In this study, we evaluated AQT as a test for detecting early ChEI treatment response in AD and compared it with the MMSE. After 8 weeks of treatment, AQT had improved significantly more than the MMSE when accounting for disease progression (Figure
2). Further, AQT identified twice as many treatment responders as did the MMSE (34% compared with 17%; p = 0.02; Figure
3). The increased number of responders cannot be explained by low reliability or random changes of AQT scores, because AQT classified only 5% (false) responders during the 8-week period before treatment (Figure
3). Finally, when comparing the AQT responders and nonresponders from the 8-week visit, the responders still showed a significantly better treatment response after 6 months of treatment. This indicates that AQT detects early treatment responders who seem to continue to benefit from ChEI treatment.
Evaluation of treatment
Good clinical practice and cost-benefit considerations require that all AD patients be evaluated before and after the initiation of treatment to determine whether the treatment shall continue [
5]. The most common test for this evaluation in clinical practice is the MMSE, and this is also the recommended test according to the NICE guidelines [
5]. In clinical trials, the ADAS-cog is the most commonly used cognitive test [
28,
29]. It measures a broader span of cognitive functions, but has the disadvantage of taking 45 minutes to administer compared with 3 to 5 minutes for AQT and 10 to 15 minutes for the MMSE. Because of the length of the ADAS-cog, it cannot really be regarded as a brief cognitive test suitable for clinical practice. ADAS-cog and the MMSE are well studied for ChEI evaluation of AD, but no previous studies of AQT were performed in this context. However, in a recent randomized, placebo-controlled, multinational study, AQT was used to evaluate the treatment effect of memantine on dementia with Lewy bodies and Parkinson disease dementia [
30]. In that study, both AQT and the global cognitive measure CGIC improved significantly after 24 weeks of treatment, compared with placebo, whereas the MMSE failed to improve significantly.
Future evaluation issues
In the future, it is likely that more patients with mild cognitive impairment (MCI) will be included in therapeutic trials and treated in clinical practice. It is then essential to have a sensitive test with no ceiling effect. The MMSE and the ADAS-cog have detected in MCI studies significant cognitive changes [
31,
32], but they have also been criticized for their ceiling effects and inability to detect small cognitive changes [
28]. In the only study in which AQT has been used to evaluate MCI treatment, AQT improved significantly, whereas the other cognitive tests failed to do so (WAIS III Digit Span, WAIS-R NI Spatial Span, Digit Symbol Modalities, and Rey's Complex Figure Test) [
33]. Further, AQT has no ceiling effect and, in this study, was able to significantly detect the subtle disease progress of AD during the nontreatment period of 8 weeks (Table
1). Although the results are promising, more studies are needed to warrant the sensitivity of AQT to cognitive change and to systematically compare it with the MMSE and the ADAS-cog.
Another important future issue is that by 2040, it is predicted that 71% of all dementia patients will be in developing countries [
34]. Therefore, the International Psychogeriatric Association (IPA) and the Alzheimer's Association have pointed out the need for a culturally independent test [
7,
28]. AQT has so far been validated in Western, Arabic, and African countries and does not exhibit any culturally dependent questions or exercises [
8,
9,
35], whereas the MMSE is affected by ethnicity [
36,
37].
Detecting treatment responders
In the present study, we evaluated the treatment response after 8 weeks. Previously, it was shown that 4 to 8 weeks of AD treatment results in a significant treatment effect compared with placebo [
38‐
41]. This supports our evaluation of treatment effects already after 8 weeks. It is also is in agreement with the guidelines by NICE and the American College of Physicians (ACP) [
5,
7]. When evaluating the treatment response, the ACP has suggested that a 3-point change in the MMSE indicates a clinically significant change [
7]. This is also the same result as the present study found to indicate a significant change (Figure
3). Unfortunately, no comparable studies are available regarding individual change on AQT.
This study used a statistical method (RCI) to determine treatment responders according to the MMSE and AQT. The clinical relevance of an MMSE improvement of at least 3 points or an AQT improvement of at least 16 seconds is uncertain. In the entire population, the clinical relevance of a mean AQT improvement of 10.8% and a mean MMSE improvement of 3.7% is also uncertain. It is important to note that these values were only used to compare the MMSE and AQT as evaluation instruments. To determine a clinically meaningful AQT or MMSE change, a minimal clinically important difference (MCID) must be defined. One approach to determine the MCID for AD could be to measure the natural history of decline over 12 months or longer in a large group of patients by using AQT, the MMSE, and a global rating of the cognitive performance. A definition of MCID could then be the percentage of change on the MMSE or AQT that is anchored against the natural history of global change in AD.
According to the cut-off values, AQT detected significantly more responders after 8 weeks of treatment than did the MMSE (34% compared with 17%;
P = 0.026), while falsely detecting 5% responders when no treatment was given (Figure
3). This indicates that AQT is a more-sensitive evaluation tool, which is further emphasized by the changes on a group level. AQT improved significantly more after treatment than did the MMSE when accounting for disease progression (Figure
2). The more-pronounced sensitivity of AQT compared with the MMSE might be explained by their different scales and the different cognitive functions that are measured. Studies have shown that ChEI mostly improves attention [
13,
14], which is one of the main cognitive domains measured by AQT. It is possible that the treatment response in our study could have been higher if all ChEI doses had been increased after 4 weeks of treatment (the dose was often increased after 8 weeks). This should, however, not affect the comparison between AQT and the MMSE.
Intuitively, it seems that patients who exhibit the right characteristics initially to have a positive treatment response would continue to benefit from the medication. This assumption has been debated, and to determine whether it is true, the reliability and validity of the evaluation instrument must be high. In our study, we found that the AD patients who were classified as treatment responders by AQT after 8 weeks of treatment still performed significantly better on AQT after 6 months, compared with the patients classified as nonresponders after 8 weeks (22.6 seconds in mean difference; P < 0.0001). This indicates that AQT might be used after 8 weeks of ChEI treatment to identify those who will continue to benefit from the treatment.
Two advantages of this study are that the treatment was evaluated prospectively and that the same population was used both as controls and as cases. The latter is the most important factor for reliable RCI results, as most confounding factors are eliminated. A shortcoming was that this was not a randomized study with a placebo group, but instead a study with a control group. The treatment effect can therefore not with certainty be separated from the placebo effect. However, placebo treatment in clinical trials of AD patients has not resulted in significant improvements of any cognitive tests [
38‐
41]. Furthermore, the lack of a placebo group should not affect the comparison of the MMSE and AQT.