Associated Factors and Dose-Response of Steroid-Induced Diabetes Mellitus in Dermatology Inpatients: A Retrospective...

To investigate the prevalence and associated factors of steroid-induced diabetes mellitus (SDM) in dermatology inpatients receiving systemic glucocorticoid (GC) therapy, with particular focus on characterizing the dose-response relationship.

Retrospective observational cohort study.

Department of Dermatology, The Affiliated Hospital of Xuzhou Medical University, Xuzhou, China, from August 2019 to October 2024.

A total of 293 patients treated with GCs were included, with 42 classified as having SDM and 251 as controls based on blood glucose monitoring. Data such as demographics, GC dosage, laboratory indices, and follow-up outcomes were analyzed by SPSS 26.0, R 4.4.1, and logistic regression analysis was applied to explore factors associated with the development of SDM. We modeled the nonlinear dose-response between average daily GC dose (methylprednisolone-equivalent) and SDM using multivariable logistic regression with restricted cubic splines, setting 0.6 mg/kg/day as the reference. All tests were two-sided, p < 0.05 indicates that the difference is statistically significant.

Among 293 patients, 42 (14.3%) developed SDM and 17 (5.8%) had impaired glucose regulation. SDM cases were older and had higher BMI, hypertension, family history of diabetes, higher average daily GC dose, and more frequent immunosuppressant use. Multivariate regression identified family history of diabetes (OR: 11.16), immunosuppressant use (OR: 3.48), average daily GC dose (OR: 29.13), and serum uric acid (OR: 1.007) as independent factors associated with SDM. The dose-response between average daily GC dose and SDM was significant (Wald χ ² = 18.48, df = 3, p = 0.0004); risk rose steeply beyond 1.0 mg/kg/day (e.g., 1.2 mg/kg/day: OR: 3.94, 95% CI: 1.44-10.66).

SDM is prevalent among dermatology inpatients receiving GCs. High-risk patients may benefit from closer glucose monitoring and individualized dose adjustment; prospective studies are warranted to confirm these associations.