Why Good Drugs Sometimes Fail: The Bexxar Story
Quite a few biotech entrepreneurs go to work each day operating under the assumption that clinical trial data is king. If you put together a rock-solid dataset from a big clinical trial that says your experimental drug beats today’s standard of care, then you’re onto something big, or so the line of thinking goes.
Everything should fall into place. Patients will demand the drug. Doctors will prescribe it. Your stock will soar. You’ll make a fortune.
Wrong. Sometimes great drugs reach the market, and flop. That sad story played out over the past decade with a drug from GlaxoSmithKline for non-Hodgkin’s lymphoma. It could have helped tens of thousands of patients, but never came close to fulfilling its potential.
The drug was tositumomab combined with iodine 131 (Bexxar). GlaxoSmithKline said recently that it will quit making and selling the drug in February 2014, a little more than a decade after it was approved by the FDA. Usage peaked in 2006, but sales have been dropping 30 percent annually since then. Fewer than 75 patients in the U.S. got the drug last year, a Glaxo spokeswoman recently told Oncology Times.
“It’s a shame to see this being pulled off the market,” said Anas Younes, the chair of the lymphoma service at Memorial Sloan-Kettering Cancer Center in New York, when we spoke by phone last week. “If you look at this drug, it has about a 70 percent response rate, in one shot. You don’t have to keep giving it for 2-3 years. It’s simple, with not much toxicity. Many patients could have benefitted.”
There are many reasons why Bexxar failed, which I’ll get to in a minute. But the story starts, like so many in biotech, with a tantalizing scientific concept.
The idea, driven by Mark Kaminski, Richard Wahl and colleagues at the University of Michigan in the 1980s, was simple. It was to take a genetically engineered antibody, aim it at a specific marker found in abundance on cancer cells, and tack on a dose of radiation to give it extra tumor-killing punch. This was the kind of product that molecular biologists had dreamed of since the late ‘70s, when there was talk of “magic bullets” or “targeted missiles” against cancer. Bexxar and its competitor—ibritumomab loaded with a radioactive isotope called yttrium-90 (Zevalin)—arrived on the U.S. market in the early 2000s as the pioneers of this new class of treatment known as “radioimmunotherapies.”
Bexxar, developed in the late ‘90s by South San Francisco-based Coulter Pharmaceutical and acquired in 2000 by Seattle-based Corixa, had a lot going for it. The drug was aimed at a protein marker called CD20, which was already a validated molecular target for cancer, based on the success a couple years earlier of a so-called “naked” antibody from Genentech and Idec Pharmaceuticals, rituximab (Rituxan). Corixa had a well-respected CEO in Steve Gillis who attracted scientific talent, and raised lots of cash. It had a Big Pharma partner in Glaxo to help it manufacture and market the drug to the fullest.
Most importantly, Bexxar had clinical trial data on its side. The pivotal study of Bexxar was small, but offered strong evidence. It enrolled 40 patients with non-Hodgkin’s lymphoma who had no other options left at the time, because their disease had worsened after they got rituximab and several rounds of chemotherapy. The trial showed that 63 percent of those very ill patients had significant tumor shrinkage on Bexxar, and the benefit lasted a median time of more than two years (25 months). About 29 percent of patients went into complete remissions, and researchers weren’t sure exactly how long they’d last because so many remissions were still ongoing at the time of FDA approval. Several patients from the earliest clinical trials in the early ‘90s were still alive more than a decade later, and became forceful advocates for the drug.
Like all cancer drugs, Bexxar came with some baggage—significant depletion of infection-fighting white blood cells, oxygen-carrying red blood cells, and clot-forming platelet cells. But those tended to be manageable side effects. There were warnings that the drug might cause nasty immune reactions sometimes associated with antibody infusions, but that’s also true for rituximab. It wasn’t a show-stopper for Bexxar.
Still, the drug was held up by a series of FDA questions and requests for information. The new drug application was first filed in June 1999, and wasn’t approved by the FDA until four years later. That delay, an excruciating time for Corixa, allowed a combo of Rituxan and chemotherapy more time to cement its position as the standard of care in non-Hodgkin’s lymphoma.
Despite the delays, it looked like Bexxar had a good opportunity when it was approved in June 2003. Chris Rivera, a former sales executive at Genzyme, joined Corixa that year as vice president of sales to spearhead the U.S. market push. Rivera, now the president of the Washington Biotechnology & Biomedical Association, was excited. The first year sales goal was $20 million, although analysts at the time forecasted it could easily top $30 million and grow from there.
“Corixa had outstanding clinical data, and I thought that with outstanding clinical data you can almost sell anything,” Rivera said.
It didn’t take long to see things were more complicated. In his early days on the job, Rivera said he recalled holding focus groups with 15-20 or more cancer physicians. The meetings typically started with a pitch about the data. “After 20-30 minutes, we’d have people shaking their heads, saying ‘Wow, we had no idea the drug was this good,’” Rivera recalls.
But there was a catch. Oncologists who saw these non-Hodgkin’s lymphoma patients could prescribe rituximab at an infusion center, along with chemotherapy. These doctors made money on every patient that went through their infusion center. Prescribing Bexxar meant they’d have to forgo that revenue stream, and refer the patient to a nuclear medicine pharmacy or radiation oncologist who could handle Bexxar or Zevalin.
“There were complicated logistics with having oncologists refer to another part of the healthcare system they normally didn’t interact with,” Rivera says. “We couldn’t get them to change their habits. The doctor would usually say ‘Oh, I’ll give the patient another course of R-CHOP’ (Rituxan plus a specific chemo regimen) instead.”
Younes, the chair of lymphoma at Memorial Sloan-Kettering, has heard the story about oncologists rejecting Bexxar because they didn’t want to refer patients to medical centers that might be seen as competitors. He says that point is “exaggerated” and notes that oncologists refer patients to other specialists all the time. He points to other problems with Bexxar’s commercialization. “It’s almost a comedy of errors,” he says.
There was a muddled clinical trial strategy, Younes said. Multiple trials were opened up to expand Bexxar usage, which may have been well-intended, but the plan ended up confusing physicians about where the drug was most useful, Younes said. A lot of clinical trials were sponsored, making it possible for many patients who might have paid to get the drug to instead get it for free. Then at one point, Glaxo abruptly shut down all the trials, Younes said.
“They ended up pissing off a lot of people,” he said.
There were headaches in manufacturing an antibody that was linked to radiation. The radioactive piece of the drug came from a supplier in Canada, and the occasional snowstorm would throw the whole supply chain out of whack, causing patients infusions to be delayed, Rivera said. That was a big inconvenience for some patients who sometimes had to drive hours for a scheduled infusion at a big academic medical center, Rivera said.
Those kind of blunders were minor compared to the nightmares of reimbursement. Bexxar’s original wholesale price was set at $26,000, which sounded expensive to a lot of people a decade ago (although it looks like a bargain by cancer drug standards today). Younes said he didn’t recall any problems with reimbursement for Bexxar at his previous institution, MD Anderson Cancer Center in Houston, TX. But there were complaints from other physicians that they weren’t getting enough money from Medicare to justify prescribing the drug.
Reliable data on pricing and reimbursement are always hard to come by, and it was further complicated in this case because of variations from place to place. But in 2007, a radiologist at Northwestern Memorial Hospital in Chicago, Gary Dillehay, did a survey on the radioimmunotherapies for the Society of Nuclear Medicine. He found that Zevalin typically cost hospitals $22,000 to $24,000. Medicare, at that time, said it planned to reimburse hospitals $21,850 for a course of Zevalin and even less for Bexxar. Corixa, unable to turn Bexxar into a profit center, ended up being acquired by GlaxoSmithKline in 2005.
“You just can not do business if you pay $30 for something and all you get back is $24,” Dillehay told me when I was at Bloomberg News.
Of course, while Bexxar and Zevalin were struggling, science continued advancing. A competing drug from Cephalon, bendamustine (Treanda), offered a new alternative for non-Hodgkin’s lymphoma. Other promising agents started wending through the pipeline, such as Pharmacyclics and Johnson & Johnson’s ibrutinib and Gilead Sciences’s idelalisib. Many companies are working feverishly today on different types of souped-up antibodies, that combine these targeted drugs with toxins as the warhead, rather than radiation.
Bexxar missed one last big chance in 2011. That year, data was presented at the American Society of Hematology from a long-term study of 554 patients who were randomly assigned to get rituximab and chemotherapy or Bexxar plus chemotherapy. The conclusion: Both treatments were excellent. There was no statistically significant improvement in complete response rate, or survival time, for Bexxar patients.
So it’s no big surprise that GlaxoSmithKline has chosen to cut bait on Bexxar. A lot of time, money, and talent were invested in this drug that never panned out. Younes, who was involved in clinical trials of the drug in the ‘90s, estimates he only prescribed it about 10 times when it was on the market, although he continued enrolling patients in trials.
When you look at the whole story, there’s no single reason for failure. There were regulatory delays, manufacturing snafus, strong competition, reimbursement challenges, and issues around physician referral patterns.
If this story sounds familiar, it should—there are some striking similarities to what happened more recently with Dendreon’s sipuleucel-T (Provenge). If there’s a lesson here, it’s that cool science and hard medical evidence aren’t enough. When companies fail to understand the markets they are entering, the results can be quite ugly, especially as insurers tighten the screws on reimbursement. If more companies fail to pay proper attention to these issues, you can count on more promising drugs like Bexxar ending up on the industry scrap heap.