Bad seed? Critics say drug study more marketing than medicine

Merck says research was legitimate science

Many in IRB circles have worried about the potential existence of "seeding" trials, which are defined as clinical trials that seek to market a product rather than answer a legitimate scientific question.

Now a recent paper in the Annals of Internal Medicine has brought the issue front and center, accusing a 1999 clinical trial of the painkiller Vioxx of being a marketing-driven seeding trial.1 An accompanying editorial in the Annals describes the article as the "first to provide documentary evidence that proves the existence of seeding trials."2

The article references a study called ADVANTAGE (Assessment of Differences between Vioxx and Naproxen To Ascertain Gastrointestinal Tolerability and Effectiveness). The results of the study ran in the Annals in 2003. 3

The ADVANTAGE study's stated purpose was to compare the GI tolerability of Vioxx to naproxen. It was carried out in 600 sites among more than 5,500 osteoarthritis patients and found Vioxx to be as effective as naproxen, with a superior GI tolerability.

But Kevin Hill, MD, MHS, a clinical fellow in psychiatry at McLean Hospital in Belmont, MA, says the real purpose of the ADVANTAGE study was to get the drug, which was in the midst of being launched, into the hands of prescribing physicians and to promote its use. Hill and his colleagues point to documents they found when working as paid consultants to plaintiffs who sued Vioxx's maker, Merck & Co. Inc., over cardiovascular problems associated with the drug.

Merck has strongly denied that ADVANTAGE was a seeding trial, saying it was intended to answer important scientific questions regarding Vioxx. It notes the authors' work for plaintiffs' attorneys and calls the seeding trials article "biased" and "inaccurate" in a letter posted on one of its websites. (

In any case, the controversy has certainly drawn the attention of IRBs. Mark Schreiner, MD, chairman of the Committee for the Protection of Human Subjects at the Children's Hospital of Philadelphia, says he distributed a copy of the Hill article to all three of his institution's IRBs.

"I thought that was very revealing," Schreiner says. "The importance of this was that everybody has believed that this was going on."

Analyzing document database

Hill says that he and his colleagues weren't thinking about seeding trials when they analyzed a database of about a million documents that had been submitted by Merck during the discovery process of the Vioxx lawsuit. Their job was to prepare an expert witness for his testimony in the case.

"An important piece of the case was whether Merck failed to warn patients about the dangers of Vioxx," Hill says. "As a result, there was a lot of specific inquiry into Merck's practices in relation to Vioxx – how they proved the medication was effective, how they decided to market it, that sort of thing.

"We came across documents that led us to believe that Merck had conducted seeding trials, in particular ADVANTAGE, which involved Vioxx."

First, Hill's team searched the database using keywords such as "seeding trial" and "marketing," then individually reviewed about 2,000 documents that turned up in that search. They eventually identified about 100 documents that were relevant to the issue, mostly Merck internal correspondence, marketing memos and presentations.

Among them was a form nominating Merck marketing employees for their work on ADVANTAGE, stating that an objective of the study was to "provide product trial among a key physician group to accelerate uptake of Vioxx." It stated that the study was "designed and executed in the spirit of Merck marketing principles."

The authors also cited a memo by the head of the research division at Merck Research Laboratories, who criticized the ADVANTAGE study, calling it a "marketing clinical stud(y)" and "intellectually redundant." They noted a memo by a marketing employee that referred to ADVANTAGE and wrote, "It may be a seeding study, but let's not call it that in our internal documents."

Hill's group also pointed to the relatively large number of doctors and small number of patients per/physician (six) targeted by the ADVANTAGE study.

Hill and his colleagues say that the marketing aims of the ADVANTAGE study were not revealed to the FDA, to the physician-investigators who recruited patients for the study, to the institutional review boards that reviewed it and to the subjects who eventually enrolled in it.

Ironically, Hill says, the actual science behind the ADVANTAGE study was sound, although he says it was not as large or well-designed as the better known VIGOR (Vioxx GI Outcomes Research) trial, which also compared Vioxx to naproxen and revealed an increased risk for heart attack associated with use of Vioxx. But he says a study that masks its true objective may cause people to enroll who might not have done so if they'd known all the facts.

"I think it's quite clear that if patients were told that this was a trial designed by marketing and it was primarily a marketing trial, they would have made other choices about participating," he says. "People elected to place themselves at risk where they probably wouldn't have if they had known the objectives of the trial."

He says the most important factor in determining whether ADVANTAGE was a seeding trial was whether there was clinical equipoise, or true scientific disagreement about the question the trial was designed to answer.

"I would say for ADVANTAGE, the answer would be no," Hill says. "Because at the time of ADVANTAGE, Merck had already run trials with Vioxx looking at this idea of GI tolerability and they were in the process of running other trials, namely VIGOR which is probably a superior study in terms of design."

Merck defends ADVANTAGE

Jonathan Edelman, MD, executive director of the Global Center for Scientific Affairs, Merck Research Laboratories, Whitehouse Station, NJ, has defended the ADVANTAGE study in a letter to the Annals.4 He says that when the study first was conceived in 1998, the question as to whether Vioxx was superior to naproxen in GI tolerability was not yet settled, despite the fact that there had been a set of studies comparing Vioxx with various non-steroidal anti-inflammatory drugs (NSAIDS).

"The feedback that Merck got up to that point was that the information that was provided was insufficient for us to make a conclusion about the GI safety of this new mechanism of action for treating pain," he tells IRB Advisor. "There was good scientific reason to believe that the GI safety would be different from NSAIDS and it was necessary to demonstrate that through clinical research."

He says Merck conceived of two complementary trials, VIGOR and ADVANTAGE. VIGOR was devoted to GI hard outcomes such as peptic ulcers and bleeding.

"The ADVANTAGE study in concept was meant to answer the question: How does this new product Vioxx compare to the most commonly used NSAID in the commonly prescribed offices of primary care physicians?" Edelman says. "And that is all in complete contrast to the preceding studies and also to the VIGOR trial."

He notes that the Annals chose to publish the ADVANTAGE study because they saw value in the results that it produced.

Edelman says the study was designed and executed by the medical and scientific affairs department of Merck's U.S. Human Health Division, not its marketing department.

"It went through all the rigorous, routine reviews by Merck Research Laboratory committees," he says. "It was signed off on by the head of clinical research, by the head of biostatistics, by the head of regulatory affairs."

He says the nomination form cited in the seeding trials article merely praised marketing employees for "leveraging the research for marketing purposes."

"They chose to talk about it in marketing terms," Edelman says.

Similarly, he says the memo by Merck's research head was misinterpreted – he was referring to the cardiovascular results coming out of ADVANTAGE, not the question that prompted the study. Edelman says the memo calling ADVANTAGE a seeding trial was written by a public affairs employee who didn't understand the meaning of the term. "The fact that she wrote it is unfortunate but it doesn't make it so," he says.

Edelman noted that while the study originally targeted six patients per physician's office, the final numbers ended up being closer to 10 patients, which he says is a fairly typical number for a multi-site study.

And he says that while Merck clearly had commercial aims for Vioxx and ADVANTAGE did serve a marketing purpose, that didn't make it a seeding trial.

"The authors start out by saying that every bit of clinical research that's done to support an approved product is essentially a marketing study," Edelman says. "I think that's an unsophisticated and inflammatory way of talking about it.

"At the core of this issue is whether the primary reason for doing the trial is strictly commercial and there's no scientific benefit at all," he says. "In the case of the ADVANTAGE trial, there were legitimate, unanswered scientific questions that would influence the way potential prescribers would understand the data and make decisions about a new product."


  1. Hill KP, Ross JS, Egilman DS et al. The ADVANTAGE seeding trial: a review of internal documents. Ann Intern Med. 2008 Aug 19;149(4):251-8.
  2. Sox HC, Drummon R. Seeding Trials: Just Say "No" Editorial. Ann Intern Med 2008; 149(4): 279-280
  3. Lisse JR, Perlman M, Johansson G, et al. Gastrointestinal tolerability and effectiveness of rofecoxib versus naproxen in the treatment of osteoarthritis: a randomized, controlled trial. Ann Intern Med. 2003;139:539-46.
  4. Edelman JM. ADVANTAGE: Merck does say "no" Letter in Annals Online, Aug 22, 2008 at: