Conceived and designed the experiments: LO AB ME HH. Performed the experiments: LO ME AB. Analyzed the data: LO FL. Contributed reagents/materials/analysis tools: LO CT BM DP. Wrote the paper: LO. Clinical diagnosis: YF MB TC DRC DO.
The authors have declared that no competing interests exist.
Previous PET and MRI studies have indicated that the degree to which pathology translates into clinical symptoms is strongly dependent on sex with women more likely to express pathology as a diagnosis of AD, whereas men are more resistant to clinical symptoms in the face of the same degree of pathology. Here we use DTI to investigate the difference between male and female white matter tracts in healthy older participants (24 women, 16 men) and participants with mild cognitive impairment (21 women, 12 men). Differences between control and MCI participants were found in fractional anisotropy (FA), radial diffusion (DR), axial diffusion (DA) and mean diffusion (MD). A significant main effect of sex was also reported for FA, MD and DR indices, with male control and male MCI participants having significantly more microstructural damage than their female counterparts. There was no sex by diagnosis interaction. Male MCIs also had significantly less normalised grey matter (GM) volume than female MCIs. However, in terms of absolute brain volume, male controls had significantly more brain volume than female controls. Normalised GM and WM volumes were found to decrease significantly with age with no age by sex interaction. Overall, these data suggest that the same degree of cognitive impairment is associated with greater structural damage in men compared with women.
Many studies have found that the clinical symptoms of Alzheimer’s disease (AD) are not tightly linked to brain pathology
While AD usually entails significant memory loss, an intermediate state between healthy ageing and AD, termed mild cognitive impairment (MCI) has become widely recognized and a significant proportion of those with MCI may represent a prodromal state of AD
Diffusion tensor imaging (DTI) offers a powerful tool for the investigation of WM
In addition to the scarcity of DTI studies examining sex differences in the WM of MCI and AD patients, there are also very few studies which have examined WM differences between healthy older men and women. Studies in younger cohorts have shown that men have higher FA than women in the corpus callosum
The incidence of MCI and AD dementia in men and women has also been subject to much debate. For those greater than 80 years of age, increased incidence of AD dementia for women may be due to the greater longevity of women versus men. Nevertheless, even when correcting for age, women appear to be more affected by dementia than men
In elderly cohorts, sex differences in brain areas that are essential to higher cognitive functions have been found, with one study indicating that age-related cerebral volume loss and glucose metabolic deficit were significantly different between men and women
In the current study, tract based spatial statistics (TBSS) was used to assess sex differences in healthy controls and MCI participants. TBSS obviates the need for
The study was approved by the St. James’ Hospital and Adelaide & Meath Hospital incorporating the National Children’s Hospital Research Ethics Committee and was in accordance with the Declaration of Helsinki. All participants provided informed written consent.
Scans were obtained from the following groups of participants: 40 healthy older people, 33 MCI participants. The total number of participants was 73. MCI patients were diagnosed using criteria for both amnestic and non-amnestics sub-groups
objective impairment on any neuropsychological test from the CERAD battery based on a cut-off of −1.5 SD below published normative data corrected for age and education of the subject;
cognitive impairment corroborated by a close family member;
essentially normal activities of daily living;
must not meet criteria for dementia as defined below.
MCI individuals with objective memory impairment were diagnosed as having MCIa and those with non-memory impairment were diagnosed as having MCIna.
MCI participants were recruited at the Adelaide and Meath Hospital incorporating the National Children’s Hospital (AMNCH), Dublin, Ireland. Healthy participants were recruited among relatives of patients and also through advertisements in the local community.
Participants were excluded if they had cortical infarction, excessive subcortical vascular disease, space-occupying lesions, depression, and any other psychiatric or neurological disease. Participants were also excluded on magnetic resonance imaging criteria such as pacemaker implant, recent metallic implants, and claustrophobia. The DTI and structural scans of the cohort used in the current study were previously used in a study of mixed-effects models
In the healthy older control group there was no significant difference between men and women in terms of age, education, and all of the psychological tests, except for Praxis Recall where men scored significantly better than women (
n | Age | Education | MMSE | Boston | ||||||
Male | Female | Male | Female | Male | Female | Male | Female | Male | Female | |
Con | 16 | 24 | 68.31 (8.41) | 64.08 (6.81) | 14.43 (6.91) | 12.14 (2.00) | 29.25 (1.44) | 29.33 (0.82) | 14.50 (0.73) | 17.79 (0.41) |
MCI | 12 | 21 | 71.33 (6.97) | 66.24 (6.46) | 11.92 (3.48) | 13.14 (3.00) | 27.75 (3.02) | 28.33 (3.34) | 12.67 (1.87) | 12.67 (1.68) |
|
|
|
|
|
||||||
|
|
|
|
|
|
|
|
|
|
|
Con | 16 | 24 | 7.50 (0.92) | 7.39 (1.17) | 8.19 (1.22) | 8.50 (1.29) | 10.75 (0.58) | 10.54 (0.78) |
|
|
MCI | 12 | 21 |
|
|
|
|
9.75 (1.48) | 9.62 (1.69) | 8.67 (3.08) | 9.43 (3.75) |
Values are mean (SD). Abbreviations: Con, control; MCI, Mild cognitive Impairment; MMSE, Mini-Mentral State Examination. Significant differences between male and female groups are marked in bold where p<0.05 using a t-test.
For the MCI group there was no significant difference between men and women in terms of age, education, MMSE, or the Boston test. Female MCI participants scored significantly better than males MCI participants in Word List Average, Word Recall, Praxis and Praxis Recall (
Magnetic resonance imaging (MRI) was conducted with a Philips Achieva 3.0 Tesla MR system (Best, The Netherlands). A parallel SENSitivity Encoding (SENSE) approach
DTI was acquired using an echo planar imaging (EPI) sequence with the following pulse sequence: TR = 12396 ms; TE = 52 ms; acquisition voxel size = 2×2×2 mm3; rec voxel size = 1.75×1.75×2 mm isotropic, 60 axial adjacent slices; slice thickness = 2 mm (no gap); FOV = 224 mm×224 mm×120 mm; acquisition matrix = 112×112; SENSE reduction factor = 2, combined with a half-scan acquisition; 1 image without diffusion weighting and 15 diffusion-encoding gradients applied in 15 noncollinear directions; b-value = 800 s/mm2; both the b0 and the 15 diffusion weighted images were averaged twice; bandwidth = 2971 Hz/pixel; total acquisition time = 7 min 34 sec.
A T2-weighted fluid attenuation inversion recovery (FLAIR) sequence was also acquired to ensure that vascular pathology was not significant. All images were rated using the Fazeka scale
Images were skull stripped with the Brain Extraction Tool (BET) from the FSL library
DTI analysis was performed using TBSS
TBSS performs a non-linear registration which aligns each FA image to every other one and calculates the amount of warping needed for the images to be aligned. The most representative image is determined as the one needing the least warping for all other images to align to it. The FSL library also provides a 1 mm isotropic FA target image (FMRIB58_FA) in standard space which sometimes used instead of the most representative image from the study cohort. This can be problematic as the target image is based on a young healthy brain. Using the method of “all subject to all subject” registration is more computationally intensively, but highly desirable when dealing with populations other than young healthy controls.
After this registration step, warped versions of each subject’s FA image were generated which were then averaged and a white matter “skeleton” is created suppressing all non-maximum FA values in each voxel’s local-perpendicular direction and subsequently comparing all remaining non-zero voxels with their nearest neighbours, thus searching for the centre of fibre bundles. The skeleton was then thresholded at an FA value of 0.2 which limits the effects of poor alignment across subjects and ensures that GM and CSF voxels are excluded from the skeleton. The skeleton that is now created contains WM tracts that are common to all subjects. A “distance map” is then created which is used to project each FA image onto the mean FA skeleton that is common to all subjects
For statistical analysis, the images were analyzed using the “randomise” tool from FSL with a standard GLM design which controlled for the effects of age by including it as a covariate in the GLM design. The GLM also included additional explanatory variables which controlled for the differences between groups in terms of scores from the following tests: Word List Average, Word Recall and Praxis Recall. The GLM examined the main effect of sex and diagnosis, as well as a sex by diagnosis interaction. Pairwise differences between diagnostic groups, and also between subgroups (e.g. male controls versus male MCIs) were calculated. Randomise computes permutation tests on the assumption that the null hypothesis implies complete exchangeability of the observations
An ANOVA test for normalised GM modelled against diagnosis by sex (i.e. normalised GM ∼ diagnosis × sex) returned a significant main effect of sex (F(1,68) = 7.86, p = 0.0066) and a significant main effect of diagnosis (F(1,68) = 7.87, p = 0.0066). No significant age by sex interaction was found. A post-hoc Tukey test found that normalised male MCI GM was significantly lower than normalised female control GM (p = 0.00085) (
(A) No significant differences in white matter volume were found between males and females for control or MCI conditions. (B) Male MCI subjects had significantly lower GM volume relative to male controls. *** p<0.001, with post-hoc Tukey test, following an ANOVA.
For an ANOVA test of normalised WM ∼ diagnosis × sex, no significant main effect of diagnosis or sex was found, and no significant interaction was found between these two factors (
For absolute brain volume (i.e. non-normalised, gross brain volume), an ANOVA test returned a significant main effect of sex on absolute WM volume (F(1,68) = 18.88, p<0.0001) for the model WM volume ∼ diagnosis × sex. No significant main effect of diagnosis and no diagnosis by sex interaction was found. A post-hoc Tukey test, reported significant differences between male controls and female controls (p = 0.001), and also between male controls and female MCIs (p = 0.003) (
(A) Absolute male control WM volume is significantly larger than absolute female control white matter volume. Absolute male control WM volume is also significant larger than absolute female MCI WM volume. (B) Absolute male control GM volume is significantly larger than absolute female control grey matter volume. Absolute male control GM volume is also significant larger than absolute female MCI WM volume. *p<0.05, **p<0.01, *** p<0.001, with post-hoc Tukey test, following an ANOVA.
For absolute GM volume, an ANOVA test returned a significant main effect of sex (F(1,68) = 8.4958, p = 0.0048). The main effect of diagnosis on absolute GM volume was found to be borderline significant (F(1,68) = 3.91, p = 0.052). There was no significant interaction between diagnosis and sex. A post-hoc Tukey test revealed a significant difference between male and female controls (p = 0.022). There was also a significant difference between male controls and female MCIs (p = 0.0034) (
Control participants were found to have significantly less FA values compared with MCI (
FA (green) was significantly decreased in MCI subjects relative to controls, while DR (yellow), MD (blue) and DA (red) were significantly increased in MCI subjects relative to controls. The plots show values for each index of diffusion taken from the regions of the WM identified in the TBSS images. The TBSS images show results at p<0.05 corrected for multiple comparisons. The leftmost images show axial slices (z = 88), images in the central panel show coronal slices (y = 109) and the rightmost images show sagittal slices (x = 112).
For DA, DR and MD, these indices of absolute diffusion were all found to be significantly increased in MCI relative to controls (
The general linear model applied to TBSS results indicated a significant main effect of sex for FA, DR and MD, with males having significantly decreased FA and significantly increased DR and MD relative to women in the anatomical regions indicated in
Top: Male subjects have lower FA than female subjects in the regions indicated in green. The plot shows the FA values in the significant TBSS regions highlighted in green. Middle: Male subjects have greater DR than female subjects in the regions indicated in yellow. The plot shows the DR values in the significant TBSS regions highlighted in yellow. Lower: Male subjects have greater MD values in the significant TBSS regions highlighted in blue. The plot shows the MD values in the significant TBSS regions highlighted in blue. No significant main effect of sex was found for the DA index. The TBSS images show results at p<0.05 corrected for multiple comparisons. For the FA images the axial slice is located at z = 88, the coronal slice is located at y = 89 and the sagittal slice is located at x = 93. For the DR images the axial slice is located at z = 97, the coronal slice is located at y = 86 and the sagittal slice is located at x = 65. For the MD images the axial slice is located at z = 101, the coronal slice is located at y = 84 and the sagittal slice is located at x = 66. Abbreviations in the boxplots: M: male; F: female.
As can be seen in
The GLM returned a significant main effect of age. FA values decreased significantly and DR, MD and DA values decreased significantly in the regions indicated in
FA (green) was found to decrease significantly with age. The plot of FA versus age utilises FA values taken from the green regions highlighted. DR (yellow) was found to increase significantly with age. The plot of DR versus age utilises DR values taken from the yellow regions highlighted. MD (blue) was found to increase significantly with age. The plot of MD versus age utilises MD values taken from the blue regions highlighted. DA (red) was found to increase significantly with age. The plot of DA versus age utilises DA values taken from the yellow regions highlighted. The leftmost images show axial slices (z = 88), images in the central panel show coronal slices (y = 109) and the rightmost images show sagittal slices (x = 112).
An ANOVA test for normalised GM volume ∼ age × sex × diagnosis, returned a significant main effect of age (F(1,64) = 34.16, p<0.00001), and a significant main effect of diagnosis (F(1,64) = 5.89, p = 0.018). No significant interactions were returned. A regression analysis for normalised GM volume decreases with age is shown in
An ANOVA test for normalised WM volume ∼ age × sex × diagnosis returned a significant main effect of age (F(1,64) = 24.25, p<0.00001). No significant interactions were returned. Similar to normalised GM volume, normalised WM volume decreases significantly with age (r = −0.50, p<0.001) (
An ANOVA model with normalised GM volume as the dependent variable and age, sex and diagnosis as the independent variables, returned a significant main effect of age, and no significant interactions. Here, a regression line for GM decline with age (including men and women) is plotted (left). An ANOVA model with normalised WM volume as the dependent variable and age, sex and diagnosis as the independent variables, also returned a significant main effect of age, with no significant interactions. A regression line for WM decline with age (including men and women) is plotted (right). The r value denotes the correlation as calculated by Pearson’s product-moment correlation using data from all subjects. Correlations are significant as indicated by the p values.
To the best of our knowledge, this is the first study to examine sex differences in those with MCI in terms of DTI measures of WM structural integrity. Our initial findings of significantly lower FA, and significantly higher DR, MD and DA in MCI participants relative to healthy older controls are broadly consistent with previous DTI studies
There have been extremely few studies investigating WM differences between men and women in older populations, but those that exist have reported a limited degree of sexual dimorphism
Previous PET studies in AD dementia patients have noted that men have a pronounced hypermetabolism relative to women, particularly in the right hemisphere
While the underlying reasons for this laterality are not known it is of interest that a model of aging, termed HAROLD (hemisphere asymmetry reductions in older adults) posits that there is an age-related loss of hemispheric asymmetry of task-related activation in the prefrontal cortex (PFC). While healthy young subjects show greater activation in left PFC compared with the right PFC during the execution of a variety of cognitive tasks, many studies indicate that this laterality is reduced in older subjects
Another possible explanation for the overall greater structural damage in men may be that brain reserve is greater in older men compared with older women. Although men have more structural damage than women, they are at the same cognitive level as women, suggesting that women may manifest the symptoms of cognitive decline at a level of structural damage which would not cause the same amount of cognitive decline in men. This can be seen in the scheme proposed in
Healthy older subjects (male and female) have an MMSE of ∼29. However, at an equivalent MMSE score, healthy older women (1) have less microstructural damage than healthy older men (2). As we enter the MCI spectrum, female MCIs (3) again have less microstructural damage than male MCIs (4). It is worth noting that the level of microstructural damage of female MCI subjects in the current study was equivalent to the level of damage found in male control subjects.
Brain reserve refers to the fact that larger brains can sustain more structural damage before clinical deficits emerge because sufficient neural substrate for normal function is retained in larger brains following an insult
In the current study, DA was found to increase in MCI, a finding which is at odds with animal studies and human callosotomy studies which have shown that axonal pathology results in a decrease in axial diffusion
WM and GM volumes were also found to have sex specific profiles. Male MCI participants showed significantly lower normalised GM volume than female MCI participants. Again, this suggests that for an equivalent diagnosis of MCI, men must sustain more GM atrophy than women in order to be diagnosed as MCI. Therefore, the threshold of structural damage which precipitates cognitive decline leading to a diagnosis of MCI appears to be lower in women. No sex-specific differences were found in normalised WM volume which suggests that while there are significant WM microstructural changes in male MCI participants, this has not reached the point of gross tissue loss. This result also highlights the fact that DTI has the ability to detect WM microstructural changes which may develop before global WM changes occur. This is particularly true in MCI, where the structural changes are subtle. In an earlier study we also noted diffusion changes in healthy older, MCIna and MCIa subjects, in the absence of significant changes in normalised WM volume
Absolute WM and GM volume were both greater in male controls relative to female controls, which is consistent with previous studies
Although there have been conflicting results, the majority of studies also indicate that the incidence of AD dementia is greater in women than in men
Genetic and hormonal characteristics might contribute to the extra vulnerability of older women. Oestrogen has been shown to have a potent neuroprotective effect, and it is possible that there is a profile of increased protection from neurodegeneration while circulating oestrogen levels are normal, followed by increased risk of neurodegeneration in older women as oestrogen levels decline following menopause
In addition to brain reserve, the role of cognitive reserve should also be noted. While brain reserve is a passive model that highlights the role of brain size or neuronal volume, cognitive reserve is an active model which suggests that the brain responds to brain damage by recruiting existing cognitive processes or by enlisting compensatory processes
Strikingly, Barnes and colleagues recently noted, using data from the Religious Orders Study, that the association between AD pathology and clinical AD was significantly stronger in women than in men
In terms of the rate of deterioration of WM microstructure with age (as measured by FA, DA, MD and DR), no age by sex interaction was found. The areas implicated in microstructural deterioration with age were dispersed through the brain, with a slightly more pronounced effect in frontal regions which is consistent with the concept of an anterior-posterior gradient of WM deterioration with age
Similarly, the rate of decline of GM with age was not found to be significantly influenced by sex which is also consistent with earlier work that has shown similar rates of GM volume decline in men and women which agrees
A limitation of the current study is the relatively small number of participants included. A study with a larger sample size of MCI participants is warranted to verify and further probe the impact of sex on mild cognitive impairment. This is being pursued as part of the European DTI Study in Dementia (EDSD) initiative. The small size of the current sample may weaken the ability to reliably identify interactions between sex and age, as well as, interactions between sex and diagnosis. It is now planned to assess these interactions in an independent sample with larger numbers of participants.
The cross-sectional nature of this study should be noted. We do not have follow-up data and thus do not know which participants subsequently developed AD or alternatively remained stable without deteriorating further. Therefore, the heterogeneity of the current cohort and the lack of diagnostic certainty in the cohort should be borne in mind. Future studies should also consider sex differences between those with MCI that remain stable and those that deteriorate to AD dementia. Technical limitation of the methodology should also be noted. Although TBSS strives to avoid the problems of voxel based morphology relating to partial voluming, some of these issues may still remain. Small WM tracts may be contaminated with GM if the tract width is smaller than the original voxel size, although applying a threshold of 0.2 or less to the WM skeleton in the TBSS preprocessing steps is thought to remove the potential occurrence of GM.
Overall our results show that in both control and MCI groups, men have greater WM damage than women with the same clinical diagnosis. This suggests that the threshold of WM damage that must be reached before entering the MCI spectrum is lower in women compared with men. The identification of older people with mild cognitive impairment that are at high risk of evolving to AD is vitally important for early treatment. The concept of MCI is now widely used to describe this high-risk group, and numerous research programmes have been undertaken with the aim of providing therapeutic intervention at the MCI stage which might reduce or prevent the possibility of progressing on to AD dementia. The current results indicate that sex differences need to be carefully considered when developing therapeutic strategies for the treatment of AD. While sex may have been overlooked in the past by simply including sex matched groups, the current study points to a need to take sex into account in all drug trials which seeks to target AD dementia and reduce its incidence in the population.
We thank Céline Bourdon and Felix Geser for comments, discussion and assistance relating to the manuscript. We thank Brendan Cody-Kenny for assistance with programming. We also thank the Trinity Centre for High Performance Computing for assistance with computing and programming on their supercomputing facilities.