The beneficial effects of β blockers in younger patients with heart failure (HF) due to systolic dysfunction are well established. However, data from patients ≥70 years old with diabetes mellitus and HF are lacking. The Study of Effects of Nebivolol Intervention on Outcomes and Rehospitalization in Seniors with heart failure [SENIORS] tested the efficacy of the vasodilator β blocker nebivolol in patients ≥70 years old with HF and impaired or preserved left ventricular ejection fraction. In the present analysis, we evaluated the association between diabetes mellitus and baseline glucose levels on the primary outcome (all-cause mortality and cardiovascular hospitalization) and secondary end points, including all-cause mortality, cardiovascular hospitalizations, and cardiovascular mortality. Of 2,128 patients, 555 (26.1%) had diabetes mellitus. Of the 555 patients with diabetes mellitus, 223 (40.2%) experienced the primary end point compared to 484 (30.8%) of the 1,573 nondiabetic patients (p <0.001). For the nondiabetic patients, the rate of the primary outcome for placebo compared to nebivolol was 33.7% for the placebo group and 27.8% for the nebivolol group (hazard ratio 0.78, 95% confidence interval 0.65 to 0.93; p = 0.006). In the diabetic subset, the rate was 40.3% for the placebo group and 40.1% for the nebivolol group (hazard ratio 1.04, 95% confidence interval 0.80 to 1.35, p = 0.773). The subgroup interaction p value was 0.073. The baseline glucose levels in the nondiabetic patients did not significantly affect the outcomes. The effect of diabetes mellitus on outcome was independent of the left ventricular ejection fraction and was most pronounced in those with HF due to a nonischemic etiology. In conclusion, in patients ≥70 years old with HF, diabetes mellitus was associated with a worse prognosis. Nebivolol was less effective in the patients with diabetes and HF than in those with HF but without diabetes who were ≥70 years old.
Diabetes mellitus is a very common co-morbidity in patients with heart failure (HF). The prevalence of HF in those with diabetes is also about 20%. The outcomes in patients with diabetes and HF have been far worse than those in patients with HF but without diabetes. Although earlier data suggested that β blockers might increase blood glucose level and thus aggravate diabetes mellitus, more recent data have shown that β blockers are well tolerated in patients with HF and diabetes mellitus, although β blockers might be less effective in patients with diabetes and HF than in nondiabetic patients with HF.
The Study of Effects of Nebivolol Intervention on Outcomes and Rehospitalization in Seniors with heart failure (SENIORS) specifically included older patients (≥70 years old) with HF, with either impaired or preserved left ventricular ejection fraction, and tested the efficacy of the third-generation β blocker nebivolol on outcome (mortality or hospitalization for cardiovascular causes). In the present post hoc analysis, we report the association between diabetes mellitus and mortality and morbidity in the SENIORS cohort and the effects of nebivolol in patients with and without diabetes mellitus.
Methods
The present study was a substudy of the SENIORS, which was published in 2005. The present study conformed with the Declaration of Helsinki. The local ethical committees approved the protocol, and all study participants provided written informed consent.
The SENIORS was a parallel group, randomized, double-blind, multicenter, international trial comparing nebivolol and placebo in patients with HF receiving optimal standard therapy. To be eligible, patients were required to be ≥70 years old and have a clinical history of HF with ≥1 of the following features: documented hospital admission within the previous 12 months with a discharge diagnosis of HF or documented left ventricular ejection fraction of ≤35% within the previous 6 months. The main exclusion criteria were any recent change in cardiovascular drug therapy, contraindications to β blockers, and significant hepatic or renal dysfunction. The study medication (nebivolol or placebo) was titrated during a 16-week period from a starting dose of 1.25 mg/day to a target of 10 mg/day. The mean duration of patient follow-up was 21 months (SD 9, interquartile range 14 to 29). The baseline left ventricular ejection fraction (LVEF) was measured by echocardiography in 94%, nuclear imaging in 4%, and magnetic resonance imaging in 2% of the patients.
Diabetes mellitus was recorded according to the known diagnosis and/or the use of antidiabetic therapy. At baseline, the fasting or nonfasting blood glucose level was measured according to the SENIORS protocol and was available for 2,108 patients. The mean fasting blood glucose level (116 ± 44 mg/dl; n = 1,312) and nonfasting blood glucose level (116 ± 48 mg/dl; n = 796) were very similar; thus, we chose to include all baseline blood glucose values in our analysis.
The influence of diabetes mellitus on the effect of nebivolol in the SENIORS population was measured as a dependent variable (categorical). The categories were prespecified as diabetes mellitus or no diabetes mellitus. For the analysis on the importance of the blood glucose level (fasting), we divided the nondiabetic patients into 3 groups. The blood glucose level was categorized into tertiles: first and lowest tertile, 16.7 to 91.4 mg/dl; second tertile, 91.5 to 103.0 mg/dl; and third tertile, 103.0 to 296.5 mg/dl. The baseline characteristics of the patients with and without diabetes mellitus are presented as the mean ± SD or median (interquartile range) for the continuous variables and by frequency (percentage) for the categorical variables. The statistical tests were 2-sided, and p <0.05 was taken as the level of significance. Hazard ratios (HR) were calculated using Cox proportional hazards models, with 95% confidence intervals (CIs).
We used multiple logistic regression analysis to model the influence of diabetes mellitus with other baseline characteristics (age, gender, LVEF, New York Heart Association functional class, serum creatinine, blood pressure, assigned therapy, hypertension, and previous myocardial infarction) to understand any interaction of these variables on outcomes using the original randomized assignment of nebivolol versus placebo. The primary outcome of interest was the composite of all-cause mortality or cardiovascular (CV) hospital admission (interval to first event). The secondary outcomes were all-cause mortality, CV hospitalization, and CV mortality and are also presented when considered appropriate. We used an intention-to-treat analysis throughout.
Results
A total of 2,128 patients were randomized, of whom 555 (26.1%) had diabetes mellitus. Of the 2,128 patients, 1,067 were randomized to nebivolol, of whom 287 (26.9%) had diabetes mellitus, and 1,061 were randomized to placebo, of whom 268 (25.3%) had diabetes mellitus. The baseline demographics and clinical parameters of all patients are listed in Table 1 . Those with diabetes were slightly younger and had greater rates of coronary artery disease, myocardial infarction, hypertension, and hyperlipidemia. The severity of HF was higher in those with diabetes, as reflected by a greater New York Heart Association class. The LVEF was comparable between the diabetic and nondiabetic patients. As expected, the renal function in diabetic patients was worse than in the nondiabetic patients, as measured by the serum creatinine and estimated glomerular filtration rate (calculated using the Modification of Diet in Renal Disease formula). Diabetic patients more often used lipid-lowering drugs and aldosterone receptor antagonists ( Table 1 ).
| Characteristic | Diabetic Subjects | p Value Comparing Nebivolol and Placebo Groups | Nondiabetic Subjects | p Value Comparing Nebivolol and Placebo Groups | Whole Group | p Value Comparing Diabetic and Nondiabetic Groups | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Nebivolol | Placebo | Whole Group | Nebivolol | Placebo | Whole Group | |||||
| Patients (n) | 287 | 268 | 555 | 780 | 793 | 1573 | 2128 | |||
| Demographics | ||||||||||
| Age (years) | ||||||||||
| Mean (SD) | 76 (4) | 76 (5) | 76 (4) | 0.754 | 76 (5) | 76 (5) | 76 (5) | 0.917 | 76.1 (4.7) | 0.014 |
| Median (IQR) | 75 (72, 78) | 75 (72, 78) | 75 (72, 78) | 75 (72, 79) | 76 (73, 79) | 75 (72, 79) | 75 (72, 79) | |||
| Women (n) | 109 (38%) | 102 (38%) | 211 (38%) | 0.984 | 301 (39%) | 273 (34%) | 574 (37%) | 0.086 | 785 (37%) | 0.521 |
| Clinical | ||||||||||
| New York Heart Association class | ||||||||||
| I | 7 (2%) | 4 (2%) | 11 (2%) | 0.588 | 25 (3%) | 25 (3%) | 50 (3%) | 0.469 | 61 (2%) | 0.023 |
| II | 150 (52%) | 139 (52%) | 289 (52%) | 453 (56%) | 458 (58%) | 911 (58%) | 1,200 (56%) | |||
| III | 122 (43%) | 121 (45%) | 243 (44%) | 291 (37%) | 290 (37%) | 581 (37%) | 824 (39%) | |||
| IV | 8 (3%) | 4 (2%) | 12 (2%) | 11 (1%) | 20 (3%) | 31 (2%) | 43 (2%) | |||
| Ejection fraction (%) | ||||||||||
| Mean (SD) | 35 (12) | 36 (12) | 36 (12) | 0.057 | 37 (13) | 36 (12) | 36 (13) | 0.269 | 36 (12) | 0.211 |
| Median (IQR) | 32 (27, 40) | 34 (29, 43) | 33 (28, 42) | 34 (29, 42) | 33 (29, 41) | 34 (29, 42) | 33 (28, 2) | |||
| Hemodynamics | ||||||||||
| Heart rate (beats/min) | 80 (13) | 79 (14) | 80 (13) | 0.899 | 79 (14) | 79 (14) | 79 (14) | 0.671 | 79 (14) | 0.379 |
| Sitting systolic blood pressure (mm Hg) | 139 (21) | 139 (20) | 139 (20) | 0.924 | 138 (20) | 140 (22) | 139 (21) | 0.288 | 139 (21) | 0.86 |
| Sitting diastolic blood pressure (mm Hg) | 80 (10) | 80 (11) | 80 (11) | 0.919 | 81 (11) | 81 (12) | 81 (11) | 0.852 | 81 (11) | 0.033 |
| Medical history | ||||||||||
| Smoker | 15 (5%) | 10 (4%) | 25 (5%) | 0.396 | 37 (5%) | 47 (6%) | 84 (5%) | 0.299 | 109 (5%) | 0.441 |
| History of coronary artery disease | 209 (73%) | 189 (71%) | 398 (72%) | 0.548 | 526 (67%) | 528 (67%) | 1,054 (67%) | 0.719 | 1,452 (68%) | 0.041 |
| Previous myocardial infarction | 152 (53%) | 121 (45%) | 273 (49%) | 0.066 | 315 (40%) | 342 (43%) | 657 (42%) | 0.27 | 930 (44%) | 0.002 |
| Previous percutaneous coronary intervention | 20 (7%) | 11 (4%) | 31 (6%) | 0.142 | 27 (4%) | 23 (3%) | 50 (3%) | 0.526 | 81 (4%) | 0.011 |
| Previous coronary artery bypass surgery | 40 (14%) | 25 (9%) | 65 (12%) | 0.092 | 61 (8%) | 69 (9%) | 130 (8%) | 0.526 | 195 (9%) | 0.016 |
| Cerebrovascular accident in previous 3 months | 1 (0%) | 0 (0%) | 1 (0%) | 0.333 | 0 (0%) | 0 (0%) | 0 (0%) | NA | 1 (0%) | 0.092 |
| Hypertension | 182 (63%) | 188 (70%) | 370 (67%) | 0.093 | 470 (60%) | 472 (60%) | 942 (60%) | 0.766 | 1,312 (62%) | 0.005 |
| Hyperlipidemia | 145 (51%) | 135 (50%) | 280 (51%) | 0.972 | 345 (44%) | 349 (44%) | 694 (44%) | 0.93 | 974 (46%) | 0.01 |
| Atrial fibrillation | 84 (29%) | 85 (32%) | 169 (31%) | 0.531 | 277 (36%) | 292 (37%) | 569 (36%) | 0.589 | 738 (35%) | 0.015 |
| Medications | ||||||||||
| Diuretic | 256 (89%) | 233 (87%) | 489 (88%) | 0.411 | 668 (86%) | 677 (85%) | 1,345 (86%) | 0.88 | 1,834 (86%) | 0.127 |
| Angiotensin-converting enzyme inhibitor | 231 (81%) | 229 (86%) | 460 (83%) | 0.121 | 645 (83%) | 651 (82%) | 1,296 (82%) | 0.755 | 1,756 (83%) | 0.793 |
| Angiotensin II antagonist | 26 (9%) | 29 (11%) | 55 (10%) | 0.488 | 61 (8%) | 62 (8%) | 123 (8%) | 0.999 | 178 (8%) | 0.126 |
| Aldosterone antagonist | 91 (32%) | 85 (32%) | 176 (32%) | 0.998 | 210 (27%) | 194 (25%) | 404 (25%) | 0.264 | 580 (27%) | 0.006 |
| Cardiac glycoside | 125 (44%) | 120 (45%) | 245 (44%) | 0.772 | 311 (40%) | 339 (43%) | 650 (41%) | 0.247 | 895 (42%) | 0.247 |
| Antiarrhythmic | 44 (15%) | 45 (17%) | 89 (16%) | 0.639 | 122 (16%) | 149 (19%) | 271 (17%) | 0.098 | 360 (17%) | 0.52 |
| Lipid-lowering drug | 78 (27%) | 75 (28%) | 153 (28%) | 0.832 | 147 (19%) | 163 (21%) | 310 (20%) | 0.394 | 463 (22%) | <0.001 |
| Vitamin K antagonist | 67 (23%) | 63 (24%) | 130 (23%) | 0.964 | 166 (21%) | 199 (25%) | 365 (23%) | 0.073 | 495 (23%) | 0.916 |
| Aspirin | 157 (55%) | 148 (55%) | 305 (55%) | 0.902 | 408 (52%) | 392 (49%) | 800 (51%) | 0.254 | 1,105 (52%) | 0.097 |
| Calcium antagonist | 40 (14%) | 42 (16%) | 82 (15%) | 0.565 | 89 (11%) | 110 (14%) | 199 (13%) | 0.142 | 281 (13%) | 0.204 |
| Dose nebivolol achieved (mg) | 7.3 (3.8) | 8.3 (3.1) | 7.8 (3.5) | 0.001 | 7.69 (3.6) | 8.4 (3.2) | 8.0 (3.4) | <0.001 | 7.9 (3.4) | 0.203 |
| Creatinine (mg/dl) | 1.2 (0.4) | 1.2 (0.5) | 1.2 (0.4) | 0.272 | 1.1 (0.4) | 1.2 (0.4) | 1.1 (0.04) | 0.608 | 1.2 (0.4) | 0.004 |
| Modification of Diet in Renal Disease (ml/min) | 64 (22) | 62 (21) | 63 (21) | 0.371 | 66 (20) | 66 (20) | 66 (20) | 0.769 | 65 (20) | 0.011 |
Of the 555 patients with diabetes, 223 (40.2%) experienced the primary outcome (all-cause mortality or CV hospitalization). Of the 1,573 nondiabetic patients, 484 (30.8%) experienced the primary outcome (p <0.001; Figure 1 ). All-cause mortality was also increased in those with diabetes (p <0.01; Figure 1 ).
On univariate and multivariate analyses, diabetes mellitus was one of the independent predictors for developing the primary outcome (HR 1.36, 95% CI 1.15 to 1.60, p <0.001) and all-cause mortality (HR 1.25, 95% CI 0.99 to 1.58, p = 0.056), after adjustment for age, gender, LVEF, serum creatinine, blood pressure, and previous myocardial infarction ( Table 2 ).
| Variable | Univariate Analysis | Multivariate Model | ||||||
|---|---|---|---|---|---|---|---|---|
| HR | 95% CI | p Value | Chi-Square | HR | 95% CI | p-Value | Chi-Square | |
| Primary outcome | ||||||||
| Diabetes | 1.44 | 1.23–1.69 | <0.001 | 16.24 | 1.36 | 1.15–1.60 | <0.001 | 10.25 |
| Women | 0.76 | 0.65–0.89 | 0.001 | 11.44 | 0.85 | 0.72–1.01 | 0.069 | 3.20 |
| Age (1-year increase) | 1.03 | 1.02–1.05 | <0.001 | 16.55 | 1.02 | 1.01–1.04 | 0.003 | 7.27 |
| Left ventricular ejection fraction (≤35% vs >35%) | 0.86 | 0.73–1.00 | 0.056 | 3.72 | — | — | — | — |
| Left ventricular ejection fraction (1% increase) | 0.99 | 0.98–1.00 | 0.003 | 8.95 | 1.00 | 0.99–1.01 | 0.865 | 0.03 |
| New York Heart Association class (1-class increase) | 1.71 | 1.51–1.94 | <0.001 | 68.47 | 1.65 | 1.45–1.87 | <0.001 | 68.47 |
| Creatinine (greater than vs less than median) | 1.48 | 1.28–1.72 | <0.001 | 26.75 | — | — | — | — |
| Creatinine (continuous, 0.1-mg/dl increase) ⁎ | 1.07 | 1.05–1.09 | <0.001 | 47.55 | 1.05 | 1.03–1.07 | <0.001 | 47.61 |
| Systolic blood pressure (10-mm Hg increase) | 0.91 | 0.88–0.95 | <0.001 | 22.37 | 0.94 | 0.90–0.98 | 0.002 | 10.93 |
| Diastolic blood pressure (10-mm Hg increase) | 0.86 | 0.80–0.92 | <0.001 | 18.35 | — | — | — | — |
| Hypertension | 0.91 | 0.78–1.06 | 0.223 | 1.48 | 1.06 | 0.90–1.26 | 0.489 | 0.45 |
| Previous myocardial infarction | 1.51 | 1.30–1.76 | <0.001 | 29.58 | 1.30 | 1.11–1.51 | 0.001 | 17.78 |
| Assigned therapy (nebivolol vs placebo) | 0.84 | 0.73–0.98 | 0.025 | 5.03 | 0.85 | 0.73–0.99 | 0.034 | 4.72 |
| All-cause mortality | ||||||||
| Diabetes | 1.33 | 1.06–1.66 | 0.014 | 5.79 | 1.25 | 0.99–1.58 | 0.056 | 3.32 |
| Women | 0.58 | 0.46–0.74 | <0.001 | 21.43 | 0.73 | 0.57–0.94 | 0.015 | 9.71 |
| Age (1-year increase) | 1.05 | 1.03–1.08 | <0.001 | 22.66 | 1.04 | 1.02–1.07 | 0.000 | 10.22 |
| Left ventricular ejection fraction (≤35% vs >35%) | 0.74 | 0.59–0.93 | 0.009 | 7.05 | — | — | — | — |
| Left ventricular ejection fraction (1% increase) | 0.98 | 0.97–0.99 | <0.001 | 18.02 | 0.99 | 0.98–1.00 | 0.149 | 2.32 |
| New York Heart Association class (1-class increase) | 1.64 | 1.38–1.95 | <0.001 | 30.18 | 1.58 | 1.32–1.89 | 0.000 | 30.13 |
| Creatinine (greater than vs less than median) | 1.95 | 1.57–2.42 | <0.001 | 37.92 | — | — | — | — |
| Creatinine (continuous, 0.1-mg/dl increase) ⁎ | 1.10 | 1.08–1.12 | <0.001 | 59.63 | 1.07 | 1.05–1.10 | 0.000 | 59.63 |
| Systolic blood pressure (10-mm Hg increase) | 0.86 | 0.81–0.91 | <0.001 | 31.14 | 0.92 | 0.87–0.98 | 0.006 | 20.26 |
| Diastolic blood pressure (10-mm Hg increase) | 0.78 | 0.70–0.86 | <0.001 | 26.36 | — | — | — | — |
| Hypertension | 0.72 | 0.58–0.89 | 0.002 | 9.23 | 0.93 | 0.74–1.18 | 0.547 | 0.36 |
| Previous myocardial infarction | 1.64 | 1.33–2.02 | <0.001 | 21.48 | 1.30 | 1.05–1.61 | 0.017 | 6.93 |
| Assigned therapy (nebivolol vs placebo) | 0.87 | 0.71–1.07 | 0.188 | 1.73 | 0.88 | 0.71–1.08 | 0.217 | 1.48 |
The rate of the primary outcome in nondiabetic patients was significantly reduced with the use nebivolol compared to placebo (HR 0.78, 95% CI 0.65 to 0.93, p <0.01; Table 3 ). However, a similar decrease was not observed in the diabetic patients (HR 1.04, 95% CI 0.80 to 1.35, p = 0.77). The subgroup interaction showed borderline significance (p = 0.073), suggesting that the presence of diabetes mellitus might modulate the effect of nebivolol negatively. The potential lesser response to nebivolol in diabetic patients seemed to be consistent for the other secondary outcome domains (all-cause mortality, HF hospitalization, and CV mortality; Table 3 ). Figure 2 shows the Kaplan-Meier curves, showing the cumulative proportion of patients experiencing the primary outcome in diabetic and nondiabetic patients, stratified by the use of placebo and nebivolol.
| Outcome | Diabetic Subjects | Nondiabetic Subjects | Interaction p value | ||||||
|---|---|---|---|---|---|---|---|---|---|
| Nebivolol (n = 287) | Placebo (n = 268) | HR (95% CI) | p Value | Nebivolol (n = 780) | Placebo (n = 793) | HR (95% CI) | p Value | ||
| All-cause mortality or cardiovascular hospitalization (primary outcome) | 115 (40%) | 108 (40%) | 1.04 (0.80–1.35) | 0.773 | 217 (28%) | 267 (34%) | 0.78 (0.65–0.93) | 0.006 | 0.073 |
| All-cause mortality | 59 (21%) | 54 (20%) | 1.03 (0.71–1.48) | 0.892 | 110 (14%) | 138 (17%) | 0.80 (0.62–1.02) | 0.077 | 0.275 |
| All-cause mortality or heart failure hospitalization | 90 (31%) | 79 (30%) | 1.10 (0.81–1.49) | 0.542 | 164 (21%) | 192 (24%) | 0.84 (0.69–1.04) | 0.112 | 0.161 |
| Cardiovascular mortality | 42 (15%) | 42 (16%) | 0.93 (0.61–1.43) | 0.753 | 81 (10%) | 103 (13%) | 0.79 (0.59–1.06) | 0.113 | 0.534 |
