Received: 16 December 2020 Revised: 9 March 2021 Accepted: 9 March 2021 Published online: 19 April 2021 DOI: 10.1002/dad2.12184 RESEARCH ARTICLE Markers of early changes in cognition across cohorts of adults with Down syndrome at risk of Alzheimer’s disease Andrew J. Aschenbrenner1 R. Asaad Baksh2,3 Bessy Benejam4 Jessica A. Beresford-Webb5 Antonia Coppus6 Juan Fortea4,7,8 Benjamin L. Handen9 Sigan Hartley10 Elizabeth Head11 Judith Jaeger12,13 Johannes Levin14,15,16 Sandra V. Loosli14 Anne-Sophie Rebillat17 Silvia Sacco17 Frederick A. Schmitt18,19 Kate E. Thurlow2 Shahid Zaman5,20 Jason Hassenstab1 Andre Strydom2,3,21 1 Washington University in St. Louis, Department of Neurology, St. Louis, Missouri, USA 2 Institute of Psychiatry, Psychology, and Neuroscience, Department of Forensic and Neurodevelopmental Sciences, King’s College London, London, UK 3 The London Down Syndrome (LonDownS) Consortium, London, UK 4 Barcelona Down Medical Center, Fundació Catalana Síndrome de Down, Barcelona, Spain 5 Cambridge Intellectual and Developmental Disabilities Research Group, Department of Psychiatry, University of Cambridge, Cambridge, UK 6 Department of Primary and Community Care, Radboud University Medical Center, Nijmegen, The Netherlands 7 Memory Unit and Biomedical Research Institute Sant Pau (IIB Sant Pau), Neurology Department, Hospital de la Santa Creu i Sant Pau, Barcelona, Spain 8 Centro de Investigación Biomédica en Red en Enfermedades Neurodegenerativas (CIBERNED), Madrid, Spain 9 Department of Psychiatry, University of Pittsburgh, Pittsburgh, Pennsylvania, USA 10 Department of Human Development & Family Studies, University of Wisconsin-Madison, Madison, Wisconsin, USA 11 Department of Pathology & Laboratory Medicine, University of California, Irvine, California, USA 12 CognitionMetrics, LLC., Wilmington, Delaware, USA 13 Deptment of Psychiatry and Behavioral Sciences, Albert Einstein College of Medicine, Bronx, New York, USA 14 Department of Neurology, Ludwig-Maximilians-Universität München, Munich, Germany 15 German Center for Neurodegenerative Diseases, Munich, Germany 16 Munich Cluster for Systems Neurology (SyNergy), Munich, Germany 17 Jerome Lejeune Institute, Paris, France 18 Sanders-Brown Center on Aging, University of Kentucky, Lexington, Kentucky, USA 19 Departments of Neurology, Neurosurgery, Behavioral Science, Psychology, Psychiatry, University of Kentucky, Lexington, Kentucky, USA 20 Cambridgeshire & Peterborough NHS Foundation Trust, Cambridge, UK 21 South London and the Maudsley NHS Foundation Trust, London, UK Correspondence Andrew J. Aschenbrenner, PhD, Washington Abstract University in St. Louis, Department of Neurol- ogy, St. Louis, Missouri, USA. Introduction: Down syndrome (DS), a genetic variant of early onset Alzheimer’s dis- Email: a.aschenbrenner@wustl.edu ease (AD), lacks a suitable outcome measure for prevention trials targeting pre- dementia stages. This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring published by Wiley Periodicals, LLC on behalf of Alzheimer’s Association Alzheimer’s Dement. 2021;13:e12184. wileyonlinelibrary.com/journal/dad2 1 of 10 https://doi.org/10.1002/dad2.12184 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2 of 10 ASCHENBRENNER ET AL . Funding information the LuMind IDSC Foundation; the Jérôme Methods: We used cognitive test data collected in several longitudinal aging studies Lejeune Foundation; the Wellcome Trust Strategic Award, Grant/Award internationally from 312 participants with DS without dementia to identify compos- Number: 098330/Z/12/Z; the Medical ites that were sensitive to change over time. We then conducted additional analyses to Research Council, Grant/Award Numbers: MRC MR/S011277/1, MR/S005145/1, provide support for the utility of the composites. The composites were presented to an MR/R024901/1; the Verum Foundation; expert panel to determine the most optimal cognitive battery based on predetermined the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) Munich Clus- criteria. ter for Systems Neurology, Grant/Award Results: There were common cognitive domains across site composites, which were Number: ID 390857198; the National Insti- tutes of Health/Eunice Kennedy Shriver sensitive to early decline. The final composite consisted of memory, language/executive National Institute of Child Health and Human functioning, selective attention, orientation, and praxis tests. Development NIH/NICHD, Grant/Award Number: R01HD064993; the National Insti- Discussion: We have identified a composite that is sensitive to early decline and thus tute for Health Research (NIHR) Cambridge may have utility as an outcome measure in trials to prevent or delay symptoms of AD Biomedical Research Centre, Grant/Award Number: BRC-1215-20014; National Insti- in DS. tutes of Health, Grant/Award Numbers: R01AG031110, U01AG051406 KEYWORDS Alzheimer’s disease, clinical trial outcome, composite measure, Down syndrome, early cognitive decline 1 BACKGROUND declines in memory (recall of new information) and attention occur first in DS,9,13 similar to sporadic AD. Other DS studies have found that Down syndrome (DS), most commonly caused by a triplication of chro- impairments in executive functioning and behavioral and psychological mosome 21, is considered a genetic variant of early onset Alzheimer’s changes precede difficulties in memory.14 Determining which abilities disease (AD). As a consequence of trisomy 21, individuals with DS first show AD-related decline during the transition to the prodromal have three copies of the amyloid precursor protein (APP) gene. phase and identifying the tests most sensitive to change in this period This significantly increases the risk of amyloid beta (Aβ) deposition is vital to determine the most optimal point for an intervention.15 If as plaques and consequently the development of dementia,1 with a given test is found to be the most sensitive (ie, earliest to change), over 95% of individuals eventually developing clinical features of then by implication it will likely be a cognitive modality quickest to AD.2 change. The ultra-high risk for AD in association with a high diagnostic cer- To facilitate the development of the first AD prevention trials in tainty for an underlying Alzheimer pathology in cases with dementia DS, there is a need to refine and adapt current tests of cognition and makes people with DS an important population to consider for ran- identify those which are most sensitive to early AD-related impair- domized controlled trials (RCTs) for interventions that seek to prevent, ments and, therefore, predictive of dementia before the diagnosis can delay, or halt the development and progression of dementia.3 Although be made. Such a test battery would also be valuable to clinicians by pro- there is growing interest in including people with DS in intervention tri- viding them with predictive measures for tracking AD-related decline. als, barriers remain for RCTs, including the need for reliable cognitive To this end, we utilized data from existing longitudinal studies associ- outcome measures of progression during the preclinical to prodromal ated with the Horizon 21 DS consortium (H21 consortium),4 as well as spectrum stages of dementia.4 from two DS research cohorts in the United States. The advantages of Identifying early subtle changes in cognition and diagnosing AD this approach include providing a more diverse sample of cognitive data in DS can be challenging because of the presence of developmen- from individuals with DS than is typically possible in single-site studies. tal cognitive impairments associated with lifelong intellectual disabil- It also capitalizes on the expertise of multiple research groups and pro- ity (ID) and the variability in baseline cognitive functioning across vides an opportunity to cross-validate the findings of one cohort with individuals.5 However, there are several cognitive tests measuring another in the presence of population, cultural, and language differ- memory, verbal fluency, planning, inhibition, attention and visuo-motor ences. abilities which appear to be appropriate for discriminating between Our aim was to use a data-driven approach to identify cognitive those with and without dementia and tracking AD-related decline and tests or test items that are the most sensitive to detecting early cog- progression.6–12 Although these tests show promising results in dis- nitive change in adults with DS. We then sought to use these results tinguishing individuals with and without dementia and may be use- as well as expertise from clinicians and researchers familiar with the ful as cognitive endpoints for RCTs, it is currently unclear which tests cognitive tools to identify the optimal constellation of tests or test show the earliest decline (before dementia can be diagnosed clini- paradigms to constitute a composite cognitive assessment battery to cally) for use in prevention trials. There is evidence to suggest that use in future RCTs in DS and in clinical settings. 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License ASCHENBRENNER ET AL . 3 of 10 2 METHODS RESEARCH IN CONTEXT 2.1 Cohorts 1. Systematic review: A literature search was conducted using the Scopus database. People with Down syndrome Longitudinal data were used from five observational studies on age- (DS) are at an ultra-high risk of developing Alzheimer’s associated cognitive change in DS. These sites included data from disease (AD) as they age. Although there are currently the Horizon21 study group including the London Down Syndrome tests that are effective at identifying those with and with- (LonDownS) Consortium in London, UK,12,13 the Dementia in Down out AD, there is limited research on tests, which are sensi- syndrome (DiDs) research cohort in Cambridge, UK,16,17 and the tive to prodromal changes for use in preventative clinical Down Alzheimer Barcelona Neuroimaging Initiative (DABNI) clinical trials of AD in DS. cohort in Barcelona, Spain,6,18 in addition to several DS longitudinal 2. Interpretation: We provide an important component in research cohorts from the United States (Neurodegeneration in Aging expediting the inclusion of people with DS in preven- Down Syndrome [NiAD] from the Universities of Pittsburgh and tative clinical trials with the development of the H21 Wisconsin, Madison19 ), and the Aging in Down syndrome (ADS) study AD test battery. It consists of measures of memory, lan- from University of Kentucky20 ). Informed consent was obtained from guage/executive functioning, selective attention, orienta- participants in all cohorts. Because the goal of this project was to tion, and praxis and is a short and comprehensive out- define a cognitive composite endpoint that is sensitive to change in come measure. clinically healthy participants, we restricted our analyses to individuals 3. Future directions: Further examination of the H21 AD who were 35 years of age or older (as this is the age group within which battery is now required to examine its association with demonstrable AD pathology is present in DS) and who did not have known AD biomarkers. a clinical diagnosis of dementia at their baseline visit. Furthermore, as we were interested in rates of cognitive decline, we included only participants who had at least two assessments (at baseline and one or more follow-ups) so that a rate of change could be estimated; all participants with a minimum of two assessment timepoints (at least 6 verted this outcome to a Cohen’s d statistic for ease of comparison months apart, with mean length of follow-up between study cohorts across cohorts. varying from 1.97 to 3.9 years) were included regardless of other We repeated steps 2-4 iteratively to evaluate a variety of cogni- comorbidities to ensure a representative sample (those with severe ID, tive composite scores. We first analyzed each cognitive test in isola- significant sensory impairments, or other acute illness that precluded tion; then we averaged two tests together, then three and so on up to cognitive testing were excluded from assessments at the site level). a maximum of six tests in a single composite. This process occurred for For participants with multiple assessment timepoints, a slope was all tests available for a given cohort, resulting in dozens of composite estimated across all timepoints. In total, 312 participants met inclusion scores that consisted of the average of between one and six cognitive criteria; the number of participants by site, number of visits, and other tests. For each composite, the Cohen’s d scores were extracted from demographic variables are listed in Table 1. Each site administered the LME model and then rank ordered in terms of absolute magnitude. a unique test battery spanning several different cognitive domains, Composites with the highest scores were retained for further analy- and the specific tests used in each battery are listed in Table S1 in the sis/discussion. Any test that appeared in three of the five top composite supplementary material. scores was assumed to tap a cognitive domain (eg, attention, memory, executive function) that shows large and consistent decline in individu- als who likely have preclinical AD. We then used these tests to form an 2.2 Statistical analysis “optimal” composite consisting of the z-scored average of each of the measures. Due to differences in the cognitive tests that were administered as well as the length and frequency of follow-up, all analyses were con- ducted separately within each available cohort. We were interested in 2.3 Evaluation of the optimal composite the rates of change on each cognitive test and therefore conceptual- ized “years since baseline visit” as the time variable (hereafter referred After selecting the composite within each cohort with the greatest sen- to as “time”). Our modeling strategy then proceeded in several steps. sitivity to decline, we conducted additional analyses to provide sup- First, all neuropsychological scores were z-scored to the baseline visit. port for the utility of such a composite in a global clinical trial. First, Second, a linear mixed-effects model was constructed using the “lme4” we examined individual rates of change on the composite score within package21 to predict scores on a given cognitive test from the “time” each cohort as a function of critical demographic variables. Specifically, variable. Random intercepts across participants were included in all annualized rates of change for each participant were extracted from a models. Third, we extracted the beta weight of “time” from the model linear mixed-effects model predicting change in the composite score that indexes the annualized rate of change in z-scores. Fourth, we con- over time. These rates of change were then further regressed onto 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 4 of 10 ASCHENBRENNER ET AL . TA B L E 1 Demographic data from each cohort Barcelona London Pittsburgh/Wisconsin Cambridge Kentucky N 128 103 31 19 31 Age (y) 43.8 (6.5) 47.4 (6.6) 42.8 (4.6) 44.1 (5.2) 44.6 (7.0) Sex M = 63 (49%) M = 59 (57%) M = 19 (61%) M = 11 (58%) M = 9 (29%) F = 65 (51%) F = 44 (43%) F = 12 (39%) F = 8 (42%) F = 21 (68%) Level of Intellectual Mild = 34 (27%) Mild = 40 (39%) Mild = 14 (45%) Mild = 7 (37%) Mild = 18 (58%) Disability Mod = 72 (56%) Mod = 41 (40%) Mod = 7 (23%) Mod = 11 (58%) Mod = 13 (42%) Severe = 22 (17%) Severe = 22 (22%) Severe = 2 (6%) Severe = 0 (0%) Severe = 0 (0%) Hearing Problems No = 112 (88%) No = 23 (22%) NA No = 12 (63%) No = 4 (13%) Yes = 14 (11%) Yes = 78 (76%) Yes = 6 (32%) Yes = 25 (81%) Vision Problems No = 31 (24%) No = 80 (78%) No = 16 (52%) No = 3 (16%) No = 8 (26%) Yes = 34 (27%) Yes = 22 (21%) Yes = 15 (48%) Yes = 5 (26%) Yes = 21 (68%) Psychotropic Medication No = 51 (40%) No = 73 (71%) No = 23 (74%) No = 17 (89%) NA Yes = 38 (30%) Yes = 22 (21%) Yes = 8 (26%) Yes = 2 (11%) APOE NA No e4 = 75 (73%) No e4 = 28 (90%) No e4 = 13 (68%) No e4 = 11 (35%) Has e4 = 22 (21%) Has e4 = 3 (10%) Has e4 = 4 (21%) Has e4 = 3 (10%) Number of Visits 3.1 (1.0) 2 (0) 2.3 (0.5) 3.4 (0.8) 3.6 (0.8) Mean length of 2.5 (1.1) 1.97 (0.06) 3.9 (1.1) 3.7 (1.3) 2.6 (0.7) Follow-up (y) Note: Variables are listed as mean (standard deviation) for continuous variables and N (percentage) for categorical variables. Percentages may not sum to 100% due to the presence of missing demographic information. age, sex, pre-study level of ID at baseline, and length of follow-up to 3 RESULTS determine which, if any, entry criteria influence rates of change. Sec- ond, exploratory power analyses were conducted using the longpower The average Cohen’s d of the top five cognitive composites in each package22 in R. Specifically, we estimated the number of participants cohort ranged from small: 0.28 for Barcelona, to moderate: 0.44 needed to detect either a 50% slowing or a 30% slowing in the rate of for Kentucky, to large: 1.13 for London, 1.20 for Cambridge, and change for either a 2-year or 3-year clinical trial, as this range repre- 1.8 for Pittsburgh/Wisconsin. Cognitive tests, which appeared in at sents a realistic and clinically significant effect size for AD prevention least three of the top five composites are shown in Table 2, colored trials. by cognitive domain. Measures of memory, language, attention, and praxis-type tests were consistently represented across all cohorts. The means of the “optimal” composite at the baseline and follow- 2.4 Consensus discussion up visits are plotted in Figure 1. Annualized rates of change (Fig- ure 2) extracted from linear mixed effects showed cognitive decline Finally, considering the above information, an expert panel of clinicians on this composite in all five samples (Pittsburgh/Wisconsin: β = −0.10, and researchers from each site represented in the consortium includ- SE = 0.02, P < .001; Kentucky: β = −0.09, SE = 0.05, P = .055; London: ing co-authors (RAB, BB, JBW, BLH, JJ, SVL, SS, FAS, JH, AS) met to β = −0.11, SE = 0.02, P < .001; Barcelona: β = −0.03, SE = 0.01, discuss the feasibility of utilizing specific tests in a global clinical trial. P = .03; Cambridge: β = −0.07, SE = 0.03, P = .02). Regression anal- The clinicians and researchers who took part in the consensus meeting yses of the individual rates of change on this composite indicated were all well established and experienced in working with adults with that age at baseline (β = −0.001, SE = 0.0003, P < .001) and pre- DS, DS and AD, and clinical trials involving patients with AD. After a study level of ID (β = 0.01, SE = 0.003, P = .003) significantly pre- series of group discussions, an optimal cognitive battery was defined dicted rates of change. Sex and length of follow-up were not significant based on the following criteria: (1) it must measure domains that the predictors. statistical analyses indicated show the largest decline over time; (2) the tests must be amenable to administration in a global trial (ie, not culturally specific, limit language effects if applied across different lan- 3.1 Power analysis guages); (3) it must be feasible in terms of administration; (4) it must have face validity as outcome measures of decline related to AD in As the London cohort had a consistent length of follow-up of ≈2 years RCTs; and (5) it must have low floor and no major ceiling effects in and a good sample size, this cohort was used as the natural history healthy adults with DS. group to evaluate power. Power analyses were conducted to estimate 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License ASCHENBRENNER ET AL . 5 of 10 TA B L E 2 List of tests that appeared in the top five composites for each cohort Barcelona London Pittsburgh/Wisconsin Cambridge Kentucky CAMCOG-DS Orientation CAMCOG-DS Forward Corsi (−0.43) SIB Orienting to Name (−0.03) Orientation (−0.78) CRT Cued Delayed Recall (−0.24) Digits span backward (−.09) CANTAB Pal (−0.58) (−0.87) CAMCOG-DS Praxis (−0.13) Finger-nose pointing Purdue Pegboard Test (−1.13) CAMCOG-DS Praxis SIB Praxis (−0.16) (−0.62) (−0.71) CAMCOG-DS Expressive one-word (−1.1) CAMCOG-DS Fluency Comprehension (−.11) NEPSY Fluency (−0.90) (−0.42) CAMCOG-DS Language (−0.71) Cancellation task (−.09) CANTAB SRT (−0.82) CAMCOG-DS Attention (−0.31) CEFA Tower of London CEFA Cats and Dogs (0.17) SIB Social Interactions (−0.84) _(−0.15) SIB Visuospatial ability (−0.38) Cohen’s D of the rate of change for each test is listed in parentheses. CAMCOG, Cambridge Cognition Examination; CANTAB SRT, Cambridge Executive Functioning Assessment Simple Reaction Time; CANTAB PAL, Cambridge Executive Functioning Assessment Paired Associate Learning; CEFA, Cambridge Executive Functioning Assessment; NEPSY, A Developmental NEuroPSYcho- logical Assessment; CRT, Cued Recall Test; SIB, Severe Impairment Battery. Cognitive domains. Memory & Orientation. Attention/Praxis. Executive Functions. Language. Praxis. Visuospatial abilities Optimal Composite Means at Annualized change for each cohort Baseline and Follow−up 0.00 0.2 Annual Change cohort −0.05 0.0 Composite Barcelona Cambridge −0.10 −0.2 Kentucky London Pittsburgh/ −0.15 Wisconsin −0.4 Barcelona Cambridge Kentucky London Pittsburgh/ 1 2 Wisconsin Time Cohort F I G U R E 1 Optimal composite means from each cohort at the F I G U R E 2 Annualized rates of change extracted from the linear baseline and follow-up visit mixed effects 3.2 Consensus discussion the required sample size to detect a given effect size assuming 80% Taking into account the above data, and the predefined parameters for power, with trial durations of either 2 or 3 years. Results of the power selecting tests of interest, the following decisions were made from the analyses are summarized in Table 3. An expected 210 participants (con- tests listed in Table 2: two tests were chosen as the consensus tests for fidence interval [CI] = 110-545) would be needed in each group (treat- memory abilities; a modified version of the Cued Recall test (mCRT)6,23 ment and placebo) to detect a 30% slowing of decline. This number and the CANTAB paired associate learning test (PAL)24 ; this was due to could be reduced to 150 per group (CI = 79-389) if assessments were the importance of memory (by accepted diagnostic criteria) in the early collected every 6 months rather than annually due to improvement in stages of AD-related decline and the variety of different tests that were stability of measurements. present in the composites from each cohort. With regard to language 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 6 of 10 ASCHENBRENNER ET AL . TA B L E 3 Estimated sample sizes (means and confidence intervals) needed to detect a given effect size for a given trial duration and assessment frequency in the London cohort 50% Effect Size 30% Effect Size Frequency 2y 3y 2y 3y 6 mo 151 (79-392) 54 (28-141) 419 (220-1090) 150 (79-389) Annual 189 (99-490) 75 (40-196) 524 (275-1362) 210 (110-545) Note: estimates are per treatment arm. abilities, verbal fluency (a simple animal fluency task with 60-second issues associated with single research group studies such as admin- time limit) was the consensus test. There was a consensus that exist- istration and language or cultural effects. Earlier studies highlighted ing executive functioning tests may not be reliable in this population the importance of decline in memory,23,28 attention,13,25 and executive and frequently demonstrate floor effects, as they are too difficult in functioning,29 with declines involving memory and attention occurring DS populations. Because ideational fluency is often regarded as a mea- before that of declines in measures of executive function in machine- sure of executive functioning, it was felt that verbal fluency could serve learning models.9 as a measure of this domain as well. The Cancellation task,25 a test of Because AD in DS has, just like in other populations, a strong rela- selective attention, was selected as the measure of attentional abilities. tionship with age, it is to be expected that older adults (and cohorts Although the CANTAB simple reaction time test (SRT) would have been with higher mean age) would show larger changes on cognitive mea- ideal given the inclusion of the CANTAB PAL, the SRT task has been sures over time, which was confirmed in our analyses. We also showed discontinued and is no longer available for use in research studies. The that the degree of premorbid ID influenced effect sizes, which may Purdue Pegboard test26 was chosen as the test of praxis-type abilities be due to those with more severe ID having lower baselines scores, and finally; the orientation subtest of the CAMCOG-DS27 was the con- thus limiting the amount of decline that can be measured over time, sensus test for orientation (see Table 4 for full details). and/or due to greater variability of scores for a given individual, as it is harder to administer the test reliably for those who are more intellectually impaired. Although we did not demonstrate any signif- 4 DISCUSSION icant relationships between length of follow-up and effect sizes of decline on measures when age, ID level, and sex are taken into account, This is the first comprehensive analysis of cognitive decline associated this may become apparent in studies with longer follow-up and could with AD in DS across cohorts regardless of assessment tools used, with potentially explain differences between previously reported studies, as a focus on decline during the earliest transition from the preclinical to the effect size could be small. Other reasons for differences in effect the prodromal stage of AD before a clinical diagnosis of dementia. We sizes between cohorts are the different tests used within the selected demonstrated a range of effect sizes for different cognitive composites domains; specific tests used within one cohort but not another may and between cohorts; the latter was partially explained by baseline dif- assess subdomains, which are more sensitive to early decline, creating ferences in age and intellectual impairment. Not surprisingly, partici- a discrepancy in the magnitude of the effect reported. Other poten- pants who are older at entry were more likely to decline over the course tial reasons may be that people who start to show decline have been of the study. We then identified cognitive domains and specific neu- dropped out from longitudinal assessments and the threshold at which ropsychological tests that were consistently represented as important people are not offered testing could have differed between sites; or (ie, sensitive to change) across all cohorts. Our final composite score potential differences in thresholds for clinical dementia diagnosis that included measures of memory, language/executive functioning, selec- have determined selection of participants included in this analysis. tive attention, orientation, and praxis. Annualized rates of change on We excluded participants with an AD diagnosis at baseline, but cross- this composite showed cognitive decline in all the cohorts, thus demon- country differences between cohorts in their criteria of diagnosing AD strating its validity in tracking change during the early stages of AD. could have an impact on whether the remaining participants are likely Expert opinion was employed to select tests to use in a final compos- to show change over time, if, for example, those in early prodromal ite for use as outcome measure in clinical trials based on feasibility and stage have been already given an AD diagnosis. known properties in individuals with DS. 4.2 Outcome measure in clinical trials of 4.1 Patterns of decline associated with treatment to delay cognitive decline in DS individuals development in AD in DS There is renewed interest in the need to target the earlier stages of AD This study confirms a pattern of decline that has been emerging in in the context of a series of failed therapeutics in later stage disease studies of cognitive AD-related change in DS. For the first time, data (including of symptomatic therapies in DS30,31 ;). DS represents a rel- from several data sets have been combined, thus avoiding some of the atively large population in which such trials are more feasible (due to 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License ASCHENBRENNER ET AL . 7 of 10 TA B L E 4 Details of the recommended test battery Test and administration Primary outcome of Domain(s) time Administration & scoring interest Memory Cued Recall test, The mCRT consists of a learning phase and a testing Cued delayed recall modified version phase; during the learning phase, 12 items (mCRT)6,23 representing distinct semantic categories are 20 min presented on 3 four-item cards, with each item accompanied by a unique category cue (Buschke, 1984). Learning is repeated up to a maximum of three times if necessary. The testing phase consists of three trials of free and cued immediate recall, generating two measures, a free immediate recall score (FIRS; spontaneous recall of the list of 12 items for each trial) and a total immediate score (TIS; FIRS plus items recalled when the category cue was provided). A 20-min delayed recall trial has also been included, generating two additional scores: free delayed recall score (FDRS) and a total delayed score (FDRS plus items recalled after category cue was provided). Scoring as per Devenny et al., 2002, and Benejam et al., 2015. CANTAB paired associate Computerized measure of visuospatial short-term PAL first trial memory learning test (PAL)24 memory from the CANTAB battery (CANTAB, 2016); score 10 min extended clinical version, with rater rather than automated prompts; Startin et al., 2015). First trial score was used. Participants are required to remember locations of an increasing number of patterns in progressive stages, hidden behind “boxes” on the screen. The main outcome from this test is the first trial memory score: the number of pattern locations correctly remembered on the first trial for each stage attempted. The secondary outcome is the number of stages completed. Language/ executive Verbal Fluency34 Animal fluency test; number of animals named in 60 s. Total raw score. Adjusted functioning tests 1 – 2 min Total correct score and adjusted score (0 – 4) are scores based on the used. CAMCOG-DS scoring can also be used. Selective attention Cancellation task25 Participants are shown a piece of paper with a clutter of Total number correct black and white items, and asked to cross-out each occurrence of a target item, following a practice trial. Total time to complete the task and total number of correct targets crossed-out are recorded. Praxis/ attention/ Purdue Pegboard test26 The Purdue Pegboard Test consists of two rows of 25 Dominant hand raw score dexterity 2 min vertically aligned holes and requires participants to place as many pegs as possible in the holes. Participants do this with their dominant hand, then with their nondominant hand, and finally with both hands, within 30 s per condition. The number of pegs placed in the holes within the time frame is scored. Orientation CAMCOG-DS27 Participants are asked several questions regarding Total score orientation (name, place and time) with possibility of prompts; potential range of scores between 0 and 12. smaller numbers required), and avoiding inclusion of “non-converters a broad range of baseline cognitive abilities. We identified a brief test with mild cognitive impairment” as trial participants as often happens battery, with elements that have been adapted for use in people with in sporadic AD trials.4 A reliable outcome measure that can be used DS, with considerable face validity for use to track AD-related decline. to track treatment effects at early stages of decline would help enable We then demonstrated observable effect sizes for longitudinal cogni- successful clinical trials. However, such outcome measures should also tive change across all study groups for the identified brief battery, in be feasible, easy to administer internationally (independent of lan- individuals at risk for but not yet presenting with diagnosable demen- guage or cultural effects), and allow for a scores from participants with tia due to AD. 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 8 of 10 ASCHENBRENNER ET AL . 4.3 Further development comparable to decline in other forms of AD, including sporadic AD and autosomal-dominant AD. We have identified a composite that is sensi- The H21 AD test battery will be included in our longitudinal studies of tive to cognitive change during the early prodromal stage of AD in DS cognitive decline associated with AD in DS. We will collect data on test- prior to a diagnosis of dementia that can be used as an outcome mea- retest reliability of this combination of subtests across cohorts, as well sure in clinical trials of treatment to prevent or delay decline associated as change in performance over time. This will allow for further analyses with the disease in DS. to consider refinement of scoring; for example, to consider the degree each individual test contributes to an overall score to deliver addi- ACKNOWLEDGEMENTS tional data to inform the design of future trials. Finally, an important The authors acknowledge all the participants in this study for their next step would be to relate changes on the new battery to biomarkers time. We would also like to thank Katherine Van Pelt and Allison associated with the development of AD in DS, including fluid biomark- Caban-Holt from the Kentucky site; Sarah Hamburg, Rosalyn Hither- ers such as neurofilament light, plasma, and neuroimaging measures of say, and Carla Startin from the London site; and Laura Videla from tau and amyloid.17,32,33 It will also be important to establish the clin- Barcelona for their work on initial data collection. We would like to ical meaningfulness and minimal clinically important difference of the thank Rosalyn Hithersay and Sarah Pape from the London site, Isabel propose composite by relating change in cognition to important func- Clare from the Cambridge site, and Anna Esbensen from the Thomas tional or clinical measures such as the rate of progression to demen- Center for Down Syndrome at Cincinnati Children’s Hospital for par- tia or standard measures of activities of daily living using measures ticipating in the consensus discussion. This work was funded by grants such as the CAMDEX-DS informant interview27,34 and the functional from the LuMind IDSC Foundation, the Jérôme Lejeune Foundation, behavior scales,35 and by using anchor-based and distribution-based the Wellcome Trust Strategic Award (grant number: 098330/Z/12/Z), approaches. the Medical Research Council (MRC MR/S011277/1; MR/S005145/1; MR/R024901/1), the Verum Foundation, the Deutsche Forschungs- gemeinschaft (DFG, German Research Foundation) under Germany’s 4.4 Strengths and limitations Excellence Strategy within the framework of the Munich Cluster for Systems Neurology (EXC 2145 SyNergy – ID 390857198), the National Although we had access to a unique and large sample, some within- Institutes of Health/Eunice Kennedy Shriver National Institute of Child group analyses had limited power. Furthermore, study groups used dif- Health and Human Development (NIH/NICHD R01HD064993), the ferent assessment batteries and, in some cases, different scoring cri- National Institute for Health Research (NIHR) Cambridge Biomedi- teria. Thus our statistical analyses are aimed at the identification of cal Research Centre (BRC-1215-20014), and NIH funding was also important cognitive domains rather than the selection of specific neu- received to support the study (R01AG031110 and U01AG051406). ropsychological tests. The discussion by our consensus group aimed The study funders and sponsors had no role in study design, data col- to supplement the statistical modeling in order to identify the spe- lection, data analysis, data interpretation, writing of the report, or the cific tests that would be most feasible to employ in a global clinical decision to submit for publication. The views expressed are those of the trial (eg, easy to score, minimal cultural, or language differences). How- author(s) and not necessarily those of the NIHR or the Department of ever, our main objective was to identify robust patterns of decline Health and Social Care. regardless of assessment tools used, and therefore diversity in terms of batteries were beneficial rather than a limitation. Despite the rela- CONFLICTS OF INTEREST tively small sample sizes of some data sets, the findings were consis- Dr. Juan Fortea is on the advisory board for AC Immune and Lündbeck tent across the cohorts. Length of follow-up also differed between the and is a consultant for Novartis. Dr. Juan Fortea has received compen- data sets. We minimized the impact that this may have had by limiting sation for consultancies to Novartis, Lündbeck, and AC Immune. Dr. the cases by length of follow-up and used annualized rates of change Benjamin L. Handen received research funding from Roche. Dr. Andre on the final composite. Finally, the included cohorts differed slightly in Strydom received funding from AC Immune and is an advisor to ProMIS terms of demographic variables such as age, which explains some of the neurosciences. All other authors declare no conflicts of interest. observed differences in effect sizes, as age is a strong predictor of cog- nitive decline in individuals with DS. This was particularly evident in AUTHOR CONTRIBUTIONS the London cohort, which included the oldest individuals (and is proba- Andrew J. Aschenbrenner: Conducted all analyses and produced ini- bly more representative of the range of intellectual ability amongst the tial drafts of the final manuscript. R. Asaad Baksh: Collated and pro- cohorts) and showed the largest annualized change in performance. cessed the data from each cohort for analyses, organized and par- ticipated in the consensus meeting, and was a major contributor to drafting the manuscript. Bessy Benejam: Participated in discussions 5 CONCLUSIONS in expert panel (group discussions), collected neuropsychological data considered for inclusion, and critically reviewed manuscript. Jessica A. The early markers of cognitive change of AD in DS include prominent Beresford-Webb: Participated in expert group discussions, collected decline in memory, language, attention, and praxis, and appear to be and processed data from Cambridge site, and edited the manuscript. 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License ASCHENBRENNER ET AL . 9 of 10 Antonia Coppus: Participated in discussions, collected data considered 8. de Sola S, de la Torre R, Sánchez-Benavides G, et al. A new cognitive for inclusion, and critically reviewed manuscript. Juan Fortea: Partici- evaluation battery for Down syndrome and its relevance for clinical trials. Front Psychol. 2015;6:708. pated in discussions, collected data considered for inclusion, and criti- 9. Firth NC, Startin CM, Hithersay R, et al. Aging related cognitive cally reviewed the manuscript. Benjamin L. Handen: Contributed data changes associated with Alzheimer’s disease in Down syndrome. Ann to analysis, led to collection of neuropsychological data at Pittsburgh, Clin Transl Neurol. 2018;5:741-751. and assisted with manuscript preparation. Sigan Hartley: Wiscon- 10. Liogier d’Ardhuy X, Edgin JO, Bouis C, et al. Assessment of cognitive scales to examine memory, executive function and language in indi- sin neuropsychologist lead; collected and processed data from Pitts- viduals with Down syndrome: implications of a 6-month observational burgh/Wisconsin site and edited the manuscript. Elizabeth Head: Con- study. Front Behav Neurosci. 2015;9:300. tributed data and assisted with manuscript preparation. Judith Jaeger: 11. Sinai A, Hassiotis A, Rantell K, Strydom A. Assessing specific cognitive Provided high-level guidance on study rationale and objectives, partic- deficits associated with dementia in older adults with Down syndrome: use and validity of the Arizona Cognitive Test Battery (ACTB). PLoS ipated in expert panel discussions, and critically reviewed manuscript. One. 2016;11:e0153917. Johannes Levin: Contributed to the conception and set up of the study, 12. Startin CM, Hamburg S, Hithersay R, et al. The LonDownS adult cog- participated in discussions in expert panel (group discussions), col- nitive assessment to study cognitive abilities and decline in Down syn- lected clinical data considered for inclusion, and critically reviewed drome. Wellcome Open Res. 2016;1. manuscript. Sandra V. Loosli: Participated in discussions in expert panel 13. Startin CM, Hamburg S, Hithersay R, et al. Cognitive markers of preclinical and prodromal Alzheimer’s disease in Down syndrome. (group discussions), collected neuropsychological data considered for Alzheimers Dement. 2019;15:245-257. inclusion, and critically reviewed the manuscript. Anne-Sophie Rebil- 14. Lautarescu BA, Holland AJ, Zaman SH. The early presentation of lat: Participated in discussions, collected data considered for inclusion, dementia in people with Down syndrome: a systematic review of lon- and critically reviewed manuscript. Silvia Sacco: Paris neuropsychol- gitudinal studies. Neuropsychol Rev. 2017;27:31-45. 15. Ritchie CW, Russ TC, Banerjee S, et al. The Edinburgh Consen- ogist lead; collected and processed data from Paris site and critically sus: preparing for the advent of disease-modifying therapies for reviewed the manuscript. Frederick A. Schmitt: Contributed data and Alzheimer’s disease. Alzheimers Res Ther. 2017;9:85. assisted with manuscript preparation. Kate E. Thurlow: Collated and 16. Annus T, Wilson LR, Hong YT, et al. The pattern of amyloid accumula- sorted data from all cohorts and coordinated necessary additions or tion in the brains of adults with Down syndrome. Alzheimers Dement. 2016;12(5):538-545. clarifications from each site of both data and methodology and partic- 17. Cole JH, Annus T, Wilson LR, et al. Brain-predicted age in Down ipated in group discussions. Shahid Zaman: Participated in discussions syndrome is associated with beta amyloid deposition and cognitive and critically reviewed manuscript; PI for Cambridge site. Jason Has- decline. Neurobiol Aging. 2017;56:41-49. senstab: Involved in the study design, oversight of analyses, and critical 18. Fortea J, Carmona-Iragui M, Benejam B, et al. Plasma and CSF biomarkers for the diagnosis of Alzheimer’s disease in adults with review of manuscript. Andre Strydom: Contributed to the conception Down syndrome: a cross-sectional study. Lancet Neurol. 2018;17:860- and set up of the study and analysis, and produced initial drafts of the 869. final manuscript. 19. Hartley SL, Handen BL, Devenny DA, et al. Cognitive functioning in relation to brain amyloid-β in healthy adults with Down syndrome. Brain. 2014;137:2556-2563. ORCID 20. Powell D, Caban-Holt A, Jicha G, et al. Frontal white matter integrity R. Asaad Baksh https://orcid.org/0000-0001-6596-2145 in adults with Down syndrome with and without dementia. Neurobiol Aging. 2014;35:1562-1529. REFERENCES 21. Bates D, Mächler M, Bolker B, Walker S. Fitting linear mixed-effects 1. Wiseman FK, Al-Janabi T, Hardy J, et al. A genetic cause of Alzheimer models using lme4. J Stat Software. 2015;67:1-48. disease: mechanistic insights from Down syndrome. Nat Rev Neurosci. 22. Donohue MC, longpower: Power and sample size calculations for lon- 2015;16:564-574. gitudinal data. R package version 1.0-21. 2020. 2. McCarron M, McCallion P, Reilly E, Dunne P, Carroll R, Mulryan N. 23. Devenny D, Zimmerli E, Kittler P, Krinsky-McHale S. Cued recall in A prospective 20-year longitudinal follow-up of dementia in persons early-stage dementia in adults with Down’s syndrome. J Intellect Dis- with Down syndrome. J Intellect Disabil Res. 2017;61:843-852. abil Res. 2002;46:472-483. 3. Castro P, Zaman S, Holland A. Alzheimer’s disease in people with 24. CANTAB®. Cambridge Neuropsychological Test Automated Bat- Down’s syndrome: the prospects for and the challenges of developing tery, (CANTAB®). Cambridge, UK: Cambridge Cognition Ltd; preventative treatments. J Neurol. 2017;264:804-813. 2016. 4. Strydom A, Coppus A, Blesa R, et al. Alzheimer’s disease in Down 25. Krinsky-McHale SJ, Devenny DA, Kittler P, Silverman W. Selective syndrome: an overlooked population for prevention trials. Alzheimers attention deficits associated with mild cognitive impairment and early Dement (N Y). 2018;4:703-713. stage Alzheimer’s disease in adults with Down syndrome. Am J Ment 5. RE Nieuwenhuis-Mark. Diagnosing Alzheimer’s dementia in Down Retard. 2008;113:369-386. syndrome: problems and possible solutions. Res Dev Disabil. 26. Tiffin J, Asher EJ. The Purdue Pegboard: norms and studies of reliabil- 2009;30:827-838. ity and validity. J Appl Psychol. 1948;32:234. 6. Benejam B, Fortea J, Molina-López R, Videla S. Patterns of perfor- 27. Hon J, Huppert FA, Holland A, Watson P. Neuropsychological assess- mance on the modified cued recall test in Spanish adults with down ment of older adults with Down’s syndrome: an epidemiological study syndrome with and without dementia. Am J Intellect Dev Disabil. using the Cambridge Cognitive Examination (CAMCOG). Br J Clin Psy- 2015;120:481-489. chol. 1999;38:155-165. 7. Benejam B, Videla L, Vilaplana E, et al. Diagnosis of prodromal and 28. Krinsky-McHale SJ, Devenny DA, Silverman W. Changes in explicit Alzheimer’s disease dementia in adults with Down syndrome using memory associated with early dementia in adults with Down’s syn- neuropsychological tests. Alzheimers Dement (Amst). 2020;12:e12047. drome. J Intellect Disabil Res. 2002;46:198-208. 23528729, 2021, 1, Downloaded from https://alz-journals.onlinelibrary.wiley.com/doi/10.1002/dad2.12184 by Readcube (Labtiva Inc.), Wiley Online Library on [11/04/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 10 of 10 ASCHENBRENNER ET AL . 29. Ball SL, Holland AJ, Treppner P, Watson PC, Huppert FA. Execu- 35. Hatton C, Emerson E, Robertson J, Gregory N, Kessissoglou S, Perry tive dysfunction and its association with personality and behaviour J, et al. The adaptive behavior scaleresidential and community (part I): changes in the development of Alzheimer’s disease in adults with towards the development of a short form. Research in Developmental Down syndrome and mild to moderate learning disabilities. Br J Clin Disabilities. 2001;22:273-88. Psychol. 2008;47:1-29. 36. Korkman M, Kirk U, Kemp S, NEPSY II: Clinical and interpretive man- 30. Hanney M, Prasher V, Williams N, et al. Memantine for dementia in ual: Harcourt Assessment, PsychCorp; 2007. adults older than 40 years with Down’s syndrome (MEADOWS): a ran- domised, double-blind, placebo-controlled trial. Lancet North Am Ed. 2012;379:528-536. 31. Livingstone N, Hanratty J, McShane R, Macdonald G. Pharmacologi- SUPPORTING INFORMATION cal interventions for cognitive decline in people with Down syndrome. Additional supporting information may be found online in the Support- Cochrane Database Syst Rev. 2015:1-55. 32. Fortea J, Vilaplana E, Carmona-Iragui M, et al. Clinical and biomarker ing Information section at the end of the article. changes of Alzheimer’s disease in adults with Down syndrome: a cross- sectional study. Lancet North Am Ed. 2020;395:1988-1997. 33. Mengel D, Liu W, Glynn RJ, et al. Dynamics of plasma biomarkers in Down syndrome: the relative levels of Aβ42 decrease with age, How to cite this article: Aschenbrenner AJ, Baksh RA, whereas NT1 tau and NfL increase. Alzheimers Res Ther. 2020;12:1-11. Benejam B, et al. Markers of early changes in cognition across 34. Roth M, Tym E, Mountjoy CQ, Huppert FA, Hendrie H, Verma S, et al. cohorts of adults with Down syndrome at risk of Alzheimer’s CAMDEX: a standardised instrument for the diagnosis of mental dis- order in the elderly with special reference to the early detection of disease. Alzheimer’s Dement. 2021;13:e12184. dementia. The British Journal of Psychiatry. 1986;149:698-709. https://doi.org/10.1002/dad2.12184