Plasma pro-brain natriuretic peptide and electrocardiographic changes in combination improve risk prediction in persons without known heart disease

BackgroundThough the electrocardiogram(ECG) and plasma pro-brain-natriuretic-peptide (pro-BNP) are widely used markers of subclinical cardiac injury and can be used to predict future cardiovascular disease(CVD), they could merely be markers of the same underlying pathology. We aimed to determine if ECG changes and pro-BNP are independent predictors of CVD and if the combination improves risk prediction in persons without known heart disease.MethodsPro-BNP and ECG were obtained on 5454 persons without known heart disease from the 4th round of the Copenhagen City Heart Study, a prospective cohort study. Median follow-up was 10.4 years. High pro-BNP was defined as above 90th percentile of age and sex adjusted levels. The end-points were all-cause mortality and the combination of admission with ischemic heart disease, heart failure or CVD death.ResultsECG changes were present in 907 persons and were associated with high levels of pro-BNP. In a fully adjusted model both high pro-BNP and ECG changes remained significant predictors: all-cause mortality(high pro-BNP, no ECG changes: HR: 1.43(1.12-1.82);P = 0.005, low pro-BNP, ECG changes: HR: 1.22(1.05-1.42);P = 0.009, and both high pro-BNP and ECG changes: HR: 1.99(1.54-2.59);P < 0.001), CVD event(high pro-BNP, no ECG changes: HR: 1.94(1.45-2.58);P < 0.001, low pro-BNP, ECG changes: HR: 1.55(1.29-1.87);P < 0.001, and both high pro-BNP and ECG changes: HR: 3.86(2.94-5.08);P < 0.001). Adding the combination of pro-BNP and ECG changes to a fully adjusted model correctly reclassified 33.9%(26.5-41.3);P < 0.001 on the continuous net reclassification scale for all-cause mortality and 49.7%(41.1-58.4);P < 0.001 for CVD event.ConclusionCombining ECG changes and pro-BNP improves risk prediction in persons without known heart disease.