Exercise modality and dose effects on renal function in non–dialysis chronic kidney disease: a Bayesian model-based network meta-analysis

作者：Wenlong Li, Xiaomeng Lin, Xudong Cai, Minjun Liang, Shirui Shao, Xuanzhen Cen, Yang Song, Guanghui Zhong · 发表于：BMC Nephrology · 年份：2026 · DOI：10.1186/s12882-026-04977-z · 研究领域：Medicine

Chronic kidney disease (CKD) is a major global health challenge. Exercise is a key non-pharmacological therapy, but the most effective modality and dose for improving renal function in non-dialysis CKD is unclear. To compare exercise modalities and quantify dose–response effects on estimated glomerular filtration rate (eGFR) in adults with non-dialysis CKD. We searched PubMed, Web of Science, Cochrane Central, EBSCO, Embase, and CNKI from inception to July 2025 for randomized controlled trials (RCTs) involving adults with non-dialysis CKD. Exercise interventions were classified as aerobic training (AT), resistance training (RT), or mixed physical activity (MPA), which included combined training and other multicomponent interventions. Doses were converted to MET-min/week. A Bayesian dose–response network meta-analysis estimated effects and dose thresholds for eGFR. Twenty-five RCTs involving 1,827 adults with non-dialysis CKD were included. Studies were published between 2012 and 2024, with sample sizes 20–163 and intervention durations 12–52 weeks. Participants were mostly CKD stages 2–4, with common comorbidities including hypertension, diabetes and obesity. Exercise sessions lasted 30–90 min/week, with doses ranging 208–1,869 MET‑min. Compared with usual care, RT yielded the greatest improvement in eGFR (MD = 5.77 mL/min/1.73 m²; 95% CrI 3.27–8.25), followed by AT (MD = 3.66; 95% CrI 1.25–6.03), whereas MPA showed a smaller, non-significant effect (MD = 1.76; 95% CrI − 0.33...