We appreciate the letter by Sergio Barra and Rui Providencia titled ‘Too ill for an ICD?’.1 Our study2 cohort, similar to the PACE3 cohort and unlike the MADIT-RIT4 trial, consisted of a significant proportion (40.6%) of secondary prevention patients. Given the different presenting ventricular tachycardia cycle lengths in a large portion of our patients, programming of implantable cardioverter defibrillator (ICD) therapies was physician driven and non-standardized. Among our primary prevention patients implanted after 2008, the PREPARE5 protocol was frequently utilized. Given that the MADIT-RIT4 study showed that appropriate device programming is able to reduce mortality among primary prevention patients, it would be interesting to study if a similar programming strategy would be beneficial among secondary prevention patients and patients with multiple organ dysfunction.
In our cohort, there were patients with and without organ dysfunction that received ICD therapies. However, without prospective randomization, whether the antitachycardia pacing or shocks delivered were lifesaving is difficult to determine given that some ventricular tachyarrhythmia episodes treated could have been self-terminating. Furthermore, we note that only 2 of the 24 deaths within 1 year in our cohort were due to recurrent ventricular tachycardia/ventricular fibrillation with inability to restore pulse. An analysis of these two patients showed one had no organ dysfunction and the other had ≥2 markers of organ dysfunction. The overwhelming majority of deaths (n = 22, 91.7%) in our ICD cohort were from causes that an ICD would not have been able to prevent. Hence, within the limits of our study, the high 1-year mortality among multiple organ dysfunction ICD patients is likely not due to low efficacy in preventing arrhythmic deaths, but rather these patients are ill and often die from multiple other causes such as heart failure, infections, or stroke (Table 2 of our article2). Whether the mortality rate would have been significantly higher (due to arrhythmic deaths) if an ICD had not been implanted among these high risk patients—we believe that further prospective studies are required. Nonetheless, a 38.1% 1-year mortality (87.5% or seven out of eight deaths non-arrhythmic) among patients with ≥2 markers of organ dysfunction warrants serious consideration prior to ICD implantation in this group.
Conflict of interest: none declared.