Accounting for cognitive practice effects results in earlier detection and more accurate diagnosis of MCI: Biomarker confirmation: Neuropsychology: Longitudinal cognitive assessment in early stages of AD. (7th December 2020)

• Record Type: Journal Article
• Authors: Kremen, William S.; Sanderson‐Cimino, Mark E.; Elman, Jeremy A.; Tu, Xin M.; Gross, Alden L.; Panizzon, Matthew S.; Eglit, Graham M.L.; Jak, Amy J.; Edmonds, Emily C.; Thomas, Kelsey R.; Eppig, Joel S.; Williams, McKenna E.; Bondi, Mark W.; Lyons, Michael J.; Franz, Carol E.
• Is Part Of: Alzheimer's & dementia. Volume 16(2020)Supplement 6
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-12-07
• Subjects: Alzheimer's disease -- Periodicals; Alzheimer Disease -- Periodicals; Dementia -- Periodicals; Démence; Maladie d'Alzheimer; Périodique électronique (Descripteur de forme); Ressource Internet (Descripteur de forme); 616.83
• DOI: 10.1002/alz.044883
• Languages: English
• ISSNs: 1552-5260

Abstract

Abstract: Background: Practice effects (PEs) mask true cognitive decline. Even with declines at follow‐up, PEs can still obscure even steeper declines. Accounting for PEs means that impairment cutoffs are reached earlier. Importantly, mild cognitive impairment (MCI) would be detected earlier. If these diagnoses are valid, there should be more biomarker‐positive MCI cases and fewer biomarker‐positive cognitively normal subjects (CNs). We tested this hypothesis using a replacement‐subjects method. Method: We examined 722 Alzheimer's Disease Neuroimaging Initiative (ADNI) subjects who were CN at baseline (mean age=74.9±6.8). MCI at follow‐up was diagnosed by Jak‐Bondi and Petersen criteria. We identified "pseudo‐replacements" whose baseline age was matched to the age of returnees at their 1‐year follow‐up. Education and estimated premorbid IQ were also matched. Difference scores were calculated at Time 2 between returnees on their second testing and the demographically‐matched pseudo‐replacements on their first testing. Attrition effects were estimated by comparing returnee baseline scores with the full baseline sample. PEs = difference scores – attrition effects. PEs are then subtracted from follow‐up scores, with the resultant scores used in diagnosing MCI. CSF Aβ42 and phosphorylated (p)‐tau were measured 6‐12 months prior to retesting. Result: There were significant PEs on 5 of 6 tests (average Cohen's d=.27). For Jak‐Bondi diagnoses, accounting for PEs at follow‐up resulted in a 24.8% increase in the proportion of MCI cases (p<.000001). There were increased proportions of Aβ42 ‐positive (18.8%) and p‐tau‐positive cases (29.8%), and reduced proportions of Aβ42 ‐positive (5.8%) and p‐tau‐positive (7.9%) CNs (McNemar χ 2 : Aβ42 : p<.0014, p‐tau: p<.028; Figure 1). Results were similar for Petersen diagnoses. Of 26 false negatives (i.e., CNs whose diagnosis changed to MCI after accounting for PEs), 9 (35%) were MCI without PE adjustment at 2‐year follow‐up. Conclusion: Accounting for PEs led to earlier detection of MCI at 1‐year follow‐up. Biomarker data support these being more accurate diagnoses because more cases were biomarker‐positive and more CNs were biomarker‐negative. Longer‐term follow‐up indicated that MCI cases may be detected a year or more earlier if performance is adjusted for PEs. Accounting for PEs has meaningful, real‐world implications for early identification, which may, in turn, improve opportunities to slow disease progression.