Hippocampal subfield volumes are uniquely affected in PTSD and depression: International analysis of 31 cohorts from the PGC-ENIGMA PTSD Working Group

PTSD and depression commonly co-occur and have been associated with smaller hippocampal volumes compared to healthy and trauma-exposed controls. However, the hippocampus is heterogeneous, with subregions that may be uniquely affected in individuals with PTSD and depression. We used random effects regressions and a harmonized neuroimaging protocol based on FreeSurfer (v6.0) to identify sub-structural hippocampal markers of current PTSD (C-PTSD), depression, and the interaction of these conditions across 31 cohorts worldwide (N=3,115; Mage=38.9, SD=13.9 years). Secondary analyses tested these associations by sex and after modeling the simultaneous effects of remitted PTSD, childhood trauma, mild traumatic brain injury, and alcohol use disorder on hippocampal subfields. A significant negative main effect of depression (n=800, vs. no depression, n=1456) was observed in the hippocampal tail (beta=-0.13) and CA1 (beta=-0.09) after adjusting for covariates and multiple testing (FDR-adjusted p-values (q)=0.028). A main effect of C-PTSD (n=1042, vs. control, n=1359) was not significant, but an interaction between C-PTSD and depression was significant in the CA1 (beta=-0.24, q=0.044). Pairwise comparisons revealed significantly smaller CA1 volumes in individuals with C-PTSD+Depression than controls (beta=-0.12, q=0.012), C-PTSD-only (beta=-0.17, q=0.001), and Depression-only (beta=-0.18, q=0.023). Follow-up analyses revealed sex effects in the hippocampal tail of depressed females, and an interaction effect of C-PTSD and depression in the fimbria of males. Collectively our results suggest that depression is a stronger predictor of hippocampal volumetry than PTSD, particularly in the CA1, and provide compelling evidence of distinct and more pronounced hippocampal phenotypes in comorbid PTSD and depression compared to either condition alone.