Construction and validation of a risk prediction model for metabolic syndrome in patients with polycystic ovary syndrome
Objective:To establish a nomogram model for the risk prediction of metabolic syndrome in patients with polycystic ovary syndrome(PCOS),and to verify its predictive efficiency.Methods:The clinical data of 332 patients with PCOS admitted to the Second Clinical Hospital of Shanxi Medical University from December 2022 to December 2023 were retrospectively analyzed.According to the proportion of 7:3,they were randomly assigned to the modeling group(n=218)and the validation group(n=114).The modeling group was divided into the metabolic syndrome group(n=46)and the non-metabolic syndrome group(n=172)according to whether the patients had metabolic syndrome.The general data were compared and analyzed between the modeling group and the validation group as well as the metabolic syndrome group and the non-metabolic syndrome group.Multi-factor logistic regression analysis was used to analyze the risk factors of metabolic syndrome in PCOS patients.The R4.2.3 software package was used to draw a nomogram model,and receiver operating characteristic(ROC)curve was drawn to evaluate the predictive efficacy of the nomogram model.Hosmer-Lemeshow test was used for the goodness-of-fit of the model,and decision curve analysis(DCA)was conducted to evaluate the clinical practicability of the model.Results:There were no statistically significant differences in basic data and levels of biochemical indicators including sex hormone,glucose metabolism,lipid metabolism and liver function between the modeling group and the validation group(P＞0.05).There were no significant differences in age,exercise status,testosterone(T),thyroid-stimulating hormone(TSH),total cholesterol(TC),and homocysteine(HCY)between MS group and non-MS group(P＞0.05).Waist circumference,neck circumference,body mass index(BMI),percentage of family history of hypertension or diabetes,Homeostasis Model Assessment of insulin resistance(HOMA-IR),triglyceride glucose(TyG)index,alanine transaminase(ALT),aspartate transaminase(AST),triglycerides(TG),low-density lipoprotein(LDL)levels in MS group were higher than those in non-MS group,while high-density lipoprotein(HDL),LH,FSH and 25-hydroxyvitamin D3(25-OH-D3)levels in MS group were lower than those in non-MS group,with statistical significance(P＞0.05).According to logistic stepwise regression analysis,waist circumference,HOMA-IR and TyG index were all risk factors for developing metabolic syndrome in PCOS patients in the modeling group(P＜0.05).The area under the curve(AUC)of PCOS patients was 0.937[95％CI(0.906,0.968)],the sensitivity was 89.1％and the specificity was 84.3％.The Hosmer-Lemeshow test showed that there was no statistical difference between the predicted and actual values of the model(x2=-159.1,P=1.000).In DCA,when the threshold probability was 30％-100％,the nomogram model could obtain a higher net benefit value.Conclusions:Waist circumference,HOMA-IR,TyG index were risk factors for developing metabolic syndrome in PCOS patients,and on which the nomogram model constructed could be more intuitive and effective in predicting the risk of developing metabolic syndrome in PCOS patients,with good predictive efficacy.
万方数据
维普
