Participants
A total of 182 cognitively normal 40–65 year-old women were enrolled. Nine participants were excluded due to neuropathological conditions encountered in the MR images (n = 7; e.g. neoplastic condition, aneurysm), or due to artifacts (n = 2). Three participants with incomplete clinical data and 9 with unclear menopause status were also excluded. Our final sample consisted of 161 women, including 30 pre-menopausal (PRE), 57 peri-menopausal (PERI), and 74 post-menopausal (POST) participants. There were no group differences in demographic or cognitive measures except for an age difference between PRE and POST groups (Supporting Information Table 1). Our procedures to address age effects are described in the Methods, and included comparing each MT group to an age-matched male group: 30 males age-matched to PRE (MALEPRE), 50 age-matched to PERI (MALEPERI), and 45 age-matched to POST (MALEPOST; Supporting Information Table 2). 42% of participants were APOE-4 positive, with comparable distributions between groups (Supporting Information Tables 1 and 2).
Biomarker differences between MT groups
We examined MT-stage effects on a panel of biomarkers examining brain (a) structure: MRI-based gray and white matter volume (GMV, WMV) and fractional anisotropy (FA, an index of WM integrity and structural connectivity12); (b) energy metabolism: 18F-FDG PET glucose metabolism (CMRglc), ASL cerebral blood flow (CBF), and 31P-MRS ATP production; and (c) 11C-PiB-PET Aβ deposition. Unless otherwise specified, all results are significant at p < 0.05, cluster-level corrected for family-wise type error (FWE), adjusted for age, APOE-4 status, and modality-specific confounders.
Structural biomarkers
GMV: MT stage differences were observed in inferior temporal gyrus, precuneus, and fusiform gyrus of the right hemisphere (Fig. 1a). On post-hoc analysis, the temporal cluster reflected lower GMV in the POST group compared to the PRE group, while the precuneus and fusiform clusters reflected lower GMV in the PERI group compared to the POST group (Fig. 1b and Supporting Information Table 3).
WMV: Significant MT stage differences were observed in anterior and posterior corona radiata (Fig. 1c). On post-hoc analysis, these clusters reflected lower WMV in the POST group vs. PRE and PERI groups (Fig. 1d and Supporting Information Table 4).
FA: The only region that survived FWE correction was a cluster of lower FA in the right external capsule of POST compared to PERI (Fig. 1e, f, g, and Supporting Information Table 5).
Metabolic biomarkers
CMRglc: Significant MT stage differences were observed in supramarginal gyrus, middle and inferior temporal gyri (Fig. 2a). On post-hoc analysis, all clusters reflected lower CMRglc in POST vs. PRE and PERI groups (Fig. 2b and Supporting Information Table 6). The PERI group also showed lower CMRglc than PRE in right middle temporal gyrus (Fig. 2b and Supporting Information Table 6).
CBF: Significant MT stage differences were observed in supramarginal gyrus, middle and superior temporal gyrus, superior and inferior frontal gyrus (Fig. 2c). On post-hoc analysis, these clusters reflected higher CBF in POST vs. PERI (Fig. 2d and Supporting Information Table 7). Figure 2e displays the regional overlap of CMRglc and CBF effects.
ATP production: We examined regional ATP to phosphocreatine (PCr) ratios in parieto-temporal regions showing MT-stage effects on CMRglc and CBF (Table 1). Multivariate general linear models (mGLM) revealed significant ATP/PCr group differences (Wilk’s Lambda p = 0.030), which on post-hoc examination, were driven by POST exhibiting higher ATP/PCr in temporal regions (p’s < 0.047) and borderline higher ATP/PCr in parietal regions (p’s ≤ 0.12) compared to PRE. The PERI group showed intermediate ATP/PCr levels. Results were unchanged including APOE status as a covariate (Table 1).
Table 1
31P-MRS ATP/PCr measures by menopausal stage group
|
Pre-menopausal group
|
Peri-menopausal group
|
Post-menopausal group
|
Superior temporal gyrus
|
1.11(0.05)
|
1.18(0.03)
|
1.26(0.03)*
|
Adjusted by APOE status
|
1.14(0.04)
|
1.18(0.03)
|
1.24(0.02)*
|
Middle temporal gyrus
|
1.09(0.04)
|
1.14(0.02)
|
1.20(0.02)*
|
Adjusted by APOE status
|
1.09(0.03)
|
1.13(0.02)
|
1.19(0.02)*
|
Inferior temporal gyrus
|
1.07(0.04)
|
1.11(0.02)
|
1.13(0.02)*
|
Adjusted by APOE status
|
1.07(0.03)
|
1.10(0.02)
|
1.14(0.02)*
|
Inferior parietal lobule
|
0.77(0.06)
|
0.76(0.05)
|
0.83(0.04)
|
Adjusted by APOE status
|
0.78(0.04)
|
0.75(0.03)
|
0.83(0.03)*
|
Values are age-adjusted means (SE). *Different from pre-menopausal group, p < 0.05. |
Amyloid-β load
There were no significant differences in PiB uptake between MT groups (Table 2). However, adding APOE-4 status (positive vs. negative) as a covariate enhanced group differences, which reached significance for POST vs. PRE (mean difference, 27%, p = 0.005), and for PERI vs. PRE groups (18%, p = 0.048).
Table 2
PiB-PET amyloid-β load by menopausal stage group
|
Pre-menopausal group
|
Peri-menopausal group
|
Post-menopausal group
|
AD-mask SUVR
|
1.03(0.12)
|
1.20(0.06)
|
1.27(0.07)
|
adjusted by APOE status
|
1.00(0.09)
|
1.21(0.06)
|
1.29(0.06)*
|
AD-mask SUVR by APOE-4 status
|
APOE-4-
|
0.94(0.12)
|
1.12(0.07)
|
1.32(0.07)*
|
APOE-4+
|
1.07(0.12)
|
1.39(0.11)*
|
1.25(0.08)*
|
PiB scans, % positive
|
5
|
16
|
24*
|
PiB scans, % positive by APOE-4 status
|
APOE-4-
|
0
|
10
|
22
|
APOE-4+
|
9
|
31*
|
26*
|
Values are age-adjusted means (SE), unless otherwise specified. *Different from pre-menopausal group, p < 0.05. |
Abbreviations: APOE-4-, APOE-4 non-carriers; APOE-4+, APOE-4 carriers; SUVR, standardized uptake value ratio to cerebellar gray matter PiB uptake. |
Twenty (18%) female participants undergoing PiB-PET imaging were brain amyloid positive. The frequency of PiB-positive scans was borderline higher in POST vs. PRE (24% vs. 5%; Pearson Chi Square p = 0.056), and intermediate in PERI (16%, n.s.; Table 2). Among APOE-4 carriers, the frequency of PiB-positive scans was higher in POST (26%) and PERI (31%) vs. PRE (9%; Pearson Chi Square p = 0.049). Among non-carriers, the frequency of PiB-positive scans was highest in POST (22%), intermediate in PERI (10%), and lowest in PRE (0%), which did not reach significance.
Comparisons to age-matched males
Unless otherwise specified, all results are significant at p < 0.05, cluster-level FWE corrected, adjusted for modality-specific confounders.
Structural biomarkers
GMV: Results are presented in Fig. 3a and Supporting Information Table 8. Analyses confirmed presence of lower GMV in inferior temporal gyrus of the POST group vs. MALEPOST, and of lower GMV in precuneus and fusiform gyrus of the PERI group vs. MALEPERI. In the entire search volume, both POST and PERI exhibited widespread areas of lower GMV compared to males, involving inferior, middle and medial temporal (MTL) regions, fusiform gyrus, superior, middle, and orbital frontal gyrus, anterior cingulate (ACC), insula, and putamen. Additionally, the PERI group exhibited lower GMV in precuneus vs. MALEPERI. The PRE group showed less extended areas of lower GMV compared to vs. MALEPRE involving MTL, frontal gyrus, putamen, and temporal cortex.
WMV: Results are presented in Fig. 3b and Supporting Information Table 9. Analyses confirmed presence of lower WMV in corona radiata of the POST group vs. MALEPOST, and revealed additional areas of lower WMV in several tracts (Fig. 3b). The PERI group exhibited a similar pattern of lower WMV compared to vs. MALEPERI, and the PRE group showed lower WMV in internal capsule vs. MALEPRE.
FA: Results are presented in Fig. 3c and Supporting Information Table 10. The POST group exhibited no regions of lower FA vs. MALEPOST, and showed instead higher FA in corona radiata and fornix. The PERI group exhibited higher FA than MALEPERI in corona radiata, fornix, external capsule and uncinate fasciculus; and lower FA in superior longitudinal fasciculus (SFL) and posterior thalamic radiation. The PRE group also showed higher FA in corona radiata, and lower FA in SLF, compared to vs. MALEPRE.
Metabolic biomarkers
CMRglc: Results are presented in Fig. 3d and Supporting Information Table 11. Analyses confirmed presence of lower CMRglc in temporo-parietal areas of the POST group vs. MALEPOST, and revealed additional areas of hypometabolism in superior and middle frontal cortex. The PERI group showed lower CMRglc in frontal cortex and insula vs. MALEPERI, and the PRE group showed lower CMRglc in left supramarginal gyrus vs. MALEPRE. All results were significant at p < 0.05, FWE-corrected.
CBF: Results are presented in Fig. 3e and Supporting Information Table 12. Analyses confirmed presence of elevated CBF in frontal regions of the POST group, and revealed areas of lower CBF in ACC, MTL, and basal ganglia vs. MALEPOST. The PERI and PRE groups also had areas of higher and lower CBF compared to males. Both groups exhibited lower CBF than males in ACC, MTL, superior temporal gyrus, and basal ganglia, and higher CBF in parietal lobe. All results were significant at p < 0.05, FWE-corrected.
ATP production: We examined ATP/PCr using two separate mGLMs: model 1 included regions with higher CBF in POST vs. vs. MALEPOST (ACC, temporo-parietal regions, and insula); model 2 included regions with lower CBF and FDG in POST vs. vs. MALEPOST (frontal areas). As shown in Table 3, the POST group had higher ATP/PCr than MALEPOST in model 1 (Wilk’s Lambda p = 0.003, APOE-adjusted p = 0.002) and borderline higher ATP/PCr in model 2 (Wilk’s Lambda p = 0.062, APOE-adjusted p = 0.082). The PERI group had higher ATP/PCr than MALEPERI in model 2 (Wilk’s Lambda p = 0.049, APOE-adjusted p = 0.023). There were no differences between PRE and vs. MALEPRE.
Table 3
MRS ATP/PCr measures by age-matched group comparisons
|
Pre-menopausal group
|
Age-matched males
|
Peri- menopausal group
|
Age-matched males
|
Post- menopausal group
|
Age-matched males
|
Model 1
|
|
|
|
|
|
|
ACC
|
1.06(0.07)
|
1.11(0.07)
|
1.10(0.04)
|
1.12(0.05)
|
1.13(0.04)
|
1.12(0.05)
|
Insula
|
1.21(0.06)
|
1.29(0.07)
|
1.23(0.04)
|
1.24(0.04)
|
1.26(0.03)*
|
1.20(0.05)
|
IPL
|
0.78(0.07)
|
0.85(0.08)
|
0.76(0.04)
|
0.84(0.04)
|
0.86(0.03)
|
0.83(0.04)
|
MiTG
|
1.10(0.04)
|
1.17(0.05)
|
1.15(0.02)
|
1.12(0.3)
|
1.19(0.02)*
|
1.10(0.03)
|
MTL
|
1.16(0.05)
|
1.24(0.05)
|
1.22(0.03)
|
1.20(0.04)
|
1.21(0.03)
|
1.19(0.04)
|
Model 2
|
|
|
|
|
|
|
MFG
|
1.05(0.06)
|
0.99(0.06)
|
1.04(0.03)
|
0.98(0.04)
|
1.07(0.03)*
|
0.95(0.04)
|
SFG
|
0.95(0.05)
|
0.86(0.06)
|
0.94(0.03)*
|
0.84(0.03)
|
0.96(0.03)*
|
0.82(0.04)
|
Values are means (SE). *Different from age-matched males on univariate post-hoc analysis, p < 0.05. |
Abbreviations: ACC, anterior cingulate cortex; IPL, inferior parietal lobule; MFG, medial frontal gyrus; MiTG, middle temporal gyrus; MTL, medial temporal lobe; SFG, superior frontal gyrus. |
Amyloid-β load
POST and PERI groups exhibited higher PiB uptake in AD-mask than the corresponding male groups (p’s ≤ 0.001; Table 4), while no differences were found between PRE and MALEPRE. These effects remained unchanged after including APOE status as a covariate, which nonetheless enhanced differences between PERI and MALEPERI APOE-4 carriers (p = 0.020).
Table 4
PiB-PET amyloid-β load by age-matched group comparisons
|
Pre- menopausal group
|
Age-matched males
|
Peri- menopausal group
|
Age-matched males
|
Post- menopausal group
|
Age-matched males
|
AD-mask SUVR
|
1.01(0.05)
|
0.92(0.05)
|
1.20(0.05)*
|
0.92(0.06)
|
1.29(0.05)*
|
0.93(0.06)
|
adjusted by APOE-4 status
|
1.01(0.05)
|
0.92(0.05)
|
1.20(0.05)*
|
0.91(0.06)
|
1.29(0.05)*
|
0.93(0.06)
|
AD-mask SUVR by APOE-4 status
|
APOE-4-
|
0.94(0.07)
|
0.86(0.07)
|
1.12(0.06)
|
0.88(0.08)
|
1.32(0.07)*
|
0.93(0.09)
|
APOE-4+
|
1.07(0.07)
|
0.96(0.07)
|
1.39(0.09)*
|
0.96(0.08)
|
1.26(0.07)*
|
0.95(0.10)
|
PiB scans, % positive
|
5
|
0
|
16*
|
0
|
24*
|
3
|
PiB scans, % positive by APOE-4 status
|
APOE-4-
|
0
|
0
|
10
|
0
|
22
|
6
|
APOE-4+
|
9
|
0
|
31*
|
0
|
26*
|
0
|
Values are age-adjusted means (SE). *Different from age-matched males, p < 0.05. |
Abbreviations: see legend to Table 2. |
The frequency of PiB-positive scans was higher in POST vs. MALEPOST (24% vs. 3%, respectively, p = 0.015), and in PERI vs. MALEPERI (16% vs. 0%, respectively, p = 0.012). An effect of APOE-4 status was noted, with POST and PERI carriers exhibiting greater frequency of PiB-positive scans than male carriers (p’s < 0.050), whereas no differences were found among non-carriers. There were no differences in PiB-positivity for PRE vs. MALEPRE, independent of APOE status.
Biomarker changes post-menopause
To examine whether biomarker effects were persistent post-menopause, and to test their specificity for menopause, we performed 2-year follow-up volumetric MRI and FDG-PET scans in a subset of 17 POST women and 12 MALEPOST (Supporting Information Table 13). All results are significant at p < 0.05, cluster-level FWE corrected, adjusted for time to follow-up, APOE-4 status, and modality-specific confounders.
GMV: Results are presented in Fig. 4a and Supporting Information Table 14. There were no GMV changes in inferior temporal and fusiform gyri of the POST group, whereas GMV increases were observed in the precuneus. These changes were significant compared to MALEPOST.
WMV: No significant changes were observed in the POST group, or in comparison to MALEPOST (Fig. 4b, d).
CMRglc: Results are presented in Fig. 4c, d and Supporting Information Table 15. Within temporo-parietal regions that were hypometabolic in the POST group, CMRglc declines were observed only in left inferior parietal lobule, which were significant compared to MALEPOST. There were no further longitudinal differences between groups.
Associations between biomarkers and cognition
As shown in Supporting Information Table 16 and Supporting Information Fig. 1, GMV in precuneus was positively associated with memory scores cross-sectionally (Pearson’s r = 0.311, p = 0.007) and longitudinally (Pearson’s r = 0.586, p = 0.013). The rate of GMV increase in precuneus was also positively associated with memory performance (Pearson’s r = 0.521, p = 0.032). ATP/PCr measures in temporal regions were positively associated with global cognition (Spearman’s Rho = 0.329, p = 0.012).
Menopausal hormone therapy (HT) and hysterectomy status
We performed a sensitivity analysis to examine the effects of HT and hysterectomy status on brain biomarkers, included in Supporting Information Appendix. Briefly, including these variables as covariates left MT effects substantially unchanged. Restricting analysis to HT non-users confirmed results from the entire cohort, with two exceptions: CBF differences between groups were more pronounced among non-users, and FA differences between PERI and POST did not reach significance. As compared to HT users, non-users exhibited lower GMV in inferior temporal gyrus, and lower CMRglc in parietal regions, independent of hysterectomy status.