Abstract
Purpose:
To determine the early predictors of surgical success 6 months after XEN surgery among clinical parameters, including anterior segment optical coherence tomography (AS-OCT).
Methods:
A total of 31 eyes with medically uncontrolled primary open-angle glaucoma or pseudoexfoliation glaucoma was enrolled retrospectively. Using AS-OCT, XEN tip location was categorized into intraconjunctival, intratenon, or uviform at day 1 and blebs were classified into no or low, high sparse, high thick, cystic, or mixed walls at month 6. Using slit-lamp photography, blebs were classified into no or low, localized avascular, diffuse avascular, localized vascular, or diffuse vascular blebs at month 6. Surgical success was defined as an intraocular pressure (IOP) of less than 14 mm Hg.
Results:
Intraconjunctival and intratenon locations of the tip mostly created a high sparse wall, whereas the uviform type mostly created no or low wall and no or low bleb. The uviform type was linked to pseudoexfoliation glaucoma. A high sparse wall and diffuse avascular bleb showed a lower mean IOP than a high thick wall and localized vascular bleb. In the multivariate analysis, female sex and IOP at week 1 were early predictors of surgical success (8.45 times and 33.1% per 1 mm Hg-decrease, respectively).
Conclusions:
Bleb evaluation using AS-OCT is valuable to correlate tip location and bleb morphology with clinical profiles, considering that a lower early postoperative IOP is linked to surgical success.
Translational Relevance:
Bleb analysis using AS-OCT on day 1 could help to predict bleb morphology after 6 months, which is important to maintain the functioning bleb in the longer term.
All surgical procedures were performed by a skilled surgeon (S.R.) after informed consent was obtained from all patients. After topical anesthesia, the eye was prepared for the operation. We mixed 0.05 mLof 2% lidocaine with epinephrine (1:10000, 0.1 mL) was injected using a 30G needle into the superior subconjunctival space located approximately 6 mm apart from the region that the XEN tip is going to be occupied. After an injection of viscoelastics to maintain the anterior chamber via a 1-mm side port, the XEN injector was advanced through a 1.5-mm clear corneal incision toward the opposite superonasal target angle. A XEN implant was then inserted through this angle into the subconjunctival space 2 mm apart from the limbus. After confirming the allocation of XEN in the subconjunctival area, the injector was moved backward and removed gently out of the corneal incision. The proper location and length (approximately 1 mm) of XEN in the angle were checked using a surgical gonioscope, and the mobility and length (approximately 3 mm) of the subconjunctivally located part of the XEN were confirmed. Irrigation and aspiration was done to remove the viscoelastics. The corneal wounds were secured by hydrosealing using a balanced salt solution. 0.05 mL of mitomycin C (MMC) 0.4 mg/mL or 0.2 mg/mL (they were unintentionally determined owing to a national shortage of MMC supply for a couple of months in early 2019) was injected into the superonasal subconjunctival space using a 30G needle.
The postoperative medication included topical corticosteroid and antibiotic use four times a day for 1 month, followed by tapered down over 3 to 6 months. During the whole follow-up, needling treatment with or without 5-fluorouracil on the bleb was performed at the slit-lamp when the treatment target IOP was not achieved in at least one visit after 1 month postoperatively. YAG fibrinolysis was planned to be performed to relieve the obstruction of the XEN tip in the anterior chamber if it is noted. Glaucoma medications were reintroduced, considering each patient's tolerated eye drops after sufficient interviews. However, a revision surgery was performed, such as an alternative filtering surgery, in refractory cases.
This study assessed the early surgical factors of successful XEN implantation regarding the relationship between IOP control and bleb morphology. The XEN gel stent is a relatively new device in glaucoma surgery, showing promising efficacy data over the last decade. Because of the academic background of its physical structure, which is 6 mm long and has a 45 µm-lumen size, there is a major difference in the blebs created by this device in the initiation spot amid the conjunctival geographics compared with blebs created by trabeculectomy. This difference makes the XEN bleb start more posteriorly at the beginning and, therefore, shapes them flatter, which is possibly affected by eyelid movement. According to previous studies, XEN blebs were found to be significantly flatter than trabeculectomy blebs.
6,13 Our data also showed that the maximal bleb height in the success group was significantly lower than that in the failure group at postoperative month 2, which can be additionally explained by the relative outflow restriction of the XEN that limits the outflow pressure into the subconjunctival space. However, the XEN bleb is expected to follow the characteristics of a filtering bleb which has a high correlation between the IOP-lowering efficacy and bleb characteristics (e.g., size, vascularity). It is reasonable to say that the XEN bleb is also likely to fail if the bleb shows less of an internal capacity to manage the aqueous humor outflow.
In our study, the correlation of the MMC concentration to success rates (0.02% vs. 0.04%) was not statistically significant, although the 0.04% group showed higher success rates than the 0.02% group. Jampel
14 described that the viability of a human Tenon's capsule fibroblast was not affected by the exposure time of MMC. Seol et al.
15 reported that the MMC concentration did not influence the characteristics of the trabeculectomy bleb such as vascularity and bleb height. Still, there is a suggestion that a higher MMC dose may lead to higher success rates; however, the heterogeneity of the study’s design makes it difficult to confirm based on the current evidence, and additional researches are needed to clarify this. We reported that the early hypotony rate that might be linked to MMC exposure was 60% to 70%, which was higher than that reported in previous studies. Hypotony in the early postoperative period after the XEN procedure is relatively common, which seems to happen owing to peritubular flow created by the difference between the diameter of XEN and the lumen diameter of the injector needle containing XEN gel stent. In addition, the higher rate of early hypotony in our study can be further elucidated by the ethnic difference in the scleral elasticity between Korean and Caucasian populations, as reported by Park et al.
16
The goal of our study was to determine the early predictors of surgical success. In that regard, we found that IOP at postoperative week 1 could be a good predictive value, which was consistent with the findings of Cutolo et al.,
17 determining that an IOP of more than 9 mm Hg on postoperative day 1 was a strong predictor of surgical failure after XEN. Similarly, Okimoto et al.
11 reported that an IOP of less than 8 mm Hg at postoperative week 2 was linked to a high probability of surgical success after filtering surgery. In our additional analysis, we also found that an IOP of less than 7.5 mm Hg at postoperative week 1 was significantly linked to higher qualified and complete success rates, although the area under the curve was small. These results raised the issue of the possible need for early needling after XEN, in contrast with conventional trabeculectomy, which allows more early postoperative options for IOP management (e.g., releasable suture technique, laser suture lysis) than XEN. Needless to say, the intraoperative mobility check of the XEN tip should be guaranteed before discussing this issue on surgical failure.
In our study, the tip location was not related to surgical success, which was defined as IOP of less than 14 mm Hg (complete success: 33.3%, 38.9%, and 42.9% in the subconjunctival, intratenon, and uviform groups, respectively). However, it is intriguing that the success rates in the intraconjunctival and intratenon groups were relatively higher if the complete success criteria were defined as an IOP of less than 18 mm Hg (data not shown; 83.3%, 66.7%, and 42.9%, respectively). This metric may imply that the IOP of the intraconjunctival and intratenon group could be controlled within the mid-teen level, whereas that of the uviform group may widely vary. Lenzhofer et al.
18 reported that the intratenon or subtenon location of the XEN tip determined by AS-OCT images was related to surgical success. A uniform bleb in the early postoperative period and microcystic multiform bleb at postoperative month 3 according to their classification, which were similar to no or low wall and high cystic wall by our classification, respectively, seemed to be predictive of greater surgical failure at postoperative month 12. As with their report, in our post hoc analysis, we also found that the intraconjunctival and intratenon type classified by the tip location on day 1 became high sparse wall (AS-OCT) or avascular blebs (slit-lamp image), which were more linked to lower mean IOP at month 6 postoperatively, than the other types, such as high thick wall or vascular blebs. So, the proper XEN tip location may be the precondition of a successful XEN bleb.
Because the AS-OCT imaging technique enables us to distinguish Tenon's fascia from the conjunctival stroma, the XEN tip location can be determined. In our current perspective, Spectralis OCT is one of the most useful OCT tools to evaluate the filtering blebs in terms of image acquisition and resolution next to Visante OCT (Carl Zeiss Meditec AG, Germany) that used by Lenzhofer et al. Although we cannot directly compare our AS-OCT classification with that of Lenzhofer et al. in terms of different devices and settings, we agree with their opinion that a deeper XEN implant position in the conjunctiva could be more beneficial for IOP control, given that the mean maximal IOP of the intratenon group during the first 6 months was lowest among the three groups from our data.
Needling is one of the frequently performed procedures after filtering surgery, which can affect bleb appearance. The need for this procedure is a limitation of our study, similar to many bleb-related observational studies. Therefore, we considered the needling procedure as a natural course of bleb management, so that all the variables were analyzed to determine the factors affecting successful bleb formation at month 6 postoperatively, rather than considering it as a failure. Although the needling rates were not correlated to the bleb characteristics at month 6, the correlation would change if they were followed up for 12 months or more, resulting in a consequent change in bleb distribution. Bleb morphology follows a natural course over time in trabeculectomy or glaucoma drainage device blebs, as found in previous studies.
19–21 Lenzhofer et al.
5 reported that the proportion of subconjunctival separation type was highest in the first week and decreased continuously in the course of bleb development over time in XEN blebs. The changeability of a filtering bleb could be not only temporal, but also be event wise (e.g., needling, discontinuation of steroid eye drops, or readministration of glaucoma eye drops), which might be a reason that the tip location was not directly linked to surgical success in our study. Regardless of this perspective, it seems to be reasonable to exclude needling in our multivariate analysis of which the goal was to determine the early predictors for successful XEN bleb, because we generally perform needling at least 1 or 2 months postoperatively, when a filtering bleb might have developed to a certain degree.
Regarding safety profiles, the ECC and best-corrected visual acuity were checked postoperatively (ECC at months 1, 2, 3, and 6; best-corrected visual acuity at days 1 and 2, weeks 1 and 2, months 1, 2, 3, and 6). ECCs decreased in the qualified success, qualified failure, complete success, and complete failure groups by 4.7%, 7.8%, 3.8%, and 8.1%, respectively, in our study. Given that Lass et al.
22 described that ECCs decreased by 9.6% and 8.9% after a suprachoroidal microstent and cataract surgery only at month 6, respectively, our results showed that the adverse effect of XEN implantation on the corneal endothelium is minimal compared with that of routine cataract surgery. The best-corrected visual acuities of the four groups decreased the most on day 1, before slowly recovering to their preoperative levels after 2 months postoperatively, which were changes that were clearly linked to the features of IOP changes.
There is an additional point that could potentially be biased. The preoperative IOP in the intraconjunctival type on day 1 was higher than that in the other groups. It is not clear whether the intraconjunctival location of the XEN tip could be inversely affected by other variables, such as the baseline IOP. However, the delta changes in IOP from baseline to week 1 postoperatively were statistically different in between intraconjunctival type and others (−31.83 ± 13.20 mm Hg vs. −18.64 ± 7.80 mm Hg; P = 0.020), although the ratio of the decrement between the two groups did not reach statistical difference (68.2 ± 13.4% vs. 75.9 ± 15.2%; P = 0.174). This finding could imply that, by all means, the success rates might be insufficient to represent the pros of interventions over controls if we are focused on the delta changes in IOP after certain anti-glaucoma interventions.
Interestingly, we found that 7 of the 31 patients showed a rather unexpected bleb morphology on day 1, which we decided to name uviform bleb because there were numerous round-shaped protuberances covering the inner bleb wall. To the best of our knowledge, this report is the first to describe an early bleb morphology after XEN implantation that has unique features in terms of its moderate association with PXG (
n = 10; Cohen's kappa = 0.439;
P = 0.022). We speculate that these protuberances were found by the benefit of the minimal invasiveness of the XEN procedure, whereas traditional filtering surgeries such as trabeculectomy almost always disrupt the subconjunctival space and transform it into totally different limbal conjunctival geographies. Want et al.
23 reported that fibroblasts of pseudoexfoliation syndrome displayed significantly decreased autophagic flux rates, indicating a decreased clearance of autophagosomes. A diminished capacity for degradation of denatured protein may have generated the protuberances in the XEN blebs; however, two of the seven patients with uviform blebs did not have typical pseudoexfoliation materials on the pupil, which could be explained by a couple of hypothetical reasons. First, they might have subclinical signs that could not be detected with the slit-lamp examination. Some pseudophakic eyes with mild PXG are fastidious to detect and may have been overlooked. Second, the tube tips might have been deployed right in the middle of the lumen of conjunctival lymphatics. In this case, the round materials may reflect the cross-sections of the luminal raphes inside the conjunctival lymphatics or debris created during XEN implantation. Because the creation of cystic blebs can possibly be explained via this theory, the number of the implantations actually happened in this way could be more than two if the theory would be applied to other types of blebs.
In conclusion, our data suggest a possible link between XEN location in the early phase and better IOP control at postoperative month 6. Based on our analyses, an intraconjunctival and intratenon location of the XEN tip can affect the distribution of bleb morphology (a higher rate of avascular bleb or sparse wall), which is more likely to have better IOP control at month 6, although it was not directly related to surgical success. The uviform type was more likely to become no or low wall after 6 months. Additionally, female sex and a lower IOP at week 1 postoperatively indicated a greater probability of successful XEN bleb in the first 6 months. Our findings can help to identify techniques and patients that are more favorable to early surgical success of the XEN procedure by igniting a functioning bleb. Further evaluation in a long-term result would help back our study.
The authors thank Sohyeon Lee for her excellent support in data management in this study.
Conceptualization and design of the study (S.R.), collection and management of data (Y.K.), analysis and interpretation of data (Y.K., S.L., S.R.), writing of the article (Y.K., S.R.), approval of manuscript (Y.K., S.L., S.R.)
Disclosure: Y. Kim, None; S.-H. Lim, None; S. Rho, None