Construction of a Nomogram prediction model for obstructive sleep apnea hypopnea syndrome complicated with coronary heart disease based on four-variable screening tool and conventional risk factors

YAO Yanli; QIU Xuan; CHEN Yulan; Gulimire Aimaiti; Ayiguzaili Maimaitimin; LIANG Zeyu

doi:10.3969/j.issn.1674-3806.2025.02.11

Abstract：Objective To construct a Nomogram prediction model for obstructive sleep apnea hypopnea syn-drome(OSAHS)complicated with coronary heart disease(CHD)based on four-variable screening tool and conventional risk factors.Methods A total of 1 224 OSAHS patients who were admitted to the First Teaching Hospital of Xinjiang Medical University from February 2020 to March 2024 were enrolled.The patients were randomly divided into a training set(857 cases)and a validation set(367 cases)in a ratio of 7:3.According to the complicating CHD of the patients in the training set,these patients were divided into OSAHS+CHD group(277 cases)and OSAHS+non-CHD group(580 cases).The independent variables were screened by LASSO regression,and the influencing factors of OSAHS complicated with CHD were analyzed by using multivariate logistic regression,and a Nomogram prediction model was established.Receiver operating characteristic(ROC)curve,calibration curve,Hosmer-Lemeshow test,and decision curve analysis(DC A)were used to evaluate the differentiation degree,calibration degree,fitting degree and clinical effectiveness of the model,respectively.Results Eight independent variables were screened out by using LASSO regression analysis,which were male,age≥45 years,a history of hypertension,body mass index(BMI)≥25 kg/m2,high-density lipoprotein cholesterol(HDL-C),apnea-hypopnea index(AHI)≥30 times/hour,mean arterial oxygen saturation(MSaO2)≥90％,and four-variable screening tool scores≥10.5 points.The results of multivariate logistic regression analysis showed that male,age≥45 years,a history of hypertension,BMI ≥25 kg/m2,AHI ≥30 times/hour and four-variable screening tool scores≥10.5 points were independent risk factors for OSAHS complicated with CHD(P＜0.05).Higher HDL-C level was an independent protective factor for OSAHS complicated with CHD(P＜0.05).A Nomogram prediction model for OSAHS complicated with CHD was constructed based on the 7 influencing factors.The ROC curve showed that area under the curve(AUC)(95％CI)of the model in the training set was 0.778(0.747-0.810),with a sensitivity of 71.80％,and a specificity of 71.20％,and AUC(95％CI)of the model in the validation set was 0.797(0.752-0.842),with a sensitivity of 78.80％,and a specificity of 70.80％,showing a high differentiation degree.The calibration curve showed that the prediction probability risk of OSAHS complicated with CHD predicted by the model was basically consistent with the actual probability,with a good calibration degree.The results of Hosmer-Lemeshow test(training set:x2=8.100,P=0.524;validation set:x2=16.880,P=0.051)proved that the model had a good fitting degree.The decision curve verified the clinical effectiveness of the model.Conclusion Male,age≥45 years,a history of hypertension,BMI≥25 kg/m2,AHI≥30 times/hour,and four-variable screening tool scores ≥10.5 points are independent risk factors for OSAHS complicated with CHD.Higher HDL-C level was an independent protective factor for OSAHS complicated with CHD.The Nomogram prediction model constructed based on the 7 influencing factors has better predictive value for OSAHS complicated with CHD.