Diabetes in heart failure: prevalence and impact on outcome in the population

Purpose: Little is known on the prevalence and prognostic importance of diabetes mellitus (DM) among individuals with heart failure (HF) in community-based cohorts. Methods: Within Olmsted County, Minnesota, a random sample of all subjects with a first diagnosis of HF between 1979 and 1999 was validated using Framingham criteria. DM was validated using glycemic criteria. Results: Among 665 subjects with HF (mean age 77+/-12 years, 46% male), 20% had prior DM. Subjects with DM were younger, had greater body mass index (BMI), and lower left ventricular ejection fraction than subjects without diabetes. The prevalence of DM increased markedly over time (3.8% per year; 95% confidence interval [CI], 0.8 to 6.9; P=.024), independently of BMI, particularly in older subjects (odds ratio of having DM in 1999 compared with 1979 was 3.93 [95% CI, 1.57 to 9.83] in subjects > or = 75 years vs. 1.11 [95% CI, .40 to 3.05] in subjects <75 years). Five-year survival was 37% among subjects with DM versus 46% among subjects without (P=.017). The risk of death associated with DM differed markedly according to clinical coronary artery disease (CAD) (P=.025). Subjects with DM and no CAD had a higher risk of death (relative risk [RR]=1.79 [95% CI, 1.33 to 2.41]) than those with CAD (RR=1.11 [95% CI, .81 to 1.51]), independently of age, sex, BMI, renal function, calendar year of HF, comorbidity and EF. Conclusions: Among community-dwelling patients with HF, the prevalence of DM increased markedly over time. DM is associated with a large increase in mortality, particularly among subjects without clinical CAD, underscoring the importance of aggressive management of DM in HF.