Association of longitudinal serum 25-hydroxyvitamin D trajectories with subsequent liver stiffness increase in patients with early-onset type 2 diabetes and metabolic dysfunction-associated steatotic liver disease

Background A single serum 25-hydroxyvitamin D [25(OH)D] measurement may not adequately reflect long-term vitamin D exposure. We evaluated the associations of longitudinal 25(OH)D trajectories with subsequent liver stiffness measurement (LSM) increases in patients with early-onset type 2 diabetes and metabolic dysfunction-associated steatotic liver disease. Methods This retrospective study included 286 patients managed between 2022 and 2025. Latent class mixed-effects modeling identified 25(OH)D trajectories before a 1-year landmark. The primary outcome was a follow-up LSM value that was both at least 20% (relative) and at least 2.0 kPa (absolute) higher than the landmark value; the conservative sensitivity outcome required increases of at least 30% and at least 3.0 kPa, respectively. Logistic regression models were fitted across 50 pseudo-class draws and pooled using Rubin’s rules. Follow-up LSM was analyzed by analysis of covariance adjusted for landmark LSM and prespecified covariates. A segmented linear mixed-effects model compared LSM slopes before and after the landmark. Results Three trajectories were identified: persistently low ( n = 95), increasing ( n = 84), and stable high ( n = 107). The primary-outcome rates were 54.7%, 33.3%, and 18.7%, respectively. Compared with the stable-high trajectory, the odds ratios in the prespecified primary model were 3.34 (95% confidence interval [CI], 1.68–6.61) for the persistently low trajectory and 1.98 (95% CI, 1.03–3.82) for the increasing trajectory. After additional adjustment for vitamin D supplementation, cumulative dose, metabolic factors, and medications, the corresponding odds ratios were 2.87 (95% CI, 1.38–5.96) and 1.84 (95% CI, 0.91–3.72). Under the conservative outcome definition, the corresponding odds ratios were 3.18 (95% CI, 1.39–7.27) and 1.71 (95% CI, 0.76–3.84). Adjusted follow-up LSM was higher by 1.07 kPa (95% CI, 0.66–1.48) and 0.52 kPa (95% CI, 0.10–0.94), respectively. The post-landmark LSM slope exceeded the pre-landmark slope by 0.98 kPa/year (95% CI, 0.80–1.16). Conclusions The persistently low 25(OH)D trajectory showed the most consistent association with subsequent LSM increase. The increasing trajectory showed weaker, outcome-definition-dependent associations and coincided with more frequent vitamin D supplementation and a greater reduction in body mass index. These findings should be interpreted as longitudinal associations rather than causal effects or evidence of validated individual-level predictive performance.