Yoga for Knee Arthritis?

Randomized Trial Tests Yoga as Knee Osteoarthritis Therapy

— Results from 6-month study suggest it’s a reasonable option
by John Gever, Contributing Writer, MedPage Today
April 8, 2025

Patients with osteoarthritis (OA) of the knee got at least as much benefit from yoga as from conventional strength training in a randomized trial, researchers said.

Pain was reduced about equally with yoga and strength training in the 117-person study, and yoga proved superior by certain other measures, according to Benny Antony, PhD, of the University of Tasmania in Hobart, Australia, and colleagues.

Overall, the results indicated that yoga was noninferior to strength training and “that integrating yoga as an alternative or complementary exercise option in clinical practice may help in managing knee OA,” the group reported in JAMA Network Open.

Exercise therapy has long been recommended for patients with knee OA, as many studies have shown that it can reduce pain and delay the need for invasive treatment such as arthroplasty. Yoga is well recognized as a good whole-body workout and is recommended in published guidelines. But the same guidelines noted that the evidence base is extremely weak, because the underlying studies had problems such as small samples, short follow-up, and unclear comparators. Thus far, Antony and colleagues wrote, no studies had compared yoga to strength training, the exercise modality with the best support.

For their trial, the researchers initially recruited 129 knee OA patients in the Australian island of Tasmania, 12 of whom were excluded or withdrew prior to the study’s start; 58 were then randomized to yoga and 59 to strength training. Mean ages were 61 in the yoga group and 64 among those assigned to strength training. More than 70% were women. Body mass index values averaged 29 and 28 in the yoga and strength training groups, respectively. Pain was self-rated on a 100-point scale, with baseline averages of 54 in the yoga arm and 53 in the strength training group.

Both programs were delivered in two hourlong in-person group sessions plus one at-home session each week for 12 weeks, with an additional 12 weeks of thrice-weekly at-home sessions. In-person training was given to groups of 10, each led by a trained professional. Strength training was centered on 45 minutes of various leg exercises, some with elastic bands and weights, conducted like a so-called circuit class. The yoga program was designed specifically for knee OA patients. Sessions began with chanting and then rotated through sun salutations and asana and savasana poses (standing and floor-supported); some also utilized elastic bands.

The interventions in both groups lasted 24 weeks, with the primary endpoint — between-group difference in self-reported pain — assessed at week 12. Secondary outcomes included pain at week 24 and measures from the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Osteoarthritis Research Society International system, and other OA evaluation scales at weeks 12 and 24.

Thirteen other outcome measures were assessed at weeks 12 and 24, for a total of 26 secondary endpoints. Of these, seven showed a significant advantage for yoga; none favored strength training (no statistical corrections were made, however, for the multiple comparisons). Most of these differences were seen at week 24: WOMAC pain, function, and stiffness; patients’ global self-assessment; a “utility score” for overall health; and a 40-meter fast walk. One secondary outcome also indicated superiority for yoga at week 12, on the nine-item Patient Health Questionnaire for depression symptoms.

Adverse events were not infrequent — about half of each group reported something — but none were serious and only a small fraction were believed to stem from the interventions. Two patients in each group dropped out because of adverse events.

Attrition was a problem in the trial, though. By week 12, each group had lost 13 participants, and two more quit in each group during the final 12 weeks. Most of the withdrawals were for nonspecified “personal” or “other” reasons. Although the main analyses were performed on an intention-to-treat basis, Antony and colleagues said they refrained from imputing data for those lost to follow-up, as their statistical models “assumed data are missing at random.” In any event, per-protocol analyses including only those participants who completed their 12- and 24-week assessments yielded similar results.

Limitations to the study included the relatively small number of participants who all lived in one Australian region. Also, it involved specific strengthening and yoga protocols, details of which may have been important for the outcomes.

“While our findings are promising, further research is needed to investigate the long-term effects of yoga and strengthening exercises beyond the 24-week period, providing insights into the sustainability of benefits,” Antony and colleagues observed. “Additionally, investigating the mechanisms underlying the observed improvements, such as pain, function, stiffness, physical performance, and depression, could yield a deeper understanding of how these interventions exert their effects.”