This is the first of two stories on monitoring the safety of new drugs — read the second here.
When a new remedy for rheumatoid arthritis arrived, ads called it a “unique” breakthrough that would “transform expectations” for patients and doctors. “If I knew then what I know now about rheumatoid arthritis, I would have been more proactive,” said one young woman, pictured happily kayaking.
Treatments for the disabling disease afflicting about 1.5 million Americans can have terrifying side effects, so doctors and patients were excited when Actemra reached the U.S. market in 2010. Unlike competing drugs, it wasn’t associated with heart attacks, heart failure, or life-threatening lung complications.
Yet hundreds of patients taking Actemra have died from such problems, and many more have suffered harm. STAT analyzed more than 500,000 side-effect reports on rheumatoid arthritis drugs, and found clear evidence that the risks of heart attacks, strokes, heart failure, and other conditions were as high or higher for Actemra patients than for patients taking some competing drugs.
Most of those medications warn about these risks on their labels. Actemra does not.
Consumers are barraged every day with drug ads accompanied by numbing lists of side effects, but STAT’s investigation shows that the risks to patients might be greater than they are led to believe. The Food and Drug Administration has received reports on 1,128 people who died after taking Actemra, and has reviewed its safety several times since it was approved. But the agency doesn’t have sophisticated tools to determine whether the drug was a culprit or a bystander in those deaths.
Though the agency is charged with monitoring the safety of prescription drugs, it doesn’t verify the side-effect reports it receives. The documents often lack crucial information, and they don’t prove that Actemra was the cause. Still, they can be telling.
In one striking example, obtained through the Freedom of Information Act, a doctor said no factor other than the drug could have explained a 73-year-old man’s fatal brain bleed two days after receiving an intravenous Actemra treatment. Another said of a 62-year-old German woman’s heart attack in 2014: “The company assessed fatal myocardial infarction as related to (Actemra).” That company was Roche (RHHBY), Actemra’s manufacturer.
But neither Roche nor the FDA has moved to change Actemra’s label to alert patients and doctors that potential risks turned up in the reports, as well as in clinical studies completed after the drug went on the market.
Experts who examined the data at STAT’s request said the FDA should immediately consider warnings for heart failure and pancreatitis — an inflammation of the pancreas that in its acute form can kill up to 50 percent of patients. They said the evidence that Actemra might increase the risk of heart attacks, strokes, and interstitial lung disease, a sometimes-fatal scarring of lung tissue, is less convincing but warrants further review.
The failure to warn the public, experts say, highlights the FDA’s inability to adequately scrutinize the safety of drugs after they have been approved, and to act promptly when potential danger signs appear.
“We’ve done a very good job of making it easier to approve drugs, often based on very preliminary evidence. But we haven’t ramped up the standards of post-marketing surveillance to make sure that what’s been out there for several years is safe and effective,” said Dr. Vinay Prasad, an oncologist and medical ethicist at the Oregon Health and Science University. “The system is broken, and all the financial incentives are lined up to keep it broken.”
The FDA has been trying for years to strengthen its monitoring of drugs, so far without much success. In 2015, the Government Accountability Office reported that the “FDA lacks reliable, readily accessible data” needed for systematic oversight and to ensure that drug companies comply with agreements to track safety after a drug comes to market.
It has spent $207 million since 2009 to build a troubled big-data system called Sentinel that scours insurance company records for serious side effects of recently approved drugs. Among its major shortcomings, critics say: It’s missing most data on deaths related to prescription drugs. The agency would not say if a Sentinel assessment of Actemra, also known by the generic name tocilizumab, has ever been initiated.
Sentinel’s track record is all the more worrisome amid the rise of an approve-first, monitor-later philosophy in Washington. The recently passed 21st Century Cures Act is intended to speed federal sign-off on new medicines by streamlining pre-approval reviews of drug safety and efficacy. The Trump administration and prominent members of Congress are itching to push drugs through the regulatory pipeline faster than ever, then hope to catch any missed safety problems after they’re on the market.
Taken intravenously or by injection, Actemra has been used by more than 760,000 patients globally and generated sales of $1.7 billion last year, making it Roche’s fifth highest-grossing drug. While primarily used to treat rheumatoid arthritis — an autoimmune disease that causes pain, swelling, and stiffness in joints — doctors prescribe the drug “off-label” for about 60 other conditions for which it has not completed testing for efficacy and safety.
The FDA declined to comment about Actemra. In a written statement, a spokeswoman said the agency “continually monitors postmarketing safety of approved drug products and remains committed to informing the public in a timely manner when the FDA identifies safety issues.”
Dr. Jeffrey Siegel, senior medical director for rheumatology products at Roche and its subsidiary Genentech, said STAT’s examination of Actemra included “important questions that we think about all the time. … We try very hard not to be complacent, and to fully explore these issues.”
In an interview at Genentech’s headquarters in South San Francisco, he cited a recent study as “definitive” proof that the drug does not increase cardiovascular risk. But experts consulted by STAT disputed that claim.
In 2008, when FDA scientific advisers met at a Silver Spring, Md., hotel near the agency’s headquarters to debate whether to recommend approval of Actemra, they were haunted by a notorious error just a few years earlier.
Vioxx, another arthritis drug, had been pulled from the market after it was implicated in tens of thousands of heart attack deaths, a problem that hadn’t shown up in the short-term clinical trials used for approval. Those early studies suggested Vioxx would be safer for patients than existing medicines. The new drug, Actemra, similarly seemed relatively safe based on short-term studies.
“I can foresee the possibility that in five years there’s another hearing like the one on Vioxx, where the cardiologists … say to us, what were you guys thinking when you approved this drug?” Dr. David Felson, a Boston University rheumatologist, worried aloud, according to a transcript of the Actemra meeting. He described patient blood test data showing elevated levels of the blood lipids cholesterol and triglyceride, suggesting that Actemra might cause serious heart problems over time.
Felson and every other scientific adviser ultimately recommended approval, on one condition: Roche would sponsor multiyear studies to monitor for unseen problems and cardiovascular events. In the meantime, no mention of those possible hazards would appear on Actemra’s label — the key reference doctors and patients use to weigh a drug’s risks and benefits.
In a recent interview, Felson called STAT’s findings “noteworthy and concerning.” He said Siegel’s assertion that Actemra’s cardiovascular safety had been proved “sounds like a drug company trying to defend themselves.”
‘I just want it straight’
Actemra might not be more dangerous than other arthritis drugs — many of which can have lethal side effects — but patients and doctors are misled into believing it might be safer because frequently reported, serious problems aren’t noted on its warning label.
Alicia Airs of Boardman Township, Ohio, said her doctor told her that Actemra had minimal side effects when prescribing it for her rheumatoid arthritis. But Airs, 40, said she suffered heart palpitations for days immediately after her first Actemra infusion in April.
This symptom — often reported by other patients on social media and in complaints to the FDA — is not listed on the drug’s warning label. When she sought help, her rheumatologist told her he’d never heard of such a side effect and referred her to her general practitioner.
“When I went to the [general] doctor, they offered me an antidepressant, but I said I’m not anxious,” said Airs, still angry. “If you give them a symptom they don’t know, they treat you like you’re a little crazy … I felt like I was dismissed.”
Rheumatoid arthritis is different from the more common osteoarthritis, which is primarily a disease of old age. It often starts in middle age, but it also affects children and young adults like Alejandra Calzadillas, a 25-year old student in Seattle. She said that after starting Actemra, she experienced memory lapses and mental sluggishness she called “brain fog.”
“I’ve had moments where I’ve gone to start my car … and not remembered how to turn it on,” or at other times forgotten how to put on makeup, she said. Such cognitive side effects are not listed on the drug’s label but are a common complaint among Actemra users.
“It’s terrifying with a drug when you come to the realization that you haven’t been warned by your doctors,” Calzadillas said. “It kind of ruins your life.”
In interviews and reports to the FDA, patients described other unlabeled adverse events, including the heart-rhythm disorder tachycardia, small strokes, and tremors.
“I don’t want any sugar coating — I just want it straight,” Michelle Lucci, a 53-year-old banking consultant near Tampa, Fla., said in an interview. She suffered something else doctors don’t warn about: hair loss — another common complaint.
“Why would you use something that would do that when there are 15 other drugs to try? That was really depressing,” Lucci said. “[Actemra] has been a miracle for some people. But I think that on the whole, more people have serious side effects.”
Signals of harm
In a review of millions of reports to the FDA involving more than 100 drugs approved since 2010, Actemra stood out. It showed that Actemra patients had experienced an unusually large number of serious side effects that didn’t appear on the drug’s warning label.
The initial review was performed for STAT by Advera Health Analytics in Santa Rosa, Calif., a company that collects and curates drug-related complaints to the FDA Adverse Events Reporting System, known as FAERS. The company then provided comparison data for all major rheumatoid arthritis drugs.
STAT’s analysis of that data, including more than 13,500 FAERS reports on Actemra, showed higher than expected numbers of several serious problems when compared to competing drugs. These included the blockbusters Humira and Remicade, which have many more users.
For example, about the same number of cases of interstitial lung disease have been reported in Actemra users as have been experienced by Humira users, and many more cases than with Remicade. The other drugs’ labels warn about that possible side effect. Actemra’s does not.
The results were similar for heart attacks, strokes, and heart failure. None of these conditions appears on Actemra’s label, while both Humira’s and Remicade’s warn about heart attacks and heart failure, and Humira’s also lists stroke as a risk.
Pancreatitis was reported in 132 patients taking Actemra, including 26 who died.
“Pancreatitis is very, very rare” and potentially life-threatening, Felson said, so that many cases is a serious matter. The disorder is known to be triggered by high cholesterol levels, and Actemra is known to increase these lipids, he noted. “So if I see a signal for pancreatitis among [Actemra] users, I would be worried.”
The actual number of deaths among Actemra users is likely higher, because the FDA monitoring system captures a fraction of adverse events that patients experience. Patients and their doctors are not obligated to report to the FDA or to drug companies, and the agency and outside experts have estimated that these filings represent only about 10 percent of cases, although some say the rate might have increased in recent years.
Dr. Eric Brodsky, associate director of the agency’s drug labeling development team, said there are numerous other challenges in using FAERS data. Inaccuracies, concurrent illnesses, and effects of other drugs a patient may be taking mean FAERS reports can “show an association, but not a causal relationship,” he said — signals of harm, not proof.
Adding warnings based on uncertain evidence from FAERS, said Roche’s Siegel, “would be a disservice to clinicians, because it would raise the concern of something where there really isn’t a risk.”
One difficulty in assessing whether arthritis drugs cause cardiovascular problems is that the disease itself is a risk factor for heart disease. That’s where longer-term and larger post-marketing safety studies can help.
In one of these studies, required by the FDA, Actemra was compared head-to-head with another arthritis drug — Enbrel, whose label strongly cautions about use by patients with cardiovascular disease, particularly heart failure. That was the study Siegel called “definitive” evidence of Actemra’s safety.
It found rates of stroke and heart failure were about 1.5 times higher in Actemra patients. That difference wasn’t statistically significant, but experts told STAT it was still worrisome.
“Since Enbrel includes a high-profile warning and precaution in the label about heart failure, it is concerning that Actemra might be similar or worse,” said Dr. Steven Woloshin, a professor of medicine at The Dartmouth Institute and an expert on risk communication.
Felson concurred. While the actual risk remains statistically uncertain, he said, if Enbrel’s label warns for heart failure, the FDA should consider that warning for Actemra.
Dr. David Pisetsky, a professor at Duke University who served with Felson on the FDA advisory panel for Actemra, cited another recent study, based on thousands of insurance claims, which concluded there was no increased cardiovascular risk among patients on the drug. But the study found nearly identical rates of heart attacks and strokes among users of competing drugs — some of which are labeled for those problems.
“If the rates are similar, it should be on the (Actemra) warning label,” although the company and FDA might have additional data to consider, Pisetsky said.
There was a lone dissenter on the 2008 FDA advisory panel that recommended Actemra’s approval — the single nonexpert, consumer representative Diane Aronson. She called STAT’s findings “very troubling.”
“As a ‘no’ voter, I felt there wasn’t enough data, it was too short-term,” Aronson said in an interview, regarding the research supporting approval. “There were some red flags.” Others on the panel nonetheless supported the drug because they “really believe that the long-term studies will be acted upon” and the warning label adjusted if needed, she said. “That’s why they voted ‘yes.’”
That hasn’t happened.
Actemra has appeared on the FDA’s radar at least three times since it was approved in 2010, but most safety concerns have been set aside — once after Genentech pushed back against an agency recommendation to expand the warning label.
In 2011, “fatal anaphylaxis” was added to the label, at Roche’s request, after two deaths occurred.
A year later, Roche asked the FDA to allow wider use of Actemra, for patients who have failed to get relief from only one other type of rheumatoid arthritis drug, rather than two. In its review, the agency found 258 cases of pancreatitis and 185 of interstitial lung disease among Actemra users in clinical trials, FAERS, and epidemiological data, according to a 2012 report.
An FDA clinical expert said that despite uncertainty in the pancreatitis data, a warning should be included on the drug label “given the potential seriousness of the event.” The expert wrote that interstitial lung disease should be monitored closely.
“The system is broken, and all the financial incentives are lined up to keep it broken.”
Dr. Vinay Prasad, an oncologist at the Oregon Health and Science University
But Genentech convinced the agency not to label for pancreatitis, arguing that the illness, like interstitial lung disease, had occurred at expected rates for rheumatoid arthritis patients. Without providing specifics about the company’s arguments, the FDA report said there wasn’t “sufficient evidence to support causality and include information in the product label at this time.” The agency approved the broader use of Actemra.
In 2013, the FDA drew the same conclusion about pancreatitis in a full safety review of Actemra, required by law 18 months following approval or after 10,000 patients have taken a medicine. STAT obtained the report — heavily redacted to protect commercial secrets — under the Freedom of Information Act.
It addressed progress on ongoing studies and about 3,500 reports to FAERS that had arrived as of August 2012, including 118 deaths — 42 linked to cardiac arrest or myocardial infarction. The FDA concluded that no further label changes were warranted.
The report said the heart attack cases were confounded by such factors as age or a history of heart problems, and by rheumatoid arthritis itself. And it noted that rates of serious adverse events, including heart disorders, “have remained stable or decreased numerically over time.”
The agency said it would revisit the issue when a long-term cardiovascular safety study was completed. That was the study published last year that compared Actemra to Enbrel. The FDA and Genentech declined to say whether this latest data had been analyzed by the agency.
Federal law requires companies to ensure accurate labeling. The FDA can force changes when serious hazards come to light after approval. Companies that don’t comply can be fined up to $10 million. But an FDA spokesperson said the agency had never fined a company for labeling failures.
That might be because once a drug has been approved, “the leverage that the FDA has is infinitely smaller,” said Dr. Caleb Alexander, co-director of the Johns Hopkins Center for Drug Safety and Effectiveness.
“We are so reliant on a strong, impartial, and scientifically driven Food and Drug Administration,” he said, calling the agency’s unwillingness to talk about Actemra troublingly typical. “This is exactly the type of setting where greater transparency on the part of the FDA would be welcome” — ongoing evaluations of “extremely serious risks associated with specific products.”
Some critics point to a revolving door culture — with FDA scientists regularly moving to more lucrative jobs with the companies they once regulated — as another concern, saying it can foster a cozy relationship between the agency and the drug industry.
Siegel, for example, was an FDA manager who guided Actemra through its approval before leaving a few months later for Roche and Genentech to oversee further development of the drug. He said the timing was coincidental and unrelated to his review of Actemra for the FDA.
Federal rules prohibit him from representing his employer before the agency on specific projects he worked on as a government official, including Actemra’s use for rheumatoid arthritis, he said. But he does work with the FDA to gain approval for other uses of the drug, based partly on safety data he evaluated as a government official.
Abundant conflicts in safety studies
The FDA relies heavily on post-marketing studies to monitor drugs’ safety, research that is almost always funded and conducted by the companies making the medications. That was the case with Actemra.
“One of the recurring criticisms of our drug safety system is that we rely upon the very companies that are so financially invested in a product’s success to conduct the studies that explore safety concerns,” Alexander said.
All 11 authors of the Roche-funded study Siegel called “definitive” disclosed financial conflicts with either Roche or Genentech, including five who worked for these companies. Similarly, three of the seven authors of the insurance claims study, funded by Genentech, were company employees, and three others disclosed financial conflicts with Genentech.
The companies also pay consulting fees and offer perks to many other academic researchers, and fund leading rheumatoid arthritis registries that track patient outcomes over many years.
For example, they paid Dr. Joel Kremer, founder and chief medical officer of the Corrona registry — the largest such organization — about $130,000 for research, consulting, and travel from 2013 to 2015, according to the federal Open Payments database. Corrona, a Southborough, Mass., company owned primarily by a private equity firm, guards its data closely. Corrona CEO Raymond Hill would not identify the group’s scientific advisers, citing privacy concerns, and said its data are restricted primarily to paying customers, such as Genentech.
Dr. David Blumenthal, a Veterans Affairs rheumatologist and professor at Case Western Reserve University in Cleveland, said he doesn’t think the evidence of harm from Actemra is yet strong enough to warrant label changes. But Blumenthal, who also served on the FDA advisory committee before Actemra was approved, said he worries about manufacturers’ influence over research.
“The companies don’t want to kill the golden goose,” he said. “I always wonder whether industry is basically commissioning a paper, knowing what the results are going to be, to get a certain point of view out in the literature. In the world of politics, this is like the difference between real news and fake news.”
Prasad, of Oregon Health and Science University, argued that the FDA should design and manage large post-market trials, paid for but not influenced by the industry.
“So much of the data is in the hands of people who are conflicted,” he said. “We have to ask ourselves if the current system is in the interests of public health.”
For all the questions about Actemra, its market and Genentech’s profit potential keep growing.
In May, the FDA approved Actemra to treat giant cell arteritis, an inflammation of the blood vessels that, like rheumatoid arthritis, is an autoimmune disease. It based the decision on a one-year study by Genentech that involved just 149 patients who took the drug. In a press release, the FDA said the trial’s “overall safety profile … was generally consistent with the known safety profile of Actemra.”
Reporting contributed by Sheila Kaplan.