TL;DR
Researchers have developed a novel nomogram, a graphical tool for predicting risk, to better stratify patients undergoing lower extremity orthopedic surgery for deep vein thrombosis (DVT). This model, incorporating five key clinical and laboratory factors, demonstrates high accuracy and clinical utility, offering a more personalized approach to DVT prevention and management.
What Problem Does This Paper Address?
Deep vein thrombosis (DVT) is a serious complication following lower extremity orthopedic surgery (LEOS), leading to prolonged hospital stays, impaired recovery, and potentially fatal pulmonary embolism. While DVT is a known risk, accurately predicting which patients are most susceptible remains a challenge. Existing methods often lack precision or are not universally applicable. This paper aims to bridge this gap by developing a sophisticated clinical nomogram specifically for LEOS patients, providing a more accurate and personalized risk assessment tool.
Methods
The study employed a retrospective observational design, analyzing clinical records of patients who underwent LEOS. To build the predictive model, a backward stepwise logistic regression analysis was used to identify significant predictors from a pool of 31 initial variables. The performance of the developed nomogram was rigorously evaluated using metrics like the area under the receiver operating characteristic (ROC) curve (AUC), calibration curves, and decision curve analysis. To handle missing data, a multivariate imputation by chained equations (MICE) method was utilized, a robust approach for improving the reliability of predictive models.
Datasets
The study utilized a dataset of 5773 patients who underwent lower extremity orthopedic surgery at the Wuxi TCM Hospital Affiliated to the Nanjing University of Chinese Medicine between January 2017 and October 2019. This dataset was split into a training cohort (4041 patients) and a validation cohort (1732 patients) to ensure the model's generalizability. No new benchmarks were introduced; the focus was on developing a predictive model using existing patient data.
Key Insights and Ideas
- The Power of a Nomogram: The study highlights the effectiveness of nomograms as intuitive and clinically applicable tools for risk prediction, translating complex statistical models into easily interpretable graphical representations.
- Beyond D-dimer: While D-dimer is a known DVT marker, the research demonstrates that combining it with other factors like age, mean corpuscular hemoglobin concentration (MCHC), platelet distribution width (PDW), and thrombin time (TT) significantly enhances predictive accuracy.
- Addressing Data Gaps: The successful application of MICE for handling missing data underscores its importance in building robust predictive models, especially in retrospective clinical studies where data completeness can be an issue.
- Clinical Utility Matters: The decision curve analysis confirms that the developed nomogram offers a tangible net benefit, suggesting it can guide clinical decisions more effectively than simply intervening for all or none.
Key Results
The developed nomogram identified five key predictors for DVT risk after LEOS: age, MCHC, D-dimer, PDW, and TT. The model achieved a high AUC of 85.9% in the training cohort and 85.7% in the validation cohort, indicating strong discriminative ability. Calibration curves showed good concordance between predicted and observed probabilities. The nomogram demonstrated clinical utility, providing a positive net benefit for threshold probabilities between 0.01 and 0.16. The incidence of DVT in the study population was approximately 1%.
Limitations and Future Work
The authors acknowledge that the study was conducted at a single center, which may limit the generalizability of the model. They suggest that future research should involve multicenter studies to validate and refine the nomogram. Additionally, they propose that future work could incorporate preoperative DVT risk prediction to provide a more comprehensive perioperative assessment.
Summary updated