Cyberspace is spurring demand for a new leukemia treatment

Just as gratifying, side effects were few. Researchers began laying plans for a second and larger phase of testing. After that, in the time-honored calendar of drug development, would come a final round of human trials that might last for years, and then, if all was still well, an application to regulators for permission to market it.

But that leisurely schedule wasn't made for the Internet, or for results as dramatic as these. Word of the initial success, against a deadly disease for which there are few alternatives, quickly began to fly through cyberspace. In short order, patients from around the world were clamoring for the compound known only as STI-571. Politicians and celebrities lobbied Novartis on behalf of friends and relatives desperate to join a clinical trial -- and right away.

Yet Novartis had nowhere near enough of the substance to meet this demand, even if it wanted to. Scaling up production is an elaborate process, all the more so for a complex, lab-invented compound like this.

Meanwhile, inside Novartis, competing interests were tugging in different directions. While researchers saw STI-571 as a potential scientific coup and wanted to plunge ahead, others were concerned about the small size of its market. They wondered if the hundreds of millions of dollars needed to develop what might still fail would be better wagered otherwise.

STI-571, while still experimental and not a cure, has excited leukemia researchers like nothing before. But STI-571's story also shows how, in an age of instant global communications, a potential breakthrough treatment poses entirely new business and moral questions for a drug company.

The disorder that STI-571 targets is chronic myelogenous leukemia, or CML, one of four forms of leukemia. CML progresses over four to six years from a chronic stage with few symptoms to an intermediate "accelerated" phase and finally a so-called blast crisis, which kills most patients within a year. Alpha interferon can delay its progression, and risky bone-marrow transplants greatly help some patients. But for most, once the disease reaches its acute stage, no really effective therapy is available.

CML was one of the first kinds of cancer traced to a genetic flaw, in this case, mismatched copies of a chromosome in white blood cells. The mutated gene in this "Philadelphia" chromosome -- it was identified in Pennsylvania in 1960 -- gets stuck in the "on" position and causes cells to divide uncontrollably. They gradually crowd out healthy white blood cells and cripple the immune system.

Researchers long dreamed of a drug to block a mutant enzyme that this defect produces. But because it would have to be highly selective -- able to hit this precise target while sparing hundreds of close chemical cousins --many companies doubted such a drug could be devised. One of the few willing to tackle the challenge was Ciba-Geigy, the Swiss company that merged with Sandoz in 1996 to create Novartis. After seven years and hundreds of ineffective compounds, Ciba synthesized one that seemed to meet the test.

Just one problem: It caused liver toxicity in rats and dogs. Was this the end of the line? Ciba scientists ordered one extra, make-or-break round of tests in a species closer to man, trying STI-571 on monkeys. It passed.

Still, the compound languished while research executives focused their attentions on drugs for diseases affecting far more patients. CML strikes only about 10,000 people in the U.S. and Europe each year. First-stage human trials finally got going at sites that included the laboratory of Brian Druker at Oregon Health Services University in Portland, a longtime advocate of STI-571.

One of his first patients was a man named Ed Crandall. Although Mr. Crandall got only a low preliminary dose and it didn't help him, he affected the development of STI-571 nonetheless: He created the first Web site devoted to it, posting reports from other clinical-trial patients and chronicling his own battle with CML, which ended with his death in February 1999.

By then, Dr. Druker and others were seeing a striking number of remissions. In all, the blood counts of 57 of 59 chronic-stage CML patients in the Phase I trials rapidly returned to normal.

Roughly a third of them registered an even bigger benefit: sharp drops in the proportion of white blood cells that carried the defective Philadelphia chromosome. Although STI-571 still hasn't been tested over the long term, this was seen as a heartening result, raising hopes that progression to the acute stages of CML might be delayed significantly, perhaps even indefinitely. In any case, driving CML into remission would make the patient eligible for a bone-marrow transplant if a match was available.

The unexpected success played havoc with Novartis's plans for the rest of the testing program. The company had calculated the amount of STI-571 needed for the early phases of testing by assuming that many patients would drop out when the drug stopped working. But with almost every chronic-phase patient continuing to respond after months of treatment, says drug-development chief Joerg Reinhardt, "nobody could be removed from the drug, which limited the amount of STI-571 free for new patients."

And the number of would-be patients was surging. Sandy Craine, a 50-year-old London restaurant owner, was all set to have a bone-marrow transplant for her advanced CML last year when she stumbled onto the Web site of a U.S. support group telling of more clinical trials to come. "I decided to sign up, just out of the blue really, but I didn't expect anything to come of it," she recalls. Within days she got an e-mail with a detailed account of the STI-571 trials. "My oncologist was amazed I'd looked up the research but told me to put off the bone-marrow transplant and go to Portland right away if I could get in," Ms. Craine says. She did, and today she is in remission.

Peter Rowbotham, who sent her the e-mail, estimates he has read 13,000 messages about STI-571 on CML chat boards over the past year. Mr. Rowbotham, the husband of a CML patient in Vancouver, British Columbia, says, "The huge amount of information that's flowed through the Internet has become quite sophisticated, and it's given patients a tremendous feeling of power."

Their voices added to the pressure on Novartis to increase production. But besides being logistically difficult, this would carry some financial risk. If the substance never became a marketable drug, the costly effort would go to waste.

Novartis Chairman Daniel Vasella, a physician himself, took personal charge of the situation early last summer, ordering a steep increase in production. "I told people not to worry about excess supplies of STI-571 that might never be sold," Dr. Vasella recalls. "People had been trying to manage the testing program in a controlled way," he adds. "We want to get this drug available to patients quickly, and to do that you simply can't stick to bureaucratic rules."

Output for this year was originally planned at just a few hundred kilograms. The new schedule calls for 20 tons.

Novartis began making up for lost time. Initial results of clinical trials normally remain under tight wraps within the company, but a big production boost required other tactics. Gregory Burke, global head of oncology clinical research, assembled production executives in August and told them that in a few dozen patients, STI-571 had produced responses "unprecedented for any cancer compound at a comparable stage of development."

Morale soared, recalls Andreas Rummelt, head of global technical operations." After hearing such results, people volunteered to work Saturdays and Sundays to make the whole thing go much faster," he says.

That wasn't so easy. Swiss labor laws include tight curbs on overtime. So the STI-571 development had to be moved to a pilot plant already authorized to work round-the-clock shifts.

And scaling up manufacture of a new drug from a few kilograms in the lab to tons in gleaming steel reaction vessels is a delicate process. Development teams usually spend two to three years fine-tuning each step in chemical synthesis before production shifts to the company's main pharmaceutical factory in Ireland. Even though STI-571 requires a marathon 12-step chemical synthesis, the handover was completed in just over a year, Dr. Rummelt says.

Even that wasn't fast enough to keep pace as news about STI-571 spread around the world at Internet speed -- to people like Tracey van Houwelin.

Ms. van Houwelin, a Dutch CML patient for whom interferon therapy had failed, heard about STI-571 in early 1999 after joining a U.S. support group via the Internet. She says her hematologist wouldn't help her find a clinical trial to join, so she talked her way into one. Learning through the support group that London's Hammersmith Hospital would take part in the next round of trials, the 37-year-old housewife and mother phoned staff members several times a week for months "so they wouldn't forget my name," she says. When the hospital enrolled patients, she was the first one in.

She now travels to London every other month for treatment, while her family doctor in Holland does blood tests and faxes the results to Hammersmith. Although Mrs. Van Houwelin now is in remission, she remains irked by the cautious attitude of her doctors in the Netherlands. "They say it isn't right to tell patients about STI-571 until they see what the long-term effects are," she says. But "patients don't sit still with Band-Aids over their mouths waiting to die any more. I don't have time to wait on the long-term effects."

Even the pace of the London trials owes something to patient activism, in this case by a 33-year-old Montreal woman named Suzan McNamara. By mid-1999, her CML was outwitting interferon and on the verge of progressing to the accelerated stage. "I was very sick and in a panic because once you go into accelerated stage, STI-571 isn't as effective," she says.

Ms. McNamara also had heard about STI-571 from Internet chat groups, and in September she called Dr. Druker, hoping to be accepted into the Portland trial. He warned that chances were slim because limited supplies made it impossible to expand testing as rapidly as he wanted.

Ms. McNamara drafted an online petition pressing Novartis to step up production of STI-571. "My goal was 500 signatures," she says. My mid-October, she had more than 4,000, and sent the petition to Dr. Vasella.

"Before that," the Novartis chairman says, "I'd never had any contact with the power of the Internet."

He was able to offer the kind of reply the petitioners wanted: Novartis told them it had already stepped up production.

Moreover, it said it had decided to open Phase II trials in Europe and the U.S. in January 2000, several months ahead of schedule, and set up a hotline to help patients find the nearest trial site. Ms. McNamara joined Dr. Druker's trial and has had a strong response to STI-571.

Not everybody was satisfied with Novartis. Internet chat groups debated whether to adopt the more aggressive tactics of AIDS activists and try to embarrass the company into moving even faster. Working behind the scenes, Dr. Druker persuaded patients to hold off until December 1999, when he was scheduled to present preliminary data from Phase I trials at a meeting of the American Society of Hematology in New Orleans.

His presentation there electrified an audience of nearly 10,000 physicians. Since the conference, Novartis executives have been deluged with calls and letters, including overtures from a queen and a prime minister, on behalf of friends or relatives desperate for access to STI-571.

Production still trails demand, and as 32 cancer centers move into Phase II trials, rationing the supplies has become a delicate problem. Novartis rejected the idea of a patient lottery and instead set strict eligibility criteria -- incurring criticism from advocacy groups and excluded patients. The question, says Novartis Research Director Paul Herrling, was "Who are people with CML who can wait another month and who should have it tomorrow to save their life?"

The criteria ended up excluding the operator of a key CML Web site, Jerry Mayfield, because his disease is still controlled by interferon.

And the company has adopted an aggressive, two-track strategy for gaining regulatory approval. Instead of proceeding through all three phases of human testing before seeking any approval -- the normal practice -- it will try to get the drug approved for advanced cases based just on Phase II tests. The U.S. Food and Drug Administration has agreed to such a fast-track approach, reflecting greater flexibility the agency has shown lately in getting important cancer drugs on the market.

Meanwhile, Novartis will proceed with elaborate Phase III trials aimed at winning broader approval of STI-571, for patients whose CML is still in the initial, chronic stage. These trials, which are about to begin, will compare STI-571 against alpha interferon, the current standard treatment. They could take three to four years before yielding statistically significant data. Long before that, Novartis hopes to have the drug on the market in the U.S. for advanced cases.

It might not be limited to them. Once a drug is approved for any condition, doctors are free to prescribe it for other cases.

Some Novartis scientists worry about this testing speed-up. "One of the benefits of going slow in a trial is that the number of patients at risk of bad things happening at any one time is small," says Dr. Burke. "Without having information about long-term exposure, you could put a whole mess of patients at risk."

But luminaries in the cancer establishment are keeping up the pressure. The director of the U.S. National Cancer Institute, Richard D. Klausner, recently called Dr. Vasella with an offer to collaborate on tests of STI-571 against certain solid tumors, based on indications it might help there, too.

Recruitment into clinical trials has exploded, with more than 1,000 CML patients now getting STI-571 and hundreds more about to, in the Phase III trial. It will be done simultaneously in 14 countries.

Despite the huge risks associated with a drug still in early stages of testing, Novartis officials now expect to submit applications for regulatory approval early next year and STI-571 could reach pharmacy shelves before the end of 2001, a pace previously matched only by a handful of AIDS medicines.