Background: Clinical trials to date have not provided definitive evidence regarding the effects of glucose lowering on the incidence of congestive heart failure (CHF). We synthesized available prospective epidemiological data on the association between glycaemia measured by haemoglobin A(1c) (HbA(1c)) and incident CHF in individuals with diabetes.
Methods and results: We searched electronic databases and reference lists of selected articles for relevant prospective epidemiological studies. We abstracted data from relevant studies using standardized forms and obtained additional data from investigators when required. We pooled study-specific relative risk estimates using random-effects model meta-analysis. Of the 1044 citations identified, we included 10 studies comprising 178 929 participants with diabetes and 14 176 incident CHF cases. Five studies included only patients with type 2 diabetes, four studies had predominantly patients with type 2 diabetes, and one study included only patients with type 1 diabetes. All studies except one showed an increased risk of CHF with higher HbA(1c). The overall adjusted risk ratio (RR) for CHF was 1.15 [95% confidence interval (CI) 1.10-1.21] for each percentage point higher HbA(1c). There was substantial heterogeneity across the 10 studies (I(2): 83%; 95% CI 69-91%; P < 0.001) not explained by available study-level characteristics such as study design or average HbA(1c) level. In seven studies reporting RRs with more than one degree of adjustment, the association was minimally altered after adjustment for several cardiovascular risk factors.
Conclusions: In observational studies of individuals with diabetes, a higher HbA(1c) level was associated with a significantly increased incidence of CHF.