Body weight change trajectories following the treatment of hyperthyroidism: a prospective cohort study

Hyperthyroidism is a well-recognized cause of significant metabolic changes with most patients experiencing weight loss in the acute illness. Weight recovery occurs following treatment of hyperthyroidism with some studies reporting excessive weight regain.1, 2 An understanding of the weight trajectories in these patients may explain the reduced quality of life,3 and increased cardiovascular risk profile. Measure and model weight change across time following the treatment of hyperthyroidism. Model the effect of self-reported disease-related pre-treatment weight loss on the treatment-related weight change. A secondary outcome was to examine the longitudinal change in body mass index (BMI) categories. We conducted a prospective cohort study of consecutive patients with hyperthyroidism, as per the inclusion and exclusion criteria (Supporting Information: Table 1), at three secondary care outpatient settings in Cyprus. We estimated a sample size of 87 for a power of 80% at an α of .05. Longitudinal weight change was examined by fitting crude and adjusted two-level linear mixed-effects models. The outcome variable was the change in weight (nested within participants). Level 1 or the fixed part of the model included the natural logarithm of time, ln(time) and level 2 or the random part of the model was the individual participants. The adjusted model included the covariates sex, age, weight and BMI (at presentation), disease-related weight loss, longitudinally measured free thyroxine (fT4) concentration and the presence of iatrogenic hypothyroidism, diagnosis of Graves' disease and the interactions of weight loss with time and fT4 with hypothyroidism. For the secondary analysis, an adjusted two-level mixed-effects ordered logistic regression model was fitted. Analysis was performed with STATA 16.1 (StataCorp). We prospectively studied 103 participants (median age 49 years) with median duration of follow-up of 12 months (87.4% of patients had at least 6 months' follow-up) and a total of 344 sets of observations. Anti-thyroid drugs (ATDs) were the sole treatment in 89.3% of patients, and ATDs followed by surgery in 10.7%. No patients opted for radioiodine therapy despite ready availability due to cultural and logistical reasons. The median (interquartile range [IQR]) presentation weight and BMI were 66.2 (58.0, 72.7) kg and 23.9 (22.2, 27.4) kg/m2, respectively. Disease-related pre-treatment weight loss was documented at presentation in 101 patients; their median (IQR) weight loss was 4.5 (0.0, 8.3) kg and BMI loss 1.7 (0.0, 3.0) kg/m2. Overall, 73% of patients lost weight over 8 (4, 20) weeks. The weight and BMI followed an incremental trend across the follow-up time points. The median (IQR) gain in weight and BMI following treatment was 6.4 (3.6, 10.5) kg and 2.2 (1.3, 4.0) kg/m2, respectively. An increase in the prevalence of excess weight (BMI ≥ 25.0 kg/m2) was observed from 40.8% at presentation to 56.9% at 18 months and 69.9% at the final measurement. The mean number of repeated weight change measurements within each participant was 3.3 (range, 1.0–8.0) in both crude and adjusted mixed-effects models (p < .001 for both). The time variable [ln(time)] was a significant predictor of weight change in both the unadjusted (b = 2.96, p < .001) and adjusted (b = 2.62, p < .001) models. The adjusted model predicted 1.82 kg of weight gain for every doubling in time in months (Figure 1A). The weight increase was significant at all time points beyond the first month (all p < .001) and ranged from 3.28 kg at 3 months to 10.56 kg at 48 months. The disease-related pre-treatment weight change covariate was not a significant predictor of post-treatment weight change (b = 0.020, p = .825), whereas its interaction with time was (b = 0.16, p < .001) (Figure 1B). The other variables in the model which were statistically significant were (longitudinally measured) fT4 (b = −0.087, p < .005) and its interaction with hypothyroidism (b = 0.13, p = .048). In a post-hoc analysis, fT4 at presentation did not alter the model significantly. Participants tended to move up BMI categories with time (b = 3.08, p = .001) following the treatment of hyperthyroidism. In a post-hoc analysis, at 6 months onwards their probability of being allocated to the ‘overweight' category (BMI ≥ 25 kg/m2) was significantly higher. In this prospective observational study of 103 patients with hyperthyroidism followed-up for a median 12 months, the mean body weight change from presentation to the last follow-up was 7 kg. The predicted weight increase was 3 kg at 3 months, 7 kg at 12 months and 11 kg at 4 years. Some studies have previously shown a positive correlation between disease-related weight loss and treatment-related weight gain,2, 4 but this is the first study to demonstrate the modifying effect of disease-related pre-treatment weight loss on the treatment-related weight changes with time. A pronounced reversal of the normoweight and overweight/obesity balance occurred within 3 months and persisted thereafter; the prevalence of overweight and obesity at the final follow-up was about 70%. This is much higher than the 46% prevalence observed in the general population in Cyprus.5 Our results are consistent with an observational controlled study in patients with hyperthyroidism treated with ATDs and radioiodine with an average follow-up of 23 months.2 Using multilevel mixed-effects models, we show a significant predicted weight gain with the treatment of hyperthyroidism. We observed a high prevalence of overweight and obesity at approximately 50% higher than expected for the background population. The disease-related weight loss was a significant modifier of the effect of time on the treatment-related weight gain. Limitations of our study include the relatively short follow-up and reliance on self-reported disease-induced pre-treatment weight loss, which is subject to recall bias. We conclude that a notable proportion of patients experience excessive and progressive weight gain following treatment of hyperthyroidism. Potential mechanisms include alterations in the central regulation of appetite, hunger and satiety, iatrogenic hypothyroidism, psychological influences and predisposing genetic and/or environmental factors.1 Future interventional studies should explore which management strategies may reduce weight gain. Angelos Kyriacou conceived the study and wrote the first draft. Angelos Kyriacou, Akheel A. Syed and Alexis Kyriacou designed the study. Angelos Kyriacou and Michalis Picolos recruited the patients. Alexis Kyriacou and Angelos Kyriacou performed the statistical analysis. All authors revised and approved the final manuscript. Angelos Kyriacou would like to thank the Society for Endocrinology (UK) for giving him the opportunity to present in the Research Incubator Meeting of the British Endocrine Societies conferences, in both 2017 and 2018, on the topic of ‘Hyperthyroidism and weight changes'. The feedback received during those sessions has helped shape the design of this study. This research did not receive any specific grant from funding agencies in the public, commercial or not-for-profit sector. The authors declare no conflict of interest. The study received ethics approval from the Cyprus National Bioethics Committee (ΕΕΒΚ/ΕΠ/2020.01.270). Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.