Resistance training improves eGFR in non-dialysis diabetic kidney disease while mind-body exercise reduces UAESpecific types of exercise may help manage diabetic kidney disease

ObjectiveTo evaluate the comparative efficacy and safety of structured exercise modalities on renal function and glycemic control in non-dialysis diabetic kidney disease (DKD).DesignSystematic review and network meta-analysis.Data sourcesSix databases (PubMed, Embase, Cochrane CENTRAL, Web of Science, CNKI, Wanfang) were searched from inception to July 15, 2026.Eligibility criteria for selecting studiesRandomized controlled trials (RCTs) evaluating structured exercise (aerobic, resistance, combined, or mind–body) versus usual care or an active comparator in adults with non-dialysis DKD (duration ≥4 weeks).Outcome measuresRenal endpoints were estimated glomerular filtration rate (eGFR) and urinary albumin excretion (UAE); glycemic endpoints were glycated hemoglobin (HbA1c) and fasting blood glucose (FBG).ResultsTwenty RCTs (1,746 patients) were included. Compared with usual care, resistance training (RT) was the sole modality significantly increasing eGFR (mean difference [MD] 11.36 mL/min/1.73 m², 95% CI 4.47 to 18.25; low certainty) and outperforming aerobic training (MD 8.61, 2.22 to 15.00). Mind–body exercise (MBE) was the only intervention significantly reducing UAE (standardized mean difference –1.41, –2.67 to –0.16; low certainty). For glycemic regulation, RT (MD –1.65%), aerobic training (MD –1.07%), and combined training (CT; MD –0.70%) significantly lowered HbA1c (all low certainty). Significant FBG reductions occurred only with RT (MD –1.78 mmol/L; low certainty) and CT (MD –1.71 mmol/L; moderate certainty). In sensitivity analyses, RT’s eGFR benefit lost significance after excluding high-risk trials (MD 8.24, –1.17 to 17.64), and CT’s HbA1c reduction was attenuated in pure-exercise models (MD –0.60%, –1.39 to 0.19). Subgroup analyses showed CT uniquely sustained FBG reductions across short-term (≤12 weeks: MD –1.50) and longer-term (>12 weeks: MD –2.39) regimens. Structured safety was monitored in only five trials (25.0%): three reported mild, transient events without serious harm, while two explicitly declared zero adverse events. Thus, current evidence is insufficient to draw definitive conclusions regarding comparative safety.ConclusionsStructured exercise confers distinct adaptations in non-dialysis DKD. Combined training provides a reliable foundation for sustained glycemic management, while mind–body exercise represents a viable strategy for albuminuria. Resistance training’s apparent renal benefit warrants caution due to fragility against bias. Prescriptions should be individualized rather than seeking a single optimal modality. Given the sparse monitoring of adverse events, comparative clinical safety cannot be definitively established and mandates routine surveillance.Systematic review registrationhttps://www.crd.york.ac.uk/PROSPERO/, identifier CRD420261489091.