Construction And Application Of A Predictive Model For Cognitive Decline Risk In Elderly Patients With Chronic Heart Failure

Objectives1.To investigate the status of cognitive decline in elderly patients with chronic heart failure(CHF),and explore independent risk factors and protective factors that affect the occurrence of cognitive decline in CHF patients.2.To establish a risk model for predicting cognitive decline in elderly CHF patients,providing a basis for medical personnel to develop screening tools for cognitive decline in elderly CHF.MethodThis study extracted relevant risk factors for cognitive decline in elderly CHF patients through literature review.This study used convenience sampling method and recruited 261 elderly CHF patients who met the inclusion and exclusion criteria as the modeling group research subjects between May 2021and March 2022.The sample selection was from the Cardiovascular Department and Geriatric Department of two tertiary A general hospitals in Henan Province.This study used general information questionnaires,biochemical related indicators,Clinical Dementia Rating(CDR),Frailty Phenotype(FP),Montreal Cognitive Assessment(Mo CA),and related scale data to investigate.According to whether the patient has cognitive impairment,they are divided into cognitive impairment group and non-cognitive impairment group.Describe the clinical data of elderly CHF patients using frequency and percentage,usingχ~2Compare the clinical data of elderly CHF patients between two groups for differences between two tests.By using Lasso(Least Absolute Shrinkage and Selection Operator,Lasso)regression,important predictive factors with non-zero coefficients affecting cognitive decline in elderly CHF patients were screened,and then included in the multivariate logistic regression equation to determine independent predictive factors affecting cognitive decline in elderly CHF patients.Based on the analysis of independent predictive factors,a risk column chart model is constructed.Evaluate the calibration and discrimination of the model through calibration curves and receiver operating characteristic Receiver Operating Characteristic curves.And evaluate the clinical practical value of the column chart through Clinical Decision Curve Analysis(DCA).The convenience sampling method was used for external validation.The 132 elderly CHF patients who met the inclusion criteria and were admitted to the cardiovascular and geriatric departments of two tertiary.A general hospitals in Henan Province from April 2022 to September 2022 were selected as the validation group research subjects.Data was collected using the same research tools,and the validation group data was used to draw ROC curves and calibration charts to reflect the predictive performance of the prediction model.The clinical practical value of using clinical decision curves to analyze column charts simultaneously.Result1.In the training group,there were 261 elderly CHF patients,90 of whom had cognitive impairment,with an incidence of 34.48%.Among the 132 elderly CHF patients in the validation group,47 had cognitive impairment,with an incidence of 35.61%.There were statistically significant differences between the cognitive impairment group and the non-cognitive impairment group in terms of age,education level,sleep duration,whether to engage in intellectual activities,NYHA(New York Heart Association)rating,LVEF(Left Ventricular Ejection Fraction)value,whether to have coronary heart disease,nutritional status,whether to have depression,and social support level(P<0.05).2.The results of the Lasso multivariate logistic regression analysis showed that age≥75 years(OR=2.42,95%CI:1.18-4.94,P=0.015),NYHA grade III-IV(OR=2.87,95%CI:1.43-5.77,P=0.003),LVEF≤50%(OR=3.07,95%CI:1.43-6.62,P=0.004),sleep duration≤6 hours(OR=2.29,95%CI:1.16-4.55,P=0.017)Malnutrition(OR=2.49,95%CI:1.21-5.12,P=0.013)and depression(OR=19.99,95%CI:6.13-65.18,P<0.001)are risk factors for cognitive decline in elderly CHF patients(P<0.05);Education level(secondary/high school OR=0.37,95%CI:0.17-0.81,P=0.013;junior college OR=0.16,95%CI:0.06-0.46,P=0.001),whether to engage in intellectual activities(OR=0.37,95%CI:0.19-0.72,P=0.004),and social support(high level OR=0.18,95%CI:0.06-0.55,P=0.003)are protective factors for cognitive decline in elderly CHF patients(P<0.05).3.A nomogram risk prediction model was established based on the risk factors and protective factors that affect cognitive decline in elderly CHF patients.The area under the model ROC curve and the C-index value are 0.854(95%CI:0.809-0.899);The calibration diagram shows that the calibration curve of the model is close to the ideal model;DCA shows that using the nomogram model can bring net clinical benefits.4.The external validation results show that the specificity of the prediction model is 0.694,the sensitivity is 0.872,and the accurate prediction value is 0.758.The area under the ROC curve is 0.852(95%CI:0.786～0.918);The calibration chart shows that the predicted occurrence probability of cognitive impairment is in good agreement with the actual occurrence probability,and DCA shows that it can also bring net benefits to clinical practice.Conclusion1.The incidence of cognitive decline in middle-aged and elderly patients with CHF in this study is high.2.Age≥75 years old,NAHA grade(III-IV),LVEF≤50%,sleep duration≤6 hours,malnutrition and depression are risk factors for cognitive decline in elderly CHF patients.Education,intellectual activity,and social support levels at or above junior high school are protective factors for cognitive decline in elderly CHF patients.3.The nomogram model constructed in this study to predict cognitive decline in elderly CHF patients can effectively predict cognitive decline in elderly CHF patients,with good predictive efficacy and high clinical application value.