In this retrospective study, we compared and analyzed the effect of AGV implantation with and without guided tube insertion into the AC on corneal ECD and surgical outcomes. Previous studies have reported corneal ECD loss after AGV implantation in the AC.14,15 Kim et al.14 reported a 10.5% reduction in the central ECD after 12 months, and Lee et al.15 reported a 15.3% and 18.6% reduction in average ECD at 12 and 24 months after surgery, respectively.
Various hypotheses have been proposed to explain the mechanism leading to corneal ECD reduction after tube shunt surgery. Mechanical factors include the foreign body effect of the silicone tube, progressive tube migration, peripheral anterior synechiae (PAS), and transient tube-corneal/uveal contact with blinking and rubbing.12,14,16−18 Jets of aqueous humor fluid through the tube occurring in sync with heartbeats could result in ECD loss near the tube.19 Chronic inflammation caused by the silicone tube and chronic trauma results in increased endothelial cell permeability and depletion of nutrients and oxygen, leading to corneal edema.19,20 Moreover, the aqueous humor protein concentration was increased 10-fold in a previous study, suggesting an impairment in the blood-aqueous barrier to allow oxidative, apoptotic, and inflammatory proteins to enter the AC and alter the aqueous humor environment.21,22
In addition to tube-related factors, other factors can influence corneal ECD, including age, toxicity from glaucoma medication, high preoperative IOP, longer duration of high IOP before surgery, higher number of previous intraocular surgeries, and history of uveitis.23 Differences in ECD reduction have been reported according to the type of glaucoma, wherein eyes with angle closure glaucoma and pseudoexfoliation glaucoma showed significantly decreased ECD.24,25
Efforts to minimize the reduction in corneal ECD after tube shunt surgery are ongoing. AGV tube insertion into the CS using the iris as a mechanical barrier has significantly lower rates of ECD change than tube insertion in the AC in previous studies.26–28 However, tube implantation in the CS may be related to a higher incidence of postoperative hemorrhage, since the CS shows greater vascularization than the AC,29 as well as intraocular lens (IOL)-related complications and capsular bag stability and zonular weakness should also be considered, and the structural difficulty of tube insertion in the CS may be present in Asian eyes with relatively small axial lengths. Placing the tube on the par plana can minimize the effect of the tube on the corneal endothelium but is limited by the prerequisite that previous vitrectomy must be performed.
In our study, no significant differences were observed in the demographic data between the gAGV and ngAGV groups, indicating that the influence of nonsurgical factors such as age, preoperative IOP, or glaucoma types that affect corneal ECD loss can be excluded. Comparisons of surgical results also showed no significant differences in postoperative IOP and the number of glaucoma medications, including topical CAIs between the two groups, thereby excluding the influence of these factors on the corneal endothelium. We also assessed postoperative inflammation by grading AC cells up to 1 month and found that postoperative inflammation in the AC did not increase in the gAGV group in comparison with that in the ngAGV group.
Postoperative complications and surgical times were compared to ensure the safety and efficiency, respectively, of the guidance technique. Surgical time was measured in seconds only for the part where the surgical technique of the two groups was different; a spatula and a guiding stent were used in the gAGV group while the tube was directly inserted into the ngAGV group. Guided AGV implantation took an average of 30 s longer than simple tube insertion because the former required more steps; however, based on the standard deviation, the guided procedure could be completed without much deviation from the mean time. The surgical time was relatively short in many eyes in the ngAGV group, but the procedure took much longer in some eyes in which the tube was not inserted into the desired position, or the tube was bent and did not enter the AC well. Postoperative complications were not significantly different between the two groups except for flat AC, which might be caused by leakage of the sclerotomy site after multiple punctures to obtain the desired tube position in the ngAGV group. Therefore, guided AGV implantation can be considered a relatively safe surgical technique that requires little additional time and sometimes saves time by facilitating tube insertion.
A guidance technique was proposed with the goal of ideal positioning of the AGV tube in the AC as far as possible from the cornea and parallel to the iris. To assess the tube position, the tube parameters were measured and compared between the two groups. The TCD and TCA are both considered important tube parameters, but the TCD is not a fully independent variable since it is affected by the TL and TCA.9 In this study, the mean TLs in both groups were not significantly different; therefore, the influence of TL on TCD could be excluded.
Postoperative ECDs at the final visit were significantly lower than the corresponding preoperative value in eyes of the gAGV and ngAGV groups (p < 0.0001). However, the percentage of postoperative ECD loss was lower in the gAGV group, and the rate of ECD change was also lower in the gAGV group, which implied that less corneal endothelial damage occurred in the gAGV group within 2 years after surgery. The rate of ECD change with guided AGV implantation in our study appeared to be comparable to the rate of ECD change with tube insertion in the CS in previous studies (-0.36%/month in Kim et al.27 and − 0.72%/month in Zhang et al.28). However, the overall degrees and rates of ECD change in this study were relatively higher than those reported previously. One possible explanation is that approximately 30% of the eyes included in our study had concurrent phacoemulsification, although lens status was not associated with treatment failure in the TVT study.1
Changes in residual corneal ECD over time after AGV implantation showed a clear difference in the slope between the two groups, and the gap in the remaining ECD widened as the follow-up period increased. The transient rapid decline in ECD at 1 month and recovery at 3 months suggest that lost endothelium due to corneal injuries during surgery might be recovered by stem cells from a niche at the posterior limbus.30 The mean percentage loss in corneal ECD was more prominent during the first 12 months than after 12 months. In fact, the mean follow-up period until tube repositioning was 13.40 ± 6.15 months in the ngAGV group, indicating the importance of careful monitoring of the remaining ECD in the first 12 months after surgery. Eyes that underwent tube repositioning within 2 years in this study had a mean ECD loss of 54.42% and a mean rate of ECD change of 4.80%/month. These figures were similar to those in the eyes included in this study that developed corneal decompensation after 2 years, with a mean ECD loss of 50.30% and a mean rate of ECD change of 4.09%/month. This result suggests that our criteria for tube repositioning are effective in selecting eyes at risk of corneal decompensation for early management.
This study had some limitations due to its retrospective nature. First, the follow-up interval and specular microscopic examination interval were not the same among the patients. Linear regression analysis was used to calculate the rate of ECD change in each eye to compensate for this limitation, and at least three ECD measurements were performed in most eyes to obtain a reliable slope. Second, the mean follow-up period in the gAGV group was shorter than that in the ngAGV group, although the difference was not significant. As the surgical method was gradually changed from simple tube insertion to a guidance-based technique, the ngAGV group had a greater chance of undergoing long-term follow-up examinations. However, we tried to compensate for this limitation by strictly obeying the inclusion criteria and limiting the data to within two years. In addition, the mean follow-up period in the gAGV group was longer than that until tube repositioning. An advantage of the gradual change in the surgical method is that there was no selection bias between the two groups because patients who visited in the specific period underwent surgery using the same surgical method regardless of their age, sex, or glaucoma diagnosis. Therefore, we excluded cases one month after the first guided AGV implantation in consideration of the learning curve and selection bias. Third, the postoperative PAS score, which may affect the corneal endothelium was not assessed in this study. We examined the iridocorneal contact by slit-lamp examination but did not use gonioscopy. Fourth, the AGV tube may change its position and length over time in the AC; therefore, tube positioning assessment using AS-OCT may vary depending on the time of evaluation.31,32 In this study, AS-OCT was performed 3–6 months after surgery to access a relatively early tube position. Finally, as the mean time from GDD surgery to corneal decompensation was over 24 months in a previous study,8 longer follow-up periods would be necessary to further investigate the association with corneal decompensation.
In conclusion, in comparison with the conventional method, guided AGV implantation resulted in a lower corneal ECD loss and frequency of tip repositioning within 24 months. Guided AGV implantation was associated with less corneal ECD loss and a lower rate of postoperative ECD change in comparison with non-guided AGV implantation. Tube parameter analysis showed that the guidance technique could be used to position the tube further from the corneal endothelium at both distances and angles. The two groups showed no difference in the frequency of postoperative complications, and tube insertion was consistently completed within the mean surgical time. Thus, guided AGV implantation is a safe and time-efficient surgical technique that may contribute to the prevention of postoperative corneal decompensation when an AGV tube is inserted into the AC.