Chicago, IL - The reliability of implantable cardioverter-defibrillators (ICDs) has followed a bumpy course since the early 1990s, showing general improvement until malfunction-related replacement rates started to climb later in the decade, according to an analysis of industry data submitted to the US Food and Drug Administration [1]. But in a separate review of registry data that came to similar conclusions but covered additional years, ICD reliability started to emerge from its slump in 2002 [2]. In contrast, according to both studies, published in the April 26, 2006, Journal of the American Medical Association, pacemaker-only devices have been increasingly dependable year after year.

These studies involve large patient groups and not individuals. Each patient and each device system pose unique issues.

An accompanying analysis of Canadian data cautioned that replacement of ICDs that have been the subject of safety advisories, including recalls, entail "substantial" risk of major complications—primarily pocket infection or hematoma [3].

"These investigations contribute important new data to the information mosaic involving pacemakers and ICDs and should enable clinicians and patients to make better-informed decisions about these devices," Dr Bruce L Wilkoff (Cleveland Clinic, OH) writes in an accompanying editorial [4]. "However, these studies involve large patient groups and not individuals. Each patient and each device system pose unique issues," he observes, noting that pacemakers and ICDs have "dramatic life-sustaining and lifesaving effectiveness despite their frailties."

In one of the two analyses, both led by Dr William H Maisel (Beth Israel Deaconess Medical Center, Boston, MA), the annual reports submitted by industry to the FDA from 1990 to 2002 disclosed generator malfunctions in 17 323 of 2.25 million pacemakers and 415 780 ICDs in use. Only single-chamber devices were included in the analysis. Lead systems were not considered. A malfunction was defined as a problem leading to device explantation and return to the manufacturer plus subsequent confirmation of inappropriate function.

Maisel, the current chair of the FDA Circulatory System Medical Devices Advisory Panel, prepared the report with coauthors from the agency's Center for Devices and Radiologic Health.

According to industry figures for the US:

• Annual pacemaker insertions grew from nearly 95 000 to >267 000 over the 13 years starting in 1990, while annual ICD implants exploded from about 6500 to nearly 100 000.
• Annual malfunction rates ranged from 1.4/1000 to 9.0/1000 for pacemakers and from 7.9/1000 to 38.6/1000 for ICDs. The means were 4.6/1000 and 20.7/1000, respectively (p<0.001).
• Malfunction rates for pacemakers fell continuously throughout the study period (p<0.006 for trend) and for ICDs trended unevenly downward for the first half of the 13 years before climbing sharply in the remaining period.
• Battery or capacitor problems accounted for almost a fourth of all device malfunctions and were, proportionally, about twice as common in ICDs as in pacemakers.
• Malfunctions were "directly responsible" for 61 confirmed deaths overall, involving 30 patients with pacemakers and 31 with ICDs. But additional, unwitnessed deaths likely added to the total, the authors write.
• Although confirmed malfunctions were about equally split between the two device types, ICD patients had 5.6 times the associated risk of death compared with the much larger pacemaker population.

Because ICDs are substantially more sophisticated than pacemakers, it is not surprising that they have a higher device-malfunction replacement rate.

"Because ICDs are substantially more sophisticated than pacemakers, it is not surprising that they have a higher device-malfunction replacement rate," write Maisel and colleagues. He attributed many of the ICD problems to certain device components that were undergoing rapid evolution. "During this latter period, ICDs shrank in size, maintained their high energy output, and provided more therapeutic features," according to the group. "The decrease in size, in particular, has necessitated modifications to battery, capacitor, and circuitry design that appear to account for some of the observed malfunctions."

The device, the leads, the chosen programmed parameters, and the patient's response are all interdependent.

In contrast, they write, pacemaker technology was relatively mature in 1990 and so the devices "did not undergo the same degree of integral component modification during the study period."

The analysis's most important limitation, according to Wilkoff's editorial, may have been omission of malfunctions related to lead systems and other variable aspects of ICD therapy. "The device, the leads, the chosen programmed parameters, and the patient's response are all interdependent," he writes. "The system malfunction rate may be the most important measure of reliability, but it is also the most elusive, because no single measure captures all elements in a verifiable fashion."

Still, Wilkoff adds, the Maisel et al report's likely underestimated malfunction rates are "useful" and appear consistent with Maisel's meta-analysis in the same issue of JAMA, one exploring similar event rates using "substantially different methodology."

After screening Medline and other databases, Maisel found three prospective pacemaker and ICD registries that met his standards for device-reliability information, including one each from Denmark, the UK, and North America. The meta-analysis defined device malfunction more broadly than the FDA report: pacemaker or ICD failure involving the primary device components but excluding the lead system. Still, its findings were largely consistent with the other report:

• Pacemaker malfunction rates reached 12.4/1000 person-years in 1983. By 1988, they had plummeted (p=0.009) to about 2.0/1000 person-years or lower until the final year of the analysis, 2004.
• ICD malfunction rates from 1987 to 2004 were more variable, falling from about 52.5/1000 person-years to a low of 5.6/1000 person-years in 1998 (p<0.001). But then they rose again to 26.3/1000 person-years by 2001 (p<0.001).
• From 2002 to 2004, ICD malfunction rates declined again to their lowest levels.
• Once again, regardless of device type, battery failure was the most common malfunction.

Our data suggest that strategies focusing on ICD-generator replacement involve meaningful risk. These issues are amplified in primary-prevention patients in whom the risk/benefit ratio may be more difficult to assess.

Reiterating that progressive ICD design changes may have increased their malfunction rates, Maisel also writes that that most of the sixfold reduction in such rates for pacemakers during the 1980s corresponded to a time of "rapid technological advances, including the introduction of dual-chamber pacing, improved programmability, increased longevity, reduced size, and more comprehensive diagnostics." Yet there was little change in pacemaker reliability, he notes. "Pacemakers, therefore, are evidence that significant technological advancements may be achieved without sacrificing device reliability."

Ongoing surveillance of implantable-device reliability is clearly needed, according to Maisel, who proposes its integration into the registry recently established by the Centers for Medicare and Medicaid Services as part of its requirements for primary-prevention ICD reimbursement. "Efforts to adapt this existing registry for the purpose of actively monitoring pacemaker and ICD performance should be undertaken and could be achieved at relatively little increased expense."

In the accompanying study of outcomes from replacement of ICDs that have been the target of safety advisories, based on 2004-2005 data from 17 centers in Canada, at least one major complication associated with replacement was observed in 5.8% of cases.

Of the 2915 patients with such "advisory devices," about 18% were replaced electively an average of 26 months after implantation, write the authors, Dr Paul A Gould and Dr Andrew D Krahn (University of Western Ontario, London). Complications developed in 8.1% of replacements within a mean of 2.7 months. Minor complications in 2.3% of cases included medically managed pain or incision infections. Major complications were more frequent.

According to Wilkoff's editorial, the report shows "the difficulty physicians and patients have in deciding when to intervene with a replacement surgery, as well as some of the consequences of device replacement. Consequently, this analysis does not help physicians answer the question of whether an ICD should be replaced in the next safety-alert situation."

Universal guidelines for management of advisory-device patients are lacking and would be "difficult" to develop, according to Gould and Krahn, because individual advisories point to unique problems. "Nevertheless, our data suggest that strategies focusing on ICD-generator replacement involve meaningful risk. These issues are amplified in primary-prevention patients in whom the risk/benefit ratio may be more difficult to assess."