OBJECTIVE: To investigate (1) whether the association of thyroid hormone with frailty risk is linear or nonlinear and (2) which range of thyroid hormones or thyroid stimulating hormone (TSH) is more associated with a higher risk of frailty in older adults. DESIGN: Systematic review and dose-response meta-analysis. METHODS: Medical electronic databases were searched for cross-sectional or longitudinal studies, published from database inception to February 2022. We focused on the relationship between TSH and frailty. Data on TSH reference range, TSH exposure categories, sample size of each exposure category, and adjusted odds ratios (ORs) for frailty with 95% confidence interval (CI) were extracted. In the dose-response meta-analysis, we set the OR for frailty as 1 at 0.3 mIU/L TSH. RESULTS: The systematic review included 10 studies, whereas the meta-analysis included 3 studies (n = 6388). TSH levels ranged from 0.3 to 4.8 mIU/L, and the dose-response meta-analysis revealed a significant J-shaped association (p = 0.0071). Frailty OR (95% CI) increased from 1.30 (1.06-1.59) for 2.7 mIU/L TSH to 2.06 (1.18-3.57) for 4.8 mIU/L TSH. CONCLUSIONS: A significant nonlinear, J-shaped association was noted between TSH level and frailty. TSH levels within the upper half (2.7-4.8mIU/L) of reference range was noted to significantly higher risk of frailty
by contrast, those in the lower half (0.6-1.5 mIU/L) had a lower risk of frailty, though not significantly so. TRAIL REGISTRATION: This systematic review was registered with PROSPERO (registration number: CRD42022299214).