Risk Prediction Of New Vertebral Compression Fractions (VCF) Following Percutaneous Vertebroplasty (PVP) In Old Women After Menopause With Osteoporotic Vertebral Compression Fractures (OVCFs)

Objective: To determine the risk predictors of new vertebral compression fractions following percutaneous vertebroplasty(PVP)in old women after menopause with osteoporotic vertebral compression fractures(OVCFs)and build the risk prediction model.Methods: We retrospectively collected the clinical and imaging data of postmenopausal old women patients who are diagnosed with OVCF and treated with PVP between January 2008 and December 2016 in Zhongda Hospital Southeast University.The patients between January 2008 and December 2013 were grouped into the training cohort.The remainder were grouped into validation cohorts.Every patients were examined with computer tomography(CT)and magnetic resonance imaging(MRI)before treatment to be diagnosed with OVCF.3 days after PVP,all patients need to review the CT.We collect the clinical,radiological and follow-up data.In the training cohort,we use the Log-Rank test analysis and the Cox proportional hazard regression analysis to identify the risk predictors of the new VCF following PVP.Finally,we create the nomogram to show the visualization of the short-term and long-term probability of new VCF according to these risk factors.The accuracy of the model was validated internally and externally using the training and validation cohorts respectively.Results: We enrolled 324 patients(training cohort: n=196,validation cohort:n=128)and 441 vertebras(training cohort: n=274,validation cohort: n=167).In training cohort,there were 64 patients who suffered from the new VCF following the first PVP(32.7%).After univariate and multivariate analysis,the fracture severity(P=0.001),the number of vertebra involved(P=0.001),the intradiscal cement leakage(p=0.001),the history of fracture(p=0.001),the age group(P<0.001),hypertension(p=0.007)and the average CT value of vertebra(p<0.001)are identified as the risk factors of new VCF(P<0.05).The c statistic of the model was 0.82 in the training cohort and 0.76 in the validation cohort.The new VCF 1 months,1 year and 2 years after PVP would not happen,when patients have no risk factors.But when having the all risk factors,the short-term probabilities of new VCF were over 80%.Conclusion: Age older than 75 y,hypertension,history of fracture,greater fracture severity,a greater number of vertebra involved,intradiscal cement leakage and a smaller CT value of adjacent vertebra are the independent risk factors of new vertebral fractions after PVP.The nomogram is helpful to predict the short-term and long-term probability of the new VCF after PVP.