Wednesday, May 07, 2014

Some Authoritative Skepticism about the "Triumph" of Sovaldi to Add to Outrage about "Blood Money"

Three weeks ago we posted about the continuing controversy over the stratospheric price of the new antiviral drug for hepatitis C (HCV), sofosbuvir (Sovaldi, by Gilead).  While the amazing $1000 per pill price got deserved attention, at that time no one but your humble blogger seemed to be publicly skeptical about published assertions that the drug was some sort of modern miracle, and a triumph of medical science.  At that time, we noted that there is no "strong clear evidence that this drug is extremely effective and remarkably safe, i.e., that the drug is a 'triumph' that will cure nearly everyone without risk or harm to them, and therefore warrants a princely price."

Since then, two other skeptical opinions have become public.

The German Institute for Quality and Efficiency in Health Care (IQWiG) Early Benefit Assessment

The German agency performed this assessment based on a dossier submitted by the drug manufacturer (presumably Gilead).  The assessment found some reason to think the drug beneficial, but that the evidence was sparse, left many questions unanswered, and was inadequate to assess the drug for some important patient populations.  At this point, only a summary is available in English.  It includes links to further information in German.

The overall assessment was,

The dossier submitted by the drug manufacturer provides indications of added benefit for non-pretreated patients infected with the virus of genotype 2. However, the extent cannot be quantified. There were no suitable data in the dossier for patients who are infected with other virus types (genotype 1 and 3 to 6) or who are coinfected with HIV.
Limited Generalizability

Particular problems with data about genotypes other than 2 were:

The manufacturer presented no adequate analyses for infection with type 1 and type 3 to 6 viruses and for HIV coinfection. It analysed results from studies in which the respective comparator therapy was tested in at least one study arm and compared these in a 'historical' comparison.

It included both randomized controlled trials (RCTs) and one-arm studies on the sofosbuvir side, but only RCTs on the comparator side. It justified this by claiming that it wanted to reduce the number of hits of its literature search. However, the database for the comparison was different because of this and the comparison itself was therefore unsuitable. A first literature search by IQWiG showed that a number of studies were not considered in the dossier.
Thus the assessment concluded that the drug company dossier included at best irrelevant data that it tried to pass off  as important, and inexplicably left out other data that might have been relevant.

Problem with Random Allocation

The German assessment focused on the single randomized controlled trial which compared sofosbuvir and ribavirin to peg-interferon and ribavirin whose results were published in the New England Journal and discussed in our previous posts (here and here).   However, the German assessment analyzed what was submitted in the company's dossier about this trial, rather than the published results in the NEJM.  It found problems in addition to those we found, starting with problems with the randomization,

Overall, IQWiG assessed the risk of bias of the FISSION study as high. The main reason was that the manufacturer only included those participants in the analysis who had received at least one dose of the medication they were randomized to. However, particularly in the control arm, where not the new drug, but conventional drugs were administered, some patients refused to have their planned treatment.

This is a general problem of open-label studies, where it is known who receives which treatment. This poses the risk that patients discontinue the study prematurely depending on which treatment they were randomized to. This compromises the aim of randomization: the comparability of the treatment groups.  This is exactly what happened in the FISSION study, which may lead to (highly) biased results.

Thus, the assessment concluded that the study had a major problem with randomized allocation.  This should be added to the other problems we found: unblinded ascertainment of adverse effects potentially leading to bias, apparently deliberate obfuscation of adverse effects, especially deaths occurring during the trial, and lack of long term followup.  This suggests that the study was manipulated to increase the likelihood of favorable results

Failure to Consider Important Outcomes

The German assessment also concluded that the trial provided no information about mortality or quality of life benefits,

 As no deaths occurred in the FISSION study in the therapeutic indication genotype 2, there could be no differences between the treatment groups, and therefore no proof of added benefit, in the outcome 'mortality'. This also applies to health-related quality of life, but in this case because the dossier contained no evaluable data for this outcome.

Note, however, that this suggests that the data submitted in the dossier was different than that reported in the New England Journal of Medicine.  We noted that in the supplemental data from the trial available on the web appeared data on mortality.  However, it showed higher (albeit still very minimal) mortality in the sofosbuvir group than in the peg-interferon group. The fact that the dossier submitted in Germany omitted all mortality results should be added to the fact that the NEJM report buried the mortality data in the on-line supplement, increasing the likelihood that the dissemination of this part of the results was manipulated to forestall the questions that it ought to provoke.

The German assessment further noted that the dossier included no good data about possible harms or adverse effects of the new drug,

The data on side effects contained in the dossier could only be assessed to a limited extent. The manufacturer presented these data on the basis of the proportions of patients with at least one event. However, this type of analysis is only suitable to a limited extent because the observation period of the patients was different in the two study arms.

Note that is suggests another type of study manipulation, unequal assessment of harms in the study arms perhaps meant to minimize the apparent harms of the new drug.  

Furthermore, again the data in the dossier appeared discrepant from that in the NEJM article (which we discussed here),

 Severe adverse events only occurred once in each of the two study arms. A statistically significant difference in the outcome 'treatment discontinuation due to adverse events' in favour of sofosbuvir was not robust in the sensitivity analyses performed by IQWiG. With regard to side effects, IQWiG therefore considers the added benefit as not proven.

The severe adverse event rates reported in the New England Journal article were higher, and the reason for this discrepancy is unclear.  

A Non-Quantifiable Benefit?

So ultimately only a qualitative assessment of the drug's benefits for genotype 2 patients was possible,

 For treatment-naive patients with genotype 2 chronic hepatitis C, an overall positive effect of sofosbuvir in comparison with the appropriate comparator therapy remains with regard to serious secondary diseases. The extent of this added benefit – which was determined with the surrogate 'SVR' – is non-quantifiable, however, because it is unclear how often the development of liver cancer can in fact be prevented. 

Note that mention of the IQWiG assessment has been made so far in only one major US media outlet, Bloomberg.  

 The Institute for Clinical and Economic Review Evidence Report

This report just appeared online in JAMA Internal Medicine [Ollendorf DA, Tice JA et al.  The comparative clinical effectiveness and value of simeprevir and sofosbuvir in chronic hepatitis C viral infection.  JAMA Inte Med 2014.  Link here.] 

The review found again that there were major limitations in the available data, and so much of it was based on a network meta-analysis and simulation studies,

The review was limited to patients with genotype 1, 2, and 3 infections, which account for most HCV cases in the United States. At the time of the review, no head-to-head trials of the direct-acting antiviral drugs had been performed, and most of the data on sofosbuvir came from uncontrolled studies We therefore used network meta-analytic techniques to perform indirect comparisons of sustained virologic response outcomes across treatment options and used these results as the input to a simulation model that compared the estimated clinical and economic outcomes of treatment after 1 and 20 years.

It is very hard for me to independently evaluate the credibility of these indirect methods of analysis of what amounts to sketchy data.  However, I should note that Ollendorf et al did not some other specific problems with the available data on sofosbuvir:

Yet at the time of this review there were several reasons for caution. First, the evidence base on the comparative clinical effectiveness of these 2 new drugs remained thin in most areas and notably incomplete in others. For example, in some subgroups, such as patients who have failed earlier treatments, there were little or no data available. There were no long-term data demonstrating the durability of short-term sustained virologic response rates for either drug; nor had either drug been compared head-to-head in a randomized clinical trial with each other or with a first-generation direct-acting antiviral drug. Finally, the evidence necessary for sofosbuvir to gain marketing approval through the FDA’s breakthrough designation was particularly sparse and did not include requirements for controlled trials.

Furthermore, it suggested reduced enthusiasm for treatment since most patients with hepatitis C will not develop severe complications, limiting the potential benefits of treatment,

 HCV is an indolent infection, and most adults with this infection will never develop negative health effects during their lifetime. Many people with HCV infection have no evident liver dysfunction. For such patients, the decision about whether and when to initiate treatment should involve careful consideration of the balance of risks and benefits and shared decision-making with experienced clinicians.

Note that at least to date, I have seen no discussion of the Ollendorf report in the media.

Pricing Unrelated to Production or Research and Development Costs

Meanwhile, there has been continued outrage in the media over the pricing of this probably not so miraculous drug.  Some has focused on the likelihood that most of the price represents pure profit, excess marketing (and perhaps huge payments to corporate insiders), rather than production or research and development costs.  For example, Bernard Munos wrote in Forbes

To compound the outrage, there is hardly a relationship between drug prices and their actual costs, which makes many patients feel as if they were held to ransom. Gilead, to its credit, has negotiated a deal that makes Sovaldi available to Egyptian patients at a 99% discount, but it may have also bolstered suspicions of profiteering.

Similarly, Julian Urrutia wrote in the Bill of Health blog from the Harvard Law School Petrie-Flom Center,

Gilead is working to offer lower prices in dozens of developing countries (of the 150 million patients with HCV, only about 1% are in the USA).

To me, this implies one of two things. Either Gilead will be selling the drug in developing countries at a price that allows it to make normal profits (in which case Gilead is capturing enormous rents in the USA market from $84K price tag), or whoever pays for the treatment of patients with HCV in developed countries will be subsidizing the treatment of patients in the developing world.

 The Editor of Modern Healthcare, Merrill Goozner wrote, an analysis of how Gilead's immense revenue was built on US government funded research, and academics who could privatize the research they did with such funding on university time,

 The National Institutes of Health invested heavily in university-based scientists to understand the genetic weak points of hepatitis C, just as it did for HIV. It gave grants to build virology labs, come up with potential drugs and conduct clinical trials.

One of those scientists was Raymond Schinazi, director of the Laboratory of Biochemical Pharmacology at the Emory School of Medicine in Atlanta. During the past 20 years, Schinazi's laboratory received at least $7.7 million from the National Institute of Allergy and Infectious Diseases. He wasn't alone. Some, like Kevin Raney, chair of the biochemistry and molecular biology department at the University of Arkansas for Medical Sciences, even developed drugs on their government grants, although his never made it into clinical trials.

Schinazi, a serial entrepreneur, took a different path. In the late 1990s, he formed the pharmaceutical company Pharmasset, to develop an antiviral drug invented at Emory. The company also worked on cures for HIV and hepatitis B. The original drug candidate went nowhere, he told me in a telephone interview. But chemists working inside Pharmasset—at the time of its sale to Gilead in 2011 it had only 82 employees—developed several novel therapeutics for the diseases, including one that eventually became Sovaldi.

Securities and Exchange Commission filings for Pharmasset during its years as a stand-alone research and development company revealed it spent under $7 million on R&D in 2003. In its early years, it received $1 million in government grants. Its total losses through 2011 came to $314.8 million. It was also investigating drugs for HIV and hepatitis B throughout that period.

The company hit the jackpot with hepatitis C and Sovaldi though, which led Gilead, whose initial success came from HIV drugs, to pay $11 billion for the company. Schinazi, who had retained a sizable stake in the firm after cutting his ties in 2006, walked away with over $400 million, according to a published report.

Gilead spent tens of millions of dollars to complete Sovaldi's clinical trials. Yet last August, the NIAID issued a news release touting its role in running one of those trials when positive results were reported in JAMA. The government, it turns out, had kept its hand in the development of Sovaldi.

So it appears that the pricing of Sovaldi has almost nothing to do with production or research and development costs.  Much of the initial research needed to develop Sovaldi was underwritten by the government.  Gilead paid royally for the results of the research, but most of the money went to investors and executives, including the former academic who did some early development work on government money and became tremendously rich from Gilead's largesse.

Growing Outrage over Extortionate Behavior

Some has focused on how the price represents unethical, if not extortionate behavior by corporate executives.  For example, Munos also wrote,

What has changed is perhaps the fact that the industry no longer realizes – or accepts – that it has a covenant with society. This is no ordinary business as it can only operate by the will of the people. Drug makers get intellectual property rights, which are essentially a license to print money, but in exchange society expects affordable innovation. For most of the past century, that bargain has worked remarkably well. But, if drug companies fail to live up to it, society can also revisit its part, and scale back patent rights, or deny reimbursement. And this is happening in a rapidly-growing list of countries – not just India and China, but Germany, the UK, France, Canada, and soon… coming to a drugstore near you.

The legendary leaders who built the pharmaceutical industry as one of the most respected understood that bargain. One remembers George Merck's admonition ('Medicine is for the people. It is not for the profits'), Jonas Salk’s quaint remark about the polio vaccine patent, Robert Wood Johnson’s Credo, and Roy Vagelos’ more recent warning about the industry’s ill-conceived pricing policies.  They understood that self-policy is always better than regulation, and such moderation served them well. But their successors’ failure to uphold their advice has led to an onslaught of regulations and policy changes that threatens to permanently remake the industry.

Elisabeth Rosenthal wrote in the NY Times, quoting Steve Francesco, a pharmaceutical consultant,

'To understand drug pricing you have to shed your sense of value as a consumer and as a noble human being,' Mr. Francesco said. 'You have to put on the lens of the health care industry, where what you’re doing is looking for opportunities to maximize return.'

Also in Forbes, Dr Robert Pearl wrote this about Sovaldi pricing,

 But at $1000 a pill, its pricing is exorbitant, monopolistic, and disrespectful to the purchasers and patients who will bear the brunt of the massive cost.

Dr Frank Huyler wrote in the New York Daily News about blood money,

This sort of blood money is nothing new. But it is among the worst of recent examples; yet another evil act, yet another predation on mostly poor, mostly desperate people, who inevitably will ask taxpayers to save them.


It is even bloodier money if the assumption that the drug is a "well-tolerated and effective cure," which  Dr Huyler held, proves not to be true.  It is clear that most of the money that Gilead is now scooping up in the US is not to pay retrospectively for research and development or drug production. Instead, it seems likely to be supporting marketing, public relations, some investors' profits, and huge executive compensation.  When the public realizes that the money may not be buying miracles, the outrage should increase.  

The Sovaldi case is a signal example of how our health care system is awash in marketing hype and public relations buzz that has swamped rational skeptical thinking about logic and evidence.  That marketing and PR is ever enriching managers while it will send the rest of us, health care professionals included, to the poor house.  And all the money we spend will not buy us the promised miracles and triumphs.  

True health care reform would revisit the pact society once made with drug, biotechnology and device companies meant to promote reasonably priced innovation, but now promoting oligarchy; support transparency and honesty in clinical research; and challenge how health care managers can make millions or billions from unproven, and sometimes worthless or dangerous products.

1 comment:

Steve Lucas said...

Drug development cost has been a topic of discussion for some time. Pharma may include such cost as the initial government investment, opportunity cost of their investment, along with corporate overhead at the time of development to bolster their number.

Some multi-billion dollar estimates have to be discounted as they would exceed the profits of the corporation making it impossible to continue to exist.

Economically the total development cost is born by the very first pill produced and all following pills only incur the marginal cost of the ingredients. Thus supporting the statement of a businessman who said: “the worse day in pharma is better than the best day in any other business.”

Steve Lucas