Abstract
The WHO definition of anemia, Hgb<12g/dL (W) and <13g/dL (M) is largely based on studies in younger populations. As our population is aging it is important to explore the impact of varying Hgb levels on mortality and function in representative populations of older persons. The Duke Established Populations for Epidemiologic Studies for the Elderly (EPESE) is a random household sample, over sampled for African Americans (AA), initiated in 1986. In 1992 we obtained Hgb levels for 1744 subjects age >71 and evaluated functional and cognitive status at baseline and 4 years later, and mortality for 8 years. Using WHO criteria, the prevalence of anemia was 24%. It was age related; 17% (65–74), 25% (75–79), 32% (>80) p<.0001, but not gender related. There was a strong racial difference, 33.6% AA, 13.8% Caucasians (p=<.0001) with an odds ratio, adjusted for age, education, renal insufficiency and comorbidity of 3.0 (95%CI 2.3 – 3.9) Baseline functional status was worse in anemic subjects (p<.0001) but cognition was not (p=.06). The risk ratio for 8 year mortality was 1.7 (95% CI 1.5, 2.0) for anemic subjects (p=.0001). This did not differ by gender nor, despite the increased prevalence, by race (RR AA: caucasian = 0.85; 95% CI 0.6–1.2). Assessing survival by Hgb level, a nonlinear relationship was seen, with highest % survival between 13–14 g/dL for women (p<.0001) and 14–15g/dL for men (p=.09). Using the WHO standard as reference, controlling for baseline characteristics, women with Hgb 10–11 g/dL had higher mortality (RR 2.1, 95% CI. 1.5 –3.1) but with 11–12 g/dL did not. Similarly men with Hgb 10–11 g/dL had increased mortality (RR2.0, 95% CI 1.1, 3.7). In women there was a progressively greater decline in function at 4 years with each decrement of Hgb level from 16 to 10g/dL, but not in men. Both women and men with mild anemia (11–12 g/dL and 10–11g/dL respectively) had greater decreases in cognition than those at higher levels (p=.01). Thus in an elderly community based population, anemia is significantly more common in AA independent of disease status, and is independently associated with increased mortality over eight years for both races and genders, and with declines in function and cognition most prominently in women. Further investigation of the etiology of these relationships is warranted.
Author notes
Corresponding author