Lost Medtronic data showed serious risks with Infuse surgical implant

A study that disappeared into Medtronic's archives for more than five years contained evidence of potentially deadly problems with the use of the company's Infuse bone-growth product during neck surgery.

A review of more than 1,000 injury reports involving Infuse that regulators have released to the public for the first time found more than 100 cases with post-surgical problems in the neck, including multiple cases of throat swelling that could close airways or damage nerves.

The Star Tribune first reported the existence of Medtronic's lost study in April 2016, but the U.S. Food and Drug Administration kept details confidential at the time. Since then, the Star Tribune obtained the injury reports through a Freedom of Information Act request.

Dr. Eugene Carragee, a Stanford spine surgeon and journal editor who has been critical of Medtronic, said he was shocked by what he saw — especially in the neck surgeries, which have been shown to be riskier than other uses of Infuse.

Asked to review the data for the Star Tribune, Carragee said the frequency and gravity of the problems turned up in Medtronic's in-house study of real-world data should have immediately set off alarms. Nerve damage, excessive bone growth and breathing trouble can all be life-threatening, especially when the complications are overlapping, as they are in many of the cases in the database.

"These are serious complications that should [have been] spread throughout the [medical and patient] community," Carragee said.

Medtronic officials have acknowledged that they should have given the data to the FDA much sooner. But they said that the project didn't turn up any problems that weren't known from other reports and that the study had an unconventional design that limited the data's value.

Medtronic also pointed out that Carragee, while generally an Infuse critic, wrote in 2012 that the product should not be banned in neck surgery because there are special cases in which using Infuse could be less risky than doing the procedure without it.

"Your analysis, which solely analyzes adverse event reports in order to reach conclusions about problems associated with the device, is fundamentally incomplete, flawed and inappropriate," the company said last week in a written response to the Star Tribune's findings.

The company also says it did work with the FDA to notify doctors about risks from using the novel biotech product near airways and cervical nerves in patients' necks. For example, it was included in the fine print on the product label, and it was the subject of a 2004 "Dear Doctor" letter that said an investigation into "occasional complaints" was ongoing.

"The 'potential for ectopic, heterotopic or undesirable exuberant bone formation' with Infuse has always been included in the IFU [instructions for use]," Medtronic spokesman Eric Epperson said via e-mail. "Beginning in August 2002, the IFU warned that, 'The safety and effectiveness of the use of the Infuse Bone Graft … at locations other than the lower lumbar spine … have not been established.' "

But doctors can legally use approved products any way they think will help a patient, including "off-label" uses the FDA hasn't specifically reviewed. They continue to use Infuse in neck surgery today — including in children — even though the optimal doses and relative risks remain up for debate, according to recent peer-reviewed journal articles.

Infuse neck surgery remains "a source of controversy" in the medical community, researchers reported in the October 2017 Journal of Clinical Neuroscience. "Further investigation is warranted to determine if clinical gains outweigh the potential harms of … use in this setting."

Infuse contains a potent growth factor called rhBMP-2 that causes bone formation. It's often abbreviated as "BMP" in medical records, which stands for bone morphogenetic protein. One of the product's key benefits is that it avoids the pain, expense and surgical time of doing a second surgery on a patient to harvest bone-graft material, which is needed to fuse vertebrae and discs in the lower back. The FDA approved that use in 2002.

But doctors have been using rhBMP-2 in the upper spine since at least September 1999, one early article shows. Although that first study documented no device-related complications, other researchers soon discovered problems.

One group of doctors from the University of Louisville looked at the medical records of 151 patients who had had Infuse used in neck surgery in 2003 and 2004, finding that higher doses of Infuse seemed to lead to more problems.

At the University of Iowa, another group of doctors reviewed medical records for 69 people who had Infuse neck surgeries from 2002 to 2004 and found they were 10 times more likely than a control group to have clinically significant neck swelling, their 2006 paper concluded.

Some patients had breathing tubes inserted to stop their airways from closing. And swelling associated with Infuse surgeries tended to happen four days after the procedure — often after the patients were back home.

Medtronic was also privately hearing from doctors in the field about problems, like golf-ball sized swelling in the neck of one patient. Swelling in the cervical spine was the "No. 1 complaint" about Infuse, according to a private June 14, 2004 e-mail from Medtronic clinical and regulatory executive Richard Treharne that later became public as part of a Senate investigation.

Sometime in 2006, Medtronic began a large-scale review of real-world use of Infuse. It was an unusual study in that it was "retrospective," like the Iowa and Kentucky research, which meant that people underwent surgeries first and later gave Medtronic permission to look at their medical records.

The review included surgeries at 26 hospitals and clinics across the country, including at least one site in Minneapolis that did lower-back surgeries. It was the most wide-ranging review of Infuse outcomes anyone had done up to that point.

In at least 1,039 of the surgeries there were adverse events, as the FDA calls any problems a patient encounters after a procedure, or situations later classified as device "malfunctions."

Medtronic executives say the research team shut down the study in early 2008, earlier than planned, and someone on staff misplaced the data in the internal archives.

Because the data went missing, Medtronic didn't hand them over six months later, when the company received a subpoena for Infuse data from federal prosecutors in Boston in late 2008. When the documents resurfaced in 2013, the company says it promptly reported the information to the FDA, as required by law. The data, once rediscovered, didn't reveal new types or numbers of adverse events that weren't already reflected in the medical literature by then.

"While we've publicly acknowledged that the [study] was not properly archived when it was discontinued back in 2008, our employees care deeply about their commitment to safety, quality and improving the lives of patients," Medtronic's Epperson said via e-mail.

Medtronic's project examined some 3,647 surgeries that took place between 2002 and 2006, the vast majority of which were off-label. The database of complications, which Medtronic submitted in response to an FDA request for a summary of "unreported serious injury events," included 1,923 individual complications among the 1,039 surgeries.

Half of the surgeries at issue had two or more complications; one case had 10. Each surgery is listed as a single "adverse event," per FDA rules.

Looking at the neck surgeries, the most common complication documented in the review of patient records was "neurological deficit/dysfunction" — defined as abnormal brain function resulting from injury to the nervous system, including injury from lack of oxygen or damage to nerves.

Stanford spine surgeon Carragee, who edits the peer-reviewed Spine Journal, said many of the patients in the database had "constellations" of throat problems classified as edema, hematoma or simply swelling.

Among the 118 Infuse neck surgeries in the Medtronic injury database were 39 in which patients developed a severe swallowing problem called dysphagia. The database also showed more than a dozen reports of swelling or hematoma, and 43 reports of neurological dysfunction.

The following year, in July 2008, the FDA published a warning about life-threatening complications associated with using BMP in neck surgery. The reports included at least 38 cases where unexpected swelling compressed airways or nerves in the neck. The reported problems included difficulty swallowing, breathing and speaking.

Until the warning went out, the number of Infuse neck surgeries had been growing swiftly. Federal data show that by 2008, when Medicare was paying hospitals roughly $22,000 per Infuse procedure, about 1,500 people a month were getting Infuse to fuse the spine bones in their necks. Medtronic had annual Infuse sales of more than $800 million by then, securities filings say.

The FDA warning in 2008 immediately flattened Infuse sales growth. "We believe growth was negatively impacted by physician and payor response to an FDA public health notice regarding use of bone morphogenetic protein (BMP) in cervical procedures and the overall regulatory scrutiny of off-label use in medical devices," Medtronic told investors at the time.

By 2015, revenue from BMP had dropped to about $500 million, according to estimates by Kevin Gainer, an analyst with the Massachusetts technology research firm BCC Research.

The first comprehensive national review of real-world data on Infuse and other BMP products, published in July 2009 in the Journal of the American Medical Association (JAMA), concluded that adding BMP to cervical-fusion procedures seemed to create more risks than adding it to other off-label surgeries.

A 2010 report in the peer-reviewed Laryngoscope journal would show that Infuse neck surgery was associated with higher costs and greater mortality.

"BMP induces a robust inflammatory reaction that is likely responsible for both the success in promoting bone formation as well as the adverse effects of soft-tissue swelling," the researchers wrote in Laryngoscope. "It is not surprising that use of BMP in the cervical spine has resulted in catastrophic consequences that have been identified recently through case reports."

The 90-day death rate in the Laryngoscope report was more than double the control group, 4.2 percent vs. 1.7 percent. But Medtronic said in a report to the FDA that the article never suggested that the treating physicians felt that Infuse caused or contributed to the 11 deaths in the study population of 260 at Henry Ford Hospital in Michigan.

Experiments and research continue, 19 years after the first reported study patient received Infuse in the neck.

In October 2017, doctors in the Chicago area examined 400 patients and found that Infuse in cervical fusion reduced patients' average time in the hospital, but increased rates of swallowing problems.

Indiana surgeon Dr. Jim Rickert, founder of the Society for Patient Centered Orthopedics, said Medtronic's handling of its Infuse injury reports highlights the importance of immediately sharing all injury data, especially when the uses are not approved by the FDA.

"I don't use things off-label because oftentimes there is not an evidence base strong enough to do that," Rickert said. "Having the most data about something used off-label is really helpful."

Staff writer MaryJo Webster contributed to this report.

Joe Carlson • 612-673-4779

Jim Spencer • 202-662-7432