Black Edge

by Sheelah Kolhatkar

  Looking through the data, Gilman felt a sense of excitement. Upon first glance, it seemed that the drug had indeed exerted significant benefits on the group of patients that did not have a gene that made them predisposed to Alzheimer’s. Even the group that did have the genetic weakness showed a positive trend. He couldn’t help but feel inspired by the potential implications.

  But as he scrutinized the data more closely, a problem jumped out at him. There was a lack of dose response, meaning that the symptoms of the disease did not respond to increasing doses of the drug. Gilman liked to explain this phenomenon by using the example of someone with a headache reaching into their medicine cabinet. Usually, he would say, “if you have a headache and you take one aspirin, you get a little bit of relief. If you take two aspirin, you get a little better relief. Take three aspirin, you get even more relief.” He saw no evidence of that, however, in the bapi results. There was no way that the treatment was working if its effects didn’t get stronger with larger doses.

  He tried to focus on the more positive aspects, which suggested that the Phase III trial, which had already started, was worth continuing. Bapi also appeared to be safe for patients to take, which was important. There were a few minor side effects and one major one, vasogenic edema, the brain swelling. It was potentially serious, but the study seemed to show that doctors could manage it by lowering the dose of the medication until the symptoms went away; once they subsided, they could ramp the dose back up again.

  Editing the enormous amount of data down to a twelve-minute PowerPoint presentation that would best capture the results was a task that Gilman tackled earnestly, along with the rest of the Elan team. Gilman flew back east the next day. Dr. Hulme, of Elan, emailed him a draft of the presentation the following day so they could work on polishing it. “ICAD presentation confidential. Do not distribute,” read the subject line. The presentation was enclosed as a password-protected attachment. “I’ll send the password by separate email,” Hulme wrote.

  Another email arrived an hour later.

  “Dear Sid,” it read. “The password is ‘nuggets.’ ”

  —

  At home that evening, Gilman looked through the presentation again, trying to draw conclusions from the tables and charts. These were the moments of scientific inquiry that he lived for, the excitement of discovery and revelation. He couldn’t wait to share the news with someone.

  Then his telephone rang. It was Martoma. The first thing he asked about were the trial findings.

  “I’m very excited about these results,” Gilman told him. “I do have some concerns about the decline in the placebo group, and the lack of a dose effect.” He also explained that bapi had shown stronger effects in patients who did not carry the gene that made them predisposed to Alzheimer’s than in those who were carriers of the Alzheimer’s gene, which wasn’t ideal. Still, the results showed that the way the drug functioned, with a protein designed to attack the plaque buildup in the brain, could work.

  They spent an hour and a half going over the slides in detail. Martoma displayed little emotion as he asked dozens of questions. Gilman prattled on: “This would certainly warrant a Phase III trial….I am very, very optimistic.”

  Then Martoma made a noise, signaling a change in subject.

  “Sid,” Martoma said, “my uncle passed away a couple of months ago and I was too busy to go to the funeral. I feel really bad.” He let his voice trail off. His uncle had lived in Ann Arbor, Martoma said, and he was planning a visit to his family. “I’m going to be there on Saturday. Will you be around? I’d like to drop by.”

  Gilman thought about it for a moment. “Yes, sure, you can drop in,” he said. He would be in his office, working, as he had been nearly every Saturday that he could remember.

  Two days later, Gilman drove to the campus and parked his car in the lot adjacent to the North Ingalls Building. Just after 10 A.M., he swiped his badge to get into the deserted complex through the garage and made his way to his office, where he quickly became absorbed in his work. Around 2 P.M., his phone rang. Martoma was outside.

  “Have you already had some lunch?” Gilman said as he led Martoma down the hall to his office. “We could go out and get something.”

  “No thanks,” Martoma said.

  “Shall we just sit and chat, then?” Gilman asked.

  Then, Gilman later recounted, Martoma got straight to the point: “I’d like to see the ICAD slides.”

  Gilman hesitated. Then he went to his computer and pulled up the latest version of the presentation, stepping aside so that Martoma could scroll through it. Gilman narrated while Martoma squinted at the charts and images. Gilman paused over the slide showing the dose response to the medication.

  “What is your view of the results?” Martoma asked.

  Gilman, who was still feeling hopeful about the data, and perhaps a bit defensive, reiterated his concerns about the lack of dose response and the decline of symptoms in the placebo group. But he didn’t think either of these negated the obvious benefits of bapineuzumab. “These are relative concerns, not huge concerns. They’re the first results showing a treatment that is effective, that appears to be effective, at least in the Phase II study,” he said, becoming lively, as he always did when he was talking about neuroscience. “I think these are certainly results worthy of Phase III, another larger trial. I’m very excited.”

  In spite of Gilman’s enthusiasm, Martoma could see as he flipped through the slides that the results were a huge disappointment from an investor perspective. Bapi had proven to be effective for only a small subgroup of patients; based on the methodology of the trial, even that was uncertain. More testing would be necessary to know for sure.

  Expectations for bapi had become massively inflated. The results wouldn’t be released to the public for another nine days, but Martoma could see clearly how this would be interpreted by the market: Elan and Wyeth shares would crash once this news got out.

  When they finished looking through the slides, Martoma asked to use Gilman’s phone. He placed a call to the cabdriver who had collected him at the Detroit airport that morning, who was waiting nearby. He and Gilman walked out together and said goodbye. Martoma climbed into the waiting car and headed back to the airport, where he boarded a 4 P.M. Delta Airlines flight back to JFK.

  Gilman returned to his office and continued working for the rest of the afternoon. That night he and his wife went out to dinner with their friends Maxine and Ronnie, and the doctor tried his best to forget about what had just happened.

  —

  “Is there a good time to catch up with you this morning?” Martoma typed in the subject line of an email. “It’s important.”

  Weekend mornings were times when most people lounged around with coffee and the newspaper, but early on Sunday, July 20, 2008, the morning after his return from Michigan, Martoma was as alert as he’d ever been. He needed to talk to Steve Cohen right away.

  Cohen replied by emailing his home phone number.

  At 9:45 A.M., Martoma called Cohen. They talked for twenty minutes. After they hung up the phone, Martoma emailed Cohen again, listing all of their holdings in Elan and Wyeth stock, which by now totaled more than $1 billion in value. Then Cohen typed out a message to his top trader, Phillipp Villhauer. He wanted to start selling off their Elan shares as soon as possible. And he wanted it done quietly.

  Cohen then wrote out an email to Wayne Holman, his healthcare guru. “I am driving back early today,” he wrote. He explained that he was returning sooner than expected because bad weather was going to prevent him from taking the helicopter later on, as he normally would. “We need to talk on eln, wye,” he wrote, using shorthand for Elan and Wyeth.

  The next morning, Villhauer arrived at the office early to prepare to execute Cohen’s orders. Cohen had instructed him to find some accounts that weren’t visible to the rest of the firm, so that no one could see the Elan and Wyeth sales taking place. It was an unusual way to sell off a position, not s
omething Villhauer had been asked to do before in his twelve years at SAC. But it wasn’t easy for Cohen to make a big trade without people noticing and without affecting the stock price.

  Like Cohen’s other traders, Villhauer had also learned through experience not to stick his nose into things that didn’t directly concern him. The stock sales at issue on this day were particularly likely to attract attention. Elan and Wyeth were among the firm’s largest holdings, so both SAC portfolio managers and brokers at outside firms would notice. SAC had over four hundred brokerage accounts open, and any one of them was likely to start talking.

  He checked with the operations guys, who told him that two accounts were reserved for “low visibility” sales and that he should execute the trades there. Villhauer then told Doug Schiff, the backup execution trader, that he was stepping out of the office and that Schiff should take over. He reminded Schiff to maintain close communication with Cohen.

  The Elan shares were in three accounts: Cohen’s, Martoma’s, and a general CR Intrinsic account. Villhauer stayed in contact with Schiff as he made the trades over the course of the day.

  After the market closed, Villhauer sent Martoma an update: They’d sold roughly 1.5 million shares for an average price of $35.

  “Obviously no one knows except me you and Steve,” Villhauer wrote.

  The next morning, Cohen asked him to sell 1.5 million more.

  “400k at 34.97 all dark pools,” Villhauer wrote to Cohen by instant message at 8:50 A.M., referring to private stock exchanges where trades were executed anonymously. The market hadn’t even opened yet.

  Then at 9:11 A.M.: “550k at 34.93,” indicating that he’d sold 150,000 more.

  Then: “660k 34.91.”

  “Keep selling,” Cohen wrote back.

  Over the next nine days, Cohen’s trader sold 10.5 million shares of Elan. Another trader did the same thing with SAC’s Wyeth position. No one in the market could see that SAC was selling.

  Even after liquidating his position, Cohen wasn’t finished. Once the selling was done, Cohen shorted 4.5 million shares of Elan, worth $960 million. In just over a week, he had turned his bet on the stock completely around.

  —

  The same day that Villhauer started selling off the firm’s Elan, a man named Harvey Pitt arrived at SAC’s offices in Stamford. It was early evening, and the stock market had just closed. After stepping down as chairman of the SEC in 2003, Pitt had founded a consulting firm called Kalorama Partners, selling his services to the private sector as an expert in compliance and regulatory issues. SAC had hired him to give a presentation to its staff on insider trading, a subject Pitt knew well. During his career in private practice as a corporate lawyer at Fried, Frank, Harris, Shriver & Jacobson, he played a central role in the Michael Milken case, when he served as Ivan Boesky’s personal defense lawyer and negotiated a plea bargain for the notorious arbitrageur.

  Pitt wasn’t terribly familiar with Cohen and his company. All he knew was that SAC was a very successful, very large hedge fund, and if history was any guide, insider trading training was likely overdue.

  Pitt had noticed in particular that younger traders on Wall Street who hadn’t lived through the scandals of the 1980s and 1990s needed constant reminding about what was legal and what wasn’t. Otherwise, the past was destined to repeat itself. He planned to deliver a talk he’d given many times before, reviewing the current state of securities law and the definition of material nonpublic information, as well as teaching some commonsense strategies for staying out of trouble. Senior SAC managers made it clear that everyone at the firm was expected to attend, no exceptions. One of SAC’s compliance officers personally rousted traders from their desks, practically dragging them to the cafeteria.

  Pitt was led into the packed room, where cameras had been set up to broadcast his presentation to SAC’s satellite offices. He spoke in a low, rumbling growl that projected authority. As he started, he looked around the room, and noticed instantly that Cohen wasn’t there. Pitt began by reminding the traders and portfolio managers in the room that insider trading was illegal in every state in the United States as well as every other country where SAC did business. “Insider trading is a hot topic for regulators,” he said, explaining that over one hundred insider trading cases had been brought against hedge funds in the previous year alone.

  The press, he pointed out, was also eager to publish stories about traders accused of breaking the law—there was no more perfect villain. “Don’t write or send any email or other electronic communication, or leave any voicemail message for anyone, if you wouldn’t want to see it in the media or have it read by regulators,” he warned. Cohen’s absence through all of it struck him as strange. Usually, the head of a company made a point of sitting in the front row at Pitt’s talks. “I’ve done these all over, and everybody usually attends,” Pitt later recalled. “Particularly the CEOs, because they want to make a statement.”

  But Pitt didn’t let Cohen’s absence hold him back. “Stop and think before trading,” he told his audience. “If the trade seems too good to be true, it probably is.”

  As the former SEC chairman launched into a meditation about the dangers of using outside expert network consultants because they might inadvertently disclose confidential information, a small drama was playing out elsewhere in SAC’s offices. A corporate announcement appeared on the news wires just before 5 P.M.: Brocade Communications Systems, a data networking company based in Silicon Valley, revealed that it was buying Foundry Networks, which produced switches and routers for Internet service providers. SAC’s CR Intrinsic unit owned 120,000 shares of Foundry stock—an investment that had been recommended by the analyst Ron Dennis, who was a member of Jon Horvath and Jesse Tortora’s “fight club” of information-sharing. Dennis had learned about the Foundry takeover three days earlier from a friend at a hedge fund in California, who had heard about it from Foundry’s chief information officer.

  “When is information ‘nonpublic’?” read one of the slides that Pitt had just displayed to SAC’s staff. “If it is not widely disseminated or if received with the expectation it will remain confidential.” A merger that hadn’t been announced yet was clearly inside information, based on Pitt’s or anyone else’s definition. After getting the tip, Dennis had told his portfolio manager at SAC about the impending deal, and he had bought a lot of stock. When Foundry went up 32 percent after the takeover was announced, SAC made $550,000.*1

  After Pitt’s lecture was over, he was told that the greatest trader on Wall Street wanted to meet the former SEC chairman, and he was escorted to Cohen’s office. The two shook hands.

  “Thanks for coming,” Cohen said. Then he turned back to his trading station, and Pitt was shown his way out.

  —

  Joel Ross strolled through the lobby of the Hyatt McCormick Place Hotel in downtown Chicago on the afternoon of July 28. A large, bland tower overlooking Lake Michigan that was a popular venue for large industry gatherings, the hotel was at that moment overflowing with scientists. It was the first day of the International Conference on Alzheimer’s Disease, and Gilman’s presentation of the bapi results was scheduled for the following afternoon. Ross was a gerontologist with a thriving medical practice in New Jersey, and he had treated twenty-five patients enrolled in the study. He had a flair for the theatrical, with a thick mustache and a taste for eye-catching ties.

  All of the physicians who had participated in the bapi study were invited to a private dinner that night in a small conference room in the hotel, where they would be treated to an early presentation of the Phase II results. Ross took his seat and signed a confidentiality reminder that was distributed to everyone in the room. Allison Hulme, from Elan, appeared at the podium and explained how bapi worked and what the study had hoped to achieve. She reiterated the design of the trial and the demographics and characteristics of the patients who participated. Then came one of the slides Ross had been waiting for: “Post Hoc Efficacy Analysis.”


  He felt depressed as he stared at the numbers on the screen. There was no dose response. Ross leaned over to another doctor, a well-known Alzheimer’s researcher from Harvard Medical School. “Pardon me,” he said. “I’m not statistically oriented. What’s the bottom line here—does this mean the drug is a failure, it doesn’t work?”

  The other doctor nodded somberly. “The drug did not work,” she said. “The drug failed.”

  Ross rushed out to the hotel lobby. He had arranged ahead of time to meet with Mathew Martoma, who was also attending the conference. Martoma had contacted him during the drug trial, and the two had been holding consultations through an expert network Ross worked with.

  Martoma was waiting for him in front of a glass wall that overlooked the hotel’s atrium. “How was it?” he asked.

  “The results were negative,” Ross said. “The drug failed to show efficacy.” Ross thought about the billions of dollars that had been funneled into bapi research, the dozens of hopeful patients and their families who participated in the trial—all that effort, for nothing. He also felt personally discouraged. He was sure that some of his patients had benefited from the drug, and he hoped they’d be able to continue taking it. He told Martoma that in spite of the negative results, he was still hopeful that bapi might work, because he had observed some improvements in his own patients who were taking it.

  “I don’t know how you can say that when the statistical evidence shows otherwise,” Martoma said. He cited the exact p-values, a number that indicated whether a result was statistically significant or not, and a handful of other specific figures that had just been included in the presentation to the investigators. The results still hadn’t been publicly released.