Acupuncture and other physical treatments for the relief of pain due to osteoarthritis of the knee: network meta-analysis

This network meta-analysis represents the most comprehensive comparative study of physical treatments for knee osteoarthritis ever conducted, analyzing data from 156 eligible studies and 114 studies with adequate data for analysis, totaling 9,709 patients. The objective was to compare the effectiveness of acupuncture with 21 other available physical treatments for pain relief in knee osteoarthritis, using network meta-analysis methodology that allows indirect comparisons between treatments not directly compared in individual studies.

The methodology involved systematic searches in 17 databases through January 2013, including randomized clinical trials in adults with knee osteoarthritis and mean age of 55 years or older. Eligible treatments included acupuncture, balneotherapy, braces, aerobic and strengthening exercises, thermal therapies, insoles, interferential therapy, laser therapy, manual therapy, electrical stimulation, electromagnetic fields, Tai Chi, TENS, and weight loss. Study quality was assessed using an adapted checklist, classifying studies as excellent, good, satisfactory, or poor.

The main results showed that eight interventions produced statistically significant pain reduction when compared with standard care: interferential therapy, acupuncture, TENS, pulsed electrical stimulation, balneotherapy, aerobic exercise, sham acupuncture, and strengthening exercises. In the analysis of all studies, acupuncture demonstrated a standardized mean difference of -0.89 (95% credible interval: -1.18 to -0.59) compared with standard care, corresponding to a 14.69-point reduction on the 0-100 WOMAC pain scale.

When directly compared with other active treatments, acupuncture proved superior to sham acupuncture (difference of 6.93 points), strengthening exercises (8.14 points), weight loss, pulsed electromagnetic fields, placebo, insoles, neuromuscular electrical stimulation, and no intervention. In the sensitivity analysis including only studies of satisfactory or good quality (35 studies, 3,499 patients), acupuncture maintained statistically significant superiority over strengthening exercises, with a standardized mean difference of 0.49 (0.00 to 0.98).

The ranking analysis showed that acupuncture and balneotherapy were the treatments most likely to be the most effective, although with considerable overlap of credible intervals with other treatments. It is important to note that most studies (110 of 156) were classified as poor quality, mainly due to the impossibility of adequately blinding patients in physical treatments.

The clinical implications suggest that acupuncture can be considered one of the most effective physical treatment options for short-term pain relief in knee osteoarthritis. Compared with established criteria of clinical relevance, the results indicate that acupuncture produces a 'minimal perceptible clinical improvement' and possibly a 'minimal clinically important change,' especially in patients with moderate to severe pain levels.

Limitations include the predominance of short-term evidence, variable methodological quality of studies, impossibility of blinding in many physical treatments, and clinical heterogeneity inherent in comparing diverse treatments. The question of specific versus nonspecific effects of acupuncture remains, since sham acupuncture comparison studies showed smaller effects, suggesting that nonspecific (placebo) effects contribute significantly to the observed benefits. However, considering that other physical treatments also cannot be adequately blinded, the comparisons remain valid for clinical decision-making.