A Nomogram Model For Individualized Prediction Of Mild Cognitive Impairment In Patients With Parkinson’s Disease

Objective: To establish and evaluate a nomogram model for individualized prediction of mild cognitive impairment(MCI)in patients with Parkinson’s disease(PD)by analyzing the relative factors of MCI in patients with PD.Methods : 95 patients with PD who were admitted to the Department of Neurology,Affiliated Hospital of Guilin Medical College from October 2016 to October 2018 were successively enrolled.Clinical medical data collecting,blood biochemical detection and motor function evaluation were consummated,and the cognitive function assessment was conducted by utilizing Montreal Cognitive Assessment(MoCA).The Lasso regression model and multi-factor Logistic regression model were used to analyze the relative factors of MCI occurrence in PD patients.R-software was used to establish a nomogram model for predicting MCI in PD patients.Harrell’s Cstatstic calculation consistency coefficient(C-index)was used to evaluate the discrimination of the nomogram model.Calibration curve and decision curve were plotted to check the consistency and clinical applicability of the nomogram.Bootstrap method was used to repeat the sampling 1000 times to internally validate the nomogram model.Results: Among 95 cases of PD patients,there were 69 cases of patients with MOCA score ≥26,26 cases of patients with MOCA score ≥21 and <26,while 0 case of patients with MOCA score <21.On the basis of PD-MCI diagnostic standard,the 69 cases of patients had no cognitive impairment,while MCI occurred to the 26 cases of patients.The incidence of PD-MCI in 95 patients with PD was 27.37%.Lasso regression and multivariate Logistic regression results suggested that PD course,Hoehn/Yahr stage(HY stage),diabetes,blood uric acid(BUA),glycosylated hemoglobin(HbA1c),and plasma homocysteine(HCY)were important relative factors of MCI in patients with PD.The original C-index of this model was 0.927(95% CI:0.8192-1.0348),which proved that this model had favorable degree of distinction.The C-index after independent internal verification was 0.873,which indicated that this model had excellent conformity.The calibration curve revealed that this model had good consistency.The decision-making curve showed that in this model,the clinical net profit rate for the patients reached the highest point when the probability thresholds of MCI were between 8.0% and 87.0%.Conclusions: PD course,H-Y stage,diabetes,BUA,HbA1 c and plasma HCY are important relative factors of MCI in patients with PD.According to the above six clinical important relative factors,a nomogram model for individualized prediction of MCI in patients with PD is established.It has good degree of discrimination and consistency,with high clinical application value.