Risk Prediction Models for Mortality in Ambulatory Patients With Heart Failure

Abstract

Background— Optimal management of heart failure requires accurate assessment of prognosis. Many prognostic models are available. Our objective was to identify studies that evaluate the use of risk prediction models for mortality in ambulatory patients with heart failure and describe their performance and clinical applicability. Methods and Results— We searched for studies in Medline, Embase, and CINAHL in May 2012. Two reviewers selected citations including patients with heart failure and reporting on model performance in derivation or validation cohorts. We abstracted data related to population, outcomes, study quality, model discrimination, and calibration. Of the 9952 studies reviewed, we included 34 studies testing 20 models. Only 5 models were validated in independent cohorts: the Heart Failure Survival Score, the Seattle Heart Failure Model, the PACE (incorporating peripheral vascular disease, age, creatinine, and ejection fraction) risk score, a model by Frankenstein et al, and the SHOCKED predictors. The Heart Failure Survival Score 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 Seattle Heart Failure Model was validated in 14 cohorts (16 057 patients), describing poor-to-acceptable discrimination (0.63–0.81), remaining relatively stable over time. Both models reported adequate calibration, although overestimating survival in specific populations. The other 3 models were validated in a cohort each, reporting poor-to-modest discrimination (0.66–0.74). Among the remaining 15 models, 6 were validated by bootstrapping ( c -statistic, 0.74–0.85); the rest were not validated. Conclusions— Externally validated heart failure models showed inconsistent performance. The Heart Failure Survival Score and Seattle Heart Failure Model demonstrated modest discrimination and questionable calibration. A new model derived from contemporary patient cohorts may be required for improved prognostic performance.

Magnitude of and Prognostic Factors Associated With 1‐Year Mortality After Hospital Discharge for Acute Decompensated Heart Failure Based on Ejection Fraction Findings

“…Prior epidemiological studies have identified a number of important prognostic factors associated with poor long‐term outcomes for patients with ADHF including advanced age, male sex, hyponatremia, lower systolic blood pressure, poorer kidney function, and several comorbid conditions 3, 4, 5, 6, 7, 8, 9, 10, 11, 12. Many of these earlier studies examined the role of various prognostic factors in patients with HF that had not further been stratified according to EF findings, and among those that did,11, 13, 14, 15, 16, 17 none used the 2013 AHA/ACC guidelines recommending specific EF cut points 2.…”

New Management Strategies in Heart Failure

“…The Seattle Heart Failure Model is frequently used in the United States as a method to assess risk or need for heart transplant or VAD surgery; critics have suggested that further validation of the score in this population might be needed. 106,107 The MAGGIC score, developed in Europe, is currently used to raise awareness of the poor prognosis of patients with HF who appear well compensated. 108 There have been many analyses comparing one risk score with another in a variety of HF patient populations.…”

Model Building as an Educational Hobby

“…External validation sets often share similar limitations with the derivation set, particularly if from clinical trials or limited to a few selected sites. Risk prediction models that have undergone external validation studies in HF were recently reviewed by Alba et al 6 , who found disappointing performance. Calibration of absolute event rates was variable, overestimating mortality in some populations and underestimating in others.…”

“…The review by Alba et al 6 indicated that higher use of beta blockers and implantable cardioverter defibrillators in the validation sets was associated with worse model performance. However, it is dangerous to link on additional therapies that were not included in the models when derived.…”