Medical Device Daily Washington Editor

The Annals of Emergency Medicine recently published an article that examines data from the Manufacturer and User Facility Device Experience (MAUDE) database at FDA to examine the types of problems experienced with the use of automated external defibrillators, but the analysis lacks information on device maintenance. On the other hand, one of the authors of the article told Medical Device Daily that adverse events are most likely under-reported, and suggested that manufacturers take more steps to enable AEDs to alert owners of the units when the batteries are unlikely to retain sufficient charge to defibrillate a patient.

FDA said late last year that it was looking into the number of adverse events associated with AEDs, which are applied to patients at risk of sudden cardiac death, stating that the agency was in possession of more than 28,000 medical device reports in the five years prior to the November conference call (Medical Device Daily, Nov. 16, 2010).

Jeff Shuren, MD, director of the Center for Devices and Radiological Health at FDA, said on the Nov. 15, 2010, conference call that it was “difficult to assign a cause of death“ in most instances, but he nonetheless said that the number of reported adverse events “has been increasing each year.“ Shuren made note of recalls announced prior to the press briefing, “most [of which] are related to design issues,“ he said, claiming that manufacturers “have engaged in a fixed-on-failed strategy“ consisting of “repairing individual devices rather than communicating the problem to all users.“

Both Cardiac Sciences (CS; Bothell, Washington) and Zoll Medical (Chelmsford, Massachusetts) have recalled thousands of units over the past few years, with CS announcing last year that it would replace roughly 24,000 AEDs in the U.S. due to concerns expressed by FDA over what the agency said was a failure of resistors in the CS units. In 2009, Zoll engaged in a global correction of about 180,000 of its AED Plus units (MDD, April 6, 2009).

At the time of the November 2010 conference call, Shuren announced that the agency would convene an advisory committee to inquire as to whether AEDs should be subject to PMA approval rather than 510(k) clearance, but the panel was unwilling to recommend a full-blown PMA route replete with randomized clinical trials, citing among other factors the cost of such an approach and the delays incurred in replacing units currently in place. Zoll took in an April 22 warning letter from FDA addressing what the agency said were conditions conducive to an unanticipated level of battery passivation.

The article in the Annals is based on data downloaded from FDA covering the period June 1993-October 2008, and the authors state that the MAUDE data show a total of 40,787 events associated with AED usage over that time, “including 1,284 events in which a fatality occurred,“ which comes to roughly 3% of all adverse events. Of those 1,284 events, 134 were jettisoned from the analysis after they were adjudicated as “not a failed defibrillation attempt, leaving 1,150 adverse events in the final analysis.“

One of the charts in the article indicates that the total annual number of adverse events tailed off after 2005. Figure 3 in the article lists the number of instances of unexpected shut-downs and instances of low battery messages, but the third line labeled “all reported events“ peaks at roughly 120 in 2001, dips to less than 80 five years later, and drops to less than 50 in 2008.

The authors of the article, a group that includes Lawrence DeLuca, Jr., MD, of the University of Arizona Department of Emergency Medicine (Tucson, Arizona), write that “the bulk of the failures,“ or 45%, “occurred during the attempt to charge and deliver a recommended shock,“ while “rhythm analysis discrepancies“ between the operator and the machine accounted for 13%. Another 22% of these incidents were chalked up to a failure to complete an analysis of the patient's rhythms, while “nearly one in six reports“ were classified as “other.“ This last category captures incidents in which the AED either displayed an error message the user could not interpret or instances in which the authors were unable to draw a conclusion based on the MAUDE data.

One of the interesting findings in the paper is that despite all the consternation over batteries, defibrillator leads were blamed for a statistically identical number of device failures, 23.7% compared to the 23.2% blamed on battery issues. Fifty-four of the incidents involving battery problems were described in the Annals article as being associated with an indication of low battery reserve, while another 110 incidents were associated with units that “powered off unexpectedly.“ The authors also state that “only 17 of the adverse events reports included documentation of any sort of device maintenance program or schedule.“

DeLuca told Medical Device Daily that the data available from MAUDE was inconsistent and hence “we didn't do any analysis on“ the age of the defibrillators. When asked whether the lack of information on how many AEDs are available for use in the U.S. – and how many times an AED had been used over that period – meant that the analysis lacks a denominator, he said, “that's an excellent question. I don't think anyone has a true answer for that.“

“We do recognize that this is a relatively rare experience,“ DeLuca said of AED failures, but he expressed a concern about under-reporting. “The quality of supervision of public access to AEDs is highly variable,“ and without local supervision, “it could be that there's under-reporting.“

“There's always a matter of doubt about what role operator error or maintenance might have played“ in an adverse event, DeLuca acknowledged, mentioning an incident in which the backup battery was dead and the review of the event disclosed that the charger had come unplugged. “That's very, very bizarre,“ he said, and an atypical finding.

DeLuca pointed out that the discharge behavior of lithium batteries is such that the ordinary operation of the device gives little warning of an impending collapse of charge unless the device in question is set up with software designed to account for such discharge characteristics. “You have full power output, and then you have nothing,“ he remarked of lithium batteries. On the other hand, he said his iPod has a lot of software to tackle such problems, and gives a reading of remaining battery power that accounts for the behavior of such batteries. “I should be able to do that with my AEDs,“ he said, adding that an AED could be equipped with a USB port to allow it to be connected to a computer for download of event data. “And it should have a voice recorder or a webcam“ for further data collection, he said.

“The machine should be able to routinely upload the information to the manufacturer,“ DeLuca said, adding, “if these machines talk to FDA and the manufacturer every time they're used,“ a lot of data would come forth and clarify the issues. Still, he acknowledged, “it's amazing that these things work as well as they do the majority of the time,“ but “like all [med-]tech devices, they fail sometimes, and when they do, the results are usually catastrophic.“

DeLuca said that given all the variables, it's difficult to apply the culprit label to any one party. “This is a classic system failure. There are no bad players, just a system that isn't getting the information packaged in a way that is useful“ to FDA and industry, he said.

A spokesperson for the Metropolitan Washington Airport Authority (MWAA; Sterling, Virginia), which operates both Dulles (Sterling, Virginia) and Reagan National (Alexandria, Virginia) airports, told Medical Device Daily that the AEDs installed at both airports are checked twice annually. Kimberly Gibbs, an assistant media relations manager for MWAA, said Reagan airport has 66 units and Dulles another 129, adding that the units are spaced 197 feet apart per FDA's recommendation that units be placed no farther apart than two minutes walking distance.

Gibbs said the units placed at the two airports “will send information via Bluetooth“ when the software detects a potential problem, adding “in 2009, we replaced all our defibrillators and bought another 30.“

“We use Zoll [AEDs] and they have a five-year warranty on the battery and device,“ Gibbs said, although she also mentioned that FDA “recommends the devices be replaced every seven years.“

Mark McCarty, 703-268-5690