Risk Prediction Models for Mortality in Ambulatory Heart Failure Patients: A Systematic Review

PurposeOptimal management of heart faikure (HF) requires accurate assessment of prognosis. Many prognostic models are available. The aim was to identify studies evaluating the use of risk prediction models for mortality in ambulatory HF patients, describe their performance and clinical applicability.Methods and MaterialsWe searched for studies in MEDLINE, EMBASE and CINAHL in May 2012 and scanned references of included studies. Two reviewers selected citations including HF patients and reporting on model performance in derivation and/or validation cohorts. We abstracted data related to population, outcomes, study quality and model discrimination and calibration.ResultsOf the 9952 reviewed, we included 32 studies testing 18 models. Only 4 models were validated in independent cohorts: the Heart Failure Survival Score (HFSS), the Seattle Heart Failure Model (SHFM), the PACE risk score and a model by Frankenstein et al. The HFSS was validated in 8 cohorts (2240 patients) showing poor to modest discrimination (c-statistic 0.56-0.79), being lower in more recent cohorts. The SHFM was validated in 14 cohorts (16057 patients) describing poor to acceptable discrimination (0.63-0.81), remaining relatively stable over time. Both models reported adequate calibration, though overestimating survival in specific populations. The other two models were validated in one cohort each, reporting poor discrimination (0.66-0.69). Among the remaining 14 models, six were validated by bootstrapping (c-statistic 0.74-0.85); the rest were not validated.ConclusionsExternally validated HF models showed inconsistent performance. The HFSS and SHFM demonstrated modest discrimination and questionable calibration. A new model derived from contemporary patient cohorts may be required for improved prognostic performance.