ICD Shocks : Not just the “Straw That Broke The Camel’s Back”

In this issue of the Journal, Bhavnani et al ¹ report on the prognostic impact of shocks delivered for induced arrhythmias at the time of device implant on morbidity and mortality of patients receiving implantable defibrillators (ICD). This is a very important study and attempts to address the question of whether the observed increase in mortality and morbidity after ICD shocks is due to the shock itself, or because of the substrate that led to the arrhythmia. The authors compared heart failure occurrence and mortality in patients who had ICD shocks during non invasive programmed stimulation (NIPS) testing, with patients who received shocks during follow up for spontaneous arrhythmia. All of the study patients had defibrillation threshold (DFT) testing at the time of the device implant. DFT testing is routinely done at many centers including ours, although the enthusiasm is dwindling. The low failure rate for defibrillation during DFT testing and the very high reliability of the current generation ICDs make DFT testing superfluous. Nevertheless, the question remains as to whether it is the shock or the arrhythmia that is harmful and the authors have picked the right model to address this question. The importance of this question deserves some introduction. High energy DC shocks are no doubt life saving, but it comes at a cost. Basic studies on ventricular myocytes have shown cellular injury by way of electroporation of the bi-lipid membrane and abnormal cellular calcium loading ^{2, 3}. Other effects of DC shocks that are well characterized include, proarrhythmia including ventricular fibrillation⁴, depression of mechanical function, myocardial stunning and electromechanical dissociation^{5, 6}, abnormal cellular conduction⁷, all of which are related to the shock strength. Spontaneous ICD shocks have been associated with elevated serum markers of myocardial injury such as Troponin I serum levels⁸. Although the majority of these effects have been characterized in cellular models, one can only postulate that these detrimental effects are further magnified in the presence of pre-existing myocardial dysfunction. In the MADIT-II (Multicenter Automatic Defibrillator Implantation Trial-II) study ⁹, patients who received an appropriate ICD shock had a 3-fold increase in later mortality compared with those who did not. Similarly, in the SCD-HeFT (Sudden Cardiac Death in Heart Failure Trial) study ¹⁰, an appropriate shock was associated with a 5-fold increase in later mortality. Thus, there is overwhelming evidence that DC shocks are probably not entirely safe and there is a dose response to the myocardial injury due to DC shocks.

Given this background, the purpose of the study by Bhavnani et al. was to assess whether the “controlled” test shock given during NIPS testing has any bearing to long term outcomes compared to shocks delivered for spontaneous arrhythmia. The authors retrospectively analyzed data collected from 1372 patients who underwent ICD testing (DFT testing) at the time of ICD implantation performed between 1997 and 2007 at a tertiary care center. After the initial implantation test shock, some patients underwent additional NIPS testing within a few weeks of the implant. Patients were followed after implant for both appropriate and inappropriate ICD therapies. Only the first occurrence was recorded and no information is available on the subsequent therapies. Heart failure related admissions and all cause mortality were the end points. Based on the type of shock received, the study patients were stratified into four groups: 1. Implantation shocks only, 2. Additional NIPS shock only, 3. Inappropriate shock only and 4. Appropriate shock only. Groups 2–4 had already received DFT shocks during device implantation. Over a mean follow up period of 807 ± 805 days, 26% of the study population died. Death occurred in 20%, 27%, 46%, and 32% of patients with implantation shocks only, NIPS, appropriate, and inappropriate shocks respectively. In multivariate analysis, appropriate shock was an independent predictor of mortality with a greater than 2-fold increase in probability of death as compared to the implantation shocks [adjusted hazard ratio (AHR) 2.09 (95% CI 1.38 – 2.69) p<0.001] or NIPS shocks [AHR 2.61 (95% CI 1.86 – 3.67) p<0.001] groups. Recipients of NIPS or inappropriate shocks had similar mortality as compared [AHR 0.91 (95% CI 0.69– 1.20), p=0.491] and [AHR 1.16 (95% CI 0.78 – 1.74), p=0.461] to recipients of implantation shocks respectively. Based on these findings, the authors argue that shocks delivered in a controlled setting for device testing may not have any bearing on the long term outcome of the patients who get ICDs. On the other hand, patients with high recurrence of ventricular arrhythmia, the appropriate ICD shocks group, have a higher incidence of heart failure and mortality by virtue of the substrate. These data seem to suggest that the appropriate shock itself may be a surrogate of patients with high risk of mortality, but does not contribute significantly to its occurrence. Adding to this notion is the lower heart failure admission rate in patients who received inappropriate shocks compared to those who had appropriate shocks. The study is elegantly done and the authors postulate that the factors that underlie the occurrence of arrhythmias such as ischemia, progressive LV dysfunction and electrolyte imbalances might also contribute to the observed mortality.

Although the data are convincing, the retrospective nature of the study inherently introduces selection bias on the “control group”, i.e. the group that had NIPS testing. Other factors such as lack of standardized programming, and the use of antiarrhythmics, which affect ICD firing and mortality, need to be taken into account while interpreting the results. The patients grouped under the “appropriate shock only” are clearly the sickest group of. This group is populated by patients with heart failure and low LVEF, with a significantly higher proportion receiving biventricular devices at implant. The fact that the mortality risk persisted even after adjusting for the known variables is not surprising. This might be because of the segregation of other non-cardiac co-morbidities in this group that have not been accounted for or because of lack of follow up information as to the course of their structural heart disease/vascular disease. No information is available on the cumulative number of appropriate therapies in the “appropriate therapy only” group to decide if this is comparable to the NIPS only group where the median number of shocks was 3 per patient. Also, given the long follow up period, nearly all of the survivors would have had device replacement with repeat DFT testing and possibly NIPS which further confounds the conclusions of the study. Finally, there is no true control group of patients who at implant had no DFT or NIPS testing and had no ICD firings during follow up. Comparing this group with the NIPS only group would have teased out whether ICD shocks themselves contribute to mortality.

The “appropriate shock only” group by virtue of advanced heart failure at implant would be expected to have a higher number of heart failure related admissions during follow up. Ventricular and atrial arrhythmias are often exacerbated during episodes of decompensated heart failure and ICD shocks during this time are more likely to trigger an admission for heart failure. The authors rightly state that aggressive treatment of heart failure might reduce the risk of mortality and in fact may also decrease the burden of ICD shocks.

The overall message of the study is consistent with the current literature in that it shows a clear association between “appropriate ICD shocks” and mortality¹¹. As arrhythmia recurrence is synonymous with ICD therapy in patients with ICDs, the issue of what actually contributes to the observed mortality is hard to resolve in a prospective study. Even inappropriate shocks have been shown to increase mortality in prospective studies. This, along with the basic studies, has shifted the burden of proof to the believers of the substrate hypothesis. The truth may be somewhere in between these two hypotheses. A high energy shock may be more detrimental to a failing heart where the cellular reparative mechanisms are already overburdened. Electrophysiologic and mechanical consequences of the shock may just be more than the “straw that broke the camel’s back” and sends the patients in a downward spiral of heart failure and ultimately death.

Unfortunately, there is no alternative to DC shock. The last modification in shock waveform to a biphasic waveform is almost 2 decades old. It seems that we have embraced biphasic DC shocks and stopped looking further. The growing number of ICD implants and the increasing incidence of heart failure in patients with ICDs underscore the need to explore novel ways to terminate fibrillation. However, until such time, physicians should pay attention to aggressive medical management of heart failure and appropriate device programming to avoid unnecessary shocks.