Context. Amiodarone-induced thyrotoxicosis (AIT) is a recognized complication of amiodarone treatment with limited management options. Its predisposing factors are incompletely defined yet a higher prevalence was reported in adults with congenital heart disease (CHD).
Objectives. We sought to determine the incidence and risk factors for AIT in adults with CHD, to construct a risk predictor index model and to describe the response to therapy in the identified AIT cases.
Design. Historical cohort study for the period 1987-2009. Patients were followed until diagnosis of AIT or last thyroid assessment on amiodarone. Identified risk factors were included in Cox proportional hazard model for construction of risk predictive index. AIT cases were analyzed for response to therapy in a descriptive manner.
Setting. Tertiary care center.
Patients. Adult CHD Clinic patients treated with amiodarone for ≥ 3 months. Patients without follow-up thyroid tests or with history of hyperthyroidism or thyroidectomy were excluded. 169 patients met criteria from 219 treated with amiodarone.
Results. AIT developed in 23/169 patients or 13.6%. The AIT incidence peaked in the 3rd year at 7.7%. AIT patients had a lower body mass index (BMI) at AMIO initiation compared with rest of the cohort (mean ± standard deviation: 21.9 ± 2.9 vs. 25.1 ± 5.0; p < 0.001). Patients with BMI < 21 were more likely to develop thyrotoxicosis (RR=6.1) compared to those with BMI >25 (p<0.001). Presence of goiter was strongly associated with AIT (RR 3.6, p=0.002). From univariate analysis we selected BMI in 3 categories (high if > 25, normal for 21 to 25 and low if < 21), gender, cyanotic status, age at start of amiodarone therapy, goiter and average amiodarone dose mg/kg body weight for further analysis. In the final model we included age at initiation of amiodarone (HR=0.98, 95% CI 0.94-102), cyanotic heart disease (HR=1.54, 95% CI 0.63-3.74) and BMI 3 categories (HR=2.29, 95% CI 1.23-4.26). The formula derived led to 3 categories of risk with the following likelihood ratios (LR): low risk LR=0.37 (95% CI 0.15-0.92), medium risk LR=1.12 (95% CI 0.65-1.91) and high risk LR=3.47 (95% CI 1.7-7.11). In the AIT series 12/23 patients (52%) were initially observed, 10 were treated medically (43%) and one had thyroidectomy (5%). 4 patients from the observation group required treatment (3 medically, one surgically) and another 3 patients from the medical group required surgery. In 20 patients hyperthyroidism resolved, one patient expired with hyperthyroidism and for 2 patients outcome assessment was not available. Complications rate was 57% (arrhythmias, hospitalizations, congestive heart failure). Median duration of disease was 184 days (inter quartile range 119-259) and discontinuation of amiodarone, AIT type or use of perchlorate did not impact duration. AIT recurrence was low at 9% (1/11 patients).
Conclusions. Low BMI at initiation of amiodarone and presence of goiter are strong predictors of AIT in this population. Its incidence is time-dependent. Our model provides a quantification of the AIT risk in amiodarone-treated adults with CHD using BMI categories and cyanotic status of the patient. This could lead the clinician to reconsider the antiarrhythmic management choice. AIT in CHD patients exhibits high prevalence, wide range of severity and relative resistance to medical therapy compared to non-CHD patients. Amiodarone continuation did not affect management outcome or disease duration. Additional studies in this high-risk population could identify elements of pathophysiology that would point towards better disease prevention and treatment.