Chinese Journal of Tissue Engineering Research ›› 2026, Vol. 30 ›› Issue (3): 749-759.doi: 10.12307/2025.867

Previous Articles     Next Articles

Risk prediction models of recurrence after percutaneous endoscopic lumbar discectomy: a systematic review and meta-analysis

Yu Weijie1, 2, Cao Dongdong1, 2, Guo Tianci1, 2, Niu Puyu1, 2, Yang Jialin1, 2, Wang Simin1, 2, Liu Aifeng1, 2   

  1. 1First Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300381, China; 2National Clinical Research Center of Chinese Medicine Acupuncture and Moxibustion, Tianjin 300381, China

  • Received:2024-10-15 Accepted:2024-12-18 Online:2026-01-28 Published:2025-07-09
  • Contact: Liu Aifeng, Chief physician, Doctoral supervisor, First Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300381, China; National Clinical Research Center of Chinese Medicine Acupuncture and Moxibustion, Tianjin 300381, China
  • About author:Yu Weijie, Doctoral candidate, First Affiliated Hospital of Tianjin University of Traditional Chinese Medicine, Tianjin 300381, China; National Clinical Research Center of Chinese Medicine Acupuncture and Moxibustion, Tianjin 300381, China
  • Supported by:
    Tianjin Health and Family Planning Industry High-level Talents Selection and Training Project (Jinmen Medical Talents), No. TJSJMYXYC-D2-028 (to LAF); Tianjin Science and Technology Plan Project, No. 23KPXMRC00170 (to LAF)

Abstract: OBJECTIVE: Postoperative recurrence is a common complication of percutaneous endoscopic lumbar discectomy for lumbar disc herniation, which can significantly increase the risk of reoperation. A well-performing risk prediction model can help identify high-risk groups early and prevent postoperative recurrence. This study systematically evaluated the risk prediction model for postoperative recurrence after percutaneous endoscopic lumbar discectomy to provide a reference for surgical decision-making. 
METHODS: The PubMed, Embase, Web of Science, CNKI, WanFang Data, VIP, and CBM were electronically searched to collect studies on the recurrence risk prediction models after percutaneous endoscopic lumbar discectomy from inception to July 1, 2024. Two reviewers independently screened the literature and extracted data. The models’ risk of bias, applicability, and report quality were assessed using prediction model risk of bias assessment tool (PROBAST) and Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis (TRIPOD) tools, respectively. Meta-analysis of postoperative recurrence rate of percutaneous endoscopic lumbar discectomy and related predictors was performed using Revman 5.4 software. 
RESULTS: (1) A total of 15 studies were included, all of which were retrospective studies, including 24 models for predicting the risk of recurrence after percutaneous endoscopic lumbar discectomy. (2) The PROBAST evaluation results indicated that all 15 studies exhibited a high risk of bias. Regarding applicability, two studies demonstrated a low risk, while 13 presented a high risk. (3) Regarding the TRIPOD reporting quality, the overall quality across the 15 studies was low. The primary reasons for this low compliance included the failure to report blinding, a lack of explanation for the sample size calculation method, lack of detailed description of missing data processing methods, and lack of information such as introduction to the model used. (4) Furthermore, the area under the receiver operating characteristic curve for the model ranged from 0.684 to 0.972, with the number of potential predictor variables varying from 15 to 28. (5) The results of meta-analysis showed that the postoperative recurrence rate of lumbar disc herniation patients treated with percutaneous endoscopic lumbar discectomy was 12% (95%CI=9.0%-15.0%), Modic changes (OR=6.72, 95%CI=3.90-11.59), body mass index (OR=1.28, 95%CI=1.10-1.49), work intensity (OR=3.22, 95%CI=1.85-5.59), age (OR=2.28, 95%CI=1.50-3.48), and smoking history (OR=2.65, 95%CI=1.75-4.00) were independent influencing factors for postoperative recurrence of percutaneous endoscopic lumbar discectomy (all P < 0.05).  
CONCLUSION: The overall predictive performance of the recurrence risk prediction models after percutaneous endoscopic lumbar discectomy is satisfactory; however, the model exhibits a high overall risk of bias and applicability, coupled with low reporting quality. Additionally, there is a lack of prospective research and external validation. Future, risk prediction models should consider factors such as Modic changes, body mass index, work intensity, age, and smoking history as potential predictors. 

Key words: percutaneous endoscopic lumbar discectomy, lumbar disc herniation, recurrence, Modic change, risk prediction model, influencing factor, systematic review

CLC Number: