Amiodarone and Pulmonary Toxicity in Atrial Fibrillation
Study Questions:
What is the association of continuous low-dose amiodarone exposure and the risk of interstitial lung disease (ILD), primary lung cancer (PLC), and all-cause mortality in patients with atrial fibrillation (AF)?
Methods:
This was a retrospective nationwide cohort study using electronic health records from Israel’s largest health maintenance organization. The final study population was developed in four steps. First, a preliminary cohort of patients with a new diagnosis of AF who were treated with amiodarone between 1999 and 2001 was identified. Next, amiodarone dispensing was verified and verified patients then matched with unexposed controls based on time elapsed since AF diagnosis, age, sex, and ethnicity and those were given the exact same start date. Next, pairs were excluded if they had a diagnosis of ILD or PLC prior to the start date. Next, patients/pairs were excluded if the amiodarone daily dose dispensed was not between 190 mg and 210 mg daily. Therefore, those taking higher doses or long periods of interruption were not included.
The primary analysis compared continuous exposed versus never exposed as an intention to treat analysis of where patients were censored at either 10 years of follow-up, 10 years after diagnosis of AF, or occurrence of a study outcome. A target trial emulation sensitivity analysis was performed accounting for short-term amiodarone use in either group and early diagnosis of ILD or PLC during the first year. Treatment effect estimates were calculated using the inverse probability treatment weighting with variables included determined by a directed acyclic graph using numerous patient covariates (e.g., congestive heart failure, anticoagulation, and prior myocardial infarction, among others). Covariate balance was determined by Kolmogorov Smirnoff, Kaplan-Meier curves constructed, and confidence intervals calculated. Events were considered from first dispensing date to 42 days following the last recorded dispensing date.
Results:
The final study population was 6,039 amiodarone patients and 6,039 matched controls. The median time from AF diagnosis to amiodarone exposure was 242 days and the median amiodarone daily dose was 200 mg. There were numerous differences between groups at baseline including amiodarone patients being more likely to have diabetes, chronic kidney disease, be prescribed beta-blockers as well as other antiarrhythmic drugs and oral anticoagulation. These covariates were well balanced in the model. As would be expected, amiodarone-treated patients were more likely to be referred earlier and more often for pulmonary assessments.
During a median follow-up of 4.2 years, ILD was diagnosed in 2% of all patients but was not more frequent in amiodarone-treated patients (hazard ratio [HR], 1.45; 95% CI, 0.97-2.44). During a median follow-up of 4.1 years, PLC was diagnosed in 0.8% of all patients and not different between amiodarone-treated patients and control (adjusted HR, 1.18; 95% CI, 0.76-2.08). All-cause death occurred in 18.1% and was lower in amiodarone-treated patients (adjusted HR, 0.65; 95% CI, 0.60-0.72) at a median follow-up of 4.9 years. The sensitivity analyses had similar findings with the exception that when confined to a median follow-up of 242 days, there was a small increased risk in ILD (HR, 1.15; 95% CI, 1.02-1.27) in the as-treated group. Additional exploratory analyses did find that amiodarone was associated with other well-known adverse effects including hyperthyroidism, hypothyroidism, and hepatic disorders.
Conclusions:
In a modern era population, long-term, low-dose amiodarone use was not associated with ILD or PLC and may be associated with lower mortality.
Perspective:
Amiodarone has been used off-label for AF for >30 years and is a first-line rhythm-control strategy for patients with heart failure. Both direct cytotoxicity and immune-mediated inflammatory responses have been proposed as mechanisms. The incidence of pulmonary toxicity with amiodarone is also known to be dose related. This retrospective observational health records research sought to evaluate the risk of ILD in a contemporary AF patient population between patients taking amiodarone continuously over several years versus matched unexposed controls. The fact that patients taking amiodarone had increased referrals for pulmonary evaluations suggests that practitioners were following current recommendations for amiodarone toxicity monitoring, which include chest X-ray and pulmonary function testing only if toxicity is suspected. The incidences of both ILD and PLC were both low.
Only low-dose amiodarone was evaluated, so the research does not exclude factors other than dose, which could be responsible for reduced pulmonary toxicity such as the lower use of intravenous amiodarone, which at one time contained polysorbate 80 and is also associated with pulmonary toxicity. While all potential differences between the treatment group and control cannot be addressed, the authors are to be commended for careful consideration and refinement of the comparator groups to most closely as possible simulate a randomized controlled trial. Another strength of the study lies in the long duration of follow-up for patients under a single insurer, which most likely would not have been possible if conducted in the United States in this group of older persons over the age of 70 years.
Clinical Topics: Arrhythmias and Clinical EP, Atrial Fibrillation/Supraventricular Arrhythmias
Keywords: Atrial Fibrillation, Lung Diseases, Interstitial, Lung Neoplasms
< Back to Listings