Risk factors for adverse outcomes by left ventricular ejection fraction in a contemporary heart failure population

Abstract

Background: Although heart failure (HF) is a syndrome with important differences in response to therapy by left ventricular ejection fraction (LVEF), existing risk stratification models typically group all HF patients together. The relative importance of common predictor variables for important clinical outcomes across strata of LVEF is relatively unknown.

Methods and results: We identified all members with HF between 2005 and 2008 from 4 integrated healthcare systems in the Cardiovascular Research Network. LVEF was categorized as preserved (LVEF ≥ 50% or normal), borderline (41%-49% or mildly reduced), and reduced (≤ 40% or moderately to severely reduced). We used Cox regression models to identify independent predictors of death and hospitalization by LVEF category. Among 30094 ambulatory adults with HF, mean age was 74 years and 46% were women. LVEF was preserved in 49.5%, borderline in 16.2%, and reduced in 34.3% of patients. During a median follow-up of 1.8 years (interquartile range, 0.8-3.1), 8060 (26.8%) patients died, 8108 (26.9%) were hospitalized for HF, and 20272 (67.4%) were hospitalized for any reason. In multivariable models, nearly all tested covariates performed similarly across LVEF strata for the outcome of death from any cause, as well as for HF-related and all-cause hospitalizations.

Conclusions: We found that in a large, diverse contemporary HF population, risk assessment was strikingly similar across all LVEF categories. These data suggest that, although many HF therapies are uniquely applied to patients with reduced LVEF, individual prognostic factor performance does not seem to be significantly related to level of left ventricular systolic function.

Keywords: heart failure; outcomes assessment (health care); prognosis; risk factors.

Figure

Multivariable predictors of various clinical outcomes among 30,056 adults with heart failure and documented left ventricular systolic function assessment (2005-2008). Panel A: death from any cause across LVEF strata; Panel B: hospitalization for heart failure across LVEF strata; Panel C: hospitalization for any cause across LVEF strata. HDL included in modeling, but not shown in Forest plots; see supplemental material for full model details.

Figure

Multivariable predictors of various clinical outcomes among 30,056 adults with heart failure and documented left ventricular systolic function assessment (2005-2008). Panel A: death from any cause across LVEF strata; Panel B: hospitalization for heart failure across LVEF strata; Panel C: hospitalization for any cause across LVEF strata. HDL included in modeling, but not shown in Forest plots; see supplemental material for full model details.

Figure

Multivariable predictors of various clinical outcomes among 30,056 adults with heart failure and documented left ventricular systolic function assessment (2005-2008). Panel A: death from any cause across LVEF strata; Panel B: hospitalization for heart failure across LVEF strata; Panel C: hospitalization for any cause across LVEF strata. HDL included in modeling, but not shown in Forest plots; see supplemental material for full model details.