Prediction model for tibial plateau fracture combined with meniscus injury

Abstract

Purpose: To investigate a prediction model of meniscus injury in patients with tibial plateau fracture.

Methods: This retrospective study enrolled patients with tibial plateau fractures who were treated in the Third Hospital of Hebei Medical University from January 1, 2015, to June 30, 2022. Patients were divided into a development cohort and a validation cohort based on the time-lapse validation method. Patients in each cohort were divided into a group with meniscus injury and a group without meniscus injury. Statistical analysis with Student's t-test for continuous variables and chi square test for categorical variables was performed for patients with and without meniscus injury in the development cohort. Multivariate logistic regression analysis was used to screen the risk factors of tibial plateau combined with meniscal injury, and a clinical prediction model was constructed. Model performance was measured by examining discrimination (Harrell's C-index), calibration (calibration plots), and utility [decision analysis curves (DCA)]. The model was validated internally using bootstrapping and externally by calculating their performance in a validation cohort.

Results: Five hundred patients (313 [62.6%] males, 187 [37.4%] females) with a mean age of 47.7 ± 13.8 years were eligible and were divided into development (n = 262) and validation (n = 238) cohorts. A total of 284 patients had meniscus injury, including 136 in the development cohort and 148 in the validation cohort We identified high-energy injuries as a risk factor (OR = 1.969, 95%CI 1.131-3.427). Compared with blood type A, patients with blood type B were more likely to experience tibial plateau fracture with meniscus injury (OR = 2.967, 95%CI 1.531-5.748), and office work was a protective factor (OR = 0.279, 95%CI 0.126-0.618). The C-index of the overall survival model was 0.687 (95% CI, 0.623-0.751). Similar C-indices were obtained for external validation [0.700(0.631-0.768)] and internal validation [0.639 (0.638-0.643)]. The model was adequately calibrated and its predictions correlated with the observed outcomes. The DCA curve showed that the model had the best clinical validity when the threshold probability was 0.40 and 0.82.

Conclusions: Patients with blood type B and high-energy injuries are more likely to have meniscal injury. This may help in clinical trial design and individual clinical decision-making.

Keywords: meniscus injury; nomogram; prediction model; risk factors; tibial plateau fracture.