Rheumatoid arthritis (RA) and ulcerative colitis (UC) are two of the most prevalent autoimmune diseases and are both projected to have increasing incidence and prevalence rates globally (1, 2). Autoimmune diseases represent a large and heterogeneous group of disorders that afflict specific target organs (3). However, there are common links between these different diseases. This includes the presence of both individual and familial polyautoimmunity, which is defined as the presence of more than one autoimmune disease in a single patient or within a familial line (4).

There is also significant overlap in the treatment of these two diseases both in the acute, flare stage and long-term disease management (5–7). The common genetic markers and antigen patterns observed may provide some indication of shared disease pathogenesis specifically in genes, including CCL3 and CXCL10, that are related to T-cell activation and leukocyte migration (8–12).

Traditionally, both in clinical drug development and clinical practice of autoimmune diseases, the primary focus has been in the management of disease burden in the periphery (13–16). However, the presence of central symptoms, such as pain and fatigue, present a distinct challenge in effectively treating patients with these diseases (17). Patients with RA and UC have increased risk for psychiatric disorders including anxiety and depression (18–21). One study reported that 23% of UC patients suffer from depression and 33% suffer from anxiety (22). This has been reported to be even higher in RA with one study showing 55% of patients reporting depressive symptoms (22, 23). This is considerably higher than the base rate of anxiety and depression in the UK population at the time this data was acquired which was 4.65 and 4.12%, respectively (24, 25). In addition, RA patients with depression report increased autoimmune disease activity and lower response to treatment (6, 19, 26).

A previous study in RA has shown smaller hippocampal volume in patients is associated with more severe functional disability and higher pain perception both on a visual analog scale and in functional pain response to pressure stimulus (27). There have also been previous studies which have looked at volumetric differences in gray matter in both UC and RA that suggest subcortical differences in both diseases that may involve the pain processing pathways (28, 29). These central manifestations combined with increased interest in improving our understanding of the neuro-immune axis, is motivating research into whether systemic inflammation from these autoimmune pathways has a central effect. Following on from that, this raises the question whether there is a common or discrete central effect caused by different autoimmune diseases.

The aims of our study were to determine if there were structural brain differences between patients with RA compared to matched controls and patients with UC compared to matched controls. The primary analysis focused on the total hippocampal volume differences between groups with a secondary analysis looking at further subcortical regions. The hippocampus was chosen as the region of interest for the primary analysis due to previous research linking the hippocampus to systemic inflammation and autoimmune diseases (30–32). Additionally, we looked to see if there was overlap in the subcortical regions affected between these two diseases.

In patients with UC compared with controls, we found a moderately smaller total hippocampal volume. Additional analyses were conducted looking at the right and left hippocampus separately and the resulting p-values did not indicate a strong unilateral affect. No similar, significant reduction in hippocampal volume was observed in the RA population. In the secondary analysis, we found a significantly lower amygdala volume in patients with RA compared to controls. This same result was not seen in the UC population.

These findings may suggest there are different areas of the brain effected in these two diseases. At the same time, the close relationship between the amygdala and hippocampus may also suggest a common pathway affecting the brain in these diseases. This is particularly interesting as both the amygdala and the hippocampus are associated with chronic pain and depression (37, 38).

The fact that in this work both regions were not affected in the same way in these two diseases suggests potential differences in the specific disease processes that may preferentially affect one region over another and may result in varying degrees of these central symptoms. The significant difference in amygdala volumes in RA may be linked to the increased rates of anxiety in RA versus UC patients. Previous rs-fMRI research has suggested that the cognitive impairments and reported “brain fog” in UC may be predominantly linked to the limbic system which would be consistent with these findings (39).

Increased rates of cardiovascular risk factors are a known complication of autoimmune diseases (40, 41). The prevalence of hypertension seen in our RA population is consistent with previously published literature (42). In the case of UC, it is more complicated with some of the literature reporting higher rates of hypertension and some reporting similar rates of hypertension to the general population, but higher rates of other cardiovascular risk factors (43). In both diseases there is extensive literature pointing to a host of higher cardiovascular risk factors with unknown etiology (44). It has been proposed that this could be a result of medication usage, lack of physical activity due to disease burden and/or a function of the disease process directly (45, 46). Irrespective of the origin of these risk factors there is clear evidence outlined here in this study, of localized, lower subcortical volumes in both diseases.

This is the largest dataset, to date, investigating subcortical brain volume measures in both RA and UC. Previously published work has looked at total gray matter volume differences in IBD, an umbrella term including both Crohn’s disease and UC, and explored volumetric differences correlated to systemic inflammatory markers and clinical disease status (n = 18 in both studies) (47, 48). Similar work has been published in RA linking increased peripheral inflammation with lower GM volumes and larger ventricle-to-brain ratios to disease duration (n = 33 and n = 54) (49, 50). There is some precedent for looking at subcortical volumes in RA (n = 31 and n = 54) (28, 51). A meta-analysis and literature review by these authors provides a more comprehensive outline of previously published volumetric work in both RA and UC (30).

A limitation to this study is the lack of information on disease duration and severity. Both UC and RA are relapsing–remitting diseases and can vary widely in their individual presentation. In a recent paper Zhang et al. sub-categorized UC patients into active stage versus remission and reported fewer regions of decreased neuroanatomical volume in those currently in remission (52). There is also previously published data that suggests that the age of onset can significantly affect disease severity (53). The lack of this information prevents us from stratifying this patient population based on their clinical disease activity and management. How well controlled an individual’s disease is and on what medications may have an important impact on these central effects which we are unable to quantify with this data set.

One ongoing point of debate with regards to the UK Biobank data set is the lack of heterogeneity and the question of whether it is truly representative of the wider population. Participants tend to be healthier than the general population reporting lower rates of cancer and overall all-cause mortality in addition to being more health conscious. This is a well-established effect seen in volunteer-based cohort studies (54). While this is important to acknowledge, the breadth of lifestyle, genetic and demographic information allows us to appropriately contextualize our patient and control populations. The UK Biobank is continually adding lifestyle, wellness and cognitive information on subjects and could be a valuable additional next step to expanding this research. As a resource this allowed us access to neuroimaging data in these two diseases that is substantially larger than anything published to date.

A potential next step for this work would be to look at volumetric changes in these populations over time. There is an ongoing project to acquire follow up scans in a subset of up to 10,000 UK Biobank Imaging cohort participants. This may make it possible to analyze atrophy over time in these patient populations. This would provide additional insight into how the presence of these systemic autoimmune diseases may affect atrophy and aging in these populations compared to their peers.

These findings play a potentially important role in further understanding brain volume differences in preferential areas of the brain in RA and UC. With increased focus on understanding and mitigating risk factors for neurodegenerative diseases this provides the potential foundation for future work in exploring the link between these autoimmune diseases and the development of future neurodegenerative diseases in these populations. There is a known link between UC and Parkinson’s Disease (PD) and between RA and Alzheimer’s Disease (AD) (55–58). There is also documented research suggesting accelerated atrophy in PD in the hippocampus, as seen in our results for UC, and in the amygdala in AD, as seen here in our RA results (59, 60).

The results of this study highlight the central effect of rheumatoid arthritis and ulcerative colitis and the importance of continued monitoring and treatment of central symptoms for brain health long term in people with chronic autoimmune conditions.