Lower regional gray matter volume in the absence of higher cortical amyloid burden in late-life depression

Abstract

Late-life depression (LLD) is associated with a risk of developing Alzheimer's disease (AD). However, the role of AD-pathophysiology in LLD, and its association with clinical symptoms and cognitive function are elusive. In this study, one hundred subjects underwent amyloid positron emission tomography (PET) imaging with [¹⁸F]-flutemetamol and structural MRI: 48 severely depressed elderly subjects (age 74.1 ± 7.5 years, 33 female) and 52 age-/gender-matched healthy controls (72.4 ± 6.4 years, 37 female). The Geriatric Depression Scale (GDS) and Rey Auditory Verbal Learning Test (RAVLT) were used to assess the severity of depressive symptoms and episodic memory function respectively. Amyloid deposition was quantified using the standardized uptake value ratio. Whole-brain voxel-wise comparisons of amyloid deposition and gray matter volume (GMV) between LLD and controls were performed. Multivariate analysis of covariance was conducted to investigate the association of regional differences in amyloid deposition and GMV with clinical factors, including GDS and RAVLT. As a result, there were no significant group differences in amyloid deposition. In contrast, LLD showed significant lower GMV in the left temporal and parietal region. GMV reduction in the left temporal region was associated with episodic memory dysfunction, but not with depression severity. Regional GMV reduction was not associated with amyloid deposition. LLD is associated with lower GMV in regions that overlap with AD-pathophysiology, and which are associated with episodic memory function. The lack of corresponding associations with amyloid suggests that lower GMV driven by non-amyloid pathology may play a central role in the neurobiology of LLD presenting as a psychiatric disorder.Trial registration: European Union Drug Regulating Authorities Clinical Trials identifier: EudraCT 2009-018064-95.

Figure 1

Brain regions showing significantly lower GMV in patients with late-life depression compared to healthy controls following whole-brain voxel-wise analysis. Three regions were identified: the left temporal, left parietal and right occipital regions. Areas associated with lower GMV in LLD are shown in yellow in (A) the left lateral view and (B) the bottom view. Boxplots of GMVs and SUVRs in each region are also presented: GMV in (C) the left temporal, (D) the left parietal, and (E) the right occipital regions; SUVR in (F) the left temporal region, (G) the left parietal region, and (H) the right occipital region. GMVs in these regions were significantly lower in LLD than in healthy controls (C–E), whereas SUVRs in these regions did not differ between groups (F–H). Y-axis represents mean voxel values in each cluster. HC healthy control, LLD late-life depression.

Figure 1

Brain regions showing significantly lower GMV in patients with late-life depression compared to healthy controls following whole-brain voxel-wise analysis. Three regions were identified: the left temporal, left parietal and right occipital regions. Areas associated with lower GMV in LLD are shown in yellow in (A) the left lateral view and (B) the bottom view. Boxplots of GMVs and SUVRs in each region are also presented: GMV in (C) the left temporal, (D) the left parietal, and (E) the right occipital regions; SUVR in (F) the left temporal region, (G) the left parietal region, and (H) the right occipital region. GMVs in these regions were significantly lower in LLD than in healthy controls (C–E), whereas SUVRs in these regions did not differ between groups (F–H). Y-axis represents mean voxel values in each cluster. HC healthy control, LLD late-life depression.

Figure 2

Associations of clinical factors with regional GMV decrease in the left temporal region. Episodic memory scores correlated with lower GMV in the left temporal region. To illustrate the results of the MANCOVAs, scatter plots of (A) episodic memory scores and (B) GDS, and boxplots of (C) onset age and (D) the presence of psychosis are presented. Episodic memory scores (A) had a significant main effect on regional GMV difference, whereas other clinical factors (B–D) did not have significant main effect. In all figures, data from healthy controls are also overlaid for reference. Y-axis in (C,D) represents mean voxel values in the left temporal cluster. EOD early-onset depression, GDS geriatric depression scale, GMV gray matter volume, HC healthy control, LLD late-life depression, LOD late-onset depression, NPMD non-psychotic major depression, PMD psychotic major depression.