Comparative effectiveness and safety of acupuncture treatments for primary insomnia: a systematic review and network meta-analysis of randomized trial

Background: This study employed a Bayesian network meta-analysis (NMA) to systematically evaluate the efficacy and safety of various acupuncture therapies compared to conventional medication, sham acupuncture, and other interventions for primary insomnia. Methods: PubMed, Embase, Cochrane Library, Web of Science, CNKI, VIP Chinese Scientific Journals, Wanfang, and China Biology Medicine were searched from inception to July 16, 2025. Literature quality was assessed using the Cochrane Risk of Bias Tool v 2.0 (RoB 2.0). Statistical analyses were performed using Stata 18 and R 4.5.1. Results: In total, 80 studies involving 7,791 patients were included. Among these, 60.0% were rated as low RoB, 26.3% as unclear RoB, and 13.8% as high RoB. Statistical analysis showed that, compared with conventional medication, abdominal acupuncture (Weighted Mean Difference (MD) -3.73; 95% Credible Interval (95% CrI) [-6.88, -0.55]), acupuncture (MD -1.96; 95% CrI [-2.64, -1.27]), and catgut embedding (MD -3.08; 95% CrI [-5.18, -0.93]) significantly reduced the short-term Pittsburgh sleep quality index (PSQI) scores. Compared with acupuncture, warm acupuncture (MD -2.55; 95% CrI [-4.88, -0.21]) significantly reduced the long-term PSQI scores. Compared with sham acupuncture, abdominal acupuncture (Standardized Mean Difference (SMD) -3.06; 95% CrI [-6.08, -0.09]) and acupuncture (SMD -2; 95% CrI [-3.05, -0.98]) significantly reduced anxiety scores; meanwhile, acupuncture (SMD -1.52; 95% CrI [-2.79, -0.26]) significantly reduced depression scores. Compared with conventional medication, acupuncture (Relative Risk (RR) 1.19; 95% CrI [1.12, 1.27]) and catgut embedding (RR 1.25; 95% CrI [1.05, 1.52]) significantly improved clinical efficacy rates. However, no significant differences were observed in the relative effectiveness among different acupuncture therapies. The cumulative sample size included in the safety analysis was 1,772, from which 99 adverse events were reported (5.59%). No significant differences were detected across interventions; based on the surface under the cumulative ranking curve (SUCRA), wrist-ankle needle may show higher potential safety. Conclusion: Currently, no single intervention has emerged as optimal across all outcomes. Abdominal acupuncture, catgut embedding, electroacupuncture, and wrist-ankle needle ranked relatively high for certain outcomes based on SUCRA and showed potential advantages. However, given the potential publication bias, variations in acupuncture protocols, and insufficient long-term follow-up data, further validation is required. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251040450, Identifier: CRD420251040450.