Background: Multi-marker risk scores accurately predict prognosis in heart failure patients but calculation is complex.
Aims: To compare the prognostic accuracy of the Seattle Heart Failure Survival Score (SHFS) and a model derived from the CHARM programme, with laboratory parameters NT-proBNP and glomerular filtration rate (GFR).
Methods and results: In a sample of 290 heart failure patients, 39 patients died, 22 were hospitalised with acute heart failure and 4 underwent urgent cardiac transplantation during a median follow-up of 498 days. NT-proBNP, GFR, CHARM and SHFS showed an AUC for an endpoint during 1-year of 0.80, 0.72, 0.79 and 0.69, respectively. The hazard ratio for an endpoint during follow-up was 2.1, 2.6, 1.9 and 2.1 per 1 SD increase of log NT-proBNP and CHARM and per 1 SD decrease of GFR and SHFS, respectively. In multivariate analysis, log NT-proBNP and GFR added independent prognostic information to CHARM and SHFS, respectively.
Conclusion: NT-proBNP and GFR independently predicted endpoint-free survival in systolic heart failure patients, with NT-proBNP being superior and equally predictive to the SHFS and CHARM score, respectively. Assessment of both laboratory markers can simplify prognostic stratification, addition to multi-marker scores should be evaluated.