in a randomized trial.
The Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain score was 6.7 out of a possible total of 20 for patients who continued meloxicam for 4 weeks versus 7.8 in those who stopped and switched to a placebo. The estimated mean difference in pain score was 1.4 (P = .92 for noninferiority), which is below the threshold of 2.1 that is considered to be the minimum clinically important difference.
Furthermore, patients who had switched to placebo and then subsequently participated in a telephone-based cognitive behavior therapy (CBT) program for another 10 weeks had higher pain levels compared with those who continued meloxicam. WOMAC scores were 12.1 and 11.8, respectively with a mean difference of 0.8 (P = .28 for noninferiority).
“Among patients with knee osteoarthritis, placebo and CBT (after placebo) are inferior to meloxicam,” Liana Fraenkel, MD, MPH, of Yale University, New Haven, Conn., and coinvestigators concluded in their article, published in JAMA Internal Medicine.
They observed that the WOMAC pain score differences between the two groups were small, however, and that there were no statistically significant differences in participants’ global impression of change or function after 14 weeks.
“Although the overall results of the trial are negative, they provide clinicians with data to support shared decision-making and reassure patients willing to taper NSAIDs and consider self-management approaches such as CBT,” Dr. Fraenkel and coauthors suggested.
The Stopping NSAIDs for Arthritis Pain trial had ultimately included 364 participants, 86% of whom were men, recruited from four veterans affairs health care systems. All had been taking NSAIDs for knee OA pain for at least 3 months and had participated in a 2-week run-in period where the NSAID they had been taking was switched to meloxicam, 15 mg once daily.
The aim of the trial had been to see if discontinuing NSAIDs and starting a CBT program would be noninferior to continuing NSAIDs in patients with knee OA.
The trial does not provide robust information on the use of CBT, David Walsh, a rheumatologist and director of the Pain Centre Versus Arthritis at the University of Nottingham, England, said in an interview.
“It can’t tell you about efficacy of CBT,” Dr. Walsh said as the CBT part of the study was not randomized, was not controlled, and was unblinded. ”It would be a different task to design a CBT trial aiming to help people to stop taking tablets,” he added.
Dr. Fraenkel and coinvestigators had reported that, at week 14, the adjusted mean difference in WOMAC pain score between the placebo (followed by CBT) and meloxicam groups was 0.8 (P = .28 for noninferiority).
“What the trial’s really doing is seeing whether people who’ve been on long-term nonsteroidals, can they just stop them without getting any worse? The conclusion for that is actually they are more likely to get worse than not if you just stop the nonsteroidals,” Dr. Walsh said.
“The withdrawal trial protocol is an important one. You can’t run a prospective trial for years to see whether something works for years. It is just not feasible. So actually, the protocol they’ve got of switching to placebo, or continuing with a nonsteroidal, is probably the best way of working out if an anti-inflammatory still has a pharmacological effect after actually being on it for X years,” Dr. Walsh said.
Dr. Walsh, who was not involved in the trial, observed that while the difference in pain scores between the groups was small, the deterioration in scores might be important for individual patients. Some may do worse, although granted that there may be some that might do better, he said.
“It is suggesting to me that nonsteroidals are still working in people who are on long-term treatment. It is not a very big pharmacological effect, but we already know from the RCTs of anti-inflammatory tablets, that they can be beneficial,” Dr. Walsh noted.
He also pointed out that patients’ pain had been improved after being switched from their current NSAID to meloxicam – the overall WOMAC pain score at recruitment was 9.6 and was 5.6 after the 2-week meloxicam run-in phase.
“Now, whether that’s because they’ve been switched to meloxicam, or whether it’s because they’re in a trial,” is an important question, Dr. Walsh suggested, adding that “it looks as though it’s more likely to be because they’re in a trial, because improvement was maintained during the following 4 weeks on placebo.”
Another point he made was that there was a higher percentage of patients in the placebo group that started taking other types of painkillers, just under half (46%) used acetaminophen versus a quarter (26%) of those who continued using meloxicam.
It is an interesting trial, “trying to tackle some really difficult questions and I think that there are really important implications from it that we can build on, but is it actually going to change the lives of patients at the moment? Not massively,” Dr. Walsh said, ”but it’s another step in the right direction.”
Dr. Fraenkel disclosed receiving research funding from the VA Office of Research and Development, the sponsor of the trial.
SOURCE: Fraenkel L et al. JAMA Intern Med. 2020 Jul 20. doi:10.1001/jamainternmed.2020.2821.