Defibrillation Testing of the Implantable Cardioverter Defibrillator: Habit or Necessity?


Luis A. Pires, MD, FACC
Heart Rhythm Center and Cardiovascular Medicine, St. John Hospital and Medical Center, and Wayne State University School of Medicine
Detroit, MI

The implantable cardioverter-defibrillator (ICD) is an effective therapy for ventricular tachycardia (VT) and ventricular fibrillation (VF) because of its ability to detect and promptly terminate spontaneous VT/VF. From its beginning, ICD testing at the time of device implantation to confirm proper detection and successful termination of induced VT/VF has been considered standard practice. Indeed, the clinical trials which established the role of the ICD were based on patients who underwent some form of intraoperative testing. The amount of energy (Joules) needed to terminate VF is used to establish a given patient’s defibrillation energy requirement (DER).

Advances in device and lead technology facilitate successful device implantation (i.e., DER that is ≥10 J lower than the device’s maximum output) in nearly 100% of cases.1-5 As a result, some investigators6 have recently questioned the necessity of ICD testing altogether, noting that, among other considerations, forfeiting ICD testing might raise its usage by allowing procedures by non-electrophysiologists with reduced training requirements.7 Existing data, however, does not support wholesale abandonment of ICD testing.

When To Test

In the absence of specific contraindications to testing (Table 1), VF is induced to ensure that it can be reliably sensed and promptly terminated by the device — preferably by the first programmed shock. Some of the contraindications are absolute (e.g., cavity thrombus), but others can be overcome. Among a cohort of 835 consecutive patients at our center, defibrillation testing was withheld in 203 (24%), of which ~70% were due to the presence of cavity thrombi, inadequate anticoagulation and intraoperative hypotension.1 Some have postponed testing during cardiac resynchronization therapy defibrillator (CRT-D) implantation (Figure 1) for fear of coronary sinus lead dislodgment, but we have not found this to be a significant problem among a cohort of >700 patients (unpublished observations). Unless a patient receives appropriate ICD therapies for VF, the need for testing should be reassessed at ~6 to 8 weeks, when system revision, if needed, can still be done fairly easily.

Defibrillation testing at times other than the initial implant is reasonable in cases where DER is expected to rise, such as amiodarone therapy, and at times of generator replacement in patients who had not received appropriate therapies against VF in the near past or had shown changes in clinical status that may adversely raise DER.

How To Test

After determining adequate sensing (>5 mV), pacing (

The defibrillation threshold (DFT), defined as the lowest amount of energy capable of terminating an episode of induced VF, is most commonly determined through a step-down method (i.e., successive lowering of shock strength). Because success of defibrillation is probabilistic, a true DFT cannot be established with certainty. For this reason, plus the potential risk of repeated defibrillations,8 determining a true DFT is now rarely undertaken. The goal of ICD testing is to determine an energy level, or defibrillation safety margin (DSM), that has a reasonable chance of success against spontaneous VT/VF events. This can be established by one or rarely two episodes of induced VF, and should be low enough to be at least ≥10 J less than the device’s maximum output.1-5 This 10 J “safety margin,” achievable in nearly all patients with modern ICDs,1-5 has been a common practice because patients with elevated DER (and earlier monophasic devices) were thought to have higher mortality. With modern, biphasic devices, elevated DER has not been shown to adversely impact patient survival.9,10 A single defibrillation success at 10-15 J adequately predicts successes with stronger shocks, thereby eliminating the need for additional tests. We follow a simple testing protocol: we test once at 10 or 15 J depending on the device’s maximum output; and if the first shock fails, a second shock of 10 J greater (20 or 25 J) is delivered (Figure 2), still within a ≥10 J safety margin provided by current devices. If the second shock fails, simple system modification(s) usually results in DER that meets the ≥10 J safety margin. To minimize risk of testing, some investigators have advocated the use of “shock-free” testing, or upper limit of vulnerability, as an alternative.5

Risk and Benefit of ICD Testing

Defibrillation testing can lead to serious complications and even death. A recent Canadian study involving 19,067 ICD implants reported three testing-related deaths (0.016%), five strokes (0.026%) and 27 post-testing prolonged resuscitations (0.14%), two of which had significant clinical sequelae.8 Interestingly, two of the deaths occurred from PEA following the second fibrillation/defibrillation episodes that were preceded by successful first shocks with energies ≥10 J lower than the devices’ maximum output. Given that a single test is more than adequate,1-5 the deaths may have been prevented if the second tests were not performed. The five testing-related strokes all occurred in patients with persistent atrial fibrillation (and post defibrillation sinus rhythm in all but one patient), suggesting that despite having been adequately anticoagulated, it might be reasonable to postpone testing until 3+ weeks of stable postoperative anticoagulation in such patients — a practice we follow at our center. However, the actual adverse event rates related to ICD testing may be higher than those reported, since in many cases patients with advanced heart failure and/or severe left ventricular (LV) systolic dysfunction are often not tested. It is not clear at what level of LV dysfunction defibrillation testing should be forgone; we use a cutoff of LV ejection fraction (EF) of ≤10%, but we have not seen more adverse complications in patients with LVEF of, say, 20%.

First shock successes for VT/VF are quite high regardless of mode of testing, DFT/DSM results, and whether or not testing was performed at all.1-5 We reported similar sudden death free survival for patients who had VF testing (DFT or DSM) versus no testing at the time of implantation, although overall survival, while similar for DFT and DSM groups, was significantly worse in those who did not undergo testing.1 It is not clear, however, if the higher mortality in the non-tested group reflects the fact that they may have been “sicker,” a potential reason why, in some cases, they were not subjected to testing. Others have reported similar two-year mortality in tested versus non-tested patients who received ICDs for primary prevention.11 The potential long-term risk of no testing, if any, is unknown and would require a randomized trial involving thousands of patients. The comparable survival rates in the DFT and DSM tested groups, on the other hand, are reassuring given that for many years now, DSM testing has replaced DFT testing in clinical settings.

The ICD’s primary function is to abort would-be fatal VT/VF events, but such protection is not absolute. With rare exceptions, device malfunctions (specifically, defibrillation failure) very infrequently account for sudden death in ICD recipients,12,13 but rather the result of acute cardiac and non-cardiac factors that cannot be anticipated or tested for at the time of implant. The likelihood of shock failures can also be minimized by a greater use of antitachycardia pacing therapies, since nearly half of “VF” events are actually rapid VTs.

Who Should Test

ICD testing has always been the domain of cardiac electrophysiologists, even when devices were implanted by cardiac surgeons. In light of recent proposals sanctioning alternative training pathways for device implantation,7 and heavy industry promotion, device testing (usually DSM) is now also performed by non-electrophysiologists. There exists no testing-related safety data of patients treated by non-electrophysiologists, making specific recommendations as to whom should test ICDs problematic. Nonetheless, it should be noted that non-electrophysiologists are not expected to participate in the implantation of devices in patients with documented VT/VF (i.e., secondary prevention) for whom careful testing is especially important to assure proper device function. Moreover, non-electrophysiologists must be fully aware of patient- and device-specific factors that influence defibrillation success and proper troubleshooting methods to ensure successful implantation in each case.


With current ICDs, successful device implantation (i.e., a DER energy allowing for a ≥10 J programming safety margin), determined through a single test, can be expected in nearly all patients. Since such minimal testing rarely results in adverse events and there are no prospective data on the outcome of ICD recipients whose devices are not tested intraoperatively, we feel that, in the absence of contraindications, a minimum of testing is still appropriate, with each patient approached individually. Certainly we do not believe testing should be abandoned to simply facilitate more device implantations by non-electrophysiologists who may be in some cases uncomfortable with defibrillation testing.


Would like to see opinions re. the necessity for annual
induced ventricular fibrillation testing in the absence of any detected abnormalities, and not associated with potential lead defects.This is a standard recommendation to all patients with no contra-indications by some electrophysiologists, and unheard of by cardiologists in other parts of the country.

Add new comment