Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 Sep 1.
Published in final edited form as: Neurobiol Aging. 2020 Mar 7;93:124–130. doi: 10.1016/j.neurobiolaging.2020.03.002

Amyloid-beta burden predicts prospective decline in body mass index in clinically normal adults

Jennifer S Rabin a,b,c,d, Zahra Shirzadi a,e, Walter Swardfager a,f, Bradley J MacIntosh a,e, Aaron Schultz g,h, Hyun-Sik Yang g,i, Rachel F Buckley g,i,j, Jennifer R Gatchel d, Dylan Kirn g, Jeremy J Pruzin g,i, Trey Hedden h,k, Nir Lipsman a,b,l, Dorene M Rentz g,i, Sandra E Black a,c, Keith A Johnson h,i,m, Reisa A Sperling g,h,i, Jasmeer P Chhatwal g,h,i,*; Harvard Aging Brain Study Alzheimer’s Disease Neuroimaging Initiative
PMCID: PMC7429980  NIHMSID: NIHMS1581693  PMID: 32249013

Abstract

In the present study, we tested the hypothesis that higher amyloid-beta (Aβ) burden at baseline is associated with greater longitudinal decline in body mass index (BMI) in clinically normal adults. Participants from the Harvard Aging Brain Study (n = 312) and the Alzheimer’s Disease Neuroimaging Initiative (n = 336) underwent Aβ positron emission tomography at baseline. BMI was assessed longitudinally over a median of >4 years. Linear mixed models showed that higher baseline Aβ burden was significantly associated with greater decline in BMI in both the Harvard Aging Brain Study (t = −1.93; p = 0.05) and Alzheimer’s Disease Neuroimaging Initiative cohorts (t = −2.54; p = 0.01), after adjusting for covariates, including cognitive performance and depressive symptoms. In addition, the association of Aβ burden with longitudinal decline in BMI persisted in both cohorts after excluding participants with diabetes/endocrine disturbances and participants classified as underweight or obese (BMI <18.5 or >30). These findings suggest that decline in BMI in clinically normal adults may be an early manifestation related to cerebral amyloidosis that precedes objective cognitive impairment. Therefore, unintentional BMI decline in otherwise healthy individuals might alert clinicians to increased risk of Alzheimer’s disease.

Keywords: ADNI, Amyloid-beta, Body mass index, HABS, Longitudinal, Preclinical Alzheimer’s disease, Weight loss

1. Introduction

Studies suggest that higher body mass index (BMI) during midlife is associated with an increased risk of dementia, including Alzheimer’s disease (AD) (Fitzpatrick et al., 2009; Kivipelto et al., 2005; Whitmer et al., 2005). However, the inverse relationship has also been reported in late-life, whereby lower BMI at older ages is associated with an increased risk of cognitive decline (Besser et al., 2014; Corrada et al., 2006; Fitzpatrick et al., 2009; Hughes et al., 2009; Johnson et al., 2006). Given that weight loss may co-occur with the development of AD pathology before the diagnosis of dementia, a low BMI in late-life may be a consequence of the disease (Buchman et al., 2005; Johnson et al., 2006; Kivimaki et al., 2018; Mathys et al., 2017).

AD has a long “preclinical” phase defined by elevated amyloid-beta (Aβ) in the absence of significant cognitive symptoms (Jack et al., 2018; Sperling et al., 2014, 2011). Cross-sectional studies using positron emission tomography (PET) or cerebrospinal fluid have found associations between AD pathology (Aβ and/or tau) with lower BMI in clinically normal adults (Hsu et al., 2016; Thirunavu et al., 2019; Vidoni et al., 2011). However, longitudinal studies are needed to understand the temporal nature of the association, that is, whether decreases in BMI represent an early manifestation of AD pathology. To our knowledge, only one longitudinal study has examined this question in clinically normal adults and found no association between AD pathology and prospective BMI change, perhaps due to the modest sample size (n = 100) and/or limited follow-up data (up to 2 years; Vidoni et al., 2011). Larger sample sizes and longer follow-up periods may be needed to detect associations in the preclinical phase of AD.

In the present study, we tested the hypothesis that higher Aβ burden at baseline would be associated with greater longitudinal decline in BMI in individuals who were clinically normal at baseline. To do so, we leveraged Aβ-PET and longitudinal BMI data (1–8 years of longitudinal follow-up; median of more than 4 years) from 2 large cohorts of clinically normal individuals: the Harvard Aging Brain Study (HABS) and the Alzheimer’s Disease Neuroimaging Initiative (ADNI).

2. Materials and methods

2.1. Participants

Data were obtained from 2 independent cohorts: HABS and ADNI. For HABS, study protocols were approved by the Partners HealthCare Institutional Review Board. All participants in HABS provided written informed consent before study procedures. For ADNI, the institutional review boards of all participating sites provided approval for the study. All participants in ADNI provided written informed consent.

Inclusion criteria for both studies have been published previously (Aisen et al., 2010; Dagley et al., 2017). To be included in the present study, participants were required to be classified as clinically normal at the time of their first available Aβ-PET scan, which was considered baseline in this study. Clinically normal was defined by a global Clinical Dementia Rating of 0 (Morris, 1993), Mini-Mental State Examination score ≥27 with educational adjustment (Folstein et al., 1975), and performance within education-adjusted norms on Logical Memory Delayed Recall (Wechsler, 1987). In the ADNI sample, the subjective memory complaint group was also considered clinically normal, given that they met the aforementioned requirements. We required that participants contribute at least 2 BMI data points to the analysis (baseline and at least one follow-up data point), given that the present study focused on change in BMI. We excluded BMI values that changed more than 5 standard deviations from baseline to avoid possible outlier effects or data recording errors. This value was chosen based on a visual inspection of the distributions of BMI change values; large gaps in the distribution were identified around 5 standard deviations above and below the mean in both HABS and ADNI. This resulted in exclusion of <1% of the data points in each cohort (11 longitudinal data points in HABS and 4 longitudinal data points in ADNI).

2.2. Aβ positron emission tomography

Detailed Aβ-PET protocols have been previously described for both HABS and ADNI (Johnson et al., 2016; Landau et al., 2012). Different Aβ tracers were used in each cohort. HABS used 11C–Pittsburgh Compound B (PiB), and ADNI used 18F–florbetapir. For consistency purposes, we used non–partial volume corrected Aβ PET data for all analyses, given that partial volume corrected values were not publicly available in ADNI. Because analyses were carried out in each cohort independently, we used the cortical Aβ-PET composite that has been routinely used in each cohort. In HABS, PiB-PET measurements were represented as a distribution volume ratio across a composite of frontal, lateral temporal and parietal, and medial parietal regions, with cerebellar gray matter serving as the reference region. In ADNI, florbetapir measurements were represented as a cortical summary standard uptake value ratio across a composite of lateral and medial frontal, anterior, and posterior cingulate, lateral parietal, and lateral temporal regions. The whole cerebellum served as the reference region in ADNI. The standard uptake value ratio values for ADNI were accessed and downloaded from the LONI website (http://adni.loni.usc.edu/) in March 2019.

2.3. Clinical and cognitive assessments

In HABS, each participant had their height and weight measured annually. In ADNI, height and weight were measured at study entry and weight was measured at subsequent follow-up visits. These values were converted to metric units to calculate BMI (kilograms divided by meters squared).

Cognition was assessed longitudinally in both cohorts using the Preclinical Alzheimer Cognitive Composite-5 or PACC-5 (Donohue et al., 2014; Papp et al., 2017). This composite was developed to be sensitive to early cognitive changes in AD (Donohue et al., 2014). For HABS, the PACC-5 was calculated using the Mini-Mental State Examination, Logical Memory Delayed Recall, Digit Symbol Coding Test, Free and Cued Selective Reminding Test (Free plus Cued Recall), and Animal Fluency. For ADNI, the PACC-5 was calculated using the Mini-Mental State Examination, Logical Memory Delayed Recall, Trail-Making Test B, the Delayed Word Recall from the Alzheimer’s Disease Assessment Scale–Cognitive Subscale, and Animal Fluency. To calculate the PACC-5 composite, z scores of the individual components were computed for each time point and then averaged together. This average was then standardized to the mean and standard deviation of the baseline sample within each cohort.

Depressive symptoms were measured at baseline and each follow-up visit in HABS and ADNI using the Geriatric Depression Scale (GDS). In HABS, depressive symptoms were measured with a full-length, 30-item questionnaire (Yesavage et al., 1982) and in ADNI they were measured with an abbreviated, 15-item questionnaire (Sheikh and Yesavage, 1986). To harmonize across the 2 cohorts, we calculated a 15-item GDS score for HABS participants and used this score in all analyses.

Metabolic and endocrine conditions that could influence BMI were identified (Bays et al., 2007). For HABS, we relied on self-reported diabetes or the use of diabetes medication. In ADNI, information on diabetes was not specifically collected, and therefore, we relied on self-reported endocrine dysfunction (yes or no), which was collected in the medical history questionnaire. In both cohorts, systolic blood pressure was measured in participants’ dominant arm in a seated position.

2.4. Apolipoprotein E genotyping

A blood sample was collected in each study for genotyping of apolipoprotein E (APOE). Heterozygotes and homozygotes for the ε4 allele were collapsed into a single category. Data on APOE were available for 300/312 HABS participants and all 336 ADNI participants.

2.5. Statistical analyses

Statistical analyses were performed using R, version 3.2.4. T-tests and χ2 tests were used to determine differences in demographic and clinical variables between HABS and ADNI. We ran a series of linear mixed models to investigate the association of baseline Aβ burden with BMI change over time. In all analyses, longitudinal BMI values served as the outcome variable and the interaction of baseline Aβ burden with time served as the independent variable of interest. A random intercept for BMI and rate of change over time (slope) were included as model parameters for each participant. Time was operationalized as years from baseline. Baseline age and sex, and their interactions with time, were included as covariates. Because poorer cognition and depressive symptoms might lead to reduced BMI in late life (McMinn et al., 2011; Morley and Kraenzle, 1994), we further adjusted for cognitive performance and depressive symptoms at each visit. In addition, given known associations between blood pressure and BMI (Domonkos Tarnoki et al., 2013), we also adjusted for systolic blood pressure at each visit.

In post hoc analyses, we examined whether APOE ε4 status moderated the association between Aβ burden with BMI change because previous studies have suggested a stronger relationship in APOE ε4 carriers compared to noncarriers (Blautzik et al., 2018; Hsu et al., 2016). We also tested whether the association of Aβ burden with BMI change was driven by metabolic or endocrine disturbances. To do so, we repeated the primary analysis after excluding participants with diabetes in HABS or endocrine dysfunction in ADNI as well as participants who were underweight (BMI <18.5) or obese (BMI >30). This allowed us to test whether the association of Aβ burden with BMI change persisted in a normal or close-to-normal BMI sample without diabetes or endocrine disturbances, after adjusting for the aforementioned covariates (i.e., age, sex, cognitive performance, depressive symptoms, and systolic blood pressure).

All tests were 2-tailed and performed separately in HABS and ADNI. Aβ burden and BMI were considered as continuous variables in all analyses. Age was centered at the mean in each cohort; 72.2 years in HABS and 74.6 years in ADNI, and females were the reference group in both cohorts.

3. Results

3.1. Cohort differences in demographics

Table 1 summarizes the baseline characteristics of the participants included in the present study and highlights the demographic and clinical differences between HABS and ADNI. At baseline, relative to ADNI, HABS participants were significantly younger, had fewer years of education, had a higher proportion of females to males, had a lower proportion of Aβ-positive individuals, and scored higher on the GDS short form (i.e., endorsed a greater number of depressive symptoms). On average, HABS participants were followed for approximately 6 months longer than ADNI participants (HABS: 5.0 years, ADNI: 4.3 years) and had one additional visit (HABS: 6 visits; ADNI: 5 visits). There were no significant cohort differences in baseline BMI or the proportion of APOE ε4 carriers.

Table 1.

Baseline demographic and clinical characteristics by cohort

Variables HABS (n = 312) ADNI (n = 336) p value
Age in years, mean (SD) 72.23 (7.46) 74.62 (6.46) <0.001
Education in years, mean (SD) 15.86 (2.96) 16.48 (2.60) 0.005
Females, n (%) 190 (61) 177 (53) 0.05
APOE ε4 carriers, n (%) 88 (29)a 91 (27) 0.67
Aβ-PET DVR/SUVr, mean (SD) 1.17 (0.19) 1.12 (0.18) –
Aβ+, n (%) 77 (25) 111 (33) 0.02
BMI, mean (SD) 26.81 (4.50) 27.43 (5.04) 0.10
Weight, kg, mean (SD) 76.2 (15.27) 77.2 (15.76) 0.40
Height, mean (SD) 66.23 (3.91) 66.03 (3.99) 0.52
Underweight (BMI <18.5), n (%) 7 (2) 1 (0.3) 0.03
Obese (BMI >30), n (%) 77 (25) 81 (27) 0.94
GDS (short form) 1.23 (1.34) 0.88 (1.15) <0.001
Diabetes (HABS) or endocrine dysfunction (ADNI), n (%) 32 (10)b 143 (43)c –
Years of follow-up after baseline, median (SD) 5.00 (2.04) 4.34 (1.96) <0.001
Total number of visits, median (SD) 6.00 (1.94) 5.00 (1.36) <0.001
Open in a new tab

We did not directly compare the mean Aβ PET values from the 2 cohorts given that different tracers were used.

Key: ADNI, Alzheimer’s Disease Neuroimaging Initiative; APOE ε4, apolipoprotein E ε4; Aβ, amyloid-beta; BMI, body mass index; DVR, distribution volume ratio; GDS, Geriatric Depression Scale; HABS, Harvard Aging Brain Study; PET, positron emission tomography; SD, standard deviation; SUVr, standard uptake value ratio.

a

Missing APOE ε4 data on 12 participants in HABS.

b

Missing diabetes information on 13 participants in HABS.

c

Missing endocrine dysfunction information on 4 participants in ADNI.

3.2. Higher Aβ burden is associated with faster BMI decline

Higher Aβ burden at baseline was associated with greater decline in BMI over time in both cohorts (Table 2 and Fig. 1; β = −0.20 and −0.33 for HABS and ADNI, respectively; these β values are not directly comparable because different Aβ-PET tracers were used in each cohort). With respect to covariates, in HABS, older age at baseline was significantly associated with faster decline in BMI (baseline age-by-time interaction, p = 0.004). In ADNI, there was an effect of older age with lower baseline BMI (p < 0.001); however, a longitudinal association between older age and decreasing BMI was not observed (baseline age-by-time interaction, p = 0.41). In HABS, cognitive performance was not associated with BMI (p = 0.85), although a significant positive association was observed in ADNI (p = 0.05). In HABS, there was a significant negative association between depressive symptoms and BMI (p = 0.002), whereas no significant association was observed in ADNI (p = 0.64). In both cohorts, greater systolic blood pressure was associated with greater BMI (p < 0.01).

Table 2.

Associations of Aβ burden with longitudinal decline in BMI by cohort

Model variables β estimate SE t value p value
BMI ~ (Aβ × time) + (age × time) + (sex × time) + (time-varying cognitive performance) + (time-varying depressive symptoms) + (time-varying systolic blood pressure)
HABS (n = 312)
 Time 0.12 0.12 1.00 0.31
 Aβ −1.98 1.34 −1.48 0.14
 Age −0.05 0.03 −1.46 0.15
 Sex 0.29 0.52 0.55 0.58
 Cognitive performance −0.013 0.07 −0.19 0.85
 Depressive symptoms −0.07 0.02 −3.07 0.002
 Systolic blood pressure 0.01 0.001 3.01 0.003
 Aβ × time −0.20 0.10 −1.93 0.05
 Age × time −0.01 0.003 −2.88 0.004
 Sex × time 0.01 0.04 0.31 0.76
ADNI (n = 336)
 Time 0.32 0.15 2.06 0.04
 Aβ −3.29 1.46 −2.25 0.03
 Age −0.17 0.04 −4.02 < 0.001
 Sex −0.78 0.54 −1.45 0.15
 Cognitive performance 0.13 0.06 1.98 0.05
 Depressive symptoms 0.01 0.02 0.47 0.64
 Systolic blood pressure 0.01 0.002 6.09 <0.001
 Aβ × time −0.33 0.13 −2.54 0.01
 Age × time −0.003 0.004 −0.82 0.41
 Sex × time −0.05 0.05 −0.95 0.34
Open in a new tab

Separate models were run in HABS and ADNI. Bolded values are significant.

Key: Aβ, amyloid-beta; ADNI, Alzheimer’s Disease Neuroimaging Initiative; BMI, body mass index; HABS, Harvard Aging Brain Study; SE, standard error.

Fig. 1.

Fig. 1.

Open in a new tab

Higher baseline Aβ burden was associated with accelerated decline in body mass index. For visualization purposes, modeled longitudinal change in body mass index (BMI) is depicted at lower and higher levels of Aβ based on Aβ values at the 10th and 90th percentiles, respectively, in each cohort. The plots demonstrate that greater Aβ burden at baseline is associated with more rapid decline in BMI over time in (A) the Harvard Aging Brain Study and (B) the Alzheimer’s Disease Neuroimaging Initiative. Shaded regions represent the 95% confidence intervals.

3.3. Post hoc analyses

When we examined the possible 3-way interaction effects of Aβ burden, APOE ε4 status, and time on BMI change, this 3-way interaction was not significant in HABS (β= 0.04, standard error [SE] = 0.22, t = 0.19, p = 0.85). This same interaction was at trend level in ADNI (β= 0.51, SE = 0.26, t = 1.92, p = 0.06), whereby Aβ burdenwas more strongly associated with declining BMI in APOE ε4 noncarriers compared to carriers. In post hoc analyses, we also tested whether the association of Aβ burden with BMI change could be attributed to metabolic or endocrine disturbances. After excluding participants with diabetes or endocrine dysfunction as well as underweight or obese participants, the association of higher Aβ burden with BMI change remained significant in both HABS (β=−0.22, SE = −0.11, t=−1.11, p = 0.04) and ADNI (β=−0.36, SE = 0.17, t = −2.19, p = 0.03), suggesting that the primary results were not driven by these factors.

4. Discussion

The present study demonstrated that higher Aβ burden at baseline, as measured by PET, was associated with greater decline in BMI over time. This association was significant in 2 independent cohorts of clinically normal adults after adjusting for age, sex, cognitive performance, depressive symptoms, and blood pressure. The association between higher Aβ burden and greater longitudinal decline in BMI also remained significant after excluding participants with diabetes/endocrine dysfunction and participants considered underweight or obese. This latter result indicates that the association of Aβ burden with longitudinal decline in BMI was not driven by the presence of diabetes/endocrine conditions or by individuals with very high or low BMIs. Together, these findings suggest that declining BMI in clinically normal adults may be an early indicator of elevated Aβ burden.

Our results are in agreement with several cross-sectional studies in clinically normal older adults demonstrating that lower BMI is associated with higher levels of Aβ burden as measured by PET (Hsu et al., 2016; Thirunavu et al., 2019) or with cerebrospinal fluid markers of Aβ and tau (Ewers et al., 2012; Vidoni et al., 2011). Similar relationships have also been reported at autopsy in non-demented older adults, whereby lower BMI proximate to death was associated with increased AD pathology (Buchman et al., 2006). Interestingly, in that study, the relationship was specific to AD pathology, as neither cerebral infarcts nor Lewy body pathology was related to BMI. Our findings are also consistent with recent work in autosomal dominant AD, showing that mutation carriers have significantly lower BMI than noncarriers, and that this difference is evident more than 10 years before expected symptom onset (Müller et al., 2017). The present longitudinal results build on the previously mentioned cross-sectional findings by demonstrating that higher Aβ burden at baseline was associated with greater decline in BMI over time. The relatively large sample size (>300 in each cohort) and long follow-up period (a median of >4 years) likely yielded increased power to detect associations in the preclinical stage of the disease, given that a prior longitudinal study in ADNI with a smaller sample size (n = 100) and only up to 2 follow-up visits did not observe a significant association (Vidoni et al., 2011).

Some researchers have suggested that cognitive impairment itself contributes to late-life weight loss (Gustafson, 2006); however, the participants in the present study were clinically normal at baseline, which argues against this explanation. Furthermore, the association of higher Aβ burden with declining BMI was significant after adjusting for cognitive performance in both cohorts, suggesting that worsening cognition was not driving the effect. Depressive symptoms have also been linked to both late-life weight loss (McMinn et al., 2011; Morley and Kraenzle, 1994) and AD pathology across the clinical course of the disease (Donovan et al., 2018; Gatchel et al., 2017; Rapp et al., 2008). Importantly, the association between baseline Aβ burden and change in BMI was independent of GDS scores in both cohorts, suggesting that this association cannot be explained by depressive symptoms. In the HABS sample alone, we found that higher depressive symptoms were significantly associated with lower BMI. This finding is consistent with the idea that weight loss is a key feature of depression in late-life (Morley, 2001) and may be due to apathy toward normally appetitive behaviors (Mann, 2005).

When we examined the influence of APOE ε4 genotype on the association between Aβ burden and BMI decline, we found that it did not moderate the association in HABS; however, there was a trend toward significance in ADNI, whereby Aβ-related decline in BMI was stronger in APOE ε4 noncarriers compared to carriers. The literature has been mixed with regards to whether APOE ε4 genotype moderates the relationship between Aβ burden and BMI. Two prior cross-sectional studies reported no moderating effect of APOE ε4 status (Ewers et al., 2012; Thirunavu et al., 2019), whereas 2 other cross-sectional studies found that Aβ burden was associated with lower BMI in APOE ε4 carriers, but not noncarriers (including one from HABS; Blautzik et al., 2018; Hsu et al., 2016). The reasons for the discrepancies are not clear and require further investigation.

Post hoc analyses also investigated whether the association between Aβ burden and BMI decline was related to the presence of diabetes/endocrine dysfunction or underweight/obese individuals. Importantly, the results remained largely unchanged after excluding such participants despite significantly reduced sample sizes (HABS = 195 and ADNI = 155). These findings suggest that the association of elevated Aβ burden with BMI decline is observable even in individuals with normal or close-to-normal BMIs at baseline and in those without diabetes or endocrine dysfunction.

The biological mechanisms underlying the relationship between higher Aβ burden and accelerated decline in BMI remain unclear. One possibility is that Aβ accumulation (or possibly Aβ-associated tau pathology) may interfere with the functioning of the hypothalamus, which serves to regulate body weight and systemic metabolism (Ishii and Iadecola, 2016). It has been suggested that altered hypothalamic function may reduce leptin levels and cause a hypermetabolic state, which can accelerate weight loss (Hiller and Ishii, 2018; Ishii et al., 2014). Another possibility is that AD pathology and/or downstream neurodegeneration may reduce olfactory function (Bacon et al., 1998; Growdon et al., 2015), increase anxiety (Hanseeuw et al., 2018; Ramakers et al., 2013) or increase apathy (Lanctot et al., 2017), all of which can decrease food intake and cause a decline in BMI. Alternatively, it may be that metabolic changes associated with a decline in BMI instigate or exacerbate Aβ deposition or that a common upstream process drives both Aβ and decline in BMI (e.g., inflammatory processes; Morris et al., 2014).

Strengths of the present study were the replication of our findings across 2 independent cohorts with Aβ-PET imaging and BMI follow-up data over a median period of >4 years. In addition, HABS and ADNI used different Aβ tracers (PiB and florbetapir, respectively), suggesting that the results are not dependent on a specific Aβ tracer. There were also several limitations. First, we relied on BMI as our outcome measure, as this was the only anthropometric measure available in both cohorts. However, BMI does not differentiate between fat tissue and lean muscle mass and therefore may not be an optimal measure of adiposity (Burns et al., 2010). Future studies using more accurate measurements of body composition (e.g., imaging adipose tissue) may provide a better understanding of the relationship between Aβ burden and weight loss in preclinical AD. Second, we did not include measures of tau pathology in the present study, as tau PET was collected several years after Aβ PET in both HABS and ADNI. Future studies should investigate whether decline in BMI is accelerated in individuals with both elevated Aβ and tau deposition. Third, since ADNI did not collect information on diabetes specifically, we relied on self-reported endocrine dysfunction. Fourth, given the sample size of each cohort, there may have been insufficient power to conduct analyses moderated by APOE ε4 status; therefore, these analyses should be interpreted cautiously. Fifth, longitudinal physical activity was not assessed in HABS or ADNI. It may, however, be an important confounder given that reduced physical activity is associated with weight loss in older adults (Baumgartner et al., 1999) and may also accelerate Aβ-related cognitive decline and neurodegeneration (Rabin et al., 2019). Finally, participants in HABS and ADNI are generally well educated, predominantly Caucasian, and generally healthy (i.e., may underrepresent individuals with high vascular risk). These factors may limit the generalizability of our findings and it will be important to replicate our results in more diverse cohorts.

5. Conclusions

We provide new evidence from 2 independent, well-characterized cohorts that higher Aβ burden at baseline is associated with greater decline in BMI in clinically normal adults. Importantly, this association persisted after adjusting for several important covariates, including cognitive performance, depressive symptoms, and metabolic or endocrine dysfunction. These findings lend credence to the idea that declining BMI in late life may not only be due to normal age-related processes, but may also be an early indicator of elevated Aβ burden.

Acknowledgements

Some data used in the preparation of this article were obtained from the Alzheimer’s Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). The ADNI (www.adni.loni.usc.edu) was launched in 2003 as a public-private partnership, led by Principal Investigator Michael W. Weiner, MD. The primary goal of ADNI has been to test whether serial magnetic resonance imaging (MRI), positron emission tomography (PET), other biological markers, and clinical and neuropsychological assessment can be combined to measure the progression of MCI and early AD. For up-to-date information, see www.adni-info.org/.

This work was supported by the National Institute on Aging (NIA) at the National Institutes of Health (NIH) (P01 AG036694 to R.A.S. and K.A.J., K24 AG035007 to R.A.S., K23 AG049087 to J.P.C., R01 AG053509 and R01 AG054110 to T.H, and P50 AG005134 to R.A.S., K.A.J., and T.H. J.S.R. was supported by a Canadian Institutes of Health Research Postdoctoral Fellowship Award, H.-S.Y. is supported by NIA/NIH K23 AG062750; R.F.B. is supported by the NHMRC Dementia Research Fellowship (APP1105576). J.R.G. is supported by NIA/NIH K23 AG058805, AACF-16-440965 and BrightFocus Foundation ADRF A2016434 F. W.S. is supported by the Alzheimer’s Association and Brain Canada (AARG501466).

References

  1. Aisen PS, Petersen RC, Donohue MC, Gamst A, Raman R, Thomas RG, Walter S, Trojanowski JQ, Shaw LM, Beckett LA, Jack CR, Jagust W, Toga AW, Saykin AJ, Morris JC, Green RC, Weiner MW, Initiative N, 2010. Clinical core of the Alzheimer’s disease neuroimaging initiative: progress and plans. Alzheimers Dement. 6, 239–246. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bacon A, Bondi MW, Salmon D, Murphy C, 1998. Very early changes in olfactory functioning due to Alzheimer’s disease and the role of apolipoprotein E in olfaction. Ann. N. Y. Acad. Sci 55, 723–731. [DOI] [PubMed] [Google Scholar]
  3. Baumgartner RN, Waters DL, Gallagher D, Morley JE, Garry PJ, 1999. Predictors of skeletal muscle mass in elderly men and women. Mech. Ageing Dev 107, 123–136. [DOI] [PubMed] [Google Scholar]
  4. Bays HE, Chapman RH, Grandy S, Group I, 2007. The relationship of body mass index to diabetes mellitus, hypertension and dyslipidaemia: comparison of data from two national surveys. Int. J. Clin. Pract 61, 737–747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Besser LM, Gill DP, Monsell SE, Brenowitz W, Meranus DH, Kukull W, Gustafson DR, 2014. Body mass index, weight change, and clinical progression in mild cognitive impairment and Alzheimer’s disease. Alzheimers Dis. Assoc. Disord 28, 36–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Blautzik J, Kotz S, Brendel M, Sauerbeck J, Vettermann F, 2018. Relationship between body mass index , ApoE4 status , and PET-based amyloid and neurodegeneration markers in amyloid-positive subjects with normal cognition or mild cognitive impairment. J. Alzheimers Dis 65, 781–791. [DOI] [PubMed] [Google Scholar]
  7. Buchman A, Wilson R, Bienias JL, Shah R, Evans D, Bennett DA, 2005. Change in body mass index and risk of incident Alzheimer disease. Neurology 65, 892–897. [DOI] [PubMed] [Google Scholar]
  8. Buchman AS, Schneider JA, Wilson RS, Bienias JL, Bennett DA, 2006. Body mass index in older persons is associated with Alzheimer disease pathology. Neurology 67, 1949–1954. [DOI] [PubMed] [Google Scholar]
  9. Burns JM, Johnson DK, Watts A, Swerdlow RH, Brooks WM, 2010. Reduced lean mass in early Alzheimer disease and its association with brain atrophy. Arch. Neurol 67, 428–433. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Corrada MM, Kawas CH, Mozaffar F, Paganini-Hill A, 2006. Association of body mass index and weight change with all-cause mortality in the elderly. Am. J. Epidemiol 163, 938–949. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Dagley A, LaPoint M, Huijbers W, Hedden T, McLaren DG, Chatwal JP, Papp KV, Amariglio RE, Blacker D, Rentz DM, Johnson KA, Sperling RA, Schultz AP, 2017. Harvard aging brain study: dataset and accessibility. Neuroimage 144, 255–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Domonkos Tarnoki A, Laszlo Tarnoki D, Bogl LH, Medda E, Fagnani C, Nisticò L, Antonietta M, Brescianini S, Lucatelli P, Boatta E, Zini C, Fanelli F, Baracchini C, Meneghetti G, Osztovits J, Jermendy G, Gabor R, Preda I, Karlinger K, Lannert A, Agnes A, Littvay L, Garami Z, Berczi V, Pucci G, Baffy G, Schillaci G, Pietiläinen KH, 2013. Association of body mass index with arterial stiffness and blood pressure components: atwin study. Atherosclerosis 229, 388–395. [DOI] [PubMed] [Google Scholar]
  13. Donohue MC, Sperling RA, Salmon DP, Rentz DM, Raman R, Thomas RG, Weiner M, Aisen PS, 2014. The preclinical Alzheimer cognitive composite: measuring amyloid-related decline. JAMA Neurol. 71, 961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Donovan NJ, Locascio JJ, Marshall GA, Gatchel J, Hanseeuw BJ, Rentz DM, Johnson KA, Sperling RA, 2018. Longitudinal association of amyloid beta and anxious-depressive symptoms in cognitively normal older adults. Am. J. Psychiatry 175, 530–537. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ewers M, Schmitz S, Hansson O, Walsh C, Fitzpatrick A, Bennett D, Minthon L, Trojanowski JQ, Shaw LM, Faluyi YO, Vellas B, Dubois B, Blennow K, Buerger K, 2012. Body mass index is associated with biological CSF markers of core brain pathology of Alzheimer’s disease. Neurobiol. Aging 33, 1599–1608. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fitzpatrick AL, Kuller LH, Lopez OL, Diehr P, O’Meara E, Longstreth T, Luchsinger JA, 2009. Midlife and late-life obesity and the risk of dementia. Arch. Neurol 66, 336–342. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Folstein MF, Folstein SE, McHugh PR, 1975. Mini-mental state: a practical method for grading the cognitive state of patients for the clinician. J. Psychiatr. Res 12, 189–198. [DOI] [PubMed] [Google Scholar]
  18. Gatchel JR, Donovan NJ, Locascio JJ, Schultz AP, Becker JA, Chhatwal J, Papp KV, Amariglio RE, Rentz DM, Blacker D, Sperling RA, Johnson KA, Marshall GA, 2017. Depressive symptoms and tau accumulation in the inferior temporal lobe and entorhinal cortex in cognitively normal older adults: a pilot study. J. Alzheimers Dis 59, 975–985. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Growdon ME, Schultz AP, Dagley AS, Amariglio RE, Hedden T, Rentz DM, Johnson KA, Sperling RA, Albers MW, 2015. Odor identification and Alzheimer disease biomarkers in clinically normal elderly. Neurology 84, 2153–2160. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Gustafson D, 2006. Adiposity indices and dementia. Lancet Neurol. 5, 713–720. [DOI] [PubMed] [Google Scholar]
  21. Hanseeuw BJ, Jonas V, Jackson J, Betensky RA, Rentz DM, Johnson KA, Sperling RA, Donovan NJ, Donovan NJ, 2018. Association of anxiety with subcortical amyloidosis in cognitively normal older adults. Mol. Psychiatry 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hiller AJ, Ishii M, 2018. Disorders of body weight, sleep and circadian rhythm as manifestations of hypothalamic dysfunction in Alzheimer’s disease. Front. Cell. Neurosci 12, 471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Hsu DC, Mormino EC, Schultz AP, Amariglio RE, Donovan NJ, Rentz DM, Johnson KA, Sperling RA, Marshall GA, Study, for the H.A.B, 2016. Lower late-life body-mass index is associated with higher cortical amyloid burden in clinically normal elderly. J. Alzheimers Dis 53, 1097–1105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hughes TF, Borenstein A, Wu Y, Larson E, 2009. Association between late-life body mass index and dementia: the Kame Project. Neurology 72, 1741–1746. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Ishii M, Iadecola C, 2016. Adipocyte-derived factors in age-related dementia and their contribution to vascular and Alzheimer pathology. Biochim. Biophys. Acta 1862, 966–974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Ishii M, Wang G, Racchumi G, Dyke JP, Iadecola C, 2014. Transgenic mice overexpressing amyloid precursor protein exhibit early metabolic deficits and a pathologically low leptin state associated with hypothalamic dysfunction in arcuate neuropeptide Y neurons. J. Neurosci 34, 9096–9106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jack CR, Bennett DA, Blennow K, Carrillo MC, Dunn B, Haeberlein SB, Holtzman DM, Jagust W, Jessen F, Karlawish J, Liu E, Molinuevo JL, Montine T, Phelps C, Rankin KP, Rowe CC, Scheltens P, Siemers E, Snyder HM, Sperling R, 2018. NIA-AA Research Framework: toward a biological definition of Alzheimer’s disease. Alzheimers Dement. 14, 535–562. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Johnson DK, Wilkins CH, Morris JC, 2006. Accelerated weight loss may precede diagnosis in Alzheimer disease. Arch. Neurol 63, 1312–1317. [DOI] [PubMed] [Google Scholar]
  29. Johnson KA, Schultz A, Betensky RA, Becker JA, Sepulcre J, Rentz D, Mormino E, Chhatwal J, Amariglio R, Papp K, Marshall G, Albers M, Mauro S, Pepin L, Alverio J, Judge K, Philiossaint M, Shoup T, Yokell D, Dickerson B, Gomez-Isla T, Hyman B, Vasdev N, Sperling R, 2016. Tau positron emission tomographic imaging in aging and early Alzheimer disease. Ann. Neurol 79, 110–119. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Kivimaki M, Luukkonen R, Batty GD, Ferrie J, Pentti J, Nyberg S, Shipley M, Jokela M, 2018. Body mass index and risk of dementia: analysis of individual-level data from 1.3 million individuals. Alzheimers Dement. 14, 601–609. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Kivipelto M, Ngandu T, Fratiglioni L, Viitanen M, Kareholt I, Winblad B, Helkala EL, Tuomilehto J, Soininen H, Nissinen A, 2005. Obesity and vascular risk factors at midlife and the risk of dementia and Alzheimer disease. Arch. Neurol 62, 1556–1560. [DOI] [PubMed] [Google Scholar]
  32. Lanctot K, Ortiz-Augera L, Brodaty H, Francis PT, Geda YE, Ismail Z, Marshall GA, Mortby ME, Onyike CU, Palada P, Politis A, Rosenberg PB, Emma S, Sultzer DL, Abraham EH, 2017. Apathy associated with neuro-cognitive disorders: recent progress and future directions. Alzheimers Dement. 13, 84–100. [DOI] [PubMed] [Google Scholar]
  33. Landau SM, Mintun MA, Joshi AD, Koeppe RA, Petersen RC, Aisen PS, Weiner MW, Jagust WJ, 2012. Amyloid deposition, hypometabolism, and longitudinal cognitive decline. Ann. Neurol 72, 578–586. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mann JJ, 2005. The medical management of depression. N. Engl. J. Med 353, 1819–1834. [DOI] [PubMed] [Google Scholar]
  35. Mathys J, Gholamrezaee M, Henry H, Gunten A, Popp J, 2017. Decreasing body mass index is associated with cerebrospinal fluid markers of Alzheimer’s pathology in MCI and mild dementia. Exp. Gerontol 100, 45–53. [DOI] [PubMed] [Google Scholar]
  36. McMinn J, Steel C, Bowman A, 2011. Investigation and management of unintentional weight loss in older adults. BMJ 342, d1732. [DOI] [PubMed] [Google Scholar]
  37. Morley JE, 2001. Decreased food intake with aging. J. Gerentol. Ser. A Biol. Sci. Med. Sci 56, 81–88. [DOI] [PubMed] [Google Scholar]
  38. Morley JE, Kraenzle D, 1994. Causes of weight loss in a community nursing home. J. Am. Geriatr. Soc 42, 583–585. [DOI] [PubMed] [Google Scholar]
  39. Morris JC, 1993. The Clinical Dementia Rating (CDR): current version and scoring rules. Neurology 43, 2412–2414. [DOI] [PubMed] [Google Scholar]
  40. Morris JK, Honea RA, Vidoni ED, Swerdlow RH, Burns JM, 2014. Is Alzheimer’s disease a systemic disease? Biochim. Biophys. Acta 1842, 1340–1349. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Müller S, Preische O, Sohrabi HR, Gräber S, Dietzsch J, Ringman JM, Martins RN, Mcdade E, Schofield PR, Ghetti B, Rossor M, Graff-Raford NR, Levin J, Galasko DR, Quaid K, Salloway S, Xiong C, Benzinger TLS, Buckles V, Masters CL, Sperling RA, Bateman RJ, Morris JC, Laske C, 2017. Decreased body mass index in the preclinical stage of autosomal dominant Alzheimer’s disease. Sci. Rep 7, 1225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Papp KV, Rentz DM, Orlovsky I, Sperling RA, Mormino EC, 2017. Optimizing the preclinical Alzheimer’s cognitive composite with semantic processing: the PACC5. Alzheimers Dement. (N Y) 3, 668–677. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Rabin JS, Klein H, Kirn D, Schultz A, Yang H-S, Buckley RF, Viswanathan A, Hedden T, Pruzin J, Yau WW, Guzmán-vélez E, Sperling RA, Chhatwal JP, 2019. Associations of physical activity and β-amyloid with longitudinal cognition and neurodegeneration in clinically normal older adults. JAMA Neurol. 76, 1203–1210. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Ramakers I, Verhey F, Scheltens P, Hampel H, Soininen H, Rikkert MO, Spiru L, Blennow K, Trojanowski JQ, Shaw L, Visser P, Alzheimer’s Disease NeuroimaginG Initiative, Investigators and DESCRIPA Investigators, 2013. Anxiety is related to Alzheimer cerebrospinal fluid markers in subjects with mild cognitive impairment. Psychol. Med 43, 911–920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rapp M, Schaider-Beeri M, Purohit DP, Perl DP, Haroutunian V, Sano M, 2008. Increased neurofibrillary tangles in patients with Alzheimer disease with comorbid depression. Am. J. Geriatr. Psychiatry 16, 168–174. [DOI] [PubMed] [Google Scholar]
  46. Sheikh JI, Yesavage JA,1986. Geriatric Depression Scale (GDS): recent evidence and development of a shorter version. Clin. Gerontol. J. Aging Ment. Heal 5, 165–173. [Google Scholar]
  47. Sperling RA, Aisen PS, Beckett LA, Bennett DA, Craft S, Fagan AM, Iwatsubo T, Jack CR, Kaye J, Montine TJ, Park DC, Reiman EM, Rowe CC, Siemers E, Stern Y, Yaffe K, Carrillo MC, Thies B, Morrison-Bogorad M, Wagster MV, Phelps CH, 2011. Toward defining the preclinical stages of Alzheimer’s disease: recommendations from the National Institute on Aging-Alzheimer’s Association workgroups on diagnostic guidelines for Alzheimer’s disease. Alzheimers Dement. 7, 280–292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Sperling RA, Mormino E, Johnson KA, 2014. The evolution of preclinical Alzheimer’s disease: implications for prevention trials. Neuron 84, 608–622. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Thirunavu V, Mccullough A, Su Y, Flores S, Dincer A, Morris J, Cruchaga C, Benzinger TLS, Gordon BA, 2019. Higher body mass index is associated with lower cortical amyloid-b burden in cognitive normal individuals in late-life. J. Alzheimers Dis 69, 817–827. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Vidoni ED, Honea RA, Burns JM, For the Alzheimer’s Disease Neuroimaging Initiative, 2011. Alzheimer disease biomarkers are associated with body mass index. Neurology 77, 1913–1920. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Wechsler D, 1987. WMS-R: Wechsler Memory Scale-Revised. Psychological Corporation, San Antonio, TX. [Google Scholar]
  52. Whitmer R, Sidney S, Selby J, Johnston S, Yaffe K, 2005. Midlife cardiovascular risk factors and risk of dementia in late life. Neurology 64, 277–282. [DOI] [PubMed] [Google Scholar]
  53. Yesavage JA, Brink TL, Rose TL, Lum O, Huang V, Adey M, Leirer VO, 1982. Development and validation of a geriatric depression screening scale: a preliminary report. J. Psychiatr. Res 17, 37–49. [DOI] [PubMed] [Google Scholar]

RESOURCES