Development and validation of a nomogram prediction model for perioperative deep vein thrombosis risk in arthroplasty: a retrospective study.

BACKGROUND

Perioperative monitoring thrombosis has become more crucial due to the rising demand for arthroplasty and shorter hospital stays. We aimed to comprehensively explore immune-inflammatory and hypercoagulable states during perioperative periods patients undergoing arthroplasty to identify the risk factors for early postoperative deep vein thrombosis (DVT) and construct a nomogram prediction model for postoperative DVT.

METHODS

Electronic medical records of 841 patients who underwent primary arthroplasty at a single institution were retrospectively reviewed. Patients' demographic and perioperative laboratory data were collected and divided into training (73.8%) and validation sets (26.2%) based on order of procedure date. Variables were screened from the training set using the Least Absolute Shrinkage and Selection Operator (LASSO) regression; a nomogram was constructed after multivariate logistic regression. The validation set was used to evaluate its discriminatory capacity and efficacy. The model's performance was evaluated through the Brier score, receiver operating characteristic curves, area under the curve (AUC), calibration curves, decision curve analysis (DCA), and clinical impact curves (CIC).

RESULTS

We found an asymptomatic DVT incidence of 27.5% (231/841) on postoperative day three and identified seven predictors: age, chronic heart failure, stroke, tourniquet, postoperative monocyte-to-lymphocyte ratio, and postoperative alpha and D-dimer levels. The predictive model yielded an AUC of 0.737 (95% CI, 0.6933-0.7785), with an external validation AUC of 0.683 (95% CI, 0.6139-0.7716). The Brier score was 0.176, indicating the model's strong robustness in predicting perioperative DVT incidence in arthroplasty. Clinical impact and decision curve analysis revealed that using the proposed nomogram for prediction yielded a net benefit for threshold probabilities of 10-70%.

CONCLUSION

Our risk prediction model demonstrated reasonable discriminative capacity for predicting perioperative DVT risk in arthroplasty. This model may help increase the clinical benefits for patients by promptly identifying high-risk individuals early postoperatively.