OBJECTIVES: Immunosuppressive treatments used in liver transplantation(LT) often include corticosteroids. However, due to the undesirable side effects associated with steroids, many centers choose to withdraw them from the regimen. This study aims to identify the effect of steroid withdrawal on the development of diabetes mellitus(DM) following LT to provide evidence on the appropriate duration of steroid treatment.

METHODS: We conducted a retrospective study utilizing a nationwide health insurance claims database covering LT recipients from January 2009 to March 2021 in South Korea. Adult patients without a diabetes diagnosis within a year before transplant were included. Steroid effects were assessed by categorizing treatment status at 1, 3, and 6 months post-transplant using a landmarking approach. Treatment cohorts were created using propensity score matching and Cox proportional hazards models were used to estimate DM incidence and hazard ratios at each landmark time.

RESULTS: A total of 6,295 recipients were included (70.7% male, 76.4% living-donor LT, 94.9% on a Tacrolimus-based regimen), among whom 4,958 withdrew steroids within 6 months. DM occurred in 2,930 patients during the study period. The incidence rate of the withdrawal group was notably lower compared to the maintenance group at the 1-month landmark, with no significant difference at 6-month. Steroid withdrawal within 3 months showed a reduced risk of DM compared to steroid maintenance during that period (1 month HR=0.586, 95% CI=0.407-0.846; 3 months HR=0.766, 95% CI=0.611-0.960; 6 months HR=0.844, 95% CI=0.619-1.152)

CONCLUSIONS: When steroids were withdrawn within 3 months after LT, it had an impact on reducing the occurrence of DM. However, among patients who maintained steroid use beyond that period, withdrawing steroids showed less influence on the development of DM. Long-term prescription of corticosteroids should be approached cautiously considering the risk of DM in patients.