Although clinical guidelines recommend long-term β-blocker (BB) therapy to decrease mortality after acute myocardial infarction, these recommendations are based predominantly on evidence from before the reperfusion and thrombolytic eras. To investigate the effects of BB therapy for patients with acute myocardial infarctions on mortality in the percutaneous coronary intervention era, a total of 5,628 consecutive patients who were admitted <24 hours after the onset of ST-segment elevation myocardial infarction, treated with emergent percutaneous coronary intervention, and discharged alive were studied. During a median follow-up period of 1,430 days, mortality rates did not differ between patients with and without BB therapy (5.2% vs 6.2%, p = 0.786). Multivariate analysis revealed that BB treatment was not associated with a reduced risk for mortality (hazard ratio 0.935, 95% confidence interval 0.711 to 1.230, p = 0.534). The results of propensity score matching also indicated that the mortality rates did not differ between the 2 groups. However, subgroup analyses among matched populations revealed that BB treatment was associated with a significantly lower mortality risk for high-risk patients, who were defined as those with Global Registry of Acute Coronary Events (GRACE) risk scores ≥121 (hazard ratio 0.596, 95% confidence interval 0.416 to 0.854, p = 0.005) or those administered diuretics (hazard ratio 0.602, 95% confidence interval 0.398 to 0.910, p = 0.016), but not for lower risk patients. In conclusion, BB treatment was associated with reduced long-term mortality in patients after ST-segment elevation myocardial infarction at higher risk, but not in those at lower risk. Although randomized controlled studies are warranted to confirm these results, the implementation of BB therapy for discharged patients with ST-segment elevation myocardial infarction may need to be assessed on the basis of individual mortality risk in the percutaneous coronary intervention era.
Under current clinical guidelines, oral β-blocker (BB) therapy is widely recommended for indefinite long-term use in all patients who recover from ST-segment elevation myocardial infarction (STEMI) and do not have contraindications. However, these recommendations are based predominantly on evidences obtained before the reperfusion and thrombolytic eras, and few data have been collected in the percutaneous coronary intervention (PCI) era. Recent advances in the management of STEMI, particularly the use of primary PCI, have significantly reduced long-term mortality. Because these treatment advances potentially mask the mortality benefits of BB therapy, reassessing the efficacy of BB exposure for patients who survive STEMI is warranted. In this study, we investigated the relation between BB treatment at discharge and long-term mortality for consecutive patients with STEMI enrolled in the Osaka Acute Coronary Insufficiency Study (OACIS).
Methods
The OACIS is a prospective, multicenter observational study of consecutive patients with acute myocardial infarctions (AMIs) at 25 collaborating hospitals located in the Osaka region of Japan and is registered with the University Hospital Medical Information Network Clinical Trials Registry in Japan ( UMIN000004575 ). One of the main aims of the OACIS is to examine the effects of cardiovascular prevention drugs on secondary prevention after AMI in the contemporary clinical setting. A detailed description of the OACIS has been published elsewhere. The study protocol was approved by the ethics committee of each participating hospital, and each patient provided written informed consent.
Among the 10,074 patients registered in the OACIS registry from April 1998 to April 2011, we identified 5,628 consecutive patients who were admitted <24 hours of the onset of STEMI, treated with emergent PCI, and discharged alive.
Investigative cardiologists and research coordinators recorded demographic and clinical data for patients during the period of hospitalization. After discharge, further data were obtained at 3 and 12 months after AMI and annually thereafter for up to 5 years. Thrombolysis In Myocardial Infarction (TIMI) and Global Registry of Acute Coronary Events (GRACE) risk scores were calculated with multiple imputation for each patient as described elsewhere. The left ventricular ejection fraction was assessed using echocardiography before discharge using the Teichholz method. The primary end point of this study was all-cause death, which was categorized as cardiac, noncardiac, or unknown.
Categorical variables were compared using chi-square tests with continuity correction or Fisher’s exact tests. Continuous variables are presented as medians (interquartile range [IQR]) or as mean ± SD and were compared using unpaired Student’s t tests or 2-tailed Wilcoxon’s rank-sum tests between patients with and those without oral BB treatment at discharge. To minimize differences in baseline characteristics between the 2 groups, patients were matched in a 1-to-1 manner on the basis of propensity scores, which were calculated for each patient using a logistic regression model that included a total of 32 variables (baseline demographics, angiographic parameters, and medication at discharge), as listed in Table 1 . The variables inserted into the multivariate models to calculate propensity scores were determined after screening for multicollinearity. According to the propensity score, patients were selected using a 5-to-1 digit-matching technique using the nearest neighbor method. The area under the receiver-operating characteristic curve and the Hosmer-Lemeshow goodness-of-fit statistic were calculated to assess the performance and calibration of the model, respectively. Mortality rates were determined using Kaplan-Meier curves and were compared using log-rank tests. Cox regression analyses were performed to assess whether BB therapy was associated with a reduced risk for mortality. Variables with p values <0.20 before matching in univariate analyses were included in the multivariate Cox regression models. Propensity score was incorporated as a variable into the models before matching. To identify high-risk populations according to GRACE scores, classification and regression trees for survival data (survival CART) were used. Survival CART analysis revealed that the first split point to partition the mortality risk for patients without BB treatment among the matched populations was a GRACE risk score of 121 and that the second and third split points for each subgroup were risk scores of 100 and 141, respectively. Therefore, the mortality benefits of BB therapy at discharge were initially compared between patients with GRACE risk scores <121 and ≥121 and then among those with scores of <100, 100 to 120, 121 to 140, and ≥141. Subgroup analysis was performed in patients after propensity score matching to identify patients having a mortality benefit of BB treatment. All analyses were performed using PASW Statistics version 18 (SPSS, Inc., Chicago, Illinois) or SAS version 9.1.3 (SAS Institute Inc., Cary, North Carolina). Statistical significance was defined as p <0.05. For the subgroup analyses, p values <0.05 and p values for interactions <0.10 were considered as statistically significant.
| Variable | Before Matching | After Matching | ||||||
|---|---|---|---|---|---|---|---|---|
| Without BBs | With BBs | Overall | p Value | Without BBs | With BBs | Overall | p Value | |
| (n = 2,748) | (n = 2,880) | (n = 5,628) | (n = 1,923) | (n = 1,923) | (n = 3,846) | |||
| Year | <0.001 | <0.001 | ||||||
| 1998–2001 | 41.8% | 20.7% | 31.0% | 27.2% | 30.9% | 29.1% | ||
| 2002–2005 | 38.9% | 32.7% | 35.7% | 45.9% | 45.4% | 45.6% | ||
| 2006–2009 | 17.1% | 37.6% | 27.6% | 23.8% | 22.4% | 23.1% | ||
| 2010–2011 | 2.2% | 9.1% | 5.7% | 3.1% | 1.3% | 2.2% | ||
| High-volume hospital | 60.7% | 68.0% | 64.4% | <0.001 | 65.3% | 65% | 65.2% | 0.839 |
| Age (yrs) | 64.7 ± 12 | 64.7 ± 11.6 | 64.7 ± 11.8 | 0.923 | 65.1 ± 12 | 64.4 ± 11.4 | 64.7 ± 11.8 | 0.092 |
| Men | 75.7% | 79.0% | 77.3% | 0.003 | 76.4% | 77.8% | 77.1% | 0.300 |
| Body mass index (kg/m 2 ) | 23.6 ± 3.5 | 24 ± 3.5 | 23.8 ± 3.5 | <0.001 | 23.7 ± 3.5 | 23.9 ± 3.5 | 23.8 ± 3.5 | 0.078 |
| Diabetes mellitus | 33.0% | 32.6% | 32.8% | 0.716 | 32.9% | 32.4% | 32.7% | 0.756 |
| Hypertension | 53.9% | 64.8% | 59.5% | <0.001 | 59.0% | 61.0% | 60.0% | 0.222 |
| Dyslipidemia | 44.0% | 47.6% | 45.8% | 0.008 | 45.6% | 47.7% | 46.7% | 0.207 |
| Smokers | 67.2% | 64.7% | 65.9% | 0.050 | 66.4% | 64.9% | 65.6% | 0.340 |
| Previous myocardial infarction | 11.1% | 10.7% | 10.9% | 0.574 | 11.5% | 11.2% | 11.3% | 0.784 |
| Angina pectoris | 22.6% | 19.1% | 20.8% | 0.001 | 21.6% | 22.0% | 21.8% | 0.789 |
| Onset to admission time (h) | 2.4 (1.2–5.8) | 2.4 (1.1–5.3) | 2.4 (1.1–5.5) | 0.194 | 2.5 (1.2–6.0) | 2.5 (1.2–5.4) | 2.4 (1.1–5.5) | 0.205 |
| Cardiopulmonary arrest on arrival | 1.0% | 1.7% | 1.3% | 0.024 | 1.4% | 0.9% | 1.2% | 0.134 |
| Killip class ≥II | 12.7% | 16.5% | 14.6% | <0.001 | 14.4% | 15.4% | 14.9% | 0.400 |
| TIMI risk score | 5.3 ± 2.5 | 5.8 ± 2.3 | 5.6 ± 2.4 | <0.001 | 5.7 ± 2.4 | 5.5 ± 2.4 | 5.6 ± 2.4 | 0.124 |
| GRACE risk score | 100.6 ± 27.4 | 102.5 ± 26.9 | 101.6 ± 27.2 | 0.008 | 101.8 ± 28.1 | 101.6 ± 26.7 | 101.6 ± 27.2 | 0.838 |
| Initial TIMI grade 3 flow | 12.4% | 11.8% | 12.1% | 0.466 | 12.7% | 11.5% | 12.1% | 0.260 |
| Collateral circulation | 34.2% | 34.3% | 34.3% | 0.924 | 35.7% | 36.2% | 36.0% | 0.703 |
| Multivessel coronary disease | 33.6% | 38.0% | 35.8% | 0.001 | 36.3% | 35.7% | 36.0% | 0.714 |
| Left anterior descending coronary artery culprit lesion | 42.5% | 51.6% | 47.2% | <0.001 | 44.8% | 48.6% | 46.7% | 0.019 |
| Stent deployment | 71.9% | 82.5% | 77.3% | <0.001 | 78.1% | 77.9% | 78.0% | 0.869 |
| Thrombectomy | 38.0% | 55.7% | 47.1% | <0.001 | 46.6% | 44.5% | 45.6% | 0.200 |
| Emergent coronary-aorta bypass graft surgery | 1.3% | 0.6% | 1.0% | 0.010 | 1.0% | 0.8% | 0.9% | 0.461 |
| Final TIMI grade 3 flow | 88.4% | 89.6% | 89.0% | 0.174 | 89.2% | 88.8% | 89.0% | 0.693 |
| Peak creatinine phosphokinase >3,000 IU/L | 36.8% | 44.7% | 40.9% | <0.001 | 40.4% | 43.0% | 41.7% | 0.106 |
| Peak creatinine phosphokinase (IU/L) | 2,220 (1,145–3,865) | 2,709 (1,329–4,766) | 2,439 (1,235–4,328) | <0.001 | 2,394 (1,158–4,065) | 2,631 (1,311–4,562) | 2,439 (1,235–4,328) | 0.002 |
| Q-wave myocardial infarction | 75.1% | 76.3% | 75.7% | 0.334 | 75.3% | 76.2% | 75.7% | 0.532 |
| Statins | 36.1% | 52.2% | 44.3% | <0.001 | 42.4% | 42.6% | 42.5% | 0.896 |
| Aspirin | 92.8% | 96.3% | 94.6% | <0.001 | 94.2% | 95.5% | 94.8% | 0.057 |
| Dual-antiplatelet therapy | 64.0% | 76.5% | 70.4% | <0.001 | 70.5% | 70.3% | 70.4% | 0.896 |
| Angiotensin-converting enzyme inhibitors | 52.7% | 52.0% | 52.3% | 0.610 | 52.7% | 56.3% | 54.5% | 0.025 |
| Angiotensin receptor blockers | 19.5% | 33.9% | 26.8% | <0.001 | 25.0% | 25.9% | 25.5% | 0.505 |
| Angiotensin-converting enzyme inhibitors or angiotensin receptor blockers | 70.7% | 83.2% | 77.1% | <0.001 | 75.8% | 79.9% | 77.8% | 0.002 |
| Calcium blockers | 20.3% | 15.9% | 18.1% | <0.001 | 17.2% | 17.9% | 17.6% | 0.525 |
| Diuretics | 21.5% | 30.5% | 26.1% | <0.001 | 25.1% | 27.5% | 26.3% | 0.092 |
| Nitrate | 35.7% | 26.6% | 31.1% | <0.001 | 30.2% | 31.8% | 31.0% | 0.280 |
| Nicorandil | 24.9% | 25.4% | 25.2% | 0.650 | 24.9% | 25.3% | 25.1% | 0.795 |
| Propensity score | 0.426 ± 0.176 | 0.579 ± 0.189 | 0.504 ± 0.198 | <0.001 | 0.491 ± 0.165 | 0.492 ± 0.165 | 0.491 ± 0.165 | 0.873 |
Results
Among the 5,628 study patients, 2,880 (51.2%) were prescribed oral BB therapy at discharge after STEMI. In the BB group, 2,075 (72.0%), 559 (19.4%), 135 (4.7%), 33 (1.1%), and 78 (2.7%) patients received carvedilol, metoprolol, bisoprolol, atenolol, and other BBs, respectively. A trend of increased prescription of BB at discharge by year was clearly evident until 2009, as shown in Figure 1 (p <0.0001). After 2009, approximately 80% of patients received BB treatment. In addition, several significant differences in the baseline characteristics between patients in the BB and non-BB groups were detected ( Table 1 ). Notably, patients in the BB group were more often men, had higher body mass indexes and TIMI and GRACE risk scores, and displayed higher frequencies of hypertension, dyslipidemia, cardiopulmonary arrest on arrival, and Killip class ≥II. With regard to angiographic findings, a greater number of BB group patients had multivessel disease and culprit lesions involving the left anterior descending coronary artery compared to those in the non-BB group. Furthermore, BB group patients had a higher frequency of thrombectomy and stent deployment in the acute stage. Although the success rate of PCI, which was defined as final TIMI grade 3 flow, was similar in the 2 groups, peak creatinine phosphokinase was significantly higher in BB group patients, reflecting the increased severity of myocardial damage in these patients. At discharge, the prescription of dual-antiplatelet therapy, angiotensin receptor blockers, and diuretics was also more common in BB group patients, whereas calcium channel blockers and nitrate were less frequent ( Table 1 ). During a median follow-up period of 1,430 days (IQR 454 to 1,794), no significant difference was detected in the rate of all-cause death between the BB and non-BB groups ( Tables 2 and 3 , Figure 2 ). Similarly, there were no significant differences in the cause of death ( Table 2 ). Multivariate Cox regression analyses revealed that BB therapy was not associated with a decreased risk for all-cause, cardiac, or noncardiac death ( Table 3 ). To minimize differences in the baseline characteristics between the BB and non-BB groups, patients were matched using the propensity score method. The area under the receiver-operating characteristic curve was 0.725 (95% confidence interval 0.711 to 0.739), and the p value of the Hosmer-Lemeshow test was 1.000. A total of 3,846 patients with well-matched baseline characteristics, with the exception of year, culprit lesion involving the left anterior descending coronary artery, peak creatine kinase, and prescription of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers, were identified between the 2 groups, ( Table 1 ). However, no marked differences in mortality rates were detected between the groups ( Table 2 , Figure 2 ). Among patients after propensity score matching, the prescription of BBs at discharge was associated with lower long-term mortality in high-risk patients, who were defined as those with GRACE risk scores ≥121 ( Figure 3 ) or those who were prescribed diuretics ( Figure 4 ), with significant p values for interaction (p = 0.013 and p = 0.077, respectively; Figure 5 ). Patients with GRACE risk scores ≥121 or those administrated diuretics were more likely to have histories of myocardial infarction (22.2% vs 7.9%, p <0.001, for GRACE score ≥121 vs <121; 17.8% vs 9.2%, p <0.001, for diuretics vs no diuretics), Killip class ≥II on admission (45.6% vs 5.3%, p <0.001, for GRACE score; 31.1% vs 9.1%, p <0.001, for diuretics), and greater peak creatine phosphokinase values (2,709 IU/L [IQR 1,442 to 4,518] vs 2,431 IU/L [IQR 1,192 to 4,261], p <0.001, for GRACE score; 3,549 IU/L [IQR 1,995 to 5,964] vs 2,206 IU/L [IQR 1,080 to 3,809], p <0.001, for diuretics). Kaplan-Meier estimates and Cox regression analysis for the subgroups partitioned by CART analysis suggested that an association existed between BB treatment and reduced mortality for patients with GRACE risk scores ≥121 ( Figure 5 ), particularly for those with scores of 121 to 140 ( Table 4 ).
| Outcome | Before Matching | After Matching | ||||||
|---|---|---|---|---|---|---|---|---|
| Without BBs | With BBs | Overall | p Value | Without BBs | With BBs | Overall | p Value | |
| (n = 2,748) | (n = 2,880) | (n = 5,628) | (n = 1,923) | (n = 1,923) | (n = 3,846) | |||
| All-cause death | 170 (6.2%) | 149 (5.2%) | 319 (5.7%) | 0.786 | 120 (6.2%) | 108 (5.6%) | 228 (5.9%) | 0.171 |
| Cardiac death | 45 (1.6%) | 31 (1.1%) | 76 (1.4%) | 0.208 | 34 (1.8%) | 27 (1.4%) | 61 (1.6%) | 0.248 |
| Reinfarction | 19 (0.7%) | 15 (0.5%) | 34 (0.6%) | 0.702 | 16 (0.8%) | 12 (0.6%) | 28 (0.7%) | 0.348 |
| Heart failure | 15 (0.5%) | 7 (0.2%) | 22 (0.4%) | 0.117 | 11 (0.6%) | 6 (0.3%) | 17 (0.4%) | 0.176 |
| Arrhythmia or sudden death | 5 (0.2%) | 5 (0.2%) | 10 (0.2%) | 0.937 | 4 (0.2%) | 5 (0.3%) | 9 (0.2%) | 0.802 |
| Mechanical complication | 4 (0.1%) | 1 (0%) | 5 (0.1%) | 0.189 | 2 (0.1%) | 1 (0.1%) | 3 (0.1%) | 0.530 |
| Others | 2 (0.1%) | 3 (0.1%) | 5 (0.1%) | 0.558 | 1 (0.1%) | 3 (0.2%) | 4 (0.1%) | 0.353 |
| Noncardiac death | 74 (2.7%) | 81 (2.8%) | 155 (2.8%) | 0.222 | 52 (2.7%) | 54 (2.8%) | 106 (2.8%) | 0.822 |
| Malignant tumor | 11 (0.4%) | 11 (0.4%) | 22 (0.4%) | 0.840 | 4 (0.2%) | 6 (0.3%) | 10 (0.3%) | 0.656 |
| Unknown cause | 51 (1.9%) | 37 (1.3%) | 88 (1.6%) | 0.332 | 34 (1.8%) | 27 (1.4%) | 61 (1.6%) | 0.232 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
