Abstract

Problem: Subclinical hypothyroidism, characterised by slightly elevated thyroid-stimulating hormone (TSH) with normal free thyroxine (fT4) levels, is prevalent in individuals aged over 50 years and is associated with adverse health outcomes. However, the impact of levothyroxine (LT4) therapy on cardiovascular and bone health in this population remains uncertain. We aimed to investigate the cardiovascular and bone health outcomes among individuals aged over 50 years with normal fT4 levels and slightly elevated TSH, comparing those prescribed LT4 versus those not prescribed.
Approach: We conducted a retrospective cohort study and an emulated target trial using healthcare records from The Health Improvement Network. Our study conducts an emulated trial to offer insights into an area where randomised controlled trials are impractical due to expense and design constraints. We included 59,579 individuals in the cohort study and 23,516 participants in the emulated target trial for cardiovascular outcomes. For bone health outcomes, our study included 62,710 individuals in the cohort study and 25,532 participants in the emulated target trial. Time-varying hazard ratios were estimated for cardiovascular and bone health outcomes among individuals prescribed LT4 compared to those not. Cardiovascular outcomes included angina, myocardial infarction, peripheral vascular disease, stroke, and stent procedure; bone health outcomes compromised of fragility fractures and osteoporosis.
Findings: In the cohort study with a 16-year follow-up, the time-varying hazard ratio for cardiovascular events among individuals prescribed LT4 was 0.78 (95% CI: 0.75-0.82, p < 0.001), suggesting a protective effect. However, for bone health outcomes, the hazard ratio was 1.07 (95% CI: 1.01-1.12, p = 0.013), indicating a slight increase in risk among LT4-prescribed individuals. In the emulated target trial with a 5-year follow-up, the hazard ratio for cardiovascular events was 1.28 (95% CI: 1.14-1.44, p < 0.001), indicating an elevated risk associated with LT4 prescription. Similarly, the hazard ratio for bone health outcomes was 2.10 (95% CI: 1.84-2.39, p < 0.001), reflecting a substantial increase in risk among individuals prescribed LT4 compared to those not prescribed.
Consequences: Our findings highlight varying effects of LT4 therapy on cardiovascular and bone health outcomes in individuals aged over 50 years with normal fT4 levels and slightly elevated TSH levels. The cohort study suggests a potential protective effect against cardiovascular events; the emulated target trial indicates increased cardiovascular risk and adverse bone health outcomes associated with LT4 prescription. This difference may be attributed to factors such as immortal time bias and misclassification bias in the cohort study or the shorter follow-up duration of the emulated target trial.