A Longitudinal, Real-World Study of the Relationship between Levothyroxine Dose and Suboptimal Hypothyroidism Treatment

Background: Suboptimal treatment of hypothyroidism with levothyroxine (LT4) is common and associated with adverse cardiovascular outcomes. Patients treated with LT4 doses >100 µg may be at increased risk of under- and overtreatment and therefore may benefit from more frequent thyroid stimulating hormone (TSH) surveillance. Our study aim was to determine the association between LT4 dose and the occurrence of suboptimal treatment within 6 months of a normal TSH value. Methods: A total of 4449 LT4-treated adults were included in a 6-month longitudinal study. All included patients were required to have at least 2 TSH collections over the study period at least 30 days apart. Patients were grouped according to LT4 dose (≤50, 51–100, 101–150, and >150 µg), which served as the primary exposure. Outcomes included the occurrence of TSH values outside an age-adjusted normal range and a wide range (0.1–10.0 mIU/L). Model covariates included sociodemographics, index TSH value, and a range of codiagnoses and comedications that were prevalent in the population and/or known to impact LT4 treatment. Longitudinal data were analyzed using both repeated measures (generalized estimating equations) and time-to-event (Cox proportional hazards models) methodologies. Kaplan–Meier curves were plotted to examine the incidence of out-of-range TSH values stratified by LT4 dose class. Results: The incidence of at least one out-of-range TSH value approached 25% across the entire population and exceeded 45% in the LT4 > 150 µg group. In the repeated measures analysis, a dose-dependent relationship between LT4 dose and the occurrence of TSH outside the age-adjusted range (OR: 1.98, 3.39, and 5.65 for LT4 doses of 51–100, 101–150, and >150 µg, respectively, p -values <0.001). Results were similar when the outcome was defined as a TSH outside 0.1–10.0 mIU/L. In the time-to-event analysis, the hazard ratios were 1.58, 2.73, and 3.08, respectively ( p- values <0.001). Conclusions: We identified a dose-dependent relationship between LT4 dose and suboptimal treatment over a 6-month study period. It may be beneficial for medically complex patients taking high doses of LT4 to increase TSH surveillance, in particular for those with a history of suboptimal treatment.