by Alison Bass
Laughren's memo also confirmed what David Healy had told Firestein: that in the fall of 2002, it was the FDA (and specifically Andrew Mosholder) who pressed GlaxoSmithKline to clarify what it meant by the term "emotional lability." Glaxo had applied for an extension of its Paxil patent on the basis of having done the requisite pediatric trials-studies 329 (the Keller study), 377, and 701. The FDA used that opportunity to demand that Glaxo separate out the possible suicidal events it had originally subsumed under the umbrella of emotional lability. It was this demand that finally prompted GlaxoSmithKline to acknowledge that there did appear to be a heightened risk of suicide in its pediatric data, prompting British and American regulators to begin investigating all the SSRIs.
The pieces were beginning to fall into place. Consumed by a sense of urgency, Firestein scanned the press coverage of the February 2 hearing. During the all-day event, held at a Holiday Inn in Bethesda, Maryland, members of the public had once again unleashed a torrent of grief and anger, sharing stories of beloved family members who had killed themselves after taking the antidepressants. The cast of witnesses with tragic stories far outnumbered the patients and physicians who spoke up on behalf of the drugs. Patient advocacy groups entered the fray, a few like NAMI-arguing that the benefits of the antidepressants outweighed their risks. But other advocacy groups expressed concern about the risks of the SSRIs, pointing out that the drug companies had hidden clinical trial results showing that the antidepressants were ineffective and dangerous in children. Vera H. Sharav, president of the Alliance for Human Research Protection, testified that the FDA had known for years that these antidepressants did not demonstrate a benefit in children and had failed to warn physicians and consumers that they increased the risk of suicide and hostility in the pediatric population. "The FDA is foot dragging, equivocating and tinkering with definitions while children are dying," Sharav charged.
Near the end of the February 2 hearing, several members of the FDA's advisory panel called on the agency to immediately alert physicians about the possible link between SSRIs and suicidal behavior. FDA officials, however, stalled again. They announced that any definitive action would have to wait until after they had reclassified all the clinical trial data on thousands of patients according to a new definition of what constituted suicidal behavior. They promised to have the results of their reanalysis by late that summer.
ROSE FIRESTEIN WAS not about to wait for the FDA's reclassification of the data. She began writing a preliminary memo outlining the facts of what she knew about the case against GlaxoSmithKline. She found that writing memos helped her pull her thoughts together, allowing her to step back from the details so that she could see the larger picture. While she had uncovered some damning facts indicating that Glaxo had misled doctors and consumers by not publishing negative results on Paxil, she still didn't have a working hypothesis or a hook on which to hang the facts. She didn't have the basic raison d'etre for the lawsuit, the theory explaining how and why GlaxoSmithKline had incontrovertibly violated New York law.
On the evening of February 10, Firestein left the office feeling tired and dispirited. It was cold and depressingly dark outside, and the streets were wet and slippery. Piles of dirty slush littered the curb. As Firestein made her way to the subway, tapping the ground in front of her, she felt grateful she had her cane. It was a relatively recent acquisition, something Firestein had resisted for a long time. A few months earlier, however, she had walked right into a metal barrier, smashing her foot into the bottom bar. No bones were broken, fortunately, but the foot took a long time to heal. Before then, Firestein had done her level best to ignore the fact that she had no peripheral vision and couldn't really see more than a few inches in front of her face. She had clung to an illusion of normalcy despite the impediments that clogged the streets and sidewalks of Lower Manhattan-speeding cabs, people rushing by, sidewalk vendors, dogs on leashes, newspaper stands and construction fencing on almost every block. Her collision with the metal barrier finally broke through her denial. A few days later, she called a state hotline, which referred her to a nonprofit service for the blind. The agency not only promised to set her up with a cane for free but said one of its representatives would provide training in its usage at a location of her choosing. Firestein chose Lincoln Center, where a very complicated grid of streets intersected. She figured if she could negotiate Lincoln Center without getting run over, she could go anywhere in New York with some modicum of safety.
That evening, after putting Ellie to bed, Firestein couldn't stop thinking about the G1axoSmithKline investigation. Bits and pieces of facts from the case kept swimming in and out of her consciousness all night.
The next morning, she was taking a shower when suddenly it hit her. This was a case of consumer fraud! The negative study results on Paxil were material to a doctor's judgment in treating patients, and they had been concealed. Yet at the same time, GlaxoSmithKline had touted the positive results from Keller's Paxil study. According to the New York statute, you didn't have to show there had been deliberate intent to conceal relevant material in proving consumer fraud, only that it had been hidden from consumers. The AG's office could charge GlaxoSmithKline with fraud because it had provided only part of the story, giving doctors and their patients a misleading picture of Paxil.
At work that morning, Firestein waited impatiently for Shirley Stark to come in and settle down at her desk. Then she rushed into her office.
"Got a minute?" she asked, trying to contain her excitement.
"Sure," Stark said.
"I reread New York's consumer fraud statute and I think I've found the hook we need for the Paxil case," Firestein said. "We know Glaxo withheld important information about Paxil from doctors, right? So wouldn't that be considered fraud? I think we can argue that by not fully disclosing all of its research results, Glaxo is guilty of fraud under the New York statute!"
Stark slowly sipped her Starbucks Venti.
"That's not the kind of case we normally bring under that statute," she said. The AG's office usually invoked New York's consumer fraud statute in going after loan companies who tried to bilk customers, or contractors who built defective homes. The law had never before been used to sue a company selling medical products.
Firestein persisted, outlining just how Glaxo's refusal to publish negative study results might fit New York's definition of consumer fraud. But Stark didn't sound convinced.
"It might be worth looking into this," she said. "I don't know. Why don't you think about it some more?"
Firestein's throat constricted. Stark didn't seem to get how innovative, how truly powerful, this legal approach might be. Firestein knew in her gut that the consumer fraud statute would give the AG's office the perfect hammer with which to curb the marketing of off-label drugs to vulnerable children. But she had learned from long experience not to argue when she was upset. Instead, she walked back to her office, closed the door, and let loose a few words more suited to a dockyard than a law office.
ithin a week, Firestein had convinced Shirley Stark that the consumer fraud angle was the way to go. The next step was bringing Tom Conway on board. In Conway's eyes, Firestein was still a newcomer to the AG's office, a relatively unknown, untested presence. So Stark suggested that she go to Conway first and give him the bare bones of the idea. Then the three of them could sit down and hash it out.
"Rose has a really good idea," Stark remembers telling her boss. But at their very first meeting in his office, Conway seemed dubious. He had a lot of questions. First, he wanted to know how G1axoSmithKline's decision not to publish its negative study results could be considered consumer fraud if the concealment was primarily directed at doctors who were running a business.
Firestein had anticipated this question. She'd run into it before-in the drug-pricing case. In thinking through the average wholesale pricing lawsuit, she and her colleagues had concluded that because a doctor has a fiduciary relationship to his or her patient, the doctor is not really acti
ng as a business; he or she is acting as an agent of the patient's. Based on that interpretation, they had included a bribery claim in the AWP complaint. As Firestein recalled, "If you're arguing that the AWP is essentially a kickback to the doctor, then because of the doctor's fiduciary relationship with his patient, it amounts to bribery."
The same fiduciary relationship would be operative in a doctor's prescribing of antidepressants to a patient, she told Conway. "So a fraud directed at a doctor is in fact a fraud directed at a consumer," she said.
Conway pursed his lips. "You may be right. But I think you need to think this through some more."
Firestein nodded, trying not to look chagrined. Here she had come up with her most creative idea in years, and the guy who counted her boss-was not impressed. Nor was he finished.
"Isn't this the FDA's bailiwick?" Conway pressed. Firestein explained that since state government regulates the practice of medicine, this kind of legal action was entirely within the state's purview.
Conway wasn't convinced.
"Are you sure we're not treading on the FDA's authority?" he persisted. "Are you sure there's nothing that says the FDA has exclusive jurisdiction here?"
"Yes, I'm sure," Firestein snapped. Her voice sounded loud and shrill to her own ears. She flushed, and added, "But I'll check it out again and get back to you"
Over the next weeks, Firestein and Stark worked hard to satisfy their chief's concerns. Joe Baker also had some questions. But it became progressively clearer that the man they had to win over was the toughminded czar of the third floor, Tom Conway. As both Firestein and Stark would later recall, he seemed to relish the role of devil's advocate and would home in on any weakness in their arguments like a mosquito zooming toward exposed human flesh. "He seemed to do it more with Rose, and sometimes it got to the point of pickiness," Stark recalled.
By the beginning of March, Firestein still hadn't received the documents from G1axoSmithKline. She was getting antsy. Conway and Baker were on her back to flesh out the facts in the case. On an impulse, Firestein decided to see if anything new had been posted on the Web. She typed "Paxil" into the Google search engine, and to her surprise, up popped a newly published article in the Canadian Medical Association Journal titled, "Drug Company Experts Advised Staff to Withhold Data about SSRI Use in Children."
The article told of a confidential internal memo written in 1998 by the Medical Affairs team at SmithKline Beecham (which subsequently merged with Glaxo Wellcome to become GlaxoSmithKline). The document, marked "SB Confidential-for Internal Use Only," discussed the results of two recently finished clinical trials on Paxil. It concluded that these two trials were "insufficiently robust" to support the FDA's approval of the drug for the treatment of pediatric depression. (While some physicians were already prescribing the drug off-label to children, FDA approval would have allowed SmithKline to substantially increase Paxil's market share.) One of the two studies mentioned in this confidential memo was the Keller Paxil trial (study 329); the other one was referred to as study 377. The memo acknowledged that study 329 had "failed to demonstrate a statistically significant difference from placebo on the primary efficacy measures" Even so, "a full manuscript will be progressed," the memo's authors wrote. "It would be commercially unacceptable," they added, "to include a statement that efficacy had not been demonstrated, as this would undermine the profile of paroxetine."
Firestein sat back in astonishment. This memo, if it was for real, would explain why Keller and his coauthors had published an abstract and conclusion that contrasted so starkly with the actual data. The drug company had insisted on it!
According to the CMAJ article, the confidential memo also acknowledged that study 377 "failed to demonstrate any separation" of Paxil from placebo. "There are no plans to publish data from study 377," the memo authors wrote. They recommended that SmithKline Beecham "effectively manage the dissemination of these data in order to minimize any potential negative commercial impact."
"Holy shit!" Firestein exclaimed. "This is unbelievable!"
Stark's eyes widened with amazement when Firestein showed her the journal article. "Gotcha," she said, grinning. As Stark would later recall, "We finally had something we could point to and say, Look at this, they really intended to suppress these results. It was the kind of evidence you look for in a case"
There was just one problem: Firestein had to get ahold of the actual memo without revealing much, if anything, about her investigation. The AG's office didn't want to tip its hand before it was ready to. In addition, she had to get permission from her superiors before even making the attempt. Even though the attorneys in each division of the AG's office worked in a collaborative, mostly collegial way, there was an unwritten set of rules and a strict hierarchy by which the attorney general governed his 630-attorney staff. Spitzer himself stood at the top of this pyramid, insulated by several top lieutenants. Although he would often pick up the phone and check in with the heads of his various divisions (such as Baker and Conway), if they wanted access to him, they usually had to go through his lieutenants. Likewise, attorneys such as Firestein and Stark, who worked in the "trenches," had to check in with their superiors on any unusual developments in their investigations.
Once again, Stark acted as go-between, asking Conway if it would be okay for Rose to call the Canadian medical journal. He said sure, go ahead. The next day, Firestein picked up the phone. After identifying herself, she was referred to an editor at the CMAJ. Without much preamble, she asked him if she could get a copy of this internal corporate document and then held her breath.
"Sure," the editor said without a moment's hesitation. "What's your fax number?"
THE NEXT DAY, Firestein called a meeting. Once she and her three colleagues were seated around the conference table in Conway's office, she plunked down the fax she had received. Baker and Conway were impressed.
"I remember we all rubbed our hands together and said this is great," Baker said. "This memo showed that [SmithKline] knew they were holding back information and that what they were doing was wrong."
As the AG team later discovered, the incriminating memo had been passed along to the Canadian medical journal by none other than David Healy. Healy, in turn, had received a copy from the BBC, which had obtained it from an anonymous whistleblower after the British broadcaster aired its second program on paroxetine and suicide.
But while the consumer attorneys now had their "smoking gun," they still didn't have the raw data from study 377 or any of the other pediatric studies on Paxil. "We didn't know the lay of the land," Firestein said. "The [British authorities] and the FDA were saying slightly different things, so there was no way we could go forward without seeing the documents."
Finally, on March 19, the long-awaited package from Glaxo arrived. It contained thousands of pages of documents, including data on the three randomized clinical pediatric studies for Paxil and a number of internal corporate memos. Also enclosed were the medical information letters that the company routinely sent out to doctors about new developments relating to Paxil.
It took Firestein weeks to pore over all the documents. What she found was a "gold mine" of duplicity. GlaxoSmithKline knew back in the 1990s that the first two of its three randomized clinical trials on Paxil in children (studies 329 and 377) failed to statistically show the drug's effectiveness in children and adolescents. (A third randomized trial, 701, finished in 2000, also failed to show efficacy in minors.) Yet here was evidence that the drug company had gone ahead anyway and paid a medical consulting company to ghostwrite a manuscript for publication of study 329. A draft of the manuscript was prepared by a woman named Sally Laden at Scientific Therapeutics Information (STI), in Springfield, New Jersey, and sent to Marty Keller and the other researchers who conducted the study. Despite data showing that Paxil did no better than placebo on the study's primary outcome measures and that patients in the Paxil group had an unusually high number of serious adverse events, the manuscript concluded that Paxil was
well tolerated in adolescents and effective in treating depression in that age group. It would later be revealed that Keller and his coauthors made some minor changes in Sally Laden's manuscript but did not object to the wording.
The manuscript, however, did not fool everyone. Back in November 1997, during a meeting at SmithKline Beecham's headquarters in Philadelphia, officials from the drug company and several authors of the study (including its principal investigator, Marty Keller) had decided to submit the paper first to JAMA, because that journal reached beyond psychiatry to many primary care physicians and pediatricians. But JAMA rejected the manuscript. Several of the doctors who read the paper as peer reviewers for JAMA had flagged the Paxil study's inconsistencies and given it a decided thumbs-down.
Finally, in January 2001, the Journal of the American Academy of Child and Adolescent Psychiatry accepted the paper, even though several of its reviewers had also flagged problems. One reviewer noted, "Overall results do not clearly indicate efficacy. Authors need to clearly note this." Another pointed out, "A relatively high rate of serious adverse effects was not addressed in the discussion." Why the journal published the Paxil study despite such criticism from its own reviewers remains unknown. The editor of the journal, Dr. Mina Dulcan, declined to be interviewed for this book. It's worth noting, however, that one of the paper's coauthors, Dr. Graham Emslie, had been on the journal's editorial board since 1996 and thus wielded tremendous clout with the journal's editors. Emslie was also the lead author on two of the primary pediatric Prozac trials sponsored by Eli Lilly, and like Keller, he had extensive ties to drugmakers. According to a 2004 task force report by the American College of Neuropsychopharmacology, Emslie was a paid consultant to GlaxoSmithKline, Eli Lilly, Pfizer, Forest Labs, Bristol-Myers Squibb, and Wyeth, among other drug companies.