Background: Heart failure (HF) is a major cause of mortality and disability in type 2 diabetes mellitus (T2DM). This study sought to improve the assessment of HF risk in patients with T2DM - a step that would be critical for effective HF screening.

Methods: A systematic literature search was performed on electronic databases including MEDLINE and EMBASE, using MeSH terms ‘heart failure’, ‘risk factor’, ‘T2DM’, ‘cardiac dysfunction’, ‘stage B heart failure’, ‘incident heart failure’, ‘risk assessment’, ‘risk impact’, ‘risk score’, ‘predictor’, ‘prediction’ and related free text terms. The search was limited to human studies in full-length publications in English language journal from 1946 to 2014. Univariable and multivariable relative risk (RR) and hazard ratio (HR) were obtained from each study.

Results: Twenty-one studies (n = 1111,569, including 507,637 subjects with T2DM) were included in this analysis with a follow-up ranging from 1 to 12 years. Associations between incident HF and risk variables described in ≥ 3 studies were reported. This association was greatest for insulin use (HR 2.48; 1.24–4.99), HbA1c 7.0–8.0% (2.41; 1.62–3.59), 5 years increase in age (1.47; 1.25–1.73), fasting glucose (1.28; 1.10–1.51 per standard deviation) and HbA1c (1.18; 1.14–1.23 each 1% increase). After adjustment for confounders, there were strong associations with coronary artery disease (1.77; 1.31, 2.39), HbA1c ≥ 10% (1.66; 1.45–1.89), insulin use (1.43; 1.14–1.79), HbA1c 9.0–10.0% (1.31; 1.14–1.50), fasting glucose (1.27; 1.10–1.47 per standard deviation) and 5 years increase in age (1.26; 1.13–1.40).

Conclusion: Among patients with T2DM, five common clinical variables are associated with significantly increased risk of incident HF.