β-Blockers Before Percutaneous Coronary Intervention Do Not Attenuate Postprocedural Creatine Kinase Isoenzyme Rise
Background β-blocker (BB) use reduces infarct size in spontaneously occurring nonreperfused infarcts but probably does not change infarct size in patients treated with reperfusion therapy. A recent observational study suggested that BB use concurrent with percutaneous coronary intervention (PCI) decreased the risk of creatine kinase (CK)-MB elevation. The cogency of such a conclusion is dependent on the ability to risk-adjust for the multiple differences in patients treated with and without BBs.
Methods and Results Using propensity score and multivariate regression analyses, 6200 consecutive patients were analyzed to assess the relationship between BB use before PCI and per protocol-measured CK and CK-MB rise. There were several highly significant (P<0.001) differences between patients with and without BB treatment (eg, age, prior infarction, unstable angina). Maximum CK and CK-MB levels were higher in patients taking BBs (CK median, 95 U [interquartile range: 61, 175]; CK-MB, 3 U [2, 5]) than in patients not taking BBs (CK, 91 U [60, 157]; CK-MB, 3 U [2, 4]) (P=0.011 and P=0.021 for CK and CK-MB, respectively). After adjustment for significant differences in baseline characteristics there was no difference in either maximum CK rise (P=0.21) or maximum CK-MB rise (P=0.99).
Conclusions The results of this large observation study do not support the contention that BB use before PCI decreases myocardial injury.
Received November 6, 2000; revision received September 26, 2001; accepted September 26, 2001.
In the prereperfusion era, early treatment of patients with acute myocardial infarction (MI) with β-blockers (BB) appeared to reduce infarct size by ≈20%.1 The magnitude of benefit in patients treated with reperfusion therapy has been less well studied but appears to be greatly attenuated, if present at all.2 A recent observational study suggested that BB use before percutaneous coronary invention (PCI) both reduced creatine kinase (CK)-MB elevation suggestive of myocardial damage and improved long-term survival.3 Such a conclusion rests on the ability to adequately adjust for the many differences between patients treated with or without BBs. The level of microvascular perfusion seen after events that lead to the relatively low level of CK-MB rise most commonly seen after PCI has not been well characterized. However, because improvement in sluggish flow often is seen spontaneously or can be achieved by the administration of vasoactive drugs,4 one might consider the condition after PCI as more analogous to that after reperfusion therapy, and hence, the results with BB therapy and PCI surprising. Therefore, we studied the relation between prior BB use and CK-MB rise in a large and consecutive cohort of patients undergoing PCI in whom CK-MB was routinely measured.
Patient Population, Follow-up, and Data Integrity
Baseline, procedural, and outcome data on all patients undergoing PCI at the Cleveland Clinic are prospectively recorded on dedicated case report forms by trained personnel.5 Automated range and consistency checks are performed routinely. In addition, an investigator masked to BB status reviewed all charts for patients with CK ≥200 U. Because CK-MB mass unit levels were first reported routinely on patients without elevation in overall CK in January 1997, the study population consisted of all patients treated between January 1997 and February 2000, exclusive of patients presenting with acute MI and those with incomplete CK-MB data (n=402). No patient had undetermined BB status.
Measurement of CK
CK and CK-MB levels routinely were measured 6 to 8 hours after PCI, the morning after PCI, and in the event of the occurrence of symptoms suggestive of ischemia. Measurements were performed using a chemiluminescence immunoassay and an Elecsys 2010 analyzer from Roche Diagnostics (Indianapolis, Ind). Laboratory upper limits of normal were 220 U for CK and 8.8 U for CK-MB. CK-MB fraction was not reported if total CK was <100 U. For such measurements, a CK-MB of 3 U was imputed because this was the median value for those patients with measurements available after January 2000.
Data are presented as mean±SD, median and interquartile range, or a percent, as appropriate. Univariate between-group analyses were performed using χ2 test, t test, or Kolmogorov-Smirnov analyses. The primary end point of the study was defined prospectively as maximum post-PCI CK-MB, as adjusted for differences in the characteristics of the two populations with propensity analysis and multiple linear regression analyses. Logarithmic transformation of skewed variables was performed as necessary. A propensity score6 was developed using prespecified candidate variables to assess the likelihood of an individual patient being treated with BBs ≤24 hours before PCI (≥1 dose). The predictive value of the score was evaluated using receiver-operator characteristic (ROC) analyses. P<0.05 was considered statistically significant.
Baseline Characteristics and Initial Treatments
Baseline clinical and angiographic characteristics of the patients treated are presented in Table 1. Multiple highly significant differences were noted, including those for patient age, history of prior MI, and presentation with unstable angina (all P<0.001). Catheter laboratory treatments received are presented in Table 2. Patients pretreated with BBs were more likely to receive abciximab, a stent, or treatment for a stenosis in the left circumflex coronary artery (all P<0.05).
Candidate variables independently correlated with use of BBs and used for the propensity score were MI ≤2 weeks before PCI, remote MI, unstable angina, chronic obstructive pulmonary disease, intravenous heparin use, New York Heart Association congestive heart failure class, insulin-dependent diabetes, and sex. The C-statistic for the score was 0.62.
In-hospital outcomes also are presented in Table 2. There were no significant differences for death, non–Q-wave MI, or the need for emergency bypass surgery between patients treated with and without BBs. Patients treated with BBs, however, were more likely to suffer a Q-wave MI (0.47% versus 0.06%, P=0.001). Both maximum CK and CK-MB values were highly skewed (kurtosis [G2] for CK=1834.1 and for CK-MB=117.9). For patients treated with BBs, the median (interquartile range) for CK was 95 (61, 175) and for CK-MB was 3 (2, 5). For patients treated without BBs the respective values were 91 (60, 157) (P=0.011 by Kolmogorov-Smirnov testing) and 3 (2, 4) (P=0.021). Independent correlates of maximum CK-MB rise are presented in Table 3. After adjustment for imbalances in patient demographics, the apparent difference between groups in terms of maximum CK-MB rise was no longer present (P=0.99 for logarithmic transformation of CK-MB.) Similar analysis for the end points maximum CK rise and maximum nonimputed CK-MB rise also showed no significant difference (P=0.21 and P=0.19, respectively).
The principal finding of this large observational study was that there seemed to be no relationship between BB use at the time of PCI and subsequent rise in CK or CK-MB suggestive of myonecrosis or MI, in contradistinction to a recent report by Sharma and colleagues.3 Why should two groups come to opposite conclusions on this matter? First, there were substantial differences in the patient populations studied. Sharma and colleagues treated 57% of their patients with rotational atherectomy, whereas in this analysis only 16%, a figure much closer to the national average,7 were so treated. Not surprisingly, inasmuch as rotational atherectomy probably increases the risk of “slow flow” and CK-MB elevation,8,9 the incidence of CK elevation in the previous study was higher than in this study, despite comparable timing of its ascertainment. Second, there were important differences in the statistical methods used to attempt to minimize the influence of differences between patients treated with or without BBs. The Sharma et al3 analysis had the advantages of measuring CK-MB before the procedure and having fewer missing values (2% versus 6%). However, it would appear that it had several statistical limitations. First, it analyzed variables with a skewed distribution using methods for which statistical assumptions require normally distributed variables. Second, 3 of the 4 most important variables influencing the risk of CK-MB elevation in the present study were not measured by Sharma et al, whereas all of the key variables used in their multivariate analysis except for systemic atherosclerosis and diffuse coronary disease (themselves rather subjective terms) were captured in this analysis. Third, the Sharma et al analysis did not provide the reader with sufficient details of a multivariate analysis (ie, the C-statistic from ROC analysis) to allow proper evaluation of the nature of risk adjustment. Finally, the present study is considerably larger than that Sharma et al.
What other relevant evidence can be brought to bear on the question? Is difficult to be certain if studies of BBs in the setting of reperfusion therapy or without such therapy are more relevant. Although not specifically studied in this analysis, data from Sharma et al3 would suggest that slow flow and spasm are the most likely cause of CK-MB elevation in the PCI setting. Given that slow flow is often at least partially correctable,4 the reperfusion setting probably is more applicable. The randomized TIMI II BB substudy analyzed the effect of BBs after tissue plasminogen activator for acute infarction and found no benefit on global or infarct zone function.2 CK-MB rise was not studied. No other substantially sized randomized trial on this question has been reported.
It is important to acknowledge the limitations of this analysis. Of most importance is the fact that allocation to BB therapy was not randomly assigned and that there were substantial differences recognized between patients treated with or without BBs. Other differences were certainly not measured. Second, our propensity analysis for the use of BB was only modestly predictive, which is perhaps not surprising because almost one half of the patients received BBs. Third, a moderate number of patients had missing values for CK-MB (all with no elevation in total CK). However, analysis both with and without imputed values failed to show even a remote trend toward benefit with BB therapy.
Thus, these data fail to support the contention that BB therapy reduces myonecrosis after PCI. Nonetheless, because of the limitations of observational data analysis we acknowledge the possibility of such a benefit. Disparities in the two observational studies analyzing this question illustrate the difficulty of reaching firm conclusions about the potential benefit of a therapeutic intervention on the basis of observational data in which that intervention is applied with considerable selection bias on the part of the physician.
The authors thank Laura D. Reinhard for her expert assistance in the preparation of this manuscript.
Yusuf S. The use of beta-blockers in the acute phase of myocardial infarction.In: Califf RM, Wagner GS, eds. Acute Coronary Care 1986 Boston, Mass: Marinus Nijhoff; 1985: 73.
Roberts R, Rogers WJ, Mueller HS, et al. Immediate versus deferred β-blockade following thrombolytic therapy in patients with acute myocardial infarction: rsults of the thrombolysis in myocardial infarction (TIMI) II-B study. Circulation. 1991; 83: 422–437.
Sharma SK, Kini A, Marmur JD, et al. Cardioprotective effect of prior β-blocker therapy in reducing creatine kinase-MB elevation after coronary intervention: benefit is extended to improvement in intermediate-term survival. Circulation. 2000; 102: 166–172.
Piana RN, Paik GY, Moscucci M, et al. Incidence and treatment of “no-reflow” after percutaneous coronary intervention. Circulation. 1994; 89: 2514–2518.
Ellis SG, Guetta V, Miller D, et al. Relation between lesion characteristics and risk with percutaneous intervention in the stent and glycoprotein IIb/IIIa era: an analysis of results from 10,907 lesions and proposal for new classification scheme. Circulation. 1999; 100: 1971–1976.
Rosenbaum PR, Rubin DB. The central role of the propensity score in observational studies for causal effects. Biometrika. 1983; 70: 41–55.
American College of Cardiology. ACC National Cardiovascular Data Registry. Cath Lab Module [computer software]. Version 1.1. Bethesda, Md: American College of Cardiology; 2000. Cited by: American College of Cardiology. Institutional Report, January 1, 1999–December 31, 1999.
Ellis SG, Popma JJ, Buchbinder M, et al. Relation of clinical presentation, stenosis morphology and operator technique to the procedural results of rotational atherectomy. Circulation. 1994; 89: 882–892.