Risk scores of incident mild cognitive impairment in a Beijing community-based older cohort

Objective: It is very important to identify individuals who are at greatest risk for mild cognitive impairment (MCI) to potentially mitigate or minimize risk factors early in its course. We created a practical MCI risk scoring system and provided individualized estimates of MCI risk.Methods: Using data from 9,000 older adults recruited for the Beijing Ageing Brain Rejuvenation Initiative, we investigated the association of the baseline demographic, medical history, lifestyle and cognitive data with MCI status based on logistic modeling and established risk score (RS) models 1 and 2 for MCI. We evaluated model performance by computing the area under the receiver operating characteristic (ROC) curve (AUC). Finally, RS model 3 was further confirmed and improved based on longitudinal outcome data from the progression of MCI in a sub-cohort who had an average 3-year follow-up.Results: A total of 1,174 subjects (19.8%) were diagnosed with MCI at baseline, and 72 (7.8%) of 849 developed MCI in the follow-up. The AUC values of RS models 1 and 2 were between 0.64 and 0.70 based on baseline age, education, cerebrovascular disease, intelligence and physical activities. Adding baseline memory and language performance, the AUC of RS model 3 more accurately predicted MCI conversion (AUC = 0.785).Conclusion: A combination of risk factors is predictive of the likelihood of MCI. Identifying the RSs may be useful to clinicians as they evaluate their patients and to researchers as they design trials to study possible early non-pharmaceutical interventions to reduce the risk of MCI and dementia.

Automated summary

Combining baseline demographic, medical history, lifestyle, and cognitive factors can predict the likelihood of mild cognitive impairment (MCI), and adding baseline memory and language performance can improve the accuracy of MCI prediction.