Factory recalls are a nuisance
at the best of times. But when the product is an implanted
medical device, they pose a deadly dilemma. Removing
an implanted device involves surgical risks, and these
risks may outweigh the danger of the device failing.
In short, the well-meaning product recall warning may
end up doing more harm than good. That has apparently
been the case with implantable cardioverter-defibrillators
(ICDs) in Canada, according to new research in the Journal
of the American Medical Association (JAMA).
Cardiologist Dr Paul Gould of the
University of Western Ontario and Dr Andrew Krahn of
the London Health Sciences Centre reviewed data from
17 ICD implanting centres around the country and found
that the risks of dangerous surgical complications during
replacement were often much higher than the risk of
device failure.
These ICDs aren't replaced because
they are broken, but because somebody else's has failed,
resulting in a product advisory. These advisories generally
try to quantify the risk of failure, and most hover
around the 0.1% mark. But the risk of complications
in removing them was 8.1%, the researchers found, and
the risk of major complications requiring reoperation
was 5.8%.
Of 2,915 patients at these centres
whose ICDs were subject to advisories, 553 had them
replaced. Thirty-one suffered major complications, and
two died. But the failure rate in these "advisory" ICDs
was just 0.1%, and these failures, three in number,
produced no clinical effect. It doesn't require much
calculation to see that the principle of "first, do
no harm" is not being served by this approach.
SHOCKING
DEFENCE
None of this is an indictment of the ICD, say the authors.
Since its invention in 1980, the implantable device
has saved thousands of lives in this country. Indeed,
45% of the patients in this study were believed to have
already benefited from at least one life-saving shock.
But ICDs are vastly more complex
than the humble pacemaker, and they get more complex
each year. In fact, another study in JAMA has just found
that the malfunction rate in ICDs is five times as high
per implant as it is in pacemakers. Even that underestimates
their relative unreliability, because the average pacemaker
in that study has been in place four times longer than
the average ICD. So the annual failure rate in ICDs
is a whopping 20 times higher.
That study found an alarming decline
in the reliability of ICDs during the period 1998-2002,
though there have been improvements since. Lead author
Dr William Maisel of Harvard Medical School told NRM he suspected that the drive for miniaturization, at
a time when new features were being added, was a key
factor. Dr Krahn agreed that progress brought "issues
related to reliability".
"There's a huge increase in biventricular
pacing, which involves three and even four leads in
patients. There's just much more to go wrong," he said.
But the threat remains modest.
Of 415,780 ICDs implanted in the US from 1990 to 2002,
31 had malfunctions that killed their hosts. "Most of
these malfunctions involved the ICD failing to do its
job," said Dr Maisel. "A couple involved the device
actually triggering a fatal cardiac event."
Dr Krahn illustrates the balance
of risk. "Years ago there was a device that had a J-shaped
lead which went into a corner of the heart, and it was
held in that shape by a weld," he said. "A problem developed
where the weld would pop off, and the lead would straighten
out, tearing a hole in the heart, and some people died
from that. This was rare, but you could prevent it by
removing that lead. The irony was that, over time, about
two or three patients died from the lead, and 25 died
from the removal. Our study is history revisited, in
the sense that we're causing more trouble by removing
them than by waiting to see if there are problems."
Guidelines on replacing "advisory"
ICDs, said Dr Krahn, must balance benefit against this
higher-than-expected rate of complications. "I think
with the information we've got now, if the risk of leaving
it in exceeds 2%, then it should be replaced."
But of six current Canadian device
advisories included in the study, only one device had
a failure rate over 2%. "If you ask me," said Dr Maisel,
"we've probably been replacing more than we needed to.
But we did that in the absence of information. We didn't
know the downside. We did what we thought best."
|
