Prevalence of anemia in elderly patients in primary care: impact on 5-year mortality risk and differences between men and women

Background: Increased mortality in patients with anemia has been demonstrated in disabled, seriously ill or hospitalized patients. In industrialized nations with their aging societies, however, elderly but apparently healthy family-physician patients are an important demographic group from a public-health perspective. We therefore set out to evaluate the prevalence of anemia in this group and associations between anemia and 5-year all-cause mortality, adjusted for multiple other established risk factors and chronic diseases. Methods: This was a monitored, prospective cohort study in Germany with 344 representative family physicians who documented, consecutively, elderly patients (aged >= 65 years). Extensive fasting plasma parameters were collected at baseline. Anemia at inclusion was defined according to World Health Organization criteria (hemoglobin below 12 g/dl in women and 13 g/dl in men). All participants were followed up for death of any cause for 5.3 years. Results: Among the 6880 individuals, 2905 men and 3975 women, aged 65-95 (mean age 72.5), mild anemia (hemoglobin levels >= 10 g/dl) was found in 6.1% of women and 8.1% of men. Among those patients, 36.1% of anemic men and 15.0% of anemic women died. In a Cox proportional hazards analysis, multiple adjusted for potential confounders including major comorbidities, a near doubling of the 5-year mortality risk in anemic men (hazard ratio [HR] 1.9; 95% confidence interval [CI] 1.5-2.4) was found, while in anemic women there was no risk increase at all (HR 1.1; 95% CI 0.8-1.6). Even if patients with the lowest hemoglobin concentration (511 g/dl for women, 512 g/dl for men) are singled out for multiple-adjusted analysis, anemia in men was related to a significant mortality risk (HR 3.3; 95% CI 2.1-5.1), but not in women (HR 1.85; 95% CI 0.97-3.53). Conclusion: In typical elderly patients without severe comorbidities, mild anemia was significantly associated with greater mortality in men but not in women. Given the impact of sex on outcomes of older subjects with mild anemia, the current definition of anemia should be adjusted for elderly males towards a higher hemoglobin threshold. Interventional trials will be needed to determine whether a consistent correction of anemia improves mortality of older men.