Neuroimging to Predict.... Alzheimer's: Amyloid Eclipses Genetic Risk
Download the PDF here |
By Crystal Phend, Senior Staff Writer, MedPage Today
Published: October 15, 2012
Memory may fade faster in brains clogged by amyloid plaques than those at genetic risk of Alzheimer's disease, researchers found.
High levels of beta-amyloid protein in the brain were associated with greater decline in working memory and verbal and visual episodic memory over 18 months among cognitively normal individuals, Yen Ying Lim, MPsych, of the University of Melbourne in Victoria, Australia, and colleagues reported in the Oct. 16 issue of Neurology.
Carriers of the APOE _e4 genotype also showed accelerated decline in visual memory, but to a lesser degree than the high-amyloid group in the longitudinal cohort study. The two factors didn't appear to interact.
These findings support amyloid imaging as a marker in preclinical Alzheimer's disease, Lim's group noted.
The "subtle decline in memory that characterizes such progression [to mild cognitive impairment] can be detected over relatively short time intervals (i.e., 18 months), even in the absence of any change in clinical status," they explained.
Thus, amyloid imaging plus monitoring decline in memory could be used to catch the Alzheimer's disease process before it ravages the brain, paving the way for early intervention.
Anti-amyloid vaccines and antibody therapies haven't proven successful so far in tackling clinical symptoms in established Alzheimer's disease, commented Brian Appleby, MD, of the Cleveland Clinic's Center for Brain Health. While that has led to questions about the amyloid hypothesis overall, Appleby noted that accumulation of these plaques may still be key but that they need to be targeted earlier in the disease process to make a difference.
The FDA has already approved a tracer for amyloid imaging in diagnosis of dementia, but clinicians have hesitated to use the expensive scans for individuals without cognitive decline without knowing what the results would mean in this group, he noted.
"This is a very good first step in understanding what amyloid accumulation means for normal people," he said in an interview.
The Alzheimer's gene isn't as good a discriminator in targeting these therapies because at least half of those who get the disease aren't APOE e4 carriers, Appleby added.
Lim and colleagues' analysis included 141 healthy older adults (mean age 76) who were cognitively normal at the baseline round of testing and PET neuroimaging in the Australian Imaging Biomarkers and Lifestyle study.
Overall, 45 of the individuals had high beta-amyloid levels marked by a standardized Pittsburgh compound B tracer uptake value ratio at or above _1.
They showed significantly more decline at the 18-month reassessment compared with those who had less amyloid buildup for verbal and visual episodic memory scores on the California Verbal Learning Test II, both overall (P<0.001) and delayed (P<0.05).
The same was true for working memory assessed on the Working Memory-Learning Composite test and visual episodic memory on the visual Paired Associate Learning task (both P<0.001).
The researchers called these effects of high cerebral beta-amyloid burden moderate in magnitude.
The only cognitive disadvantage for the 63 APOE e4 carriers was greater decline in visual learning as assessed with the One Card Learning task, compared with noncarriers of the risk allele (P<0.001).
Nether the amyloid levels or genotype correlated with psychomotor function or attention.
Clinical status with regard to dementia or Alzheimer's disease didn't change in these patients with relatively high IQ levels at baseline; high IQ has been suggested as being protective against amyloid-related cognitive decline, the researchers noted.
Adjustment for baseline intelligence didn't change the results.
The researchers cautioned, though, that they used memory tests that could be performed quickly without improvements in scores solely from more practice at them, rather than more extensive assessments of cognitive function.
Co-authors reported financial relationships with CogState (the company that provided the battery of tests used in this study), Novartis, Eli Lilly, Janssen, Pfizer, Lundbeck, GlaxoSmithKline, Forest Laboratories, CSIRO, Alzhyme, Bayer Schering Pharma, Elan, AstraZeneca, Avid Radiopharmaceuticals, the Alzheimer's Association, sanofi-aventis, Mayne Pharma, NHMRC and NEDO in Japan.
One co-author is employed by CogState.
Stronger effect of amyloid load than APOE genotype on cognitive decline in healthy older adults
"The aim of the current study was to examine associations between cerebral Aß load, APOE genotype, and cognitive change over 18 months in a large group of healthy older adults. We hypothesized that increased cerebral Aß load and APOE ∈4 genotype would be associated with episodic memory decline over 18 months. ......... Healthy and cognitively normal older adults (n = 141; mean age 76 years) underwent PET neuroimaging for cerebral Aß, APOE genotyping, and cognitive assessment as part of their baseline assessment in the Australian Imaging Biomarkers and Lifestyle study. Cognitive function was reassessed 18 months later......In this prospective study of healthy older adults, high cerebral Aß load was associated with greater decline in episodic and working memory over 18 months. The APOE ∈4 genotype was also associated with a decline in visual memory, although the effect was less than that observed for cerebral Aß load......An important methodologic issue needs to be considered when interpreting the current results because an extensive investigation of cognitive function was not conducted. The tasks used were chosen based on brevity, test-retest reliability, demonstrated sensitivity to change, and ability to be applied repeatedly without the generation of practice effects.21,40 Exploration using more detailed neuropsychological tests will be useful in further elucidating the nature of cerebral Aß-related decline in cognition."
Objective: Although the APOE ∈4 allele is associated with more rapid decline in memory in healthy older adults, the significance of elevated cerebral ß-amyloid (Aß) load for longitudinal changes in cognition is unclear.
Methods: Healthy and cognitively normal older adults (n = 141; mean age 76 years) underwent PET neuroimaging for cerebral Aß, APOE genotyping, and cognitive assessment as part of their baseline assessment in the Australian Imaging Biomarkers and Lifestyle study. Cognitive function was reassessed 18 months later.
Results: Linear mixed-model analyses adjusted for baseline cognitive function indicated that, relative to individuals with low cerebral Aß, individuals with high cerebral Aß showed significantly greater decline in working memory and verbal and visual episodic memory at 18 months. Compared with noncarriers, APOE ∈4 carriers showed a greater decline in visual memory at the 18-month assessment. No interaction between APOE ∈4 and cerebral Aß load was observed for any measure of cognitive function.
Conclusions: In this prospective study of healthy older adults, high cerebral Aß load was associated with greater decline in episodic and working memory over 18 months. The APOE ∈4 genotype was also associated with a decline in visual memory, although the effect was less than that observed for cerebral Aß load.
ADNI=Alzheimer's Disease Neuroimaging Initiative;
AIBL=Australian Imaging Biomarkers and Lifestyle;
ANCOVA=analysis of covariance;
CVLT-II=California Verbal Learning Test, Second Edition;
LMM=linear mixed model;
MCI=mild cognitive impairment;
OBK=One Back task;
OCL=One Card Learning task;
OCL-OBK=Working Memory-Learning Composite;
PAL=Paired Associate Learning task;
PiB=Pittsburgh compound B;
SUV=standardized uptake value;
SUVR=standardized uptake value ratio
Neuroimaging studies reveal abnormally high cerebral ß-amyloid (Aß) loads in a substantial proportion of healthy older adults.1,-,4 However, associations between cerebral Aß load and cognitive function are generally only small in magnitude, with the strongest relationships found for episodic memory.5,-,8 Recent cross-sectional studies of healthy older adults reported moderate associations between cerebral Aß burden and episodic memory when analyses were restricted to carriers of the APOE ∈4 allele. No such relationships were observed in noncarriers,5,6 suggesting that in healthy older adults, elevated cerebral Aß load may be a more important prognostic factor in individuals genetically at risk for Alzheimer disease (AD). However, the implications of a high cerebral Aß load for cognitive function should be determined in prospective studies. Further, whereas it is well known that healthy older adults who carry the APOE ∈4 allele show accelerated cognitive decline,9,10 it is not clear whether this allele may also moderate the relation between cerebral Aß level and cognitive decline.
Two prospective Pittsburgh compound B (PiB) PET neuroimaging studies have reported no differences between healthy older adults with high and low cerebral Aß on the rate of cognitive change over 24 months. However, in both studies, the samples were small (n ≤ 30), and neither examined the effect of the APOE genotype.8,11 The aim of the current study was to examine associations between cerebral Aß load, APOE genotype, and cognitive change over 18 months in a large group of healthy older adults. We hypothesized that increased cerebral Aß load and APOE ∈4 genotype would be associated with episodic memory decline over 18 months.
Results of this study support the hypothesis that in healthy older adults high cerebral Aß load is associated with a decline in episodic memory over 18 months. Compared with the cognitive performance of healthy older adults with normal cerebral Aß load, older adults with high cerebral Aß load showed greater decline across each aspect of episodic memory assessed and the magnitudes of these were by convention moderate26 (i.e., Cohen d = 0.40- 0.60). A decline of comparable magnitude was observed for visual working memory (figure). Because the 95% confidence intervals for these different effect sizes showed substantial overlap, the current data are interpreted most parsimoniously as showing that high cerebral Aß load gave rise to equivalent decline across all aspects of episodic and working memory. In contrast to memory, no effect of cerebral Aß load was observed for psychomotor function or visual attention. Hence, the decline in episodic memory observed could not have been due to changes in arousal or attention.
Consistent with previous reports,1,4,11,28 we observed no differences between the low and high cerebral Aß load groups or between APOE ∈4 carriers and noncarriers, in performance on any of the cognitive tests at the time of neuroimaging (table 2). Healthy older adults with low and high cerebral Aß load who were APOE ∈4 carriers and noncarriers were also equivalent on each demographic characteristic (table 1). Furthermore, when the healthy adults were considered as a single group, no decline from the baseline to the 18-month assessment was observed in any aspect of cognition (table 2), and review at 18 months confirmed that their clinical status had not changed. The relatively high estimated premorbid intelligence of the control group in the AIBL cohort has been noted previously,12 and this was also evident in the current subsample. It has been suggested the high premorbid intelligence could protect against cerebral amyloid-related cognitive decline.27 However, in the current sample, there was no difference between high and low SUVR groups in estimated premorbid intelligence. Furthermore, the effects of cerebral Aß load and APOE ∈4 on decline in episodic and working memory did not change when age and estimated premorbid intelligence were controlled for statistically in the analyses. These aspects of the results support the conclusion that the decline in memory and working memory observed after 18 months in healthy older adults was related specifically to high cerebral Aß load.
The decline in episodic and working memory associated with high cerebral Aß load was not moderated at all by the APOE ∈4 status of the healthy older adults. However, when considered by itself, the APOE ∈4 allele was associated with statistically significant decline over 18 months in visual memory, although the magnitude of this decline was slightly less than that related to differences in cerebral Aß load. Inspection of the magnitudes of differences between APOE ∈4 carrier and noncarrier groups for the other cognitive functions indicated moderate26 (i.e., Cohen d = 0.24-0.38) but nonsignificant decline for visual episodic memory, for verbal learning, and for the OCL-OBK (figure). This consistency suggests that these effect sizes did reflect true differences in the level of cognitive decline between APOE ∈4 carriers and noncarriers, although the magnitudes of these differences were not sufficient to be rendered statistically significant in the current study. The presence of an APOE ∈4 allele-related decline in memory in healthy older adults is consistent with previous observations,9,10 although the more subtle effect of APOE ∈4 positivity observed here could reflect the fact that the current sample was younger than some samples studied recently10 and studied over a shorter interval than that in the benchmark study of this issue.9 Thus, although these data do suggest that both cerebral Aß load and APOE ∈4 status predict decline in episodic memory over 18 months, the magnitude of the effect of cerebral Aß load was greater than that of APOE ∈4 status. More important, though, LMM analyses and inspection of the group mean change scores (table 4) suggest strongly that the presence of the APOE ∈4 allele did not moderate cerebral Aß-related decline in memory in healthy older adults.
Two small prospective studies have investigated the extent to which cerebral Aß load, measured using PiB PET neuroimaging, is related to change in cognitive function in healthy older adults.8,11 As in the current study, the entire group of healthy older adults from the Melbourne Healthy Aging Study cohort showed no change in memory or nonmemory composite scores 18 or 36 months after a baseline PiB PET scan. However, when adults in this group were classified as having high or low cerebral Aß load, using the same criterion as that used in the current study, healthy older adults with high cerebral Aß load showed statistically significant decline of a moderate magnitude on both memory and nonmemory composite scores 36 months after PET scanning.8 However, by the 36-month assessment, 25% of healthy older adults with high cerebral Aß load met the clinical criteria for mild cognitive impairment (MCI). Thus, the general decline in cognition observed at 36 months postscan reflected this change in clinical status. Conversely, in a subset of healthy control subjects recruited from the Alzheimer's Disease Neuroimaging Initiative (ADNI), individuals with low or high cerebral Aß load did not differ in the rate at which their global cognitive ability and verbal episodic memory deteriorated over 15 months. Given the magnitudes of the effects observed here, the sample studied in the ADNI cohort was probably too small to allow for the detection of any subtle decline in memory.11 Taken together, the results from these 2 previous prospective studies and the current study suggest that in healthy older adults, high cerebral Aß does increase the risk of progression to MCI and that the subtle decline in memory that characterizes such progression can be detected over relatively short time intervals (i.e., 18 months), even in the absence of any change in clinical status. Our results are consistent with prospective studies of healthy older adults with increased cerebral amyloid determined from analysis of CSF.29,30 These studies also report that pathologic levels of CSF Aß42 are associated with increased rates of cognitive decline, although like the neuroimaging studies, these changes were observed over much longer intervals than those observed here (e.g., 4-10 years).29,30 In the current cohort of individuals with high cerebral Aß, we expect that the decline in memory will continue until it reaches levels that would be considered as reflecting cognitive impairment (compared with normative data) and the clinical status of individuals changes from cognitively normal to indicative of MCI.
The finding that cerebral Aß-related decline in memory can occur without a change in clinical status also raises an important issue about the methods for classifying levels of cognitive function in healthy older adults. Our group has suggested that the earliest stages of AD may be most accurately detected through objective evidence for decline in cognitive function, in particular memory function, rather than from the detection of impairment in cognition.31,-,33 Recent recommendations for the definition of the preclinical stage of AD have also recommended that the identification of objectively defined cognitive decline may assist in the detection of preclinical AD.34,35 Thus, the combination of objectively defined cognitive decline with putative AD biomarkers should considerably strengthen the ability to detect preclinical AD.34
Although these are the first studies to investigate the extent to which amyloid imaging biomarkers predict future cognitive decline, previous research has sought to determine the extent to which decline in cognitive function predicts an increase in cerebral Aß load. For example, in a series of studies of healthy adults followed prospectively with brief batteries of cognitive tests, we identified that statistically reliable decline in aspects of verbal and visual episodic memory tests over 6 years13 2 years,36 and even 1 year37 was associated with increased risk of high cerebral Aß load. Other studies have observed similar declines in cognition in individuals with high cerebral Aß load.28,38 When considered with the results of the current study, these data suggest that the combination of decline in memory with parameters from amyloid imaging may be useful for the identification of the AD process in individuals who do not meet any clinical criteria for cognitive impairment. In healthy older adults with high cerebral Aß load, an objectively defined decline in memory may also be a worthy target for anti-amyloid therapies.39
An important methodologic issue needs to be considered when interpreting the current results because an extensive investigation of cognitive function was not conducted. The tasks used were chosen based on brevity, test-retest reliability, demonstrated sensitivity to change, and ability to be applied repeatedly without the generation of practice effects.21,40 Exploration using more detailed neuropsychological tests will be useful in further elucidating the nature of cerebral Aß-related decline in cognition.
Of the total sample, there was one noncompletion for DET and PAL, 2 noncompletions for OCL, and 2 noncompletions for OBK at the 18-month assessment. Complete baseline and 18-month data were available for all other tasks. LMM analyses indicated that, relative to older adults with SUVR <1.5 at baseline, older adults with SUVR ≥1.5 showed significantly greater decline at the 18-month assessment on all measures of verbal and visual episodic memory as well as working memory (table 3). No decline at the 18-month assessment was observed for measures of psychomotor or attentional function in either SUVR group. The magnitudes of the differences in baseline-adjusted performance between the low and high SUVR groups at the 18-month assessment are shown in the figure. Inspection of group mean raw change scores also indicate that when averaged across APOE ∈4 status, individuals in the high SUVR group showed greater decline on tasks of verbal and visual episodic memory, as well as working memory, at the 18-month assessment (table 4).
When APOE ∈4 status was added to the LMM analyses, no statistically significant interaction between SUVR status and APOE ∈4 status was observed for any cognitive measure. However, when considered by itself, the presence of the APOE ∈4 allele was associated with a greater decline in visual memory after 18 months. Magnitudes of differences in baseline-adjusted performance between APOE ∈4 carriers and noncarriers at the 18-month assessment are shown in the figure (light green bars). The greater decline in performance on memory scores in the APOE ∈4 carriers, averaged over SUVR groups, is also shown in table 4.
Reanalysis of the data with age and premorbid intelligence included as covariates did not change the pattern of results. No interactions between SUVR and APOE ∈4 status were observed (all p > 0.5). However, statistically significant main effects of SUVR were observed for the OCL, OBK, PAL, CVLT-II Total and Delayed Recall, and OCL-OBK (all p < 0.05). Statistically significant main effects of APOE ∈4 status were observed for the OCL (p < 0.05).
Participants were recruited from the healthy control group enrolled in the Australian Imaging Biomarkers and Lifestyle (AIBL) flagship study of aging.3,12 The process of recruitment and diagnostic classification of healthy older adults enrolled in the AIBL cohort has been described in detail elsewhere.12 Healthy participants who volunteered were excluded from the AIBL study if they had any of the following: schizophrenia; depression (Geriatric Depression Score of 6 or greater); Parkinson disease; cancer (other than basal cell skin carcinoma) within the last 2 years; symptomatic stroke; uncontrolled diabetes; or current regular alcohol use exceeding 2 standard drinks per day for women or 4 per day for men. A clinical review panel chaired by D.A. reviewed all available medical, psychiatric, and neuropsychological information to confirm the cognitive health of individuals enrolled in the healthy control group. In this study, only the subgroup of healthy older adults who had undergone PiB neuroimaging and who had completed the cognitive battery at baseline and 18-month assessment (n = 141) were included. Demographic and clinical characteristics of the healthy control subjects are shown in table 1. The clinical status of all participants did not change at 18 months.
Standard protocol approvals, registrations, and patient consents.
The study was approved by and complied with the regulations of 3 institutional research and ethics committees,12 and all participants provided written informed consent before participating in the study.
PiB PET neuroimaging and APOE ∈4 genotyping.
PiB PET imaging methodology has been outlined in detail previously.3,13 PET standardized uptake value (SUV) data acquired 40-70 minutes after PiB injection were summed and normalized to the cerebellar cortex SUV, resulting in a region/cerebellar ratio termed the standardized uptake value ratio (SUVR). An 80-mL blood sample was also taken from each participant, 0.5 mL of which was forwarded for APOE genotyping at a clinical pathology laboratory.
All participants were assessed with the clinical rating scales and neuropsychological battery from the AIBL study (table 2). These have all been described in detail elsewhere and were administered according to standard protocols by trained research assistants.7,12,14 The clinical status of participants was determined by data that included the Mini-Mental State Examination15 and Clinical Dementia Rating Scale.16 Premorbid intelligence was estimated using the Wechsler Test of Adult Reading,17 and levels of depressive and anxiety symptoms were assessed using the Hospital Anxiety and Depression Scale.18 Verbal learning and verbal episodic memory were measured using the California Verbal Learning Test, Second Edition (CVLT-II),19 and visual episodic memory was measured using the visual Paired Associate Learning (PAL) task.20 All individuals also performed a set of computerized cognitive tests, which included measures of visual learning (One Card Learning task [OCL]), working memory (One Back task [OBK]), attention (Identification task [IDN]), and psychomotor (Detection task [DET]) function from the CogState battery.14,21,22 Performance on the tests from the CogState battery was not used to classify individuals' clinical status.
All healthy participants in this study underwent an extensive medical, psychiatric, and neuropsychological assessment upon enrollment into the AIBL study. The same assessments were repeated 18 months later. In this study, we report PiB neuroimaging and APOE ∈4 genotyping data obtained at baseline and neuropsychological data obtained at baseline and 18 months to examine the rate of cognitive change in relation to cerebral Aß load and APOE ∈4 status.
Each cognitive task provided a single performance score (table 2). A Working Memory-Learning Composite (OCL-OBK) score was generated by standardizing the OCL and OBK scores and then averaging them. A Psychomotor-Attention Composite (DET-IDN) score was generated by standardizing the DET and IDN scores and then averaging them. Consistent with observations from other studies,5,8 the distribution of PiB SUVR data was skewed negatively and could not be normalized with data transformations. Thus, SUVR was classified dichotomously as either low (SUVR <1.50) or high (SUVR ≥1.50) in accordance with established criteria.8,23
There were no statistically significant differences between SUVR and APOE groups for any of the demographic characteristics (table 1). A series of linear mixed-model (LMM) analyses of covariance (ANCOVA) was conducted to examine the relationship between SUVR status (SUVR <1.5 vs SUVR ≥1.5), APOE ∈4 status (APOE ∈4 carrier vs APOE ∈4 noncarrier), and cognitive change between baseline and 18-month assessment. The LMM procedure was used because of its ability to model both fixed and random effects, which accounts for various sources of variability, and because it provides improved estimates of within-subject coefficients (i.e., random effects) in longitudinal studies.24,25 For each task, an LMM ANCOVA with SUVR status, APOE ∈4 status, and the SUVR X APOE ∈4 status interaction were entered as fixed factors, participant as a random factor, baseline cognitive test score as the only covariate, and cognitive test score at the 18-month assessment as a dependent variable. For each performance measure, the magnitude of the difference in adjusted means between low and high SUVR groups and between APOE ∈4 carrier and noncarrier groups at the 18-month assessment was expressed using the Cohen d.26 To assist with interpretation of the LMM, group mean and SD change scores for each SUVR and APOE group, as well as for SUVR high and low groups averaged over APOE ∈4 status and APOE ∈4 carriers and noncarriers averaged over SUVR status, were calculated. Although estimates of premorbid intelligence were not different between the study groups (table 1), analyses were recomputed with age and premorbid intelligence included as covariates to examine the extent to which cognitive reserve may affect the relationship between cerebral Aß level, genotype, and change in cognition.27