Open AccessCognitive practice effects delay diagnosis of MCI: Implications for clinical trials
Author(s) -
SandersonCimino Mark,
Elman Jeremy A.,
Tu Xin M.,
Gross Alden L.,
Panizzon Matthew S.,
Gustavson Daniel E.,
Bondi Mark W.,
Edmonds Emily C.,
Eglit Graham M.L.,
Eppig Joel S.,
Franz Carol E.,
Jak Amy J.,
Lyons Michael J.,
Thomas Kelsey R.,
Williams McKenna E.,
Kremen William S.
Publication year - 2022
Publication title -
alzheimer's and dementia: translational research and clinical interventions
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 2.49
H-Index - 30
ISSN - 2352-8737
DOI - 10.1002/trc2.12228
Subject(s) - alzheimer's disease neuroimaging initiative , neuroimaging , incidence (geometry) , cognitive impairment , cognitive decline , cognition , medicine , clinical trial , psychology , effects of sleep deprivation on cognitive performance , clinical psychology , dementia , disease , psychiatry , physics , optics
Abstract Introduction Practice effects (PEs) on cognitive tests obscure decline, thereby delaying detection of mild cognitive impairment (MCI). Importantly, PEs may be present even when there are performance declines, if scores would have been even lower without prior test exposure. We assessed how accounting for PEs using a replacement‐participants method impacts incident MCI diagnosis. Methods Of 889 baseline cognitively normal (CN) Alzheimer's Disease Neuroimaging Initiative (ADNI) participants, 722 returned 1 year later (mean age = 74.9 ± 6.8 at baseline). The scores of test‐naïve demographically matched “replacement” participants who took tests for the first time were compared to returnee scores at follow‐up. PEs—calculated as the difference between returnee follow‐up scores and replacement participants scores—were subtracted from follow‐up scores of returnees. PE‐adjusted cognitive scores were then used to determine if individuals were below the impairment threshold for MCI. Cerebrospinal fluid amyloid beta, phosphorylated tau, and total tau were used for criterion validation. In addition, based on screening and recruitment numbers from a clinical trial of amyloid‐positive individuals, we estimated the effect of earlier detection of MCI by accounting for cognitive PEs on a hypothetical clinical trial in which the key outcome was progression to MCI. Results In the ADNI sample, PE‐adjusted scores increased MCI incidence by 19% ( P < .001), increased proportion of amyloid‐positive MCI cases (+12%), and reduced proportion of amyloid‐positive CNs (–5%; P ’s < .04). Additional calculations showed that the earlier detection and increased MCI incidence would also substantially reduce necessary sample size and study duration for a clinical trial of progression to MCI. Cost savings were estimated at ≈$5.41 million. Discussion Detecting MCI as early as possible is of obvious importance. Accounting for cognitive PEs with the replacement‐participants method leads to earlier detection of MCI, improved diagnostic accuracy, and can lead to multi‐million‐dollar cost reductions for clinical trials.