Classification, Prediction, and Concordance of Cognitive and Functional Progression in Patients with Mild Cognitive Impairment in the United States: A Latent Class Analysis
Julie Mouchet, Keith A. Betts, Mihaela Georgieva, Raluca Ionescu‐Ittu, Lesley M. Butler, Xavier M Teitsma, Paul Delmar, Thomas Kulalert, Jingjing Zhu, Neema Lema, Urvi Desai
Abstract
BACKGROUND: Progression trajectories of patients with mild cognitive impairment (MCI) are currently not well understood. OBJECTIVE: To classify patients with incident MCI into different latent classes of progression and identify predictors of progression class. METHODS: Participants with incident MCI were identified from the US National Alzheimer's Coordinating Center Uniform Data Set (09/2005-02/2019). Clinical Dementia Rating (CDR®) Dementia Staging Instrument-Sum of Boxes (CDR-SB), Functional Activities Questionnaire (FAQ), and Mini-Mental State Examination (MMSE) score longitudinal trajectories from MCI diagnosis were fitted using growth mixture models. Predictors of progression class were identified using multivariate multinomial logistic regression models; odds ratios (ORs) and 95% confidence intervals (CIs) were reported. RESULTS: In total, 21%, 22%, and 57% of participants (N = 830) experienced fast, slow, and no progression on CDR-SB, respectively; for FAQ, these figures were 14%, 23%, and 64%, respectively. CDR-SB and FAQ class membership was concordant for most participants (77%). Older age (≥86 versus≤70 years, OR [95% CI] = 5.26 [1.78-15.54]), one copy of APOE ɛ4 (1.94 [1.08-3.47]), higher baseline CDR-SB (2.46 [1.56-3.88]), lower baseline MMSE (0.85 [0.75-0.97]), and higher baseline FAQ (1.13 [1.02-1.26]) scores were significant predictors of fast progression versus no progression based on CDR-SB (all p < 0.05). Predictors of FAQ class membership were largely similar. CONCLUSION: Approximately a third of participants experienced progression based on CDR-SB or FAQ during the 4-year follow-up period. CDR-SB and FAQ class assignment were concordant for the vast majority of participants. Identified predictors may help the selection of patients at higher risk of progression in future trials.