When the pharmaceutical company Gilead launched its hepatitis C drug sofosbuvir (Sovaldi) in 2013, it justified the $1,000 per-pill price by pointing to ironclad data on its life-saving potential. Chiefly, clinical trials clearly demonstrated the drug's ability to "cure" more than 90% of patients, as measured by a widely accepted surrogate endpoint called sustained virological response (SVR).
But earlier this month, a pooled analysis by the prestigious Cochrane Collaboration questioned whether sofosbuvir and other direct-acting antivirals (DAAs) can actually save lives in the long run -- a finding that could deal a massive blow to hepatitis C patients and drug companies alike.
The problem is that it's simply not true, according to a chorus of experts who have called the Cochrane review "outdated," "fundamentally flawed," and "absurd."
In an extraordinary show of unity, hepatitis C researchers, clinicians, and advocates have denounced the Cochrane analysis and its implications, pointing to clear-cut evidence that current DAAs not only cure hepatitis C, but also reduce the risk of liver transplant, cancer, and other complications later in life.
Cochrane, meanwhile, stands by its analysis and has declined a request by the World Hepatitis Alliance to clarify its results.
DAAs "are marketed as a revolution," so it's "natural" that Cochrane would be interested, said first author Janus Jakobsen, MD, PhD, chief physician at the Copenhagen Trial Unit at Rigshospitalet in Denmark.
In the 744-page meta-analysis, Jakobsen and his group reviewed 138 trials assessing the effects of 51 different DAAs. They disputed the value of SVR as a clinically meaningful endpoint, and noted that all studies had a "high risk of bias" because of industry funding.
In sum, "the results indicated that [DAAs] do not have any clinical effect," said Jakobsen. "Regarding the long-term effects, we are totally in the dark," he added.
Jakobsen's results sparked backlash from hepatitis C experts and advocates, including the World Hepatitis Alliance, the Hepatitis C Trust, the American Association for the Study of Liver Diseases (AASLD), and the Infectious Diseases Society of America (IDSA).
Critics of the Cochrane review argue that SVR is a proven surrogate endpoint, that the studies included were not designed to determine long-term benefits and that the analysis should not have included drugs that have since been withdrawn from the market, including boceprevir (Victrelis) and telaprevir (Incivek).
It is an "irresponsible interpretation of the data" that carries "so-called evidence based medicine to an illogical extreme," said Douglas Dieterich, MD, of Mount Sinai Hospital in New York.
"My immediate reaction was extreme anger because this is clearly poor science," said Charles Gore, President of the World Hepatitis Alliance, who contributed to rebuttals printed in The Guardian and The Lancet. "Their research says there is no evidence that DAAs work, but that is because they looked at research that was not designed to show they work in their terms."
Chief among criticisms is that the Cochrane paper characterized the relationship between SVR and mortality as "questionable." The FDA and other drug regulators approved DAAs based on the idea that SVR, a surrogate endpoint, correlates to other more tangible benefits such as overall mortality and morbidity.
"They're the only ones in the world that I know of that don't accept SVR," said Paul Pockros, MD, director of the Liver Disease Center at Scripps Clinic in La Jolla, Calif. Pockros pointed to a 2012 JAMA paper demonstrating that SVR among patients treated with the older generation of interferon-based treatments correlated with lower all-cause mortality after an average follow-up of 8.4 years.
"An SVR with DAAs, in principle, would have the same benefits as interferon," said Harry Janssen, PhD, MD, of Toronto General Hospital Research Institute. " The only difference is with these drugs which have no side effects, we are able to treat very sick patients," said Janssen, who was a co-author of the 2012 JAMA paper.
Jakobsen asserted that the correlation between SVR and all-cause mortality for DAAs has still not been established. As well, it's "an assumption" to say that SVR for interferon-based regimens is the same as SVR for DAA-based treatment, he said.
Gore called this "splitting hairs." The Cochrane Collaboration is essentially claiming "the FDA doesn't know what it is talking about," he said.
Another critique of the Cochrane study was that it included 57 trials on prior generations of DAAs that have since been withdrawn or discontinued. However, Jakobsen notes that "we reported the results separately, so we emphasized the results of the drugs that are still on the market." Still, the review found "no evidence of a difference" in morbidity and mortality for both currently approved and discontinued drugs.
In addition, because the most effective DAAs were approved after 2013, much of the data indicating their life-extending properties has not yet been published, critics argue. Because hepatitis C is such a slow-moving disease, the most immediate effects will be seen among the sickest patients, such as those with decompensated cirrhosis and those on the liver transplant list.
In 2015 the introduction of DAAs was associated with an 11% decrease in hepatitis C-related deaths, according to Public Health England. In addition, a dramatic number of patients with decompensated cirrhosis have been taken off liver transplant lists since the introduction of DAAs in the U.S., said Pockros, citing data from an abstract presented at the AASLD meeting last November.
Jakobsen, meanwhile, noted non-randomized trials do not meet Cochrane's evidence standards. In addition, some of the critics have contradicted themselves, he said.
"They say we need 30 years [of data], but at the same time they say they can see in the data that it already works," Jakobsen said.
"I don't see a contradiction," said Pockros. "I just think they're a little early. If they come back and do it again in 2020, that would be right."
For Pockros this recent controversy around DAAs echos a 2008 Cochrane analysis in which a drug called ursodiol was found to have no demonstrated mortality benefit for patients with primary biliary cirrhosis (PBC), a liver condition that was the leading indication for liver transplant in the 1990s.
"When that study came out with PBC, I went ballistic," said Pockros. "We had started using this drug for this disease in the late 1990s, and by 2008 it had basically disappeared. So why didn't the Cochrane review show that? They didn't have enough follow-up data on patients."
"I'm not saying they are wrong," he said. "They are just extremely bound by published literature that they deem acceptable for review."
"A lot of people are irritated, and rightly so because these people do not know what they are talking about and they're causing a lot of damage," said Janssen. "What they've been doing is very dangerous, because there will be people who read this and say, 'well I don't need any treatment.'"
Yet Jakobsen maintains that there's not enough evidence to support the use of DAAs: "It becomes a question of faith," he said, "and I don't think we can use that to treat patients."