In late January, something unsettling happened at the Minneapolis Heart Institute. On two successive days, patients came to the clinic after their heart defibrillators had jolted them with huge, unnecessary and painful electric shocks. One 65-year-old woman said she'd been zapped 14 times in an hour.
Doctors checked the hospital's records and discovered four similar cases had occurred in recent months. Each stemmed from a broken wire — called a lead — that tells a defibrillator when to send an electric shock to a malfunctioning heart. All six cases involved the Sprint Fidelis 6949, manufactured by Medtronic Inc., a leading medical-device maker.
Within days, the Heart Institute concluded that the Sprint Fidelis wasn't safe enough, told the company of its concerns, and stopped using the product.
Across the country, physicians at leading hospitals from Chicago's Children's Memorial Hospital to Boston's Brigham and Women's Hospital came across similar problems and some took similar steps.
But it wasn't until this month that Medtronic of Minneapolis reached the same conclusion. On Oct. 7, Medtronic President and Chief Executive Bill Hawkins convened a meeting of top executives who decided that the company should suspend sales of the Fidelis leads. In one of the biggest recalls of a medical device, it pulled all Sprint Fidelis models from the market, citing five deaths in the devices' three years on the market.
The events surrounding the Medtronic recall expose a hole in the U.S.'s medical safety system: Medical devices are regulated under different standards from those applied to prescription drugs. The Food and Drug Administration requires that almost all new medications be tested in human trials before they go on the market. But some devices, like the Sprint Fidelis leads, are subject to lighter guidelines because they are considered modifications of earlier products. The FDA, in most cases, also doesn't mandate major studies of medical devices after they've hit the market.
As a result, both the federal agency and the company were handicapped in evaluating whether a widespread public health threat was emerging.
Pieces of Information
Daniel Schultz, director of the FDA's device center, says that over the last several months, "we had pieces of information that suggested there were certain problems associated with the Fidelis lead." But, he says, "there was nothing we could point to specifically to say this is a violative product that needs to come off the market."
Sen. Charles Grassley, an Iowa Republican who has been critical of the FDA, is examining its handling of the Sprint Fidelis leads. The agency, meanwhile, says it is developing a new surveillance program that will help it independently monitor the safety of heart devices.
"Are there ways to identify problems more quickly?" Dr. Schultz asks. "I think the answer is yes." But, he adds, "if you require a clinical trial for every design change, what does that do to the ability of bringing new technologies to market?"
As the Fidelis leads mystery unfolded, the FDA relied almost solely on Medtronic's limited data. That left private cardiologists, including those at the Minneapolis Heart Institute, to ring the alarm bells, pressing both Medtronic and the FDA for action.
Throughout 2007 — and as recently as late September — Medtronic sought to reassure doctors, at times blaming the problems largely on physicians' technique. In March, a Medtronic letter to doctors said its investigation "suggests that variables within the implant procedure may contribute significantly to these fractures." In announcing that it was suspending sales of the Sprint Fidelis leads on Oct. 15, the company said they were failing at a rate of 2.3% over 30 months, more than twice as often as its other top lead.
Although the difference wasn't yet statistically significant — that is, it could still have plausibly been due to chance — the safety gap between the two types of wires was widening.
Rob Clark, a Medtronic spokesman, says the company didn't hide the problems or delay revealing them. He says Medtronic couldn't rely on data from individual hospitals and needed time to complete its statistical analysis and confirm that there truly was an excessive fracture rate with Sprint Fidelis. "Physicians disagree on the data," Mr. Clark says. "Some still think this device should stay on the market."
Defibrillators are life-saving devices that dispatch shocks to treat cardiac arrest and restore normal heart rhythm. To work properly, they depend on leads, the complex wires that connect defibrillators to the heart muscle. The wires sense aberrant heart rhythms and deliver jolts of electricity to revive a dying patient.
Fractured wires can deliver unneeded and frightening shocks — and, on occasion, can cause a lethal heart rhythm. Yet thick-diameter leads have been known to pose risks. So for years, doctors clamored for ever-narrower designs, partly because blood clots tend to form around broader ones.
Medtronic answered their call with the Sprint Fidelis, among the thinnest leads, with a diameter of 2.2 millimeters — about the thickness of a nickel. In late 2004, the device gained FDA approval and it quickly became the world's most widely used defibrillator lead.
By early this year, about 90% of new Medtronic defibrillators used Fidelis leads. Some 268,000 of the devices have been implanted in people around the world, and about 235,000 remain in patients' chests. The leads have brought in about $1 billion in revenue for Medtronic, which has annual sales of more than $12 billion.
Like other leads made by Medtronic and its competitors, the Fidelis leads occasionally broke. But the issue went largely unnoticed until those two patients walked into the Minneapolis Heart Institute's pacemaker and defibrillator clinic, in January.
In both cases, doctors at the clinic determined that the patients' Fidelis leads had fractured and misfired. It worried Linda Kallinen, the clinic's technical director. "We wondered if this was happenstance, or not," she says. Adrian K. Almquist, the doctor who treated the patients, found the cases odd because the fractures had occurred within roughly two years of implant.
Scouring electronic logs of other clinic patients, Ms. Kallinen found reports of four other Fidelis fractures in the previous seven months. She and Dr. Almquist went to Robert G. Hauser, a senior consulting cardiologist at the Heart Institute who has made a career of studying defects in heart devices.
In 2005, Dr. Hauser, 68 years old, was instrumental in triggering the recalls of more than 200,000 defibrillators and pacemakers made by Guidant Corp., now part of Boston Scientific Corp. Eight years ago, he organized other cardiologists to create a private database of failures in defibrillators, pacemakers and leads.
After hearing from Ms. Kallinen and Dr. Almquist, Dr. Hauser combed through his multi-hospital database. He found similar trends of fractures in that database as well as multiple Sprint Fidelis lead failures in a separate federal database. The Heart Institute decided to stop implanting the Fidelis leads altogether and substitute an older Medtronic lead that appeared to be safer, the Sprint Quattro.
Dr. Hauser contacted Medtronic. In February, he and several other clinic physicians met at the Heart Institute with Warren Watson, a Medtronic vice president, and an engineer. Dr. Hauser says he told Mr. Watson that Medtronic had a serious problem with its Fidelis lead. Three identical device defects at one hospital, he argues, can signify a broader problem.
Mr. Watson disagreed that there was enough evidence yet that the Fidelis lead performed worse than others, several participants in the meeting recall. Medtronic officials also suggested that lea
d fractures could have resulted from doctors' mishandling the devices, according to participants. "They were blaming the implanters," says Dr. Almquist, who says he was offended by the suggestion.
Mr. Watson and other Medtronic officials declined to be interviewed for this story. Mr. Clark, the Medtronic spokesman, wouldn't discuss specifics of the meeting, but said there were hospitals that had implanted hundreds of Fidelis leads with no fractures.
At the same meeting, Medtronic officials shared their own internal analysis of Fidelis leads that had been returned to the company. The data showed that from late 2004 through February of this year, there had been 226 fractures, for a 0.15% failure rate, in the Fidelis. That compared with 64 fractures, or a failure rate of 0.05%, in Medtronic's other defibrillator lead.
The return-rate data are imperfect, however, because few leads are ever returned to the company. (Doctors often leave faulty leads in bodies and insert new ones because removing leads risks torn veins and dangerous bleeding.)
The FDA uses its discretion to determine whether a manufacturer should perform a safety or surveillance study after a device goes on the market. In the case of Sprint Fidelis leads, an agency spokeswoman said, the FDA didn't require any such study because "no issues were raised in premarket review that suggested the need for a post-market study." Since the Fidelis was similar to an earlier design, the FDA hadn't required pre-marketing testing in human patients, either.
Agency officials also say that engineering bench tests can be more valuable than small human trials which might not identify relatively rare or longer-term problems.
As with all devices, Medtronic has to file reports of Fidelis problems, reported by doctors or others, to the FDA's safety database.
Monitoring Patients Remotely
Medtronic had begun on its own a study as the leads went on the market. By March it had data on 487 patients. The month before, a team of Medtronic analysts had begun the laborious process of tapping into CareLink, a proprietary system that allows doctors to monitor patients and their defibrillators remotely by phone or computer. Using computer files on 25,000 patients fitted with Fidelis leads, the analysts set out to measure fractures and pre-fracture conditions. They also had to contact doctors and hospitals to verify that what they were assembling matched doctors' own records.
In part due to information from Dr. Hauser, Medtronic on March 21 sent out a "dear doctor" letter saying, "Medtronic has received reports from a limited number of implanting physicians indicating they have experienced higher than expected" fracture rates. The letter cautioned doctors about how to handle the device to avoid problems.
Meanwhile, Dr. Hauser submitted a manuscript to the journal Heart Rhythm. Based on his analysis of his multihospital database as well as a federal database, Dr. Hauser concluded that "the Sprint Fidelis high-voltage lead appears to be prone to early failure." He sent an early copy of the manuscript — whose findings would be published by Heart Rhythm online in April — to the FDA and Medtronic.
The manuscript "put Sprint Fidelis on our radar," says the FDA's Dr. Schultz. Still, the implications of one bad shock — even one death — in isolation were hard to discern, FDA officials say. The agency lacked details about some incidents. Given the lack of information, it couldn't put them in context, or be sure they were all tied to a specific pattern of failure.
By this spring, doctors were reaching their own conclusions. Frank Mazzola, an electrophysiologist at Long Island Jewish Medical Center in New Hyde Park, N.Y., stopped using the Sprint Fidelis in April after seeing patients with lead fractures.
Physicians at Brigham and Women's Hospital did the same after they too saw problems. At Western Pennsylvania Hospital in Pittsburgh, cardiologist Leonard I. Ganz says he stopped using the Fidelis leads in May after two patients suffered multiple shocks. "Even though there was no statistical trend yet, I was concerned enough that it might be" in time, Dr. Ganz says.
Medtronic maintains that many hospitals using Sprint Fidelis implants weren't experiencing any problems with fractures.
On July 19, Medtronic officials met again with Dr. Hauser and other physicians at the Heart Institute. Dr. Hauser urged Medtronic to stop selling the leads. Medtronic's vice president for quality and regulatory issues, Reggie Groves, demurred, using a PowerPoint presentation to show that the incidence of fractures still wasn't statistically significant, according to people present. The company declines to elaborate.
According to the Heart Institute's Ms. Kallinen, Medtronic's Ms. Groves said the company had identified a problem and was working on a possible remedy, but had no intention of pulling the leads off the market.
The company was trying to get to the bottom of what was becoming a crisis. Medtronic says it learned about the five deaths potentially linked to Fidelis leads between August 2006 and this September. The patient study it had begun in 2004 by late July had data on 654 patients, and the separate, eight-month CareLink analysis of 25,000 patients was well under way. Using that information, Medtronic analysts by October determined that the Fidelis overall failure rate — 2.3% over 30 months on the market — was higher than the 0.9% rate for one of its Quattro models.
Medtronic consulted its outside advisory committee of heart doctors, who thought the company had to act. Just after midnight on Oct. 15, the company issued a news release saying it was withdrawing all Sprint Fidelis leads from the world-wide market. The release quoted Mr. Hawkins as saying the recall "is the right thing to do given currently available information."