Linking And Comparing The Montreal Cognitive Assessment And Mini-Mental State Examination In Elderly Surgical Patients

BackgroundCognitive impairment a common comorbidity in elderly surgical patients,which may increase the risk of postoperative complications,such as delirium,prolong the hospital stay,and affect the surgical prognosis.On the other hand,factors related to disease or treatment such as anesthesia and chemotherapy may lead to further cognitive decline.Therefore,preoperative screening and dynamic assessment of cognitive impairment are necessary for the optimization of perioperative management of elderly patients.At present,there is no standardized cognitive assessing tools for the elderly surgical patients,and Montreal Cognitive Assessment scale(Montreal Cognitive Assessment,MoCA)and simple Mental State scale(Mini-getting the State Examination,MMSE)is commonly used in surgery clinical practice and research.To establish a conversion table between the two scales,and to compare the psychometric properties of them in the assessment of elderly surgical patients may help the surgical clinicians and researchers to better understand and apply these two scale.ObjectivesThis study aims to(1)establish a conversion table between the MoCA and MMSE,and validate its applicability in elderly surgical patients;(2)compare the psychometric properties of the MoCA and MMSE in the assessment of elderly surgical patients,and compare their effectiveness of screening for cognitive impairment(CI);(3)explore the effect of MCI and amnestic MCI(aMCI)on postoperative cognitive decline in elderly surgical patients.MethodsA total of 129 elective surgery patients aged 60 years or older were enrolled from a tertiary hospital in Guangzhou by convenience sampling method.The MMSE,MoCA and a battery of four neuropsychological tests were administered to the patients at least 1 day before and 7 days after surgery.At the preoperative assessment,all patients completed the MMSE and MoCA,and 113 patients of them also completed the neuropsychological tests.Fifty-four patients completed the postoperative cognitive assessments.Rasch analysis was used to establish a conversion tables for the MMSE and MoCA,then the Rasch model-based conversion tables and three previously published conversion tables were validated in the current sample.The psychometric properties of the MMSE and MoCA were compared in the co-calibration Rasch model.Furthermore,the sensitivity,specificity,positive/negative predictive value and accuracy of the observed and converted scores of the MMSE and MoCA in screening for cognitive impairment were compare.Cognitive impairment was defined as deficit in at least one neuropsychological test.Finally,repeated measures ANOVA was performed on the cognitive function test scores,with "MCI/non-MCI " or "aMCI/non-aMCI" as between-subject variable and the pre-/post-operative test score as within-subject variable.ResultsRasch analysis of 22 items from the MoCA and MMSE showed that the Rasch dimension of the item set explained 59.4%of the variance,while the eigenvalue of the first contrast was 2.33,indicating that the two scales shared a common dimension.All items fitted well to the Rasch model.Based on the MMSE-MoCA score conversion table established in this study,the intra-class correlation coefficient(ICC)and root mean squared error(RMSE)between the MoCA converted score and the observed scores were 0.82 and 2.56,respectively,and 60.5%MoCA converted scores is within a±2 points range of the observed scores;the ICC and RMSE between the MMSE converted score and the observed scores are 0.84 and 1.36,respectively,and the 95.2%MMSE converted scores is within a ±2 points range of the observed scores,indicating a slightly better accuracy than the 3 previously published MoCA?MMSE conversion tables in the current sample.In the co-calibrated Rasch model of the MMSE and MoCA,The item difficulty of the MMSE range from-3.5 to 0.8 Logit,and the scale has the highest measurement precision in the original score range of 14?25 points.The item difficultiy of the MoCA ranges from-0.7 to 3.0 Logit,and the MoCA has the highest measurement precision in the original score range of 8-24 points.The cognitive ability of the elderly surgical patients ranged from-2.5 to 6.2 Logit,with a mean of 2.0(±1.2)Logit.87(73.1%)patients were rated as cognitive impairment."MoCA observed score<26 points" has the highest sensitivity(78.7%),negative predictive value(50.0%),and accuracy(78.8%)in identified cognitive impairment,while "MMSE converted score<28 points" has the highest specificity(95.8%)and positive predictive value(98.2%).Of the 55 patients who underwent postoperative cognitive assessment,13 were rated as MCI preoperatively,8 of whom were rated as aMCI.With aMCI/non-aMCI as a grouping factor,signifiticant group xmeasurement time interactions were detected for the MoCA total score,MoCA's "memory ability module" and "language ability module" scores.The aMCI group showed a greater decline on these three indicies compared to the non-aMCI group.Conclusion(1)The Rasch model-based MoCA?MMSE conversion table has good applicability in the elderly surgical patients,which provides a new method to compare and integrate the cognitive assessment results of surgical elderly patients more accurately and effectively.(2)The item difficulty hiarachy of the MoCA better matched to the cognitive abilities of elderly surgical patients than those of the MMSE,therefore the MoCA was more precise than MMSE in the assessment of surgical elderly patients.MoCA has good sensitivity and accuracy for cognitive impairment screening,while the MMSE converted score,which were obtained from MoCA observed scores,has high specificity and positive predictive value.Therefore,combining the two may help to improve the accuracy of cognitive impairment screening.(3)aMCI patients tend to be more vulnerable to cognitive decline after surgery,further studies need to pay more attention to this population.