Diabetes, Glycemic Control: Results
FILED IN Other No Comments
Baseline characteristics of the study population (n = 839) are outlined in Table 1. Mean age was 67 years, and 200 patients (23.8%) had diabetes.
Table 1
| No diabetes | Diabetes | P value | |
|---|---|---|---|
| n | 639 | 200 | |
| Age (years) | 67.4 ± 11.0 | 65.3 ± 10.3 | 0.02 |
| Sex, male (%) | 528 (82.6) | 161 (80.5) | 0.49 |
| Race, white (%) | 416 (65.2) | 88 (44.0) | <0.001 |
| BMI (kg/m2) | 27.9 ± 4.7 | 29.9 ± 6.0 | <0.001 |
| Smoking (%) | 124 (19.4) | 34 (17.0) | 0.44 |
| Heavy alcohol use (%) | 215 (33.8) | 38 (19.0) | <0.001 |
| Physical inactivity (%) | 221 (33.1) | 85 (42.5) | 0.02 |
| LDL cholesterol (mg/dl) | 106 (34) | 100 (32) | 0.03 |
| A1C (%) | 5.5 ± 0.5 | 7.1 ± 1.4 | <0.001 |
| Systolic blood pressure (mmHg) | 132 ± 20 | 137 ± 23 | 0.003 |
| Medical history | |||
| Myocardial infarction (%) | 313 (49.1) | 106 (53.8) | 0.25 |
| Revascularization (%) | 367 (57.4) | 104 (52.3) | 0.20 |
| Medication use | |||
| ACE inhibitor/ARB (%) | 261 (40.8) | 136 (68.0) | <0.001 |
| β-Blocker (%) | 343 (53.7) | 129 (64.5) | 0.007 |
| Baseline LVEF (%) | 62.7 ± 8.6 | 63.1 ± 8.8 | 0.59 |
| Diastolic function (%) | 0.22 | ||
| Normal | 361 (63.2) | 116 (63.7) | |
| Impaired | 151 (26.4) | 40 (22.0) | |
| Pseudo/restricted | 59 (10.3) | 26 (14.3) | |
| Exercise-induced wall motion abnormalities (%) | 128 (21.5) | 44 (24.4) | 0.40 |
| Creatinine clearance (ml/min) | 82.4 ± 26.9 | 82.0 ± 31.2 | 0.87 |
| CRP (mg/l) | 4.0 ± 6.8 | 4.8 ± 7.0 | 0.11 |
Data are means ± SD unless otherwise indicated. CRP was log-transformed for statistical analysis.
Diabetes as a predictor of heart failure hospitalizations
During a mean ± SD follow-up of 4.1 ± 1.2 years, 30 (15.0%) patients with diabetes and 47 (7.4%) patients without diabetes developed heart failure. Between baseline and end of follow-up (either heart failure event or end of study), 52 patients (6.2%) had a myocardial infarction. In Fig. 1, Kaplan-Meier analysis shows the proportion of patients with hospitalizations for heart failure divided into patients with and without diabetes. In Table 2, results of the Cox regression models are presented. Diabetes was a significant predictor of heart failure hospitalization (HR 2.17 [95% CI 1.37–3.44]; P = 0.001). Diabetes remained a strong predictor of heart failure while adjustments were made for other predefined predictors of heart failure. Thus, adjustment for age, sex, race, smoking, physical inactivity, BMI, LDL cholesterol, systolic blood pressure, myocardial infarction during follow-up, LVEF, exercise-induced wall motion abnormalities (i.e., ischemia), diastolic dysfunction, or CRP did not attenuate the strength of the relationship between diabetes and heart failure. In the fully adjusted model, diabetes was associated with an increased HR for hospitalization because of heart failure (3.34 [1.65–6.76]; P = 0.001). Other significant multivariable predictors were age (years, HR 1.06), smoking status (3.01), physical inactivity (2.18), LVEF (percent, 0.94), exercise-induced wall motion abnormalities (2.34), diastolic dysfunction (1.26–4.97, depending on the grade of diastolic dysfunction), and logCRP (2.10).
Figure 1
Proportions of patients free of hospitalization for heart failure divided into patients with diabetes (· · · ·) and patients without diabetes (——).
Table 2
Diabetes and A1C as risk factors for heart failure hospitalization (multivariable Cox regression)
| Diabetes as predictor for heart failure | P value | A1C ≥6.5% as predictor for heart failure | P value | A1C (%) as predictor for heart failure | P value | |
|---|---|---|---|---|---|---|
| n | 839 | 832 | 832 | |||
| Univariable analysis | 2.17 (1.37–3.44) | 0.001 | 1.61 (0.96–2.71) | 0.071 | 1.36 (1.17–1.58) | <0.001 |
| Model 1 | 2.50 (1.57–4.01) | <0.001 | 1.72 (1.02–2.92) | 0.043 | 1.46 (1.24–1.73) | <0.001 |
| Model 2 | 2.65 (1.61–4.36) | <0.001 | 1.58 (0.90–2.78) | 0.114 | 1.50 (1.26–1.79) | <0.001 |
| Model 3 | 2.53 (1.58–4.07) | <0.001 | 1.72 (1.02–2.92) | 0.043 | 1.48 (1.25–1.76) | <0.001 |
| Model 4 | 2.79 (1.74–4.50) | <0.001 | 2.03 (1.18–3.47) | 0.010 | 1.46 (1.24–1.71) | <0.001 |
| Model 5 | 2.19 (1.29–3.71) | 0.003 | 1.50 (0.82–2.73) | 0.189 | 1.33 (1.09–1.61) | 0.004 |
| Model 6 | 2.60 (1.55–4.36) | <0.001 | 2.02 (1.16–3.52) | 0.014 | 1.48 (1.24–1.75) | <0.001 |
| Model 7 | 2.42 (1.50–3.90) | <0.001 | 1.71 (1.01–2.92) | 0.047 | 1.39 (1.17–1.64) | <0.001 |
| Model 8 | 2.49 (1.52–4.08) | <0.001 | 1.67 (0.98–2.84) | 0.061 | 1.45 (1.22–1.72) | <0.001 |
| Model 9 (full) | 3.34 (1.65–6.76) | 0.001 | 2.27 (1.06–4.87) | 0.036 | 1.40 (1.13–1.74) | 0.003 |
Data are HR (95% CI) unless otherwise indicated. Model 1: age, sex, and race. Model 2: age, sex, race, smoking, BMI, physical inactivity, LDL cholesterol, and systolic blood pressure. Model 3: age, sex, race, and myocardial infarction during follow-up. Model 4: age, sex, race, and LVEF. Model 5: age, sex, race, and exercise-induced wall motion abnormalities. Model 6: age, sex, race, and diastolic dysfunction. Model 7: age, sex, race, and logCRP. Model 8: age, sex, race, ACE inhibitor/ARB and β-blocker use. Model 9: age, sex, race, smoking, BMI, physical inactivity, LDL cholesterol, systolic blood pressure, myocardial infarction during follow-up, LVEF, exercise-induced wall motion abnormalities, diastolic dysfunction, logCRP, and ACE inhibitor/ARB and β-blocker use.