Survival advantage in Asian American end-stage renal disease patients

Abstract

Survival advantage in Asian American end-stage renal disease patients.

Background: An earlier study documented a lower mortality risk for end-stage renal disease (ESRD) patients in Japan compared with the United States. We compared the mortality of Caucasian (white) and Asian American dialysis patients in the United States to evaluate whether Asian ancestry was associated with lower mortality in the United States.

Methods: The study sample from the U.S. Renal Data System census of ESRD patients treated in the United States included 84,192 white or Asian patients starting dialysis during May 1995 to April 1997, of whom 18,435 died by April 30, 1997. Patient characteristics were described by race. Relative mortality risks (RRs) for Asian Americans relative to whites were analyzed by Cox proportional hazards regression models adjusting for characteristics and comorbidities. Population death rates were derived from vital statistics for the United States and Japan by age and sex.

Results: Adjusting for demographics, diabetes, comorbidities, and nutritional factors, the RR for Asian Americans was 0.75 (P = 0.0001). Race-specific background population death rates accounted for over half of the race-related mortality difference. For whites, mortality decreased as the body mass index (BMI) increased. For Asians, the relationship between BMI and survival was u-shaped. The ratio of Asian American/white dialysis death rates and the ratio of Asian American/white general population death rates both varied by age in a similar pattern. The population death rates of Asian American and Japanese were also similar.

Conclusion: Among dialysis patients, Asian Americans had a markedly lower adjusted RR than whites. The effect of BMI on survival differed by race. Compared with the respective general population, dialysis patients had the same relative increase in death rates for both races. The difference in death rates between the United States and Japan does not appear to be primarily treatment related, but rather is related to background death rates.