MON-597 Association of Serum TSH with Atrial Fibrillation in the Elderly

INTRODUCTION: Hypothyroidism and atrial fibrillation (Afib) are both common conditions in elderly patients. Levothyroxine, the most common therapy for hypothyroidism, has a narrow therapeutic window and requires monitoring to ensure that thyroid-stimulating hormone (TSH) remains within the reference range. Subclinical hyperthyroidism whether endogenous or induced by over-treatment with levothyroxine, has been associated with higher rates of Afib. Therefore, the American Thyroid Association has proposed that thyroid hormone replacement be adjusted to maintain a higher TSH in the elderly, with a proposed goal of 4-6 mcIU/mL; however, the evidence behind this target is limited. The purpose of our study was to determine whether TSH outside this target, but within the reference range for young adults, is associated with higher rates of AFib in elderly patients treated with levothyroxine. METHODS: This was a retrospective cohort study of patients aged 65 and older who were treated with levothyroxine therapy. Exclusion criteria included: pre-existing diagnosis of AFib; use of amiodarone in the prior year; recent diagnosis of thyroid cancer; history of pituitary disorders. Patients were followed from their first visit in 2010 through the end of 2017 and stratified into four groups based on median TSH value obtained in 2010. This grouping was maintained throughout the course of the study regardless of changes in TSH values in subsequent years. Primary outcome was incidence of AFib. Descriptive and comparative statistics were performed with continuous variables being assessed with one-way ANOVA or Kruskal-Wallis tests, as appropriate. Categorical variables were assessed with the chi-squared test. Rates of AFib between groups were assessed via Poisson regression with control of underlying confounders. RESULTS: Patients (n=11274) were divided into the following groups: TSH <0.5 (n=989), TSH 0.5-2 (n=3827), TSH 2.0-6.0 (n=5574), and TSH >6.0 (n=883). Age and sex distribution between TSH quartiles was fairly matched, although a trend was seen toward older age and male sex with higher TSH values. Incidence of AFib in the entire cohort was 17%. No association was found between TSH quartile and incidence or severity of AFib. Increased risk of AFib was seen with increasing age and with male sex. DISCUSSION: When compared to patients with subclinical hyperthyroidism, no significant difference was found in incidence of AFib in any of the TSH quartiles. Interestingly, patients with subclinical hyperthyroidism did not have a higher rate of AFib. A limitation of the study is that only TSH during one calendar year was used to form cohorts, which does not allow correlation with TSH changes over time. However, it is reassuring that at least in the short term, a TSH outside recommended ranges does not convey an increased risk of AFib in the elderly.