By Makoto Ishii, MD, PhD
Assistant Professor of Neuroscience and Neurology, Feil Family Brain and Mind Research Institute, Department of Neurology, Weill Cornell Medical College
Dr. Ishii reports he is a stockholder in Regeneron.
SYNOPSIS: In a prospective, controlled study of 49 patients with dementia and 25 healthy controls, marked hyperphagia is restricted to behavioral-variant frontotemporal dementia patients that is likely due to differing neural networks, while increased sucrose preference is likely controlled by a similar network in both behavioral-variant frontotemporal dementia and semantic dementia patients.
SOURCE: Ahmed RM, Irish M, Henning E, et al. Assessment of eating behavior disturbance and associated neural networks in frontotemporal dementia. JAMA Neurol 2016:73:282-290.
Changes in eating behavior are common in frontotemporal dementia (FTD). Hyperorality (i.e., oral exploration of inedible objects) and dietary changes are criteria for the diagnosis of behavioral-variant frontotemporal dementia (bvFTD), while semantic dementia (SD) patients are reported to have rigid eating behaviors and changes in food preference. However, eating behavior typically is assessed by caregiver questionnaires, which may be biased by subjective interpretation and reporting errors. Furthermore, neuroimaging studies suggest that atrophy in different brain regions contributes to the behavior abnormalities in FTD, but the precise neural networks involved in eating behavior are not known. Therefore, Ahmed et al conducted a prospective controlled study in bvFTD and SD patients to 1) rigorously examine the changes in eating behavior using ecologically valid methods and 2) identify the associated neural networks using voxel-based morphometry analyses of structural brain MRI.
The study was conducted from Nov. 1, 2013, to May 31, 2015. Forty-nine study participants with dementia (19 bvFTD, 15 SD, and 15 Alzheimer’s disease [AD]) were recruited in Australia. Twenty-five healthy controls were recruited in Australia and the United Kingdom. Mean age for the four groups ranged from 62 to 66 years. The dementia and control groups were matched for age, sex, and body mass index to remove potential effects on eating behavior. Caloric intake and food preference were assessed by the ad libitum breakfast test meal. Sucrose preference was assessed by measuring the liking ratings of three desserts of varying sucrose content (26%, 39%, and 60%). To identify the neural networks, all participants underwent whole brain 3T MRI on the day of the eating experiments. The MRI data were analyzed with voxel-based morphometry and voxelwise general linear models to identify correlations between performances on the two eating behavior tests and gray matter volume.
In this study, the bvFTD group was more functionally impaired relative to the AD (Frontal Rating Scale; P = 0.009) and SD groups (P < 0.001). The bvFTD group also had more severe eating disturbances based on caregiver surveys (P < 0.005 for all). The two eating behavior tests revealed important similarities and differences between bvFTD and SD groups. All patients with bvFTD had markedly elevated total caloric intake and hyperphagia that was not seen in the other groups (P < 0.001). The SD group did not have increased caloric intake, but a number of patients had rigid eating behavior, including refusing to eat, that affected their food preferences. Both bvFTD and SD patients had a strong sweet preference compared with AD and control groups, which was not due to altered sweetness perception, as they were similar among all the groups.
Voxel-based morphometry analyses found complex mechanisms underlying the changes in eating behavior in FTD patients that suggested disturbed functional neural networks involved in reward, visual, autonomic, and neuroendocrine processes, with subtle but important differences between the bvFTD and SD patients. Specifically, in the bvFTD group, increased caloric intake was associated with loss of gray matter intensities in the bilateral cingulate gyri, thalami, bilateral lateral occipital cortex, lingual gyri, and right cerebellum but not in the orbital frontal cortex, suggesting that the altered eating behavior is not due to a loss of inhibitory control. In the SD group, similar brain regions were associated with caloric intake with the addition of bilateral orbitofrontal cortices, nucleus accumbens, and more left-sided structures involved in the semantic networks, suggesting a loss of knowledge concerning foods. Increased sucrose preference correlated with decreased gray matter intensities in bilateral orbitofrontal cortices, right-sided insula-striatal reward structures, including the nucleus accumbens, amygdala extending into the temporal occipital cortex, lingual gyrus, and cerebellum. This analysis combined both bvFTD and SD groups, as they had similarly increased sucrose preferences.
This important paper advances our understanding of the eating behavior changes in bvFTD and SD patients and supports the diagnostic value of assessing for hyperphagia in bvFTD. The paper suggests that differences in the behavioral abnormalities of bvFTD and SD patients can be explained by changes with the associated neural networks. Strengths include well-controlled study groups, prospective assessments using ecologically valid methods from obesity research, and detailed voxel-based morphometry analyses to identify for the first time the neural networks that correlate with the changes in eating behavior in FTD. A limitation is the interpretation of behavioral tests and voxel-based morphometry using a relatively small number of subjects. Additionally, caloric intake and sucrose preference correlated with disease severity, making it difficult to determine whether the results are specific to eating behaviors or simply reflect disease severity. Therefore, these results need to be validated by others. Additional studies using functional MRI also would be useful in determining if the brain regions that correlated with these eating behaviors are truly functionally connected and how alterations in these networks are linked to the behavioral changes. Delineating the neural networks involved in the eating behavior changes eventually could lead not only to improved diagnosis and treatment of these changes in FTD patients, but also to an overall better understanding of brain structures that control eating behavior. This would be clearly important in other medical conditions affected by eating behavior such as obesity and anorexia nervosa.