Smoking history and underlying lung disease are associated with poor outcome in patients developing interstitial pneumonia during low-dose amiodarone therapy

Abstract

Background

Amiodarone is a useful antiarrhythmic drug, especially in patients with serious heart diseases, but amiodarone-induced interstitial pneumonia (AMD-IP) is sometimes lethal.

Methods and results

We retrospectively analyzed the clinical characteristics of 280 patients who were treated with daily oral amiodarone. Among the patients, 29 (10.2%) developed AMD-IP during the follow-up period of 66.0±38.2 months. The average dose of amiodarone in the AMD-IP group (173±10 mg) was significantly higher than that in the non-AMD-IP group (150±3 mg). The prevalence of smoking history (AMD-IP group: 70.0%; non-AMD-IP group: 42.2%; P<0.01) and underlying lung disease (AMD-IP group: 17.2%; non-AMD-IP group: 5.6%; P<0.05) was significantly higher in the AMD-IP group than in the non-AMD-IP group. Furthermore, multiple stepwise logistic regression analysis demonstrated that smoking history was an independent predictor of AMD-IP (OR, 3.56; 95% CI, 1.08–10.23; P<0.001). Among patients who developed AMD-IP, those with chronic obstructive pulmonary disease (COPD) had a higher relapse rate and a worse prognosis.

Conclusions

During the mean follow-up period of >5 years, 10.2% of patients receiving low-dose amiodarone therapy developed AMD-IP. Higher maintenance doses, smoking history, and preexisting lung disease were related to the development of AMD-IP. The AMD-IP patients with COPD had a poor prognosis.

1 Introduction

Sudden cardiac death (SCD) in patients with low ventricular function remains a substantial problem despite improved medical therapies [1–3]. Many patients require antiarrhythmic drugs to prevent recurrent arrhythmias and reduce the number of shocks from implanted cardioverter-defibrillators (ICDs) [4]. Amiodarone, a class III antiarrhythmic drug, which prolongs myocardial repolarization homogeneously via potassium channel blockade, has been used extensively to treat life-threatening arrhythmias.

The major forms of amiodarone toxicity during chronic therapy are extracardiac [5]. Several adverse effects in many organs have been reported [5]. Among these, amiodarone toxicity, interstitial pneumonia, and pulmonary fibrosis are the most life-threatening [5,6]. According to previous studies, amiodarone-induced pulmonary toxicity has been observed in 1%–20% of patients receiving this drug [5,7–9], and there may be ethnic or racial differences in the susceptibility [7,10].

Early detection is needed to prevent amiodarone-induced interstitial pneumonia (AMD-IP). However, it has been difficult to predict which patients are at the greatest risk of the development of AMD-IP [11]. The aims of this study were to evaluate the incidence and prognosis of AMD-IP and to identify the predictors of AMD-IP in Japanese patients treated with low-dose oral amiodarone.

2 Methods 2.1 Study subjects

From January 2004 to May 2009, 373 patients were treated with oral amiodarone in Saiseikai Kumamoto Hospital Cardiovascular Center. Many previous studies have shown that AMD-IP shortly after starting amiodarone was rare [5,7–9]; therefore, we focused on the patients undergoing maintenance therapy with daily amiodarone. We retrospectively reviewed 280 consecutive patients who continued amiodarone therapy for more than 2 months.

2.1 Diagnosis of AMD-IP and extrapulmonary side effects

Several previous studies have shown that underlying lung disease was associated with the development of AMD-IP [12]. In patients who had underlying lung diseases such as emphysema, we avoided amiodarone therapy if possible. However, some patients with underlying lung disease needed amiodarone therapy because of lethal ventricular arrhythmias or frequent ICD shocks. At the beginning of amiodarone therapy, we evaluated the computed tomographic (CT) images and the pulmonary function tests in all cases, and we avoided amiodarone therapy in patients who had a background of obvious pulmonary fibrosis. Underlying lung disease in this study was defined as chronic obstructive pulmonary disease (COPD), bronchial asthma, and old tuberculosis.

The maintenance doses were adjusted according to the individual patient's conditions at a range of 50–200 mg. All patients received follow-up every 3 months at our outpatient clinic. Routine tests included physical examination, chest radiography, and serum KL-6 measurement. The diagnosis of AMD-IP was performed by more than 2 physicians on the basis of clinical data, namely, new onset of pulmonary symptoms such as dyspnea, cough, and pleuritic chest pain with pulmonary fibrosis on the CT image. Chest radiographic abnormalities, including pulmonary edema caused by congestive heart failure, infective pneumonia, and malignancy, were excluded. The patients who discontinued amiodarone only because of a marked increase of serum KL-6 without other evidence were not considered as having AMD-IP.

We also evaluated thyroid functions and serum transaminase routinely every 6 months. Thyroid abnormalities and liver dysfunction, which required dose reduction or additional medical therapy, were defined as extrapulmonary side effects.

2.3 Statistical analysis

Continuous variables, presented as mean±standard deviation, were compared by unpaired Student's t-test. Categorical variables, presented as frequency and percentage, were compared using chi-square test or Fisher's test as appropriate. The cumulative incidence was estimated using Kaplan–Meier analysis from the initiation of drug administration to the last observation day or the day of AMD-IP onset. P<0.05 was considered to indicate statistical significance. All statistical analyses were performed using StatView version 5.0 (SAS Institute, Cary, NC, USA).

3 Results 3.1 Incidence of AMD-IP

Among 280 patients who received daily amiodarone for more than 2 months, 29 (10.2%) developed AMD-IP during the observation period. The Kaplan–Meier curve, which demonstrates the incidence of AMD-IP, is shown in Fig. 1A. Mean follow-up period was 66.0±38.2 months, and mean duration from initiation of amiodarone therapy to onset of AMD-IP was 31.9±25.2 months. In all patients diagnosed as having AMD-IP, amiodarone was discontinued, and they received corticosteroid therapy.

Fig. 1. (A) Kaplan–Meier analysis showing the incidence of amiodarone-induced interstitial pneumonia (AMD-IP). Among 280 patients who received daily amiodarone longer than 2 months, 29 developed AMD-IP during the observation period. (B) Survival ratio in AMD-IP and non-AMD-IP. Kaplan–Meier analysis revealed that the patients who developed AMD-IP had a poor prognosis (P[less than]0.05). Only 5 patients died from acute respiratory failure directly due to AMD-IP.

3.2 Patient characteristics and dosage of amiodarone

The characteristics of patients who received amiodarone therapy are shown in Table 1A. Mean age at the start of amiodarone therapy was 67.3±12.5 years. There were no significant differences between the AMD-IP and non-AMD-IP groups in age, gender, body mass index (BMI), and left ventricular ejection fraction (EF). A large percentage of patients had organic heart diseases, such as dilated cardiomyopathy, hypertrophic cardiomyopathy, and ischemic cardiomyopathy (mean EF, 38.9%), but the causes of organic heart diseases were not significantly different between the AMD-IP and non-AMD-IP groups (Table 1A).

1 (A) Patient characteristics. The prevalence of smoking history (AMD-IP group: 70.0%; non-AMD-IP group: 42.2%; P<0.01) and underlying lung disease (AMD-IP group: 17.2%; non-AMD-IP groups: 5.6%; P<0.05) were significantly higher in the AMD-IP group than in the non-AMD-IP group. Data are presented as mean value±SD or number (%) of patients. EF=left ventricular ejection fraction, BMI=body mass index, DCM=dilated cardiomyopathy, HCM=hypertrophic cardiomyopathy, ICM=ischemic cardiomyopathy, VT=ventricular tachycardia. (B) Medications were not different between the AMD-IP and non-AMD-IP groups. ARB=angiotensin II receptor blocker. (C) Multiple stepwise logistic regression analysis of clinical variables demonstrated that only smoking history was a significant independent predictor of AMD-IP (Table 1C). OR=odds ratio, CI=confidence interval.

(A)
	AMD-IP	non-AMD-IP	P
n	29	251
Agge (y.o.)	69.6±13.6	67.0±12.3	NS
Gender (Male)	25 (86.2%)	187 (74.5%)	NS
EF (%)	36.8±12.7	39.2±15.3	NS
BMI	22.5±3.6	22.9±3.7	NS
DCM	8 (27.5%)	74 (29.5%)	NS
HCM	6 (20.7%)	56 (22.3%)	NS
ICM	13 (44.8%)	73 (29.1%)	NS
Valvular disease	2 (6.9%)	19 (7.6%)	NS
Congenital disease	0 (0%)	6 (2.4%)	−
Idiopathic VT	0 (0%)	6 (2.4%)	−
Others	0 (0%)	17 (6.8%)	−
Lung disease	5 (17.2%)	14 (5.6%)	<0.05
Smoking history	20 (70.0%)	106 (42.2%)	<0.05
(B)
	AMD-IP	non-AMD-IP	P
n	25	246
ACE inhibitors	8 (32%)	95 (38.6%)	NS
ARB	5 (20.0%)	81 (32.9%)	NS
β blockers	21 (84.0%)	183 (74.4%)	NS
Statins	5 (20.0%)	64 (26.0%)	NS
Diuretics	19 (76.0%)	144 (58.5%)	NS
Digitalis	5 (20.0%)	25 (10.2%)	NS
(C)
Valiables	OR	95% CI	P value
Lung disease	3.08	1.39–7.37	0.07
Smoking history	3.56	1.08–10.23	<0.01

Underlying lung disease was correlated with a higher incidence of AMD-IP (AMD-IP group: 17.2%; non-AMD-IP group: 5.6%; P<0.05). Furthermore, the prevalence of smoking history was significantly higher in the AMD-IP group than in the non-AMD-IP group (AMD-IP group: 70.0%; non-AMD-IP group: 42.2%; P<0.01; Table 1A). Multiple stepwise logistic regression analysis of clinical variables demonstrated that only smoking history was a significant independent predictor of AMD-IP (Table 1C).

Medical treatments were not significantly different between the AMD-IP and non-AMD-IP groups (Table 1B). However, the maintenance doses of amiodarone were significantly higher in the AMD-IP group than in the non-AMD-IP group (AMD-IP group: 173±10 mg; non-AMD-IP group: 150±3 mg; P<0.05; Fig. 2A).

Fig. 2. Maintenance doses of amiodarone. (A) Maintenance doses of amiodarone were significantly higher in the AMD-IP group than in the non-AMD-IP group (AMD-IP group: 173±10mg, n=29; non-AMD-IP group: 150±3mg, n=251; P[less than]0.05). (B) Among patients who developed AMD-IP, patients with COPD had higher mortality than those without COPD.

3.3 Extrapulmonary side effects and AMD-IP

Thyroid abnormality, which required additional medical therapy, was observed in 35 of 280 patients (12.5%). Liver dysfunction, which required a dose reduction of amiodarone, was seen in 3 of 280 patients (1.1%). Neither the incidence of thyroid abnormalities nor that of liver dysfunction was different between the AMD-IP and non-AMD-IP groups (thyroid dysfunction: AMD-IP, 5 [18.5%]; non-AMD-IP, 30 [11.8%]; liver dysfunction: AMD-IP, 0 [0%]; non-AMD-IP, 3 [1.2%]). Extrapulmonary side effects were not common in AMD-IP patients.

3.4 Survival rates and causes of death

Among 280 patients treated with oral amiodarone as maintenance therapy, 55 (19.6%) died of any cause during our follow-up period. As shown in Fig. 1B, the Kaplan–Meier analysis revealed that the patients who developed AMD-IP had a poor prognosis. Five patients died of acute respiratory failure directly due to AMD-IP (1.8%), even though all patients who developed AMD-IP received oral or intravenous corticosteroid therapy. Furthermore, we found that the incidence of SCD was significantly higher in the AMD-IP group than in the non-AMD-IP group (AMD-IP group: 23.5%; non-AMD-IP: 10.5%; Table 2A).

2 (A) Causes of death in AMD-IP patients. The number of sudden cardiac deaths was significantly higher in the AMD-IP group than in the non-AMD-IP group. Five AMD-IP patients died directly because of respiratory failure caused by AMD-IP. (B) AMD-IP patients with COPD had a higher relapse rate and poor prognosis.

(A)
	AMD-IP	non-AMD-IP	P
n	29	251
Total death	17	38
CHF	5 (29.4%)	21 (55.3%)	NS
Sudden death	4 (23.5%)	4 (10.5%)	<0.01
AMD-IP	5 (29.4%)	0 (0%)	−
Others	3 (17.6%)	13 (34.2%)	NS
(B)
	with COPD	without COPD	P
n	5	24
Relapse of AMD-IP	5 (100%)	1 (4.1%)	<0.01
Death for AMD-IP	3 (60%)	2 (8.3%)	<0.05
Total death	5 (100%)	12 (50%)	<0.01

3.5 AMD-IP and underlying lung disease

In this study, the patients who developed AMD-IP were more likely to have underlying lung disease (AMD-IP group: 17.2%; non-AMD-IP group: 5.6%; P<0.05; Table 1A). Furthermore, AMD-IP patients with COPD had a higher relapse rate and poor survival (Fig. 2B). We observed 6 relapse cases of AMD-IP after initial treatment including corticosteroid therapy. In these relapse cases, 5 patients had COPD (Table 2B). The deaths of 3 of these patients were directly attributed to respiratory failure caused by AMD-IP (Table 2B.) All AMD-IP patients with COPD died during our observation period.

4 Discussion

Amiodarone is an antiarrhythmic drug, which is widely used around the world. However, its pulmonary side effects are sometimes lethal. In this study, we found that 10.2% of patients developed AMD-IP during prolonged low-dose amiodarone therapy (mean follow-up period, 66.0±38.2 months). We show 2 major findings. First, although the maintenance doses of this observation was relatively low compared to that of previous studies (mean dosage, 152 mg/day) [1–5], higher maintenance doses were related to the occurrence of AMD-IP. Second, the patients with a smoking history and preexisting COPD were more likely to develop AMD-IP. In particular, we showed for the first time that smoking history was an independent predictor of AMD-IP. In addition, the AMD-IP patients with COPD had a poor prognosis resulting from relapsing AMD-IP.

Neurologic and gastrointestinal tract toxicity of amiodarone are dose related [6,13], but pulmonary toxicity has been reported to be related in part to drug dosage [6,13–15]. It has been reported that pulmonary toxicity is uncommon at maintenance doses less than 300 mg/day [14,15]. Some studies indicate that there is no statistically significant difference in the rate of pulmonary adverse effects between placebo and low-dose amiodarone [13,16]. However, AMD-IP can occur during prolonged therapy with doses at below 200 mg/day [17,18]. The average dose of amiodarone in the present study was relatively low (152 mg/day) compared with previous studies; however, the doses used in the AMD-IP group were significantly higher than those in the non-AMD-IP group. Yamada et al. reported the incidence of AMD-IP to be 7.8% and 10.6% at 3 and 5 years, respectively, even though the mean maintenance dose was only 141 mg/day in their study [10].

AMD-IP was significantly associated with a poor prognosis. Mortality in AMD-IP is reported to range between 21% and 33% in patients who were admitted to the hospital [9,19–21]. In the present study, AMD-IP was not only a major cause of acute respiratory failure leading to death but also increased the risk of SCD. We use amiodarone commonly for ventricular arrhythmia patients with organic heart disease and poor left ventricular function, rather than atrial arrhythmias. For patients with poor left ventricular function, AMD-IP might worsen heart failure caused by hypoxemia and secondary pulmonary hypertension, and withdrawal of amiodarone may result in sudden death or poor prognosis.

Higher age [7,9], cumulative dosage [5], reduced predrug diffusion capacity [12], underlying lung disease [5,7–9,22,23], and recent pulmonary insults [5] are reported to be risk factors of AMD-IP. Although it is arguable whether pulmonary toxicity occurs to a greater degree in patients who have underlying pulmonary disease, such as COPD [6,11], we clearly showed that the likelihood of developing AMD-IP was higher in patients with COPD and a smoking history than in patients without COPD or a smoking history. Many studies support our results that COPD is a risk factor for AMD-IP [23–25]. However, we showed for the first time that smoking history per se could predict AMD-IP. The mechanisms of pulmonary toxicity are unclear, but amiodarone may interfere with lysosomal phospholipases, leading to phospholipidosis [26]. It has been suggested that an immunologic reaction may also produce a hypersensitivity response [22]. Some studies have suggested that preexisting lung disease might predispose the patient to lung toxicity [12,27], even though this has not been verified by other studies [25,28]. We also showed that patients with COPD had a higher relapse rate and poor prognosis.

This is the first report demonstrating that smoking history itself might predict the development of AMD-IP. Patients on amiodarone who have COPD and a smoking history should be monitored carefully and routinely for pulmonary side effects.

Conflict of interest

None.