Accelerated Approval and the FDA

Excerpted from: Drugs and the FDA: Safety, Efficacy, and the Public’s Trust, by Mikkael A. Sekeres, MD (The MIT Press 2022).

The 130-acre White Oak campus of the U.S. Food and Drug Administration is located one mile north of the Capital Beltway in Montgomery County, Maryland, making it convenient to get to by car. Its fragmented buildings house the Office of the Commissioner, the Office of Regulatory Affairs, the Center for Drug Evaluation and Research, the Center for Devices and Radiologic Health, the Center for Biologics Evaluation and Research, and offices for the Center for Veterinary Medicine.

The site, acquired by the US Department of the Navy in 1944, originally housed the Dahlgren Division of the Naval Surface Warfare Center, White Oak Detachment, which developed and tested ballistics, torpedoes, mines, and other explosives. In 1993, the Navy moved its operations and transferred the site to the General Services Administration; the FDA then occupied the campus in 2009. True to its military origins, the main ingress to the FDA—Building 1—resembles a fortress, and still has Naval Ordnance Laboratory etched in stone on its frontispiece.

The quarter-mile-long Mahan Road, named for the nineteenth-century naval strategist Alfred Thayer Mahan, leads from New Hampshire Avenue and ends with a circular driveway in front of the building. The low-slung glass entrance, accessible only by navigating a row of cement pillar barricades, fronts a four-story, imposing red-brick building with long, vertical windows, opaque when viewed from the outside. On the inside, TSA-level security awaits visitors, who must also wear black-and-white photo identification stickers while on the premises. At any given time, taxis or Uber or Lyft vehicles line the drive, along with the occasional catering service delivery car or corporate limousine.

But early in the morning on June 28, 2011, seven police cruisers bearing the insignia of the Department of Homeland Security Federal Protective Service and the Montgomery County Police Department blocked the campus entrance from casual visitors. A crowd had started to form on New Hampshire Avenue. Many people wore pink T-shirts with a white ribbon on the front and the words “Avastin Works! Let Doctors and Patients Decide!” on the back. Some carried home-made signs with pictures of families and messages of “Breast Cancer Patients are NOT Statistics” and “It’s MY cancer! It should be MY decision!”

On this day the FDA would hold a hearing to decide whether to withdraw its previous approval of the drug bevacizumab (brand name Avastin)—the anti-angiogenesis wunderkind born from studies conducted by the Harvard researcher Judah Folkman—for the treatment of metastatic breast cancer. Studies from his laboratory showed that killing the blood vessel growth and, thus, the blood supply to tumors could eliminate the tumors themselves. Representatives of the FDA would face off against the attorneys and leadership of Avastin’s manufacturer, Genentech, who desperately wanted to keep their blockbuster drug on the market for breast cancer.

Protest organizer Terry Kalley from Troy, Michigan, whose wife Arlene had battled breast cancer for over 30 years and eventually was treated with Avastin herself when the cancer became metastatic, stood in front of a crowd. Wearing one of the pink T-shirts, he held up a white megaphone to speak to those gathered, using language more commonly invoked at abortion rights rallies, pitting a woman’s choice—in this case, to have access to Avastin—against a government agency’s proposed restrictions to that choice.

“Today, it’s Avastin for women with breast cancer. Tomorrow, it’s the next drug that you or a loved one may need… This is a national disgrace,” Kalley shouted.

A folk singer, Andrew Katz, sported one of the same pink T-shirts along with dark shades as he strummed a guitar and played the harmonica, singing the Bob Dylanesque “Avastin Protest Anthem:”

Now I tell this story and I tell it right,
Tell about women in a hell of a fight.
Mothers, sisters, daughters too,
Americans just like me and you,
17,500 of ’em—
Cancer in their breast.
Tell you what son, boy, dad, pops, Mr. President
It’s a hell of a test.

These sorts of issues—a government agency’s regulatory reach; patient autonomy and access to drugs for life-threatening conditions; testimonials from patients and doctors; assessment of drug safety and efficacy; and the precedent set by preventing the sales and marketing of a drug—were not new to the FDA, and in fact had both colored and informed its authority for over a century.

But it begged the question: Had the FDA made a mistake in approving the drug in the first place?

—

Avastin was approved by the FDA under its accelerated approval program. Accelerated approval allows drugs for life-threatening diseases to reach the market faster, on the basis of data that is reasonably likely to translate to a clinically meaningful benefit, often defined as patients who “live longer” or “live better.” In exchange for a faster approval, though, the regulatory mechanism requires the conduct of a post-marketing confirmatory trial to “verify and describe the anticipated effect on irreversible morbidity or mortality”—usually a randomized, phase III trial comparing the new drug to standard treatments—and that this trial be completed in a timely fashion.

Accelerated approval was born from the AIDS activism of the 1980s, particularly groups like ACT UP, which staged protests on the FDA campus, similar to the Avastin protestors, that shut down the agency on more than one occasion. This type of militant activism, in turn, was born from the appalling numbers of people who were dying from a strange new virus as an often uncaring, tone deaf, and homophobic government failed to respond.

It was not until May 1983—almost two years after the first report of an unusual pneumonia occurring in five gay men—that Congress passed its first bill committing $12 million of funding for AIDS research and treatment. Compare that to the weeks it took Congress in 2020 to fund $1.25 billion in research on Covid-19, in which sexual orientation couldn’t be implicated as an etiology. Even accounting for inflation, this amount is still 42 times what was earmarked for AIDS.

Meanwhile, in 1982, 853 people died from AIDS in the United States. This number jumped to 2,304 deaths in 1983, 4,251 in 1984, and 5,636 in 1985.

According to the CDC, more than 50,000 people were diagnosed with AIDS in the United States from 1981 to 1987, and 96 percent of those people died.

From 1988 to 1992, more than 200,000 people in the United States were diagnosed with AIDS, 90 percent of whom died.

It wasn’t until 1987 that the first drug to treat people with AIDS—AZT—was approved by the FDA. AZT was no panacea, but at least it was something. In the words of Samuel Broder, MD, whose lab at the NCI led the seminal studies showing AZT’s ability to control HIV in test tubes:

“We strongly believed then, as we do now, in Voltaire’s maxim: “Le mieux est l’ennemi du bien,” which translated for the AIDS pandemic means, “The perfect is the enemy of the good.” Rather than wait for the perfect antiretroviral drug to be developed, we decided to proceed with what we had in hand as rapidly as possible.”

The drug approval process at the time was lugubrious at best, typically taking 8-10 years for a drug to make it to the market from its first studies in humans. Members of ACT UP demanded that it be shortened, and that placebo-controlled randomized trials as a prerequisite for approval be eliminated. They also wanted to ensure that patients enrolled in trials be representative of those who had AIDS, and that insurance be required to pay for experimental therapies.

By January 1992, public and political pressure for the FDA to accelerate the drug development and approval process for people with life threatening diseases had reached a fevered pitch. Groups like ACT UP staged thousands of demonstrations from 1987 through 1996. Congress started looking into allegations of a “drug lag” at the FDA. Although the FDA continued to claim it was under-resourced to meet the nation’s need for new and guaranteed safe drugs to be approved for marketing, the agency responded by amending its internal prioritization scheme. It gave drugs indicated for the treatment of AIDS or HIV-related disease an “AA” designation—its top category. First in line.

But while the FDA recognized the importance of drugs to treat AIDS, it still had to get them through their regulatory processes faster.

In 1992 Congress finally passed the Prescription Drug User Fee Act (PDUFA), legislation that could finally deliver on hastening the pace of approval and giving the FDA the funding it needed to make that process viable. PDUFA established a system of user fees, to be paid for by the pharmaceutical manufacturer when it applied for the New Drug Application (NDA), which the FDA required before considering a drug for marketing approval. In return, the FDA committed to a timetable for drug approval that would be quicker than before, and it created the accelerated approval mechanism for qualifying drugs.

To be eligible for accelerated approval, a drug was required to meet three criteria:

1. It had to treat a serious condition
2. It had to provide a meaningful advantage over available therapy
3. It had to demonstrate an effect on a surrogate endpoint that is reasonably likely to predict clinical benefit, or on a clinical endpoint that can be measured earlier than irreversible morbidity (illness) or mortality—what’s called an intermediate clinical endpoint.

—

At around 7:30 a.m. on that hot June day in 2011, many of the pro-Avastin protestors drifted toward the entrance of Building 1. After passing the many police cars, barricades, the formidable building’s entrance, and the security checkpoint, they walked by a small historical display of some of the FDA’s accomplishments, headed down a hallway, and then exited through a small courtyard to reach Building 31, where the hearing on Avastin would take place in the Room 1503, known as the Great Room.

The white shuttle bus bringing me and the six other members of the Oncologic Drugs Advisory Committee (ODAC) of the FDA took an alternate route from the Hilton Hotel in Silver Spring, Maryland, where we had spent the night. It drove past the protestors and our usual destination at FDA Building 1, depositing us instead at the entrance to the nondescript Building 32 to receive our temporary badges and run the gauntlet through a smaller security area before heading to the hearing.

Our job as ODAC members was typically to evaluate data concerning the safety and efficacy of marketed and investigational drugs for use in the treatment of cancer, and then to make recommendations to the FDA commissioner. But that day we would serve as the hearing’s “jury.” A member of the FDA itself had already informed us that it might be helpful to think of this debate—between the FDA’s Center for Drug Evaluation and Research (CDER, the arm of the FDA that regulates over-the-counter and prescription drugs) and Genentech—as a trial.

Both the agency and the company would present witnesses who would try to make the case for their respective positions—for Genentech, that Avastin should remain on the market, retaining its breast cancer indication; and for the FDA, that the breast cancer indication should be withdrawn. The opposing party would then question those witnesses. We would then have an opportunity to question those witnesses, as would the presiding officer for the FDA, Karen Midthun, MD. Finally, the party presenting the witnesses would have a chance to ask clarifying questions of its own witnesses.

We would have a busy couple of days ahead of us: reviewing the evidence supporting Avastin’s safety and efficacy track record across multiple clinical trials; then hearing the testimonies of physician experts and individual patients who have taken the drug (the women I would consider to be particular experts about the drug’s side effects and its impact on beating back breast cancer); and, finally, asking questions of the people making presentations at the hearing. We would then provide advice and recommendations to the FDA: advice on whether it should continue to allow the company to market Avastin to the tens of thousands of women with metastatic breast cancer who had few other treatment options (Genentech’s contention), or whether the toxicities of the drug outweighed its potential benefit and it should therefore be withdrawn from the market (FDA’s position). We would vote publicly, for all to see in real time, whether we were for or against Avastin. The FDA Commissioner would then make the final decision on whether or not to withdraw Avastin’s label for metastatic breast cancer.

A few times each year, the FDA holds a public hearing and calls together our committee, ODAC, to consider a cancer drug’s safety and efficacy data in making a recommendation for approving, or not approving, that drug. And a few times a year, a line maybe four or five people deep waits in front of a folding table to sign in before being allowed to enter into the rear of the Great Room. On June 28, 2011, however, a line of people 25 or 30 long snaked around the vestibule, disrupting another line of 20 people waiting in front of the small snack bar that typically struggled for business.

Usually, these open hearings were attended by those required to be there—maybe 50 to 75 people comprising FDA representatives, pharmaceutical company employees, and members of ODAC, along with other FDA personnel who may have evaluated the drug’s application documents, other company staff, a few employees from unrelated pharmaceutical companies preparing for their own upcoming ODAC meeting, some reporters, and a lot of representatives from financial investment firms.

Financial investment firms? That’s because drugs are big business.

They are such big business—and an ODAC recommendation is so impactful on a pharmaceutical manufacturer’s bottom line—that trading of a company’s stocks on Wall Street is suspended during a public hearing because swings in stock prices can became too volatile, depending on positive or negative comments made by individual members of ODAC, in real time.

On that unusual day in June 2011, hundreds of people packed the room that appeared, not so much “Great,” but just the right size for the number of attendees.

I walked down a side hallway that runs the length of the room, passing television cameras and encamped cameramen, and reporters from local and network news stations jockeying for position near an open door. Another door at the far end of the hallway was marked “ODAC Members Only.” I swung it open and walked in.

Chaos.

The seats were already three-quarters filled. A cacophony of heated conversation filled the room, where usually it was respectfully quiet before hearings began. A scrum of cameramen thronged near the table where I and the six other members of ODAC would sit, like encroaching jackals surrounding fresh kill.

—

Since its inception in 1992, the FDA’s accelerated approval mechanism was used to grant more than 275 marketing approvals for drugs (with some drugs receiving approval for multiple indications). The very first, in June 1992, was zalcitabine, used in combination with AZT to treat patients with advanced HIV infection and significant deterioration in their health and/or immune system. The following year the antibiotic clarithromycin took advantage of this new regulatory mechanism for the treatment of mycobacterial infections—one of the opportunistic infections to which people with more advanced HIV were susceptible. The year after that, in 1994, stavudine received accelerated approval for HIV patients with advanced disease. And in 1995, five drugs were approved: three for HIV or complications of the virus, and for the first time two for cancer; Casodex for advanced prostate cancer; and zinecard to prevent heart failure from certain types of chemotherapy.

From 1993 to 1995, 257,000 people in the United States were diagnosed with AIDS. Whereas previously the mortality rate had been 90 percent or higher, now 62 percent died. New drugs, with better efficacy and fewer side effects, were getting to people with HIV/AIDS faster, and they were starting to have an impact on meaningful outcomes, like survival. And this could actually be seen on a population level.

AIDS activists desperately fought over the years for a treatment—any treatment—to combat this terrible disease that had laid waste to a generation. But use of the accelerated approval mechanism had been commandeered for the first time by pharmaceutical companies making drugs to treat the far vaster—and far more lucrative—cancer patient population.

In 1996, eight drugs received accelerated approval; one for low blood pressure, four directed to HIV or complications of AIDS and three were for cancer. By 1999 the ratio had shifted, with five drugs focusing on cancer or cancer prevention, and only one—Agenerase—on HIV. From that point through 2008, cancer drugs received accelerated approval at a ratio of three-to-one compared to drugs for HIV/AIDS.

In 2008, the drug Etravirine was approved for the treatment of patients with HIV that was resistant to other HIV medications. It would be the last HIV medicine approved under the accelerated mechanism at the FDA until May 2020 when pomalidomide—already on the market to treat the cancer multiple myeloma—was approved for another cancer associated with HIV, Kaposi’s sarcoma. That same year, in 2008, four drugs were approved for oncology indications, one for lung cancer, another for a soft tissue cancer, one for leukemia, and another for breast cancer.

The one for breast cancer was Avastin.

By 2009, and for the next decade, medications that treated HIV or complications from the virus disappeared entirely from the list of drugs that received accelerated approval, which was now almost completely dominated by cancer therapies. By the end of that decade, cancer indications accounted for 17 out of 21 accelerated approvals in 2017; 16 of 18 approvals in 2018; 11 of 14 approvals in 2019; and 40 of 44 approvals (some involving one drug for multiple indications) in 2020.

From the standpoint of accelerated approval, cancer was king.

A critical aspect of the accelerated approval program is the requirement for a post-marketing confirmatory trial to “verify and describe the anticipated effect on irreversible morbidity or mortality,” and for this trial to be completed in a timely fashion. In other words, when a drug receives accelerated approval based on a surrogate marker of efficacy (such as a tumor that shrinks – often called a “response rate”) that is reasonably likely to translate to a clinically meaningful benefit (such as someone living longer), the FDA wants to then see a confirmatory trial demonstrating that that clinically meaningful benefit wasn’t a fluke.

The FDA’s accelerated mechanism has approved some wildly effective drugs. In the 1980s and 1990s people diagnosed with chronic myeloid leukemia—a cancer caused by a specific genetic mutation known as the “Philadelphia chromosome”—could expect to live, on average, for about three years after their diagnosis. Then the drug imatinib (brand name Gleevec), which targeted that genetic mutation, was discovered and tested in 532 CML patients. Almost all everyone who received the drug had a complete remission from their leukemia, and in almost one-third the Philadelphia chromosome—the root of all evil in CML—was eradicated entirely.

This set the stage to consider accelerated approval: CML was a serious condition that profoundly affected the survival of a person suffering from it; the standard therapy for that condition was inadequate; and it had a measurable surrogate endpoint (the Philadelphia chromosome), the elimination of which could reasonably translate to the clinically meaningful benefit of improved survival.

And in May 2001, the FDA granted accelerated approval to Gleevec. The agency based the decision not on survival but on elimination of the Philadelphia chromosome. In a remarkable show of efficiency, the entire FDA review process took just 72 days—at the time, the fastest approval ever for a cancer drug.

The confirmatory study for Gleevec, reported in March 2003, randomized over 1,100 patients with CML to receive Gleevec or standard chemotherapy. The results were equally dramatic as the initial studies that led to the drug’s accelerated approval: After a year and a half of treatment, 95 percent of patients treated with Gleevec improved and 85 percent had some degree of eradication of their Philadelphia chromosome.

Based on the results of this study, Gleevec was granted full approval by the FDA in December 2003.

What happens, though, when a manufacturer is not able to actually confirm the initial benefit that led to accelerated approval, or fails to complete such a confirmatory trial?

Well, then the FDA can—and will—exercise its authority to withdraw that marketing approval, what one FDA representative referred to as “accelerated withdrawal.”

—

Avastin had received accelerated approval by the FDA, in combination with chemotherapy—a drug called paclitaxel—for the treatment of women with metastatic breast cancer. At the time, Avastin had already been approved for the treatment of patients with colorectal cancer or lung cancer.

In the study that led to its breast cancer indication, which started enrolling patients in 2001, 772 women with breast cancer that had spread to other organs were randomized to receive Avastin combined with paclitaxel, or paclitaxel alone. Women receiving Avastin along with chemotherapy demonstrated progression-free survival a full year–about six months longer than women receiving chemotherapy alone. This was a statistically significant difference. Progression-free survival meant that women receiving Avastin went a longer period of time without their tumors worsening (progressing) than women receiving the chemotherapy alone.

The average overall survival between the groups did not differ, though: Studies showed an average of 26.7 months for those receiving Avastin, versus 25.2 months for those who didn’t. Additionally, patients receiving Avastin combined with chemotherapy were more likely to experience serious infections, high blood pressure, headache, and kidney problems.

Despite a vote against approval by ODAC, the FDA felt this was enough of a signal that a surrogate endpoint (progression-free survival) was reasonably likely to predict clinical benefit (overall survival). Avastin received accelerated approval for breast cancer in 2008, requiring that follow-up studies confirm the initial benefit, and hopefully extend it.

One follow-up study, abbreviated AVADO for Avastin And Docetaxel, was structurally similar to the first trial, and enrolled patients from March 2006 through October 2007. A total of 736 women with metastatic breast cancer were randomly selected to receive a chemotherapy cousin of paclitaxel called docetaxel, or to be treated with docetaxel combined with two different doses of Avastin. Women receiving the chemotherapy docetaxel alone had a progression-free survival average of 7.9 months—a bit more than women receiving paclitaxel alone on the initial study.

Those receiving docetaxel combined with Avastin on the AVADO study had an improvement in progression-free survival that amounted to just a little over three-and-a-half weeks more than those getting docetaxel alone.

Weeks. Not even months. In fact, a full five months less of an improvement for those receiving Avastin than observed in the original study. Overall survival was no different between the groups. Serious side effects were, again, all higher among women receiving Avastin.

In the second follow-up study, called RIBBON-1 for Regimens In Bevacizumab for Breast Oncology (the uninspired acronym alone may have presaged where this trial was headed), enrolled 1,237 women with metastatic breast cancer to receive one of three different types of chemotherapy alone, or combined with Avastin. The first woman entered the trial in December 2005, and the final woman was enrolled in August 2007.

The average progression-free survival for women treated with chemotherapy alone used in the other studies was almost identical to that seen in the AVADO study, at eight months. Those who also received Avastin enjoyed an average improvement in progression free survival of an additional 5 weeks for those treated with a chemotherapy similar to previous studies, and 12 weeks for women treated with other types of chemotherapy combined with Avastin.

That is still far inferior to what was seen in the original 2001 study (the basis for Avastin’s accelerated approval). Again, overall survival did not differ between the groups. And again, toxicities were higher for those treated with Avastin, including bleeding, a suppressed immune system, fever, and high blood pressure.

As the Avastin hearings got underway on that hot and humid day in June 2011, a day when the temperature hit 90 degrees, the FDA’s presiding officer, Dr. Midthun, outlined the four issues we would be considering over the course of June 28 and 29:

1. Did two recent trials fail to verify the initial benefit that was seen with Avastin in women with metastatic breast cancer—the benefit that led to its initial approval?
2. Did the total available evidence for Avastin demonstrate that it was not, in fact, effective for the indication (metastatic breast cancer) for which it was approved?
3. Did the total available evidence for Avastin demonstrate that it was not, in fact, safe for the indication, and that the clinical benefit did not justify these risks?
4. If we decided that the drug should be withdrawn from the market, should the FDA allow continued approval of the drug for breast cancer while Genentech designed and conducted additional studies to try to demonstrate Avastin’s benefit?

Each of these questions, in and of itself, had enormous implications. We were to decide if clinical trials enrolling thousands of women with breast cancer—which Genentech and a number of world-renowned breast cancer specialists had touted as undeniable successes in support of Avastin—were in fact abject failures. Then we were to vote on whether Avastin was effective at all in treating women with metastatic breast cancer. From a public health perspective, the third question was probably the most important; if the drug fulfilled the criterion for safety required by the 1938 Food, Drug, and Cosmetic Act. And finally, if our conclusions fell somewhere in a gray zone of safety and benefit, was Avastin safe enough and did it demonstrate at least a modicum of benefit to justify keeping it on the market while a better clinical trial was designed?

Members of the public were invited to speak at the hearings, to tell us about their experience with Avastin. All told, 13 women with breast cancer treated with Avastin and other chemotherapy drugs shared their stories with us, and we heard from their husbands, friends, and doctors, and from other advocates for women with breast cancer.

Thirty-four people in total. It was awe-inspiring.

Heraleen Broome, a 74-year-old woman from Oakland, California, was the first patient to approach the microphone to talk about her experience with Avastin. She wore glasses that arched up at each side and had a face that was warm and expressive. She spoke for many other women in the FDA’s Great Room that day when she told us:

“I don’t think it is reasonable for you to set a number of people that need to be alive as a result of this drug in order to allow it to be sold. It seems to me that my life should be enough, and it’s not just my life but the lives of my family, friends, coworkers and everyone I meet.”

Her comment was met with thunderous applause.

These women were living their lives. They had children and grandchildren. They were not aggregates of data. They were women with breast cancer who were healthy enough to participate in a clinical trial with an investigational new drug and to then get in a car or on a plane and travel to the White Oaks facility in Rockville, Maryland, to remind everyone at this hearing what was at stake. And they credited Avastin with still being alive.

Then, an FDA representative came up to the mic to tell us another story, one that countered the stories we had heard from patients and their advocates:

“This morning we have heard from patients and their families describing how they feel they have benefited from Avastin. However, there are other voices that need to be heard. Those voices include a 53-year-old woman with metastatic breast cancer who suffered severe abdominal pain caused by gastrointestinal perforation that led to her death after 4 doses of Avastin; or an asymptomatic 33-year-old woman with metastatic breast cancer who suffered a massive fatal pulmonary hemorrhage after 11 doses of Avastin.”

We heard from the living. She spoke for those who died.

In 2021, another FDA advisory committee voted overwhelmingly against the Alzheimer’s drug aducanumab. The drug had been studied in two randomized trials: one was felt to demonstrate a modicum of clinical benefit to patients by Biogen, the manufacturer; the other trial was a failure. Aducanumab was associated with reduction of the amyloid plaques that typify Alzheimer’s, though other analyses showed no correlation between plaque reduction and improvement in cognition. The drug also was associated with significant side effects, including bleeding into the brain.

Despite the committee’s vote and the study results, and similar to what happened with Avastin, the FDA granted accelerated approval to aducanumab, based largely on the plaque reduction. The agency likely felt pressure to grant the approval as the first new therapy for Alzheimer’s in 18 years. Immediately, three members of that committee resigned, and reprobation from the scientific community was swift. Some hospitals have refused to allow the drug on their formularies, as its use remains controversial despite the approval

In June 2012, we weighed some of the same issues as the aducanumab committee, and as future FDA advisory committees will, about drugs that don’t work quite as well, or have toxicities that are unacceptably high, for people with life-threatening diseases.  We would have to vote our conscience on whether Avastin allowed women with a life-threatening disease to live longer or live better; or whether it threatened their health even further.