The Impact of Hemoglobin on Total Mortality in Elderly Asian-American Men: The Honolulu Heart Program.

Anemia in the elderly is frequently unrecognized and untreated. Anemia prevalence increases with age and has been reported to occur in >20% of elderly persons. Anemia has also been shown to be a significant and independent predictor of mortality. There are no published data, however, regarding prevalence of anemia or its association with mortality in an elderly Asian-American population. In this study, we examined the relationship between hemoglobin (Hb) and mortality in elderly Asian-American men in the Honolulu Heart Program (HHP). The HHP began in 1965 following a cohort of 8006 Asian-American men aged 45-68 living on Oahu, Hawaii. These men have been followed for over 40 years(y). The present analysis is based on 14 years of follow-up for mortality events beginning at examinations that were given from 1991–1993 when the men were aged 71-93y. Anemia (Hb<13g/dL) and borderline anemia (Hb:13-13.9 g/dL) were defined using WHO criteria. Age-adjusted mortality rates in person-years(y) of follow-up were calculated across ranges of Hb using logistic regression and standard analysis of covariance methods. Statistical tests for an association between Hb and mortality were based on proportional hazards regression models with adjustments for age, total cholesterol, alcohol intake, HDL cholesterol, fibrinogen, BMI, forced expiratory volume, physical activity, diabetes, hypertension, cigarette smoking, and a history of cancer and CVD. Linear and nonlinear relationships were examined along with efforts to identify ranges of Hb that were associated with the lowest incidence of mortality. In the latter instance, differential calculus yielded a combination of regression coefficients that resulted in a single Hb value that was associated with the lowest mortality. Bootstrap sampling further provided a corresponding 95% CI. Of the total subjects, 9.1% had anemia with 14.7% having borderline anemia. Anemia prevalence increased from 4.3% in men 71-75y to 27.2% in those ≥86y (p<0.001). In 14 years of follow-up, there were 2,311 deaths (6.7/100 person-y). The mortality rate in those with anemia was 11.8/100 person-y vs 6.4/100 person-y in those without anemia (p<0.001). The association persisted for men with and without cancer and cardiovascular disease, across age strata, for deaths that occurred <1y, 1-5y and >5y and after adjustment for other risk factors. There was a significant (p<0.001) inversed J-shaped relationship with a minimum in mortality incidence occurring when Hb was 15.5g/dL. The corresponding 95% CI (15–15.9g/dL) was used as a reference for comparing rates of mortality with other Hb ranges. Compared to this latter group, there was a 2.4-fold excess in mortality (95% CI, 1.8–3.1) in men with a Hb<12g/dL (p<0.001). For men with Hb ≥17g/dL, there was a 1.3–fold excess (95% CI, 1.2–1.6). The latter, however, was not significant after excluding former and current smokers. These results are the first to describe prevalence of anemia in elderly Asian-American men. These findings indicate anemia is a strong independent predictor of mortality in this elderly population with the risk similar to that reported from other studies of the elderly.