Abstract / Summary
Objectives Identifying patients at high risk of low anterior resection syndrome (LARS) after rectal cancer surgery may facilitate early management and improve postoperative quality of life. Several LARS prediction models have been developed, but their predictive performance and methodological quality vary, and their predictors have not been systematically synthesized. This study aimed to systematically review and meta-analyze the predictive performance of existing LARS prediction models, focusing on the area under the curve (AUC), and to synthesize the predictors incorporated into these models. Methods PubMed, Web of Science, Embase, the Cochrane Library, CINAHL Plus with Full Text, China National Knowledge Infrastructure, Wanfang Database, and VIP Database were systematically searched from inception to April 25, 2026. The Prediction Model Risk of Bias Assessment Tool (PROBAST) was used to assess the risk of bias and applicability of the included prediction models. Meta-analyses were conducted using R software (version 4.2.1) to synthesize model discriminative performance using AUC and predictors associated with LARS. Results A total of 29 studies were included. The pooled AUC for external validation of major LARS prediction models was 0.83 (95% confidence interval [CI]: 0.78-0.87). Significant predictors of LARS included low anastomotic height (odd ratio [OR] = 2.83, 95% CI: 1.26-7.32), low tumor height (OR = 3.13, 95% CI: 2.18-4.49), diverting stoma (OR = 2.01, 95% CI: 1.33-3.03), anastomotic leakage (OR = 6.38, 95% CI: 1.49-27.25), neoadjuvant therapy (OR = 3.19, 95% CI: 1.88-5.39), and body mass index ≥24 kg/m² (OR = 2.49, 95% CI: 1.81-3.42). Conclusions Existing LARS prediction models showed promising discriminative performance, but their clinical applicability remains uncertain because all included studies were judged to have a high risk of bias and external validation was limited. Future studies should prioritize the development and external validation of robust LARS prediction models using large prospective cohorts. Implications to Nursing Practice This study synthesized the predictive performance and key predictors of existing LARS prediction models, providing a reference for future model development. Once adequately validated, such models may help nurses identify patients at higher risk of postoperative LARS and support timely counselling, bowel-function monitoring, and targeted management to reduce its impact on recovery and quality of life.