Background: The importance of assessing liver function in patients with type 2 diabetes (T2DM) has been well established, but few studies have examined this from a time-series view. We investigated the effect of seasonal fluctuations in liver function on long-term glycemic control on T2DM.
Method: A total of 9,949 individuals (20-80 yo) were enrolled in a large registry study of T2DM in Japan from Jan 2014 (baseline) to Dec 2020. After excluding missing data on covariates, 9,438 were included in the analysis. Seasonal fluctuations of AST/ALT was defined as the amplitude of the seasonal component of each individual extracted by STL decomposition after multiple imputations. HbA1c at the final visit of each individual (final-visit A1c) was used as the index of glycemic control. Multivariate regression analysis was performed using standardized seasonal fluctuations in AST/ALT as explanatory variables and final-visit HbA1c as the objective variable. Covariates included sex, follow-up time, and age at baseline, BMI, HbA1c, log-transformed AST or ALT, drinking and smoking habits, diabetes duration, and insulin use. A sensitivity analysis was performed by adding seasonal BMI fluctuation during follow-up to the above model.
Result: During the follow-up of 6.0 ± 1.8 yrs, AST/ALT tended to be lower in summer and higher in winter. Multivariate analysis showed that the seasonal fluctuations of AST/ALT were significantly associated with final-visit A1c, with an estimated effect of +0.044% [95% CI: 0.017, 0.071] and +0.028% [0.002, 0.054] per 1SD increase, respectively. The results of the sensitivity analysis remained consistent.
Conclusion: This study indicated that seasonal fluctuations in liver enzymes might be a risk factor for poor glycemic control independent of traditional risk factors, as well as BMI fluctuation in Japanese patients with T2DM. Our results suggest that year-round assessment of liver enzymes may be important in clinical practice.
R. Toki: None. M. Sakamoto: None. M. Iida: None. M. Yuki: None. T. Takebayashi: None.