Earlier this week at the American Heart Association Scientific Sessions 2016 in New Orleans, researchers presented data – which was simultaneously published in the New England Journal of Medicine – from the PRECISION trial, which evaluated the impacts of celecoxib – commonly known as Celebrex – on risk of heart attack and stroke as compared to prescription doses of other anti-inflammatory drugs, naproxen and ibuprofen. While the study found celecoxib to be just as safe, if not safer for patients with osteoarthritis and rheumatoid arthritis, some believe there is little to no evidence in the paper to substantiate this conclusion.
News coverage of the results ranged from a more straightforward account of the data, as reported by the New York Times, to more robust pieces, such as that in Forbes, which addressed the questions and limitations of the study.
Over a 10-year span the PRECISION trial followed more than 24,000 participants with arthritis who were randomly assigned to take either Celebrex, ibuprofen, or naproxen (an over-the-counter painkiller sold under brand names including Aleve), and evaluated the rates of heart attacks and risk of stroke, as well as other effects such as anemia caused by bleeding in the gastrointestinal tract and risk of kidney problems or high blood pressure. At the conclusion of the study, more than 16,5000 patients had stopped taking their prescribed medication, and researchers had lost track of more than 6,500 participants – over 25 percent of those enrolled. When the results were tabulated, Forbes reporter Matthew Herper said “the problem is that the study is flawed in ways that some scientists think may make those results meaningless.”
One of the most influential scientists Herper is referring to is Garret A. FitzGerald, MD, FRS, director of the Institute for Translational Medicine and Therapeutics at Penn Medicine, who has spent decades researching NSAIDs and their effect on cardiovascular disease, and who authored an editorial which was published in Circulation detailing what he says are extensive limitations and shortcomings of the data.
“Unfortunately, there are many sources of bias that favor Celebrex and interact with each other to undermine our ability to draw useful conclusions from this study,” FitzGerald said. Chief among the biases he cites as impacting the quality of the results include, the failure to compare equally effective doses amongst the drugs and the remarkably high dropout rate, those who were not followed for the duration of the study, and those who switched to different drugs during the course of the research
All of these factors, FitzGerald says “undermines the validity of this study as a cardiovascular outcome trial.”
There is a trend among the points listed, of which many revolve around the overall study design, which led FitzGerald and others in the field to question this why this study was ever initiated.
“Given that we had clear evidence for a hazard from Celebrex from placebo controlled trials before it started, the question was whether performing such a large comparative trial like this was a good idea in the first place,” he said. “It was predictable that doctors would not put their high-risk patients with an established cardiac risk into such a trial, and that researchers would wind up without the information that it was designed to address – and such issues were discussed publicly before the first patient was recruited for this trial.”
FitzGerald continued, “I feel sorry for the wasted effort on the part of the patients who enrolled in the trial, and of course the effort of many who were involved in more than a decade-long research project which yielded, in my opinion, inconsequential data. As long as study went on, the argument around the relative uncertainty about the cardiovascular risk of NSAIDs was being made in direct-to-consumer advertising, while sales of Celebrex averaged two and a half to three billion dollars a year. Indeed, despite the serious limitations of PRECISION, the study is now being touted as "Study Findings Dispel Long Standing Perceptions of Excess Cardiovascular Risk Associated With Long Term Use Of Celebrex" on the sponsor’s website, with the inference that Celebrex is devoid of the same risk profile of Vioxx.”
Vioxx is the blockbuster drug which was first approved for the treatment of arthritis pain in1999, along with Celebrex. Both drugs worked by blocking the enzyme cyclooxygenase-2 (cox-2) to treat inflammation without some of the traditional side effects of ibuprofen and naproxen, which also blocked cyclooxygenase-1 (cox-1), causing gastrointestinal damage including GI bleeding. It was later found that Vioxx was linked to greater risk of heart attack compared to naproxen, it was removed from the market in 2004. While there are no direct comparisons between Celebrex and Vioxx, an overview analysis led by Oxford researchers demonstrated that at equally effective dose, the cardiovascular risk Celebrex or Vioxx were nearly identical, FitzGerald said.
“Sadly, I believe this new study provides precisely no additional useful information to guide clinical practice,” FitzGerald said.
So another question remains, if the study design and the data won’t change the prescriptions doctors write for patients still dealing with arthritis pain, what is next for both clinicians and researchers?
“I think we need to move away from commercially driven, poorly designed large trials to try to understand the factors that underlie variability in drug response,” said FitzGerald. “We can then parse differences between patients, determine which factors render them susceptible to cardiovascular risk from NSAIDs, and identify a combination of biomarkers that reflect that risk and seek to validate their utility in risk prediction.”
In other words, FitzGerald concluded, “we should be trying to bring more precision to the way we deploy these medicines.”