ABSTRACT
Background
The Addenbrooke's Cognitive Examination‐III (ACE‐III) is a common cognitive screening measure. We provide performance deciles, descriptive performance bands and indices of reliable change between repeat assessments using the ACE‐III in a large cohort of dementia participants and healthy controls.
Methods
Baseline data from 727 participants diagnosed with a dementia syndrome and 157 healthy controls were used to calculate performance deciles and descriptive bands. A subset of 393 participants diagnosed with a dementia syndrome completed two annual assessments. These data were used to calculate 95% confidence intervals of characteristic yearly change in each dementia syndrome. Data from a subset of 74 healthy participants who completed two assessments were used to calculate reliable change indices.
Results
Baseline performance was grouped into five descriptive performance bands across the entire spectrum of scores: Very Mild, Mild, Moderate, Severe and Very Severe. Deciles and performance bands are provided for all dementia participants combined, for each syndrome where n > 50, and for healthy controls, stratified by group, sex and years of education. A decline of −7 to −9 points yearly was characteristic for the dementia population, and this varied between syndromes. Reliable change calculations indicate that a 5‐point decline from within the ‘normal’ range (88–100) is the minimum clinically important decline on the ACE‐III.
Conclusions
This study presents a suite of reference data on the ACE‐III in a range of dementia syndromes, providing important insight and support to clinicians who work with people living with cognitive impairments.
Keywords: Alzheimer's disease, cognitive screening, frontotemporal dementia, neurodegeneration, primary progressive aphasia
Performance on the Addenbrooke's Cognitive Examination‐III is stratified into 5 performance bands, and performance deciles are provided for a sample of 727 dementia participants and 157 healthy controls. An improvement of +4 points or a decline of −5 points is described as clinically meaningful, and metrics of characteristic annual change are provided for multiple dementia syndromes.
1. Introduction
Accurate monitoring of cognitive changes over time in dementia requires lengthy comprehensive cognitive assessments. These examination methods may be unfeasible due to resource or time limitations within clinical settings; and some individuals may lack the capability of completing such rigorous testing. In these circumstances, brief cognitive screening measures offer a useful method of monitoring longitudinal cognitive performance [1, 2]. The Addenbrooke's Cognitive Examination‐III (ACE‐III) [3] is a brief cognitive screening measure widely used for diagnosing and monitoring dementia [1, 4, 5, 6, 7] and other neurological conditions, such as motor neuron disease [8, 9]. The ACE‐III provides an overall measure of cognitive ability, as well as performance across five cognitive subdomains: attention and orientation, memory, verbal fluency, language and visuospatial ability. The overall and subdomain scores demonstrate good convergent validity with other cognitive screening measures such as the Mini Mental State Examination (MMSE) [10], and with standard neuropsychological tests assessing the same domains [1, 3]. Importantly, the ACE‐III is sensitive to cognitive changes across a wide spectrum of severity levels, from mild cognitive impairment to severe dementia [1, 7, 11], making it useful for tracking cognitive decline in these populations. The ACE‐III has been shown to outperform other brief screening measures, such as the MMSE and the Montreal Cognitive Assessment [12], making it particularly useful for providing an overview of a person's overall cognitive functioning when comprehensive neuropsychological testing is not feasible [1, 2].
Standardised stratification of cognitive performance into delineated bands can help clinicians and researchers detect when a cognitive change crosses a meaningful threshold, either due to disease progression or as an effect of treatment. Such metrics can also assist in the challenging process of accurately characterising cognitive components of disease staging within clinical research studies. It can also assist in enabling thorough monitoring of prognosis and selection of suitable treatment options within public health settings. So and colleagues [7] proposed ACE‐III cut‐off scores, with scores of ≥ 88/100 points reflecting ‘normal’ performance, and < 84/100 indicating possible dementia [7], as well as cut‐off scores according to dementia severity based on the Clinical Dementia Rating (CDR) [13] scale. Although informative, these CDR‐based cut‐off scores offer little information to contextualise an individual's cognitive performance relative to that of other people with a similar diagnosis and are not informative across the entire spectrum of possible scores. Indeed, these severity cut‐off scores were based on the relationship between the ACE‐III and CDR scores, but this relationship was variable and was noted to violate some statistical assumptions. For example, the CDR score explained 26% of the ACE‐III score variance in patients with Alzheimer's disease (AD) and 49% in those with logopenic progressive aphasia (LPA), but only 17% in patients with behavioural‐variant frontotemporal dementia (bvFTD) [7]. This is likely to be due to the greater weight the CDR places on the presence of memory impairment when determining severity, which varies in severity in other dementia syndromes [14]. Whilst the ACE‐III provides a holistic assessment of an individual's cognitive performance, there is limited data describing characteristic ACE‐III performance across a range of diseases and demographics. Likewise, no data exist documenting the magnitude of change between repeat assessments of the ACE‐III that is considered clinically meaningful.
This study aims to address these issues by providing a set of detailed reference data for the ACE‐III, including performance deciles and severity bands, stratified by diagnosis, sex and education in large well‐characterised cohorts of dementia and healthy participants. These bands allow clinicians to contextualise their patient's performance relative to others with a dementia diagnosis. We also provide metrics of characteristic yearly change in ACE‐III score across various dementia syndromes. Finally, this study offers indices of reliable change between repeat assessments of the ACE‐III to determine the minimal clinically important difference in scores required beyond measurement artefacts. Overall, this study aims to provide a series of reliable reference data to support clinical decision‐making when using the ACE‐III.
2. Methods
2.1. Participants
Seven‐hundred‐and‐twenty‐seven individuals diagnosed with dementia between November 2007 and February 2020 from FRONTIER, the younger‐onset dementia research clinic based in Sydney, Australia, were included this study. Assessments were typically completed within a 1–2 month period, and patients were reviewed at approximately 12‐month intervals, with an average of 2–3 visits. Diagnosis and clinical staging was made after clinical examination, comprehensive neuropsychological assessment, structural brain magnetic resonance imaging and consensus meeting according to the relevant criteria: bvFTD [15], semantic dementia (SD; also known as semantic variant primary progressive aphasia), progressive nonfluent aphasia (PNFA; also known as nonfluent variant primary progressive aphasia) and logopenic progressive aphasia (LPA; also known as logopenic variant primary progressive aphasia) [16], typical ‘amnestic’ AD [17], corticobasal syndrome (CBS) [18], motor neuron disease [19], frontotemporal dementia‐motor neuron disease [20], posterior cortical atrophy [21], progressive supranuclear palsy [22], vascular dementia [23, 24] and multiple systems atrophy [25]. All data were collected by a qualified clinical neuropsychologist trained in the administration of the ACE‐R [26] and ACE‐III. These clinicians were not blinded to participant diagnosis, as they provided insight and expertise to the consensus diagnosis itself. Participants were excluded if they were not fluent in English, did not have a reliable informant, were unable to complete the ACE‐R or ACE‐III at first visit, or had > 10 years of disease duration at baseline visit.
Additional data from 157 healthy individuals were also included as a control comparison. These participants were recruited from a volunteer panel and underwent identical cognitive and brain imaging evaluation to the dementia participants. Healthy controls were excluded if they had a family history of dementia or other neurological disease, subjective cognitive complaints, current alcohol or substance misuse, history of brain injury with loss of consciousness for more than 5 min, or an ACE‐R/ACE‐III score of < 88 at first assessment.
All participants completed the ACE‐R or ACE‐III at each timepoint as part of a standard comprehensive neuropsychological battery. Assessments conducted between 2007 and 2012 (inclusive) utilised the ACE‐R, and assessments conducted between 2013 and 2020 (inclusive) utilised the ACE‐III. ACE‐R scores were converted to ACE‐III scores using published formulae [7]. So and colleagues [7] found, on average, only a one‐point difference between original ACE‐R and the converted ACE‐III total scores, indicating a negligible difference between original ACE‐R and converted ACE‐III scores.
Three‐hundred‐and‐ninety‐three participants with a dementia diagnosis and 74 healthy controls completed two annual assessments (T1: first visit; T2: second visit). These participants were included in longitudinal analyses examining changes in ACE score over time. Version A of the Australian ACE was used on all T1 assessments, whereas version B was used on all T2 assessments, in accordance with best practice [3].
2.2. Methods and Statistical Analyses
All analyses were conducted using R version 4.3.0 [27] and R Studio version 2023.03.1 + 446 [28]. Data were analysed in two ways: (1) by combining all participants diagnosed with dementia into a single group (‘dementia combined’), and (2) by conducting separate analyses in each dementia syndrome where n ≥ 50. The dementia combined group is intended to broadly represent the possible range of performance for an individual with a suspected dementia diagnosis when additional information is not available. In instances where a specific diagnosis is suspected, the separate syndrome data enable clinicians to interpret metrics that are, in theory, more representative of that diagnosis. Table 1 presents the demographic characteristics for the dementia combined group, the six dementia syndromes with at least 50 participants, and the healthy control group at T1 assessment. One‐way analysis‐of‐variance (ANOVA) revealed a significant group difference in age at baseline, F(6, 767) = 6.0, p < 0.001. Post hoc comparisons revealed that bvFTD participants were younger, on average, than the healthy control, LPA and PNFA groups, but no other differences in age were found. A significant group difference in education level was also identified, F(6, 767) = 4.0, p ≤ 0.001. Post hoc comparisons revealed that healthy control participants had higher education, on average, than the AD, bvFTD, LPA and CBS groups, but no other differences in education were found. Demographic data for all the dementia syndromes, including those with fewer than 50 patients, are shown in the Supporting Information (Table S1).
TABLE 1.
Demographic characteristics of study samples with at least one timepoint.
| Diagnosis | n | Sex (m: f) | Age | Education | ACE‐III total (/100) |
|---|---|---|---|---|---|
| Dementia (combined) | 727 | 391: 336 | 65 (8.5) | 12 (3.1) | 67 (20) |
| AD | 170 | 85: 85 | 65 (8.9) | 12 (3.1) | 62 (20) |
| bvFTD | 168 | 110: 58 | 62 (8.6) | 12 (3) | 75 (17) |
| SD | 72 | 35: 37 | 65 (7.2) | 12 (3.1) | 58 (18) |
| LPA | 68 | 30: 38 | 67 (7.2) | 12 (3) | 58 (19) |
| PNFA | 64 | 31: 33 | 68 (10) | 12 (3.2) | 71 (18) |
| CBS | 51 | 19: 32 | 66 (7.3) | 11 (3) | 69 (19) |
| Controls | 157 | 71: 86 | 66 (7.3) | 13 (3.1) | 94 (3.2) |
Note: Values are mean (standard deviation). Mean values are rounded to the nearest integer.
Abbreviations: ACE‐III, Addenbrooke's Cognitive Examination – III; AD, Alzheimer's disease; bvFTD, behavioural variant frontotemporal dementia; CBS, corticobasal syndrome; f, female; LPA, logopenic progressive aphasia; m, male; PNFA, progressive nonfluent aphasia; SD, semantic dementia.
2.3. Describing ACE‐III Performance Profiles
Percentiles of ACE‐III performance at T1 were calculated for (1) the combined dementia group, (2) the six dementia syndromes with at least 50 participants and (3) the healthy controls. Percentiles were calculated separately for male and female participants, and in two education bands: < 10 years of education, and ≥ 10 years or more. Due to group size variation after stratification, results are presented for each diagnostic group with sex and education levels combined. The complete data stratified by sex and education are available in the Supporting Information (Table S2). In line with previous publications, ‘normal’ performance on the ACE‐III was defined as scores ≥ 88/100 [7]. In addition to the ‘normal’ range, five descriptive performance bands based on ACE‐III total score were defined: Very Mild: 75th percentile to a score of 88/100, Mild: 60th to 74th percentile, Moderate: 40th to 59th percentile, Severe: 20th to 39th percentile and Very Severe: < 20th percentile. These performance bands were selected to represent the distribution of the total ACE‐III score in easily interpretable categories that are clinically meaningful. Performance bands stratified by sex and education are available in the Supporting Information (Table S3).
2.4. Quantifying Characteristic Yearly Change in Dementia
To establish metrics of characteristic yearly change of ACE‐III total score for the dementia groups, data from the participants with a dementia diagnosis who completed two annual assessments were selected. T1 score was subtracted from T2 score to obtain a change score. To best represent the variability in the magnitude of change between T1 and T2, all datapoints were included in the dataset, including those from outliers. Due to these outliers, the negatively skewed distribution of change scores, and attrition from T1 to T2, a bootstrapped resampling method was applied to obtain 95% confidence intervals of characteristic yearly change per diagnostic group. Details of this attrition are provided per group in the Supporting Information (Table S1).
2.5. Understanding Clinically Meaningful Change
To understand clinically meaningful change on the ACE‐III, we calculated indices of reliable change (RCI). A reliable change between ACE‐III assessments is a change that extends beyond measurement artefacts; that is, measurement error or practice effects. A change of this magnitude is therefore likely due to some external factor (e.g., a disease process or effective treatment). Data from the healthy controls who completed two assessments were utilized to calculate these indices. The ACE‐III and other cognitive screeners (e.g., MMSE) are designed such that healthy individuals score at or near ceiling, and their performance should not change over time. Previous studies using similar cognitive tests in older adults show the RCI calculation method used here is appropriate for detecting reliable changes in performance beyond measurement error and practice effects [29, 30, 31]. The magnitude of change in ACE‐III total score which represents a reliable improvement or decline was calculated for each possible score within the ‘normal’ range (88–100). Further details regarding the calculation of these indices, as well as examination of the presence of regression to the mean in this dataset, are provided in the Supporting Information (Tables S4–S6 and Figure S8).
3. Results
3.1. Describing ACE‐III Performance Profiles
ACE‐III performance deciles at the T1 assessment are presented in Table 2 for the combined dementia group (n = 727), for each dementia syndrome where n ≥ 50 and for healthy control participants (n = 157). The overall ACE‐III total score ranges for each performance band are presented in Table 3 for each dementia group. Descriptive performance bands were not derived for healthy participants due to the very small range of scores in this group. Scores for each group stratified by sex and education are provided in the Supporting Information as deciles (Table S2) and performance bands (Table S3). Descriptive performance bands were then applied to a density plot of the ACE‐III T1 performance distribution in the combined dementia group to visualise the bands (Figure 1). Similar plots for each dementia syndrome where n ≥ 50 are provided in the Supporting Information (Figures S1–S6).
TABLE 2.
ACE‐III T1 performance profiles in dementia groups and healthy controls.
| ACE‐III total score | ||||||||
|---|---|---|---|---|---|---|---|---|
| Percentile | Dementia (combined) | AD | bvFTD | SD | LPA | PNFA | CBS | Healthy controls |
| n = 727 | n = 170 | n = 168 | n = 72 | n = 68 | n = 64 | n = 51 | n = 157 | |
| 100th | 99 | 93 | 99 | 96 | 91 | 96 | 95 | 100 |
| 90th | 89 | 84 | 92 | 79 | 81 | 88 | 88 | 98 |
| 80th | 83 | 80 | 87 | 74 | 74 | 85 | 83 | 97 |
| 70th | 80 | 75 | 84 | 68 | 71 | 83 | 80 | 96 |
| 60th | 76 | 71 | 82 | 64 | 67 | 81 | 77 | 95 |
| 50th | 72 | 67 | 79 | 60 | 62 | 77 | 74 | 94 |
| 40th | 67 | 60 | 77 | 52 | 57 | 70 | 70 | 94 |
| 30th | 60 | 53 | 72 | 48 | 47 | 65 | 66 | 93 |
| 20th | 50 | 43 | 64 | 45 | 39 | 55 | 59 | 92 |
| 10th | 39 | 34 | 57 | 41 | 34 | 44 | 37 | 90 |
Note: All values are ACE‐III total score/100. Scores are rounded to the nearest whole number for ease of interpretation.
Abbreviations: ACE‐III, Addenbrooke's Cognitive Examination‐III; AD, Alzheimer's disease; bvFTD, behavioural variant frontotemporal dementia; CBS, corticobasal syndrome; LPA, logopenic progressive aphasia; PNFA, progressive nonfluent aphasia; SD, semantic dementia.
TABLE 3.
ACE‐III T1 score performance bands in dementia groups.
| Descriptor | ACE‐III total score | ||||||
|---|---|---|---|---|---|---|---|
| Dementia (combined) | AD | bvFTD | SD | LPA | PNFA | CBS | |
| n = 727 | n = 170 | n = 168 | n = 72 | n = 68 | n = 64 | n = 51 | |
| Normal range | ≥ 88 | ≥ 88 | ≥ 88 | ≥ 88 | ≥ 88 | ≥ 88 | ≥ 88 |
| Very Mild | 83–87 | 79–87 | 86–87 | 70–87 | 73–87 | 84–87 | 82–87 |
| Mild | 77–82 | 71–78 | 82–85 | 63–69 | 67–72 | 81–83 | 77–81 |
| Moderate | 67–76 | 60–70 | 77–81 | 52–62 | 57–66 | 70–80 | 70–76 |
| Severe | 51–66 | 43–59 | 64–76 | 45–51 | 39–56 | 55–69 | 59–69 |
| Very Severe | ≤ 50 | ≤ 42 | ≤ 63 | ≤ 44 | ≤ 38 | ≤ 54 | ≤ 58 |
Note: Very Mild 75th+, Mild = 74th—60th %ile, Moderate = 59th—40th %ile, Severe = 39th—20th %ile, Very Severe ≤ 19th %ile. Where a category split was made in between a whole number (e.g., 75th percentile = 77.78, 74th percentile = 77.34), the upper score was rounded up to a whole number and the lower percentile was rounded down to a whole number to ensure no ambiguity of score ranges between bands.
Abbreviations: ACE‐III, Addenbrooke's Cognitive Examination‐III; AD, Alzheimer's disease; bvFTD, behavioural variant frontotemporal dementia; CBS, corticobasal syndrome; LPA, logopenic progressive aphasia; PNFA, progressive nonfluent aphasia; SD, semantic dementia.
FIGURE 1.
ACE‐III first visit (T1) performance distribution for all dementia syndromes (combined). Derived T1 performance bands are overlaid. The X‐axis represents ACE‐III total score, and the Y‐axis represents the proportion of the sample that obtained that score. Demographic data for these participants is described in Table 1 (‘dementia [combined]’).
3.2. Quantifying Characteristic Yearly Change in Dementia
Characteristic annual ACE‐III performance change was examined using data from the 393 dementia participants who completed two annual assessments. Group demographics, average time between T1 and T2, and 95% confidence intervals of mean annual change are presented in Table 4. These data are presented using bar plots in the Supporting Information (Figure S7). Unsurprisingly, total ACE‐III score for all groups decreased from T1 to T2, given the progressive nature of the diseases. The large standard deviations and bootstrapped 95% confidence ranges of mean annual change per group between T1 and T2 reflect the high degree of variability in ACE‐III scores, likely due to the varied severity of disease within each group.
TABLE 4.
Demographic characteristics and average time between T1–T2 in the dementia groups with two assessments.
| Diagnosis | n | Sex (m: f) | Age | Education | ACE‐III (T1) | ACE‐III (T2) | Time T1 – T2 | T1 – T2 ACE‐III Delta |
|---|---|---|---|---|---|---|---|---|
| Dementia (combined) | 393 | 228: 165 | 64 (8) | 12 (3.2) | 73 (15) | 65 (19) | 1.06 (0.41) | −7 < > −9 |
| AD | 68 | 40: 28 | 64 (8) | 13 (3.3) | 72 (13) | 65 (16) | 1.06 (0.24) | −5 < > −9 |
| bvFTD | 112 | 78: 34 | 62 (8.5) | 12 (2.9) | 78 (12) | 71 (18) | 1.11 (0.61) | −5 < > −9 |
| SD | 57 | 30: 27 | 64 (6.9) | 12 (3.2) | 63 (16) | 54 (18) | 1.03 (0.25) | −6 < > −10 |
| LPA | 30 | 13: 17 | 66 (7.8) | 13 (3.4) | 67 (12) | 54 (16) | 1.12 (0.33) | −9 < > −17 |
| PNFA | 37 | 17: 20 | 65 (9.2) | 13 (3.1) | 77 (14) | 68 (22) | 1.05 (0.24) | −6 < > −17 |
| CBS | 29 | 13: 16 | 66 (7.3) | 12 (3.1) | 75 (16) | 67 (18) | 1.09 (0.35) | −6 < > −11 |
Note: Values are mean (standard deviation). T1 – T2 ACE‐III Delta presented as bootstrapped 95% confidence intervals. T1—T2 ACE‐III Delta and 95% confidence intervals are presented graphically in the Supporting Information (Figure S7).
Abbreviations: ACE‐III, Addenbrooke's Cognitive Examination‐III; AD, Alzheimer's disease; bvFTD, behavioural variant frontotemporal dementia; CBS, corticobasal syndrome; f, female; LPA, logopenic progressive aphasia; m, male; PNFA, progressive nonfluent aphasia; SD, semantic dementia. T1, first assessment; T2, 12‐month review assessment; Time T1—T2, Time elapsed between T1 and T2 (in years).
3.3. Understanding Clinically Meaningful Change
Reliable change indices were calculated using data for the 74 healthy control participants who completed two assessments (2.5‐year interval on average). Healthy control participants were included in these analyses if they scored above 88 at T1, regardless of their T2 score. Little to no practice effect was observed over 2.5 years in these data. Although a trend towards some degree of regression to the mean was observed (see Supporting Information), this effect was smaller than the threshold for measurement error. As such, for each score within the ‘normal’ range (88–100), a change of −5 points is unlikely to be due to measurement error, and an improvement of +4 points is unlikely to be due to measurement error or practice effects.
4. Discussion
This study is the first to quantify and delineate characteristic performance on the ACE‐III across the entire spectrum of scores for a range of dementia syndromes. Previous research has established cut‐off scores for ‘normal’ performance and possible dementia [7]. Beyond these cut‐offs, however, there is little guidance with regards to the clinical meaning behind a given score. By referring to the deciles or to the five descriptive bands of ACE‐III performance decline defined here (Very Mild to Very Severe), clinicians are now able to ground their decision‐making in sound empirical data. Moreover, these metrics are provided for different levels of stratification—diagnosis, sex and education—to allow even more precise contextualisation of an individual score relative to other individuals with similar characteristics. To further support a clinician's decision‐making, this study is the first to provide 95% confidence intervals of characteristic yearly change across a range of dementia syndromes. These values provide an empirically derived reference point for clinicians to gauge whether a cognitive change warrants extra concern or investigation given a known or suspected diagnosis.
The performance distribution of the dementia (combined) group is broadly in line with other large studies utilising a mixed dementia cohort [32]. The data presented here describe a similar average score (3‐point difference), but a greater range of scores (higher ceiling, lower floor). This difference is likely due to the inclusion of more types of dementia and a greater range of severity levels in the current study. We argue this is a positive feature—the deciles and descriptive performance bands defined here reflect the entire range of scores and do not neglect the most severe or the least severe cases. This also allows the data to be used to inform clinical practice during both diagnosis and post‐diagnosis follow‐up. Likewise, the 95% confidence intervals of characteristic change described here are within 2–3 points of those described in previous research, with the current study finding slightly greater characteristic yearly change compared to previous research [32]. Again, this may be due to the inclusion of more dementia syndromes in the current study, and as such all metrics are provided for the combined dementia cohort as well as each diagnostic group separately so that clinicians can utilise the normative data that are most appropriate to their clinical situation.
The reliable change values described here represent the minimum clinically important difference at which a decline or improvement should be considered meaningful. Although the lower limit of ‘normal’ performance is already defined (88/100), the reliable change values allow clinicians to gauge whether a decline within the normal bounds is a cause for concern, improving the tools available for early diagnosis of cognitive decline. Results indicate that a decline of −5 points or more from within the ‘normal’ range is not due to measurement error nor regression to the mean, and an improvement of +4 points or more is not due to measurement error or practice effects. Indeed, the characteristic decline of each dementia syndrome in the current study aligns with these markers: all syndromes exhibited a decline of 5 or more points yearly, and the 95% confidence bounds did not cross the −5‐point marker. This is corroborated by prior research, where studies examining longitudinal cognitive decline using the ACE‐III also report a yearly change of −5 or more points in all AD and frontotemporal dementia syndromes [33], the primary progressive aphasias [34, 35] and in a mixed dementia cohort [32]. Likewise, the −5 point threshold for reliable decline on the ACE‐III has been recently been noted elsewhere, though this study did not examine reliable improvement [36]. The current research codifies these values through the calculation of reliable change indices using a large sample of normative data.
The reference data presented here are relevant in all clinical settings. For example, where the main question is centred around differentiation from ‘normal’ performance, the provided metrics will support clinicians in determining severity of cognitive impairment (via the overall or sex/education stratified performance bands and deciles), as well as determining whether a given change over time may indicate a possible disease process (via the reliable change indices). Although monitoring the cognitive performance of healthy individuals was not the focus of the current study, the healthy deciles and clinically meaningful change indices provided here may also be useful in the early detection of cognitive changes before the onset of other symptoms in otherwise healthy individuals. Future research may wish to examine the utility of the ACE‐III in early detection and prevention of cognitive decline in healthy individuals, particularly in individuals that are not as highly screened as the cohort presented here. In a more specialist setting where the focus will be on establishing a differential diagnosis, clinicians can now utilise diagnosis‐stratified performance bands or deciles as well as diagnosis‐stratified 95% confidence intervals of characteristic yearly change to help inform their decisions regarding diagnosis, prognosis and management.
In providing these tools, we wish to also emphasise that the ACE‐III is best used within the context of a comprehensive clinical assessment and should not be used in isolation for the diagnosis of dementia. The reference data provided here are intended to support the clinician's expert judgement and formulation, not override it. Although the data presented here reflect a variety of dementia syndromes, these data may not reflect other clinical populations, such as Parkinson's disease, dementia with Lewy bodies, vascular dementia, or brain injury. As such, clinicians need to be mindful that discrepancies may occur when using these data in different clinical populations. Likewise, we advise caution when applying these metrics to more culturally diverse populations or in individuals outside of the relatively young age range presented here. Nonetheless, as the overall performance profiles described here cover the entire spectrum of possible scores on the ACE‐III, we argue that these profiles will be broadly applicable for a variety of clinical populations.
Some caveats must be noted when interpreting the data presented here. Some group differences in age and education were noted in this cohort (e.g., healthy controls had higher education, on average, than the AD and bvFTD groups). Although these variables may interact with cognitive performance, performance deciles and bands stratified for education and separated by dementia group will minimize the potential effects of these variables. Nonetheless, clinicians should be mindful of this detail when interpreting the data presented here. Likewise, we acknowledge that the variability in sample size of each dementia group may introduce some bias, particularly within the dementia (combined) data. Nonetheless, relative to other similar studies utilizing well characterized samples of rare dementia syndromes, the ratio presented here represents a more balanced distribution of sample sizes [32]. By providing performance bands and deciles for each group separately, as well as providing bootstrapped confidence intervals for characteristic annual change, we aimed to minimize the potential effects of this disproportionate sample size, as well as the potential bias introduced via selective attrition in the change data. Future research should expand these reference data by examining sub‐domain performance deciles, performance bands, and characteristic change indices in dementia syndromes. As the ACE‐III is sensitive to cognitive changes across a broad range of severities, expanding these reference data to include profiles and change indices for other diagnostic groups not represented here, such as in Parkinson's disease, dementia with Lewy bodies, vascular dementia, or brain injury, would also be valuable. In summary, the reference data presented here will assist clinicians in making informed decisions based on cognitive performance and change using the ACE‐III, enabling better management and clinical care for people with cognitive impairments.
Author Contributions
James Carrick: conceptualization, methodology, writing – original draft, writing – review and editing, formal analysis, data curation, visualization. Sau Chi Cheung: writing – review and editing, investigation, conceptualization, data curation. David Foxe: investigation, conceptualization, writing – review and editing. Olivier Piguet: conceptualization, methodology, funding acquisition, writing – review and editing, writing – original draft, supervision, resources, project administration.
Ethics Statement
This study was approved by the Human Research Ethics Committee of the South Eastern Sydney Local Area Health District (HREC 10/126).
Consent
All participants or their person responsible provided written informed consent in accordance with the Declaration of Helsinki.
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Data S1
Acknowledgements
We thank all participants and their families for their time and contribution to this study.
Carrick J., Cheung S. C., Foxe D., and Piguet O., “Interpreting Addenbrooke's Cognitive Examination‐III Scores in Dementia: Performance Distributions and Clinically Meaningful Change,” European Journal of Neurology 32, no. 7 (2025): e70257, 10.1111/ene.70257.
Funding: This work was supported in part by funding to ForeFront, a collaborative research group dedicated to the study of frontotemporal dementia and motor neuron disease, from the National Health and Medical Research Council (NHMRC) (GNT1037746) and the Australian Research Council (ARC) Centre of Excellence in Cognition and its Disorders Memory Program (ce11000102). JC is supported by an Australian Government Research Training (RTP) Scholarship. DF is supported by The Edwards Fund for Dementia Research. OP is supported in part by an NHMRC Leadership Fellowship (GNT2008020).
James Carrick and Sau Chi Cheung contributed equally to the work.
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. No part of the study procedures or analyses was preregistered prior to the research being undertaken. The Addenbrooke's Cognitive Examination‐Third edition (ACE‐III) is freely available at https://frontierftd.org (accessed on 03 June 2025).
References
- 1. Matias‐Guiu J. A., Cortés‐Martínez A., Valles‐Salgado M., et al., “Addenbrooke's Cognitive Examination III: Diagnostic Utility for Mild Cognitive Impairment and Dementia and Correlation With Standardized Neuropsychological Tests,” International Psychogeriatrics 29, no. 1 (2017): 105–113, 10.1017/S1041610216001496. [DOI] [PubMed] [Google Scholar]
- 2. Giebel C. M. and Challis D., “Sensitivity of the Mini‐Mental State Examination, Montreal Cognitive Assessment and the Addenbrooke's Cognitive Examination III to Everyday Activity Impairments in Dementia: An Exploratory Study,” International Journal of Geriatric Psychiatry 32, no. 10 (2017): 1085–1093, 10.1002/gps.4570. [DOI] [PubMed] [Google Scholar]
- 3. Hsieh S., Schubert S., Hoon C., Mioshi E., and Hodges J. R., “Validation of the Addenbrooke's Cognitive Examination III in Frontotemporal Dementia and Alzheimer's Disease,” Dementia and Geriatric Cognitive Disorders 36, no. 3–4 (2013): 242–250, 10.1159/000351671. [DOI] [PubMed] [Google Scholar]
- 4. Elamin M., Holloway G., Bak T. H., and Pal S., “The Utility of the Addenbrooke's Cognitive Examination Version Three in Early‐Onset Dementia,” Dementia and Geriatric Cognitive Disorders 41, no. 1–2 (2016): 9–15. [DOI] [PubMed] [Google Scholar]
- 5. Foxe D., Hu A., Cheung S. C., et al., “Utility of the Addenbrooke's Cognitive Examination III Online Calculator to Differentiate the Primary Progressive Aphasia Variants,” Brain Communications 4, no. 4 (2022): fcac161, 10.1093/braincomms/fcac161. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Leyton C. E., Hornberger M., Mioshi E., and Hodges J. R., “Application of Addenbrooke's Cognitive Examination to Diagnosis and Monitoring of Progressive Primary Aphasia,” Dementia and Geriatric Cognitive Disorders 29, no. 6 (2010): 504–509, 10.1159/000313980. [DOI] [PubMed] [Google Scholar]
- 7. So M., Foxe D., Kumfor F., et al., “Addenbrooke's Cognitive Examination III: Psychometric Characteristics and Relations to Functional Ability in Dementia,” Journal of the International Neuropsychological Society 24, no. 8 (2018): 854–863, 10.1017/S1355617718000541. [DOI] [PubMed] [Google Scholar]
- 8. Long Z., Irish M., Hodges J. R., Piguet O., and Burrell J. R., “Distinct Disease Trajectories in Frontotemporal Dementia–Motor Neuron Disease and Behavioural Variant Frontotemporal Dementia: A Longitudinal Study,” European Journal of Neurology 29, no. 11 (2022): 3158–3169, 10.1111/ene.15518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Xu Z., Alruwaili A. R. S., Henderson R. D., and McCombe P. A., “Screening for Cognitive and Behavioural Impairment in Amyotrophic Lateral Sclerosis: Frequency of Abnormality and Effect on Survival,” Journal of the Neurological Sciences 376 (2017): 16–23, 10.1016/j.jns.2017.02.061. [DOI] [PubMed] [Google Scholar]
- 10. Folstein M. F., Folstein S. E., and McHugh P. R., ““Mini‐Mental State”: A Practical Method for Grading the Cognitive State of Patients for the Clinician,” Journal of Psychiatric Research 12, no. 3 (1975): 189–198, 10.1016/0022-3956(75)90026-6. [DOI] [PubMed] [Google Scholar]
- 11. Jubb M. T. and Evans J. J., “An Investigation of the Utility of the Addenbrooke's Cognitive Examination III in the Early Detection of Dementia in Memory Clinic Patients Aged Over 75 Years,” Dementia and Geriatric Cognitive Disorders 40, no. 3–4 (2015): 222–232. [DOI] [PubMed] [Google Scholar]
- 12. Nasreddine Z. S., Phillips N. A., Bédirian V., et al., “The Montreal Cognitive Assessment, MoCA: A Brief Screening Tool for Mild Cognitive Impairment,” Journal of the American Geriatrics Society 53, no. 4 (2005): 695–699, 10.1111/j.1532-5415.2005.53221.x. [DOI] [PubMed] [Google Scholar]
- 13. Morris J. C., “Clinical Dementia Rating: A Reliable and Valid Diagnostic and Staging Measure for Dementia of the Alzheimer Type,” International Psychogeriatrics 9, no. S1 (1997): 173–176, 10.1017/S1041610297004870. [DOI] [PubMed] [Google Scholar]
- 14. Mioshi E., Flanagan E., and Knopman D., “Detecting Clinical Change With the CDR‐FTLD: Differences Between FTLD and AD Dementia,” International Journal of Geriatric Psychiatry 32, no. 9 (2017): 977–982, 10.1002/gps.4556. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Rascovsky K., Hodges J. R., Knopman D., et al., “Sensitivity of Revised Diagnostic Criteria for the Behavioural Variant of Frontotemporal Dementia,” Brain 134, no. 9 (2011): 2456–2477, 10.1093/brain/awr179. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Gorno‐Tempini M. L., Hillis A. E., Weintraub S., et al., “Classification of Primary Progressive Aphasia and Its Variants,” Neurology 76, no. 11 (2011): 1006–1014, 10.1212/WNL.0b013e31821103e6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. McKhann G. M., Knopman D. S., Chertkow H., et al., “The Diagnosis of Dementia due to Alzheimer's Disease: Recommendations From the National Institute on Aging‐Alzheimer's Association Workgroups on Diagnostic Guidelines for Alzheimer's Disease,” Alzheimer's & Dementia 7, no. 3 (2011): 263–269, 10.1016/j.jalz.2011.03.005. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Armstrong M. J., Litvan I., Lang A. E., et al., “Criteria for the Diagnosis of Corticobasal Degeneration,” Neurology 80, no. 5 (2013): 496–503, 10.1212/WNL.0b013e31827f0fd1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Costa J., Swash M., and de Carvalho M., “Awaji Criteria for the Diagnosis of Amyotrophic Lateral Sclerosis:A Systematic Review,” Archives of Neurology 69, no. 11 (2012): 1410–1416, 10.1001/archneurol.2012.254. [DOI] [PubMed] [Google Scholar]
- 20. Strong M. J., Grace G. M., Freedman M., et al., “Consensus Criteria for the Diagnosis of Frontotemporal Cognitive and Behavioural Syndromes in Amyotrophic Lateral Sclerosis,” Amyotrophic Lateral Sclerosis 10, no. 3 (2009): 131–146, 10.1080/17482960802654364. [DOI] [PubMed] [Google Scholar]
- 21. Tang‐Wai D. F., Graff‐Radford N. R., Boeve B. F., et al., “Clinical, Genetic, and Neuropathologic Characteristics of Posterior Cortical Atrophy,” Neurology 63, no. 7 (2004): 1168–1174, 10.1212/01.WNL.0000140289.18472.15. [DOI] [PubMed] [Google Scholar]
- 22. Litvan I., Agid Y., Calne D., et al., “Clinical Research Criteria for the Diagnosis of Progressive Supranuclear Palsy (Steele‐Richardson‐Olszewski Syndrome): Report of the NINDS‐SPSP International Workshop,” Neurology 47, no. 1 (1996): 1–9, 10.1212/WNL.47.1.1. [DOI] [PubMed] [Google Scholar]
- 23. Chui H. C., Victoroff J. I., Margolin D., Jagust W., Shankle R., and Katzman R., “Criteria for the Diagnosis of Ischemic Vascular Dementia Proposed by the State of California Alzheimer's Disease Diagnostic and Treatment Centers,” Neurology 42, no. 3 (1992): 473. [DOI] [PubMed] [Google Scholar]
- 24. Gold G., Bouras C., Canuto A., et al., “Clinicopathological Validation Study of Four Sets of Clinical Criteria for Vascular Dementia,” American Journal of Psychiatry 159, no. 1 (2002): 82–87. [DOI] [PubMed] [Google Scholar]
- 25. Gilman S., Wenning G. K., Low P. A., et al., “Second Consensus Statement on the Diagnosis of Multiple System Atrophy,” Neurology 71, no. 9 (2008): 670–676. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Mioshi E., Dawson K., Mitchell J., Arnold R., and Hodges J. R., “The Addenbrooke's Cognitive Examination Revised (ACE‐R): A Brief Cognitive Test Battery for Dementia Screening,” International Journal of Geriatric Psychiatry 21, no. 11 (2006): 1078–1085, 10.1002/gps.1610. [DOI] [PubMed] [Google Scholar]
- 27. R Core Team , “R: A Language and Environment for Statistical Computing [Computer Software],” (R Foundation for Statistical Computing, 2023), https://www.R‐project.org/.
- 28. Posit Team , “RStudio: Integrated Development Environment for R [Computer Software],” (Posit Software, PBC, 2023), http://www.posit.co/.
- 29. Chelune G. J., Naugle R. I., Lüders H., Sedlak J., and Awad I. A., “Individual Change After Epilepsy Surgery: Practice Effects and Base‐Rate Information,” Neuropsychology 7 (1993): 41–52, 10.1037/0894-4105.7.1.41. [DOI] [Google Scholar]
- 30. Frerichs R. J. and Tuokko H. A., “A Comparison of Methods for Measuring Cognitive Change in Older Adults,” Archives of Clinical Neuropsychology 20, no. 3 (2005): 321–333, 10.1016/j.acn.2004.08.002. [DOI] [PubMed] [Google Scholar]
- 31. Heaton R. K., Temkin N., Dikmen S., et al., “Detecting Change: A Comparison of Three Neuropsychological Methods, Using Normal and Clinical Samples,” Archives of Clinical Neuropsychology 16, no. 1 (2001): 75–91, 10.1016/S0887-6177(99)00062-1. [DOI] [PubMed] [Google Scholar]
- 32. Martyr A., Ravi M., Gamble L. D., et al., “Trajectories of Cognitive and Perceived Functional Decline in People With Dementia: Findings From the IDEAL Programme,” Alzheimer's & Dementia 20, no. 1 (2023): 410–420, 10.1002/alz.13448. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Foxe D., Irish M., Cheung S. C., et al., “Longitudinal Changes in Functional Capacity in Frontotemporal Dementia and Alzheimer's Disease,” Alzheimer's & Dementia: Diagnosis, Assessment & Disease Monitoring 16, no. 4 (2024): e70028, 10.1002/dad2.70028. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Foxe D., Irish M., Hu A., et al., “Longitudinal Cognitive and Functional Changes in Primary Progressive Aphasia,” Journal of Neurology 268, no. 5 (2021): 1951–1961, 10.1007/s00415-020-10382-9. [DOI] [PubMed] [Google Scholar]
- 35. Landin‐Romero R., Kumfor F., YSLee A., Leyton C., and Piguet O., “Clinical and Cortical Trajectories in Non‐Fluent Primary Progressive Aphasia and Alzheimer's Disease: A Role for Emotion Processing,” Brain Research 1829 (2024): 148777, 10.1016/j.brainres.2024.148777. [DOI] [PubMed] [Google Scholar]
- 36. Murphy D. F., Scott J. P., and Noad R. F., “Measuring Change Over Time on Cognitive Screening Measures: An Evaluation of the Psychometric Properties of ACE‐III, CBI‐R, and EMQ for the Purpose of Dementia Screening,” Dementia and Geriatric Cognitive Disorders 53, no. 2 (2024): 47–56, 10.1159/000534313. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data S1
Data Availability Statement
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. No part of the study procedures or analyses was preregistered prior to the research being undertaken. The Addenbrooke's Cognitive Examination‐Third edition (ACE‐III) is freely available at https://frontierftd.org (accessed on 03 June 2025).