Data Manipulation and Incriminating Evidence About the Safety of Vioxx

ImagePhoto: Reuters/Mike Segar

“How many people is 100,000?…Imagine we’re talking about jetliners. If there were an average of 150 to 200 people on an aircraft, this range of 88,000 to 139,000 would be the rough equivalent of…two to four aircraft every week [dropping from the sky], week in, week out, for the past five years….”

–Dr. David Graham from the FDA Office of Drug Safety, speaking of deaths due to Vioxx in his testimony to the U.S. Senate Finance Committee on November 18, 2004

In September 2004, the pharmaceutical giant Merck withdrew its drug Vioxx from the market after five years, because it was shown to double the risk of heart attack and stroke in patients. Vioxx is in a class of drugs called COX-2 inhibitors that is used to treat arthritis and pain. Unlike COX-1 inhibitors such as ibuprofen, COX-2 inhibitors were prescribed for long term use because they cause fewer stomach ulcers and other gastric distress. Worldwide sales of Vioxx brought Merck nearly $2.4 billion, and was used by over 20 million people before it was pulled from the shelves.

By 2007, Merck had paid $4.85 billion to settle over 50,000 lawsuits, and in 2011 the U.S. Justice Department charged Merck $321.6 million in criminal fines and $628.4 million in a civil settlement agreement. In the end, Merck only had to plead guilty to marketing Vioxx as a treatment for rheumatoid arthritis prior to getting FDA approval, but it has consistently denied any other wrongdoing. Their payment into the settlement fund protects them from future class-action suits.

The basis of the many lawsuits was that Merck was aware of the dangers of Vioxx long before 2004. In fact, the earliest data of elevated risk of heart attack and stroke was evident in the results of its own study, called the VIGOR study, within the first year of receiving FDA approval. However, at that time, a Merck statistician and a lead scientist delayed the specific analysis of the data relating to heart problems. Documents and emails between Merck employees obtained during litigation (much of which is not publicly available), suggest that the VIGOR data revealed a high incidence of cardiac, vascular, or thoracic events (CVT) in patients given Vioxx as compared to those given Aleve. The results of VIGOR were even presented in the New England Journal of Medicine, but those results were presented as being ambiguous as to whether Vioxx caused increased CVT events, or whether Aleve somehow protected the study subjects from CVT events.

The Food and Drug Administration played a role in the unfolding nightmare. As background, how FDA approval works is, drug companies must go through laboratory and animal tests, and then three phases of clinical trials monitored by the FDA. The company sends all of its data to the FDA’s Center for Drug Evaluation and Research (CEDR). If the drug’s benefits outweigh the known risks, it is approved for market. The FDA Office of New Drugs then tracks the safety once the drug is available. Also, during clinical trials, it is standard for there to be a Data and Safety Monitoring Board (DSMB) to review all results.

Many events and factors contributed to what is possibly the world’s largest violation of the public trust concerning health. Systemically, we find that:

  • Drug companies administer and report on their own studies. DSMBs are usually involved, but they are not necessarily impartial. In this case, the chairman of the DSMB that was monitoring the VIGOR study was on Merck’s payroll and owned significant amounts of Merck stock.
  • Dr. David Graham, an associate director within the FDA Office of Safety Research and whistle-blower who had conducted an independent investigation into Vioxx, pointed out in his testimony to the U.S. Senate Finance Committee, there is an inherent conflict of interest that the FDA both approves drugs and then later is responsible for policing drug safety. When a drug is not safe it essentially finds itself in the position of having to prove itself wrong.
  • The FDA relies heavily on voluntary reporting from doctors when a drug is in the ‘post-marketing’ phase, so there is general under reporting of adverse events.
  • In the aftermath of the Vioxx scandal, the Journal of the American Medicinal Association, in a series of articles, found that nearly half of the post-marketing studies that drug companies commit to perform as a condition of drug approval do not happen.

In terms of Merck’s manipulation of information:

  • The VIGOR results published in the New England Journal of Medicine was written internally by Merck, but published under ghost authors.
  • Merck chose to eliminate data from the last two months of the VIGOR study that would have caused the results to accurately reflect the risks associated with Vioxx. In the relatively small sample sizes used in drug studies, a few incidents can be enormously significant. Note too, that what appear to be extremely small measures of statistical significance, amplified among millions of users, represents a lot of people who are potentially affected.
  • The data that Merck shared with the FDA was presented separately, making it appear less significant than if it had been shown in the aggregate, where it would have been obviously significant. It is normally standard to present all such data together, and there is some evidence that Merck did aggregate it for other purposes, but not for the FDA.
  • The FDA, presumably under pressure from Merck, told Dr. Graham to soften his findings and even tried to discredit him and prevent his findings being published in a major medical journal in the U.K. Further anonymous attacks were made when Dr. Graham sought protection from retaliation at the Government Accountability Project.
  • Other doctors and researchers were kicked off of FDA review committees after they spoke or wrote about the dangers of COX-2 inhibitors.

The Vioxx scandal and the subsequent judicial process played out over many years and permanently affected countless individuals and families. It also shed light on the weaknesses of the FDA, which were described as a broken system that is incapable of protecting the public from unsafe drugs. No doubt much of the reason for such protracted litigation was the process of e-discovery that involved insurance companies, Merck, and many third party care providers. Smart investigative practices allowed lawyers for the plaintiffs to obtain emails, primary data sources, and even outtake footage of a Merck corporate video that was circulated to the media as news. The video captures Merck employees and their paid experts discussing how to avoid interview topics that the company was ‘cagey’ about. But all of this was after the fact.

Critics and experts suggest that drug companies should not be allowed to design, conduct and report on their own drug trials, that DSMBs are not enough. But if drug companies are not going to pay for studies (and have the opportunity to influence the data), then who will bear that expense? Even though the cost of running drug trials is a mere drop in the bucket as compared to the generally overinflated costs of healthcare, it is still prohibitive for the FDA.

Public interest groups, such as Consumers Union Prescription for Change, in response to Vioxx and other drugs are launching. This grassroots initiative allows the public to learn about clinical trials and drug safety and effectiveness. Proactive changes to FDA policy and more robust peer review processes are also needed, to separate the profit motive from the FDA regulatory process and from the Data and Safety Monitoring Boards. There needs to be a cap on what review chairpersons and members can be compensated for their work, and they should not be allowed to hold stock in the drug companies that they review. The FDA has since established an internal appeals mechanism that allows employees to safely report on practices that might affect public safety.

It seems that in the case of Vioxx, destruction of records relevant to litigation was not an issue. In fact, Merck produced plenty of records for opposing parties that ultimately painted a clear picture of what happened. Rather, it was an outdated review system and too close a relationship between the regulatory body and the companies they are responsible for regulating. In the move toward centralized and digital health care records, there needs to be a corresponding information aggregation and optimization among research communities, the FDA, and the drug companies. Transparency needs to be guaranteed from multiple vantage points, including public interest groups. Transparency into one company, one review board, or even one regulatory agency is not enough.


The following references were used to write the above summary:

Center for Effective Government. (November 30, 2004). Critics diagnose systemic maladies of FDA [Blog]. Accessed March 1, 2014 from (November 18, 2004). Vioxx medication withdrawal: Testimony of Dr. David Graham to the U.S. Senate Finance Committee [Video]. Accessed March 1, 2014 from (November 22, 2011). Merck pays $950M to settle Vioxx charges [Web page]. Accessed March 1, 2014 from

O’Neil, C. (February 15, 2012). How big pharma cooks data: The case of Vioxx and heart disease [Blog]. Accessed March 1, 2014 from

Prakash, S. & Valentine, V. (November 10, 2007). Timeline: The rise and fall of Vioxx [Web page]. Accessed March 1, 2014 from

U.S. Food and Drug Administration (page last updated 4/4/2012). How FDA evaluates regulated products: drugs [Web page]. Accessed March 1, 2014 from