Stabilization of macular, peripapillary and papillary vascular parameters after XEN and trabeculectomy visualized by optical coherence tomography angiography
Filtering IOP-lowering surgery (XEN or TE stent) did not significantly alter vascular parameters in the papillary, peripapillary, and macular region in preoperative eyes with moderately elevated IOP. Comparing the two treatments, the TE-treated patients showed slightly higher results in VAD (SVC, DVC, and Optic Disc) than the XEN stent group after 6 months.
Both filtering surgeries, the XEN stent and the TE, significantly reduce IOP12,14,15,16,17. Similar to previous results, the decrease in glaucoma eye drops was comparable between the two procedures, but the IOP lowering effect was stronger in the TE group than in the XEN stent group.14,15,16,17.
To our knowledge, this is the first report that has comprehensively analyzed VAD by exploring the SVC, DVC, CL, CCL and optic disc layer in filtering glaucoma surgery and thus advances the concepts of glaucoma treatment. Analyzing the influence of lowering IOP on the microvasculature, we found no significant changes in VAD over time in all perfusion regions, except the CCL layer in patients with VAD. a XEN stent. Comparable to these results, Zéboulon et al.29 detected no change in the VAD of the superficial peripapillary layer 1 month after TE. With similar results 6 months after surgery, Lommatzsch et al.25 also did not detect changes in the papillary and peripapillary layer. However, a significant increase in VAD of the peripapillary layer after 3 months after TE surgery was reported by Shin et al.23In et al.24 and Miraftabi et al.30. Kim et al.21 could detect an increase in papillary VAD 3 months after ET but not in the peripapillary layer.
Considering these irregular postoperative follow-up intervals found in the current literature, we inferred that the time period selected to detect VAD changes could be crucial due to influential factors like postoperative inflammation (e.g. reaction of the anterior chamber or corneal oedema) which can influence image quality. For example, we hypothesize that the significantly lower retinal thickness observed 3 days after XEN (p=0.018) compared to patients treated with TE, is due to the significantly lower image quality (p=0.028) during this period. For our study, we tried to include all intervals that seemed significant for the investigation, but as mentioned above, we could not detect significant changes or specific trends in VAD at any follow-up interval. .
Regarding the FAZ results, Shoji et al.31 and Park et al.32 were able to detect significant change and reduction 3 months after glaucoma surgery, while we could not. This could be related to a lack of precision of our method. Shoji et al.31 used the same OCT-A but different external software for their study which allowed for a magnified view and specific color coding making tracing potentially more suitable for detecting changes in FAZ. In contrast, our results of no FAZ change match our VAD results. Unlike In et al.24who found significantly increased VAD after TE treatment, and Park et al.32, which showed increased VAD in the optic disc and decreased FAZ after glaucoma surgery, our patients had moderately high preoperative glaucoma (about 20 mmHg). IOP amplitude was significantly lower compared to their study where IOP amplitude was high and preoperative IOP was poorly regulated (approximately 30 mmHg)24. It is possible that the effect of lowering IOP in preoperative moderately IOP-controlled glaucoma is not as influential on changes in VAD. This is further supported by our correlation analysis where none of the vessel complexes are significantly correlated with changes in IOP. The influence of higher preoperative baseline IOP values has already been discussed and it is hypothesized that “lower pressure range autoregulation can compensate for changes” in VAD adequately, as reported by Lommatzsch et al.25. Another potential influence could be the initial severity of glaucoma as measured, for example, by the MD visual field. Miraftabi et al.30, which had patients with slightly more pronounced visual field loss from the higher baseline MD values, observed a significant increase in VAD after TE. Once again, it could be argued that the severity of glaucoma could influence vascular architecture.
Kim et al.21 linked their improved diagnostic outcome of vascular parameters more to reduction in lamina cribrosa depth (LCD) rather than reduction in IOP. LCD changes were not analyzed in this study.
Glaucoma filtering surgery has diversified in recent years and, as mentioned above, offers the best IOP reduction alternative. XEN stent has less IOP lowering effect than TE treatment. Additionally, we see a slight decrease in RNFL after 6 months for the XEN group. No significant increase in VAD could be detected, but VAD also did not decrease. This allows us to infer or at least assume that in addition to lowering IOP, filtering surgery might also stabilize the vascular architecture of eyes with glaucoma. Comparing the two procedures with each other, eyes treated with TE showed significantly higher VAD values in the SVC, DVC and optic disc after 6 months. These better results, and the more effective reduction of IOP after TE treatment, could indicate that a sufficiently lowered long-term IOP could potentially better stabilize RNFL and VAD. The positive influence of the type of glaucoma surgery on VAD is probably an additional reason, besides the higher preoperative IOP levels, for the significant differences in VAD after filtering surgery, as reported by In et al.24 and Park et al.32where the majority of patients received trabeculectomy and only a different small fraction shunt implantation.
On the other hand, the preoperative RNFL was also higher in the TE group than in the XEN group without significant difference. To get more information about the long-term effects after filtering surgery, longer follow-up periods would be needed.
Early hypotonia, occurring in four patients in this study, also did not show a significant influence on VAD in the overall outcome.
Although VAD analysis may not have been as indicative of detecting a change in postoperative vascular architecture, our correlation analysis showed a significant correlation between the extent of glaucomatous damage , given in the functional parameters, and the VAD. Specifically, MD and RNFL were significantly correlated with multiple layers of perfusion before and after surgery. Our results are in line with previous proposals made by Yarmohammadi et al.4.33 where the extent of glaucomatous damage could be associated with microvasculature. OCT-A, on its own, could therefore be useful in complementing and offering a broader spectrum of glaucomatous damage. Specifically, through the detection of early changes in the microvasculature that could potentially be seen before irreversible RNFL or visual field loss. These changes, along with classic parameters like IOP, could offer an improvement in the current management and potential screening of patients with glaucoma.
The results of this study are evaluable to some extent. Systemic vascular diseases such as diabetes and high blood pressure should be excluded from future studies. On the other hand, these comorbidities are also very common in patients with glaucoma. Thus, we show real data after glaucoma surgery. The intra-statistical analysis of the patients, also carried out in our study, makes it possible to reduce this bias compared to the inter-statistical analysis. The influence and impact of image quality on the results should also be considered as it is known to significantly influence the results.34. We have tried to minimize the risk of influence by setting a minimum to the signal strength index (not below 20). Czakó et al.34 indicate that the repeatability of OCTA parameters increased with increasing image quality. Additionally, we also did not find a significant difference in image quality between the two groups, except 3 days after surgery, as previously reported. The size of the study sample could also influence the statistical significance of the results. Due to the number of patients and the ratio of TE to XEN patients, it should be emphasized that a lack of statistically significant difference does not directly imply that the two treatment groups are the same or similar in terms of results. Additionally, the analysis and anatomical definition of the perfusion layer varies slightly depending on the OCT-A device used and is therefore defined differently in other studies. A more standardized definition and classification of the perfusion layer is needed in the future, ideally with a standardized method for VAD analysis. However, we find significant correlations comparable to the results of the current literature.