Endocrine Abstracts

The TEARS database links six datasets for patients with treated and stabilised thyroid conditions. It has been validated previously and represents a population-based patient sample, as patient identification includes using regional-wide biochemical records and prescription data. For patients on long-term thyroxine we examined the outcome of ‘serious vascular events’ defined as vascular death, non-fatal myocardial infarction and non-fatal stroke as used in the Anti -thrombotic Trialists’ Collaboration.

Hazards ratios (HR) for cardiovascular events were calculated for each of the following variables, corrected for all the other co-variates. Previous thyroid condition and Carstairs deprivation score did not predict outcomes.

Overall 9493 patients (8235 female), of mean age 59.3 years were followed up for 3.5 years (IQR 1.5–6.7 years). Increasing age (HR 1.10 (1.09–1.11)), co-existing diabetes (HR 1.80 (1.35–2.39)), previous cardiovascular disease (1.72 (1.44–2.04)), thyroxine dose of <100 μg (1.25 (1.02–1.53)), or >150 μg (HR 1.50 (1.07–2.09)) were all associated with increased cardiovascular events, whilst female sex had lower rates (HR 0.71 (0.59–0.85)). A low TSH (<0.1 mU/l) was not associated with increased cardiovascular events (HR 1.02 (0.79–1.31)), whilst a raised TSH >4 mU/l was (HR 2.42 (1.79–3.26) when compared to patients with a normal TSH.

Having a low TSH whilst on long-term thyroxine was not associated with increased risk of serious vascular events. Having a raised TSH was associated with vascular events, and this may relate to under-treatment or intermittent concordance with thyroxine and other medication. Low doses and high doses of thyroxine were associated with increased vascular events, independent of serum TSH concentration.