Abstract
OBJECTIVE: We aimed to clarify the relationship between serum alanine transaminase (ALT) levels and incidence of new-onset diabetes in a Japanese general population. SETTING: Population-based retrospective cohort study using annual health check-up data for residents of Iki City, Nagasaki Prefecture, Japan. PARTICIPANTS: A total of 5330 Japanese individuals (≥30 years old) without diabetes at baseline were analysed. PRIMARY AND SECONDARY OUTCOME MEASURES: Serum ALT levels were determined using an enzymatic method and were classified into gender-specific quartile groups as follows: group 1 (3-16 U/L in men and 3-13 U/L in women), group 2 (17-21 U/L in men and 14-16 U/L in women), group 3 (22-29 U/L in men and 17-22 U/L in women) and group 4 (30-428 U/L in men and 23-268 U/L in women). The study outcome was the incidence of diabetes (fasting glucose ≥7.0 mmol/L, non-fasting glucose ≥11.1 mmol/L, glycated haemoglobin ≥6.5% or use of glucose-lowering therapies). RESULTS: After an average follow-up period of 5.0 years, 279 individuals developed diabetes. The incidence rate of diabetes increased with elevation of serum ALT levels (0.7% per 100 person-years in group 1, 0.9% in group 2, 0.9% in group 3 and 1.7% in group 4) (p<0.001 for trend). This association was significant after adjustment for other risk factors including age, sex, obesity, hypertension, dyslipidaemia, smoking, current daily alcohol intake and regular exercise (p<0.001 for trend). Comparable associations were observed between men and women (p=0.459 for interaction). CONCLUSION: Serum ALT levels were associated with future development of diabetes in the general Japanese population.