Risk Factors And Prediction Model Of Fragility Fractures In Postmenopausal Women With Osteopenia

Objective:The large proportion of postmenopausal women with osteopenia is a major contributor to the occurrence of fragility fractures,and studies have confirmed that Fracture Risk Assessment Tool(FRAX)underestimates the risk of fragility fractures in this population.The study investigated the risk of fragility fractures in postmenopausal women with osteopenia,explored the factors associated with fragility fractures in postmenopausal women with osteopenia and developed a prediction model,and compared it with FRAX to provide a basis for preventing fragility fractures and promoting bone health in this population.Methods:A retrospective study method was used.A total of 394 postmenopausal women with osteopenia who were hospitalized in Jiangxi Provincial People’s hospital during the same period were identified by reviewing our bone density examination reports from March 2020 to September 2021.Patient medical records were reviewed through the physician workstation,and additional face-to-face or telephone inquiries were conducted for women who met the criteria for inclusion in the platoon but for whom some information was incomplete,resulting in the inclusion of a total of 166 eligible patients.Grouped according to the presence or absence of fragility fractures.The 10-year major osteoporotic fractures probability(MOFP)and 10-year hip fractures probability(HFP)were calculated by logging into the FRAX scoring system and entering the relevant parameters and clicking on the calculation,divided into those with and without Bone Mineral Density(BMD).SPSS26 software was applied to perform univariate analysis on the data of both groups,and for the correlation between BMD and age at different sites,Pearson correlation analysis was used to obtain statistically significant variables(P<0.05).To determine the influencing factors associated with the occurrence of fragility fractures in postmenopausal women with osteopenia,a binary logistic regression model was established,while the efficacy of the model was tested by Receiver operating characteristic curve(ROC)and Hosmer-Lemeshaugh(H-L)goodness of fit methods respectively,and compared with FRAX.The validation of the ten-fold cross-validation method within the prediction model and the creation of nomogram were then performed with R software to obtain the final model for clinical application.Result:1.A total of 166 postmenopausal women with osteopenia were included in this study,including 42 cases in the fragility fracture group(two cases had two fractures,one at different sites and one at the same site,20 vertebral fractures,19 forearm fractures and 4 hip fractures)accounting for 25.3%and 124 cases in the non-fragility fracture group,accounting for 74.7%.All postmenopausal women with osteopenia,mean age(64.08±8.53)years.Comparing the general data of the two groups,age,year of menopause,Barthel score,albumin,osteocalcin,femoral neck BMD,femoral neck T value,total hip BMD,and total hip T value,were statistically significant(P<0.05),and the remaining indicators were not statistically significant.2.Univariate analysis of categorical data in both groups:marital status,number of births,whether living alone,daily sunshine hours,and whether exercising were statistically significant(P<0.05),and the remaining indicators were not statistically significant.3.Comparison of FRAX scores between the two groups:MOFP with BMD,MOFP without BMD,HFP with BMD,and HFP without BMD was statistically significant(P<0.05).4.Correlation analysis between BMD and age at different sites:femoral neck BMD and total hip BMD showed negative correlation with age,with statistical differences(P<0.05)and correlation coefficients of-0.236 and-0.396,respectively,but there was no correlation between lumbar spine BMD and age,with no statistical differences.5.Binary logistic regression analysis of factors influencing fragility fractures:65years and older(OR 2.454,[95%CI 1.073-5.611];P<0.05),live alone(OR 7.314,[95%CI 1.643-32.566];P<0.05)were independent risk factors for the development of fragility fractures,while exercise(OR 0.239,[95%CI 0.101-0.568];P<0.05),and high femoral neck T value(OR 0.372,[95%CI 0.148-0.934];P<0.05)were protective factors for the occurrence of fragility fractures.The formula for calculating the risk of developing fragility fractures:P=e~x/(1+e ~x)*100%,e is an exponential function,the value of e 2.718,X=0.898*age(1 or 2)-0.989*(femoral neck T value)-1.432*exercise(0 or 1)+1.990*live alone(1 or 2)-5.696.6.Evaluation of the efficacy of the prediction model and comparison with FRAX:FRAX with BMD and age,femoral neck T value,live alone,and exercise were more accurate than FRAX without BMD in distinguishing fragility fractures from non-fragility fractures;age,femoral neck T value,live alone,and exercise agreed better than FRAX for the true and predicted risk probabilities of fragility fractures,with or without the inclusion of BMD;in FRAX,at the optimal cutoff values,the corresponding values are:MOFP with BMD 3.7%;MOFP without BMD3.6%;HFP with BMD 0.7%;and HFP without BMD 1.1%.7.Ten-fold cross-validation method was used to internally validate the prediction model,and the average Area Under Curve(AUC)of the 10 times binary logistic regression validation model was 0.76(>0.7),indicating that the model has a good internal validation effect.8.Based on binary logistic regression analysis to establish the nomogram and its application:the corresponding scores for each influencing factor are summed to correspond to the total score as the predicted probability of fragility fracture risk.Age≥65 years scored 37.5,no exercise scored 60,live alone scored 82.5,and femoral neck scored 100 at a T value of-2.4.The total score of each influencing factor was280,and the predicted probability of fragility fracture risk when the total score was280 corresponded to>0.9.According to the best cutoff value of 0.337 of the prediction model,we included P≥0.337 in the population with high risk of fragility fractures and<0.337 in the population with low risk of fragility fractures.Conclusion:1.The most common site of fragility fractures in postmenopausal women is the vertebral body,followed by the forearm.The risk of fragility fractures in postmenopausal women with osteopenia is generally older,lower bone density,lack of exercise and live alone.65 and older,live alone is a independent risk factor for fragility fractures in postmenopausal women with osteopenia,while high femoral neck bone density values and exercise can reduce the risk of fragility fractures in this population.2.Bone loss increases with age,and femoral neck and total hip BMD Better reflect bone loss than lumbar BMD.3.The use of the FRAX to assess the risk of fragility fractures in postmenopausal women with osteopenia,based on intervention thresholds of HFP≥3%or MOFP≥20%,may underestimate the risk of fragility fractures in this population regardless of femoral neck BMD inclusion in the calculation.4.Postmenopausal women are at higher risk of osteoporosis and fragility fractures,This prediction model has certain guiding significance for assessing the risk of fragility fractures in postmenopausal women with osteopenia in clinical practice,but further verification of external data is needed.