Development and validation of a predictive model for increased drainage after open transforaminal lumbar posterior fusion

Abstract

Objective This study aims to investigate the risk factors associated with increased drainage volume following open transforaminal lumbar interbody fusion (TLIF) surgery and to develop and validate a predictive model. Methods We collected clinical data from 795 patients who underwent open TLIF at the Northern Theater Command General Hospital between January 2016 and December 2020. These patients were randomly divided into a training group (n = 557) and a validation group (n = 238), with no significant statistical difference between the groups (p > 0.05). Using variables selected via LASSO regression analysis, we constructed a multivariable logistic regression prediction model and developed a corresponding nomogram. The model's performance was internally validated using ROC curves, the Hosmer-Lemeshow goodness-of-fit test, and calibration curves. Its clinical utility was assessed using Decision Curve Analysis (DCA). Results Four predictive variables were identified through LASSO regression analysis: age, surgical segment, duration of surgery, and intraoperative blood loss. The ROC curve demonstrated that the model possesses excellent discriminative ability. Additionally, the Hosher-Lemeshow test and calibration curves indicated that the model's predicted probabilities align closely with actual outcomes, showing high calibration accuracy. The DCA confirmed the clinical utility of the predictive model, establishing its suitability for clinical application. Conclusion The risk factors for increased drainage volume post-open TLIF include age, surgical segment, duration of surgery, and intraoperative blood loss. The validation confirms that the constructed predictive model is robust and can be effectively applied in clinical settings.