Objective: The purpose of this systematic review (SR) is to evaluate the effectiveness and safety of Tui Na therapy for insomnia.

Methods: Two authors separately searched PubMed, the Cochrane Library, EMBASE, SinoMed Database, China National Knowledge Infrastructure (CNKI), Wanfang Database, China Science Technology Journal Database, related SR and published protocols at the same time to find randomized controlled trials (RCTs), which compared Tui Na therapy with estazolam therapy for insomnia, from their inception to January^1st 2019. Screening documents, data extraction, quality assessment of methodology and quality assessment of evidence were also conducted by two authors separately at the same time. We used Cochrane Risk of Bias tool to assess the methodological quality of included RCTs. The results of meta-analysis were made via RevMan software (5.3). The quality of evidence was assessed by on-line GRADEpro. The primary outcome: Pittsburgh Sleep Quality Index (PSQI) score, the secondary outcome: clinical effectiveness rate and the safety index: adverse events. The clinical effectiveness of these included RCTs all focused on the improvement of patients' satisfaction with sleep time and sleep quality.

Results: We included 22 RCTs(1,999 participants), meeting the inclusion and exclusion criteria. The assessment of methodological quality was not satisfied, in which "high risk", "unclear risk" and "low risk" all existed. The results of meta-analysis demonstrated that (1)for primary outcome, the PSQI score of Tui Na therapy was lower than that of estazolam therapy after treatment in subgroup1(Head)(MD-2.39,95%CI[-3.79,-0.98],I² = 82%,n = 291,3 trials) and subgroup4(Abdomen)(MD-1.7,95%CI[-2.53,-0.87],I² = 0%,n = 120,2 trials);while there was no significant difference between Tui Na therapy and estazolam therapy in subgroup2(Head and trunk)(MD-1.39,95%CI[-3.03,0.24],I² = 90%,n = 200, 2 trials) and subgroup3(Head, trunk and extremities)(MD-0.51,95%CI[-1.53,0.5],I² = 30%,n = 126,2 trials).(2) for secondary outcomes(the clinical effectiveness rate and safety index),the clinical effectiveness rate of Tui Na therapy was higher than that of estazolam therapy after treatment in subgroup1(Head)(RR1.21,95%CI[1.05,1.39],I² = 26%,n = 239,3 trials), subgroup2(Head and trunk)(RR1.15,95%CI[1.08,1.23],I² = 33%,n = 1024,10 trials) and subgroup4(Abdomen)(RR1.12,95%CI[1.01,1.23],I² = 0%,n = 180,3 trials); while there was no significant difference between Tui Na therapy and estazolam therapy in subgroup3(Head, trunk and extremities)(RR1.03,95%CI[0.94,1.13],I² = 28%,n = 346,4 trials).Safety index, 5 RCTs reported adverse events. Among them, only 1 RCT reported adverse event in Tui Na therapy, which was daytime drowsiness; all 5 RCTs reported adverse events in estazolam therapy, which were dry mouth, dizziness, daytime drowsiness etc. The evidence quality was generally low to very low.

Conclusion: Tui Na therapy appeared to be superior to estazolam therapy in treating areas(head; abdomen), while there was no significant difference between Tui Na therapy and estazolam therapy in treating areas(head and trunk; head, trunk and extremities). No serious adverse event was reported in Tui Na therapy. However the methodological quality and evidence quality were not satisfied. Therefore we could not make a convincing conclusion on the effectiveness and safety of Tui Na therapy for insomnia. Practitioners should combine their experience, evidence of our review and patients' preferences to make a proper treatment. And more high quality RCTs and well-designed protocols of Tui Na therapy for insomnia are needed in the future.