Selective Ability of Some CANTAB Battery Test Measures to Detect Cognitive Response to a Single Dose of Donepezil in Alzheimer Disease

Background

OBJECTIVES:

The objective the present study was to establish the ability of CANTAB tests and their measures to detect significant cognitive change after a single 5-mg dose of donepezil in treatment-naïve AD patients.

Material and Methods

PARTICIPANTS:

This prospective, randomized, rater-blind study was performed at the Memory Disorders Unit of the Neurology Center, Vilnius University Hospital Santariskiu Clinics. We enrolled 92 subjects in the study. We recruited 62 consecutive, de novo-diagnosed, treatment-naïve AD patients and 30 healthy controls (Control group, CG) matched according to age, education, and gender. All patients were diagnosed with AD in usual clinical practice settings by a neurologist not involved in this study. Patients started their treatment with donepezil when the medication was prescribed by a neurologist of the Memory Disorders Unit. After the assessment day (the first day of dosing), the patients continued their treatment with donepezil as per usual clinical practice rules according to the treatment guidelines established by the Lithuanian Ministry of Health in the directives No. 382 and V-156. No modifications to the patient treatment were made due to this research.

STUDY DESIGN:

Informed consent was obtained, screening evaluation (including MMSE and GDS) was performed, inclusion/exclusion criteria were verified, and both sessions of CANTAB testing were performed on the same day when the patients took their first dose of donepezil. AD patients were randomly assigned to 1 of 2 research groups with the ratio 1: 1 using the sequence of random numbers 1 or 2, produced by the on-line Research Randomizer at http://www.randomizer.org/. Thirty AD patients were assigned to the AD+ group and 32 patients to the AD− group. Patients allocated to the AD+ group received a 5-mg donepezil tablet immediately after the first CANTAB testing session. The second testing session was performed 4 hours after the AD+ group patients took donepezil. The 4-hour period was selected because this corresponds to the time at which the peak plasma concentration of donepezil is observed after oral administration. Patients in the AD− group underwent both assessment sessions without taking donepezil between sessions (i.e., both AD+ and AD− groups completed the first CANTAB testing session while treatment-naïve). The second CANTAB testing session was completed by the AD+ group 4 hours after the first single 5-mg dose of donepezil. The AD− group completed the second CANTAB testing session 4 hours after the first testing session, but were still treatment-naïve. The AD+ and AD− groups did not differ by age, education, or gender, as verified after the completion of the recruitment period. The neurologist performing CANTAB testing (JK) was blinded to the participant’s assignment to the AD+, AD−, or control group. Randomization was accomplished; MMSE, GDS, and Hachinski ischemic score were4 assessed and donepezil usage instructions were provided by another neurologist (GK). Control group participants were recruited from the spouses, relatives, and accompanying persons of other patients attending the Neurology Department, but with no medical history of AD or other dementia.

APPROVAL BY ETHICS COMMITTEE:

The study Protocol and Informed Consent Form (ICF) were approved by the Vilnius Regional Biomedical Research Ethics Committee (approval No. 158200-12-128-36). Written Informed consent was obtained from all participants.

INCLUSION AND EXCLUSION CRITERIA:

Rigorous inclusion and exclusion criteria were applied for inclusion in the study. Inclusion and exclusion criteria for AD+ and AD− groups were the same. Randomization to AD+ or AD− groups was performed after enrollment.

Inclusion criteria for AD patients were:

Exclusion criteria for AD patients were:

Inclusion criteria for Control group participants were:

Exclusion criteria for Control group participants were the same as for AD patients.

NEUROPSYCHOLOGICAL ASSESSMENT INSTRUMENTS:

Global cognitive performance of participants was assessed using the Lithuanian version of the Mini-Mental State Examination (MMSE). The Cambridge Neuropsychological Test Automated Battery (CANTAB^®, Cambridge Cognition Ltd., UK) was used as the main instrument for detailed and more sensitive research assessment. CANTAB is a computer-based battery using a touch tone screen and press pad with 2 buttons. The order of test sequence remained constant across both test sessions, because 2 special batteries with the same sequence have been set-up of tests selected for this research according to the instructions provided by CANTAB Software User Guide. Separate batteries were assembled for testing 1 and testing 2; the only difference between them was that parallel versions of most tests (where available) were used at different testing sessions to minimize any potential learning effect.

After an initial explanation, subjects were given the following tests in the following order:

Raw scores of CANTAB test measures were selected for statistical assessment. Standard scores and the internal normative database of CANTAB were not used for comparison of results. Instead, our own control group was enrolled in the study because the internal normative database of CANTAB involved healthy volunteers and provides estimations with subjects matched for age, gender, and National Adult Reading Test (NART) scores. There is no analogue to NART in the Lithuanian language and it is hard or impossible to devise something similar due to the phonemic orthography used in the Lithuanian language. Therefore, the score of NART was left blank when entering initial participant data before testing. This invalidates the comparison with the internal normative database of CANTAB. Instead of NART, the education level in years was used as a proxy of premorbid intelligence level. It should be noted that while the learning effect was impossible due to use of parallel test versions in testing 1 and testing 2, the practice effect might have some influence on the results of testing 2.

STATISTICAL ANALYSIS:

Comparisons between groups were performed using analysis of variance (ANOVA) for continuous variables and chi-square test for categorical variables. Normal distribution of data was verified using the Shapiro-Wilk test. One-way ANOVA with Bonferroni post-hoc test was used to assess differences between the results of the first testing session among the 3 participant groups. The Levene test was used to assess the homogeneity of variances across participant groups. The tables provided below indicate when the assumption of homoscedasticity (homogeneity of variance) was violated.

Correlation of CANTAB test measures with MMSE scores was assessed using Pearson correlation coefficient r.

Repeated-measures ANOVA was applied to answer the question, whether the mean change in the cognitive function of the first to second testing session differed in the three groups. Scores on the CANTAB test measures at each testing session were submitted to a repeated-measures analysis of variance with test session number (first and second) entered as independent variables, and CANTAB test measure score entered as a dependent variable. The differences in change were measured directly by the “testing session”*”group” interaction effect. The Bonferroni post hoc test was used for comparisons of the 3 independent groups. Sphericity of the variances of the differences between all possible pairs of groups was tested by using the Mauchly’s sphericity test. The assumption of sphericity has not been violated, which could be due to the fact that there were only 2 levels of repeated measures. All 3 participant groups were included in the data set for repeated-measures ANOVA. Test results of 1^st and 2^nd testing session were included as 2 within-factor levels. Belonging to 1 of the 3 participant groups was entered as a between-factor.

The statistical significance value was set at p<0.05.

Results

DEMOGRAPHIC CHARACTERISTICS, DEPRESSION LEVEL, AND OVERALL COGNITIVE FUNCTION:

Study groups did not differ significantly by age (p=0.828), education (p=0.952), or gender (p=0.948). Demographic characteristics, depression level by GDS, and MMSE scores for all groups are shown in Table 1.

:

Before proceeding to the main purpose of this study, it was necessary to establish which tests and which specific measures of the tests were able to distinguish AD patients from the Control group. For those tests that do not make this distinction, investigation of improvement is pointless. While these tests might be useful for research of cognitive enhancement, they are not relevant to AD research. One-way ANOVA was used to evaluate the significance of differences between participant groups. The Bonferroni post hoc test was used for multiple comparisons between separate groups. Normative data and standard scores (the number of standard deviations from the mean) of the peer group provided by the CANTAB internal database were not used in our research due to a problem with NART scores explained above. Standard scores were invalidated by the absence of National Adult Reading Test (NART) data in our study. The control group, matched for age, education, and gender, was used for comparison purposes.

Results of CANTAB test measures for memory functions (PAL and PRM tests) are provided in Table 2.

Results of CANTAB test measures for other cognitive domains (CRT, SOC, and SWM tests) are provided in Table 3.

CORRELATION BETWEEN CANTAB TEST MEASURES IN FIRST TESTING SESSION AND DEMENTIA SEVERITY (MMSE):

Correlations between CANTAB test measures and MMSE (as a measure of global dementia severity) were established to evaluate whether change in CANTAB test measures may be treated as a clinically relevant cognitive change (improvement). Only those test measures able to distinguish AD groups from the Control group were selected for correlation analysis. Eight measures of PAL and PRM tests showed statistically significant and strong correlation (r>0.7) with MMSE. The repeated measures ANOVA was performed with these test measures.

:

For those CANTAB test measures that were able to distinguish AD and Controls and demonstrated the statistically significant and strong correlation (r>0.7) with MMSE (as a measure of global dementia severity, or global cognitive functioning), repeated-measures ANOVA was used to assess which measures are able to detect significant cognitive change due to a single dose of donepezil.

Significant interaction effect “Testing session” * “Group” was established only for seven PAL test measures (Table 4). For all 7 PAL test measures Bonferroni post-hoc revealed that the difference is significant only between the results of the second testing session in AD+ and the second testing session in AD− group (between-group effect), while the first testing session results between AD+ and AD− groups were insignificant. If more demanding requirements for the ability of the test to detect the change are accepted, that is, significant change should be detected between AD+ group first testing and second testing results (within treatment group effect), then only 4 PAL test measures are able to fulfil this requirement. This means that only these 4 PAL test measures fulfil both criteria to be able to detect significant cognitive change after a single dose of donepezil: 1) significant difference of change in AD+ and AD− groups; and 2) significant change of mean in absolute values between testing 1 and testing 2 in AD+ group due to a single dose of donepezil (within treatment group effect). These 4 best PAL test measures are: PAL Mean trials to success, Total errors (adjusted), Total errors (6 shapes, adjusted), Total trials (adjusted).

Results of interaction effect (“Testing session” * “Group”), between group effect on second testing, and within AD+ group effect are provided in Table 4. It should be noted that Table 4 contains only indicated results, but not all effects assessed by repeated measures ANOVA.

Only some comparisons provided by repeated-measures ANOVA are presented in Table 4. The patterns of change between testing 1 and testing 2 are very variable for different test measures. Figure 1A shows the changes in PAL First trial memory score results. This measure meets criterion 1, but fails to meet criterion 2. Figure 1B and 1C illustrate the best PAL test measures (2 of 4) – PAL Mean trials to success and Total errors (adjusted). Both PAL test measures fulfil both criteria. Figure 1D illustrates that the change of PRM test measure “PRM delayed (Number correct)” was very different. This PRM test measure meets the criterion 2, but fails to meet criterion 1. This variety of changes in CANTAB test results is more explicitly discussed in the Discussion section.

Discussion

The Choice Reaction Time (CRT) test measures speed of response in a simple 2-choice paradigm using a 2-button press pad [26]. Our results did not show a difference in CRT test performance in mild and mild-to-moderate AD patients and the Control group. The CRT test measures speed of response in a simple 2-choice paradigm. It seems that the test is too simple and easy for all participant groups. A clear ceiling effect is visible in the results of this test, but this does not necessarily mean that the speed of psychomotor reactions is not altered in AD patients. Even the CRT may be informative in moderate and severe AD patients, but for our participants, CRT is unable to distinguish AD patients from normal controls. These results could not be extrapolated to other mild dementias, especially those related to dementia plus Parkinsonism syndromes, such as dementia with Lewy bodies (DLB) and Parkinson disease dementia (PDD). However, the CRT is not suitable for evaluation of attention or psychomotor speed in mild AD. The positive implications that can be drawn from our results with the CRT test are that other cognitive results of our study could not be attributed to malfunction of attention systems or psychomotor speed. Deterioration in other cognitive functions, detected by other tests, is independent of attention in our study, as the results of the test for attention (CRT) were normal in AD groups. This allows more confidence in drawing conclusions about other test results, because disorders of attention seriously distort the results of any other cognitive tests. As no clear disorder of attention was detected in our study, memory, executive, and other cognitive disorders can confidently be attributed to the corresponding cognitive systems.

The Stockings of Cambridge (SOC) test assesses spatial planning and motor control. This test gives a measure of frontal lobe function [26]. The results of SOC test in our study are interesting, unequivocal, and clearly need further investigation. The SOC test is used for evaluation of complex executive functions, working memory, and planning, which are attributed to the frontal lobe. Published reports about frontal function in AD are quite controversial. Some of them indicate early abnormalities of executive functions in AD [27,28], and others did not show a significant difference between AD patients and normal controls [11]. Problems that can be solved in 2 minimum moves are easy for AD patients and controls and provide inconclusive results. Problems that can be solved in 4 or 5 minimum moves are quite difficult for AD patients and controls. The “informative window” is very narrow – only problems that could be solved in 3 minimum moves clearly distinguished AD patients and controls (Table 3). Although this result is interesting in itself, it shows how easy is to pass over indicators of frontal dysfunction in AD. Moreover, correlation of the results of the “3 moves task” with overall dementia severity by MMSE is quite weak, albeit significant. We performed repeated-measures ANOVA for this task of the SOC test (results are not provided in this article), but the difference of changes in AD+ and AD− was not significant. These results show that frontal executive functions are affected in mild AD, but research in this field is very demanding and studies should be very cautious and well designed. Importantly, the effect of cholinergic treatment on executive dysfunction in AD is much smaller than the effect on memory, which raises a number of interesting questions, such as “Is executive dysfunction in AD dependent on cholinergic deficit? If not, what is the neurochemical basis of this dysfunction?”. Summarizing, it should be stressed that the investigation of frontal dysfunction in AD is far from over and extensive research is needed to elucidate the place and mechanisms of executive (frontal) dysfunction in AD.

The Spatial Working Memory (SWM) test assesses working memory and strategy use. This test is a sensitive measure of working memory, frontal lobe, and executive dysfunction [26]. While some, but not all, SWM test measures are significantly worse in AD than in normal controls (Table 3, only some significant SWM measures are provided), SWM test results showed only a moderate correlation with the MMSE. Repeated-measures ANOVA with the SWM test provided inconsistent and patchy results. Many of our comments about the SOC test also apply to the SWM, but while the SOC test depends heavily on planning abilities, the SWM places heavy demands on working memory. Comparison of change in AD+ and AD− groups after donepezil administration indicates that SWM is not significantly dependent on the cholinergic status of the brain. Possible use of the SWM test in AD research needs further investigation.

The Pattern Recognition Memory (PRM) test assesses visual recognition memory. PRM is a test of visual pattern recognition memory in a 2-choice forced discrimination paradigm. This test is sensitive to dysfunction in medial temporal areas of the brain and is relatively insensitive to dysfunction in the frontal lobe. PRM is a relatively simple test. Although it is a memory test, as a test for assessment of recognition memory it has different significance in AD than recall memory tests. Recognition memory is relatively well preserved in AD, while disorders of episodic recall memory are a hallmark of AD. Our results showed that the PRM test provided significantly worse results in both AD groups than in the control group (Table 2) and had a strong correlation with MMSE, but repeated-measures ANOVA did not show significant difference of change in AD+ and AD− groups after donepezil challenge (Table 4, Figure 1D), even though the change in AD+ score from testing session 1 to testing session 2 was significant. Because the dynamics of performance of the Control group on the PRM test is similar to AD−, a conclusion could not be drawn about memory change itself, rather it might be hypothesized that a single dose of donepezil somehow increased retaining of attention or decreased the tiredness in AD+ group.

The Paired Associates Learning (PAL) test assesses episodic visual recall memory and new learning [26]. This test is primarily sensitive to changes in medial temporal lobe functioning. PAL is by far the best test in mild and mild-to-moderate AD. These results are in line with previously published findings [29–32]. Most of the PAL test measures provided significantly worse results in both AD groups than in the Control group (Table 2) and had a strong correlation with MMSE. Only measures of the PAL test showed a significant difference of change in AD+ and AD− groups after donepezil challenge (Figure 1A–1C). Some measures of the PAL test were able to detect significant differences of change in AD+ and AD− and statistically significant improvement in the results of AD+ group itself due to a single dose of donepezil (Table 4), while this improvement was absent in the AD− group (Figure 1A–1C). This indicates that such PAL test measures as PAL Mean trials to success, PAL Total errors (adjusted), PAL Total errors (6 shapes, adjusted), and PAL Total trials (adjusted) are reliable for measurements of the effect in donepezil challenge tests. PAL First trial memory score, PAL Stages completed, and PAL Total errors (8 shapes, adjusted) showed significant differences of change in AD+ and AD− groups, but failed to demonstrate significant changes in the AD+ group itself. Again, the situation is ambiguous – performance due to donepezil did not increase significantly (AD+ group), but decrease of performance without donepezil (AD− group) contributed substantially to the overall result of significant difference in change. Tiredness or different ability to retain attention may be responsible for this result. Either way, these CANTAB test measures are not the best to demonstrate the clinically relevant real change in memory itself. We deliberately provide 4 separate graphs in Figure 1 – not for all 4 of the best tests, because their response patterns are quite similar (only 2 of them are shown in Figure 1 – PAL Mean trials to success and PAL Total errors (adjusted), but the graphs of change between testing session 1 and testing session 2 – which demonstrate the different patterns of change discussed above.

It should be noted that our study leaves many questions unanswered. It is not clear whether the initial impressive improvement in PAL test results after donepezil challenge will be durable in long-term treatment and after down-regulation of acetylcholine receptors. It is not clear whether improvement of PAL results after a single dose of donepezil will correlate significantly with overall cognitive function based on MMSE during long-term treatment. The question remains open whether the initial donepezil challenge effect on PAL test results can distinguish responders to donepezil from non-responders after a longer period of treatment. Many remaining questions require follow-up study with the same study population. Moreover, AD in real clinical practice usually is accompanied by cerebrovascular disease and diabetes, which may contribute to the pathogenesis of the AD itself, and also can change response to cholinergic treatment [33]. Our results are important because they indicate there are at least several CANTAB test measures able to detect significant change after a single dose of donepezil. Having more options to evaluate the results of pharmacological challenge in cognitive neurology may be useful in future clinical trials of new symptomatic treatment for AD, may help to identify responders early in the course of treatment, and could provide reliable tools for personalization of AD treatment.

Conclusions

Four CANTAB PAL test measures are able to detect reliable and clinically relevant cognitive change after a single 5-mg dose of donepezil in AD patients: PAL Mean trials to success, PAL Total errors (adjusted), PAL Total errors (6 shapes, adjusted), and PAL Total trials (adjusted). These 4 CANTAB PAL test measures detected a significant difference in cognitive performance change in the AD+ group in comparison with the AD– and Control groups. Mean scores within the AD+ group itself improved significantly following a single 5-mg dose administration of donepezil.