Predictors for development of hyperuricemia: an 8-year longitudinal study in middle-aged Japanese men.

To identify the factors responsible for increases in serum uric acid (SUA), a cohort of 1,312 hyperuricemia-free (SUA < 7.5 mg/dL and no medication for hyperuricemia or hypertension) male office workers aged 30 to 52 years were examined annually for 8 successive years. Subjects who were found to have become hyperuricemic (SUA > or = 7.5 mg/dL) or who started medication for hyperuricemia during repeat surveys were defined as incidence cases. The SUA trend was also examined in 1,062 subjects for whom 9 consecutive SUA values were available and who did not start medication for hyperuricemia or hypertension during the observation period. Multivariate analyses, excluding the baseline SUA level as a factor in the Cox proportional-hazards model, indicated that age (negative), body mass index (BMI), log triglyceride level, hemoglobin A(1c) (HbA(1c)) level (negative), white blood cell count, and alcohol intake at study entry were significantly associated with the incidence of hyperuricemia. In the model including the baseline SUA level, baseline SUA level was the strongest factor for the incidence of hyperuricemia, and BMI, white blood cell count, and alcohol intake at study entry remained as independent factors. From stepwise linear regression analyses for SUA slope, excluding the baseline SUA level as a factor, significant correlates with SUA slope were, in order of their relative importance, slopes of BMI, HbA(1c) (negative), blood urea nitrogen, log triglyceride level, total protein, and baseline levels of hematocrit (negative), white blood cells, and HbA(1c) (negative). In stepwise linear regression analyses, including the baseline SUA level as a factor, SUA level (negative) and alcohol intake at study entry emerged as significant factors for SUA slope. The cumulative percentage of variation for SUA slope was 25.6%. In conclusion, obesity, alcohol intake, and multimetabolic disorders were determined to be independent predictors for the development of hyperuricemia. In addition, the white blood cell level may be a contributory factor. Copyright 2001 by W.B. Saunders Company