Strategy to identify subjects with diabetes mellitus more suitable for selective echocardiographic screening: The DAVID-Berg study

Background Despite the burden of pre-clinical heart failure (HF) among diabetes mellitus (DM) patients, routine screening echocardiography is not currently recommended. We prospectively assessed risk prediction for HF/death of a screening strategy combining clinical data, electrocardiogram, NTproBNP, and echocardiogram, aiming to identify DM patients more suitable for selective echocardiography. Methods Among 4047 screened subjects aged ≥ 55/≤80 years, the DAVID-Berg Study prospectively enrolled 623 outpatients with DM, or hypertension, or known cardiovascular disease but with no HF history/symptoms. The present analysis focuses on data obtained during a longitudinal follow-up of the 219 patients with DM. Results Mean age was 68 years, 61% were men, and median DM duration was 4.9 years. During a median follow-up of 5.2 years, 50 subjects developed HF or died. A predictive model using clinical data demonstrated moderate predictive power, which significantly improved by adding electrocardiogram (C-statistic 0.75 versus 0.70; p < 0.05), but not NTproBNP (C-statistic 0.72, p = 0.20). Subjects with normal clinical variables or abnormal clinical variables but normal electrocardiogram had low events rate (1.3 versus 2.4 events/100-person-years, p = NS). Conversely, subjects with both clinical and electrocardiogram abnormalities (47%) carried higher risk (9.0 events/100-person-years, p < 0.001). The predictive power for mortality/HF development increased when echocardiography was added (13.6 events/100-person-years, C-statistic 0.80, p < 0.05). Conclusions Our prospective study found that a selective echocardiographic screening strategy guided by abnormal clinical/electrocardiogram data can reliably identify DM subjects at higher risk for incident HF and death. This screening approach may hold promise in guiding HF prevention efforts among DM patients.