Measurement of ECG abnormalities and cardiovascular risk classification: a cohort study of primary care patients in the Netherlands.

Abstract

GPs need accurate tools for cardiovascular (CV) risk assessment. Abnormalities in resting electrocardiograms (ECGs) relate to increased CV risk. To determine whether measurement of ECG abnormalities on top of established risk estimation (SCORE) improves CV risk classification in a primary care population. A cohort study of patients enlisted with academic general practices in the Netherlands (the Utrecht Health Project [UHP]). Incident CV events were extracted from the GP records. MEANS algorithm was used to assess ECG abnormalities. Cox proportional hazards modelling was applied to relate ECG abnormalities to CV events. For a prediction model only with SCORE variables, and a model with SCORE+ECG abnormalities, the discriminative value (area under the receiver operator curve [AUC]) and the net reclassification improvement (NRI) were estimated. A total of 2370 participants aged 38-74 years were included, all eligible for CV risk assessment. During a mean follow-up of 7.8 years, 172 CV events occurred. In 19% of the participants at least one ECG abnormality was found (Lausanne criteria). Presence of atrial fibrillation/flutter (AF) and myocardial infarction (MI) were significantly related to CV events. The AUC of the SCORE risk factors was 0.75 (95% CI = 0.71 to 0.79). Addition of MI or AF resulted in an AUC of 0.76 (95% CI = 0.72 to 0.79) and 0.75 (95% CI = 0.72 to 0.79), respectively. The NRI with the addition of ECG abnormalities was small (MI 1.0%; 95% CI = -3.2% to 6.9%; AF 0.5%; 95% CI = -3.5% to 3.3%). Performing a resting ECG in a primary care population does not seem to improve risk classification when SCORE information - age, sex, smoking, systolic blood pressure, and total cholesterol/HDL ratio - is already available.