. 18F-fluorodeoxyglucose positron emission tomography, aging, and apolipoprotein E genotype in cognitively normal persons. Neurobiol Aging. 2014 Sep;35(9):2096-106. Epub 2014 Mar 11 PubMed.


Please login to recommend the paper.


  1. This study by Knopman and colleagues from Mayo Clinic demonstrates that decreases in brain glucose metabolism are independently associated with both advancing age and ApoE4 genotype in cognitively normal individuals. The effects of age appeared quite global, while the ApoE4 effect was relatively specific to AD-vulnerable regions. Strengths of the study include the large sample size, and the use of partial volume correction to demonstrate that metabolic decline is independent of atrophy. A relative weakness is the small number of ApoE4 homozygotes—this is expected in studies of normal aging, but precluded investigation of a dose effect for ApoE4.

    A striking finding in the study is that the effect of ApoE4 was present across a broad age range (30-95), and was independent of amyloid burden as assessed by PiB-PET. While ApoE4 is clearly a risk factor for amyloid aggregation, there is increasing evidence that ApoE4 also increases the risk for AD via a multitude of Aβ-independent pathways, some of which traverse the entire life-span (see review by Huang and Mucke,  2012, and interesting model proposed by Jagust and Mormino, 2011). This has major implications as the field develops strategies for AD treatment and prevention in E4-positive individuals.


    . Alzheimer mechanisms and therapeutic strategies. Cell. 2012 Mar 16;148(6):1204-22. PubMed.

    . Lifespan brain activity, β-amyloid, and Alzheimer's disease. Trends Cogn Sci. 2011 Nov;15(11):520-6. PubMed.

  2. That [18F]-T807 uptake in FTLD syndromes co-localizes with structural changes seen on MRI scans demonstrates the sensitivity of this PET ligand for the pathology of FTLD. This finding has implications for early diagnosis as well as for tracking disease progression and potential treatment responses at the initial stages of FTLD pathology. Ideally, [18F] T807 uptake would differentiate between FTLD syndromes with TDP versus those associated with tau deposition. However, the specificity of [18F] T807 binding to the tau protein in FTLD needs to be further investigated with in-vitro and in-vivo studies and with autopsy confirmation.

    View all comments by Kejal Kantarci
  3. The interesting paper by Knopman et al. referred to in this report convincingly demonstrates statistically significant age reductions in brain glucose metabolism as measured with PET FDG in a large cohort (806) of cognitively normal subjects ranging in age from 30 to nearly 100. The conclusions differ somewhat from findings by Ibanez et al. in a study that examined statistically insignificant age changes in resting-state, atrophy-corrected glucose metabolism comparing 13 older men (55 to 82 years old) with 11 younger men (22 to 34 years).

    While the difference certainly can be ascribed to the Knopman paper's greater power due to its larger numbers and many older subjects, other methodological differences deserve to be noted. One is that the Ibanez paper determined absolute glucose metabolic rates by quantifying the arterial input function to calculate metabolism, where the Knopman study did a ratio analysis (regional to global incorporated radioactivity) to estimate glucose metabolism. Additionally, the Ibanez study excluded hypertensive subjects, whereas the Knopman paper does not mention subject blood pressure. The exclusion is relevant because Salerno et al. reported that even well-treated hypertensive subjects had reduced brain glucose metabolism compared with controls, and the prevalence of hypertension increases with aging.

    Both the Knopman and Ibanez papers agree that cerebral atrophy occurs with aging, and another study using PET FDG reported reduced correlations between regional metabolic rates, suggesting functional disconnection as well. 


    . 18F-fluorodeoxyglucose positron emission tomography, aging, and apolipoprotein E genotype in cognitively normal persons. Neurobiol Aging. 2014 Sep;35(9):2096-106. Epub 2014 Mar 11 PubMed.

    . Resting state brain glucose metabolism is not reduced in normotensive healthy men during aging, after correction for brain atrophy. Brain Res Bull. 2004 Mar 15;63(2):147-54. PubMed.

    . Brain metabolic function in older men with chronic essential hypertension. J Gerontol A Biol Sci Med Sci. 1995 May;50(3):M147-54. PubMed.

    . Age differences in intercorrelations between regional cerebral metabolic rates for glucose. Ann Neurol. 1986 Jan;19(1):60-7. PubMed.

    View all comments by Stanley Rapoport
  4. I would like to comment on this interesting article that refers to our previous studies (Yanase et al., 2005; Samuraki et al., 2012) on similar topics.

    Knopman et al. observed a modest but significant age-related FDG decline in many brain regions, including areas that are known to be affected in AD, such as the posterior cingulate/precuneus, even after correction for atrophy. In our prior study, also using atrophy-corrected FDG PET (Yanase et al., 2005), the area of age-related FDG decline was confined to fewer regions, including the anterior cingulate and inferior frontal gyrus. There are at least two differences between the studies that may have affected the results. First, their study included a much larger number of cognitively normal individuals (n=806) than our study (n=139), which is certainly beneficial for unmasking significant regions that could not be detected in smaller studies. Second, the atrophy-correction algorithms were different. Where our study used a three-compartment method that accounted for the effect of partial-volume averaging between gray and white matter in addition to the diluting effects of CSF spaces, Knopman and colleagues used a two-compartment method. Because each method has its advantages and disadvantages, it would be worthwhile to see whether the results are consistent over different methodologies.

    In their study, there was an APOE ε4-associated FDG decline in AD-vulnerable regions, which was independent of amyloid burden as measured by PiB PET. In our other prior study (Samuraki et al., 2012), however, the prevalence of AD-like imaging abnormalities, such as an FDG decline in the posterior cingulate/precuneus or hippocampal atrophy, did not differ among APOE ε4 carriers with normal cognitive function and non-carriers. In that study, we did not use atrophy correction for FDG PET, because we expected that uncorrected FDG signal would reflect both metabolism and atrophy due to partial-volume effect, and therefore may serve as a more sensitive marker of AD-associated changes. This should particularly be the case in individuals with posterior cingulate/precuneus atrophy, which is known to occur in a subset of AD patients (Shima et al., 2012). That they found the opposite—FDG decline in AD-vulnerable regions was more significant with atrophy correction than without—is interesting. I assume that the use of atrophy correction may have lessened atrophy-related variability in measured FDG signal and thereby increased the statistical power to detect APOE ε4-associated alterations, although this hypothesis needs to be addressed in further studies. 

    In addition to the atrophy-correction issue, the magnitude of APOE ε4-associated FDG decline that they found seems to be much smaller than that typically seen in AD patients. 

    Further, our study participants were much younger (mean age: 53.6 years for the ε4 carriers and 53.5 years for the non-carriers) than those in the Knopman et al. study (median age: 76 years); the majority of our participants were 30 to 60 years old: there were no significant differences between ε4 carriers and non-carriers in that age group in their study. The APOE ε 4-associated FDG decline may be more difficult to detect in such young subjects. All these circumstances, alone or in combination, may have contributed to the somewhat discordant observations between the studies.

    From a pathophysiological perspective, it would be interesting to see how APOE ε4-associated FDG decline interacts with structural alterations measured by MRI. In our study, a cognitively normal individual with AD-like FDG decline did not necessarily have hippocampal atrophy, and vice versa. In this regard, PET and structural MRI combination would be useful for more comprehensive understanding of APOE ε4-related metabolism-atrophy interaction.

    Finally and most importantly, it should be noted that many factors other than APOE ε4, such as diabetes mellitus, are likely to be involved in the development of AD-like imaging signatures in cognitively normal individuals, which has nicely been demonstrated in a recent study by the same research group (Roberts et al., 2014). In our study a subset of ε4 non-carriers showed AD-like imaging signatures, further supporting this notion. Non-ApoE factors can be genetic or non-genetic, and need to be identified in further studies.

    In summary, this study provides an important clue for better understanding of interaction between normal aging and AD-related changes in the brain. At the same time, questions still remain to be addressed. 


    . 18F-fluorodeoxyglucose positron emission tomography, aging, and apolipoprotein E genotype in cognitively normal persons. Neurobiol Aging. 2014 Sep;35(9):2096-106. Epub 2014 Mar 11 PubMed.

    . Diabetes and elevated hemoglobin A1c levels are associated with brain hypometabolism but not amyloid accumulation. J Nucl Med. 2014 May;55(5):759-64. Epub 2014 Mar 20 PubMed.

    . Glucose metabolism and gray-matter concentration in apolipoprotein E ε4 positive normal subjects. Neurobiol Aging. 2011 Dec 19; PubMed.

    . Posterior cingulate atrophy and metabolic decline in early stage Alzheimer's disease. Neurobiol Aging. 2011 Aug 18; PubMed.

    . Brain FDG PET study of normal aging in Japanese: effect of atrophy correction. Eur J Nucl Med Mol Imaging. 2005 Jul;32(7):794-805. Epub 2005 Mar 10 PubMed.

  5. I agree with the comments from Dr. Rabinovici regarding ApoE possibly playing a multifactorial role in AD risk over the lifespan. In addition to the excellent articles by Huang and Mucke, and Jagust and Mormino, I would like to mention two reviews that I was involved with that provide expanded discussion of the potential link of apoE to energy metabolism (Wolf et al., 2013; Wolf et al., 2013). Additionally, I believe it is fundamental that cellular/molecular studies be utilized to determine potential mechanisms underlying the findings in these brain imaging studies. 


    . Apolipoprotein E as a β-amyloid-independent factor in alzheimer's disease. Alzheimers Res Ther. 2013 Sep 3;5(5):38. PubMed.

    . APOE and neuroenergetics: an emerging paradigm in Alzheimer's disease. Neurobiol Aging. 2013 Apr;34(4):1007-17. PubMed.

    View all comments by Andrew Wolf

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. PET Tracers Enlighten at Neurology Conference