In the current study, we found distinct changes in the retinal architecture and vasculature in patients with RRMS and pSS. Individuals with RRMS revealed atrophy of inner retinal layers (pRNFL, GCIP) and vessel rarefication within the SVC whereas vessel loss within both the SVC and DVC was evident in patients with pSS irrespective of changes in the retinal architecture. Our findings are consistent with the idea of a disease-specific retinal vessel pathology in RRMS versus pSS.

Prior studies on changes in the retinal thicknesses in pSS are controversial [6, 7, 11, 12, 30]. As has been reported [11, 12], we did not see changes in the average retinal layer thicknesses in patients with pSS. However, a reduction of the total macular RNFL⁠ [6, 30] and the whole posterior pole [7] have been described. Further, antibody positivity, especially of anti-Sjögren’s syndrome type B (anti-SS-B) antibodies, was shown to be associated with inner retinal layer atrophy [6, 30] but was not measured in our study. As expected, we found a reduction of the SVC in patients with RRMS [15,16,17] and pSS [13]. The loss of deep retinal vessels, however, was only visible in patients with pSS and not in patients with RRMS. This findings are coherent with data from smaller pSS cohorts [12, 13, 18]. In our study, the vessel loss of both SVC and DVC in pSS was linked to low visual acuity which has not been described previously. In contrast to vision loss in RRMS patients [31], reduced visual acuity was not associated with inner retinal layer thinning in pSS suggesting a different underlying pathophysiology.

As shown by others [18], we did not find significant changes in the size of the FAZ. However, alterations in the retinal microvasculature have been previously linked to an increased size and reduced circularity of the FAZ in diabetic retinal ischemia in particular of the deep vessels [32]. In other connective tissue diseases like systemic lupus erythematosus, an increase in FAZ sizes is controversially discussed, but often assumed in connection with retinal ischemia [33]. Assuming a similar phenotype in pSS as in other connective tissue diseases, an increase in FAZ area should be investigated with a more sensitive approach using longitudinal intra-individual comparisons.

Studies have shown a dosage-dependent atrophy of inner retinal layers and the retina pigment epithelium due to hydroxychloroquine [34] while others did not find any layer differences [7, 11]. By applying OCTA, a small cohort study found retinal vessel atrophy in some sectors of the deep vascular complex and in the superior sector of the superficial complex in pSS patients under hydroxychloroquine treatment [35] while others could not find any differences [18]. Further, patients with various autoimmune diseases showed an increase in FAZ as well as decreased para-/perifoveal vessel densities in patients taking hydroxychloroquine for more than 5 years [36] while others saw an increase in vessel densities [37]. Despite the exclusion of patients under hydroxychloroquine, pSS patients still exhibited a reduced vessel density [12] like our cohort which is why we assume a hydroxychloroquine-independent effect.

We did not see any differences in serum NfL or GFAP in neither pSS nor RRMS in comparison to healthy control. In pSS, studies have shown normal NfL levels during remission and an association of NfL with active disease [38, 39]. Since the pSS patients in our cohort did not show signs of active disease, normal serum NfL levels in our study were consistent with prior reports. In contrast, elevated NfL levels in RRMS are indicative of disease activity and have demonstrated responsiveness to immunotherapy [40, 41]. Notably, RRMS exhibits a less pronounced age-related increase in NfL levels compared to healthy individuals and other neurodegenerative diseases [42]. Compared to typical RRMS studies, our cohort, with an advanced median age of 54 years, demonstrated a relatively benign disease course (median EDSS 1.8). Additionally, participating patients reported neither recent relapses nor relapse-independent progression and 75% received immunotherapy. Similar results have been found for GFAP levels in RRMS [43]. Considering this, our findings seem in line with the literature.

Different hypotheses have been proposed to explain the pathophysiological mechanisms behind superficial retinal vessel loss in MS. These include primary and secondary effects of either altered metabolic states of retinal cells or relapse-independent inflammatory processes. Firstly, inner retinal layers, like the GCIP and the RNFL, are supplied with oxygen by the SVC [44]. Acute ON [15] and retrograde axonal degeneration due to inflammation in the visual pathway [45, 46] lead to atrophy of inner retinal layers. The reduced oxygen demand might consecutively cause a rarefication or decreased perfusion of superficial vessels. Secondly, research in our group has linked inflammatory processes in the CNS to ON-independent superficial vessel loss [47]. Therefore, we suspect that inflammation-related mechanisms might be responsible for triggering significant alterations in the retinal microvasculature independent of relapses.

The pathophysiology behind pSS is complex and not fully understood [48, 49]. It involves the activation of cytokines like interferons, leading to B- and later T-cell activation and infiltration in particular of exocrine glands. It has been shown in the past that vision-threatening ocular involvement in pSS is associated with systemic disease manifestations like nephritis, peripheral neuropathy and vasculitis [9]. There are no data on retinal histology in ocular manifestations in pSS; however, nerve biopsies in peripheral neuropathy have found evidence of perivascular inflammatory infiltrates and other vessel abnormalities including (necrotizing) vasculitis [50]. Further it has been shown that anti-Sjögren’s syndrome type A (anti-SSA) antibodies are associated with vasculitis [51], in particular retinal vasculitis, as shown in a case report based on fluorescence angiography [52], and are more commonly observed in central nervous system manifestations [53]. In our study, we observed a loss of both superficial and deep retinal vessels without significant alterations of the retinal layer architecture itself. We also saw an association of vascular changes and not layer atrophy with impaired visual function. Taking this into account, we suspect a primarily vascular pathology in patients with pSS.

Our study has several limitations. Our samples sizes are limited, and we conducted an exploratory analysis without replication in a second. Therefore, we cannot rule out any false positive results. Especially pSS is a rather rare disease and larger cohorts have not been published so far. Furthermore, our study contains only cross-sectional data. A longitudinal study with a larger sample size is necessary to determine changes during the disease course. Additionally, neurological assessments of patients with pSS relied solely on comprehensive clinical examinations, lacking supplementary electrophysiological diagnostics or imaging. Consequently, a differentiated perspective on pSS patients with central in contrast to peripheral nervous system involvement is not provided and should be the subject of further studies. Differential diagnosis takes place in early stages of disease therefore data on differences during disease onset or early ocular manifestations are necessary. In accordance with the female predominance of pSS, this study only provides limited data on findings in male patients with pSS. A large male cohort would be necessary to evaluate sex differences in OCT and OCTA measures in pSS. Further, pre-existing conditions affecting the visual system like diabetes were gathered through a comprehensive medical history only. An impact of undiscovered comorbidities on our data cannot be ruled out. Moreover, we were not able to include MRI findings or further instrumental diagnostic procedures like visual evoked potentials, visual field testing or electroneurography into our analysis. Also, we did not use the EULAR Sjögren Syndrome Patient Reported Index (ESSDAI) [54] to quantify disease burden which impairs comparisons with other studies involving pSS. However, the ESSDAI comprises only two neurological of 12 items: CNS, including ocular involvement, and peripheral nervous system. Therefore, we chose the EDSS to facilitate the comparison with MS patients and better captivate the neurological burden of disease. Furthermore, OCTA is extremely susceptible to imaging artifacts, in particular in patients with impaired vision. In this study, OCTA examinations were conducted by experienced technicians and underwent rigorous quality control [25]. OCTA only provides data on vessel perfusion and not morphology. Therefore, concerning reduced vessel densities, we cannot differentiate between true vessel loss, wall thickening or constriction.

In conclusion, a distinct atrophy of retinal vessels can be observed during MS and pSS suggesting different underlying disease mechanisms. After validation in larger, longitudinal cohorts, OCTA might allow for differential diagnosis of RRMS and pSS.