Sex and age differences in atrophic rates: an ADNI study with n=1368 MRI scans

Abstract

We set out to determine factors that influence the rate of brain atrophy in 1-year longitudinal magnetic resonance imaging (MRI) data. With tensor-based morphometry (TBM), we mapped the 3-dimensional profile of progressive atrophy in 144 subjects with probable Alzheimer's disease (AD) (age: 76.5 +/- 7.4 years), 338 with amnestic mild cognitive impairment (MCI; 76.0 +/- 7.2), and 202 healthy controls (77.0 +/- 5.1), scanned twice, 1 year apart. Statistical maps revealed significant age and sex differences in atrophic rates. Brain atrophic rates were about 1%-1.5% faster in women than men. Atrophy was faster in younger than older subjects, most prominently in mild cognitive impairment, with a 1% increase in the rates of atrophy and 2% in ventricular expansion, for every 10-year decrease in age. TBM-derived atrophic rates correlated with reduced beta-amyloid and elevated tau levels (n = 363) at baseline, baseline and progressive deterioration in clinical measures, and increasing numbers of risk alleles for the ApoE4 gene. TBM is a sensitive, high-throughput biomarker for tracking disease progression in large imaging studies; sub-analyses focusing on women or younger subjects gave improved sample size requirements for clinical trials.

Figure 1

Age and sex differences in atrophic rates are shown across the entire brain and also in an analysis restricted to changes within the temporal lobes. CDF plots for the effects on atrophic rates of age (a) and sex (b) show the statistical significance of their correlations with atrophy rates. Both effects were most prominent in the MCI group, probably because it had the most subjects. CDF plots for the whole brain were reflected in the y axis to avoid clutter. CDF curves that rise more steeply at the origin generally indicate greater effect sizes. Tests for age and sex effects throughout the whole brain showed inferior statistical power relative to similar tests for effects inside the temporal lobes. This can be seen by comparing CDF curves of the same color in (a) and (b). Regression coefficient maps are shown for the age (c) and sex (d) effects in MCI across the entire brain; colors show the signs of the regression coefficients. The map of each effect (age, sex) is adjusted for the effect of the other covariate. Younger age (↓) is associated with faster tissue loss rates (↓) and faster ventricular expansion (↑). There is approximately a 1% increase in atrophic rates and 2% increase in ventricular expansion rates, for every 10-year decrease in age, as shown by positive correlations in the temporal lobes and negative correlations in the CSF, respectively. Women (↓) had faster brain degeneration (↓)by about 1–1.5%/year relative to men. These correlation coefficient maps show only the values in regions demonstrating significant correlations, after FDR correction across the entire brain. C.P.: critical P-value. n.s.: not significant

Figure 2

Average maps of atrophic rates in MCI subjects, subdivided by age and sex. Female MCI subjects (top) are divided into three age groups, 60–70 (N=24), 70–80 (N=59), and 80–90 years (N=37). Male MCI subjects (bottom) are divided into the age groups of 60–70 (N=33), 70–80 (N=102), and 80–90 years (N=77). Faster atrophic rates occur (darker blue) in younger subjects, and in women versus men; age and sex effects are clearly visible. A small number of MCI subjects (N=6) fell outside of these age ranges but were too few to form a separate sample so they are not included in the maps.

Figure 3

Whole brain and temporal lobe atrophic rates are correlated with baseline clinical measures in AD (a) and MCI (b). Significant correlations are marked with a critical P value greater than 0.01 or 0.0001. Interestingly, the ADAS-Cog, perhaps the most widely used cognitive measure in clinical trials, was most strongly correlated at baseline with future atrophic rates, in both AD and MCI groups. Again, the CDF curves for the temporal lobe data tend to rise more sharply at the origin, compared to curves computed from all the voxels in the brain. This is visually evident when comparing curves of the same color on each side of the plot. C.P.: critical P-value

Figure 4

Whole brain and temporal lobe atrophic rates correlated with rates of clinical decline, for various different clinical measures, in AD (a) and MCI (b) groups separately. Significant correlations are marked with critical P>0.01 or >0.0001. CDR-SB is the measure examined here whose change over time was the most highly correlated with atrophic rates in MCI. Comparison of curves of the same color on either side of the y-axis shows that, in general, the analyses of the temporal lobe voxels give higher effect sizes (steeper CDFs) than those that include all the voxels in the brain. C.P.: critical P-value.

Figure 5

Correlations between atrophic rates and CSF biomarker levels (biomarker and clinical labels are with and without borders, respectively). Whole brain and temporal lobe atrophic rates were correlated with biomarker levels in the following rank order, from strongest to weakest correlations: Aβ, tau/Aβ, p-tau, and tau, in CSF at baseline, in the combined group of all control, MCI, and AD subjects (blue CDF curves). CSF Aβ levels were weakly correlated with atrophic rates (critical P=0.004 in the temporal lobes and 0.001 in the whole brain) in the MCI group (cyan curves). The amount of change in tau/Aβ over 12-month was weakly linked to brain atrophic rates. Clinical correlations, computed in the common dataset, were compared with the results from CSF biomarkers. In this common subsample, baseline ADAS-cog and CDR-SB (labeled as CDR) rates of decline are more strongly correlated with structural brain atrophy, as indicated by higher CDF curves and critical P-values compared to CSF biomarker correlations. All other correlations, with baseline and longitudinal measures of CSF biomarkers, were not significant (not shown in the graph). C.P.: critical P-value.

Figure 6

Genetic influences on brain atrophy. The presence of the ApoE4 (marked by solid lines) and the GRIN2b risk gene (also known as SNP rs-10845840; dotted lines; Stein et al., 2010) were associated with faster rates of atrophy in the temporal lobes, with ApoE4 showing a greater statistical influence than GRIN2b, indicated by the rank order of the CDFs and by the FDR critical P values. When expanding the search region to the whole brain, the presence of the ApoE4 risk allele was no longer associated with higher atrophic rates in individual diagnostic groups, but the effect remained significant in the combined group. Risk genes were coded as 0, 1, and 2 for zero, 1 risk allele, and 2 adverse alleles respectively; association tests assumed an additive model of gene action. Critical P-values are shown for each test. n.s.: not significant.

Figure 7

MCI converters showed faster rates of brain atrophy in temporal lobes than MCI non-converters. The mean difference map shows regions where atrophy rates are faster in converters than non-converters (left panel; blue colors: 3% faster). Red colors show regions where ventricular expansion is faster in converters than non-converters. The right panel shows the FDR corrected p maps displaying regions with significant difference overall (critical P=0.002).