Cupping therapy versus acupuncture for pain-related conditions: a systematic review of randomized controlled trials and trial sequential analysis

This systematic review represents the first comprehensive study comparing cupping therapy with acupuncture for pain-related conditions. The investigators analyzed 23 randomized clinical trials involving 2,845 participants, covering 12 different painful conditions, including cervical spondylosis, knee osteoarthritis, lateral femoral cutaneous neuritis, lumbar disc herniation, postherpetic neuralgia, and scapulohumeral periarthritis. The study protocol was registered with PROSPERO, following rigorous methodological standards. The search was carried out in six electronic databases through March 2017, including Chinese and international databases.

All included studies were conducted in China, with 21 published in Chinese and 2 in English. The methodology involved quality assessment of the studies using standard risk-of-bias tools, meta-analysis when appropriate, and trial sequential analysis (TSA) to adjust for random errors. In the cupping group, 86.96% of the studies used wet cupping, with treatment frequency ranging from daily to weekly and per-session durations of 3-10 minutes in most cases. Cupping practitioners focused primarily on tender (Ashi) points in painful areas.

In the acupuncture group, 95.6% used manual acupuncture, with predominantly daily frequency and 20-30 minute sessions. Acupuncturists tended to use a more comprehensive approach, selecting points based on syndrome differentiation in addition to Ashi points. Methodological quality assessment revealed that all studies showed poor methodological quality according to the established criteria. No study was categorized as low risk of bias, 8 studies presented uncertain risk, and 15 studies were classified as high risk of bias.

Only one study specified a sample size calculation, and few reported adequate details on randomization, allocation concealment, and blinding. Owing to clinical and statistical heterogeneity across studies, only three meta-analyses could be conducted. For cervical spondylosis, the analysis of 6 studies with 646 participants showed that cupping therapy had a slightly greater effect on symptom improvement rate (RR 1.13, 95% CI 1.01-1.26, p = 0.04). For lateral femoral cutaneous neuritis, two studies with 102 participants demonstrated similar effects between the therapies (RR 1.10, 95% CI 1.00-1.22).

In scapulohumeral periarthritis, two studies with 208 participants favored cupping therapy (RR 1.31, 95% CI 1.15-1.51). The trial sequential analysis for cervical spondylosis revealed that the available data had not yet reached a statistically robust conclusion, indicating the need for approximately 2,847 participants for definitive evidence. With respect to safety, no serious adverse events were reported in either group, suggesting that both therapies are safe when applied appropriately. Seventeen studies did not report adverse event outcomes, while six studies confirmed the absence of adverse events during treatment.

The individual results of 14 studies that could not be included in the meta-analyses showed, for the most part, no statistically significant differences between the therapies. Although some studies reported statistically significant differences favoring one or the other therapy, those differences did not reach clinical significance according to established criteria. The clinical implications suggest that both therapies can be considered equivalent options for the treatment of painful conditions. The choice between cupping therapy and acupuncture can be based on practitioner experience, patient preference, cost-effectiveness considerations, and availability.

Cupping therapy may offer advantages such as shorter treatment duration and potentially lower cost, while acupuncture has a more established evidence base in some conditions. The main limitations include the poor methodological quality of the included studies, small sample sizes, possible publication bias (all Chinese studies with 'positive' results), and heterogeneity in interventions and outcome measures. Most studies did not use adequate methods of randomization, allocation concealment, or assessor blinding.