Comparative effectiveness of different modes of exercise interventions in diabetics with frailty in China: a systematic review and a network meta-analysis

Abstract Objective To systematically evaluate the efficacy of different training modes in patients with diabetes decline. Methods PubMed, Cochrane Library, EMbase, Web of Science, CNKI, VIP, WANFANG, SinoMed were searched in computer to collect randomized controlled trials (RCTs) of training intervention in patients with diabetes and frailty, and the search time was as of May 21, 2023. After two review authors independently screened studies, extracted data, and assessed the risk of bias of included studies, network meta-analysis was performed using Stata14.0 and R4.3.1 software. Fasting blood glucose (FGB), glycosylated haemoglobin (HbA1c), two-hour postprandial blood glucose (PBG), total cholesterol (TCH), triglycerides (TG), low-density lipoprotein cholesterol (LDL-C), Short Physical Performance Battery (SPPB), and body mass index (BMI) were used as outcome measures. Results A total of 15 RCTs were included, including 1550 patients. The results of the network meta-analysis showed that integrated training reduced FBG compared with the control group; integrated training, Pilates training, resistance training can reduce HbA1c; Pilates training and resistance training can reduce PBG; integrated training, Pilates training, resistance training can reduce TCH; Pilates training and resistance training can reduce TG; resistance training improves BMI. The results of the best probability ranking showed that multi-group training had the most significant effect on improving PBG and SPPB scores. Conclusion The current evidence suggests that multi-group training is the best way to reduce fasting blood glucose and improve physical activity before meals, and Pilates training may be the best way to reduce glycated hemoglobin, blood glucose two hours after meals, improve blood lipid level and BMI in patients with diabetes in China. Trial registration: PROSPERO registration number for this study: CRD42023427868.