A Prediction Model Study Of Osteoporosis In Early Menopausal Women With Type 2 Diabetes Mellitus

Objective:Osteoporosis(OP)is a slow-progressing bone metabolic disorder and is common in older adults.Frailty fractures are the most frightening outcome of osteoporosis and are a major cause of disability and death in many elderly people.The probability of osteoporosis fracture reached more than 50% in the population over 50 years old,and the proportion of those who died from various complications within 1 year after osteoporotic fracture was as high as 20%,while about 50% of those who survived got lifetime disabled,which significantly impairs the quality of life and caused heavy family,social and economic burden.Menopause and type 2 diabetes,as the main risk factors for osteoporosis,often co-occur in elderly women.With the increase of type 2 diabetes in recent years,the number of this osteoporotic high-risk group has also increased significantly.Current clinical guidelines recommend the age of 65 for the first time screening for osteoporosis,combined with a lack of facility of BMD testing equipment,leading to the fact that most type 2 diabetic women under 65 in early menopause do not know their bone conditions,increasing the risk of frailty fractures.Some existing clinical prediction tools for osteoporosis in the general population have not achieved ideal results in the prediction of osteoporosis in this special population.Therefore,clinical and epidemiological scholars must develop targeted osteoporosis prediction model tools for early menopausal women with type 2 diabetes.The aims of this study: 1)To screen high-risk exposure factors for osteoporosis in early postmenopausal women with type 2 diabetes 2)To exploit predictive metabolomic biomarkers for osteoporosis in early menopausal women with type 2 diabetes 3)To develop simple and effective prediction model tools for osteoporosis in early menopausal women with type 2 diabetes according to different application scenarios.Methods: This study is a cross-sectional case-control study based on people living in Shen-Yang and its ten subordinate districts.Sheng-Jing hospital affiliated to China medical university is taken as the research center,and women with type 2 diabetes in early menopause are recruited as subjects voluntarily.A total of 200 project volunteers were recruited,and 123 subjects were selected after several rounds of data analysis.Complete case data include General information,WHO-FRAX osteoporotic fracture prediction questionnaire,FI-70 questionnaire,PSQI questionnaire,MNA questionnaire,Blood tests included routine blood test,liver function,renal function,serum calcium,blood ion concentrations,glycosylated hemoglobin and glycosylated albumin series,25-hydroxyvitamin-D,lipoprotein cholesterol series,lactate dehydrogenase series,triathlon hepatitis,urine routine.Osteoporosis was diagnosed by the DXA examination.Epidata 3.1 software was used for data collection,collation,and verification.Risk factors were selected for statistical description based on medical expertise and clinical experience.T-test was used for inter-group comparison of continuous variables conforming to normal distribution and homogeneity of variances,otherwise,a non-parametric test was used.A Chi-square test was used to compare discrete variables between groups.The target compounds were separated by a Waters ACQUITY UPLC HSS T3(2.1 mm 100 mm,1.8 m)liquid chromatographic column using Agilent 1290(Agilent Technologies)high ultra-performance liquid chromatography.Thermo Q Exactive Orbitrap mass spectrometer is under the control of the software(Xcalibur,version 4.0.27,Thermo)to collect the first-level and second-level mass spectrometry data.The R program package(XCMS in the kernel)is independently written to carry out peak recognition,peak extraction,peak alignment,and integration.In the prediction model construction part,risk factors were incorporated into the Logistic regression model for multi-factor stepwise regression analysis to calculate the relationship between the clinical characteristics,blood test indicators and the incidence of osteoporosis,expressed by odds ratio(OR)and 95% confidence interval(95% CI).SPSS 25.0 was used for reliability and validity analysis,and R 3.6.0 was used for the analysis of other statistical methods.P <0.05 in the bilateral test was considered to be statistically significant.Results: The proportion of osteoporosis in 123 subjects was 33.3%,and 13 clinical characteristics and blood test indicators were finally selected after adjustment for multiple indicators.The result manifest osteoporosis subjects have a lower average level of platelet(OR=0.985,95%CI =0.975-0.996,P=0.005),25-(OH)-D(OR =0.851,95% CI=0.776-0.933,P=0.001),BMI(OR=0.720,95%CI=0.604-0.859,P<0.001),Serum potassium(OR =0.030,95%CI =0.004-0.242,P=0.001),calcium(OR<0.001,95%CI =0.000-0.010,P=0.001),and a higher level of Hba1c(OR=1.396,95%CI =1.084-1.797,P=0.010),TBil(total bilirubin)(OR =1.283,95%CI =1.120-1.469,P<0.001),Beta-glutamine transferase(OR =1.026,95%CI =1.010-1.043,P=0.001)(all P <0.05).The subjects with an FI-70 of more than 0.25 had significantly higher rates of osteoporosis than those less than 0.25(OR =3.266,95%CI =1.318-8.091,P=0.011).Those who were inactive had significantly higher rates of osteoporosis than those who exercise regularly(OR =0.365,95%CI =0.138-0.963,P=0.042)(all P <0.05).No significant statistical difference was found between other characteristics.Neopterin was selected as a predictive metabolic biomarker for early menopausal women with type 2 diabetes.Serum samples from 55 subjects were selected to make a series of cross-verify tests by ELISA on Neopterin,Adiponectin,osteocalcin,and glycated hemoglobin.Neopterin was proved to be an effective predictor of osteoporosis in early menopausal women with type 2 diabetes.The prediction models of osteoporosis for early menopausal women with type 2 diabetes in different scenarios were constructed and verified in part three.A community screening prediction model of osteoporosis for early menopausal women with type 2 diabetes is Logit(P)=7.748-0.358×(BMI)+1.511×(FI-70)-1.489×(education level)-0.954×(activity),The area under the ROC curve of the prediction model was 0.823,the specificity was 72.0% and the sensitivity was 85.4.A clinical screening prediction model(without neopterin)of osteoporosis for early menopausal women with type 2 diabetes is Logit(P)=1.109-0.076×(25-OH-D)-0.011×(Platelet)+0.165×(TBil),The area under the ROC curve of the prediction model was 0.778,the specificity was 80.5% and the sensitivity was 68.3.A clinical screening prediction model(including neopterin)of osteoporosis for early menopausal women with type 2 diabetes is Logit(P)=-6.330-0.142×(25-OH-D)+0.959×(Neopterin)+0.187×(TBil),The area under the ROC curve of the prediction model was 0.917,the specificity was 95.1%,and the sensitivity was 82.9.A precise screening prediction model of osteoporosis for early menopausal women with type 2 diabetes is Logit(P)=40.735-0.678×(BMI)-0.225×(25-OH-D)-13.906×(Calcium)+0.173×(TBil)+1.136×(Neopterin),The area under the ROC curve of the prediction model was 0.967,the specificity was 93.9%,and the sensitivity was 87.8.Conclusion: Clinical characteristics and blood test indexes of single-factor analysis results show that a normal or higher level of BMI,calcium concentration,platelet,serum potassium,25-hydroxyvitamin D,a lower level of gamma-glutamine transferase,glycated hemoglobin and TBil,is helpful to reduce the incidence of osteoporosis in early postmenopausal women with type 2 diabetes.A higher risk of osteoporosis was in the people with frailty index over 0.25,lower education level,and lack of physical activity.Neopterin is an effective metabolomic biomarker of osteoporosis in early menopausal women with type 2 diabetes.The prediction models and scoring nomograms applied in different scenarios all obtained high predictive power,which provided important references for the clinical and epidemiological screening of osteoporosis in early menopausal women with type 2 diabetes.