Trial details
Patients were eligible for PRECISION, a Pfizer-funded trial, if they had osteoarthritis or rheumatoid arthritis and were at increased cardiovascular risk.
In all, 24,081 patients from 926 centers globally were randomly assigned to double-blind treatment with celecoxib, naproxen, or ibuprofen, all of which are now available as generics in the United States.
“Patients were randomized to the FDA-approved doses of these drugs, and they could have their dose increased if they had unrelieved pain up to the maximum allowed by regulatory authorities in the local jurisdictions where the study was done,” Dr. Nissen noted, pointing out that studies initially generating some concern about celecoxib used a supratherapeutic dose of 800 mg daily.
As COX-2 inhibitors are less likely than nonselective NSAIDs to cause ulcers, which might affect compliance and outcomes, all patients additionally received esomeprazole for gastroprotection “to try to level the playing field,” he explained.
The mean treatment duration was 20.3 months, and the mean follow-up duration was 34.1 months, according to data reported at the meeting and simultaneously published (N Engl J Med. 2016 Nov 13. doi: 10.1056/NEJMoa1611593). On average, the daily dose of drug received was 209 mg for celecoxib, 852 mg for naproxen, and 2,045 mg for ibuprofen.
In intention-to-treat analyses, the rate of the primary composite outcome of cardiovascular death (including hemorrhagic death), nonfatal myocardial infarction, or nonfatal stroke was 2.3% in the celecoxib group, 2.5% in the naproxen group, and 2.7% in the ibuprofen group. The hazard ratio was 0.93 for celecoxib versus naproxen (P less than .001 for noninferiority) and 0.85 for celecoxib versus ibuprofen (P less than .001 for noninferiority).
Differences were more marked in the on-treatment population. Here, the rate of the primary outcome was 1.7% in the celecoxib group, 1.8% in the naproxen group, and 1.9% in the ibuprofen group. The HR was 0.90 for celecoxib versus naproxen (P less than .001 for noninferiority) and 0.81 for celecoxib versus ibuprofen (P less than .001 for noninferiority).
Secondary outcomes, which tested for superiority, showed that the rate of major adverse cardiovascular events was 15% higher in the ibuprofen group as compared with the celecoxib group in the intent-to-treat population. The difference translated to a near-significant reduction in risk with the latter (HR, 0.87; P = .06) that appeared greater in the on-treatment population.
However, Dr. Nissen cautioned that he could not state that celecoxib was superior. “Secondary and tertiary endpoints in a clinical trial are hypothesis generating, and they are not considered definitive evidence,” he commented. At the same time, “the FDA is going to have to deal with that because what do they do with labeling? What do they do with over-the-counter access to these various drugs?”
The rate of all-cause mortality was 25% higher with naproxen than with celecoxib (HR, 0.80; P = .052).
The rate of gastrointestinal events was 54% higher with ibuprofen (HR, 0.65; P = .002) and 41% higher with naproxen (HR, 0.71; P = .01) as compared with celecoxib. And the rate of renal events was 64% higher with ibuprofen than with celecoxib (HR, 0.61; P = .004).
In a post hoc analysis of global safety, the rate of serious cardiovascular, gastrointestinal, and renal events was 28% higher with ibuprofen (HR, 0.78; P less than .001) and 15% higher with naproxen (HR, 0.87; P = .03) than with celecoxib.
Of note, the findings for the primary endpoint were similar regardless of whether patients were taking low-dose aspirin or not. “There was no interaction with aspirin use,” Dr. Nissen stated. “This was not about the interference of ibuprofen or naproxen with aspirin use.”
Analyses of pain relief using a visual analogue scale showed no clinically meaningful differences, suggesting that the drug doses used were equally analgesic, he said. Stopping of study drug because of lack of efficacy was slightly more common in the celecoxib group.
“We didn’t study the low-dose, intermittent use of these drugs that most of the public engages in, and it’s really important that we crisply communicate that to the public because somebody who takes occasionally ibuprofen or naproxen for a headache should not look at these comparative data in a way that should necessarily influence their behavior. We just don’t know the answer,” cautioned Dr. Nissen, who disclosed that he received grant support from Pfizer during the conduct of the trial.
But the findings are relevant for individuals who take over-the-counter NSAIDs at doses exceeding the label, a phenomenon known as dose creep. “We need to reemphasize to the public that the labeled over-the-counter dose is what you should take. You shouldn’t double up or triple up on the drugs because the issue of high-dose therapy, which is what we studied, suggests that there are really potentially important gastrointestinal, renal, and cardiovascular risks,” he said.