Development and external validation of a risk prediction model for hypertension among individuals with established hypothyroidism

Background Hypothyroidism represents a multisystem endocrine disorder involving cardiovascular, renal, and metabolic alterations beyond thyroid hormone deficiency. However, validated tools for hypertension (HTN) risk stratification specifically designed for patients with hypothyroidism remain lacking. This study aimed to develop and externally validate a hypothyroidism-specific HTN prediction model incorporating routinely available clinical variables. Methods This retrospective multicohort study included patients with hypothyroidism from a hospital-based cohort and an independent external population-based cohort. Eligible participants were identified according to predefined diagnostic criteria, and individuals with missing key clinical variables or incomplete outcome information were excluded. Candidate predictors were selected using regression-based approaches, and a multivariable prediction model was developed in the derivation cohort and subsequently evaluated in internal and external validation cohorts. An interpretable HTN risk profile was developed and evaluated using internal validation, external validation, and subgroup analyses. Results Age, serum creatinine (SCr), anion gap (AG), and diabetes mellitus (DM) constituted a clinically interpretable risk profile reflecting vascular aging, renal vulnerability, and metabolic dysregulation. The model demonstrated robust discrimination, good calibration, and favorable clinical utility across training, testing, and external validation cohorts. Conclusions A disease specific HTN risk profile was developed and externally validated in patients with hypothyroidism. This model provides a practical approach for individualized cardiovascular risk stratification and supports the application of precision endocrine medicine.