Predicting Mortality in Decompensated Heart Failure: A Retrospective Study of 1,023 Veteran Admissions

Abstract

Introduction Several risk models have been developed to predict mortality in patients with heart failure to help guide therapy and facilitate shared decision making. Implementation of these predictive models has been cautious due to a lack of generalizability and limitations in estimating mortality for individuals. Hypothesis We hypothesized that the online Meta-Analysis Global Group in Chronic Heart Failure (MAGGIC) Risk calculator, derived from patient cohorts with preserved and reduced systolic heart failure, would accurately predict 1- and 3-year mortality rates in veterans hospitalized with acute decompensated heart failure. Methods A retrospective analysis was undertaken of 1023 consecutive, unique patients admitted with the primary diagnosis of heart failure over a 10-year period at the William S. Middleton Memorial Veterans Hospital. The MAGGIC risk calculator was used to predict 1 and 3-year mortality rates based on 13 patient variables. These estimates were compared to actual survival. Results The demographics of our study population were as follows: median age 73, male 99%, diabetes 57%, COPD 48%, HF diagnosis Table 1 . To further clarify if the MAGGIC model could accurately predict mortality across various risk groups, patients were stratified into quintiles based on integer score ( Figures 1a, b ). Patients in the highest risk group had precise risk predictions at 1-year (64.49% actual versus 54.26% predicted) and at 3-years (86.92% actual versus 84.89% predicted) Conclusions Our data demonstrates that the MAGGIC risk calculator is accurate in predicting 1 and 3-year mortality rates in hospitalized veterans, independent of systolic performance. Moreover, the risk calculator effectively predicted patient mortality across multiple integer risk scores. This risk model demonstrates promise in identifying hospitalized patients at high risk for mortality and therefore may assist in decision making regarding appropriateness of pursuing advanced heart failure therapy or palliative care.