关键词: 脊柱, 静脉血栓栓塞, 预测模型, 系统评价, 循证护理学

Abstract:

Objective To systematically evaluate risk prediction models for venous thromboembolism（VTE） after spinal surgery and provide evidence for clinical risk assessment and model optimization. Methods Chinese and English databases were systematically searched from inception to June 30，2026. The CHARMS checklist was used for literature screening and data extraction. Methodological quality was assessed using the PROBAST tool，and a meta-analysis was performed for models eligible for quantitative synthesis. Results A total of 15 studies were included，most of which were retrospective cohort studies. Logistic regression was the predominant modeling approach，while several studies incorporated the LASSO regression or machine learning algorithms. The main predictors included D-dimer，age，and surgery-related factors. The AUC ranged from 0.68 to 0.94，and most models demonstrated good calibration. External validation was performed in 6 studies，and 8 studies evaluated clinical utility using DCA. 4 studies were included in the meta-analysis，yielding a pooled AUC of 0.841（95%CI：0.783-0.899）. Conclusion Existing prediction models for postoperative VTE following spinal surgery generally demonstrate good predictive performance. However，their methodological quality remains suboptimal，and external validation is limited. The predictors remain focused on conventional VTE risk factors，with inadequate reflection of the specialized characteristics of spinal surgery. Future prospective multicenter studies are warranted to identify spinal-specific predictors and improve the generalizability and clinical applicability of these prediction models.

Key words: Spine, Venous Thromboembolism, Prediction Model, Systematic Review, Evidence-Based Nursing