Time-stratified prognostic factors for poor renal outcomes in childhood-onset lupus nephritis: a systematic review and meta-analysis.

Objectives: To identify prognostic factors independently associated with poor renal outcomes in childhood-onset lupus nephritis (cLN) based on multivariable analyses. Methods: Seven databases were searched up until October 2025. Adjusted odds ratio (OR), hazard ratios (HR), and pooled proportions of renal outcomes were estimated using random-effects models. Outcome-specific analyses were performed when sufficient data were available, and sensitivity analyses excluding studies at high risk of bias were conducted to assess the robustness of findings. Methodological quality was evaluated using quality in prognostic studies tool. Certainty of evidence was assessed using Grading of Recommendations, Assessment, Development and Evaluation. Results: 47 studies were included, with 23 assessed as having a high risk of bias. The included studies explored prognostic factors across follow-up periods of 1 year (n=11), 1-2 years (n=9), and >2 years (n=27). Within 1 year, 27 (OR) prognostic factors were identified. Within 1-2 years, 25 (OR) and 6 (HR) prognostic factors were identified respectively. Beyond 2 years, 26 (OR) and 55 (HR) were identified respectively. Frequently evaluated factors associated with poor renal outcomes included age, dialysis, eGFR <60, hypertension, kidney flare, male sex, and renal non-remission within 6 or 12 months. Nephrotic-range proteinuria and increased serum creatinine were associated with poor renal outcomes in the primary analysis but showed limited robustness after excluding studies at high risk of bias. Higher eGFR and hemoglobin levels showed inverse associations with poor renal outcomes. Outcome-specific analyses suggested that male sex and increased serum creatinine were mainly associated with kidney-specific outcomes comprising CKD and/or ESRD, whereas renal non-remission at 6 and 12 months was associated with composite outcomes. Except for renal non-remission (1 year: 23.00%; >2 years: 22.43%), the reported cumulative proportions of other outcomes were higher in studies with longer follow-up durations (CKD: 4.19% to 11.46%; ESRD: 3.10% to 8.58%; mortality: 2.69% to 8.71%). Conclusion: Identified prognostic factors provide a reference for stratified management of cLN. Differences in prognostic associations across studies categorized by follow-up duration and higher cumulative proportions of adverse outcomes in studies with longer follow-up highlight the need for longitudinal, dynamic risk assessment strategies. Systematic review registration: https://www.crd.york.ac.uk/PROSPERO/view/CRD420251169446, identifier CRD420251169446.