Purpose
To construct a novel nomogram model that can predict DVT and avoid unnecessary examination.
Methods
Patients admitted to the hospital with pelvis/acetabular fractures were included between July 2014 and July 2018. The potential predictors associated with DVT were analyzed using Univariate and multivariable logistic regression analysis. The predictive nomogram was constructed and internally validated.
Results
230 patients were finally enrolled. There were 149 individuals in the non-DVT group and 81 in the DVT group. Following analysis, we obtained the final nomogram model. The risk factors included age (OR, 1.037; 95% CI, 1.013–1.062; P = 0.002), body mass index (BMI) (OR, 1.253; 95% CI, 1.120–1.403; P < 0.001); instant application of anticoagulant after admission (IAA) (OR, 2.734; 95% CI, 0.847–8.829; P = 0.093), hemoglobin (HGB) (OR, 0.970; 95% CI, 0.954–0.986; P < 0.001), D-Dimer(OR, 1.154; 95% CI, 1.016–1.310; P = 0.027) and fibrinogen (FIB) (OR, 1.286; 95% CI, 1.024–1.616; P = 0.002). The apparent C-statistic was 0.811, and the adjusted C-statistic was 0.777 after internal validations, demonstrating good discrimination. Hosmer and Lemeshow’s goodness of fit (GOF) test of the predictive model showed a good calibration for the probability of prediction and observation (χ2 = 3.285, P = 0.915; P > 0.05). The decision curve analysis (DCA) and Clinical impact plot (CIC) demonstrated superior clinical use of the nomogram.
Conclusions
An easy-to-calculate nomogram model for predicting DVT in patients with pelvic-acetabular fractures were developed. It could help clinicians to reduce DVT and avoid unnecessary examinations.