Therapies tailored for people's distinct genetic makeup have for years been touted as the next big thing in drug development. But if those drugs have yet to materialize, it may be because drug companies don't want them to.

A July report based on dozens of pharmaceutical company documents and filings with the Securities and Exchange Commission says companies often abandon drugs that would benefit only a genetic subset of people, and even obstruct research on genes that would predict drug response.

Companies are using pharmacogenomics to weed out drugs at an early stage. Adam Hedgecoe, University of Sussex

“There's lots of hand waving about how big companies are going to develop pharmacogenetic drugs,” says Adam Hedgecoe, coauthor of the report, published by the UK Pharmacogenetics Study Group.

“But companies are using pharmacogenomics to weed out drugs at an early stage,” he says. “[They] don't develop a drug that's only going to work for 20% of the population.”

In one example, a leading drug company refused to provide its product to academic researchers for a clinical trial assessing genetic and lifestyle factors that predict response to a drug for Alzheimer disease. The company only agreed after the researchers threatened to sue. Companies have also prevented scientists from publishing data that stratified drug responses by genetic factors, he says.

But drug makers maintain that they are trying to identify genetic factors that affect their products' effectiveness. “Critical generalizations are unfounded,” says Stephen Lederer, senior director of media relations for Pfizer. “When we have robust data, this is shared, as appropriate, with regulators and the scientific community.”

In March 2005 the FDA amended its guidelines to have companies voluntarily submit genetic data with new drug applications. Under these guidelines, companies can apply to broadly market a drug despite what the data might suggest. But if a drug is not approved, the FDA can deny companies that hold back genetic data the chance to reapply and include the data.

Since the guidelines have been introduced, companies have been more willing to disclose data linking certain genes to drug response, says Judes Poirier, director of the McGill Centre for Studies in Aging, who has for years worked closely with industry. The success of the tailored cancer therapies Herceptin and Gleevec has also shown industry that such drugs can earn significant profits, he adds.

In the absence of tests that can single out people who would most benefit from a drug, however, most drugs continue to be prescribed indiscriminately. Some biotechnology companies such as Third Wave Technologies in Wisconsin and Massachusetts-based Genzyme Genetics are trying to develop tests that can be marketed along with drugs.

Research has already shown that customizing drugs to genetic makeup can benefit public health. For instance, British researchers have shown that individuals with type 1 diabetes who have certain mutations in two genes respond poorly to insulin and fare better on sulfonylurea drugs (N. Engl. J. Med. 355, 467–477; 2006). Prompted by those results, the UK government has begun training a national network of nurses to implement genetic testing for diabetics.

“If policymakers are serious about applying the science to public health,” Hedgecoe says, “they need to engage with what is actually happening rather than what people hope companies are doing.”