Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Jul 1.
Published in final edited form as: Am J Cardiol. 2023 May 12;198:124–132. doi: 10.1016/j.amjcard.2023.04.022

Impact of Diabetes Mellitus on Benefit of Beta-Blocker Therapy After Myocardial Infarction

Ghaith Zaatari 1,*, David Bello 2,*, Catherine Blandon 1, J Dawn Abbott 3, Haris Subačius 4, Jeffrey J Goldberger 1, Outcomes of Beta-blocker Therapy After Myocardial INfarction Investigators
PMCID: PMC10330557  NIHMSID: NIHMS1900697  PMID: 37183092

Abstract

Beta-blockers are uniformly recommended for all patients after myocardial infarction (MI), including those with diabetes mellitus (DM). This study assesses the impact of beta-blocker type and dosing on survival in patients with DM after MI. A cohort of 6,682 patients in the OBTAIN registry were discharged post-MI. In this cohort, 2137 patients had DM (32%). Beta-blocker dose was indexed to the target daily dose used in randomized clinical trials, reported as percentage. Dosage groups were: no beta-blocker, >0-12.5%, >12.5-25%, >25-50%, and >50% of the target dose. The overall mean discharge beta-blocker dose in patients with DM was 42.7±34.1% vs 35.9±27.4% in patients without DM (p<0.0001). Patients with DM were prescribed carvedilol at a higher rate compared to those without DM (27.8% vs 19.6%). The 3-year mortality estimates were 24.4% and 12.8% for patients with DM versus without DM (p<0.0001), respectively, with an unadjusted HR=1.820 (CI:1.587-2.086, p<0.0001). Patients with DM in the >12.5% to 25% dose category had the highest survival rates, while patients in the >50% dose had the lowest survival rate among patients discharged on beta-blockers (p<0.0001). In multivariable analysis among post-MI diabetic patients, all beta-blocker dose categories demonstrated lower mortality compared to no therapy; however, only the >12.5-25% dose had a statistically significant HR=0.450 (95% CI: 0.224-0.907; p=0.025). In patients with DM, there was no statistically significant difference in 3-year mortality among those treated with metoprolol versus carvedilol. In conclusion, our analysis in post-MI diabetic patients suggested a survival benefit from beta-blocker therapy with no apparent advantage to high versus low dose beta-blocker therapy, though physicians tended to prescribe higher doses in patients with DM. There was no survival benefit for carvedilol over metoprolol in patients with DM.

Keywords: Acute myocardial infarction, diabetes mellitus, beta-blockers

INTRODUCTION

Beta-blocker therapy after an acute myocardial infarction (MI) improves survival. Patients with diabetes mellitus (DM) and those with MI are at a higher risk for cardiovascular events and mortality than those without DM.1-3 Beta-blockers improve survival post-MI in patients with and without DM and are now uniformly recommended.4 However, optimal beta-blocker dosing post-MI remains unknown and the majority of patients receive doses that are substantially lower than the target doses used in randomized clinical trials.5 The OBTAIN study6 was an observational multicenter registry that evaluated the impact of beta-blocker dosing on survival following MI, showing no benefit for target dose therapy compared to lower doses, with the lowest mortality found with 25% of the target dose.6 It is plausible that underlying risk for cardiovascular events may determine the intensity of beta-blockade required for optimal outcomes. Given the increased risk associated with DM in the post-MI patient, we hypothesized that the dose-response relationship of beta-blocker therapy post-MI would differ among diabetic patients, with increased survival associated with higher dose beta-blocker therapy. Additionally, in patients with DM, there has been concern about the effects of non-selective and cardioselective beta-blockers on glycemic and metabolic control. In the GEMINI trial there was better overall glycemic control with carvedilol versus metoprolol in hypertensive patients with DM.7 In this report, we also compare the effect of carvedilol and metoprolol in post-MI patients with DM.

METHODS

Detailed methods for the OBTAIN study have been published6. Briefly, the OBTAIN registry included patients hospitalized for acute MI from 26 participating centers (25 in the United States and 1 in Canada). MI was documented by: 1) Cardiac enzymes (creatine phosphokinase elevation >2 times or troponin elevation >3 times the upper limit of normal for the laboratory value) and 2) Electrocardiographic changes and/or symptoms consistent with MI (ie, chest pain, shortness of breath). From a total cohort of 7057, 6682 were discharged alive. The 2137 subjects with DM are the focus of this substudy. Characteristics of the study population, medical history, information regarding the index MI, discharge medications, including beta-blocker type and dose were collected. All data were collected at the site and deidentified patient information was entered in a web-based electronic data capture system.

Beta-blocker type and dose was chosen by the managing physician. As performed in the original report, beta-blocker doses were indexed to the target doses (dose administered/target dose) used in prior clinical trials: metoprolol 200 mg/day,8,9 carvedilol 50 mg/day (carvedilol controlled-release equivalent dose-80 mg/day),10 propranolol 180 mg/day,11 bisoprolol 10 mg/day,12 and atenolol 100 mg/day.13 Beta-blocker doses were divided into five pre-specified groups: no beta-blocker, >0-12.5%, >12.5-25%, >25-50%, and >50% of the target dose.

The pre-specified endpoint for this study was time to all-cause mortality with survival right-censored at 3 years. Vital status was assessed by chart review, the Social Security Administration Death Master File, or direct communication with the patient/family.

Patient characteristics were summarized as mean ± SD or count. Differences among groups were compared using chi-square tests for categorical variables and analysis of variance for continuous variables. Distribution-free rank sum tests were used for variables that deviated from normality. The median (interquartile range) was used to summarize these variables. The Kaplan-Meier method was used to calculate 1-, 2-, and 3-year survival in each study group. Pre-specified analysis of the effect of the 5 pre-specified groups (no beta-blocker and each of the 4 dose groups) on 3-year survival was tested by comparing Kaplan-Meier survival curves with a log-rank test. Cox proportional hazards regression was used to test for the independent effects of beta-blocker dosing on survival. Cox proportional hazards regression was also used to test for the effects of metoprolol versus carvedilol and the interaction of reduced (≤40%) versus preserved (>40%) left ventricular ejection fraction on survival.

Multivariable analysis randomized by site included all covariates listed in Table 1, including use of carvedilol versus metoprolol. Propensity score analysis was also performed to adjust for patient differences in the 5 beta-blocker dose groups in patients with and without DM. To calculate the propensity score, we used mixed effects linear regression with random effects of the recruiting centers, categorical discharge beta-blocker dose range as a dependent variable, and the expanded control variable set reported in table 1. All tests were 2-tailed and conventional 5% significance level was used. Analyses were performed using SAS software version 9.4 and R software version 4.03.

Table 1:

Characteristics of the study population stratified by presence or absence of diabetes

No diabetes Diabetes p-value
n (%) 4545 (68.0%) 2137 (32.0%) < 0.0001
Age (years) 63.2±14.1 65.2±12.3 < 0.0001
Male 3173 (69.9%) 1366 (63.9%) < 0.0001
Race
   White 3692 (81.3%) 1594 (74.6%) < 0.0001
   Black 434 (9.6%) 287 (13.4%) < 0.0001
   Asian 98 (2.2%) 58 (2.7%) 0.16
   Indian 15 (0.3%) 14 (0.7%) 0.06
   Pacific 12 (0.3%) 3 (0.1%) 0.40
   Race Unknown 295 (6.5%) 183 (8.4%) 0.002
   Race Mixed 6 (0.1%) 2 (0.1%) 1.0000
Hispanic 257 (6.1%) 217 (10.8%) < 0.0001
Current smoker 1652 (36.9%) 551 (26.2%) < 0.0001
Medical history
   Hypertension 2734 (60.2%) 1800 (84.2%) < 0.0001
   Hypercholesterolemia 2216 (48.8%) 1396 (65.4%) < 0.0001
   Previous myocardial infarction 813 (17.9%) 579 (27.1%) < 0.0001
   Congestive heart failure history 323 (7.1%) 382 (17.9%) < 0.0001
   Congestive heart failure history diagnosed on arrival 359 (7.9%) 325 (15.2%) < 0.0001
   Chronic obstructive pulmonary disease 434 (9.6%) 286 (13.4%) < 0.0001
   Coronary artery bypass graft surgery 434 (9.6%) 450 (21.1%) < 0.0001
   End stage renal disease 69 (1.5%) 156 (7.3%) < 0.0001
   Cerebrovascular accident/Transient ischemic attack 368 (8.1%) 330 (15.4%) < 0.0001
Implanted cardioverter-defibrillator * 125 (2.8%) 101 (4.7%) < 0.0001
Myocardial infarction characteristics
   ST-segment elevation myocardial infarction 2136 (47.0%) 751 (35.1%) < 0.0001
Anterior 720 (33.7%) 241 (32.1%) 0.42
Inferior/Post 1090 (51.0%) 379 (50.4%) 0.77
   Non-ST segment elevation myocardial infarction 2404 (53.0%) 1386 (64.9%) < 0.0001
   Thrombolytic therapy 378 (8.3%) 121 (5.7%) 0.0001
   Primary percutaneous coronary intervention 2858 (63.0%) 1060 (49.6%) < 0.0001
   In-hospital revascularization (nonprimary percutaneous coronary intervention and coronary artery bypass graft surgery) 1065 (23.5%) 610 (28.5%) < 0.0001
   Diagnostic angiography 485 (10.7%) 245 (11.5%) 0.34
Vitals
   Body mass index (kg/m2) 28.3±6 31±7.2 < 0.0001
   Admission systolic blood pressure (mmHg) 139.6±29.6 142.0±30.6 0.003
   Admission heart rate (beats/minutes) 81.2±21.1 86.3±21.7 < 0.0001
Left ventricular ejection fraction (%) 47.6±12.7 45.4±13.5 < 0.0001
Troponin (ng/ml) 7.4 (2-30.3) 6.2 (1.7-22) 0.0006
Length of stay (days) 6.5±6.2 8.5±9.3 < 0.0001
Discharge medications
   Aspirin 4241 (93.4%) 1939 (90.7%) 0.0001
   Angiotensin converting enzyme inhibitor/Angiotensin receptor blocker 3045 (67.1%) 1383 (64.7%) 0.06
   Statin 3984 (87.8%) 1782 (83.5%) < 0.0001
   Clopidogrel 3309 (72.9%) 1453 (68.0%) < 0.0001
   Dual antiplatelet 3190 (70.3%) 1367 (64.0%) < 0.0001
   Beta-blockers
Beta-blocker dose (%) 35.9%±27.4% 42.7%±34.1% < 0.0001
Beta-blocker dose (%) 25.0% (12.5%-50%) 25.0% (25.0%-50.0%) < 0.0001
Not on beta-blocker 414 (9.2%) 152 (7.2%) < 0.0001
>0% to 12.5% 1049 (23.3%) 399 (18.9%) < 0.0001
>12.5% to 25% 1556 (34.6%) 688 (32.6%) < 0.0001
>25% to 50% 1008 (22.2%) 533 (25.3%) < 0.0001
>50% 472 (10.5%) 336 (15.9%) < 0.0001
Received beta-blocker therapy within 24 hours from admission 3525 (77.6%) 1648 (77.1%) 0.7
Metoprolol 2908 (64.1%) 1233 (57.7%) < 0.0001
Carvedilol 890 (19.6%) 594 (27.8%) < 0.0001
Open in a new tab

Values are mean ± standard deviation, n (%), or median (Q1-Q3).

*

Includes patients with pre-admission implanted cardioverter-defibrillator and those discharged with an implanted cardioverter-defibrillator.

RESULTS

Diabetes mellitus was present in 32% (2137) of patients with acute MI discharged from the hospital. Table 1 compares patient characteristics for those with versus without DM. Detailed patients characteristics stratified by beta-blocker dose categories for patients with and without DM can be found in table S1. Prior MI, congestive heart failure, hypertension, coronary artery bypass graft surgery, end-stage renal disease and prior cerebrovascular accident were more prevalent in patients with DM. Left ventricular ejection fraction was not significantly different. Patients with DM had a lower incidence of ST-segment elevation MI (35.1%, versus 47.0%, p<0.0001) and higher incidence of non-ST segment elevation MI (64.9% versus 53.0%, p<0.0001) compared to those without DM. There was no significant difference in beta-blocker initiation within the first 24 hours of admission in patients with compared to without DM (77.1% versus 77.6%, p=0.71). Fewer patients with DM were discharged without beta-blockers therapy (7% versus 9% , p<0.0001). The overall mean discharge beta-blocker dose was 42.7±34.1% in patients with DM vs 35.9±27.4% in nondiabetics (p<0.0001). At discharge, patients with DM tended to receive higher doses of beta-blockers with 18.9%, 32.6%, 25.3%, and 15.9% receiving >0-12.5%, >12.5-25%, >25-50%, and >50% of the target dose, respectively. For patients without DM, 23.3%, 34.6%, 22.2%, and 10.5% received >0-12.5%, >12.5-25%, >25-50%, and >50% of the target dose, respectively, p<0.0001 (table 1). More patients with DM (44.5%) were in the higher beta-blocker dose categories (>25%-50%, and >50%) compared to non-diabetics (36.3%, p<0.0001). At 1 year follow-up, 59.7%, 62.0%, 57.6%, and 64.8% patients with DM continued to receive the same dose as at discharge in the >0-12.5%, >12.5-25%, >25-50%, and >50% beta-blocker target dose groups, respectively (Figure S1). In patients without DM, there was a similar distribution (Figure S1).

Figure 1 shows the 3-year mortality estimates were 24.4% for patients with DM and 12.8% for non-diabetics (p<0.0001), respectively, with an unadjusted hazard ratio (HR)=1.820 (CI:1.587-2.086, p<0.0001). Survival curves stratified by dose categories according to diabetes status are shown in figure 2. Patients discharged without beta-blockers had the highest mortality rate. Among patients discharged on beta-blockers, patients with DM in the >12.5% to 25% dose category had higher survival rates compared to all other dose categories, while patients in the >50% dose had the lowest survival rate (p<0.0001). Similar results were observed in non-diabetic patients (p<0.0001). These results follow a similar trend to what was reported in the OBTAIN study.6 DM was associated with increased mortality across all beta-blocker dose categories (table 2). After propensity score adjustment, the effect of DM was somewhat less and significant only in the >50% and >25% to 50% beta-blocker dose categories (table 2).

Figure 1:

Figure 1:

Open in a new tab

Three-year unadjusted Kaplan-Meier survival curve for patients post myocardial infarction with or without diabetes.

Figure 2:

Figure 2:

Open in a new tab

Beta-blockers after myocardial infarction in patients with and without diabetes: Unadjusted Kaplan-Meier survival curves for the 5 discharge doses (no beta-blocker [BB] and >0% to 12.5%, >12.5% to 25%, >25% to 50%, and >50% of target dose) of beta-blockers.

Table 2:

The effect of diabetes mellitus on mortality in post-MI patients with or without beta-blocker therapy stratified by dose categories.

Unadjusted Hazard Ratios
Description HR 95% CI p-value
Diabetes Effect; Not on BB 1.756 1.228-2.510 0.002
Diabetes Effect; >0-12.5% 1.772 1.326-2.368 0.0001
Diabetes Effect; >12.5-25% 1.852 1.433-2.395 <0.0001
Diabetes Effect; >25-50% 1.921 1.472-2.506 <0.0001
Diabetes Effect; >50% 1.760 1.240-2.497 0.002
Propensity Score Adjusted Hazard Ratios
Description HR 95% CI p-value
Diabetes Effect; Not on BB 1.370 0.968-2.021 0.11
Diabetes Effect; >0-12.5% 1.330 0.940-1.827 0.078
Diabetes Effect; >12.5-25% 1.254 0.940-1.674 0.12
Diabetes Effect; >25-50% 1.353 1.009-1.816 0.04
Diabetes Effect; >50% 1.517 1.035-2.223 0.03
Open in a new tab

BB= beta-blockers; CI= confidence interval; HR= hazard ratio

As the lowest observed mortality in the OBTAIN study6 was at the >12.5-25% of target dose, we further explored the dose effect by diabetes group relative to the >12.5-25% target dose (figure 3, table S2). In the group with DM, HRs for 3-year mortality in post-MI patients when the other doses were compared with the >12.5-25% of target dose were 1.232 (95% confidence interval [CI]: 0.918-1.653; p=0.16), 1.437 (95% CI: 1.100-1.879; p=0.008), and 1.525 (95% CI: 1.132-2.054; p=0.006) for the >0-12.5%, >25-50% and >50% of target dose, respectively. After propensity score adjustment, none of the comparisons were statistically significant (table S2). In the group without DM, the corresponding HRs were 1.288 (95% CI: 1.001-1.658; p=0.049), 1.386 (95% CI: 1.075-1.789; p=0.01), and 1.605 (95% CI: 1.167-2.207; p=0.004), demonstrating a similar pattern of dose response. After propensity score adjustment, all the comparisons remained statistically significant (table S2). When comparing the >12.5-25% target dose to no beta-blocker therapy, similar HRs for 3-year mortality were noted: 2.444 (95% CI: 1.731-3.451; p<0.0001) in diabetics and 2.578 (95% CI: 1.960-3.392; p<0.0001) in non-diabetics (Figure 3). After propensity score adjustment, the results remained statistically significant in diabetics and non-diabetics patients (table S2).

Figure 3:

Figure 3:

Open in a new tab

Unadjusted Cox proportional hazard regression for post-myocardial infarction patients with or without diabetes across 4 beta-blocker discharged dose therapy relative to 25% dose.

In the multivariable analysis among post-MI patients with DM, all beta-blocker dose categories demonstrated lower mortality compared to no therapy; however, only the >12.5-25% dose had a statistically significant HR=0.450 (95% CI: 0.224-0.907; p=0.025). Moreover, variables such as: male sex, history of coronary artery bypass graft procedure, history of chronic obstructive pulmonary disease, history of end-stage renal disease, history of congestive heart failure, and in-hospital length of stay were associated with lower survival, while ST-segment elevation MI, in-hospital percutaneous coronary intervention, in hospital coronary artery bypass grafting, and revascularization were associated with improved survival in patient with DM.

Multivariable analysis of the post-MI patients without DM showed that all beta-blocker dose categories had lower mortality compared to no therapy however none were statistically significant. Table S3 details the set of variables that were associated with lower survival and lower mortality in this group.

We further explored the effect of metoprolol versus carvedilol on mortality in patients with and without DM. Metoprolol and carvedilol use was 57.7% and 27.8% in patients with DM, and 64.1% and 19.6% in patients without DM respectively (p<0.0001). The mean metoprolol and carvedilol doses were similar in patients with DM (42.2%±31.9% versus 40.4±34.9, p=0.28), while the metoprolol dose was slightly higher than the carvedilol dose in patients without DM (35.1±25.5% versus 32.8±27.8%, p=0.024). Patients with DM were discharged on higher doses of metoprolol or carvedilol compared to non-diabetics (p<0.0001). In patients with DM, there was no statistically significant difference in 3-year mortality among those treated with metoprolol versus carvedilol (22.7% vs 26.5%; unadjusted HR=0.818, 95% CI: 0.659-1.016, p=0.07). In patients without diabetes, there was lower mortality among patients discharged on metoprolol compared to carvedilol (10.9% vs 14.4%; unadjusted HR=0.714, 95% CI: 0.574-0.888, p=0.002).

After multivariable analysis, the effect of metoprolol was no longer significant (table S4). When stratified according to beta-blocker dose category, metoprolol had non-significantly lower mortality in patients with and without DM across all doses compared to carvedilol (p>0.05) (table S4). After multivariable adjustment, metoprolol was borderline significantly associated with improved survival compared to carvedilol in non-diabetics discharged on >50% dose (adjusted HR=0.551, 95% CI:0.303-1.001, p=0.05) (table S4).

We further explored the effect of left ventricular ejection fraction on mortality in patients with and without DM. In patients with DM, patients with reduced left ventricular ejection fraction of ≤40% had more than double the risk of 3-year mortality compared to preserved ejection fraction of >40% with unadjusted HR= 2.201, 95%CI: 1.806-2.682, p<0.001 (adjusted HR=1.906, 95%CI 1.483-2.450, p<0.001). Patients without DM and reduced left ventricular ejection fraction of ≤40% also had more than double the risk of 3-year mortality compared to preserved ejection fraction of >40%; however, after multivariable adjustment, the effect was no longer significant (table S5). When stratified according to beta-blocker dose categories, reduced left ventricular ejection fraction ≤40% was associated with increased 3-year mortality across all categories compared to >40% (table S5). After multivariable adjustment, the effect was no longer significant in the >12.5% to 25% and >0% to 12.5% beta blocker dose categories, while patients not receiving beta blockers had more than 10-fold increase in 3-year mortality (adjusted HR= 10.439, 95%CI 2.089-52.176, p=0.004), followed by >50% (adjusted HR=3.105, 95%CI:1.711-5.635, p<0.001) and >25% to 50% (adjusted HR=2.678, 95%CI:1.555-4.613, p<0.001) beta blocker dose categories (table S5). Figure 4 shows the adjusted mortality rate in patients with and without diabetes post-MI stratified by left ventricular ejection fraction (>40% or ≤40%). Diabetic patients post-MI with reduced left ventricular ejection fraction had the lowest mortality when prescribed >12.5%-25% beta-blocker dose and the highest mortality when they were discharged on no beta-blocker therapy, followed by beta-blocker dose >50% (figure 4). The trends were similar among diabetics with left ventricular ejection fraction >40% as well as non-diabetics with left ventricular ejection fraction >40% or ≤40%. There was no significant interaction between left ventricular ejection fraction and beta-blocker dose categories on survival among patients with or without DM.

Figure 4:

Figure 4:

Open in a new tab

Adjusted mortality rate in patients with and without diabetes post-myocardial infarction stratified by left ventricular ejection fraction (>40% or ≤40%).

DISCUSSION

In this OBTAIN substudy evaluating the effect of beta-blocker dosing on outcomes post-MI in patients with DM, there are several notable findings. As expected, we demonstrated overall worse outcomes in patients with DM compared to without DM post-MI. Interestingly, as drug dosing in this registry was by physician choice, practice patterns emerged including prescription of higher doses of beta-blockers and greater use of carvedilol versus metoprolol in patients with DM. This suggests that physicians recognized the heightened risk of MI related mortality and the potential complication of cardioselective beta-blockers in patients with DM. However, in this registry, neither of these changes in physician prescription patterns was associated with an improved outcome. In all patients, the greatest benefit was observed with >12.5-25% of the target dose used in the randomized clinical trials that established the survival benefit of beta-blockers post-MI. These findings further support the need for a critical reevaluation of optimal dosing of beta-blocker therapy and type selection post-MI.

We observed a significantly lower survival among patients with DM out to three years post-MI. A recent meta-analysis study that included 10 randomized controlled trials and 56 cohort studies showed significantly higher risk of death (unadjusted HR=1.82, 95% CI: 1.73 to 1.91) in post-MI patients with DM compared to those without.14 After adjusting for confounders, mortality remained higher in patients with DM (adjusted HR=1.48, 95% CI:1.43-1.53). The effect of DM was apparent irrespective of acute MI phenotype and modern treatment (unadjusted HR 1.82, 95% CI: 1.70-1.95).14 This is in line with our results.

While the survival benefit of beta-blocker therapy in patients post-MI has been well established, the adoption in clinical practice of prescribing the target doses used in the clinical trials that established their efficacy has not occurred. Several registries,15,16 including OBTAIN,17 show that the majority of post-MI patients receive well below these target beta-blocker doses. It is interesting to note that there are no prospective trials of variable doses of beta-blockers post-MI. One retrospective study showed there was no difference in mortality post acute coronary syndrome between high dose beta-blocker group (defined as ≥50% target dose) and low dose beta-blocker group (defined as <50% target dose) with adjusted HR =1.18, 95% CI 0.72-1.96, p=0.5118. A recent study, from the SWEDEHEART registry, looked at 97575 patients admitted for first-time MI and showed no superiority of high beta-blocker dose (≥50%) over low dose (<50%) at 1 year follow-up and up to 5 years.15 In another study,19 medium beta-blocker dose (defined as ≥25% to <50%), when compared to high dose (defined as ≥50%), was found to have lower 3 years all-cause mortality (adjusted HR 0.65, 95% CI:0.4-1.07, p=0.12) and cardiac mortality (adjusted HR 0.49, 95% CI:0.25-0.96, p=0.04). The results were similar when medium dose was compared to low dose (defined as <25%) for all-cause mortality (HR 0.59, 95% CI:0.42-0.82, p=0.005) and cardiac mortality (HR 0.46, 95% CI:0.29-0.74, p=0.002).19 Thus, despite the fact that there are no randomized clinical trial data, the existing reports do not support a benefit for the higher target dose beta-blocker therapy used in the clinical trials. De facto practice patterns have gravitated to lower doses. Yet, in post-MI patients who are at increased risk, there may be some benefit to higher doses. For example, Ajam et al20 found that higher beta-blocker dose (defined as ≥50%) was associated with lower overall mortality compared to low (defined as <50%) beta-blocker dose (HR=0.75, 95% CI:0.73-0.77, p<0.01) in heart failure patients with reduced ejection fraction independent of heart rate. While the increased mortality among patients with DM post-MI was associated with physician prescription of higher doses, in this particular subgroup there was no benefit to this practice in the current registry.

There is no consensus on the type of beta-blocker that should be prescribed in patients with DM post-MI. In the OBTAIN registry,6 metoprolol and carvedilol were the two most common beta-blockers prescribed, accounting for 93% of all beta-blockers. Metoprolol is cardioselective and acts primarily by reducing heart rate and myocardial contractility leading to compensatory peripheral vasoconstriction, and resulting in increased insulin resistance.21 In contrast, vasodilating beta-blockers such as carvedilol have shown neutral or beneficial effects on metabolic parameters in patients with diabetes and hypertension. GEMINI trial7 was a multicentre 35-week randomized trial of carvedilol versus metoprolol in patients with diabetes and hypertension, showing small but significant differences in hemoglobin A1c levels (0.13% lower with carvedilol, p=0.004), improved insulin sensitivity (7.2% lower with carvedilol, p=0.004), and lower cholesterol and triglyceride levels in carvedilol when compared to metoprolol. Also, there was more weight gain (1 kg more with metoprolol, p<0.001) and a greater progression to microalbuminuria in the metoprolol group compared with those randomized to carvedilol (10.3% vs. 6.4%, p=0.04).7 Furthermore, it was found that more diabetic patients taking metoprolol had initiation of statin therapy or their statin dose increased compared with those in the carvedilol group (32% vs. 11%, p=0.04).22 In contrast to metoprolol, carvedilol has antioxidative properties which may have beneficial effects on endothelial dysfunction caused by oxidative stress in patients with type 2 diabetes. In type 2 diabetes, an increased production of free radicals leads to an increased oxidative stress to the vascular wall.23 In one study, patients post-MI prescribed a “diabetic friendly” beta-blocker like carvedilol did trend toward a lower risk of worsened glucose control at 6 months compared to “non diabetic friendly” beta-blockers like metoprolol (RR=0.8; 95% CI:0.6-1.08).21 In another study, a lower rate of onset of type 2 diabetes was noted among post-MI patients on carvedilol compared to metoprolol (HR=0.83, 95% CI: 0.75-0.91, p<.0.0001), however there was no survival benefit over metoprolol in diabetic patients (HR-0.97, 95% CI: 0.9-1.05, p=non-significant).24 This is consistent with our findings. Despite higher rates of carvedilol prescription in patients with DM, we did not observe improved outcomes compared to metoprolol in these patients.

This analysis has several limitations. This is a non-randomized post hoc analysis with significant clinical differences between patients with and without DM. Even though multivariable adjustment and propensity score analysis were employed, there are potential confounding variables, particularly pertaining to diabetes such as hemoglobin A1c, that were not accounted for. We also recognize that there is a correlation between glycemic control and cardiovascular events.25,26 Finally, the analysis is based upon discharge beta-blocker type and dose; as we have shown in another OBTAIN substudy,17 approximately 60% of the patients remained on the same dose at one year as they were at discharge; it is possible that changing doses over time could impact the outcome.

In conclusion, post-MI patients with DM had significantly higher mortality compared to non-diabetics. In contrast to what we hypothesized, even patients with diabetes appeared to have the best outcome when treated with >12.5-25% of the target beta-blocker dose, but certainly did not derive greater benefit from higher doses. There also did not appear to be any added benefit to the use of a non-selective beta-blocker in patients with DM. These findings should be instrumental in the design of future studies assessing dose-dependent effects of beta-blockers post-MI, particularly in patients with DM.

Supplementary Material

1

FUNDING SOURCE

This research was supported by grant #5U01HL080416 from the National Heart, Lung, and Blood Institute of the National Institutes of Health. Dr. Goldberger receives funding from the Miami Heart Research Institute

Footnotes

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Disclosures: JDA: Research- Boston Scientific, Microport. Consulting -Abbott, Medtronic, Penumbra, Shockwave, Philips, Recor.

Jeffrey J Goldberger reports financial support was provided by National Institutes of Health.

References:

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

1
NIHMS1900697-supplement-1.docx (7.3MB, docx)

RESOURCES