Endocrine Abstracts

Thyrotoxicosis is a common disorder which induces many cardiovascular effects and is associated with increased cardiovascular mortality. We recruited 283 consecutive unselected thyrotoxic subjects attending our thyroid clinic and 283 age and sex-matched euthyroid controls. All completed a structured cardiovascular history and examination, resting 12-lead ECG and 24-hour Holter recording. Follow up data on mortality was collected for a year. The median age of the cohorts was 49.5years (214 females and 69 males in each group). The cohorts were matched with respect to past history and family history of vascular disease. The median serum free thyroxine (T4) concentration at presentation was 37.8picamoles/Litre (IQR: 26-51). Fifteen(5%) of the thyrotoxic cohort were in atrial fibrillation(AF) and 22(8%) had sinus tachycardia on ECG(p<0.005 vs controls for both). Cardiac rhythm abnormality on ECG (one or more atrial or ventricular ectopics or AF) was found in 31 thyroid subjects and 9 controls(p<0.0001) and rhythm abnormality on Holter was noted in 68 thyroid subjects vs. 50 controls(p<0.05); persistent or paroxysmal AF was more prevalent in thyroid subjects than controls (11 vs 1, p<0.01), as were atrial and ventricular salvoes ( >3 consecutive premature atrial or ventricular beats: 50 vs 31, p<0.02 and 8 vs 1, p<0.05 respectively). Holter recording was more sensitive than resting ECG at detecting these rhythm abnormalities(p<0.0001). Age and systolic/diastolic BP were independantly associated with presence of AF on resting ECG(p<0.005); cardiac rhythm abnormality on Holter was associated with age(p<0.0001) and presentation T4>30picamoles/Litre (p<0.02). AF and atrial salvoes on Holter were associated with age(p<0.0001); ventricular salvoes were associated with a past history of vascular disease(p<0.005). During the follow-up there were 5 deaths in the thyroid cohort and none in the control group. Death was independently associated with presence of AF on ECG(p<0.0001). The cause of this mortality is unknown. We hypothesise that overt or subclinical dysrhythmias contribute to this increased risk.