Medical Device Dailys
Modern medicine bills itself as a science-based venture. But some of its critics allege that marketing alternately hypes science that helps its bottom line and hijacks science that does not.
In April, the open-access journal PLoS Medicine published a theme issue on “disease mongering,” which guest editors Ray Moynihan, a journalist, and David Henry, professor of pharmacology at the University of Newcastle in Australia, described as “the selling of sickness that widens the boundaries of illness and grows the markets for those who sell and deliver treatments.”
One of the assertions in Moynihan's and Henry's article is that mere risk factors are being turned into diseases, and treatment is not in line with what basic science supports. In the article “The Fight Against Disease Mongering: Generating Knowledge for Action,” Moynihan and Henry assert that it is disease mongering when “risk factors, such as high cholesterol and osteoporosis” are “framed as diseases.”
Henry elaborated on the issue of osteoporosis in an e-mail interview.
“Our basic ideas about prevention were based on contagious diseases where there is great specificity between the pathogen and the disease,” he told Medical Device Daily's sister publication, BioWorld Today. “This translates less well to 'risk factors' for chronic diseases.” Risk factors like bone mineral density (BMD) “are definitely associated with outcomes” such as fractures. “But they are not predictive. So, the majority of people with low BMD do not get fractures and the majority of people who get fractures do not have low BMD.
“This lack of specificity is a problem – many have to be treated so a few will benefit, which means that a significant number will experience adverse effects, all will carry the costs and a few will benefit,” Henry said. “We need much better predictors. But companies would rather we treated the many. The signal-to-noise ratios are very poor but the companies like this – they make a lot of money treating the noise.”
Willard Dere, chief medical officer at Amgen (Thousand Oaks, California), bristled at the notion that his company treats “noise” for profit. Amgen is in Phase III trials with denosumab, and antibody against RANK ligand that could be useful for preventing osteoporosis, as well as bone metastases, multiple myeloma and rheumatoid arthritis.
As for the idea that osteoporosis (which is defined as bone mineral density values that are 2.5 standard deviations below that of a normal 25 year old female) is a risk factor rather than an illness, Dere said, “I strongly believe that it is a disease.” He pointed out that the lifetime risk of a fracture for women is nearly 50% and that the annual 1.5 million fractures in the U.S., leading to 300,000 hospitalizations, have an enormous “human impact” as well as imposing an economic burden.
Asked to comment on low bone mineral density being a poor predictor of fractures, Dere said that the one-quarter of the population with the lowest bone mineral density has a risk of fractures that is 12 times that of the highest one-quarter. Dere conceded that predictors of absolute rather than relative risk would be desirable, allowing people to make better decisions about the relative risks and benefits of treatments compared to a fracture. But he did not agree that bone mineral density is a poor predictor; in comparative terms, he said, “bone mineral density stacks up as well against the risk of fracture as blood pressure stacks up against stroke.”
As for the final arbiter – the FDA – a spokesperson told BioWorld Today that it considers BMD “an imperfect surrogate.” Prior to accepting its use as a surrogate, the agency asks for preclinical proof that the bone effect leads to normal bone accretion. For a drug to be approved on BMD, the FDA said developers should have preliminary fracture data showing evidence of an effect, though not a final effect. An approval, in such cases, is generally limited to prevention of postmenopausal osteoporosis, with treatment given when the definitive fracture data are reviewed and accepted.
Once a drug is on the market, all promotional efforts are regulated by the FDA – although it does not require that ads be reviewed before public dissemination – to ensure that prescription drug information provided by firms is “truthful, balanced, and accurately communicated,” according to the Food, Drug, and Cosmetic Act.
Sometimes ads overstep those bounds; for osteoporosis, most of the warning letters on the FDA's web site appear to have gone to “natural” foods and supplements companies, who have made claims that everything from drinking cherry juice to spraying growth hormone onto the inner cheek “and holding for 90 seconds” can help prevent or cure osteoporosis.
The FDA has a number of potential remedies for those it thinks have stepped out of line. Among them are warning letters and untitled letters: Both request that companies immediately stop disseminating promotions cited as misleading. Should a company refuse to abide, the FDA can go a step further and take injunctive actions or refer the matter to the Department of Justice.
It rarely goes that far. Overall, the agency boasts a good track record of obtaining compliance with its letters.
As for surrogate endpoints, they are used throughout all phases of drug development, principally to shorten the time it takes a product to go from lab studies into human testing. In addition, a few are used for FDA approvals.
In some instances, the agency has based regular approvals on conclusions that such markers predict a product's benefit in a broader indication. That's the case with blood pressure and cholesterol for cardiovascular drugs, for example. For accelerated approvals, surrogate endpoints are used under the agency's Subpart H regulations, which permit clearance based on evidence from adequate and well-controlled studies of a drug's effects on a surrogate endpoint that reasonably suggest clinical benefits, or on the drug's effect on a clinical endpoint other than survival, provided the company agrees to a confirmatory study. Predictors used for such approvals include viral load for HIV and tumor response rate in cancer.
But not all publicized predictors are held to be absolute.
In fact, there have been cases in which surrogates' predictive values were later discounted and the benefits of drugs approved on those surrogates were brought into question. That happened with anti-arrhythmic drugs such as encanide and flecanide more than a decade ago. Initially, reducing ventricular premature beats after heart attacks was thought to predict a clinical benefit of fewer life-threatening arrhythmias. But that was later proved false when large studies showed that the use of the drugs actually decreased survival, so they are no longer used for that purpose.
Among drugs approved under accelerated approval on the basis of a surrogate is Iressa (gefitinib) – a recent example in which its clinical benefit was disproved in required follow-on trials. The product had been approved for lung cancer on the basis of a response rate on tumor size, but large clinical trials showed that it did not confer a survival advantage in those patients. Today it is used on a restricted basis in the U.S., and only as a second or third option.