How the FDA allows faulty, and sometimes dangerous, medical devices onto the market
As a busy urologist in Madison, Wis., Michael Kuglitsch saw firsthand the pain and suffering of the more than 10 million American women afflicted with incontinence. The condition was more than embarrassing; it was debilitating. So Kuglitsch was intrigued when a salesman from Boston Scientific Corp. came by with a new product back in 1997 that promised to make corrective surgery less painful and more effective. The ProteGen Sling, a synthetic material designed to support the bladder, had just been approved by the Food and Drug Administration. And if the salesman was to be believed, it worked like a charm.
Kuglitsch began implanting the device in patients. But many soon developed problems. Instead of returning to their normal lives, as expected, most of the patients began to experience extreme discomfort and infection around their implants. Three lost all their hair and fingernails, Kuglitsch said. Later investigation revealed that many of the devices were wearing through patients' vaginal or urethral walls. "It finally reached the point," Kuglitsch said, "where I was removing them every week."
What Kuglitsch didn't know, until the product was pulled from the market and the first of some 600 lawsuits was filed, was that Boston Scientific had received more than 100 similar complaints by the time he stopped implanting the device in September 1998. Internal company documents show that the manufacturer didn't tell the FDA about nearly 400 more. More shocking to Kuglitsch was that the FDA had approved the sling for incontinence surgery without requiring any new testing. In fact, the sling, like 98 percent of the 80,000 medical devices on the market today, was approved through a streamlined process that requires data to show only that a product is similar to one already in use.
Big business. Today, 1 in 10 Americans is walking around with some kind of synthetic body part. Pacemakers, defibrillators, heart valves, knee joints, and spinal-fusion screws help make the lives of many of these people less painful and more livable. They have also created a big business. Industrywide, revenues top $78 billion a year. Unsurprisingly, lots of upstart companies want a piece of the pie. In just the past five years, the number of manufacturers of medical devices has risen from just over 9,000 to more than 13,000. An aging population and new technologies are expected to fuel substantial growth for some time to come.
Amazing as these devices are, many of them have caused chronic pain and crippled patients. In rare cases, they have killed people. There are few hard statistics, but from 1990 to 1999, according to the independent medical research organization ECRI, the number of recalls of implantable devices has nearly tripled, from 35 to 104 a year, with a single recall sometimes affecting thousands of patients. The problem, critics say, is an FDA approval process that allows faulty devices to be marketed before they have been properly tested and the agency's inability to adequately monitor devices once they are in use. "If we keep this up," says Arthur Kaplan, chairman of the Department of Medical Ethics at the University of Pennsylvania, "we will end up with a device that slips through and takes a lot of lives."
Pendulum. Not long ago, it was generally agreed that the FDA took far too long to approve new drugs and devices. Following the furor over silicon breast implants in the early 1990s, the FDA tightened controls to such an extent that the approval time for new devices doubled--often to two years--and many good products were stalled. The successful Omnicarbon Heart Valve, for instance, was approved for American patients only this year--15 years after it had been approved overseas. Manufacturers also say the old FDA was too aggressive in pushing for recalls and nitpicked during inspections. "Companies were afraid to complain about warning letters received at a factory," said Thomas Thompson, who helped form the Medical Device Manufacturers Association, "because if they said something, the FDA would shut down two more plants in retaliation." Such fears caused some manufacturers to hide bad data from the FDA, he said. "There were reasons to do it."
But now, after intense congressional lobbying by manufacturers to speed up the process, the pendulum has swung too far the other way, critics say. Under the FDA Modernization Act of 1997, the agency allows many new devices to be marketed before full-scale clinical trials--a system that essentially approves products first and asks questions later.
The system can be problematic even when companies are honest and their data thorough and true. But a U.S. News examination of several recent approvals shows that the FDA is often unable to guarantee that medical devices are safe or effective. Manufacturers can exacerbate the problem. Some have tried to push new products through with as little testing as possible. Others have fudged clinical data. Still others have failed to notify the FDA, as required, when their products malfunction. "The safety net is weak," says Robert E. Baier of the University at Buffalo, "and sometimes it seems like it is actually not installed."
By law, all of the approximately 7,500 manufacturers of the highest-risk devices must be inspected once every two years. But the FDA division charged with device safety, the Center for Devices and Radiological Health, has long complained that its annual funding--about $177 million--fails to keep up with its exploding workload. In addition to reviewing some 17,000 device submissions a year and inspecting more than 15,000 manufacturers, the division also must inspect more than 29,000 mammogram facilities, radiological health firms, and clinical trial investigators.
The result? Between 1995 and 2000 the number of inspections of medical-device firms in the United States dropped by nearly half--from 3,602 to 1,841. Clinical trial sites run by these companies are inspected even less often; FDA inspectors examined only 575 of more than 15,000 clinical sites in 1999. When the FDA does inspect manufacturers and lab sites, according to the private consulting firm warningletters.com, it finds violations in 1 out of every 7 or 8 cases. Common violations include failing to assure that a device is being made correctly and failing to submit data about deaths and injuries.
Under pressure from Congress and the Bush administration, the FDA has instituted industry-friendly measures that critics say further compromise safety. One policy requires inspectors to notify manufacturers in advance of inspections. The agency has also loosened requirements to track certain high-risk devices, although it says it intends to keep tracking many. It has exempted companies from submitting many adverse-event reports after a device has been recalled or repaired. And it is considering allowing manufacturers with good records to inspect their own facilities.
Yet it is the policy of allowing companies to collect most clinical data after a device is marketed that most troubles longtime FDA observers like Jeffrey Lerner, president of ECRI. It is, he says, nothing short of "experimenting on people." The policies look even worse, he says, in light of reports that manufacturers are violating requirements to keep patient registries and to closely monitor their products. "If a company doesn't follow through, we don't have much clout to force them," one FDA official concedes.
Medical devices can be approved in one of two ways, depending on how risky or innovative they seem. Under the pre-market approval (PMA) process--generally used for new implantable devices that might do medical harm--manufacturers must conduct rigorous clinical tests before the products can be used on the public. However, the vast majority of devices are approved under the faster process known as 510(k), for the section of the law that applies. Usually under 510 (k) approval, manufacturers need only show that their new device is "substantially similar" to an existing approved product. The intent of 510 (k) was to fast-track approval for devices--from next-generation catheters to new pectoral implants--that the FDA considered less worrisome. As the ProteGen sling shows, however, the absence of fresh data can have grave consequences.
Boston Scientific is a company known for a diverse line of products, including cardiovascular catheters and stents. In 1988, seeking to expand, the company entered a number of new fields, including urology. In 1995, it bought Vesica Medical Inc., which made a special kind of anchoring system used in incontinence surgeries. The anchor was paired with a woven polyester fabric coated with bovine collagen to make a vaginal sling. The fabric had been used predominantly as a patch in cardiovascular surgeries. The company claimed--and the FDA ultimately agreed--that the intended new use was substantially similar to its previous use. It was not.
According to a lawsuit filed against the company, Boston Scientific pushed for approval of the new use despite serious safety concerns raised by its own investigators. A class action lawsuit filed last year notes that on at least two occasions, Aug. 8, 1995, and Jan. 8, 1996, doctors urged caution with the intended new use, speaking of the "erosive qualities of synthetics" and the need for "animal data . . . before we give it serious consideration." Yet two months later, according to the complaint, Boston Scientific decided to seek immediate approval. It relied on a 90-day rat study by a company that originally tested the material as a vascular graft.
When a device contains more than one part, or is tailored for use with another device, the manufacturer is expected to inform the FDA about both pieces and how they will be used together. Boston Scientific, FDA documents show, submitted data describing only the material, not the anchoring system it would be used with. Critics say such omissions deprive regulators of important information that would be likely to lead to more scrutiny.
The FDA approved the sling on Nov. 15, 1996, agreeing that the product was "substantially equivalent" to other devices used to repair hernias, lungs, and hearts. But shortly thereafter, Kuglitsch, now a witness against the company, said he told Boston Scientific that many of his patients were not healing properly. Although he repeated these concerns throughout the year, Kuglitsch says company officials told him that he was the only one having trouble. But by November of the next year, when the company started its first study of the sling's long-term biocompatibility, the company already knew of at least 125 adverse reactions, FDA data show.
There was reason to worry, according to an April 1998 internal memo obtained by U.S. News. The author, apparently a consultant hired by Boston Scientific, complained that "[t]here is general concern among [sales representatives] regarding ongoing ProteGen complications and impact on overall business/company image." Twenty-seven percent of doctors who used the product most consistently had already decided to stop, the memo said. It went on to say that some sales representatives "question ProteGen's ability to foster tissue ingrowth," quoting one as saying he watched six removals and noted that the sling was "white and clean as out of the package; tissue doesn't want to heal into it." Far from expressing worry about patients, the memo questioned whether Boston Scientific's problems might make it lose out to competition from other "hot" sling products that were "setting [the] world on fire."
By June 12, 1998, according to another internal memo, salesmen had confirmed a 3.5 percent dehiscence (failure to adhere) and erosion rate and were saying that the rate of such complications was probably as much as 10 percent. Finally, in January, the company recalled the nearly 24,000 remaining slings on the market. The FDA later said it believed the product was not safe and effective for its prescribed use.
Boston Scientific claimed little responsibility for the sling's problems in a recall letter to doctors on Jan. 28, 1999. "Since the release of ProteGen, our reported adverse events have remained below 1 percent of units shipped," the letter says. Three months later, however, Boston Scientific sent its lawyers a study of 99 patients, which found that more than 30 percent of the patients experienced vaginal erosion after only five months. Paul Donovan, a spokesman for the company, said that "[w]hile the ProteGen Sling met or exceeded FDA requirements, the use of the product didn't produce the outcomes we wanted."
Fudging the data
Even when a medical device goes through the long and grueling post-market approval process, the product is not assured of being safe. Between 1994 and 1999, according to the FDA's Office of the Inspector General, medical device trials were twice as likely as trials for drugs and biologics to violate FDA regulations. The review of 189 regulatory letters found that 107 were issued against device manufacturers and that 78 percent of them cited violations such as missing data, poor data collection, and falsification of data. The findings "cast reasonable doubt on the efficacy and reliability of the current oversight process," the review said. Arthur M. Horowitz, a respected regulatory consultant to medical companies, estimates that research fraud and misconduct occur in up to 5 percent of all trials. "It's an industrywide problem that I believe is growing," said Horowitz.
Data falsification is particularly worrisome since it can lead to deaths and serious injuries. In 1993, C. R. Bard Inc., one of the world's largest catheter producers, pleaded guilty to 391 charges of healthcare fraud for submitting false data to the FDA when it applied for approval through the lengthy PMA process. Critics claim a defect caused some of the catheters to break, leading to at least one death and 22 emergency surgeries.
The scientific community has become increasingly concerned about data corruption, noting the large financial stakes investigators often have in the products they are testing. Unlike with capital-intensive drug research, the inventor of a medical device may also be the product's lead investigator, as well as the company CEO. Even with less severe conflicts, ethical questions remain. Just before the FDA approved the ProteGen sling, for instance, members of Boston Scientific's urological advisory board submitted a paper for publication by the American Urological Association saying the sling material was "a safe, biocompatible substance." The abstract failed to mention, although such disclosure was required, that two of the research doctors had received compensation--approximately $320,000 combined--from Boston Scientific.
The growing worry is that in hopes of a big payday or out of fear of losing their jobs, researchers with vested interests will rejigger tests to get the "right" results or hide data that raise concerns about the product. But even when investigators speak up about problems, the company paying for the research does not always want to listen. At least that was what two prominent doctors say happened when they tested an implantable weight-loss device and didn't like what they found.
The gastric LAP-BAND, made by California-based Inamed Corp., has been billed as a safe and easy weight-loss device for thousands of morbidly obese Americans. The band is inserted, wrapped around the middle of the stomach, and then tightened, making an hourglass shape that causes patients to feel full sooner. But the results of this procedure did not impress two researchers who conducted U.S. clinical trials between 1996 and 1998.
More than one third of the implants, the doctors say, had to be removed because of complications, and another third of the patients didn't lose anywhere near the weight advertised or necessary. Harvey J. Sugerman, chief of general and trauma surgery at the Medical College of Virginia and one of BioEnterics' research leaders, says the company discouraged him from publishing his data and attempted to circumvent the three-year trial mandated by the FDA. The company petitioned the FDA for approval after two years--without submitting the disagreeable data, he said. The FDA panel rejected the company's PMA application only after Sugerman presented his data without invitation.
Robert McIntyre Jr., who helped run another trial of the device at the University of Colorado-Denver, said he believed the company wanted early approval so it could "make the safety profile look better, so the true complication rate would not be discovered." Patricia Choban, one of the panel of experts brought in by the FDA to evaluate the LAP-BAND, agreed. "The data at three years was only going to get yuckier," she said. Despite these objections, the FDA granted BioEnteric's PMA application the following year. The FDA did demand new post-market trials, however. So far this year, the agency has received 469 adverse-event reports. One death, 34 injuries, and 446 malfunctions have been reported since January 2001. An FDA official says the agency is "very concerned" about these reports. The company has denied doing anything improper. Spokesman Peter Nicholson said that although the data from the U.S. studies were "not strong," other data were and the adverse events span many years.
Slaps on the wrist
Presenting the FDA with misleading or incomplete data can lead to recalls, civil fines, and even prosecution. Last year, for instance, an official of catheter maker Micro Interventional Systems was sentenced to 10 months in jail for submitting false testing data, concealing adverse events, and covering up test failures. And earlier this year, Gliatech, a Beachwood, Ohio, medical device company, agreed to pay $1.2 million in fines for submitting false data and committing other violations while testing an internally applied gel used to inhibit scarring.
But in most cases, the FDA seeks no punitive action--a reluctance to enforce the law that even its own chief counsel finds baffling. "Our review," said Daniel Troy in a recent speech to the Drug Information Association, "has often caused us to ask why [FDA investigators are] proposing a letter instead of a [law]suit."
More often, when a company fails to provide crucial data to the FDA, it gets a slap on the wrist. A case in point is the troubled history of a product known as the AneuRx aortic stent. The aortic stent is manufactured by device giant Medtronic, a leader in the development of implantable devices with $5 billion in annual revenue last year. In September 1999, Medtronic and a competitor, Guidant Corp., obtained FDA approval to market a product that both had long sought--a stent that would shore up weak areas of the aortic artery. Inserted through the groin and fished up to the aneurysm site, the stents can help patients avoid open surgery. More important for the manufacturers, they have a market estimated at 3 million people worldwide. But not long after the stents were introduced, they ran into trouble. How the two manufacturers handled the problems could not have been more different.
Guidant was the first to report difficulties. In March 2001, it revealed to the FDA that its salesmen had failed to report hundreds of device malfunctions and adverse events, mainly related to doctor error. An internal audit revealed numerous other problems. Guidant fell on its sword, recalling all of the stents it had on the market. (Later, the FDA allowed the stent for limited use with experienced doctors.) The effect on the company was devastating; it lost $15 million, and its stock plunged.
While Guidant was fighting for its survival, demand for Medtronic's stent, the AneuRx, surged. But it turns out that Medtronic's stent was not necessarily any better than Guidant's--just luckier. During the approval process, Medtronic committed its own omissions of fact. The sole purpose of the stent is to prevent ruptures, and the Medtronic data showed the device ruptured only twice among the 416 patients taking part in Phase II clinical trials. In fact, Medtronic had been informed of five other ruptures during its trials, yet it failed to report the ruptures to the FDA until at least a month after the product went on the market.
"Rampant" problems. Medtronic Vice President Tom Wilder said that the FDA and Medtronic had "differences of opinion over how these should have been reported" and that the company did nothing wrong. But FDA documents reveal that Medtronic has been cited for numerous similar and repeated violations at testing sites for other devices in recent years. One December 1997 letter noted a consistent underreporting of adverse events and called data problems "rampant." FDA officials were equally pointed after they investigated the AneuRx trials. In a sharply worded letter from May 2000, the FDA described repeated "violations of FDA regulations [and] repeated deviations from the requirements of the investigational plan." The agency concluded that it had lacked the data necessary to evaluate the safety and effectiveness of the AneuRx stent.
All the while, however, the product remained on the market. Medtronic did not recall it, nor did the FDA seek any restrictions on its use. In the months that followed, Medtronic did send letters to doctors warning of "the continuing risk of aneurysm rupture." The United Kingdom's Medical Devices Agency sent out a cautionary public-health notice about the AneuRx stent in June 2000. But the FDA waited until April 27, 2001, to send a similar letter to doctors here, citing reports of approximately 25 aneurysm ruptures, as well as leaks, suture breaks, fabric tears, and migration of the device. In what appears to have been its most serious enforcement action concerning the stent, the FDA required Medtronic to continue collecting more clinical data. An FDA spokeswoman said no further restrictions were warranted because of the "low frequency" of ruptures and the seemingly "high percentage" of successful use.
Meanwhile, independent studies are showing that while aortic stents have been called less risky than open surgery, they may be worse. One recent study of the stents, including the AneuRx, found that 95 percent of those with open surgery had survived and needed no corrective surgery a year later, versus 71 percent of patients treated with stents. While the FDA approved the devices on the basis of just 12 months of data, British researchers found an "alarming" increase in failures after one year. Medtronic counters that at three years its rupture rate is only 2 percent among approximately 400 of the patients in its clinical trials.
Yet not all of Medtronic's stent patients are pleased. William Anderson, 60, a technology consultant from Houston, had the stent implanted in November 1998, hoping to return to work soon. But problems soon emerged. The stent relies on a pressure fit to keep it in place, and in Anderson's case, the fit didn't hold; the stent started migrating right after surgery. By November 2000, he says, the device was bent, broken, and leaking, with one wire puncturing his artery wall. The next month, Anderson's doctor had to remove the stent and fix his damaged artery. Later, Anderson underwent another surgery to repair a hernia along his scar line. He says he has not been able to work steadily for four years.
Medtronic, which is now fending off numerous lawsuits related to the stent, including one filed by Anderson, contends that doctor error, among other problems, is often the cause of the device's failure. Experts agree that this often is the case. At an FDA panel meeting in July 2001, one of the original researchers, Kim Hodgson, president-elect of the Society of Clinical Vascular Surgery, said that physicians who did not take part in clinical trials were unlikely to have the skills to succeed with the stent. He said that the AneuRx was "far from a mature technology" and placed blame not just with doctors but also with manufacturers and the FDA.
Whatever the case, Medtronic argues--and courts have sometimes agreed--that the FDA's approval of the stent shields the company from liability for the device's failure. It is an argument that serves to only further outrage injured patients. "This has just stopped my life," says Anderson. "I don't know how they snuck this thing past the FDA; I just can't imagine how this got shoved through."
BETTER THAN SURGERY? The AneuRx aortic stent, shown shrinking the balloon aneurysm at left, is billed as less risky than open surgery. But the FDA approved the device despite serious questions about its clinical data and the ability of doctors to properly use it. The manufacturer, device giant Medtronic, is fighting numerous lawsuits related to the stent, including one filed by William Anderson (right). Anderson had the stent implanted in 1998. The device broke, damaging his artery. He says that he has not been able to work steadily for four years.
This story appears in the July 29, 2002 print edition of U.S. News & World Report.