OUP user menu

Do some implant too many defibrillators or others too few?

Carsten W. Israel
DOI: http://dx.doi.org/10.1093/europace/eup207 982-984 First published online: 27 July 2009

Since the landmark trials MADIT II, SCD-HeFT, and COMPANION have reported their positive results in 2002–05,13 the indications for the use of implantable cardioverter-defibrillators (ICDs) have substantially broadened. However, the ‘MADIT II shock’ affected many physicians and it has repeatedly been questioned if and how these results should be translated into daily clinical practice. Major concerns are listed in Table 1.

View this table:
Table 1

Objective and subjective reasons for not providing patients with an indication for primary prevention of sudden cardiac death with an ICD

Objective restrictions by the healthcare system (i.e. no reimbursement)
Subjective economical concerns of the physician
Number needed to treat to save 1 quality of life year clinically unacceptably high (‘too many unnecessary ICD implantations’)
Not enough qualified staff (implantation, follow-up)
Absolute or relative contraindications for ICD therapy
Significant co-morbidity of the patient
Patient's age
Non-referral (physician does not realize that the patient has an indication for ICD implantation or does not feel responsible to discuss this therapy option with the patient)
Distrust in the study results by physicians
Refusal by the patient
Concerns about lack of safety of ICD leads and/or generators
Concerns about potentially life-threatening inappropriate ICD shocks

Some more recent publications have fuelled the impression that at least in Europe and the USA, there may be too many ICD implantations. An important post hoc analysis of MADIT II suggests that patients who are (apart from their severely depressed left ventricular function) ‘too healthy’ and those who have too many co-morbidities (renal failure, atrial fibrillation, etc.) may not benefit from ICD implantation.4 Also in patients with severely depressed left ventricular function after myocardial infarction, it seems as if all-cause mortality cannot be reduced significantly if the ICD is implanted within the first month after the event.5 In addition, device and lead recalls and safety alerts question the safety of ICD therapy. Finally, a post hoc analysis of SCD-HeFT found an association between ICD shock discharges and mortality.6 Therefore, the euphoria about ICD therapy for primary prevention of sudden cardiac death has substantially diminished. Looking at the marked variability in ICD implantation rates also within developed countries in this context (Figure 17), the question arises if some countries implant too many or others too few.

Figure 1

Implantable cardioverter-defibrillator implantation rates in European countries 2005–08 (source: Eucomed, 2008.7 Reproduced with permission).

It should be clarified that the positive results of studies on ICD therapy for primary prevention were significantly in favour of ICDs, despite all inherent limitations of implantable devices. SCD-HeFT showed an impressive improvement in survival in patients equipped with an ICD compared with those without, despite the observation that the need for ICD therapy due to development of (new) ventricular tachycardia or atrial fibrillation leading to inappropriate shocks is associated with an increased risk of death. Similarly, MADIT II was positive even with all the subgroups of patients included where post hoc analyses suggest only a minor or even absent benefit. In the days of guidelines based on evidence medicine, new randomized studies rather than post hoc analyses have to prove prospectively that some subgroups of MADIT II patients may not require or may not benefit from ICD therapy.

In addition, device therapy today is safer than at the time of the MADIT II and SCD-HeFT studies since modern devices

  1. can better prevent unnecessary right ventricular pacing,

  2. prevent unnecessary shocks by better nominal programming and dedicated algorithms, and

  3. are equipped with algorithms that automatically test system integrity and provide patient alerts (or alert transmissions to a telemonitoring centre) as soon as any abnormality (e.g. fracture or insulation defect of the high-voltage lead) is detected.

Also concerns about the cost-effectiveness of ICD therapy in primary prevention of sudden cardiac death cannot be substantiated by data from published analyses. In fact, ICD implantation for MADIT II and SCD-HeFT indicated that patients are highly cost-effective, i.e. one quality of life year gained by ICD implantation in these patients is ‘cheaper’ than with many other established cardiac and non-cardiac interventions.8

There are several publications in favour of high ICD implantation rates. Parkes et al.9 demonstrated for the UK an association of ICD implantation rates and ‘ischaemic heart disease standardized mortality rate’. Regions with low ICD implantation rates had a mortality rate which was 5–15% above the average, whereas it was >10% below average in regions with a higher ICD implantation rate. One of the most intriguing analyses on this topic has been published earlier this year. Picking 1200 randomized samples of echocardiograms with a left ventricular ejection fraction <35% in an academic US hospital, authors found that even in the absence of any contraindication, only 28% of such patients received an ICD for primary prophylaxis. These ‘lucky’ patients had a mortality of 15% compared with 31% of ‘unlucky’ patients who did not receive an ICD.10 Implantable cardioverter-defibrillators were predominantly, in fact in 58%, not implanted simply because they were not referred for implantation and not because patients declined or physicians expressed concerns. Data from the Netherlands point in a similar direction. Although 91% of patients with an indication for secondary prophylaxis received an ICD, this happened only in 7% of patients with primary prophylaxis, despite the fact that only few patients (12%) were seen for the first time and the average number of physician–patient contacts was 4.6 in the preceding year.11

The study by Scott et al.12 adds some more information on the problem of non-referral. The authors observed an annual ICD implantation rate of 103 per million inhabitants in areas where the local hospital provided ICD implantation, whereas this rate was only 49 per million inhabitants in areas covered by hospitals that need to refer patients to another hospital for ICD implantation. Of note, this low ICD implantation rate in hospitals not performing this procedure themselves was independent of the presence or absence of a device specialist. The same association has also been found in a 3-year survey of the Southampton City Primary Care Trust (annual ICD implantation rate 66/million in implanting centres vs. 32–34/million in centres without this facility). In the ‘Get With The Guidelines—Heart Failure’ registry with more than 10 000 heart failure patients, Shah et al.13 found a 35-fold variation in ICD use in eligible patients between hospitals, and that the use of ICD therapy was higher in hospitals that provided cardiac procedures.

This interesting and important information is another piece in the puzzle demonstrating that the rate of ICD implantations is most likely still too low, even in well-developed industrial nations with an extensive budget for other prophylactic interventions. The reasons for ICD under-use are complex, particularly in the UK where an invasive electrophysiological study was required by the NICE guidelines before ICD implantation until 2000. Penetration of guidelines and translation of knowledge into clinical practice may be contributing factors. However, the present and other studies suggest that referral seems to be a critical barrier to appropriate diffusion of ICD therapy. Efforts should therefore be made to educate more device specialists and establish more implanting centres. Again, the scientific documentation that a therapeutic intervention reduces mortality has rarely been as convincing as in ICD implantation for primary prevention of sudden cardiac death.

Conflict of interest: C.W.I. is a member of the advisory board of Medtronic and Sorin, and a member of the speaker's bureau for Medtronic, Sorin and St Jude.


  • The opinions expressed in this article are not necessarily those of the Editors of Europace or of the European Society of Cardiology.