Our findings of reduced cpVD and cpFD of glaucomatous eyes compared with age-matched normal controls were in concordance with previous studies that showed reduction of the microvascular density of the optic disc in glaucoma patients.11, 28–31 Rao et al29 reported that VD in optic disc region in PACG eyes was significantly lower than control eyes, whereas VD in primary angle closure (PAC) with high IOP and thinner superotemporal peripapillary RNFL thickness was similar to that of the controls. They suggested that high IOP affects the RNFL measurements earlier than VD in PACG. Scripsema et al11 found that both POAG and NTG patients demonstrated decreased perfused capillary density compared to normal subjects, with POAG patients having a lower perfused capillary density than NTG patients. They attributed the latter finding to be possibly related to medication effect (POAG patients using larger number of drops) and/or the different pathophysiological processes of NTG and HTG.
To our knowledge, this is the first study to compare the peripapillary microvasculature in early PACG and early NTG eyes using OCT-A. We utilized OCT-A to compare the peripapillary microvasculature of two subtypes of early glaucoma which have likely different pathogenic mechanisms – NTG that is less IOP dependent and has possibly a stronger vascular pathogenic component,5, 32 and PACG that has likely a predominantly IOP-dependent mechanism.33 This allowed a “snapshot” comparison of the detailed microvasculature between NTG and PACG at the early stage of the disease. Our results showed a significantly reduced global cpVD in NTG eyes compared with PACG eyes, despite the comparable RNFL thickness and disease severity. We also identified a significant association between the OCT-A metrics and RNFL thickness in both glaucoma groups, with a stronger relationship between the cpFD and RNFL thickness in the NTG group compared with the PACG group. The findings might reflect a more specific and early reduction of microvascular perfusion at the peripapilliary region of NTG eyes at early stage glaucoma. Such a difference was not observed in other studies that compared patients with PACG and POAG (involved both NTG and HTG).34, 35
Given the similar disease severity and the number of medications used, the difference in cpVD between NTG and PACG eyes could be related to the different pathogenic mechanisms of the two glaucoma subtypes. For NTG at the early stage, the impairment of vascular autoregulation – postulated to be an important risk factor for disease progression in NTG5, 6, 36–38 – could lead to reduced blood flow but has yet to cause dysfunction, death, or atrophy of the RGC, as well as the consequential RNFL thinning. Therefore, in NTG eyes, there was a reduction of peripapillary microvascular perfusion prior to RGCs loss, and a further delay for the development of a detectable RNFL thinning after the RGCs’ loss. The latter is supported by in vivo study that showed an initially faster decline of RGC soma counts compared with RNFL thickness following optic nerve injury in animal model.39 Our findings of the stronger association between cpFD and RNFL thickness in NTG eyes compared with PACG eyes, as well as the association between decrease in cpVD and decrease MD of VF in NTG (which was not observed in PACG eyes), may provide a modest but concordant support of this theory. Further in vivo studies are needed to validate the role of microvasculature in the pathophysiology in NTG. In PACG eyes, the loss of RGCs was possibly mainly due to elevation of IOP that occurred prior to IOP lowering treatments (including lens extraction and/or laser iridotomy) and the reduction of the OCT-A metrics could be a secondary consequence of RGCs loss; this echoed with the suggestion of Rao et al. that high IOP affects the RNFL measurements earlier than VD in PACG.29 The atrophy of RGCs may lead to a reduced demand of blood supply and blood flow that was reflected as a reduction in cpVD. However, this secondary reduction of cpVD, unlike NTG eyes at a similar stage of disease severity, was not extensive enough to a degree that would cause further RGCs loss. Hence, this possible difference in the pathogenic algorithm of NTG and PACG might lead to the differences in cpVD measurement between the two subgroups, despite similar RNFL thickness. A longitudinal study with larger number of patients may, in the future, verify the causal relationship of vascular-RNFL thickness in these two subtypes of glaucoma. This is not only important in terms of understanding the pathophysiology of the disease, it is also clinically implicative if OCT-A is to be utilized as a diagnostic and monitoring tool in glaucoma management.
In the multivariable analyses, we found a significant association between OCT-A metrics and RNFL thickness in both glaucoma subtypes but only an association between cpVD and MD in the NTG group (Table 4). This differed from previous studies that reported a stronger association between decreased cpVD with the severity of VF damage, compared with the association between RNFL thickness and VF function in PACG34 and POAG eyes.14, 34 However, our results were expected because we only included glaucoma patients with mild severity with the MD score of better than − 6.0 D. Identifying an association between OCT-A metrics within a narrow range of VF parameters (MD of -5.73 to 1.65 dB and VFI of 81 to 100% in the glaucoma groups) is understandably difficult. Furthermore, the association between microvasculature with functional change may not be strong in these patients with early stage disease that have minimal functional loss. Indeed, a study by Shin et al showed that whilst there was a significant relationship between VD and VF function in moderate-to-advance POAG regardless of location, the relationship of VD and VF function was only significant in the superotemporal and inferotemporal regions for early stage POAG40.
The strength of our study was the inclusion of early NTG and PACG patients – that have similar age, disease severity, RNFL thickness and number of medications used – for comparison. This could better reflect the role of microvasculature in the glaucomatous pathogenic mechanisms at the early stage of the disease. We also used an objective, automated MATLAB program to quantitatively measure retinal microvasculature. Limitations of this study included a cross-sectional study design, a relatively small sample size, as well as not taking systemic diseases into consideration (e.g. obstructive sleep apnea, hypertension). We have limited the range of refractive errors (+ 3.0 to -3.0 D) in an attempt to avoid inclusion of eyes with extreme AL. We acknowledge that the AL of PACG eyes were statistically shorter than either the NTG eyes or control eyes in the current study (Table 1). However, the difference in AL between the PACG and NTG group was reasonable in the clinical point of view (22.43 ± 0.12 D vs 24.60 ± 0.17 D; P < 0.001). Nonetheless, our findings provide the basis for future longitudinal study that may review the causal relationship between retinal microvasculature change and RNFL thickness change.
In summary, the cpVD was significantly lower in early NTG eyes when compared to early PACG eyes, despite similar RNFL thickness and VF parameters. Reductions in cpVD and cpFD were associated with average RNFL thickness thinning in both NTG and PACG eyes. Longitudinal study may verify the differences of microvasculature change in different glaucoma subtypes, improve our understanding of the pathogenic mechanisms of the diseases, and also establish a role for OCT-A in the management of glaucoma.