by Alison Bass
Shortly after the Paxil paper was accepted, Martin Keller contacted Sally Laden, the paper's ghostwriter, to request that GlaxoSmithKline pick up the cost of sending him five hundred reprints of the published study. (In an e-mail to Glaxo officials, Laden estimated the cost of the reprints to be about fifteen hundred dollars). In a telling response to Laden on April 25, 2001, a Glaxo official wrote, "Because Dr. Keller is a member of our advisory board and an influential KOL, we will support his request to purchase 500 reprints." KOL is an industry acronym for key opinion leader, and it refers to doctors who, as prominent members of their profession, command the attention of their colleagues when they publish articles in medical journals or speak at conferences about new drugs or devices. Keller was a KOL in good standing with GlaxoSmithKline, as were Graham Emslie and two other coauthors on the Paxil paper: Neal Ryan, a professor of child psychiatry at the University of Pittsburgh School of Medicine, and Karen Wagner, a professor of psychiatry at the University of Texas Medical Branch at Galveston. (Like Keller and Emslie, Ryan and Wagner also had close ties to the drug companies that sold antidepressants. According to the ACNP report, Wagner was a consultant for GlaxoSmithKline, Lilly, Pfizer, Abbott, Forest Labs, and Bristol-Myers Squibb, among other drugmakers. Ryan consulted for GlaxoSmithKline, Pfizer, Wyeth, and Abbott.)
Their Paxil paper finally saw the light of print in the July 2001 issue of the JAACAP. A few weeks later, a Paxil product manager at GlaxoSmithKline sent a memo to all the sales representatives charged with selling Paxil to doctors in the United States. In the memo, the product manager, a man named Zachary Hawkins, wrote, "This `cutting edge' landmark study is the first to compare efficacy of an SSRI and a [tricyclic antidepressant] with placebo in the treatment of major depression in adolescents. Paxil demonstrates REMARKABLE efficacy and safety in the treatment of adolescent depression." (Hawkins himself capitalized the word "remarkable" in his memo.)
Firestein couldn't believe her eyes. This was further evidence that GlaxoSmithKline was deliberately trying to mislead doctors about Paxil. And that was not all. In the medical information letters that the company sent out to American doctors, the company not only spoke glowingly of study 329 but made no reference to the negative results from the two other randomized trials, studies 377 and 701. Nor, she discovered, were studies 377 and 701 ever published. Although a Canadian psychiatrist and one of the coauthors of study 377 did present some findings from the multisite trial at a sparsely attended session of the American Academy of Child and Adolescent Psychiatry meeting in 1999, he later said his purpose was not to cast light on whether Paxil was an effective treatment for pediatric depression, but rather to review some of the study's methodological flaws. Dr. Robert Milin made his presentation without any financial support from GlaxoSmithKline.
Firestein also discovered that while the drugmaker had acknowledged in letters to British and Canadian physicians that its clinical trials failed to demonstrate the efficacy of Paxil in the pediatric population, it withheld that information from American physicians. For example, in one medical letter to U.S. doctors sent shortly after the FDA's advisory in October 2003, GlaxoSmithKline omitted any reference to Paxil's lack of efficacy in children and adolescents or to the increased risk of suicidal behaviors in its own pediatric trials (even though the company acknowledged that risk in its letter to British and Canadian physicians). Instead, the company wrote, " GlaxoSmithKline stands firmly behind Paxil as a safe and effective medication that continues to help millions of patients suffering from mood and anxiety disorders."
On March 22, while Firestein was still digging through company documents, the FDA issued yet another public health advisory warning physicians and families to closely monitor both adults and children with depression, especially at the beginning of treatment with antidepressants. In this Talk Paper, the FDA continued to insist that "it is not yet clear whether antidepressants contribute to the emergence of suicidal thinking and behavior." The agency did say, however, that it was asking drug companies to change the labels of ten antidepressants to include stronger cautions about the need to monitor patients "for the worsening of depression and the emergence of suicidal ideation." It also reminded the public that it was still in the process of reviewing data from studies on Prozac, Paxil, Zoloft, Celexa, Effexor, Wellbutrin, Serzone, Luvox, and Remeron. All but one of these drugs (Luvox) was approved for use in adult depression, but not all of them were SSRIs. Drugs such as Effexor, Remeron, and Serzone were thought to operate on both the serotonin and norepinephrine systems of the brain, and Wellbutrin was believed to inhibit the uptake of serotonin and dopamine.
The day after the FDA's advisory, the American Psychiatric Association weighed in on the debate. "We are concerned that publicity around the advisory may cause some successfully treated patients to stop taking antidepressants," the APA news release said. "We believe it would be tragic if publicity around the FDA advisory kept people with depression from getting the help they deserve."
On March 24, the New York Times published an article citing a number of psychiatrists who said they needed these antidepressants in their medicinal black bag to treat children and adolescents. Dr. Harold Koplewicz, director of New York University's Child Study Center, was quoted as saying, "The fear I have about this [FDA] warning is that many teenagers will not get the medicine ... and that's a really tragic outcome." The article again gave Firestein and her colleagues pause.
"If the entire medical community was arrayed against us, saying [our lawsuit] was a terrible thing, that would have made it difficult for us to proceed," she said. So the assistant AGs decided to talk to a number of prominent physicians and opinion leaders to gauge what the reaction of the medical community might be to a legal action against GlaxoSmithKline.
Joe Baker picked up the phone and called the editor of the prestigious New England Journal of Medicine. Rose Firestein and Shirley Stark spoke to other experts, including the head of the New York State Psychiatric Institute and top officials at the American Psychiatric Association. Firestein also talked to two former editors of the New England Journal of Medicine, Arnold Relman and Marcia Angell. The assistant attorneys general made it clear that their case was not about whether the SSRIs were bad for kids, but whether doctors were getting enough information from the drug companies to make a good clinical judgment.
"Nobody said, Oh, that's a stupid idea," Firestein said. "A lot of people said, You won't be able to force Glaxo to publish all their data; the drug companies are never going to do that. But once these physicians understood that our case was about access to information and not a judgment by the New York attorney general's office on whether the drugs were good, bad, or indifferent, they thought it was a great idea. They all agreed that doctors should have this information."
Encouraged by what she heard, Firestein began drafting the official complaint against GlaxoSmithKline. The seminal paragraph accused the company of engaging in "repeated and persistent fraud by misrepresenting, concealing and otherwise failing to disclose to physicians information" about the safety and effectiveness of Paxil in treating pediatric depression.
As part of the complaint, Firestein wanted to spell out the results of all three randomized clinical trials, including the mixed results in 329. She wrote a memo to her colleagues carefully explaining the methodology and outcomes for all three studies.
At their next meeting, held in Tom Conway's crowded corner office on the third floor, the head of the consumer bureau jabbed at the memo in front of him.
"What does this mean?" Conway said. "Are you telling me that all three of these studies were negative?"
"Yes," Firestein said. Sitting in her usual place at the head of the table, which Conway jokingly referred to as the hot seat, she carefully explained what the primary outcome measures in studies 377 and 701 showed. And then she talked about the mixed results in study 329 and noted that the authors had actually changed an outcome measure late in the game, after the study blind had already been broken.
"When they didn't achieve statistical significance with the primary measures, it looks like they changed one of the primary end points to try to show statistical significance," she said, absentmindedly clicking the button on the end of her pen. "That doesn't look kosher to me"
"But how do you know?" Conway pressed. "Is this just your interpretation of the study or is this fact?"
Exasperated, Firestein picked up the copy of the memo she had written and waved it at him. "It's all right here!" she yelled.
Conway didn't blink an eye. He was used to her outbursts.
"It probably is, Rose. But we don't know this material as well as you do, and we have to be absolutely solid on the facts. Is there any way you can back up some of these allegations?"
Firestein sank back into her chair. It was clear that Conway wasn't going to take her word for it, even though she had carefully spelled it all out in her memo. So be it. She'd get him the backup he wanted.
"Well, I remember there were some letters to the editor talking about some of the same discrepancies I found," she said. "I wasn't able to get them online, but maybe there's a way I can get them from the journal. I'll try to do that. Okay?"
A FEW DAYS LATER, Firestein called Dr. Mina Dulcan, the editor of the Journal of the American Academy of Child and Adolescent Psychiatry. Could Dr. Dulcan help her understand how the process of peer review worked? The journal editor was happy to help.
"She [Dulcan] explained that the peer reviewers do this on a voluntary basis; it's not their day job," Firestein recalled. Dulcan went on to explain that her reviewers are generally committed to doing the right thing, but they may miss things. They don't see the raw data, so it would be difficult to pick up deliberate errors. The purpose of the peer review, Dulcan said, was to see whether the researchers had taken a reasonable methodological approach to the study and whether the results matched the conclusion.
"If there is a disconnect between the data and the text, isn't that something the peer reviewers would pick up?" Firestein asked.
"Oh yes," the editor replied.
Firestein mentioned that there appeared to be just such a discrepancy in a 2001 Paxil study published in the child and adolescent psychiatry journal. The editor's voice turned cold. She replied that she doubted that very much.
"That's not the kind of thing our reviewers would miss," she said. Firestein didn't press the point, and the conversation ended on a cordial note, with Dulcan promising to send Firestein the letters that pertained to the 2001 Paxil study.
Firestein found the conversation quite revealing. She now understood for the first time that peer review was not a guarantee of good science. It was an imperfect process at best. At its worst, it was a system rife with bias.
On April 15, Firestein received the journal's letters to the editor, and they confirmed her interpretation of study 329. Several of the correspondents, all psychiatric researchers, had picked up on the fact that Paxil did not show significant improvement over placebo on either of the two primary outcomes. Two of the letter writers pointed out that Keller and his coauthors had indeed changed a primary measure disturbingly late in the game.
A Canadian physician also expressed concern over the large number of patients in the Paxil group who suffered serious adverse effects, when compared with the two other groups in the study. "This finding would appear to be statistically significant, although this was not specifically addressed in the study," wrote Dr. Mitch Parsons of the University of Alberta in Canada. Parsons also wanted to know why the investigators had so readily dismissed the possibility that Paxil could have caused some of the serious side effects that emerged among patients in the study, such as suicidal thoughts or behaviors, hostility, and other conduct problems. "A degree of emotional lability and increased defiance is so common, in fact, that I routinely warn parents that this is likely to happen during the first few weeks on medication," Parsons wrote.
Raising similar objections in another letter to the editor, two researchers from universities in Adelaide, Australia, went so far as to accuse the Keller study of "distorted and unbalanced reporting." "Given that the research was paid for by GlaxoSmithKline, the makers of paroxetine, it is tempting to explain the mode of reporting as an attempt to show the drug in the most favorable light," Jon Jureidini and Anne Tonkin wrote in a letter published in the Journal of the American Academy of Child and Adolescent Psychiatry in May 2003.
Firestein sat down and typed up another memo to Conway, Baker, and Stark, quoting liberally from these letters to the editor. Over the next month, the four attorneys held several more meetings to discuss what Baker described as the "large issues" that needed to be covered in their legal complaint against GlaxoSmithKline. They also met to go over their markups of subsequent drafts of the complaint. But there were no more questions about primary and secondary outcome measures. By the time the complaint was ready for review by Eliot Spitzer and his top lieutenants, Baker and Conway had come to trust that if Rose Firestein said something was so, it was so. They had accepted her as a critical thinker who could process reams of complicated material and boil it down to a few accurate bullet points.
"There was hardly a question we asked that Rose didn't know the answer to," Baker recalled. "I don't know that we ever had our mind around [the details] the way Rose did, but at a certain point, we relented and trusted that she had it right."
They were about to find out if she did.
n the morning of June 2, 2004, a clerk from the New York State attorney general's office hand-delivered the eighteenpage complaint against GlaxoSmithKline to Foley Square and filed it with the Supreme Court of the State of New York. The document carried Rose Firestein's signature. Later that morning, Eliot Spitzer fielded press calls about the lawsuit. Conway, Baker, Stark, and Firesteinjoined Spitzer in his office on the twenty-fifth floor for the impromptu press conference. Dr. Barry Perlman, the president of the New York State Psychiatric Institute, also participated in several of the press calls. The institute, which represented psychiatrists throughout New York State, had issued a statement that day supporting the attorney general's lawsuit.
Spitzer was in his element. Talking to one reporter after another, the attorney general emphasized that his office wasn't making any pronouncement on whether Paxil was a safe or effective drug. This lawsuit, he said, focused solely on the failure of GlaxoSmithKline to disclose negative information. "Their effort to suppress [these] studies was harmful and improper to the doctors who were making prescribing decisions and it violated the law," he told the Wall Street Journal reporter Barbara Martinez.
"Similar suits are likely," he told Gardiner Harris, a reporter for the New York Times.
For their part, GlaxoSmithKline officials declined to comment on the specifics of the case. The company issued a statement saying that it "had acted responsibly in conducting clinical studies in pediatric patients and disseminating data from those studies." The day the lawsuit was filed, G1axoSmithKline's stock price dropped 3.2 percent. (Publicity over Paxil's link to suicidal behaviors had already dampened the drug's sales. In late July, Glaxo would announce that its earnings for the second quarter of 2004 had taken a dive, falling 13 percent over the previous year's earnings.)
OVER THE NEXT few days, news of the AG's legal action appeared not only in the New York Times, the Wall Street Journal, and the Washington Post but also in the Financial Times of London and in newspapers as far away as Asia and Australia. It did not go unnoticed that the lawsuit against GlaxoSmithKline came less than a month after Pfizer had settled the multistate legal actions over the marketing of Neurontin for a whopping $430 million. (As part of the settlement, Pfizer pleaded guilty to charges that its Parke-Davis/Warner-Lambert unit had illegally marketed Neurontin to doctors for the treatment of off-label conditions.) Several news reports cast Spitzer's lawsuit as part of a broad assault on the pharmaceutical industry's practices by government prosecutors throughout the nation. A few journalists also noted that it appeared to break new ground, relying on
a legal argument never used before. "The new wrinkle in Mr. Spitzer's suit is his argument that a drug maker is committing fraud if it does not tell doctors about trials of a medication that raise safety concerns," Gardiner Harris wrote in his front-page New York Times story on June 3.
On the third floor of the Equitable, Firestein and Stark were amazed at all the publicity the case was receiving. As was their boss. Conway kept going around with a wide grin on his face, saying, "Look what we did from our smelly little offices!"
The euphoria didn't last long. A few days later, Jean-Pierre Garnier, the chief executive of G1axoSmithKline, came out swinging. In an interview with the Wall Street Journal, Garnier attacked Spitzer's lawsuit and others like it as "an outrageous cost of doing business." "The legal system is getting out of control," he told the Journal's Jeanne Whalen on June 3. A few days later, Garnier upped the ante, accusing Spitzer of bullying and extortion. "The big picture here is that it's becoming too easy for many people to attack the pharma industry and hold it to standards that are higher than anywhere else. I don't have a problem with the standards but I do have a problem with extortion," he told the Sunday Telegraph of London in an interview published on June 6. In that interview, Garnier charged that Glaxo had fully cooperated with the AG's request for information, yet had been given no advance warning about the lawsuit. "We are not getting a fair chance to rebut because this is the almighty attorney office from mid-Manhattan," he fumed. "There is a certain amount of bullying in these tactics."