Association of distinct γ-glutamyltransferase trajectories with incident hyperglycemia using latent class growth mixture modeling: A longitudinal cohort study of Chinese adults

Aims To elucidate the association between distinct latent γ-glutamyltransferase (GGT) increasing trajectories in life time and incident hyperglycemia. Methods 4547 subjects were followed up for 3 years (January 2016–December 2019), and data regarding fasting plasma glucose, HbA1c, GGT, and other indices were recorded. Latent class growth mixed modeling (LCGMM) was used to analyze GGT latent trajectories. Results A three-class quadratic model was selected as the best fit by LCGMM. Subjects were categorized into three latent classes: high-increasing (n = 98, 2.16%), low-increasing (n = 364, 8.01%), and stable (n = 4085, 89.83%) classes. Adjusted hazard ratios of hyperglycemia for the high-increasing and low-increasing classes were 1.341 (1.076–2.051) and 1.264 (1.048–1.525) when compared with stable class, respectively. Odds Ratios (ORs) of GGT slopes were confirmed in the 28–57-year age group, shaped like an “M”, ranging from 1.144 (1.059, 1.237) to 2.502 (1.384–3.862). Significant differences in the associations between model-estimated GGT values and hyperglycemia incidence were observed from 28 to 51 years of age, with ORs ranging from 1.011 (1.011, 1.012) to 1.014 (1.012, 1.019). Conclusions Our study demonstrated that subjects in GGT increasing classes exhibited higher risks of developing hyperglycemia. A steeper GGT slope is a more effective predictor of hyperglycemia than the GGT value.