Another disappointing result with implantable cardioverter-defibrillator therapy in patients with Chagas disease

We disagree with Barbosa et al.'s1 conclusion from their study, published in February 2013, that implantable cardioverter-defibrillator (ICD) is ‘the standard therapy for secondary prevention in patients with Chagas disease (ChD) and the need for a randomized clinical trial (RCT) is doubtful and, maybe, unethical.’ Their conclusion was based on the largest study of the efficacy and safety of ICD therapy for secondary prevention in ChD vs. non-ChD patients. After a median follow-up of 8.7 months, all-cause mortality (the only indisputable and unbiased endpoint in ICD recipients) was similar in the two groups: 12.3% in 65 ChD patients versus 11.4% in 70 non-ChD patients (P = 0.82). But this conclusion is neither sound nor supported by the authors' findings. First, the levels of all-cause mortality reported for both groups are very high for a median follow-up of only 9 months. We are not aware of any RCTs or registries that show such a high mortality in non-ChD patients of similar age and clinical characteristics. However, in patients with ChD, Cardinalli-Neto et al.,2 in 2007, also showed high mortality in 90 patients resuscitated from cardiac arrest or with sustained ventricular tachycardia (SVT) accompanied by haemodynamic compromise who were treated with an ICD. During a mean follow-up of 25 months, 31 of 90 patients (34%) died. Total mortality rates were exceedingly high (18% at 1 year, 27% at 2 years, 40% at 3 years, 50% at 4 years, and 73% at 5 years of follow-up), and the number of shocks per patient by day 30 was the only independent predictor of all-cause mortality. Notably, the annual mortality rate was 16.6%, which is also higher than the annual mortality rates from historical series in which patients with SVT and a similar left ventricular ejection fraction (LVEF) were treated with amiodarone.3 Second, in other forms of cardiomyopathy, ICDs are effective only in patients with very poor LVEF, whereas LVEF was 37% in Barbosa et al.'s1 study and 47% in Cardinalli-Neto et al.'s2 study. According to a meta-analysis of data from three RCTs (AVID, CIDS, and CASH), in patients with LVEF of >35%, all-cause mortality was 20% higher in the ICD group than in the amiodarone group, but this difference was not statistically significant.4 Finally, we were surprised that, in a consecutive series of patients treated with an ICD, 91% had recovered from cardiac arrest.1 One may wonder if this information is accurate. In this regard, there is a misconception that patients with life-threatening ventricular tachyarrhythmias with haemodynamic instability or who have been resuscitated from cardiac arrest are at the highest risk of arrhythmic death.5 But this claim has been refuted by several large observational studies.6–8 Multivariate analysis of predictors of death in patients with SVT has shown that arrhythmia presentation has no effect on outcome.9 In addition, in patients with ChD, syncope during the presentation of SVT has not been associated with an increase in total or cardiac mortality.10 So, the view that the need for a RCT in ChD patients is doubtful and unethical should be vigorously contested. Instead, an RCT comparing ICD with amiodarone is urgently needed to establish the most appropriate form of therapy in ChD.3 Conflict of interest: none declared.