Is It the Therapy or the Therapist Who Helps With Knee Pain?

WASHINGTON — Patients with meniscal tears undergoing physical therapy (PT) had no more improvement in a randomized trial than those receiving sham PT, raising — and perhaps answering — the question of how real-world patients actually derive benefit from PT.

In the study, called TeMPO, which also tested home-based exercise with and without supportive text messaging, patients assigned to receive genuine PT saw a mean decrease of just over 20 points in Knee Injury and Osteoarthritis Outcome Score (KOOS) pain evaluations after 12 weeks, which was almost exactly the mean decrease seen for those randomized to sham PT, according to Jeffrey N. Katz, MD, of Brigham & Women’s Hospital in Boston.

With baseline KOOS values averaging about 46, sham and real PT were equally effective, separated by less than 1 KOOS point, Katz said in presenting the results here at the American College of Rheumatology’s annual meeting.

“The active ingredient in in-clinic physical therapy arms in this study appears to be interpersonal support, and not the exercises performed,” he told attendees.

That’s a conclusion many in the field have long suspected: studies have repeatedly suggested that some part of physical therapy’s benefit in orthopedic patients is a type of placebo effect, based on the attention and motivational boost they get from the personal professional contact.

In the current study, participants got that with the sham PT, in which a therapist provided fake ultrasound on the knee (using a real machine but with its intensity set to zero) and rubbed an “inert lotion” around the mid-thigh and ankle. Real PT in the study was a “straight-up” clinic-based program of strengthening and functional exercises, Katz explained. Both involved 14 sessions over the 12-week intervention period, each lasting 20-30 minutes.

Median attendance was 13 sessions in both arms. All participants were also asked to record their home exercises in a log; close to 80% of patients in all four study arms reported that they had done their exercises at least three times weekly throughout. So it wasn’t poorer adherence to real PT that explains why its participants got no more benefit than the sham PT group.

Participants assigned to home exercise alone or with regular text messaging to encourage it also saw improvements in KOOS pain score. Baseline values also averaged about 46, and both of these groups achieved decreases of about 17 points by week 12.

KOOS scores across all four arms also continued to decrease — by about five additional points, approximately — over 9 months of further follow-up. This suggests that all of them were durably effective, although the interventions not involving in-person therapist contact were slightly less so.

Study Details

As Katz explained, physical therapy is a key early treatment to recommend when meniscal tears are suspected, as several trials have shown that it’s at least as effective as immediate arthroscopic partial meniscectomy (APM). But the potential for a placebo effect is well recognized, thus deserving an explicit investigation to gauge its scale.

Participants were recruited from centers in Boston; Pittsburgh; Cleveland; and Buffalo, New York. Eligibility criteria included age 45-85, pain scores of at least 15 on the Western Ontario and McMaster Universities Osteoarthritis Index, and MRI confirmation of meniscal tear in the painful compartment as well as cartilage degeneration. Individuals with previous surgery on the index knee, daily use of “strong” opioids, Kellgren-Lawrence grade of 4 on x-ray, or evidence of inflammatory arthritis or avascular necrosis were excluded.

More than 26,000 patients were prescreened for participation, but the vast majority were excluded for either missing the eligibility criteria or meeting one of the exclusions. Most exclusions among those screened were because meniscal tears were not found on MRI, Katz said. In the end, 879 patients were randomized among the four study arms.

Most were from Boston or Buffalo. Mean age was 59, body mass index values averaged 30.1, and 90% were non-Hispanic white. Two-thirds had Kellgren-Lawrence scores of 2 or less.

About 9% of patients progressed to APM during the study’s 12-month duration, and this rate, too, was nearly the same in all study arms. Overall adverse event rates stood at 16.0%-21.6%, with the lowest seen in the real PT group.

Limitations cited by Katz included the largely white sample and the relatively short real/sham PT sessions, although he said 30 minutes is typical in ordinary U.S. practice.