Association of longitudinal measures of hemoglobin and outcomes after hospitalization for heart failure.

Background: Cross-sectional assessments of hemoglobin (Hb) are associated with mortality in patients with heart failure (HF). Our objectives were to characterize patterns of change in Hb over time in patients with HF and to evaluate the relationship between longitudinal measures of Hb and adverse outcomes.

Methods: The study included 2,478 patients with a primary discharge diagnosis of HF from January 2001 to December 2006. Outcomes included time to death and time to death or HF hospitalization. The association between baseline Hb and outcomes was evaluated using multivariable Cox regression. The longitudinal association was evaluated using a time-dependent Hb predictor variable and using anemia trajectory groups.

Results: For a median of 475 days, baseline Hb was associated with a trend toward increased mortality (hazard ratio [HR] 1.02, 95% CI 0.99-1.06 per g/dL decline). With a time-dependent approach, the magnitude of the association was greater (HR 1.35, 95% CI 1.30-1.39 per g/dL decline). In trajectory analysis, 35% of the cohort had variable patterns of anemia. Persistently low Hb (HR 1.65, 95% CI 1.27-2.14) and a progressive decline in Hb (HR 1.54, 95% CI 1.16-2.05) were associated with increased mortality risk. Patients with recovery of anemia had similar outcomes as those patients who are persistently nonanemic. Results were similar for the composite of death or HF hospitalization.

Conclusions: Variability in Hb over time is common in patients with HF, and declining Hb is associated with a poor prognosis. Longitudinal characterization of Hb levels has greater prognostic significance than a single measurement. Systematic surveillance of Hb levels may help identify high-risk patients with heart failure.