Commentary on “Smaller Hippocampal Volume in Current But Not in Past Depression in Comparison to Healthy Controls: Minor Evidence From an Older Adults Sample”

Introduction

According to the World Health Organization, the overall prevalence of unipolar major depression among the general older population is around 7%. Unfortunately, depression among older adults is frequently underdiagnosed and undertreated, especially when it coincides with other chronic medical conditions associated with aging. Moreover, it is both a risk factor and a prodrome of dementia.

The effect of depression on the brain has been extensively studied. Numerous structural magnetic resonance imaging studies examined the neuroanatomical correlates of depression, reporting both smaller hippocampus (HC) volume and, although in a lower number of studies, smaller right anterior cingulate cortex (ACC) gray matter volume in patients with major depressive disorder (MDD) in comparison to healthy controls. Similar findings were found in older adults (see our full article for relevant literature¹).

Smaller brain volume in MDD could be a cause or a consequence of the disease and be related to treatment status, depression duration, and other factors, including the severity of the depressive episode. Remission after treatment has been associated with larger pretreatment hippocampal volume. Conversely, antidepressant use reduces time to recovery and induces brain plasticity especially in HC, prefrontal cortex, ACC, and medial temporal areas, which would decrease the gray matter loss.

Our Study on a Large Population of Older Adults

In order to further understand volume differences between currently depressed individuals and people with remitted past major depressive episode (MDE), and assess if remission is associated with volume recovery, we investigated HC and ACC volume differences in a sample of French older adults (n = 516).¹ We compared HC and ACC volumes in older adults currently depressed (n = 150) or with past remitted MDE (n = 79) versus healthy controls (n = 287). Consequently, we selected nondemented individuals from a cohort of community-dwelling older adults. We used both an automated method (FreeSurfer Software, version 6.0) and a manual method to measure HC subregions: head, body, and tail, in T1-weighted magnetic resonance images.

Our results showed smaller left posterior HC volume in currently depressed participants compared to healthy controls with manual HC measurement, after adjustment for potential confounders (gender, age, education level, and total intracranial volume). When individuals with past, but not current, MDE were compared to those who never experienced depression, no volume reduction was found in either the HC or ACC.

Reduction in HC volume could play an important role in the pathophysiology of older depression by participating in affective symptoms and cognitive dysfunction that embody the core of major depression. Our results are in line with former studies supporting HC volume loss in depressed older adults, but, perhaps due to the strict methodological adjustments we made to avoid type I errors, the reduction was only significant for the left posterior HC. In agreement with our findings, a recent study showed a significantly smaller left posterior HC in individuals with current late-onset depression versus controls.² Importantly, the posterior HC has been implicated in memory function, while the anterior HC is more important for emotional responses. Therefore, a significant reduction in the posterior portion in older depressed participants could suggest a more important contribution to memory deficits in this population than what one could find in younger depressed participants. This is in line with what is clinically observed in older depressed participants, which often mimics the phenotype of dementia.

These findings suggest that studying total HC volume alone could fail to detect relevant changes in HC subregions. The posterior HC tail is especially involved in spatial learning and memory—neurocognitive skills often found to be altered in MDD.³ In the same vein, changes in HC volume in cases of depression were recently found to be more pronounced in left areas of core HC substructures such as the CA1 region and the tail.⁴

Additional Results and Insights

Although we did not present this analysis in our article, we also investigated bivariate correlations between the years in remission, since the last MDE, and the HC and ACC volumes. We found a positive correlation between years spent in remission and the left posterior HC volume (r = 0.14; P = .03). This is in line with previous literature relating remission from depression and HC volume and specifically reporting that patients in remission (during the 3-year study period) had less left HC volume decline than nonremitted ones.⁵ Remitted patients with MDD were also shown to have no HC volume differences compared with healthy controls.⁶ Hence, HC volume reduction in depression can be hypothesized as a state and not a stable trait of the illness depending on the time since remission. The longer the duration of remission, the more HC volume might be restored. Another plausible hypothesis is that lingering depression may conduct to smaller HC and that aggressive early treatment (leading to remission) may be associated with larger “recovered” HC.

According to the neuroplasticity theory, neural circuits and connections undergo permanent modifications and reorganization in reaction to external or internal environmental stimuli. Neurogenesis in adulthood involves precursors of cell proliferation, migration, and differentiation largely occurring in the dentate gyrus of the HC, therefore generating new neurons throughout life.^7,8 Stress may inhibit neurogenesis in the HC and eventually impact upon its volume in aging individuals. Hence, stress-induced reduction in neurogenesis may be a significant factor in precipitating episodes of depression. As for recovery from depression, it could involve a restoration of the original basal rate of neurogenesis (whether it is due to antidepressant treatment that positively impacts HC neurogenesis or endogenous changes). So, our results revealing a positive correlation between years since the last MDE and the left posterior HC volume may be interpreted in the direction of a brain recover of its volume. However, we need to consider that antidepressant response in terms of hippocampal neurogenesis and clinical improvement may be blunted with aging.

Regarding ACC, our results were inconsistent with previous literature since only a few studies did not find any volume reduction in the ACC of depressed patients compared to controls. The lack of reported volume reduction could be linked to our conservative approach (the Benjamini-Hochberg false discovery rate [FDR] correction): in fact, results went in the same direction of previous literature since we were able to detect a lower right caudal ACC volume in currently depressed versus past MDE individuals (P = .01); however, when we applied the FDR correction, the difference became no longer significant. This probably implies that the effect is not so strong or that other factors play a more important role than previously assumed.

It should be underlined that HC volume decrease in late-life depression could be due to neurodegeneration, and, more specifically, to prodromal Alzheimer disease. Interestingly, a smaller left HC volume was found to be associated with incident dementia in a sample of older, nondemented depressed individuals followed for up to 5 years.⁹ We specifically selected nondemented participants, but this may have been insufficient, since prodromal dementia cases among currently depressed participants were not excluded, and the differential diagnosis between these 2 conditions is controversial. Of note, amyloid binding measured with positron emission tomography has been shown not to be associated with HC volume loss, suggesting HC volume loss can occur via other pathogenetic mechanisms.¹⁰

In summary, our study reported a left posterior HC volume reduction in older adults with current depression but not in participants with past MDE compared to healthy controls, and only when manual measures were used. The small but positive correlation between time since the last MDE and left posterior HC volume—reported for the first time in this commentary—suggests that the HC volume loss would be state related and reversible after treatment and clinical improvement. This is in contrast with the hypothesis that smaller HC volume would precede the onset of MDD and be a predisposing factor rather than a consequence of the disease. Moreover, our findings suggest the presence of a neuroplasticity “reserve” allowing the brain to recover its volume after clinical remission. We hypothesize that, in depressed individuals, HC volume loss might recover over time in the absence of new episodes. Further studies using a longitudinal design are needed to verify our hypothesis. Other limitations present in our previous study should also be overcome; in particular, factors influencing brain volume changes, such as the type, dose, and duration of antidepressant treatment, as well as cognitive decline, need to be carefully considered. Finally, the issue of the different meaning of HC volume changes cannot be overlooked, since larger HC tail volume was also reported to be associated with depression.