VD reduction as determined by OCT-A is known to be highly associated with the loss of VFMS in glaucoma patients,2,3,22 which indicates the potential role of VD as an additional surrogate marker to monitor functional loss in glaucoma. As the presence of CMvD is linked to microvasculature insufficiency of the ONH/choroid and a poorer prognostic outcome,4,10,15 it is plausible to hypothesize that vasculature-function relationships may differ between OAG eyes with and without CMvD. In this current study, OAG eyes with CMvD revealed stronger global and sectoral vasculature-function relationships in the 24 − 2 and central 10° VF maps, while there were no differences in the structure-function relationships between the two groups. These findings suggest that VD parameters, as measured by OCT-A, may be useful biomarkers for detecting and monitoring glaucomatous damage in OAG eyes, particularly in those showing signs of choroidal microvasculature insufficiency such as CMvD. To our knowledge, this present study is the first to demonstrate different impacts on the vasculature-function relationship in OAG eyes according to the presence of CMvD.
Our current analyses indicated that the global mean cpVD was significantly lower in the CMvD + eyes but that there was no significant difference in the global cpRNFLT values between the two groups (Table 1). In sectoral assessments, the CMvD + eyes showed a significantly lower cpRNFLT at the IT sector and lower cpVD at both the IT and T sectors (Table 2). These findings suggest that OAG eyes with CMvD show more severe structural and microvasculature damage than those without CMvD, despite both groups having a similar age, AL, and VF MD. There are several possible explanations for these observations. First, the damaged blood-brain barrier at the choriocapillary non-perfusion area within the β-PPA zone, as represented by CMvD, can promote the release of vasoactive or toxic substances into the ONH, leading to axonal damage, RNFL loss, and cpVD reduction.23,24 Another possibility is that as the parapapillary choroid is closely linked to ONH perfusion, insufficient blood flow within this structure in the eyes with CMvD can result in diminished blood flow to ONH, thereby leading to deep ONH structural damage, such as LC morphologic changes, which can subsequently induce axonal and vascular loss in the superficial retina.4–8, 25 Of note in this regard, sectoral loss of the cpVD and cpRNFLT was found to be spatially correlated with CMvD location and was mostly observed at the T and/or IT sectors in the CMvD + eyes in the present study series (83.5%). These findings are in line with those of a previous study, which also reported an association between CMvD and reduced cpVD and RNFLT with topographic correlation,5 and with prior reports that CMvDs are most often found at the IT region in the optic disc, which is the most vulnerable area to disruption of the microvasculature and ONH structural damage.5,16,26
In our present study series, the global and sectoral mean mVD and mGCIPLT values were significantly lower in the CMvD + eyes, except for the ST and SC sectors of mGCIPLT (Tables 1 and 2). The CMvD + eyes also showed a significantly higher prevalence of central VF loss as well as a lower cVFMS at the ST and SN sectors compared to CMvD- eyes, despite both groups having similar overall VF MD values (Tables 1 and 2). One speculation in relation to these findings is that since CMvD is most often located near the macular vulnerability zone, which was first described by Hood et al,27 its presence can be directly related to the loss of macular structural parameters (i.e., mGCIPLT and mVD) and central VF defects.16,28 In addition, as the density of retinal ganglion cells (RGCs) is highest in the central macula, the RGCs in the macular region require a high oxygen supply, rendering them particularly vulnerable to microvascular hypo-or non-perfusion conditions such as CMvD.
Our comparisons of the vasculature-function relationships between the two groups demonstrated that global and sectoral vasculature-function relationship were significantly stronger in the CMvD + eyes, according to either the 24 − 2 or central 10° VF maps. (Table 3). One explanation for these observations is that despite both groups having similar VF MD values after matching for glaucoma severity (i.e., early-stage glaucoma), choroidal microvasculature insufficiency in the form of CMvD may induce generalized deficiency of blood flow in the ONH and retina of CMvD + eyes,9 thus leading to more pronounced loss of cpVD/mVD and VFMS globally and sectorally.22 This in turn result in stronger vasculature-function relationships in the CMvD + eyes based on both 24 − 2 and central 10° VF maps.
It is noteworthy that there were no significant differences in the global and sectoral structure-function relationships between the eyes with and without CMvD, according to either the 24 − 2 or central 10° VF maps (Table 3). Although the explanation for these findings is unclear, one possible reason for this observation is that RGC damage induced by microvasculature insufficiency, as observed in the CMvD + eyes, might not have completely manifested into sufficient cpRNFLT/mGCIPLT reduction, thus resulting in similar structure-function relationships between the two groups. Munguba et al29 reported that changes in RNFL thickness following crush injury lag behind the decline of RGC function, based on their histopathologic analyses. Hence, although the mechanisms of ONH injury may differ between the two studies, ONH/retinal damage induced by CMvD may first lead to microvasculature loss and functional damage, followed by a slower degeneration of the RNFL, which may explain our present findings that VD loss is better associated with the RGC dysfunction and subsequent VFMS loss than RNFL thinning. Further studies with a longitudinal design are needed to validate our hypothesis in the future.
The linear regression analyses showed that a reduced global cpVD and mVD were independently associated with the corresponding global 24 − 2 and central 10° VFMS in the CMvD + eyes, but this association was found with a reduced global cpRNFLT and mGCIPLT in the CMvD- eyes. These findings may further support our speculation that VD parameters have stronger relationship with functional loss in eyes with CMvD. Interestingly, the global cpRNFLT was not significantly associated with global 24 − 2 VFMS, whereas mGCIPLT was independently associated with cVFMS, in the CMvD + eyes (Table 4). Marshall et al30,31 recently reported that glaucoma patients with vascular risk factors such as hypertension or myocardial infarction predominantly showed mGCIPLT loss, rather than a cpRNFLT reduction. These authors hypothesized that vascular dysfunction may be important in glaucomatous damage involving the macular structure. Our current findings of a significant association between the cVFMS and mGCIPLT in the CMvD + eyes may be explained by microvasculature insufficiency present in those eyes.
There were several limitations of note in the present study. First, en face choroidal images of an OCT-A may sometimes have technical limitations in accurately detecting CMvD. For example, projection artifacts, such as shadowing effects from the optic disc hemorrhage (ODH) or projection flow from the superficial retinal layer, can make it difficult to define the presence of CMvD and/or ONH/CMvD boundary. Nonetheless, we attempted to exclude the influence of large overlying retinal vessels and/or ODH on our scanning laser ophthalmoscopy images within the \(\beta\)-PPA during the assessment of CMvD by having multiple examiners to evaluate the OCT-A images based on a method validated in previous studies.4,9,16,17 Second, the six sectors covered by the circumpapillary and macular OCT-A/OCT maps may not completely match the corresponding sectoral VFMS sectors topographically. However, these regions are automatically provided by these devices and closely matched with the corresponding VFMS based on the previously validated regionalization map of Garway-Heath et al.13,18−20 Third, as we evaluated early-stage OAG eyes in our present analyses, our results may be limited in terms of their generalizability to other patients with different types and severity of glaucoma. However, since the vasculature-and structure-function relationships vary according to disease severity, early-stage OAG eyes were selected to compare the relationship of vasculature/structure with the corresponding VFMS in the two OAG groups, while removing the confounding effects of advanced disease severity on the CMvD. Fourth, we did not evaluate the potential confounding effects of topical IOP-lowering medications on vasculature-function relationships, since ocular hypotensive medications may affect ocular blood flow,32 thus affecting VD measurements. Hence, our findings should be cautiously interpreted due to the possibility of confounding effects of topical hypotensive medications. Fifth, the homogeneous Korean population analysed in this study may have introduced a selection bias, in which our findings may not fully reflect those of general populations in other countries. Hence, validation of our data is needed in clinically similar populations outside of Korea. Finally, our cross-sectional study design did not provide information on the temporal relationship between the CMvD and vasculature-function correlations.
In conclusion, early-stage OAG eyes with CMvD show a significantly greater degree of global and regional macular VD loss, and stronger vasculature-function relationships, compared to eyes without CMvD. Our findings suggest that VD parameters derived from an OCT-A may be useful biomarkers of disease progression in early-stage OAG patients with choroidal microvasculature insufficiency.