Side Effects: A Prosecutor, a Whistleblower, and a Bestselling Antidepressant on Trial

by Alison Bass

  How could a prestigious journal like the JAACAP have published a study with such glaring discrepancies between the actual data and its conclusions? At the very least, Firestein thought, there were some serious methodological issues with this study.

  The next thing Firestein did was google Martin Keller, the lead author of the Paxil study. What she found shocked her to her core. Up popped the series of articles published a few years back in the Boston Globe that essentially cast doubt on Keller's integrity as a scientist. As a doctor's daughter, Firestein had up to this point pretty much believed in the purity of science, the tenet that its practitioners were honorable scholars laboring away in the isolation of their ivy-covered labs, impervious to the siren song of big money. But here was the chief of psychiatry at Brown essentially accused in the pages of a respected newspaper of being beholden to the drug industry. And that was not all. Firestein found an earlier article in the same newspaper charging Keller's department with falsifying invoices to obtain money from a financially strapped state agency for research that didn't appear to have happened. Still another clip noted that Brown had repaid the Commonwealth of Massachusetts more than $300,000 for the contract dollars that Keller's department had obtained under these dubious circumstances.

  "When I read the pieces about Keller, it made me look at psychiatry in a whole different light, with new cynicism," Firestein later acknowledged. "I had not realized the depth of the connection between academic researchers and the pharmaceutical industry."

  A FEW WEEKS LATER, Firestein went to see an economist who worked for the antitrust division of the AG's office. She wanted to know what he thought of the methodological problems with the 2001 Keller study.

  "Does this look right to you?" she asked him, after showing him the journal paper and explaining her concerns. "Could it be that they are claiming statistical significance where there is none?"

  "Yeah, I think you're onto something here," the economist said, squinting at the tables of data Firestein had laid on his desk. "But don't take my word for it."

  Firestein realized she would have to find an outside expert who understood data analysis and was familiar with the Paxil study. But who? During her searches online, she had come across a muckraking Web site -www.socialaudit.org.uk-that regularly posted articles critical of the SSRI antidepressants. It was run by Charles Medawar, a Brit who billed himself as a specialist on medical policy and drug safety. Firestein picked up the phone and called Medawar. He in turn referred her to a British psychiatrist by the name of David Healy.

  Near the end of August, after a round of e-mails with Healy, Firestein finally spoke with him by phone at a hotel in Philadelphia. Healy had told her he would be visiting the United States on a fact-finding mission for Baum Hedlund, a Los Angeles-based law firm then suing GlaxoSmithKline on behalf of three thousand families whose children had allegedly become addicted to Paxil. (Most of these lawsuits would not be settled for years, but in 2002, GlaxoSmithKline agreed to stop airing television commercials and distributing literature claiming that Paxil was not habit-forming.) When Firestein called that late summer evening in 2003, Healy had just returned from an exhausting day of poring through documents in the corporate archives of SmithKline Beecham, now a subsidiary of GlaxoSmithKline.

  Healy gave Firestein a quick overview. He explained that the British authorities, after reexamining the pediatric data on Paxil (which included the Keller study), had found that among eleven hundred children enrolled in those trials, children taking the drug were nearly three times as likely to consider or attempt suicide as those taking placebos. It was that reanalysis that had led the British regulators to issue their advisory against Paxil in June, compelling the FDA to begin its own reexamination of the SSRIs.

  Firestein then asked the British psychiatrist about the inconsistencies she had noticed in Keller's Paxil paper. Healy confirmed that she was correct in her analysis: the primary outcome measures had indeed failed to show that Paxil was significantly better than placebo. He told Firestein that several medical professionals had leveled the same accusation in letters to the journal's editor after the Keller et al. study was published in 2001.

  "You might be able to find those letters online," Healy said. "And when I get back to Wales, I'll look around to see if I saved any copies."

  Firestein also mentioned her confusion with the term "emotional lability." According to the medical dictionary, it meant emotional instability. Healy said that based on data he had seen of clinical trials of Paxil in both adults and children, G1axoSmithKline commonly coded participants as being emotionally labile when in fact they had expressed suicidal thoughts or made a suicide attempt.

  "It's a bit of a trick," he said. "People just reading the tables of the [published paper] might not be aware that they should look behind the term." Even FDA officials might not have understood that the term "emotional lability" was a code word for suicidal behavior when Glaxo first presented the clinical trial data on adults and children to the federal agency, he said.

  As Healy would later recall, Firestein didn't say much during that phone conversation. She asked him to go through some of the statistics again, and he had the impression that she was taking notes. "She didn't give too much away and I didn't know where she was going with this," he said. "She played things fairly deadpan"

  But back in her cluttered office at 120 Broadway, Firestein could barely contain her joy. Her reading of the data had been accurate. She still hadn't found the other pediatric studies the FDA Talk Paper alluded to, and she wasn't even sure whether the Keller study was one of the three negative studies the Talk Paper mentioned. But after talking to Healy, Firestein knew that she had sufficient ammunition to convince her colleagues that this was a case worth exploring.

  She was right. In the fall of 2003, after going over the information she had, first with Shirley Stark and then with Tom Conway and Joe Baker, Rose Firestein received the go-ahead to write an official letter to GlaxoSmithKline demanding all internal corporate documents relating to the testing and marketing of Paxil in the under-eighteen age group. Upon receiving such an official demand letter from the AG's office, corporations doing business in New York were required by law to cough up all documents, although the process could take months. Firestein was prepared to be patient. She had a feeling GlaxoSmithKline's internal documents would give her the keys she needed to unlock the mystery of why a study published in a prestigious medical journal could purport to conclude one thing when it actually showed the exact opposite.

  n Christmas Day, 2003, Rose Firestein accompanied Janice and Ellie to the corporate offices of Merrill Lynch in Midtown Manhattan (where her sister worked) to help out with an annual holiday ritual. Every Christmas, Merrill Lynch invited in seniors who lived in the New York area to use the brokerage firm's phones to call their families all over the world. Merrill Lynch employees served the visitors coffee and cookies while they waited to use the phones. Firestein had baked some cookies, and she was eager to see what her eight-year-old niece would make of it all. It was important to Rose and Janice that Ellie, growing up in privilege, far from her place of birth, understood the importance of sharing with others less fortunate than she.

  The afternoon didn't turn out the way Firestein had expected. At one point, one of her sister's colleagues asked her where she worked. Without thinking, Firestein told him. The man reared back in surprise and looked at her as if to say, What the hell are you doing here! The rest of the afternoon, Firestein could feel the stares following her around the room. It didn't matter that she personally had not been involved in Spitzer's high-profile case accusing Merrill Lynch stock analysts of misleading investors. For those who worked at the Wall Street firm, the wounds were still fresh. Firestein realized she would have to be more discreet about her place of employment: to many people in New York's insular business community, the New York State attorney general and his staff were public enemy number one.

  A FEW WEEKS LATER, Firestein was sitting in Shirley Stark's
office nursing a Diet Coke, one of many she would down in the course of the day. Stark's caffeine of choice was a Starbucks Venti, and she liked to tease Firestein about her predilection for Diet Coke. "How can you drink that stuff?" she'd say. "It's disgusting!" But today it was clear that Stark was in no mood for banter. Even though her face was a blur, Firestein could hear the distress in her supervisor's voice.

  "I just got off the phone with a child psychiatrist who is a friend of mine," Stark said. "I told him in general terms what we were doing. And he was very opposed to it."

  Stark's friend believed that if the AG's office went after a drug company for its marketing of an SSRI antidepressant, the ensuing publicity would make many doctors reluctant to prescribe the drugs at all.

  "He felt that we might be taking away the tools he needs to treat children with," she said.

  The conversation jolted Stark, making her question whether they were in fact doing the right thing in pursuing legal action against GlaxoSmithKline. After all, almost everything the AG's office had heard from the medical community seemed to indicate that SSRI antidepressants worked in the treatment of depression and other disorders like anxiety. Who were they -a bunch of attorneys-to tell physicians differently?

  Firestein listened patiently to her supervisor's concerns. The conversation helped both women understand that if they were going to pursue action against GlaxoSmithKline, they would have to do it in such a way as not to pass judgment on whether Paxil was safe and effective as a medication for children or adults. As far as Firestein was concerned, that wasn't the issue anyway. In her mind the crux was whether the drug company had somehow misled doctors and consumers in its research and marketing of Paxil.

  At this point, of course, she couldn't even prove that. True, the FDA had issued another public health advisory in October, alerting physicians to "reports of suicidal thinking (and suicide attempts) in clinical studies of various antidepressant drugs in pediatric patients." But Firestein found the FDA advisory frustratingly equivocal. On one hand, the FDA announced completion of a preliminary review of eight antidepressants, concluding that "the data do not clearly establish an association between the use of these drugs and increased suicidal thoughts or actions." But in an accompanying letter to physicians, the FDA said it could not "rule out an increased risk of suicidality in any of these drugs" Concluding that additional data were needed, the FDA Talk Paper announced that on February 2, 2004, the FDA would convene a meeting of its Psychopharmacological Drugs Advisory Committee to discuss the issue. This was the same advisory group that had brushed aside similar concerns about Prozac thirteen years earlier.

  Then in mid-December, the British health authorities issued a much stronger advisory, recommending against the use of all the SSRI antidepressants, except for Prozac, in treating depression among children and adolescents.

  In reading the press coverage, Firestein was surprised to see that among the data the FDA and British regulators had reanalyzed were nine pediatric studies on Paxil, including three randomized clinical trials. Wow, she thought. In her exhaustive search of the medical literature, she had found only one published study: the Keller report. Where were the other eight?

  As 2003 drew to a close, Firestein was still waiting for the documents she had requested from Glaxo. She wished there were some way she could quickly get her hands on the data that the British and American authorities were privy to. But the FDA had a long-standing policy of not releasing even completed clinical trial data without the consent of the company that sponsored the research.

  Firestein had no choice but to be patient. She had plenty of other work to keep her busy, having been drafted to help with several other pressing complaints. The Paxil investigation would have to wait.

  NOT FOR LONG, as it turned out. By late January, with the date of the FDA's long-awaited advisory meeting looming, press coverage of the SSRI controversy heated up. Firestein found herself drawn back to the Paxil investigation as a cat to the sound of skittering mice. Joe Baker had passed along an intriguing article published in the New York Times a few months earlier. The piece written by Gardiner Harris, confirmed Firestein's suspicions that GlaxoSmithKline had selectively published the one Paxil trial by Keller et al. and suppressed other clinical trials that found Paxil was no more effective than placebo in treating pediatric depression. In the Times article, Harris quoted Dr. Graham Emslie, chief of child and adolescent psychiatry at the University of Texas Southwestern Medical Center in Dallas and one of Keller's coauthors on the published Paxil study. Emslie, who had been involved in four of the pediatric Paxil studies, said he suspected that the other studies had gone unpublished in part because their results were unfavorable. Harris also reported that while most of the clinical studies done on the SSRI antidepressants had found them to be no more effective in fighting pediatric depression than sugar pills, much of this data had not been made public. He quoted Emslie as saying, "I know of at least a half-dozen other studies of antidepressant treatments in children and adolescents that have been completed but as yet have not been published."

  Firestein was astounded. A prominent psychiatric researcher was publicly admitting the practice of not publishing negative results. This admission would help bolster the AG's case against GlaxoSmithKline. Firestein was further heartened by what she read in the next paragraph of the article; it quoted Dr. Marcia Angell, the former editor of the New England Journal of Medicine, as saying that drug companies, by not disclosing negative research results, can actually "change the way medicine is practiced."

  Yes! She had finally found an expert who was willing to say that the drug companies, by publishing only the positive results and suppressing the negative, were actually interfering with the practice of medicine.

  A few days later, Firestein stumbled across another news story, from the Washington Post, headlined ANTIDEPRESSANT MAKERS WITHHOLD DATA ON CHILDREN. In that article, a growing chorus of critics took aim at the industry's practice of suppressing data unfavorable to their drugs. Even the director of the American Psychiatric Association's Division of Research chimed in, saying, "We need a journal of negative findings. The probability of those negative findings being published is far less than the chances of positive studies."

  Firestein began perusing the news coverage on a daily basis now. On February 1, she was appalled to read that the FDA intended to gag one of its own scientists, Dr. Andrew Mosholder, preventing him from testifying about his findings at its public hearing the next day. According to the exclusive in the San Francisco Chronicle, Mosholder, the FDA's longtime specialist in antidepressants and pediatrics, had reached a conclusion similar to that of the British: that a number of these antidepressants were indeed linked to an increase in suicidal thoughts and behavior among children and teens. But top FDA officials had decided that Mosholder couldn't present those preliminary findings at the agency's own public hearing. This was crazy, Firestein thought. "Here was someone who had been asked to do the analysis, and no one was saying it was wrong, and yet they prevented him from testifying before the very committee that was considering what to do about the SSRIs," she said.

  At the same time, the psychiatric community was mounting a vigorous defense of the SSRI antidepressants. Many prominent psychiatrists felt, as Shirley Stark's friend did, that a stronger warning on the labels for these drugs, or any restrictions on their usage, would scare physicians away from using them. And without the availability of these drugs, key practitioners believed that the suicide rate among minors would go up, not down. National statistics, in fact, showed that the suicide rate among American adolescents and adults had dropped in the 1990s, and some researchers attributed that decline to the rising use of SSRI antidepressants. Other psychiatric researchers, however, countered that there was little evidence that the two trends were connected. Dr. Ross Baldessarini, director of the Neuropharmacology Laboratory at McLean Hospital, was among those who believed there was a far stronger correlation between reduced suicide rates and environmental factor
s, such as better access to mental health care.

  While the epidemiological picture on long-term suicide trends remained unclear, many psychiatrists and primary care physicians, Baldessarini included, were convinced that the SSRI antidepressants helped depressed patients, particularly adults. They had seen the beneficial results in their own practices. By early 2004, most doctors had still not heard about the data linking the drugs to suicidal behaviors in some children and adolescents.

  On January 21, just two weeks before the FDA hearing, a task force of the American College of Neuropsychopharmacology (ACNP) issued a reassuring report. It concluded that SSRIs did not appear to increase the risk of suicide in teens and children. Only later would Firestein learn that of the ten members on the ACNP task force, nine (including three coauthors on the 2001 Paxil study) had been paid consultants to companies selling SSRIs, including G1axoSmithKline, Pfizer, and Eli Lilly. Hardly an impartial jury of experts, she thought.

  IT WASN'T UNTIL February 6, four days after the FDA hearing, that Firestein found time to open the package of background information for the hearing. The twenty-four-page memo from the FDA's Thomas Laughren was an eye-opener. Laughren was the same FDA official who had written the 2001 paper that noted (in passing) a heightened suicide risk from SSRI antidepressants. Now a team leader with the FDA's Neuropharmacological Drug Products Division, Laughren reported that out of the fifteen clinical trials done to examine the effectiveness of these drugs in treating pediatric depression, only three studies had positive results. Among the twelve studies the FDA deemed negative were all three of the randomized clinical trials done on Paxil, including the Keller study on adolescents published in 2001.

  Aha! Firestein thought. Here was the FDA reaching the same conclusion she had reached about the Keller study. Yet this particular trial had been presented to doctors as a glowing endorsement of Paxil for use in the pediatric population. In his background memo, Laughren took note of that fact, writing with dry understatement, "The published literature gives a somewhat different perspective" than the FDA's analysis. He went on to conclude, "The failure of most of these [studies] to show a benefit in major depressive disorder does heighten the concern about the possibility of certain risks that may be associated with these drugs, in particular the concern about induction of suicidality." In other words, if these drugs have not even been shown to be effective in children, that makes it all the more difficult to accept the risks of their generating suicidal behaviors in some patients.