A retrospective cohort of 115 thyroid-storm patients at a Thai tertiary center had 13% 30-day mortality. An eight-point score using male sex, hypertension, SOFA at least 2, and GCS at most 10 achieved apparent AUC 0.81 and bootstrap-corrected AUC 0.79.
Key findings
- Male sex (OR 4.34), hypertension (OR 2.54), SOFA at least 2 (OR 9.93), and GCS at most 10 (OR 1.95) entered the score. Roughly 15 deaths among 115 patients provide limited events per parameter, although bootstrap AUC 0.79 (95% CI 0.69-0.91) is promising.
Why this matters globally
Thyroid storm is uncommon but lethal, and diagnostic scores were not designed for prognosis. If multicenter validation succeeds, a simple score could support monitoring and transfer decisions in resource-limited settings, but calibration and clinical impact matter.
Thai researcher contribution
Chiang Mai University endocrinology and clinical teams used Thai patient data to address a prognostic gap in a rare endocrine emergency.
Limitations to consider
The single-center retrospective design, few deaths, and four predictors permit overfitting despite bootstrapping. Sex and hypertension may reflect case mix; dichotomized SOFA/GCS lose information. Calibration, missingness, treatment effects, and external validation are not reported.