WEST PALM BEACH, Fla. — While physical activity can be protective for developing knee or hip osteoarthritis (OA), too much vigorous activity may increase the risk, particularly for patients with a higher BMI, according to research presented at the World Congress on Osteoarthritis (OARSI) 2026 Annual Meeting.

“Exercise therapy constitutes a cornerstone of osteoarthritis management,” but the precise role of habitual physical activity on OA risk is inconsistent in the literature, Zhaohua Zhu, PhD, MS, rheumatology research fellow at the Royal North Shore Hospital, Kolling Institute, of the University of Sydney in Sydney, Australia, told attendees. “These findings highlight the importance of balancing PA [physical activity] volume with body size for effective osteoarthritis prevention.”

The researchers analyzed the relationship between accelerometer data and risk for incident knee or hip OA using data from 83,578 participants in the UK Biobank. The participants were an average of 55 years old, and 56% were women. None had knee or hip OA at baseline based on primary care and hospitalization records. 

The researchers converted the accelerometer data into average daily physical activity energy expenditure (PAEE), which should reflect both the duration and intensity of physical activity, measured in kilojoules per kilogram per day (note: 1 calorie = 4.184 kJ). They categorized the participants into three groups of low, moderate, and high PAEE at baseline.

For knee OA, low PAEE was less than 32 kJ/kg/d; moderate was between 32 and 42 kJ/kg/d; and high was above 42 kJ/kg/d. For hip OA, low was below 34 kJ/kg/d, moderate was between 34 and 44 kJ/kg/d, and high was above 44 kJ/kg/d. A total of 16,378 participants were classified as having low PAEE, 29,762 were classified as moderate, and 37,438 were classified as high.

Over a median follow-up of 6.7 years, there were 2069 incident cases of knee OA and 1551 incident cases of hip OA. Comparison of these cases with PAEE showed a U-shaped association. Risk for knee OA fell as PAEE increased up until 37 kJ/kg/d, the threshold with the lowest risk. Each single standard deviation increase in PAEE until then was associated with a 15% lower risk for knee OA (hazard ratio [HR], 0.85; P = .04) after accounting for participants’ age, sex, deprivation index, ethnicity, education level, diet, alcohol consumption, smoking status, nonsteroidal anti-inflammatory drug use, and history of joint injury. 

After 37 kJ/kg/d, the risk for knee OA gradually rose, with every single standard deviation associated with an 8% higher risk (HR, 1.08; P = .04). The researchers calculated the protective range of PAEE to be between 32 and 42 kJ/kg/d based on where the CIs overlapped.

The association between PAEE and incident hip OA was an L-shape, with risk remaining fairly stable at lower levels and then increasing sharply above 39 kJ/kg/d. Each standard deviation increase above that threshold was associated with a 16% higher risk for hip OA (HR, 1.16; P = .001), so the protective range was anything below that.

The researchers also looked at how BMI affected the associations. Participants who had a BMI between 25 and 30 and a moderate level of PAEE had a higher risk for knee OA than the reference group of individuals with a BMI below 25 and a low PAEE (HR, 2.69; 95% CI, 1.95-3.72). And those with a high level of PAEE had slightly more risk than those in the reference group (HR, 3.16; 95% CI, 2.29-4.35).

Meanwhile, participants with obesity (BMI of 30 or higher) who had a high PAEE had more than six times higher risk for knee OA (HR, 6.66; 95% CI, 4.78-9.28) than the reference group. This risk for people with obesity was attenuated for those with low (HR, 4.94) or moderate PAEE (HR, 5.48).

BMI also moderated the associations with hip OA. People with a BMI between 25 and 30 and moderate PAEE had a modestly higher risk for hip OA (HR, 1.56; 95% CI, 1.22-1.99) than the reference group, and people with obesity and high PAEE had double the risk (HR, 2.30; 95% CI, 1.69-3.13).

Types of Exercise Not Described

Neil Segal, MD, physical medicine and rehabilitation physician in the University of Kansas Health System in Kansas City, Kansas, expressed some skepticism about how to address the messaging from these data when the precise types of exercise were not described or delineated in the study.

“From a research standpoint, it’s really interesting, but the fact that the domain of exercise wasn’t included might be important for messaging for patients,” Segal told Medscape Medical News. “The sample size is huge, so it would be possible to know, Were they swimming? Were they running? It may be that, if you’re swimming, you can go above this limit, or the lower limit is actually higher to be protective.” 

Segal was not skeptical about the rigor or robustness of the findings but rather about the nuances of communicating the implications to patients.

“If you take it to an extreme, it is possible to be so active that you hurt your body, so I believe in the upper and lower limits,” Segal said. “But just from a messaging standpoint, telling people, ‘Oh, you’ve reached the threshold of 43 [kJ/kg/d of physical activity], so don’t exercise anymore’… I understand how they got that result, but it could be, ‘Don’t run anymore.’ It’s not going to be ‘don’t swim.’ Or, it could be ‘don’t run up mountains anymore.’”

No external funding was noted for the study. Zhu and Segal reported having no disclosures. 

Tara Haelle is a science/health journalist based in Dallas.