Insomnia not only decreases the quality of life and working efficiency, but also relates to other deceases. Suitable exercises may be useful to improve sleep quality. However, comparative studies on the effects of different types of exercise on sleep outcomes are currently lacked.

To compare the effects of different exercise types on sleep in insomnia patients and to provide insights for developing exercise prescriptions to improve sleep in this population.

Randomized controlled trials (RCT) investigating the effects of different types of exercises on the sleep of people with insomnia were retrieved from the PubMed, Web of Science, EBSCO, Scopus, China National Knowledge Infrastructure (CNKI), and Wanfang Data Knowledge Service Platform from database inception to December 2023. The quality of the literatures was assessed using the Jadad scale. A network meta-analysis was performed using the random-effects model of the frequency framework in Stata 17.0 software, and publication bias was observed using a funnel plot.

A total of 36 articles representing 2 706 participants were included in this study. Exercise interventions were categorized into five types, including the aerobic exercise (AE), resistance training (RT), mind-body exercise (ME), prolonged time of exercise (PTE), and AE combined with RT (AE+RT). The results of the network meta-analysis showed that there was a significant difference in the total score of the subjective evaluation tools between AE group and blank control group (SMD=2.20, 95%CI=0.35-4.05, P<0.05). The maximum surface under the cumulative ranking curve (SUCRA) of the total score of the subjective evaluation tools was detected in AE group (82.2%), followed by ME (71.1%), AE+RT (54.2%), RT (38.9%), controlled trial (CT) (29.2%) and PTE groups (24.4%). There was a significant difference in sleep latency (SL) between PTE group versus AE+RT group (SMD=-14.66, 95%CI=-28.60 to -0.72), between AE+RT group versus RT group (SMD=21.32, 95%CI=6.24-36.40), and between AE+RT group versus AE group (SMD=11.19, 95%CI=1.56-20.81) (all P<0.05). The maximum SUCRA of SL was detected in AE+RT group (96.7%), followed by CT (77.1%), ME (49.5%), AE (42.6%), PTE (27.9%) and RT groups (6.1%). There was a significant difference in sleep efficacy (SE) between AE group versus CT group (SMD=4.59, 95%CI=1.59-7.58, P<0.05). The maximum SUCRA of SE was detected in the AE group (83.1%), followed by AE+RT (59.9%), RT (50.7%), ME (46.9%), PTE (35.5%) and CT groups (23.9%). There was a significant difference in wake time after sleep onset (WASO) between PTE group versus RT group (SMD=23.39, 95%CI=14.67-32.02), between PTE group versus AE group (SMD=14.86, 95%CI=4.64-36.40), between PTE group versus CT group (SMD=12.11, 95%CI=7.62-16.59), between RT group versus ME group (SMD=-27.61, 95%CI=-40.49 to -14.73), between RT group versus CT group (SMD=-11.28, 95%CI=-20.86 to -1.70), between ME group versus AE group (SMD=19.08, 95%CI=5.18-32.98), and between ME group versus CT group (SMD=16.33, 95%CI=7.26-25.40) (all P<0.05). The maximum SUCRA of WASO was detected in the RT group (94.7%), followed by CT (80.2%), AE (41.8%), ME (31.5%) and PTE groups (1.7%). There was a significant difference in the total sleep time (TST) between RT group versus AE group (SMD=-20.84, 95%CI=-38.46 to -3.23), between AE group versus CT group (SMD=12.02, 95%CI=2.96-21.08) (all P<0.05). The maximum SUCRA of TST was detected in the AE group (87.2%), followed by AE+CT (71.6%), PTE (53.8%), ME (40.6%), CT (35.9%) and RT groups (10.8%). There was no significant difference in the sleep quality (SQ) among groups (P>0.05). The maximum SUCRA of SQ was detected in the AE group (84.2%), followed by AE+CT (75.5%), ME (24.9%) and CT groups (15.4%). Publication bias analysis of the total score of subjective evaluation tools, SL, SE, WASO, TST, and SQ revealed some asymmetry in all six outcome indicators, indicating a certain risk of publication bias.

Given the favorable effects of AE on multiple sleep parameters, we recommend insomnia patients engage in moderate-intensity AE, such as jogging or brisk walking, three times a week for approximately 30 minutes. Additionally, options include moderate-intensity RT or AE+RT four to five times a week. When engaging in the above exercises, attention should be paid to the choice of time period, and exercise within 3 hours before bedtime should be avoided as much as possible.