The Danger Within Us

by Jeanne Lenzer

  Hoffman has pounded away for years at the ways in which industry funding of clinical trials and individual doctors has undermined good medicine. He rattles off examples. There was the study of calcium-channel blockers showing that researchers who received manufacturer funding were more likely to interpret results as positive than researchers who didn’t have manufacturer funding. In another case, researchers employed by the manufacturer of dioxin were less likely than independent researchers to identify cancer in the slides of rat livers.

  Hoffman quotes the famous comment by author and political activist Upton Sinclair: “It is difficult to get a man to understand something when his salary depends upon his not understanding it.” “But that is our responsibility,” says Hoffman. “Our society rewards doctors with enormous prestige and excellent income, not to mention almost unchallenged professional autonomy. In return, our patients expect us to put their interests above our own. We cannot take money from industry with one hand and write prescriptions with the other and remain a profession.”

  Chapter Nine

  The Quest for Truth

  ON APRIL 15, 2009, I was writing for The BMJ when I received the first of what would become many e-mails from Dennis Fegan. I was just one of the “dozens and dozens” of journalists whom Fegan had reached out to.

  It took a while for me to look into his story and to reach a decision about whether the claims he was making were credible and whether his was a story I wanted to report. First, I wanted to get all his medical records. In addition, I wanted to contact the doctors who had treated him. Then I wanted Cyberonics’ account of the matter, and I wanted the FDA’s version.

  I recall clearly the day that I received copies of Fegan’s medical record. I read his ER report, and then I saw the cardiograms—electrical tracings of his heart, which gave a compelling picture of what had happened to him. In nearly thirty years of clinical practice, working in ERs and intensive care units, I’d never seen anything like this: There was a lengthy flat line, far longer than a mere pause or a few missed beats. It was the sort of flat line that one sees only as patients are dying—or already dead. I could see why Larry Johnson, the ER doctor who tried desperately to save him, remembered Fegan above all other patients in his more than thirty-year career.

  Dates would prove important in what happened next. In February of 2010, I asked Cyberonics to send me its records on Fegan. The company said it couldn’t discuss his case without his permission. On March 4, 2010, I sent Cyberonics a copy of Fegan’s written release, authorizing me to discuss his case. Then a surprising thing happened: six days later, on March 10, 2010, the FDA received a report about Fegan’s hospitalization—a report dated nearly four years earlier, July 3, 2006. Why the FDA didn’t receive the report until then is not clear. Cyberonics insists it sent the report in 2006, but the FDA marked the report as received on March 10, 2010, the same day it finally appeared in the MAUDE database.

  The report itself contained a subtle clue about its likely actual date. The company listed the date of Fegan’s hospitalization as “May 2006.” While the medical record clearly gave the actual date of Fegan’s hospitalization as July 2, 2006, there was one place that also (erroneously) listed the date as “May 2006,” and that was a report filed by Dr. Juan Bahamon, Fegan’s neurologist. This suggests that Cyberonics may have copied the date of the event from Bahamon’s report, which would give the lie to the company’s claim that it submitted its report on July 3, 2006, because Bahamon didn’t file his report until November of that year.

  Furthermore, both the report Fegan eventually filed on himself and the report filed by Bahamon were dated as received the same day by the FDA. The four-year delay from the time Fegan was hospitalized to the time the FDA received Cyberonics’ report was in stark contrast to reports sent to the FDA by Fegan and Bahamon, which were marked as received the same day they were sent.

  Fegan told me there was something else suspicious about their claim that it took the FDA four years to suddenly log the report by Cyberonics: if Cyberonics really had written its report the day after he was hospitalized, why did the company representative, Steven Parnis, demur when he was asked, a month after Fegan’s hospitalization, whether the company intended to file a report with the FDA?

  But the central point was undeniable: as Cyberonics acknowledged in its written report to the FDA, “both the cardiologist and the neurologist believe that the reported events were caused by the stimulation of the vagus nerve.”

  Cyberonics, however, wasn’t about to let the conclusion rest at that. Instead, the company went on, “Inderal had recently been added to the patient’s medication regimen.” This point was supposed to be an evidentiary bombshell. Cyberonics was suggesting that Inderal was the crucial factor in Fegan’s asystole. The company spelled this out in an “issue report” dated August 23, 2006, about the “root cause” of Fegan’s asystole. The report concluded, “At this time the exact cause of the cardiac related event is unknown. It is possible for the cardiac events to be related to the initial [sic] of Inderal.”

  About two months before he was hospitalized, Fegan had indeed been started on a low dose of Inderal, a drug that can slow the heart rate. But it’s highly unlikely that Inderal was the cause of Fegan’s asystole. Inderal belongs to the class of drugs known as beta-blockers, which can reduce the heart rate by slowing passage of the electrical impulses between the atria (the top chambers of the heart) and the ventricles (the lower chambers). When this effect occurs, it causes readily detectable and very specific changes in the cardiogram. When Fegan’s heart was beating, it was entirely normal. There was no slowing. There were no changes in the passage of impulses between the atria and the ventricles.

  Furthermore, Fegan had been on the medicine for two months, which was plenty of time for his blood levels to stabilize and for signs of any problems to emerge. None did. As his neurologist indicated, Fegan began telling his doctors about his new symptoms before he was started on Inderal, suggesting that the VNS device alone was indeed causing the problem. Nor was it possible to explain how Inderal would cause his heart to stop at exactly the same three-minute intervals that the VNS device fired or why his problem ceased immediately when the device was turned off—and while Inderal was still in his system.

  For all these reasons, the link between Inderal and Fegan’s asystole seems tenuous at best.

  The Inderal defense was especially curious: if the company believed that beta-blockers like Inderal could cause a potentially fatal reaction when used with a VNS device, why didn’t it issue any warnings to the public indicating that taking beta-blockers after being implanted with the device might cause asystole? Stephen C. Karceski, director of clinical trials at the Weill Cornell Epilepsy Center, gives talks (paid for by Cyberonics) to patients and doctors about the VNS device, and when asked if he cautions patients about using beta-blockers after implantation, he replied, “No. Of course not.”

  Warnings are a double-edged sword. If companies make cautions widely known, patients and their doctors might be reluctant to use a device. On the other hand, if a warning exists, the company is off the hook for urgent adverse-event reporting and, perhaps even more important, in the case of devices, it may be protected from lawsuits. Nothing served to immunize cigarette manufacturers against lawsuits better than the warning on cigarette packages about the serious health problems that can be caused by smoking.

  At the 2017 BIOMEDevice conference and exposition in Boston, Michael Drues, PhD, president of Vascular Sciences, gave a lively presentation to members of the medical device industry on “how to address risks you’d rather not draw attention to in documentation.” Drues is a heavyweight in the field: he is a consultant to big-name device and pharmaceuticals companies as well as to the FDA, Health Canada, US and European patent offices, and governmental agencies around the world. The audience listened intently as he described various ways to deemphasize the unpleasant side effects of medical devices. He started by telling them, “Not many people can pla
y on both sides of the fence,” but of course it is precisely his ability to navigate both sides of the fence that makes him a sought-after expert. He told the audience that presenting device risks in order of severity or frequency can draw attention to risks the manufacturer might not want to emphasize and suggested that because there is no rule about how to present risks in a submission to the FDA they should instead list them in random order—and by listing a lot of minor side effects, they could effectively bury the worrisome effects in a lengthy list.

  By issuing cautions but making them easy to overlook and difficult to interpret, manufacturers have the best of both worlds. The cautions get them off the hook—much as the warnings on cigarette packages now immunize tobacco companies against lawsuits over lung cancer—and because few people know of the caution, sales aren’t harmed.

  Of course, Drues counseled conference attendees, some things are better left unsaid. He quoted Massachusetts politician Martin Lomasney, saying, “Never write when you can speak; never speak when you can nod; never nod when you can wink.”

  Cyberonics handled the issue of beta-blockers and “heart stoppages” by simply staying mum. The company all but ensured that no one would know about the issue—not patients, not doctors, not even its own paid experts. There is no warning in any of the physician or patient manuals indicating that beta-blockers in combination with a VNS device might cause asystole and, potentially, death. That would be a bad marketing move, because beta-blockers are so widely prescribed that it would certainly cut the company’s profit margins. But what the absence of information about beta-blockers more realistically reflects is that Cyberonics never believed that the Inderal Fegan was taking was the cause of or significant contributor to his “heart stoppages.”

  Cyberonics’ failure to identify the VNS device as the cause of Dennis Fegan’s nearly fatal asystole was of a pattern with its stance on the safety of its VNS device. On March 15, 2010, Cyberonics’ vice president of clinical affairs and quality, Bryan Olin, a statistician, wrote in an e-mail, “We do not believe, nor does the data support, that there is any evidence suggesting that it [the VNS device] causes death.” Earlier that month he said, “We have received no reports of post-operative asystole related to the death of a VNS Therapy patient.” If that were true, then Fegan’s life-threatening experience should have triggered an urgent report to the FDA of a new and potentially deadly problem. That might have led to a warning to doctors and patients. But none of the several dozen doctors I spoke with was aware of any problem involving asystole with the VNS device.

  Similarly, the issue of increased seizures caused by the device was neatly obscured by the company. In an e-mail dated March 6, 2009, Joseph Tartal, the FDA’s technical branch chief, responded to Fegan’s query about why increased seizures weren’t listed as a complication of the VNS device. Tartal told Fegan that companies are exempt from urgent reporting requirements if the problem is “well known to the medical community.” But it was far from well known.

  This was the same exemption the company claimed about asystole—asserting that the side effect was indeed well known. As evidence, the company pointed to the physician manual for the VNS device, which states that during implantation and testing of the leads, “infrequent incidents of bradycardia [slowed heart rate] and/or asystole have occurred.”247 But the company also insisted that no serious problems occur once patients are implanted other than allowing that bradycardia can occur “among patients with certain underlying cardiac arrhythmias.” In other words, unless the patient already has heart problems, no worries, a point it underscored in its 171-page physician manual.248 The Cyberonics patient manual simply lists “heart rate and rhythm change” as potential side effects. But the broad and rather innocuous phrase “heart rate and rhythm change” could mean anything and is a far cry from an emphatic warning that patients might experience heart stoppages and possibly die. Nor could such a low-key statement buried in a 171-page manual ensure that the information was well known by patients or their doctors. It certainly wasn’t on the front page of the company’s website or in any of its promotional materials.

  There was something else that undermined the company’s claim that the problem was well known. I interviewed David Newman, ER physician and then director of clinical research at Mount Sinai Hospital in New York City, about the VNS device, and he was unaware of any training for ER doctors conducted by Cyberonics to educate them about the device or its effects on heart rhythm. So he called a Cyberonics representative to ask how frequently asystole occurred with the VNS device, and the representative said he’d never heard of such a case. He did say he’d “look into it.” He never did get back to Newman.

  An informal poll (a questionnaire I e-mailed to a group of emergency-room doctors across the country asking whether they had ever been given training by Cyberonics or educated about the VNS device) led to just over a dozen responses. None of the doctors said he or she had been approached by Cyberonics or been given information about the device—nor did any of them have the equipment necessary to turn off the device in an emergency. This is in contrast to manufacturers of pacemakers and defibrillators, which have provided toll-free phone numbers and training for ER doctors so they can handle emergencies.

  Had Cyberonics spent time doing the same, it might have meant the difference between life and death for thirty-six-year-old Shelly Rae Wilhite, who called her sister, Bridgette Coleman, on a Friday evening to tell her that her VNS device was suddenly giving her shocks that were so painful she fell to her knees.249 She said she’d see her doctor Monday morning, because only her doctor could adjust the device. Doctors in the ER didn’t have the equipment or the know-how to handle the problem. But Shelly didn’t make it to Monday. On Sunday evening, September 13, 2010, Shelly’s daughter found her dead on the bathroom floor.

  The company so frequently failed to include critical information in its reports to the FDA that the agency finally sent Cyberonics a letter on November 13, 2008, stating, “Many death reports do not include sufficient information to determine a cause of death.…Please make an effort as much as possible to obtain event information…to determine the relationship of the device to the event.”

  Unfortunately, nothing changed after this scolding from the FDA. The MAUDE database is filled with reports—like this one, dated July 7, 2015—providing virtually no information: “It was reported that the vns patient passed away on [date redacted] 2015. The cause of death and its relationship to vns are unknown. No further information relevant to the patient’s death has been received to date.”

  Although it’s impossible to conclude definitively from such reports what caused a patient’s death, we know from Fegan’s case that the device can cause asystole.

  * * *

  By the time Dennis Fegan solicited my help with his case, the already thin rope he was hanging his hopes on was rapidly fraying.

  The Texas Department of State Health Services had closed its investigation into Cyberonics on August 26, 2009, with no finding against the company. Since the state didn’t have the authority to investigate the cause of Fegan’s asystole, the only thing it could do was determine whether Cyberonics had followed proper procedures, filed papers, and filled out required forms, including reporting Fegan’s asystole to the FDA. *

  Politicians had proved unable or unwilling to help Dennis Fegan obtain justice. The media didn’t do much better. My own article “Watching Over the Medical Device Industry,” published in The BMJ on June 23, 2009, included a discussion of device regulation and problems with the VNS device.230 But it failed to prompt any action. I followed up with Congressman Ortiz’s office, who at Fegan’s request had asked the FDA to investigate his case, but Ortiz’s office told me I’d have to contact the FDA. His office would release no information.

  When I contacted the FDA, a representative initially told me the agency could neither confirm nor deny whether it had conducted an investigation. If I wanted anything more, I would have to file a Freedom of Informati
on Act (FOIA) request. Furthermore, if there was an investigation (which, the agency reiterated, it was neither confirming nor denying), information about it would only be available via FOIA once the investigation was closed.

  After I filed additional requests, the agency told me that it had “no information responsive to my request.”250 After even more pestering from me, the FDA told me the case was “closed via telephone.” I requested notes from the phone call between the FDA and Congressman Ortiz, and the agency told me it had no notes and that the Office of Legislation at the FDA said it was not required to keep notes.251 Finally, after more phone calls and e-mails, the agency sent me a single-page document indicating that it did, indeed, receive communication from Congressman Ortiz, but virtually everything on the page was redacted. When I said I would file a FOIA request to unredact the page, the agency reversed itself. Suddenly the case was not closed (which would have made the information open to a FOIA request). Instead, it said, the case was still open, protecting any information from release under the FOIA.

  The claim that an investigation is still open eight years after Ortiz’s letter is strange at best. But what it adds up to is simple: eleven years after Fegan’s experience, any information about an FDA investigation—or an answer to the question of whether there ever was one—is locked up so tight that even my FOIA requests have failed to unearth it.

  * * *

  When I began researching Dennis Fegan’s case and the overall track record of the VNS device, I was startled to see so many deaths among relatively young patients during the twelve years the device had been on the market. I asked Cyberonics for the postapproval study data proving the device’s safety, required by the FDA as part of its conditional approval. Conditional approval meant that if the company didn’t prove the device was safe, the FDA could order it off the market. Cyberonics forwarded five studies to me.252–256