Neuroanatomical correlates of olfactory loss in normal aged subjects

Postmortem studies have described that olfactory loss observed in normal aging is associated with Alzheimer’s type brain degeneration. We hypothesized that distinct measures of gray and white matter integrity evaluated through magnetic resonance imaging (MRI) techniques could detect degenerative changes associated with age-related olfactory dysfunction. High-resolution T1-weighted images and diffusion-tensor images (DTI) of 30 clinically healthy subjects aged 51 to 77 were acquired with a 3-Tesla MRI scanner. Odor identification performance was assessed by means of the University of Pennsylvania Smell Identification Test (UPSIT). UPSIT scores correlated with right amygdalar volume and bilateral perirhinal and entorhinal cortices gray matter volume. Olfactory performance also correlated with postcentral gyrus cortical thickness and with fractional anisotropy and mean diffusivity levels in the splenium of the corpus callosum and the superior longitudinal fasciculi. Our results suggest that age-related olfactory loss is accompanied by diffuse degenerative changes that might correspond to the preclinical stages of neurodegenerative processes.


INTRODUCTION
There is a large body of evidence linking olfactory loss and neurodegenerative processes. Olfactory impairments are strongly associated with Parkinson's (PD) and Alzheimer's (AD) diseases [1,2], and have been investigated in animal models of AD [3].
In normal aging, the association between olfactory impairment and brain degeneration has been reported in a longitudinal clinicopathological study that included a large cohort of 471 elderly subjects (Rush Memory and Aging Project). During a mean follow-up of 2.2 years, autopsies were obtained from 122 of 166 subjects who died, revealing that scores in odor identification correlated with neuropathological changes usually associated with AD: density of neurofibrillary tangles in the entorhinal cortex and in the CA1 subfield of the hippocampus and the subiculum [4].
In a 5-year follow-up study including subjects from the same cohort, initial smell identification test scores were found to be associated with the risk of developing mild cognitive impairment (MCI). Moreover, in 34 subjects with olfactory dysfunction who died without cognitive deficits, autopsy showed greater burden of AD pathology [5].
The relationship between olfactory impairment and progressive cognitive decline was also seen in an epidemiological study involving 1,920 participants with a mean age of 66.9 years. In this study, authors reported an association between olfactory impairment and the incidence of MCI 5 years later with an odds-ratio of 6.62 [6].
The purpose of the current study was to investigate the cerebral correlates of impairments in odor identification in a sample of clinically healthy subjects and to correlate the performance in odor identification with measures of MRI cerebral degeneration such as cortical thickness, gray matter volumes and measures of white matter integrity. We hypothesized that olfactory dysfunctions in older persons could be associated with brain olfactory regions but also with other brain regions sensitive to the preclinical stages of degenerative processes.

Subjects.
The sample included 30 clinically healthy subjects (12 males; mean age: 66.0 + 7.4 years, range, 51-77 years; mean years of education: 11.1 + 4.2). All were right handed. All the participants were volunteers recruited from the Institut Català de l'Envelliment in Barcelona.
General exclusion criteria were: uncorrected visual or auditory deficits, drug abuse, and history of past or current psychiatric or neurologic disorder. Specific exclusion criteria for the olfaction test were: history of nasal bone fracture, diagnosis of rhinitis or nasal polyps, and upper respiratory tract infections in the 2 weeks prior to or at the moment of evaluation. Imaging exclusion criteria included any abnormality except mild white matter hyperintensities All selected participants completed a screening interview to check relevant medical information. One subject was currently a smoker, 7 had been smokers in the past and 22 had no history of smoking.  [12].
All participants underwent a comprehensive neuropsychological assessment.
The study was approved by the ethics committee of the University of Barcelona.
All enrolled subjects signed an informed consent form before taking part in the study.

Image acquisition and analysis
Magnetic resonance images were acquired with a 3T scanner (MAGNETOM Trio, Siemens, Germany). High-resolution 3-dimensional T1-weighted images were acquired in the sagittal plane (TR 2300 ms, TE 2.98 ms, TI 900 ms; 256 x 256 matrix, 1 mm isotropic voxel). Sagittal diffusion tensor images were obtained using a single-shot EPI sequence (TR 5533 ms, TE 88 ms), with diffusion-encoding in 30 directions at b=0 and 1000 s/mm 2 .
Voxelwise statistical analysis of fractional anisotropy (FA) and mean diffusivity (MD) data was carried out using TBSS (Tract-Based Spatial Statistics, [[16]]), part of FSL. First, FA and MD images were created by fitting a tensor model to the raw diffusion data using FDT, and then brain-extracted using BET [17]. All subjects' FA and MD data were then aligned into a common space using the nonlinear registration tool FNIRT (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/FNIRT) which uses a b-spline representation of the registration warp field [18]. Next, the mean FA and MD image were created and thinned to create a mean FA and MD skeleton which represents the centers of all tracts common to the group. Each subject's aligned FA and MD data were then projected onto this skeleton and the resulting data fed into voxelwise cross-subject statistics.

Structures used as regions of interest (ROIs) for GM analysis were chosen
based on literature about the olfactory system. The included structures were the amygdala, hippocampus, parahippocampal gyrus (encompassing entorhinal and perirhinal areas), olfactory portion of the orbitofrontal cortex, gyrus rectus and insula bilaterally. The Automated Anatomical Labelling (AAL) atlas [19] was used to create the corresponding masks.
Finally, voxelwise general linear model was applied using permutation-based non-parametric testing (5000 permutations) for FA, MD and GM analyses, correcting for multiple comparisons across space using familywise error correction (FWE).
The relationship between cortical thickness and UPSIT scores was assessed The automated procedure for volumetric measures of brain structures implemented in FreeSurfer was used to obtain the volumes of subcortical structures.

DISCUSSION
In this study, we found that olfactory loss in normal subjects correlated with several brain measures of gray and white matter integrity. We observed correlations of UPSIT scores with several structures involved in olfactory function such as the perirhinal and entorhinal cortices and the amygdala. Bitter et al. [8], using a VBM approach, described that subjects with anosmia of different etiology had gray matter reductions in primary as well as secondary olfactory regions. In the same line of our correlational data, they found gray matter volume reductions in anosmic patients in the parahippocampal gryrus, but they also found decrements in associative regions such as the medial prefrontal cortex. The widespread gray matter reductions they reported could be due to the fact that their subjects were anosmic (none of our subjects 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 10 presented anosmia) and that a more liberal threshold of significance (p<0.01, not corrected for multiple comparisons) was used. In their study, at the corrected level, only the medial/anterior cingulate region remained significant. The same group of research in a sample with less severe olfactory loss (hyposmic subjects) found gray matter reductions in the insular cortex, anterior cingulate cortex, orbitofrontal cortex, cerebellum, fusiform gyrus, precuneus, middle temporal gyrus and perirhinal cortex.
However, the results were also obtained at an uncorrected level of significance. The authors interpreted their findings as secondary to the reduced olfactory input, because a subsample of subjects with hyposmia of peripheral origin of the olfactory impairment has similar gray matter reductions. The samples of both studies included young and old people thus aging effects could not be isolated.
In addition to the classical structures of the olfactory system, we also found correlations of UPSIT scores with cortical thickness in the neocortex. Specifically, we found that right postcentral gyrus thinning is related with poor performance in olfaction or in other words that thicker cortex in this region is associated with better performance in odor identification. Curiously, this result is in agreement with the data obtained by Frasnelli et al (2010) in a sample of 46 healthy young university students. These authors found that a composite olfactory score positively correlated with the performance and cortical thickness in the right dorsal postcentral gyrus. Thus, it is possible that such relationship in our sample reflected individual differences in odor perception acquired during young ages rather than aging effects. DTI analyses have provided evidence of the association between degenerative brain changes and olfactory dysfunction in our sample. At the corrected level, we found that UPSIT scores positively correlated with FA scores and negatively with MD scores mainly involving the the corpus callosum and the superior longitudinal fasciculi. These results indicate that the loss of integrity of the cerebral fibers is related to loss of olfactory efficiency. The fibers that correlated with olfaction performance are not related 1  2  3  4  5  6  7  8  9  10  11  12  13  14  15  16  17  18  19  20  21  22  23  24  25  26  27  28  29  30  31  32  33  34  35  36  37  38  39  40  41  42  43  44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63  64  65 11 to the olfactory circuitry. In our opinion, this correlation reflects a common cause of degeneration for the olfactory system and other brain structures. Aging per se doesn't seem to be responsible for olfactory loss in our sample; we haven't observed any correlations between UPSIT scores and age. It is possible that a subsample of our subjects might instead be in the preclinical stage of a degenerative illness. This subsample could have both degeneration in limbic structures directly explaining the olfactory dysfunctions and subtle changes in the neocortical circuitry reflecting a more generalized degeneration.
There is wide evidence on the olfactory loss as a preclinical or prodromal sign of brain degeneration. Carriers of the CAG repeat expansion who are not yet diagnosed with Huntington's disease [26] have decreased UPSIT scores between 9 and 15 years before the estimated onset of the disease. Hyposmia is listed among the nonmotor symptom of PD (alongside constipation, daytime sleepiness, and rapid eye movement sleep behavior disorders and affective symptoms), and is also present in undiagnosed individuals at risk for PD (first-degree relatives of PD patients) [27], being considered a nonmotor feature which may precede by years the onset of motor disease [28][29][30]. Moreover, olfaction is impaired in PD patients with leucine-rich repeat kinase (LRRK2) G2019S mutations, and also in a subset of LRRK2 carriers without PD [31]. It has also been found that severe olfactory dysfunction is a prodromal symptom of dementia [32].
Hyposmia is seen in early stages of PD and in subjects with MCI [33,34].
Higher density of entorhinal cortex and hippocampal neurofibrillary tangles correlates with greater deficits in odor identification, suggesting a role for hippocampal dysfunction in Alzheimer's disease hyposmia [4]. Olfactory dysfunctions are predictive of cognitive decline in the elderly and interact with the ApoEε4 allele [35][36][37].