An experimental cancer drug shrinks tumours in rats with no apparent side effects. The scientists behind the study plan to do a clinical trial in humans, but it could take years to complete. Meanwhile, dying patients begin taking the unapproved drug and collect their results on the web. Both groups desperately want to save lives: but which is the right route to follow?

This scenario has been playing out in recent weeks for a compound called dichloroacetate (DCA). It taps into long-running issues about whether terminally ill patients should be able to get access to drugs that have not yet had formal approval. Researchers fear that those taking the drug could suffer unanticipated side effects; patients argue they don't have the luxury of waiting for clinical trials to find out.

In January this year, Evangelos Michelakis at the University of Alberta in Edmonton, Canada, and his colleagues reported that DCA has seemingly remarkable anticancer properties (S. Bonnet et al. Cancer Cell 11, 37–51; 2007). DCA is a small molecule that blocks an enzyme in mitochondria — the energy-production centres in cells — causing more glucose to be metabolized in the mitochondria rather than by a different pathway in the cytoplasm. The compound has been in clinical trials for years as a treatment for certain mitochondrial diseases, but it has not yet been approved.

Mitochondria also control cell suicide, and Michelakis wondered whether cancer cells were suppressing these cellular structures to prevent the cells from dying — and so thought DCA might reactivate them. When his team gave DCA to rats that were growing human lung tumours, the tumours stopped growing within a week, and three months later were half the size of those in untreated animals. Other experimental drugs have had similar effects. But DCA stands out because it seems to leave healthy cells untouched, has been relatively safe in human trials, can be taken by mouth and easily penetrates tissues. “If there were a magic bullet,” wrote Newsweek about the discovery, “it might be something like dichloroacetate.”

Reactivating mitochondria seems to trigger cancer cells to commit suicide. Credit: E. MICHELAKIS

Because DCA has been around for years, its structure can't be patented and Michelakis found that pharmaceutical companies weren't interested in developing the drug. So he is raising money and hopes to start his own small clinical trial within the next few months.

In the meantime Jim Tassano, who owns a pest-control and marketing company in Sonora, California, came across DCA when researching alternative cancer therapies to help his dying ballroom-dance instructor. He wanted something that was effective, safe and that he could lay his hands on: DCA fit the bill. He ordered some from chemical supply companies, teamed up with a chemist friend and they worked out a way to synthesize the compound themselves. “I couldn't walk away from it,” Tassano says. “It could do so much good for so many people.”

Tassano set up two websites. The first of these ( hosts information on DCA and a patient chatroom. On the second ( Tassano is selling his home-made DCA — labelled for veterinary use, as drugs sold for human use in the United States must have approval from the Food and Drug Administration (FDA). Tassano says he is sure patients are buying the drug to use on themselves, and reckons that a couple of hundred of people from around the world have bought from the site. Many patients taking DCA — acquired from Tassano, chemical companies or other sources — are reporting their progress on

Some of these patients plan to set up a database on Tassano's website to collect DCA results in a more organized way. They want people to submit information including the type of cancer they suffer, medical history and the dose they are taking, says Susan Hirasawa in Seattle, Washington, who suffers from late-stage breast cancer and is one of the organizers. The idea is to provide information for others who want to take DCA, she says, but “it's not a real clinical trial”.

It's destroying efforts to do this right. Any way you look at this, it's a negative development.

Michelakis and other researchers are worried by the development. Although DCA seems safe overall, they point to a clinical trial that was stopped early because those taking the drug developed damage to their peripheral nerves (P. Kaufmann et al. Neurology 66, 324–330; 2006). Without a control group, they point out, it will be impossible to tell whether any improvement in the patients' condition is caused by the drug. Patients could also be taking DCA that is not of pharmaceutical grade and might contain harmful impurities.

Michelakis says the patients could end up undermining efforts to do a controlled clinical trial if, for example, some develop harmful side effects and the drug earns a bad reputation. “It's destroying efforts to do this right,” he says. “Any way you look at this, it's a negative development.” An FDA spokesperson told Nature that the agency is looking into the matter.

The battle between dying patients who want immediate access to unapproved drugs and doctors who urge trials and caution is a perennial one. Some patients argue that they cannot wait for trials and should have the right to take unapproved drugs, regardless of the risks.

But there are arguments against this. An estimated 95% of cancer drugs that enter clinical trials do not get approval, many because they are ineffective or unsafe, so patients risk shortening their life or making their last days more uncomfortable. “They say what do I have to lose? The truth of the matter is, you have the rest of your life to lose,” says George Annas, an expert in bioethics at Boston University School of Public Health.

And if patients can access DCA — or other unapproved drugs — there is no incentive for them to enter a clinical trial. So in terms of public health, ethicists argue, more people will be helped if access to unapproved drugs is restricted and proper trials performed.

Peter Jacobsen, an expert in ethics, health and law at the University of Michigan in Ann Arbor, doubts whether any good can come of the patients' efforts. They are so desperate to see results, he says, that there is no way they can report unbiased results and no mechanism to ensure the reports are accurate. “I don't trust the data,” he says. “It's hard enough to rely on them in clinical trials, let alone this.”