An imaging-based study identified 2 dimensions of late-life depression (LLD) which had distinct neuroanatomical, cognitive, clinical, and genetic profiles. These findings were published in JAMA Psychiatry.
The Imaging-Based Coordinate System for Aging and Neurodegenerative Diseases (iSTAGING) database has images from >35,000 participants aged 22-90 years from 13 study sites. For this analysis, data collected at 4 sites between 2017 and 2020 from participants (n=996) aged ≥60 years were evaluated for heterogeneity in LLD (cases: n=501; controls: n=495).
Findings were compared with data from all individuals (n=12,518) aged ≥60 years in the United Kingdom Biobank (general population cohort) and all healthy controls (n=1431) from the Alzheimer Disease Neuroimaging Initiative, Baltimore Longitudinal Study of Aging, and Biomarkers for Older Controls at Risk for Dementia studies (longitudinal population cohort).
The cases and controls were aged median 67.3 (range, 60.00-91.00) and 66.26 (range, 60.00-91.47) years, 66% and 67% were women, and they had an average of 14.87 (standard deviation [SD], 2.62) and 14.76 (SD, 2.68) years of education, respectively.
A clustering analysis found 2 dimensions among these data, in which 227 patients with LLD were assigned to cluster 1 and 274 to cluster 2.
Patients in dimension 1 had greater gray matter volume in the bilateral thalamus, caudate, and putamen, and dimension 2 had reduced gray matter volume in widespread cortical regions, compared with controls.
White matter integrity was similar between controls and dimension 1. Patients in dimension 2 had widespread white matter disruption in 31 of the 48 tracts but with small effect sizes. The highest effect size was observed in the middle cerebellar peduncle (Cohen f2 =.05).
Compared with dimension 2, those in dimension 1 had superior scores in fluid intelligence (Cohen d, 0.25), they had fewer errors in pairs matching test (Cohen d, -0.28), and reported fewer depressive symptoms (Cohen d, -0.45).
One de novo variant on chromosome 4 at rs13120336 was associated with dimension 1 (odds ratio [OR], 2.35; P =3.14×108).
With the general population cohort, 2269 were grouped with dimension 1, 3786 with dimension 2, and 2963 in both dimensions, and 3500 in neither dimension. Fluid intelligence scores were higher for dimension 1 (Cohen d, 0.28) and there were fewer errors in pairs matching (Cohen d, -0.13).
Using the longitudinal population cohort, 301 were grouped with dimension 1, 390 with dimension 2, 330 in both dimensions, and 410 in neither. The rate of change in gray matter decreased more rapidly among those in dimension 2, specifically in the left precentral gyrus, temporal pole, and right anterior insula. Over time, dimension 2 exhibited progression similar to Alzheimer disease (Cohen f2 =.03) and brain atrophy (Cohen f2 =.03).
The identified genetic variant should be validated among an independent cohort.
The study authors concluded, “In this study, LLD was characterized into 2 dimensions associated with neuroanatomy, cognitive functioning, and genetic profiles. The 2-dimensional representation offers a potential system for future research on the underlying etiology mechanisms and heterogeneity of genetic architectures as well as personalized clinical care.”
Disclosure: Multiple authors declared affiliations with industry. Please refer to the original article for a full list of disclosures.
Wen J, Fu CHY, Tosun D, et al. Characterizing heterogeneity in neuroimaging, cognition, clinical symptoms, and genetics among patients with late-life depression. JAMA Psychiatry. Published online March 9, 2022. doi:10.1001/jamapsychiatry.2022.0020