Complex association of serum alanine aminotransferase with the risk of future cardiovascular disease in type 2 diabetes

Background and aims We aimed to determine the prospective association between baseline serum levels of alanine aminotransferase (ALT) and the incident cardiovascular disease (CVD) in people with type 2 diabetes. Methods In an open cohort setting, people with type 2 diabetes were followed for their first ever CVD presentation from 1995 to 2015. Statistical methods included Cox regression analysis for reporting of hazard ratios (HRs), artificial neural network modelings, and risk reclassification analyses. Results We found a nearly constant CVD hazard with baseline serum ALT levels below the 30 IU/L mark, whereas baseline serum ALT levels ≥ 30 IU/L remained an independent predictor of lower CVD rates in patients with type 2 diabetes in the final multivariate Cox proportional hazards regression model (HR: 0.204, 95%CI [0.060–0.689], p _{for trend} value = 0.006). Age, male gender and fasting plasma insulin levels independently predicted baseline serum ALT ≥ 30 IU/L among the population cohort. Augmentation of serum ALT into the weighted Framingham risk score resulted in a considerable net reclassification improvement (NRI) of coronary heart disease (CHD) risk prediction in the study population (NRI = 9.05% (8.01%–10.22%), p value < 0.05). Conclusions Serum ALT could successfully reclassify about 9% of the population with type 2 diabetes across the CHD-affected and CHD-free categories. Overall, our findings demonstrate a complex and nonlinear relationship for the risk of future CVD by baseline serum ALT levels in patients with type 2 diabetes. Further studies are warranted to confirm whether this complex association could be translated into a clearly visible U or J-shaped figure.