Main menu


Online yoga: Doesn't help much with knee pain, study finds

The researchers suggested that an unsupervised, online yoga program designed to relieve knee osteoarthritis pain and improve function has shown hints of effectiveness but appears to need some tweaking.

When assessed at Week 12, when the program nominally ended, patients in the yoga group showed a mean improvement of 2.1 points from baseline on a 0–10 pain scale, while the educational control group improved only by 1.4 points, with a difference failing to reach significance. A statistic, according to Kim Pinnell, Ph.D. from the University of Melbourne in Australia, and colleagues.

There was a significant advantage of the online program in terms of physical function, as measured by the Western Ontario and McMaster Universities (WOMAC) Arthritis Index of 68: Mean improvement of 10.1 points with the yoga program versus 4.7 points among controls (median​ ​Difference 4.0 points after adjusting for baseline parameters, 95% CI 1.3-6.8), researchers reported in Annals of internal medicine.

But after another 12 weeks, during which the yoga group could continue the program if they wished, there was no significant difference between the groups, both for pain (0.3 points, 95% CI -0.4 to 1.0) or for function (2.1) points, 95 %CI -0.9 to 5.1). Pain scores and functional impairment rebounded slightly in the yoga group, while controls showed some sustained improvement.

Furthermore, differences between groups in the 12-week assessment failed to meet pre-established thresholds for improving “minimal clinical significance,” which were 1.8 points on the pain scale and 6 points on the WOMAC index.

Bennell and colleagues noted that a lack of adherence may have limited the effectiveness of the yoga program, and that a lack of blinding may have affected the results. Since the program was not moderated, it is possible that the participants did not perform the yoga maneuvers correctly or completely, although they may have made it to the video on schedule. Also, many participants in the control group performed exercises and used devices such as orthotics and knee braces, in some cases at higher rates than the yoga group. The use of analgesics was also somewhat greater in the control group.

Overall, the researchers described their results as “promising” for unsupervised online yoga, inasmuch as the improvements were numerically superior in the yoga group and there are opportunities to make the program more effective.

“Our program was low-intensity, focusing on postures and movement, with fewer or no elements included such as deep relaxation, chanting, singing, and meditation,” Pennell and colleagues write. “Thus, studies looking at different types, doses, and forms of yoga on the Internet, along with an exploration of mediators and mediators of effects, would be helpful.”

Another positive note is that the program appeared to be safe, with only a few patients showing new pain or injuries, and none of them showing seriousness.

Study details

Bennell and his colleagues had high hopes for the program, which was meant to be very low-cost and easily reachable to large numbers of people — unlike regular yoga, which is usually done in small groups with a trained leader. The program was developed jointly with five registered yoga therapists, a physical therapist, and two patients with knee osteoarthritis. The researchers explained that each session lasted 30 minutes, starting with 5 minutes of simple breathing and standing exercises, followed by 20 minutes of intense static and dynamic postures “aimed at activating, strengthening and stretching the muscles of the trunk and lower extremities.” The sessions ended with a 5-minute ‘cool down’ period with stretching, breathing and relaxation exercises.

Specific positions and exercises changed over time during the 12-week program, so that there were a total of 12 different videos available. Each featured a trained instructor and three reps following them, to simulate a regular yoga class. The sessions were scheduled to take place three times per week during the 12-week programme. After the program ended at week 12, the yoga group participants were encouraged to continue.

A total of 212 patients were randomly assigned to the two groups, both of whom received education packages on self-management of knee osteoarthritis, including recommendations for exercise, weight loss, use of pain relievers and supportive devices. Participants completed questionnaires about pain, functional ability, and general health status at baseline, week 12, and week 24.

The average patient age was about 62, and nearly 70% were women. Their BMI was around 30. Most had at least one co-morbidity. High blood pressure was most common, followed by depression. Some baseline parameters were not evenly balanced at baseline, such as current employment, joint problems elsewhere in the body (both more common in controls) and heart disease (more common in the yoga group).

Self-discipline was recorded by the yoga group participants in an electronic log book. They averaged 2.5 sessions per week; At week 12, 69% said they had accessed the program three times in the previous week. However, the investigators also counted participants’ logins to the website where the yoga videos were played: this averaged 20.5 in the yoga group, or less than twice a week over the 12-week program, and the average access to The session is 25 minutes.

The study included a range of secondary outcomes besides knee pain and WOMAC function score: depression, anxiety, stress, quality of life, fear of movement, confidence in balance, and others. Some showed better results in the yoga group at week 12, but – as with the initial results – none of these benefits persisted until week 24.

  • John Jeffer was managing editor from 2014 to 2021; He is now a regular shareholder.


The study was funded by Australian Government grants. Some authors have reported relationships with medical publishers but no drug or device companies.