Can a Clinical Trial Go to Seed?


In most clinical trials the targets are patients, volunteers with a disease who sign up for a study to help advance medical knowledge and perhaps lead to better treatments for what ails them. But this week a report in the Archives of Internal Medicine revealed that sometimes the real targets are not so much the patients as the physicians who treat them.

The doctors who agree to participate as investigators in such a trial almost never find out that that they are the trial’s subjects, the ones being studied. Studies of this sort are known as “seeding” trials–kind of like seeding a cloud with particles in order to produce precipitation. The goals of a seeding trial extend beyond measuring the safety or efficacy of a drug to persuading – some would say hoodwinking – the hundreds of doctors who take part in the study to prescribe the drug and become rainmakers for a drug company. So much for informed consent.

Such matters “seldom see the light of day,” explained Caleb Alexander, associate professor of medicine at the Medical Center, who wrote a commentary to accompany the Archives paper, which was authored by researchers at Brown and Yale Universities. “One might think that seeding trials should be illegal,” he said. “They are unethical. They are not illegal.”

No one knows how often this happens. There is only one other recent, well-documented case and it required a lawsuit to come to light.

In this case, the lawsuit was triggered by accusations that the drug gabapentin – AKA Neurontin, approved by the FDA in 1994 for use to control seizures and in 2002 for nerve-related pain – was being marketed for off-label uses, such as for psychiatric disorders. The suit opened the vaults, allowing plaintiff’s lawyers broad access to the drug makers’ marketing plans, eventually resulting in hundreds of millions of dollars in fines.

One small part of the drug’s history was the STEPS trial, short for Study of Neurontin: Titrate to Effect, Profile of Safety. This was an uncontrolled, unblinded trial, launched after FDA approval. It enlisted 772 investigators, many with little or no clinical-trial experience, and 2,759 patients. Even before it began, there were questions about the design. After it was completed, there was the lawsuit, and then there were documents – more than 3,000 of them – correspondence, memos, presentations, market research. The researchers focused on about 400.

What they found met all the requirements for a seeding trial, as spelled out in a 1994 article in the New England Journal of Medicine (PDF) by former FDA commissioner David Kessler and colleagues. “Some company-sponsored trials of approved drugs appear to serve little or no scientific purpose,” they wrote. They are, instead, “thinly veiled attempts to entice doctors to prescribe a new drug being marketed by the company.”

Features that distinguish such trials include: a design that does not support the stated research goals, the recruitment of investigators not because they are experts or leading researchers but because they are frequent prescribers of competing products, disproportionately high payments given to ‘investigators’ for their work, sponsorship of the studies by the company’s sales and marketing division, minimal requirements for data, and the collection of data that are of little or no value to the company.

These were all hallmarks of the STEPS trial. Sales representatives, the authors note, were directly involved in collecting trial data. Investigators often lacked adequate training and were rewarded for enrolling extra patients. Adherence to the study protocol was lax. Patient recruitment, according to the Archives study, was used not only as an opportunity to promote gabapentin but also to block competing medications. “The data,” even by internal assessment, were “very dirty.”

Nevertheless, it worked. Physicians in the trial increased gabapentin prescriptions by 38 percent and dose by 10 percent. Neurologists who were not part of the trial did not increase their use of the drug.

Post-approval or phase 4 trials serve a vital role in examining the real-world safety and effectiveness of approved drugs, argued Alexander. “We need more phase 4 studies,” he insists, “but they should be well designed, executed with better intentions and backed by a compelling scientific rationale. Seeding trials threaten the integrity of these studies.”

“There are many ways to market a new medication,” he added. Pharmaceutical companies should select methods “that don’t involve duplicity or ethical breaches.”


Alexander GC (2011). Seeding Trials and the Subordination of Science: Comment on “Study of Neurontin: Titrate to Effect, Profile of Safety (STEPS) Trial”. Archives of internal medicine, 171 (12), 1107-8 PMID: 21709112

Krumholz SD, Egilman DS, & Ross JS (2011). Study of Neurontin: Titrate to Effect, Profile of Safety (STEPS) Trial: A Narrative Account of a Gabapentin Seeding Trial. Archives of internal medicine, 171 (12), 1100-7 PMID: 21709111

%d bloggers like this: