blocker use during the hospitalization, and contraindications to β-blockade. Additional information included the heart rate on the first postoperative day and in-hospital mortality.

A data collection tool was developed and pilot tested to abstract relevant data from medical records, which were then entered into a computerized database (TELEform Standard version 6.2; Cardiff Software, Inc, Vista, Calif) directly via facsimile. Abstraction errors were reduced by providing detailed data definitions and by a reabstraction of a random sample of charts.

Summary statistics for the overall sample were constructed using simple frequencies and proportions for categorical data and medians and interquartile ranges for continuous variables. We used the χ² and unpaired t tests for comparisons of categorical and continuous data, and reported P values are 2-sided.

Comorbidities present at the time of surgery were used to calculate a Revised Cardiac Risk Index score as described by Lee et al.⁵ As suggested in recent reviews,¹⁶ patients were considered potential candidates for perioperative β-blockade if they had a Revised Cardiac Risk Index score of 1 or greater (with 1 point assigned for each factor), or if they had 2 or more risk factors for coronary artery disease. Risk factors for coronary artery disease were adopted from Mangano et al,¹³ and included ischemic heart disease, cerebrovascular disease, renal insufficiency, diabetes mellitus, and high-risk surgical procedures. Minor criteria included age greater than 65 years, hypertension, current smoking, and hyperlipidemia. Patients who met these criteria were considered ideal candidates for perioperative β-blockade if they had no contraindications to β-blocker use at the time of surgery. Contraindication to β-blockade included a documented allergy or intolerance to β-blockers, resting heart rate of less than 55 beats/min in the absence of a pacemaker, second- or third-degree heart block, systolic blood pressure of less than 100 mm Hg, active airway obstruction, or decompensated heart failure.

A series of bivariate analyses was used to evaluate the relationship between a limited number of baseline clinical characteristics, perioperative β-blocker therapy, β-blocker therapy within 24 hours of infarction, and in-hospital mortality. Variables were entered into a final multivariable logistic regression model if they were associated with both β-blocker treatment and mortality at the level of P<.10, or if they were generally accepted as important potential confounders. Interaction terms were evaluated using similar criteria. Odds ratios (ORs) and corresponding 95% confidence intervals (CIs) are reported. All analyses were performed using StatView (version 5.0) (SAS Institute Inc, Cary, NC).

One hundred fifty-five patients who underwent major surgery were identified as having a secondary diagnosis of acute MI in the hospital's discharge database. Fifty-six patients were excluded from study because acute MI was not documented in the medical record; 16, because laboratory results failed to confirm MI; 10, because the MI occurred before surgery; and 1, because he had not undergone a surgical procedure during the admission.

Seventy-two patients who developed postoperative MI were therefore available for analysis. The median age was 74 years, roughly half were female (53%), and most were white (Table 1). Thirty-one (43%) of the patients had a history of MI, and the median Revised Cardiac Risk Index score was 2. Twenty-seven patients (38%) were being treated with a β-blocker at the time of admission. In descending order of frequency, vascular, general, and orthopedic surgeries were most common among patients in the study. The median interval between surgery and the development of postoperative MI was 2 days, and the in-hospital mortality rate was 25%. More than half (57%) of the patients had received a consultation by an internist or cardiologist before the development of the infarction.

Seventy patients (97%) who developed postoperative MI were categorized as being at increased risk for cardiac complications on the basis of a Revised Cardiac Risk Index score of 1 or greater, or because they had 2 or more risk factors for coronary artery disease. Twelve (17%) of these potential candidates had contraindications to β-blockade at the time of admission, leaving 58 ideal perioperative β-blocker candidates. Among this group, 24 (41%) were taking β-blockers before admission, and 30 (52%) received a β-blocker before the onset of myocardial injury. Five (9%) of the ideal candidates received optimal perioperative β-blocker therapy, namely a β-blocker in the preoperative and postoperative periods.

Of the 58 ideal candidates for β-blocker therapy, a greater percentage of those who received treatment were male, and there were nonsignficant trends toward younger age and higher cardiac risk, but in other respects the patients were not different (Table 2). There were 4 (13%) in-hospital deaths among ideal candidates who were treated with β-blockers before MI compared with 9 (32%) among those who did not receive a β-blocker before their infarct (P = .10). In a multivariable logistic regression model that adjusted for differences in age and Revised Cardiac Risk Index score, perioperative treatment with β-blockers before infarction was associated with an odds of death of 0.19 (95% CI, 0.04-0.87) (Table 3). The addition to the model of a variable for β-blocker therapy within 24 hours of MI did not alter the estimate of perioperative β-blocker effectiveness.

This study was designed to estimate the potential preventability of postoperative MI by examining risk characteristics and β-blocker use among patients who underwent major noncardiac surgery and had this complication. We found that most patients who developed postoperative MI could have been identified as being at high risk at the time of surgery. More important, we found that only half of the patients who were at high risk and still had no contraindications to β-blockade actually received these medications before experiencing the infarction. Overall, some 40% of the cases of postoperative MI that we reviewed might have been prevented if β-blockade had been initiated in all ideal candidates. Only a handful of patients received what we considered to be optimal perioperative β-blockade, namely a β-blocker before and after surgery. Finally, we found that among ideal candidates who developed postoperative MI, the administration of a β-blocker before the development of the infarction was associated with a marked reduction in in-hospital mortality.

It has been known for almost 2 decades that the administration of β-adrenergic blockers can reduce the incidence of myocardial ischemia associated with the stress of surgery. More recently, randomized controlled trials have shown that perioperative β-blocker therapy can reduce the incidence of postoperative MI and death among selected patients undergoing major noncardiac surgery.^13,14 To date, few reports have described the use of β-blockers in general practice. One study carried out at a large community-based teaching hospital found that only 40% of ideal β-blocker candidates who underwent noncardiac surgery received appropriate therapy.²² In the Netherlands, Boersma et al²³ found that 27% of high-risk patients undergoing vascular surgery received a β-blocker perioperatively. Although these studies suggest that increasing the use of β-blockers may represent an opportunity to improve the safety of noncardiac surgery, they did not address the potential preventability of postoperative MI. This notion of potential preventability is particularly useful for those charged with developing quality or safety initiatives, because it presents a more realistic picture of actual opportunities to reduce complications.

Although it is unique in the context of postoperative MI, our focus on potentially preventable events is not new. Arnold et al²⁵ found that 17% of hospital-acquired venous thromboembolism was potentially preventable through appropriate use of preventive measures. Moreover, infection control practitioners have used the term potentially preventable surgical infections for close to a decade²⁶ to refer to infections that develop in the face of suboptimal prevention practice. For example, a wound infection that develops in the setting of poorly timed antibiotic prophylaxis can be considered potentially preventable. On the other hand, infections that develop even when all preventive measures have been performed can be considered apparently unavoidable. This approach to classifying complications has been adopted by the Centers for Medicare and Medicaid Services in their Surgical Infection Prevention Collaborative.²⁷

Our study has a number of strengths. First, we used a rigorous definition of postoperative infarction, which led to the elimination of greater than half of the cases that would have been included had we relied on administrative data alone. Second, through examination of the medication administration record, we were able to determine actual use of β-blockers at key points throughout the hospitalization. This revealed that although half of the patients received a β-blocker at some point before the onset of MI, few received optimal therapy. Finally, in designing this study, we addressed the potential preventability of postoperative infarction, an issue that is especially relevant to those interested in patient safety and quality improvement.

There are several weaknesses of this study. First, we were dependent on administrative data to identify potential cases of postoperative infarction, and the number of cases that we identified were below what might have been expected on the basis of the known incidence of postoperative MI. This may have resulted from underdetection by clinicians, from poor documentation, or from undercoding in our medical records department. For this to have biased our findings, there would need to be differential rates of β-blocker use between unidentified and identified infarctions. That seems unlikely. Second, eligibility for perioperative β-blocker therapy was determined retrospectively using medical records. Although this might have been anticipated to result in an underestimate of the number of potentially eligible patients, we found that 97% of the patients met our criteria for increased risk. Third, we looked only at MIs that occurred during the hospitalization, but we know that a large number of postoperative infarctions occur in the weeks and months after discharge from the hospital. Finally, many factors contribute to the actual preventability of postoperative infarction; we focused on β-blocker therapy but recognize that other, less well-studied aspects of care may be important. It was for this reason that we chose to use the term potential preventability when addressing the opportunities uncovered in this analysis. As our results demonstrate, perioperative β-blocker therapy is no panacea; MI developed in 30 patients despite receiving a β-blocker. However, our finding that prophylactic β-blocker therapy may attenuate the impact of the infarction, resulting in lower in-hospital mortality rates, is encouraging and should be explored further.

Given the often devastating consequences of postoperative MI, the finding that nearly half of such cases could be considered preventable reinforces perioperative β-blockade as an important new strategy for those interested in improving safety in noncardiac surgery. Although this is reason to be hopeful, the obstacles facing those intent on improving perioperative β-blocker use are sizable. Increasing the rates of β-blocker administration in the setting of acute MI not associated with surgery has been a priority for many years, yet progress has been slow.^28-30 In that instance, the physicians targeted for education were internists and cardiologists who had previous experience with this class of drugs. In the case of noncardiac surgery, however, the challenge is to encourage the use of β-blockers by surgeons who may be less comfortable prescribing these agents. For this reason, novel strategies such as the comanagement of surgical patients between surgeons and internists may be required to achieve success.

A large percentage of postoperative MI may be preventable through improved use of β-blockers perioperatively. Greater efforts should be made to encourage β-blocker use among high-risk patients undergoing major noncardiac surgery.

Corresponding author: Peter K. Lindenauer, MD, MSc, Division of Healthcare Quality, Baystate Medical Center, 759 Chestnut St, P-5928, Springfield, MA 01199 (e-mail: Peter.Lindenauer@bhs.org).

Accepted for publication May 29, 2003.

This study was presented at the annual meeting of the Society of General Internal Medicine; May 3, 2003; Vancouver, British Columbia.