Physical therapy versus glucocorticoid injection for osteoarthritis of the knee

Key points

1. 01Physical therapy patients had a mean WOMAC score of 37.0 at 1 year versus 55.8 in the injection group, a clinically meaningful difference of 18.8 points.
2. 02Only 10.3% of physical therapy patients failed to reach the minimal clinically important difference in WOMAC, compared with 25.6% in the injection group.
3. 03All secondary outcomes (Global Rating of Change, Timed Up and Go, Alternate Step Test) also favored physical therapy at 1 year.
4. 041-year health care costs were similar in both groups ($2,131 for physical therapy vs. $2,113 for injection).
5. 05Adverse events were rare; one patient fainted during the injection, and no other adverse events were recorded.

How it was conducted

Design

Randomized controlled trial (intention-to-treat analysis)

Setting

Two large US Military Health System hospitals (Brooke Army Medical Center, TX and Madigan Army Medical Center, WA), October 2012 to May 2017

Participants

156 adults aged 38 or older with clinical and radiographic knee osteoarthritis (Kellgren-Lawrence grade 1-4), mean age 56 years, 48% women, mean BMI 31.5

Groups

78 patients received intraarticular triamcinolone acetonide injection (up to 3 over 1 year); 78 underwent physical therapy (manual therapy plus exercise, up to 8 sessions over 4-6 weeks with optional additional sessions at 4 and 9 months)

Primary outcome

Total WOMAC score at 1 year (range 0-240, higher = worse)

Secondary outcomes

Global Rating of Change scale, Timed Up and Go test, Alternate Step Test, and 1-year knee-related health care costs

What they found

• Mean WOMAC score at 1 year: 37.0 (95% CI 30.8-44.5) for physical therapy vs. 55.8 (95% CI 45.0-69.1) for glucocorticoid injection; mean between-group difference 18.8 points (95% CI 5.0 to 32.6; P = 0.008), favoring physical therapy.
• Failure to achieve minimal clinically important WOMAC improvement (12%) at 1 year: 10.3% in physical therapy group vs. 25.6% in injection group.
• Median Global Rating of Change score at 1 year: +5 ("quite a bit better") for physical therapy vs. +4 ("moderately better") for injection; 11/78 (14.1%) in physical therapy did not reach +3 threshold vs. 26/78 (33.3%) in injection group (relative risk 0.42; 95% CI 0.23 to 0.80).
• Mean time to complete Alternate Step Test at 1 year: 8.0 sec (physical therapy) vs. 9.0 sec (injection); between-group difference 1.0 sec (95% CI 0.3 to 1.6; P = 0.003).
• Mean time to complete Timed Up and Go test at 1 year: 7.3 sec (physical therapy) vs. 8.1 sec (injection); between-group difference 0.9 sec (95% CI 0.3 to 1.5; P = 0.005).
• Mean 1-year knee-related health care cost: $2,131 (physical therapy) vs. $2,113 (glucocorticoid injection).
• Patients in the injection group received a mean of 2.6 injections (range 1-4); physical therapy patients attended a mean of 11.8 treatment visits (range 4-22).

Limitations

• Physical therapy patients had more total contact time with health care providers, which may have contributed to greater benefit independent of the specific treatment.
• 18% of patients in the injection group also received physical therapy during the trial, and 9% of physical therapy patients also received an injection, making the groups less distinct.
• The physical therapy group had a higher proportion of patients with severe radiographic osteoarthritis (Kellgren-Lawrence grades 3 and 4) at baseline despite randomization.
• The trial was conducted in a military health system population, and most patients were referred by primary care physicians, which may limit generalizability to other settings.

Why it matters

For patients

Patients with knee osteoarthritis are more likely to have sustained pain relief and functional improvement after a course of physical therapy than after a steroid injection.

For clinicians

Physical therapy should be prioritized over glucocorticoid injection as a first-line treatment for knee osteoarthritis, with comparable 1-year costs but meaningfully better pain and function outcomes.

For readers

This well-powered RCT published in NEJM provides direct evidence that physical therapy outperforms glucocorticoid injection at 1 year across both patient-reported and functional outcomes.