Conflicting Data Leave Some Doubt About OA Role for GLP-1s

Knee osteoarthritis (KOA) appears to be among a long list of conditions improved by glucagon-like peptide 1 receptor agonists (GLP-1 RAs), but data suggesting possible adverse effects are giving some OA experts pause about long-term use.

One the one hand, data from several studies have demonstrated reductions in KOA endpoints associated with the use of GLP-1 RA drugs, for the most part correlated with significant weight loss. But there have also been hints of worsening OA with GLP-1 RA use, although the studies have examined the issue in different ways and over different time periods, making comparisons difficult.

“There’s a lot of heterogeneity in the literature in terms of the study populations, and what they’re comparing. So if you’re looking at subsets of populations, like those with obesity with or without diabetes or neither, you’re going to find subtly different findings depending on the study question. Results for GLP-1s may also differ for joint-specific disease vs global disease,” Atul F. Kamath, MD, the lead author of a 5-year study suggesting OA worsening with GLP-1RAs, told Medscape Medical News.

Kamath, who is professor of surgery at Cleveland Clinic’s Lerner College of Medicine, Cleveland, noted that his study, as others, examined use of the older first-generation GLP-1 RAs such as liraglutide and exenatide, rather than the newer semaglutide and tirzepatide. “The newer ones are much more potent than the first-generation ones that we studied. That could certainly contribute to the conflicting data.”

David T. Felson, MD, a rheumatologist and professor of medicine and epidemiology at Boston University, Boston, agreed. “The studies all ask somewhat different questions. So the answer, I think, is that semaglutide is really good for osteoarthritis, and one of the main reasons is that it causes a lot of weight loss and the earlier GLP-1 agonists frankly didn’t,” he told Medscape Medical News.

Felson is skeptical of the data suggesting OA harm, noting “I find it hard to believe that the GLP-1s make things worse.”

But another OA expert, Matlock Jeffries, MD, director of the Arthritis Research Unit at the Oklahoma Medical Research Foundation, Oklahoma City, told Medscape Medical News he’s not so sure. “We don’t know because it’s so early…And we don’t have a significant degree of animal studies on this to guide us in one direction or another.”

It’s plausible that the GLP-1 RAs have competing effects, Jeffries noted. “We know they result in significant weight loss, and the OA-specific effects of weight loss are well known. We know that it improves pain and maybe slows progression. A lot of that comes from the inflammation signaling that happens within fat tissue. However, we also know that these GLP-1 drugs over the longer-term result in loss of lean muscle as well probably some other metabolic changes, and I would guess that those might have a detrimental impact on arthritis risk in certain contexts.”

In addition, “the big thing that I worry about is that we don’t have great long-term data yet for people who stay on these drugs. It looks like the weight loss probably tapers off once you stop them. So presumably, they’re going to be used long term, and we just don’t know yet what will happen.”

The Data: The Good, the Bad, and the Uncertain

In Novo Nordisk’s 68-week double-blind, randomized, placebo-controlled 68-week STEP 9 trial of 407 people with obesity and KOA, once-weekly semaglutide significantly reduced KOA-associated pain, while also promoting significant weight loss. Those findings were published in October 2024 in The New England Journal of Medicine.

A Taiwanese population-based cohort study of more than 35,000 patients with type 2 diabetes (T2D) during 2011-2015 showed significantly lower rates of either receiving a KOA diagnosis or a total knee replacement among those taking GLP-1RAs over a 5-year period.

Yet another observational study, this one of more than 40,000 adults with both T2D and KOA in Shanghai during 2011-2017, those using GLP-1RAs for more than 2 years were significantly less likely to undergo knee surgery, again associated with greater weight loss compared with those not taking GLP-1RAs.

On the other hand, in January 2025 a large “discovery approach” study looking at 175 outcomes among nearly 2 million US Veterans with diabetes over a median of 3.68 years from October 2017 through 2023 showed that those initiating GLP-1 RAs experienced plenty of benefits compared with those starting other glucose-lowering agents. But some harms were also identified, including “joint pain and arthritic disorders.” That study was published in Nature Medicine.

And in Kamath’s 5-year analysis of more than 2 million people with obesity and/or T2D but without diagnosed OA at baseline during 2015-2017, the use of GLP-1 RAs was significantly associated with progression to hip and knee OA among those with T2D, but not among those without diabetes.

A Lot Left to Learn

Felson, who wrote an editorial accompanying the STEP 9 trial results, faulted Kamath’s 5-year OA progression study for evaluating prevalent GLP-1RA use rather than new users.

But he praised the Shanghai study, noting that although it also preceded the widespread use of the newer-generation GLP-1RAs, “it showed modest effects and also shows that those on GLP-1s had less cartilage loss, which is an important finding.” 

Felson also noted that all of these studies suffer from the fact that OA typically starts long before it’s diagnosed, making correlation difficult. “You can’t tell when osteoarthritis begins. It begins insidiously over many years, so even in studies where they say they excluded people with prevalent OA, they may have had it a long time before the doctor records it in the patient’s chart. An outcome to look at that’s really legitimate is the knee or hip replacement. In the [5-year OA progression] study, there was no effect on knee or hip replacements.”

Felson noted that the Nature Medicine “discovery” study also didn’t examine those hard endpoints — possibly due to low numbers — but was impressive nonetheless. “They looked at all these different disease categories and conditions, and did a really nice job. It’s hypothesis-generating…Each of these conditions that they identify probably needs further exploration.”

As for the suggestion from some that weight loss might lead to people becoming more active, which could lead to joint injury, both Felson and Jeffries countered that there’s little evidence to support it.

Both Kamath and Jeffries cited the need for longer-term follow-up. Regarding the STEP-9 trial, Kamath said, “I would argue you can’t make any definitive conclusions about OA natural history based on a year and change. I would say this field needs to be tempered by exploring those subtleties on longer time frames in larger control populations before we can definitively say we’re right on the money. Our study serves to temper some of the over-exuberant excitement about the wonder drugs that they’re purported to be.”

His group is currently researching several aspects of this. “We’re now trying to tease out some of the subtleties, like relative weight loss over time, the durability of the weight loss, and then what is the effect on arthritis progression, conversion, and arthroplasty…We just need to keep chipping away at this.”

Clinically, Jeffries said, “I cautiously tell patients that a [GLP-1 RA] will probably help them lose weight, and we have short term data that it’s going to help the pain but that we don’t know what’s going to happen longer term…I have seen a decent number of patients that have been on these drugs lose a lot of weight, and they do have improvements in their OA pain. But of course, in clinic I’m only dealing with the short-term.” 

Kamath had received royalties/licenses from Innomed; consulting fees from Zimmer Biomet, Bodycad, Ortho Development, and United Ortho; and honoraria from Zimmer Biomet. Felson is a consultant for AposHealth. Jeffries had no disclosures.

Miriam E. Tucker is a freelance journalist based in the Washington, DC, area. She is a regular contributor to Medscape Medical News, with other work appearing in the Washington Post, NPR’s Shots blog, and Diatribe. She is on X @MiriamETucker and BlueSky @miriametucker.bsky.social.