Trajectories of memory decline in preclinical Alzheimer's disease: results from the Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing

Neurobiology of Aging 36 (2015) 1231e1238

Contents lists available at ScienceDirect

Neurobiology of Aging journal homepage: www.elsevier.com/locate/neuaging

Trajectories of memory decline in preclinical Alzheimer’s disease: results from the Australian Imaging, Biomarkers and Lifestyle Flagship Study of Ageing Robert H. Pietrzak a, b, *, Yen Ying Lim c, d, David Ames e, f, Karra Harrington c, Carolina Restrepo c, Ralph N. Martins g, h, Alan Rembach c,1, Simon M. Laws g, h, Colin L. Masters c, Victor L. Villemagne c, i, j, Christopher C. Rowe i, j, Paul Maruff c, k, for the Australian Imaging, Biomarkers and Lifestyle (AIBL) Research Group a United States Department of Veterans Affairs, National Center for Posttraumatic Stress Disorder, Clinical Neurosciences Division, VA Connecticut Healthcare System, West Haven, CT, USA b Department of Psychiatry, Yale School of Medicine, New Haven, CT, USA c The Florey Institute of Neuroscience and Mental Health, The University of Melbourne, Parkville, Victoria, Australia d Department of Neurology, Warren Alpert School of Medicine, Brown University, Providence, RI, USA e Academic Unit for Psychiatry of Old Age, St. Vincent’s Health, Department of Psychiatry, The University of Melbourne, Kew, Victoria, Australia f National Ageing Research Institute, Parkville, Victoria, Australia g Centre of Excellence for Alzheimer’s Disease Research and Care, School of Exercise, Biomedical and Health Sciences, Edith Cowan University, Perth, Western Australia, Australia h Sir James McCusker Alzheimer’s Disease Research Unit, Hollywood Private Hospital, Nedlands, Western Australia, Australia i Department of Nuclear Medicine and Centre for PET, Austin Health, Heidelberg, Victoria, Australia j Department of Medicine, Austin Health, The University of Melbourne, Heidelberg, Victoria, Australia k Cogstate Ltd, Melbourne, Victoria, Australia

a r t i c l e i n f o

a b s t r a c t

Article history: Received 21 October 2014 Received in revised form 10 December 2014 Accepted 13 December 2014 Available online 20 December 2014

Memory changes in preclinical Alzheimer’s disease (AD) are often characterized by heterogenous trajectories. However, data regarding the nature and determinants of predominant trajectories of memory changes in preclinical AD are lacking. We analyzed data from 333 cognitively healthy older adults who participated in a multicenter prospective cohort study with baseline and 18-, 36-, and 54-month follow-up assessments. Latent growth mixture modeling revealed 3 predominant trajectories of memory change: a below average, subtly declining memory trajectory (30.9%); a below average, rapidly declining memory trajectory (3.6%); and an above average, stable memory trajectory (65.5%). Compared with the stable memory trajectory, high Ab (relative risk ratio [RRR] ¼ 2.1), and lower Mini-Mental State Examination (RRR ¼ 0.6) and full-scale IQ (RRR ¼ 0.9) scores were independently associated with the subtly declining memory trajectory; and high Ab (RRR ¼ 8.3), APOE ε4 carriage (RRR ¼ 6.1), and greater subjective memory impairment (RRR ¼ 1.2) were independently associated with the rapidly declining memory trajectory. Compared with the subtly declining memory trajectory group, APOE ε4 carriage (RRR ¼ 8.4), and subjective memory complaints (RRR ¼ 1.2) were associated with a rapidly declining memory trajectory. These results suggest that the preclinical phase of AD may be characterized by 2 predominant trajectories of memory decline that have common (e.g., high Ab) and unique (e.g., APOE ε4 genotype) determinants. Published by Elsevier Inc.

Keywords: Alzheimer’s disease Memory Trajectories Ab APOE

1. Introduction

* Corresponding author at: United States Department of Veterans Affairs, National Center for Posttraumatic Stress Disorder, Clinical Neurosciences Division, VA Connecticut Healthcare System, 950 Campbell Ave 161E, West Haven, CT, USA. Tel.: þ1 860 638 7467; fax: þ1 203 937 3481. E-mail address: [email protected] (R.H. Pietrzak). 1 Dr. Rembach is deceased. 0197-4580/$ e see front matter Published by Elsevier Inc. http://dx.doi.org/10.1016/j.neurobiolaging.2014.12.015

There is consensus that in cognitively normal (CN) older adults, abnormally high levels of amyloid beta (Abþ), assessed using Ab neuroimaging or cerebrospinal fluid sampling, herald the start of the preclinical stage of Alzheimer’s disease (AD; (Rowe et al., 2014). The finding that Abþ is associated with reduced episodic memory has also raised the possibility that in CN older adults, memory

1232

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

dysfunction may indicate Abþ (Darby et al., 2011; Wagner et al., 2012). However, when compared with normative data, CN older adults with Abþ show little to no impairment on tasks of episodic memory, certainly much less than that required to warrant evaluation for possible mild cognitive impairment (MCI; e.g., z score 1.5 standard deviations [SDs]). Consequently, the extent to which measures of episodic memory, applied in a single assessment, can be used to identify the presence of Abþ in CN older adults is limited. Results of prospective studies have indicated that Abþ is associated reliably with memory decline suggests that objectively defined memory decline, which may provide a more accurate basis for identifying preclinical AD (Doraiswamy et al., 2014; Lim et al., 2014a; Mormino et al., 2014). For example, CN older adults with objectively defined memory decline could be investigated with Ab neuroimaging or cerebrospinal fluid sampling. However, in Abþ CN older adults, memory function also varies as a function of demographic (e.g., premorbid intelligence, cognitive reserve; Duff et al., 2013), genetic (e.g., APOE; Mormino et al., 2014; Naj et al., 2014), and BDNF rs6265 genotype; Lim et al., 2014b), and psychiatric (e.g., anxiety symptoms; Pietrzak et al., 2014) factors. Therefore, any memory decline detected in CN older adults might reflect one or more of these factors in addition to, or even instead of, early amyloidosis. Furthermore, different trajectories of episodic memory change may reflect different underlying neuropathologic processes (Knopman et al., 2013; Nettiksimmons et al., 2013; Wilson et al., 2010). The conclusion that memory decline is associated with AD risk factors comes from studies of CN older adults stratified according to the risk factor of interest and then evaluated for changes in memory over time (Caselli et al., 2009; Doraiswamy et al., 2014; Lim et al., 2012, 2013b; Mormino et al., 2014; Villemagne et al., 2013). Consequently, associations between AD risk factors and memory decline, observed in these studies, do not take into account the potential that heterogeneity in memory trajectories reflects the simultaneous effects of multiple risk factors (Hayden et al., 2009). Thus, to evaluate how memory decline in CN older adults reflects different AD risk factors it would be useful first to characterize predominant trajectories of memory change from a large cohort studied prospectively, and then to examine the extent to which different AD risk factors predict these trajectories. The aims of the present study were: (1) to identify predominant trajectories of episodic memory change in CN older adults; and (2) to characterize the AD risk factors that are associated with these trajectories. To evaluate these aims, we applied latent growth mixture modeling to evaluate the nature and determinants of predominant trajectories of episodic memory change over a 54-month period in a well-characterized cohort of CN older adults enrolled in the Australian Imaging, Biomarkers and Lifestyle (AIBL) Study of Ageing.

medical, psychiatric, and neuropsychological data to confirm the cognitive health and clinical classification of each participant. The study was approved by the institutional research and ethics committees of Austin Health, St. Vincent’s Health, Hollywood Private Hospital, and Edith Cowan University. All participants provided written informed consent (Ellis et al., 2009). 2.2. Positron emission tomography imaging and genotyping Ab imaging with positron emission tomography (PET) was conducted using 11C-Pittsburgh Compound B (PiB), 18F-florbetapir, or 18F-flutemetamol. A 30-minute acquisition was started 40 minutes after injection of PiB, whereas 20-minute acquisitions were performed 50 minutes after injection of florbetapir and 90 minutes after injection of flutemetamol. For PiB, PET standardized uptake value (SUV) data were summed and normalized to the cerebellar cortex SUV, yielding a region-to-cerebellar ratio termed SUV ratio (SUVR). For florbetapir, SUVR was generated using the whole cerebellum as the reference region (Clark et al., 2011), whereas for flutemetamol, the pons was used as the reference region (Vandenberghe et al., 2010). In line with previous studies, SUVR was classified dichotomously as either negative or positive (i.e., Ab or Abþ). For PiB, an SUVR threshold 1.5 was used. For florbetapir and flutemetamol, an SUVR threshold of 1.11 and 0.62 were used, respectively. An 80-mL blood sample was also obtained from each participant, 0.5 mL of which was sent to a clinical pathology laboratory for genotyping. DNA was isolated from whole blood using a QIAamp DNA blood Midi or Maxi kit (Qiagen) according to the manufacturer’s protocol. APOE genotype was determined through TaqMan genotyping assays (Life Technologies) for rs7412 (assay ID: C__904973_10) and rs429358 (assay ID: C__3084793_20). The BDNF polymorphism, rs6265 (Val66Met), was genotyped either via a custom Illumina GoldenGate assay, performed by the Beijing Genomics Institute, or through inclusion of the TaqMan genotyping assay (assay ID: C__11592758_10) in a custom designed OpenArray assay (Life Technologies). All TaqMan and OpenArray assays were performed on a QuantStudio 12K-Flex real-time PCR system (Applied Biosystems). Participants were split by the presence or absence of a Met at amino acid position 66 in the BDNF gene (BDNFMet). 2.3. Anxiety and depressive symptoms Anxiety and depressive symptoms were assessed using the Hospital Anxiety and Depression Scale (HADS; Zigmond and Snaith, 1983). 2.4. Vascular risk factors

2. Methods 2.1. Sample The participants were 333 CN older adults who had undergone Ab neuroimaging as part of the AIBL Study (Ellis et al., 2009; Rowe et al., 2010). Selection into this cohort was controlled to ensure: (1) a wide age distribution from 60 years through to the very elderly; (2) enrollment of approximately 50% with a subjective memory complaint; and (3) enrollment of approximately 30% APOE ε4 carriers (Rowe et al., 2010). Exclusion criteria included: schizophrenia; depression (15-item Geriatric Depression Score 6); Parkinson’s disease; cancer (except basal cell skin carcinoma) within the last 2 years; symptomatic stroke; uncontrolled diabetes; obstructive sleep apnea, previous head injury with >1 hour of posttraumatic amnesia; current regular alcohol use >2 standard drinks per day for women or 4 per day for men. For each assessment, a clinical review panel, blinded to Ab neuroimaging data, considered all available

A count of vascular risk factors was obtained by summing whether respondents met criteria for hypertension (blood pressure 140/90 mm Hg or currently undergoing treatment with an antihypertensive medication), dyslipidemia (fasting serum total cholesterol 6.22 mmol/L, fasting serum triglycerides 2.26 mmol/ L, or currently undergoing treatment with statin or fibrate medications), obesity (BMI >30 kg/m2), smoking (smoked >20 cigarettes per day for over 1 year), diabetes (fasting plasma glucose >7 mmol/ L or currently undergoing treatment with diabetes medication), high homocysteine levels (males >16.2 mmol/L; females >13.6 mmol/L), or chronic kidney disease (estimated glomerular filtration rate <45 mL/min; (Yates et al., 2014). 2.5. Subjective memory complaints Subjective memory complaints were assessed at the 18-month assessment using the Memory Complaint Questionnaire (MAC-Q;

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

Crook et al., 1992), a 6-item scale that asks individuals to report the extent to which they experience memory difficulties in everyday situations (e.g., remembering a telephone number) relative to when he or she was in high school. Scores range from 7 to 35, with scores 25 indicative of clinically significant subjective memory impairment. 2.6. Neuropsychological assessment Comprehensive neuropsychological evaluations were conducted at baseline and 18-, 36-, and 54-month follow-ups. The composite measure of episodic memory used in the AIBL study was the primary outcome measure. This composite measure was derived by averaging the standardized scores on the California Verbal Learning Test, Second Edition delayed recall, Logical Memory delayed recall, and RCFT delayed recall tasks (Harrington et al., 2013). 2.7. Data analysis To examine longitudinal trajectories of episodic memory, we conducted latent growth mixture modeling (LGMM) of composite episodic memory scores using robust full-information maximum likelihood estimation in Mplus version 7.11. These analyses proceeded in 3 steps. First, we compared 1- with 6-class unconditional LGMMs and assessed their relative fit using conventional indices, including the Bayesian Information Criterion, Akaike Information Criterion, entropy, the Lo-Mendell-Rubin likelihood test (LRT), and the bootstrap likelihood ratio test (Nylund et al., 2007). We identified the best-fitting models based on smaller Bayesian Information Criterion and Akaike Information Criterion values, higher entropy values, and on results of the LRT and the bootstrap likelihood ratio test, which quantify the likelihood that the data can be described by a model with one less trajectory. In addition to these formal criteria, we also considered class sizes, parsimony, and theoretical interpretability of the solutions, and to enhance generalizability, aimed to select a final model that contained a meaningful proportion of the sample (i.e., >1%) in the smallest class. Each participant was assigned the class having the greatest posterior probability. For each memory trajectory group, we computed the time to reach a clinically significant level of impairment (i.e., 1.5 SDs relative to the group mean) by interpolating slope means for each group. Second, we computed bivariate relative risk ratios between AD risk factors and memory trajectory groups. Third, AD risk factors that were significant (p < 0.05) in bivariate analyses were entered into a multinomial logistic regression analysis to identify independent predictors of declining memory trajectories; the stable trajectory group was defined as the reference group. Associations between independent AD risk factors were expressed as relative risk ratios and 95% confidence intervals (95% CIs). Individuals who developed MCI or AD over the course of the followup period were retained in analyses. Sensitivity, specificity, and positive likelihood ratio (LRþ) tests were computed for each of the significant risk factors for declining memory trajectories for which

1233

Table 1 Demographic and clinical characteristics of the sample (n ¼ 333) Mean (SD) or n (%) N at baseline N at 18 mo N at 36 mo N at 54 mo Age Sex Male Female Education <9 y 9e12 y 13e15 y >15 y Any vascular risk factor Full-scale IQ MAC-Q score HADS depression score HADS anxiety score APOE ε4 carrier BDNFMet carrier Abþ

333 321 305 296 70.0

(100.0%) (96.4%) (91.6%) (88.9%) (6.8)

Range d d d d 60e89 d

160 (48.0%) 173 (52.0%) d 23 120 63 126 272 108.6 25.3 2.6 4.2 109 123 84

(6.9%) (36.0%) (18.9%) (37.8%) (81.7%) (7.1) (4.5) (2.2) (2.8) (32.7%) (36.9%) (25.2%)

d 85e125 11e35 0e11 0e15 d d d

Key: APOE ε4 carrier, individuals with 1 or 2 copies of the apolipoprotein E epsilon 4 allele; BDNFMet, brain-derived neurotrophic factor Met allele (rs6265) carrier; HADS, Hospital Anxiety and Depression Scale; IQ, intelligence quotient assessed by the Wechsler Test of Adult Reading (WTAR); MAC-Q, Memory Complaints Questionnaire; SD, standard deviation.

clinically meaningful thresholds have been ascertained: Abþ, APOE ε4 carrier, and MAC-Q score A25; of note, these tests were not computed for MMSE scores, as none of the participants in the AIBL healthy older adults had an MMSE score <24, which is commonly used to identify possible cognitive impairment. 3. Results Table 1 shows demographic and clinical characteristics of the sample. Table 2 shows results of LGMM analyses. A 3-class solution was identified as the optimal model, because it fit the data better than the 1- and 2-class solutions, was parsimonious and theoretically defensible and contained >1% of the sample in the smallest class. As shown in Fig. 1, this 3-class model consisted of a stable memory trajectory group characterized by above average memory scores at baseline and stable trajectory of memory scores over time (hereafter termed: “stable memory trajectory group;” n ¼ 218, 65.5%; intercept ¼ 0.23 (standard error [SE] ¼ 0.066, slope ¼ 0.004 [SE ¼ 0.001]), a subtly declining memory trajectory group characterized by low average memory scores at baseline and a subtle decline in memory scores over time (hereafter termed: “subtly declining memory trajectory group;” n ¼ 103, 30.9%; intercept ¼ 0.64 [SE ¼ 0.08], slope ¼ 0.005 [SE ¼ 0.002]) and a rapidly declining memory trajectory group characterized by a lower than average memory at baseline and a rapidly declining trajectory of memory scores over time (hereafter termed: “rapidly declining memory trajectory group;” n ¼ 12, 3.6%; intercept ¼ 0.61 [SE ¼

Table 2 Fit indices for 1- to 6-class unconditional latent growth mixture models of composite episodic memory scores Class

Log-likelihood

BIC

SSA-BIC

AIC

Entropy

LMR adjusted LRT p-value

VLMR p-value

No. in the smallest class (%)

1 2 3 4 5 6

1172.56 1125.34 1112.73 1108.14 1104.17 1109.13

2397.40 2320.38 2312.59 2320.83 2330.35 2357.66

2368.85 2282.32 2265.01 2263.73 2263.71 2281.53

2363.13 2274.68 2255.47 2252.28 2250.35 2266.27

0.964 0.782 0.828 0.772 0.749

0.0001 <0.0001 0.0355 0.5635 0.2121

<0.0001 <0.0001 0.0295 0.5505 0.2121

13 12 2 2 2

(3.9) (3.6) (0.6) (0.6) (0.6)

Key: AIC, Akaike Information Criterion; BIC, Bayesian Information Criterion; LMR LRT, Lo-Mendell-Rubin likelihood ratio test; SSA-BIC, Sample size-adjusted Bayesian Information Criterion.

1234

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

Fig. 1. Trajectories of episodic memory change in cognitively normal older adults.

0.26], slope ¼ 0.048 [SE ¼ 0.005]). Means of quadratic growth factors did not differ from 0 (all p’s > 0.28), so linear functions were used to describe the data. Interpolation of latent slopes for each memory trajectory group indicated that, on average, the subtly declining memory trajectory group would meet criteria for clinically significant memory impairment in 14.3 (95% CI ¼ 8.0e66.4) years after enrollment in the AIBL study. In contrast, the rapidly declining memory trajectory group would on average meet criteria for clinically significant memory impairment in 1.5 (95% CI ¼ 1.3e1.9) years after enrollment. A total of 6 (2.8%) of the stable trajectory group and 17 (16.5%) of the subtly declining trajectory group were diagnosed with MCI over the 54-month study period. All of the older adults in the rapidly declining trajectory group (n ¼ 12) were diagnosed with MCI or dementia over the study period; 6 (50.0%) were diagnosed with AD, 5 (41.7%) with MCI, and 1 (8.3%) with vascular dementia. Rerunning LGMMs excluding older adults who progressed to MCI or dementia revealed that a 2-class solution provided the best fit to the data (2-class vs. 1-class: LRT test ¼ 3.34, p ¼ 0.0006; LMR adjusted LRT test ¼ 24.105, p ¼ 0.0009). The majority of the sample (n ¼ 196; 65.8%) were classified into a stable memory trajectory group characterized by an intercept of 0.28 and slope of 0.004. The remainder of the sample (n ¼ 102; 34.2%) were classified into a group with a subtly declining memory trajectory group, characterized by an intercept of 0.557 and a slope of 0.004. Notably, 196 (92.5%) of the participants classified as having a stable memory trajectory in the revised analysis had been also classified as having stable memory in the initial analysis. Furthermore, all 86 participants (100%) classified as having subtly declining memory trajectory in the revised analysis had also been classified in the subtly declining memory trajectory in the initial analysis (X2(1) ¼ 232.29, p < 0.001). Table 3 shows results of bivariate analyses of how individual risk factors were associated with memory trajectory groups. All the variables assessed except vascular risk factors, HADS depression scores, and BDNFMet genotype were associated with declining memory trajectory group membership. Specifically, when compared with the stable memory trajectory group, the subtly declining trajectory group was older, more likely to be male, less likely to be college educated, had lower full-scale IQ and MMSE scores, and were more likely to be Abþ. Relative to the stable memory trajectory group, the rapidly declining memory group was older, more likely to be male, had lower MMSE and higher MACQ and HADS anxiety scores, and were more likely to be APOE ε4 carriers and Abþ.

Multivariable analyses evaluating the relation between AD risk factors and declining memory trajectories indicated that, relative to the stable trajectory group, Abþ was independently positively associated with membership in both declining memory trajectory groups (Table 4). Full-scale IQ and MMSE scores were additionally associated with membership of the subtly declining memory trajectory group. APOE ε4 genotype and subjective memory impairment were additionally associated with membership in the rapidly declining memory trajectory group. As shown in Table 5, results of sensitivity, specificity, and LRþ tests revealed that Abþ at baseline predicted both subtly and rapidly declining memory trajectories relative to a stable trajectory and a rapidly declining memory trajectory relative to a subtly declining memory trajectory. MAC-Q score 25, APOE ε4 carriage, and Abþ predicted a rapidly declining memory trajectory relative to the stable trajectory and rapidly declining memory trajectory relative to a subtly declining trajectory. 4. Discussion Results of this study suggest that there are 3 predominant trajectories of memory performance in cognitively normal older adults assessed over a 54-month period. Most (65.5%) of the sample showed an above average and stable memory trajectory, whereas 30.9% showed a below average and subtly declining memory trajectory and 3.6% showed a below average and rapidly declining memory trajectory. This finding indicates that statistical modeling of between- and within-individual variability in prospectively assessed memory scores in CN older adults can identify heterogenous and qualitatively different patterns of change in episodic memory in preclinical AD. Notably, these trajectory groups contained proportionally greater numbers of older adults who converted to MCI or AD over the 54-month study period, with 16.5% of the subtly declining and 100% of the rapidly declining groups converting to MCI or AD over the study period, relative to only 2.8% of the stable trajectory group. This finding suggests that a significant proportion of older adults in the subtly and rapidly declining memory trajectory groups were likely in the late preclinical phase of AD. They further suggest that latent growth mixture modeling may be useful in identifying heterogenous subpopulations of older adults who have a high probability of developing clinically significant memory decline in the preclinical phase of AD.

Key: Ab, amyloid beta; APOE, apolipoprotein; BDNF, brain-derived neurotrophic factor; IQ, intelligence quotient; HADS, Hospital Anxiety and Depression Scale; MAC-Q, Memory Complaint Questionnaire; MMSE, Mini-Mental State Examination; RRR, relative risk ratio (95% confidence interval); SE, standard error of the mean.

0.36 1.29 (0.41e4.02) 0.90 (0.31e2.62) 0.46 0.10 0.90 1.72 (0.24e12.35) 0.20 0.40 3.75 (1.20e11.69) 1.58 (0.55e4.58) 3.23 (11.03e10.10)

RRR (95% CI) or Cohen d RRR (95% CI) or Cohen d

1.04 2.69 (0.83e8.69) 0.63 (0.21e1.90) 0.12 0.55 1.05 2.12 (0.28e15.87) 0.31 0.58 3.97 (1.23e12.79) 1.50 (0.50e4.50) 7.58 (2.38e24.39) 0.67 1.68 (1.21e2.33) 0.67 (0.48e0.92) 0.31 0.52 0.23 1.02 (0.42e1.15) 0.19 0.17 0.98 (0.70e1.38) 0.94 (0.66e1.33) 1.73 (1.27e2.37)

RRR (95% CI) or Cohen d

3>1, 2 2, 3>1 1>2 1>2 1>2, 3 3>1, 2 d d 3>1 3>1, 2 d 3>2>1 <0.001 0.003 0.007 0.018 <0.001 0.002 0.67 0.076 0.048 0.038 0.69 <0.001 (1.9) (66.7%) (25.0%) (2.0) (0.3) (1.4) (91.7%) (0.6) (0.8) (66.7%) (50.0%) (66.7%) 74.7 8 3 110.2 28.3 29.2 11 3.3 5.7 8 6 8 (0.6) (60.2%) (26.5%) (0.7) (0.1) (0.4) (82.5%) (0.2) (0.3) (31.1%) (37.2%) (35.0%) 72.5 62 27 107.0 28.4 25.5 85 2.9 4.5 32 35 36

Mean (SE) or n (%)

(0.4) (41.3%) (44.0%) (0.5) (0.1) (0.3) (80.7%) (0.2) (0.2) (31.7%) (39.4%) (18.3%)

Mean (SE) or n (%)

68.5 90 96 109.3 29.1 24.5 176 2.4 4.0 69 82 40 Age Male sex Some collegeþ education Full-scale IQ MMSE score MAC-Q score Any vascular risk factor HADS depression score HADS anxiety score APOE ε4 carrier BDNFMet carrier Abþ

Mean (SE) or n (%)

16.28, 11.75, 9.99, 4.04, 17.32, 6.59, 0.80, 2.59, 3.06, 6.52, 0.74, 21.56,

Omnibus F, p or c2, p Rapidly declining memory trajectory Subtly declining memory trajectory Stable memory trajectory

Table 3 Results of bivariate analyses evaluating baseline correlates of predominant trajectories of episodic memory change

Pairwise contrasts

Subtly declining memory trajectory versus stable memory trajectory

Rapidly declining memory trajectory versus stable memory trajectory

Rapidly declining memory trajectory versus subtly declining memory trajectory

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

1235

As some AD risk factors are themselves interrelated in CN older adults (e.g., Abþ and APOE ε4 allele carriage; Morris et al., 2010; Rowe et al., 2010), multivariable analyses were conducted to identify the extent to which the different AD risk factors were independently related to declining memory trajectories. Compared with the stable memory trajectory group, Abþ, APOE ε4 allele carriage, and subjective memory impairment were all associated with increased risk of a rapidly declining memory trajectory. Of these AD risk factors, the independent effects of both Abþ and APOE ε4 allele carriage were large. For example, older adults who were Abþ and APOE ε4 allele carriers at baseline were 8.3 and 6.1 times more likely than those without these risk factors to be in the rapidly declining memory trajectory group. Furthermore, APOE ε4 allele carriage at baseline was associated with 8.4 times increased likelihood of being in the rapidly declining memory trajectory than subtly declining memory trajectory group, although Ab status at baseline did not differentiate these 2 groups in multivariable analysis. The finding that Abþ and APOE ε4 genotype showed large magnitude associations with a rapidly declining memory trajectory is consistent with results of prior studies that examined the extent to which these same risk factors, when stratified according to their presence or absence, are related to memory decline (Lim et al., 2013a, 2014a; Mormino et al., 2014; Naj et al., 2014). The observation that even after statistical adjustment for Abþ, APOE ε4 genotype remained associated with a rapidly declining memory trajectory indicates that the combination of these AD risk factors rather than the effect of either alone increases risk for rapidly declining memory in preclinical AD. Although the risk for Abþ imposed by APOE ε4 genotype is well known, as is the effect of Abþ on memory decline, there is ongoing debate about whether the APOE ε4 genotype increases disease progression independently of Abþ. For example, data from our own initial study of CN older adults observed no interaction of Abþ or APOE ε4 genotype on memory decline over 18 months (Lim et al., 2012). In contrast, a more recent study of 490 CN older adults did find that Abþ and APOE ε4 genotype interacted to increase the magnitude of memory decline (Mormino et al., 2014). Results of the present study extend this work to suggest that Abþ and APOE ε4 carriage additively increase disease progression, and that APOE ε4 carriage may be useful in predicting rates of memory decline, in preclinical AD. The finding that increased subjective memory impairment at the 18-month assessment was related to a rapidly declining memory trajectory, and that greater subjective memory complaints was associated with increased likelihood of being in the rapidly versus subtly declining memory trajectory group, suggests that CN older adults may have some insight into their rapidly deteriorating memory (Jessen et al., 2010, 2014). Taken together, these results underscore the importance of assessing Abþ, APOE ε4 carrier status, and subjective memory complaints in preclinical AD, as all 3 of these variables were independently associated with a rapidly declining memory trajectory. Abþ and lower full-scale IQ and MMSE scores were independently associated with membership of the subtly declining memory trajectory group. Specifically, Abþ on PET scanning at the baseline evaluation was associated with more than double the likelihood of, being in a subtly declining relative to a stable memory trajectory. This finding suggests that assessment of Ab levels may have prognostic utility in identifying even very subtle memory decline in preclinical AD. Furthermore, the finding that reduced IQ and MMSE scores were associated with increased likelihood of being in the subtly declining memory trajectory group is consistent with the subtly declining memory trajectory group having lower memory scores at baseline compared with the stable memory trajectory group. Thus, this finding likely reflects the subtly declining memory trajectory group having low average memory scores and reduced IQ and MMSE scores at baseline relative to the stable memory

1236

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

Table 4 Results of multinomial logistic regression analysis of baseline determinants of predominant trajectories of episodic memory change

Full-scale IQ MMSE score MACQ score HADS anxiety score APOE ε4 carrier Abþ

Subtly declining memory trajectory versus stable trajectory

Rapidly declining memory trajectory versus stable trajectory

Rapidly declining memory trajectory versus subtly declining trajectory

LRT c2, p 6.72, 0.010 12.78, <0.001 0.01, 0.92 1.89, 0.17 0.58, 0.45 4.18, 0.041

LRT c2, p 0.47, 0.49 0.14, 0.70 5.47, 0.019 0.98, 0.32 5.85, 0.043 6.00, 0.014

LRT c2, p 0.076, 0.783 1.40, 0.24 4.79, 0.029 0.10, 0.75 6.84, 0.009 2.39, 0.12

RRR (95% CI) 0.94 (0.90e0.98)** 0.63 (0.49e0.82)*** 1.01 (0.95e1.07) 1.07 (0.97e1.19) 0.76 (0.40e1.44) 2.09 (1.06e4.12)*

RRR (95% CI) 1.00 (0.88e1.15) 0.79 (0.42e1.49) 1.19 (1.01e1.42)* 1.11 (0.89e1.39) 6.12 (1.06e35.34)* 8.33 (1.51e45.45)*

RRR (95% CI) 1.04 (0.92e1.18) 1.36 (0.70e2.61) 1.21 (1.01e1.46)* 1.03 (0.82e1.30) 8.41 (1.39e51.02)* 3.60 (0.62e20.83)

RRRs and 95% CIs are adjusted for demographic variables. Significant associations are highlighted in bold: *p < 0.05; **p < 0.01; ***p < 0.001. Key: 95% CIs, 95% confidence intervals; Ab, amyloid beta; APOE, apolipoprotein; IQ, intelligence quotient; HADS, Hospital Anxiety and Depression Scale; LRT, likelihood ratio test; MAC-Q, Memory Complaint Questionnaire; RRR, relative risk ratio.

trajectory group who had high average memory scores and higher IQ and MMSE scores at baseline. Nevertheless, this finding suggests that assessment of IQ and global cognition, as well as Abþ, may have utility in predicting subtle memory decline in preclinical AD. The identification of 3 different memory trajectories in the AIBL CN cohort, and strong association between Abþ and declining memory is consistent with results of a prospective study of 1049 CN older adults from the Religious Orders Study whose cognition was followed prospectively for up to 15 years (Hayden et al., 2011). In this study, most (64.6%) of the older adults showed only slow decline (0.04 SD/y) on a standardized measure of global cognition, whereas 27.1% showed a moderate decline (0.19 SD/y) and 8.3% a more rapid decline (0.57 SD/y). Notably, APOE ε4 genotype was associated with rapid decline in this cohort; and in an autopsy subsample analysis, the magnitude of cognitive decline was correlated with amyloid burden (moderate effect size) and tangle density (large effect size). In a smaller study, Darby et al. (2011) found that of 195 CN older adults who completed a 24-month prospective study, 7.7% had a decline in memory that was below the lower bound 95% CI for the group. In this memory decliner group (n ¼ 15), PET imaging identified 33.3% as Abþ compared with 6.1% (n ¼ 2) of a matched-control group recruited from those with no evidence of memory decline. Limitations of this study must be noted. First, the AIBL CN older adult cohort is a convenience sample, and thus, results of this study may not be representative of the general population. Second, the sample size and follow-up period were modest, and the sample comprised larger numbers of APOE ε4 carriers than would be found in general population samples. Thus, further studies of larger general population-based samples of CN older adults followed for longer periods of time are needed to evaluate the generalizability of these results. Third, given the relatively high proportion of the rapidly and subtly declining memory trajectory groups who converted to MCI or AD over the 54-month study period, the study sample comprised older adults in both the early and late preclinical phases of AD. Thus, additional studies that assess the nature and determinants of memory trajectories in earlier stages of preclinical

AD (i.e., younger adult cohorts), as well as older adults with MCI, will be useful in identifying predominant memory trajectories in different phases of preclinical AD, as well as common and unique risk factors associated with these trajectories. Fourth, other factors not assessed in the AIBL study (e.g., levels of tau proteins and other genes associated with risk of AD, such as CR-1) may additionally predict subtly and rapid declining episodic memory in preclinical AD (Desikan et al., 2012; Naj et al., 2014). Additional studies that incorporate these variables will be useful in evaluating their utility in predicting declining trajectories of episodic memory. 5. Conclusions To our knowledge, this study is among the first to evaluate the nature and determinants of predominant trajectories of episodic memory in preclinical AD. Results indicated that episodic memory change over a 54-month period in cognitively normal older adults was best characterized by 3 trajectories that were differentiated by baseline episodic memory change, as well as magnitude of decline in memory performance over time. Baseline Ab levels emerged as the strongest predictor of both subtly and rapidly declining memory trajectories, thereby underscoring the importance of assessment of Ab levels in predicting AD risk in the preclinical phase of this disorder. Further research in larger population-based samples of CN older adults that are followed for longer periods of time will be useful in evaluating the generalizability of the findings reported herein, and in identifying additional demographic and clinical variables that can inform models of risk for varying rates of episodic memory decline and risk for MCI and/or AD in preclinical AD, as well as in recruiting individuals for clinical trials of memoryenhancing interventions. Disclosure statement Yen Ying Lim, Karra Harrington, and Colin L. Masters report no disclosures. Robert H. Pietrzak is a scientific consultant to Cogstate Ltd. Paul Maruff is a full-time employee of Cogstate Ltd. David Ames

Table 5 Sensitivity, specificity, and positive likelihood ratios for risk factors of declining episodic memory trajectories with established clinical definitions

MAC-Q score 25 APOE ε4 carrier Abþ

Subtly declining memory trajectory versus stable trajectory

Rapidly declining memory trajectory versus stable trajectory

Rapidly declining memory trajectory versus subtly declining trajectory

Sensitivity

Specificity

LRþ (95% CI)

Sensitivity

Specificity

LRþ (95% CI)

Sensitivity

Specificity

LRþ (95% CI)

0.582 0.311 0.350

0.407 0.683 0.817

0.98 (0.80e1.21) 0.98 (0.69e1.39) 1.90 (1.30e2.80)

1.00 0.667 0.667

0.407 0.683 0.817

1.69 (1.36e1.91) 2.11 (1.35e3.29) 3.63 (2.23e5.92)

1.00 0.667 0.667

0.418 0.689 0.650

1.72 (1.32e2.04) 2.15 (1.31e3.51) 1.91 (1.18e3.08)

Significant LRþ values are highlighted in bold. Key: Abþ, positive amyloid beta scan; APOE, apolipoprotein; HADS, Hospital Anxiety and Depression Scale; LRþ, positive likelihood ratio; MAC-Q, Memory Complaint Questionnaire.

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

has served on scientific advisory boards for Novartis, Eli Lilly, Janssen, and Pfizer Inc; has received funding for travel from Janssen and Pfizer Inc; has received speaker honoraria from Pfizer Inc and Lundbeck Inc; and has received research support from Eli Lilly and Company, GlaxoSmithKline, Forest Laboratories Inc, and Novartis. Ralph N. Martins is a consultant to Alzhyme. Christopher C. Rowe serves on scientific advisory boards for Bayer Schering Pharma, Elan Corporation, and AstraZeneca; has received speaker honoraria from Bayer Schering Pharma; and receives research support from Bayer Schering Pharma and Avid Radiopharmaceuticals. Victor L. Villemagne served as a consultant for Bayer Pharma and received research support from an NEDO grant from Japan. Acknowledgements Funding for the study was provided in part by the study partners [Commonwealth Scientific Industrial Research Organization (CSIRO), Edith Cowan University (ECU), Mental Health Research Institute (MHRI), National Ageing Research Institute (NARI), Austin Health, and CogState Ltd]. The study also received support from the National Health and Medical Research Council (NHMRC) and the Dementia Collaborative Research Centres program (DCRC2), as well as funding from the Science and Industry Endowment Fund (SIEF) and the Cooperative Research Centre for Mental Health (CRCMH). Robert H. Pietrzak, Yen Ying Lim, and Paul Maruff conceptualized the study design, reviewed the literature, and wrote the first draft of the article, and conducted data analyses; David Ames, Ralph N. Martins, Colin L. Masters, Paul Maruff, Victor L. Villemagne, Alan Rembach, Christopher C. Rowe are senior investigators of the AIBL study and responsible for the design of the AIBL study and selection of study endpoints. Victor L. Villemagne and Christopher C. Rowe conducted and oversaw neuroimaging for all participants. Yen Ying Lim and Karra Harrington conducted neuropsychological assessments. Simon M. Laws undertook genetic analyses for all participants. David Ames, Karra Harrington, Carolina Restrepo, Ralph N. Martins, Alan Rembach, Simon M. Laws, Colin L. Masters, Victor L. Villemagne, and Christopher C. Rowe provided critical revision of article drafts. References Caselli, R.J., Dueck, A.C., Osborne, D., Sabbagh, M.N., Connor, D.J., Ahern, G.L., Baxter, L.C., Rapcsak, S.Z., Shi, J., Woodruff, B.K., Locke, D.E.C., Snyder, C.H., Alexander, G.E., Rademakers, R., Reiman, E.M., 2009. Longitudinal modeling of age-related memory decline and the APOE ε4 effect. N. Engl. J. Med. 361, 255e263. Clark, C.M., Schneider, J.A., Bedell, B.J., Beach, T.G., Bilker, W.B., Mintun, M.A., Pontecorvo, M.J., Hefti, F., Carpenter, A.P., Flitter, M.L., Krautkramer, M.J., Kung, H.F., Coleman, R.E., Doraiswamy, P.M., Fleisher, A.S., Sabbagh, M.N., Sadowsky, C.H., Reiman, E.P., Zehntner, S.P., Skovronsky, D.M., AV45-A07 Study Group, 2011. Use of florbetapir-PET for imaging beta-amyloid pathology. J. Am. Med. Assoc. 305, 275e283. Crook 3rd, T.H., Feher, E.P., Larrabee, G.J., 1992. Assessment of memory complaint in age-associated memory impairment: the MAC-Q. Int. Psychogeriatr. 4, 165e176. Darby, D.G., Brodtmann, A., Pietrzak, R.H., Fredrickson, J., Woodward, M., Villemagne, V.L., Fredrickson, A., Maruff, P., Rowe, C., 2011. Episodic memory decline predicts cortical amyloid status in community-dwelling older adults. J. Alzheimers Dis. 27, 627e637. Desikan, R.S., McEvoy, L.K., Thompson, W.K., Holland, D., Brewer, J.B., Aisen, P.S., Sperling, R.A., Dale, A.M., Alzheimer’s Disease Neuroimaging Initiative, 2012. Amyloid-b-associated clinical decline occurs only in the presence of elevated ptau. Arch. Neurol. 69, 709e713. Doraiswamy, P.M., Sperling, R.A., Johnson, K., Reiman, E.M., Wong, T.Z., Sabbagh, M.N., Sadowsky, C.H., Fleisher, A.S., Carpenter, A., Joshi, A.D., Lu, M., Grundman, M., Mintun, M.A., Skovronsky, D.M., Pontecorvo, M.J., AV45-A11 Study Group, 2014. Florbetapir F 18 amyloid PET and 36-month cognitive decline: a prospective multicenter study. Mol. Psychiatry 19, 1044e1051. Duff, K., Foster, N.L., Dennett, K., Hammers, D.B., Zollinger, L.V., Christian, P.E., Butterfield, R.I., Beardmore, B.E., Wang, A.Y., Morton, K.A., Hoffman, J.M., 2013. Amyloid deposition and cognition in older adults: the effects of premorbid intellect. Arch. Clin. Neuropsychol. 28, 665e671.

1237

Ellis, K.A., Bush, A.I., Darby, D., De Fazio, D., Foster, J., Hudson, P., Lautenschlager, N.T., Lenzo, N., Martins, R.N., Maruff, P., Masters, C., Milner, A., Pike, K., Rowe, C., Savage, G., Szoeke, C., Taddei, K., Villemagne, V., Woodward, M., Ames, D., AIBL Research Group, 2009. The Australian Imaging, Biomarkers and Lifestyle (AIBL) study of aging: methodology and baseline characteristics of 1112 individuals recruited for a longitudinal study of Alzheimer’s disease. Int. Psychogeriatrics 21, 672e687. Harrington, K.D., Lim, Y.Y., Ellis, K.A., Copolov, C., Darby, D., Weinborn, M., Ames, D., Martins, R.N., Savage, G., Szoeke, C., Rowe, C., Villemagne, V.L., Masters, C.L., Maruff, P., 2013. The association of Abeta amyloid and composite cognitive measures in healthy older adults and MCI. Int. Psychogeriatr. 25, 1667e1677. Hayden, K.M., Reed, B.R., Manly, J.J., Tommet, D., Pietrzak, R.H., Chelune, G.J., Yang, F.M., Revell, A.J., Bennett, D.A., Jones, R.N., 2011. Cognitive decline in the elderly: an analysis of population heterogeneity. Age and Ageing 40, 684e689. Hayden, K.M., Zandi, P.P., West, N.A., Tschanz, J.T., Norton, M.C., Corcoran, C., Breitner, J.C.S., Welsh-Bohmer, K.A., Cache County Study Group, 2009. Effects of family history and APOE e4 status on cognitive decline in the absence of AD: the Cache County Study. Arch. Neurol. 66, 1378e1383. Jessen, F., Wiese, B., Bachmann, C., Eifflaender-Gorfer, S., Haller, F., Kolsch, H., Luck, T., Mosch, E., van den Bussche, H., Wagner, M., Wollny, A., Zimmermann, T., Pentzek, M., Riedel-Heller, S.G., Romberg, H.P., Weyerer, S., Kaduszkiewicz, H., Maier, W., Bickel, H., 2010. Prediction of dementia by subjective memory impairment: effects of severity and temporal association with cognitive impairment. Arch. Gen. Psychiatry 67, 414e422. Jessen, F., Wolfsgruber, S., Wiese, B., Bickel, H., Mosch, E., Kaduszkiewicz, H., Pentzek, M., Riedel-Heller, S.G., Luck, T., Fuchs, A., Weyerer, S., Werle, J., van den Bussche, H., Scherer, M., Maier, W., Wagner, M., 2014. AD dementia risk in late MCI, in early MCI, and in subjective memory impairment. Alzheimers Dement. 10, 76e83. Knopman, D.S., Jack, C.R., Wiste, H.J., Weigand, S.D., Vemuri, P., Lowe, V.J., Kantarci, K., Gunter, J.L., Senjem, M.L., Mielke, M.M., Roberts, R.O., Boeve, B.F., Petersen, R.C., 2013. Brain injury biomarkers are not dependent on b-amyloid in normal elderly. Ann. Neurol. 73, 472e480. Lim, Y.Y., Ellis, K.A., Harrington, K., Kamer, A., Pietrzak, R.H., Bush, A.I., Darby, D., Martins, R.N., Masters, C.L., Rowe, C.C., Savage, G., Szoeke, C., Villemagne, V.L., Ames, D., Maruff, P., 2013a. Cognitive consequences of high Abeta amyloid in mild cognitive impairment and healthy older adults: implications for early detection of Alzheimer’s disease. Neuropsychology 27, 322e332. Lim, Y.Y., Ellis, K.A., Pietrzak, R.H., Ames, D., Darby, D., Harrington, K., Martins, R.N., Masters, C.L., Rowe, C., Savage, G., Szoeke, C., Villemagne, V.L., Maruff, P., 2012. Stronger effect of amyloid load than APOE genotype on cognitive decline in healthy older adults. Neurology 79, 1645e1652. Lim, Y.Y., Maruff, P., Pietrzak, R.H., Ames, D., Ellis, K.A., Harrington, K., Lautenschlager, N.T., Szoeke, C., Martins, R.N., Masters, C.L., Villemagne, V.L., Rowe, C.C., AIBL R.G., 2014a. Effect of amyloid on memory and non-memory decline from preclinical to clinical Alzheimer’s disease. Brain 137, 221e231. Lim, Y.Y., Pietrzak, R.H., Ellis, K.A., Jaeger, J., Harrington, K., Ashwood, T., Szoeke, C., Martins, R.N., Bush, A.I., Masters, C.L., Rowe, C.C., Villemagne, V.L., Ames, D., Darby, D., Maruff, P., 2013b. Rapid decline in episodic memory in healthy older adults with high amyloid-beta. J. Alzheimers Dis. 33, 675e679. Lim, Y.Y., Villemagne, V., Laws, S.M., Pietrzak, R.H., Snyder, P.J., Ames, D., Ellis, K.A., Harrington, K., Rembach, A., Martins, R.N., Rowe, C.C., Masters, C.L., Maruff, P., 2014b. APOE and BDNF polymorphisms moderate amyloid b-related cognitive decline in preclinical Alzheimer’s disease. Mol. Psychiatry. In press. Mormino, E.C., Betensky, R.A., Hedden, T., Schultz, A.P., Ward, A., Huijbers, W., Rentz, D.M., Johnson, K.A., Sperling, R.A., Alzheimer’s Disease Neuroimaging Initiative, Australian Imaging Biomarkers and Lifestyle Flagship Study of Ageing, Harvard Aging Brain Study, 2014. Amyloid and APOE E4 interact to influence short-term decline in preclinical Alzheimer’s disease. Neurology 82, 1760e1767. Morris, J.C., Roe, C.M., Xiong, C.J., Fagan, A.M., Goate, A.M., Holtzman, D.M., Mintun, M.A., 2010. APOE predicts Ab but not tau Alzheimer’s pathology in cognitively normal aging. Ann. Neurol. 67, 122e131. Naj, A.C., Jun, G., Reitz, C., Kunkle, B.W., Perry, W., Park, Y.S., Beecham, G.W., Rajbhandary, R.A., Hamilton-Nelson, K.L., Wang, L.S., Kauwe, J.S., Huentelman, M.J., Myers, A.J., Bird, T.D., Boeve, B.F., Baldwin, C.T., Jarvik, G.P., Crane, P.K., Rogaeva, E., Barmada, M.M., Demirci, F.Y., Cruchaga, C., Kramer, P.L., Ertekin-Taner, N., Hardy, J., Graff-Radford, N.R., Green, R.C., Larson, E.B., St George-Hyslop, P.H., Buxbaum, J.D., Evans, D.A., Schneider, J.A., Lunetta, K.L., Kamboh, M.I., Saykin, A.J., Reiman, E.M., De Jager, P.L., Bennett, D.A., Morris, J.C., Montine, T.J., Goate, A.M., Blacker, D., Tsuang, D.W., Hakonarson, H., Kukull, W.A., Foroud, T.M., Martin, E.R., Haines, J.L., Mayeux, R.P., Farrer, L.A., Schellenberg, G.D., Pericak-Vance, M.A., 2014. Effects of multiple genetic loci on age at onset in late-onset Alzheimer disease: a genomewide association study. JAMA Neurol. 71, 1394e1404. Nettiksimmons, J., Beckett, L., Schwarz, C., Carmichael, O., Fletcher, E., Decarli, C., 2013. Subgroup of ADNI normal controls characterized by atrophy and cognitive decline associated with vascular damage. Psychol. Aging 28, 191e201. Nylund, K.L., Asparouhov, T., Muthen, B., 2007. Deciding on the number of classes in latent class analysis and growth mixture modeling. A Monte Carlo simulation study. Struct. Equation Model. 14, 535e569. Pietrzak, R.H., Scott, J.C., Neumeister, A., Lim, Y.Y., Ames, D., Ellis, K.A., Harrington, K., Lautenschlager, N.T., Szoeke, C., Martins, R.N., Masters, C.L., Villemagne, V.L., Rowe, C.C., Maruff, P., Australian Imaging, Biomarkers and Lifestyle (AIBL) Research Group, 2014. Anxiety symptoms, cerebral amyloid burden and memory decline in healthy older adults without dementia: 3-year prospective cohort study. Br. J. Psychiatry 204, 400e401.

1238

R.H. Pietrzak et al. / Neurobiology of Aging 36 (2015) 1231e1238

Rowe, C.C., Bourgeat, P., Ellis, K.A., Brown, B., Lim, Y.Y., Mulligan, R., Jones, G., Maruff, P., Woodward, M., Price, R., Robins, P., Tochon-Danguy, H., O’Keefe, G., Pike, K.E., Szoeke, C., Salvado, O., Macaulay, S.L., O’Meara, T., Head, R., Cobiac, L., Martins, R., Masters, C.L., Ames, D., Villemagne, V.L., Australian Imaging, Biomarkers and Lifestyle (AIBL) Research Group, 2014. Predicting Alzheimer disease with b-amyloid imaging: results from the Australian Imaging, Biomarkers, and Lifestyle Study of Ageing. Ann. Neurol. 74, 905e913. Rowe, C.C., Ellis, K.A., Rimajova, M., Bourgeat, P., Pike, K.E., Jones, G., Fripp, J., Tochon-Danguy, H., Morandeau, L., O’Keefe, G., Price, R., Raniga, P., Robins, P., Acosta, O., Lenzo, N., Szoeke, C., Salvado, O., Head, R., Martins, R., Masters, C.L., Ames, D., Villemagne, V., 2010. Amyloid imaging results from the Australian Imaging, Biomarkers and Lifestyle (AIBL) Study of Aging. Neurobiol. Aging 31, 1275e1283. Vandenberghe, R., Van Laere, K., Ivanoiu, A., Salmon, E., Bastin, C., Triau, E., Hasselbalch, S., Law, I., Andersen, A., Korner, A., Minthon, L., Garraux, G., Nelissen, N., Bormans, G., Buckley, C., Owenius, R., Thurfjell, L., Farrar, G., Brooks, D.J., 2010. 18F-flutemetamol amyloid imaging in Alzheimer disease and mild cognitive impairment: a phase 2 trial. Ann. Neurol. 68, 319e329.

Villemagne, V.L., Burnham, S., Bourgeat, P., Brown, B., Ellis, K.A., Salvado, O., Szoeke, C., Macaulay, S.L., Martins, R., Maruff, P., Ames, D., Rowe, C.C., Masters, C.L., Group, T.A.R., 2013. Amyloid b deposition, neurodegeneration and cognitive decline in sporadic Alzheimer’s disease: a prospective cohort study. Lancet Neurol. 12, 357e367. Wagner, M., Wolf, S., Reischies, F.M., Daerr, M., Wolfsgruber, S., Jessen, F., Popp, J., Maier, W., Hüll, M., Frölich, L., Hampel, H., Perneczky, R., Peters, O., Jahn, H., Luckhaus, C., Gertz, H.J., Schröder, J., Pantel, J., Lewczuk, P., Kornhuber, J., Wiltfang, J., 2012. Biomarker validation of a cued recall memory deficit in prodromal Alzheimer disease. Neurology 78, 379e386. Wilson, R.S., Leurgans, S.E., Boyle, P.A., Schneider, J.A., Bennett, D.A., 2010. Neurodegenerative basis of age-related cognitive decline. Neurology 75, 1070e1078. Yates, P.A., Desmond, P.M., Phal, P.M., Steward, C., Szoeke, C., Salvado, O., Ellis, K.A., Martins, R.N., Masters, C.L., Ames, D., Villemagne, V.L., Rowe, C.C., 2014. Incidence of cerebral microbleeds in preclinical Alzheimer disease. Neurology 82, 1266e1273. Zigmond, A.S., Snaith, R.P., 1983. The hospital anxiety and depression scale. Acta Psychiatrica Scand. 67, 361e370.