Primary Versus Posttrabeculectomy XEN45 Gel Stent Implantation: Comparison of Success Rates and Intraocular Pressure-lowering Potential in Pseudophakic Eyes.

PRECIS
The XEN45 Gel Stent can be implanted after failed trabeculectomy without disadvantages compared with primary implantation.


PURPOSE
We aimed to compare the outcomes of XEN45 Gel Stent implantation in pseudophakic eyes after failed trabeculectomy and in pseudophakic eyes without other previous surgeries.


PATIENTS AND METHODS
In this retrospective study, we included 30 pseudophakic eyes of 30 patients who underwent XEN45 Gel Stent implantation after failed trabeculectomy (trabeculectomy group) and 60 eyes of 60 patients with primary XEN45 Gel Stent implantation (control group). The groups were matched for preoperative intraocular pressure (IOP), preoperative medication score, cup-to-disc ratio, follow-up time, visual acuity, and age at a ratio of 1:2. Eyes with concomitant eye diseases, those who had undergone previous surgery, and those with a follow-up duration <6 months were excluded. We compared the success rates on the basis of different criteria: Criteria A (IOP<21 mm Hg, IOP reduction>20%, no repeat surgery); Criteria B (IOP<18 mm Hg, IOP reduction>20%, no repeat surgery); and Criteria C (IOP≤15 mm Hg, IOP reduction≥40%, no repeat surgery).


RESULTS
After an average follow-up period of 20 months, the mean IOP was reduced from 24.4±5.7 to 14.0±4.1 mm Hg in all the study subjects (90 patients). There were no significant differences between the 2 groups in postoperative IOP, postoperative medication score, revision rate, and repeat surgery rate or success rate.


CONCLUSIONS
XEN45 Gel Stent implantation is a viable option after failed trabeculectomy. According to our results, it has no disadvantage compared with primary XEN45 Gel Stent implantation.

T rabeculectomy is regarded as the gold standard in glaucoma surgery. 1,2 Its efficacy and success rates are dependent on the development of a working filtering bleb, and therefore on the wound healing process and fibrosis in the conjunctiva. However, if trabeculectomy fails, repeat surgery may be necessary. Re-trabeculectomy can be performed at the same or at another quadrant.
It has been demonstrated that repeat surgery in eyes with previous conjunctival surgery, particularly previous trabeculectomy, results in poorer outcomes compared with those in eyes without previous surgery. [3][4][5][6] Even if repeat surgery is performed in another conjunctival quadrant, changes of the conjunctiva have been observed in apparently fibrosis-free tissue that may lead to a poorer outcome. 7 A survey of the American Glaucoma Society on treatment methods used by surgeons after failed trabeculectomy revealed an increasing trend toward glaucoma drainage devices over the last decades. 8 In most cases, glaucoma drainage devices with a wider lumen were used.
A new glaucoma drainage device with a small lumen is the XEN45 Gel Stent (Allergan, Irvine, CA). This is a new approach to filtering surgery for glaucoma. The stent is implanted ab interno and leads to the formation of a filtering bleb. Its small lumen of 45 μm was designed to avoid side effects such as hypotony. Several studies showed the intraocular pressure (IOP)-lowering potential of this technique. [9][10][11][12][13][14][15] The XEN45 Gel Stent has also been implanted after failed trabeculectomy. Hengerer et al 13 examined 226 eyes after XEN45 Gel Stent implantation, 53 of which were characterized by previous failed trabeculectomy; however, the authors did not provide an analysis of this subgroup. Karimi and colleagues reported on the implantation of XEN45 Gel Stent after failed trabeculectomy in a group of 17 eyes, but they did not compare their results with a control group. Although XEN45 Gel Stent implantation appears to be a viable option after failed trabeculectomy, it remains unknown how the success rates and IOP-lowering effect compare with those in eyes without previous surgery.
Therefore, this study compared the results of XEN45 Gel Stent implantation between eyes with failed trabeculectomy and those without previous conjunctival surgery.

Ethical Considerations
adhered to the tenets of the Declaration of Helsinki. Informed consent was obtained from all patients.

Study Design and Patients
In this retrospective observational study, we reviewed our database of patients who underwent XEN45 Gel Stent implantation and identified those with pseudophakic eyes with a previously failed trabeculectomy. We decided to focus on pseudophakic eyes to keep the group more homogenous. Patients with a follow-up duration of <6 months were excluded.
Furthermore, we identified patients potentially serving as a control group. Patients who had undergone previous eye surgery, including trabectome (Microsurgical Technology, Redmond, WA), iStent inject (Glaukos, CA), cyclophotocoaculation, vitrectomy, canaloplasty, and lasertrabeculoplasty, were excluded. Patients with concomitant eye diseases, such as uveitis and vascular glaucoma, and those with a follow-up of <6 months were also excluded. We matched the remaining patients with those who had undergone trabeculectomy on the basis of preoperative IOP, maximum known preoperative IOP, preoperative medication score (number of IOP-lowering medications), cup-to-disc ratio, follow-up time, visual acuity, and age at a ratio of 1:2.
The IOP was measured by Goldmann applanation tonometry. The preoperative value was the mean value of the last 3 preoperative measurements. Visual acuity was measured with standard Snellen charts.

Surgical Procedure
XEN45 Gel Stent implantation was performed, as described before. 10 The target for the stent was the upper nasal quadrant. If scarring of the upper nasal conjunctiva was obvious after trabeculectomy, we placed the XEN45 Gel Stent more nasally. First, 0.1 mL Mitomycin C (0.1 mg/ mL) was injected under the conjunctiva, 6 mm away from the limbus. Next, a temporal paracentesis and a paracentesis at the 5 or 7 o'clock position were made. After filling the anterior chamber with high-viscosity viscoelastic, the stent was placed with the injector. The apex of the injector was pushed through the trabecular meshwork and through the sclera, aiming a 3-mm distance from the limbus. The stent was injected under the conjunctiva and the injector removed from the anterior chamber. The position of the stent was confirmed by gonioscopy, and the viscoelastic substance was removed by irrigation.
The goal of the surgery was to regulate the IOP without IOP-lowering medication. Patients without sufficiently reduced IOP postoperatively underwent surgical revision according to their individual situation. Following our concept as described above, we performed open conjunctival revision instead of needling in all these patients. 10,16 The revision started with preparation of the conjunctiva from the limbus, as we would do during trabeculectomy. The scar tissue was removed from the stent, and the conjunctiva was fixed at the limbus with two 9.0 absorbable sutures. We allowed one revision for each patient before revision surgery was classified as repeat surgery and led to a failure according to criteria A to C. Mitomycin C was applied in 9/27 eyes during revision surgery.

Outcome Measurement
The primary outcomes of the study were the changes in the IOP and medication score at the end of the follow-up period and the surgery success rates as well. To measure the success rates, we used different criteria: Criteria A (IOP < 21 mm Hg, IOP reduction > 20%, no repeat surgery), Criteria B (IOP < 18 mm Hg, IOP reduction > 20%, no repeat surgery), and Criteria C (IOP ≤ 15 mm Hg, IOP reduction ≥ 40%, no repeat surgery). Criteria A and B were chosen according to the Tube versus Trabeculectomy Study, while Criteria C were chosen according to the criteria of the World Glaucoma Association. 17,18 Statistical Analysis Statistical analysis was performed using the software package GraphPad Prism 6.0 (GraphPad Software Inc., San Diego, CA). The Mann-Whitney U test, Fisher exact test, and log-rank tests were applied. A P-value <0.05 was regarded as statistically significant.

RESULTS
Among the 664 patients in our database who had undergone XEN45 Gel Stent implantation, 44 patients were identified to have pseudophakic eyes with a previous failed trabeculectomy. There were 14 patients with a follow-up duration <6 months, and hence they were excluded. Thus, 30 eyes of 30 patients were included for analysis. The mean time from trabeculectomy to XEN45 Gel Stent implantation was 8.4 years (SD 6.8, range: 2 to 35 y). Mitomycin C was used during the previous trabeculectomy in 25 of 30 eyes. Of the 364 pseudophakic eyes identified as the control group, 109 eyes of 109 patients were included on the basis of study criteria. Finally, after matching, there were 30 eyes of 30 patients in the trabeculectomy group and 60 eyes of 60 patients in the control group. The patients' baseline characteristics are shown in Table 1.
After an average follow-up period of 20 months, all 90 patients showed a mean IOP reduction by 42%, from 24.4 ± 5.6 to 14.1 ± 4.3 mm Hg, and a mean medication score reduction by 89%, from 2.7 ± 1.2 to 0.3 ± 0.8 mm Hg. There were no significant differences between the 2 groups in the postoperative IOP, postoperative medication score, revision rate, repeat surgery rate, or success rate (Table 2, Fig. 1). Severe complications were not recorded (Table 3). There were no cases of macular edema, leakage of aqueous humor, endophthalmitis, or retinal ablation. One patient in the trabeculectomy group showed a flat anterior chamber on the first postoperative day, choroidal effusion occurred in 2 patients in each group, intraocular bleeding occurred in 1 patient in the trabeculectomy group, and hyphema in 2 patients in each group. All these side effects were selflimiting. In 1 patient in the trabeculectomy group, it was not possible to implant the stent in the superior nasal quadrant due to extensive conjunctival scars; thus, the stent was placed in the inferior nasal quadrant.

DISCUSSION
In this study, we investigated whether a previously performed trabeculectomy leads to a worse outcome in XEN45 Gel Stent implantation. IOP reduced in all patients postoperatively, on average, by 43%, to 14.0 mm Hg, with a postoperative medication score of 0.4. The rate of revision surgery was 30%, which is concordant with that reported in previous studies. [9][10][11][12][13][14][15] Reduction in IOP after XEN45 Gel Stent implantation seems to be lower when compared with a trabeculectomy but higher than following minimally invasive glaucoma surgery modifying the anterior chamber angle. 10,17,19,20 It could be assumed that conjunctival fibrosis is of crucial importance for XEN45 Gel Stent implantation, as it is for trabeculectomy. Trabeculectomy failure is most frequently the result of subconjunctival fibrosis (bleb failure), as the postoperative wound healing process proceeds to an extent greater than desired. 6 One of the main risk factors for development of fibrosis is a previous ocular surgery, particularly conjunctival incisional surgery. 6 Therefore, retrabeculectomy has a higher risk for failure.
Nevertheless, retrabeculectomy after failed trabeculectomy is a viable option even when the surgery is performed in the same quadrant. 3,4 Law and colleagues pointed out that the long-term success rates of retrabeculectomy are lower than those of initial trabeculectomy. They still recommended retrabeculectomy, but found that the success rates dropped by ∼50% in patients after retrabeculectomy  compared with those in patients with initial surgery. It seems obvious that conjunctival fibrosis occurs at the place of previous surgery, but it was shown that the distant conjunctiva is also compromised. Broadway and colleagues analyzed samples of the conjunctiva they took at the time of retrabeculectomy. They found a significantly higher degree of fibrosis in these eyes explaining the poorer outcome after retrabeculectomy. They also found a higher number of fibroblasts, macrophages, and lymphocytes in the superficial and deep conjunctival tissues away from apparently scarred conjunctiva, which might explain the poorer outcome even if retrabeculectomy is performed in another quadrant. 6,7 Hence, it is reasonable to question the success of XEN45 Gel Stent implantation when performed in eyes with previous incisional surgery, particularly trabeculectomy. The surgery success is dependent on the state of the conjunctiva. Although high rates of postoperative needling procedures and revision surgery have been reported, ranging between 20% and 60%, [9][10][11][12][13][14][15] it remains elusive whether XEN45 Gel Stent implantation leads to poorer outcomes after failed trabeculectomy. Karimi et al 21 reported a multicenter series of 17 eyes with XEN45 Gel Stent implantation after failed trabeculectomy. Surgery was performed as a stand-alone procedure or in combination with phacoemulsification. They reported an IOP reduction of 37% with a postoperative medication score of 1 and a bleb intervention rate of 53%. The authors concluded that XEN45 Gel Stent implantation is a viable, safe, and effective option after failed trabeculectomy; however, there was no control group to compare the outcome with that in eyes without previous trabeculectomy. Hengerer and colleagues reported the results of 242 eyes after XEN45 Gel Stent implantation, of which 53 eyes had failed trabeculectomy. They reported an IOP decrease of 54%, while the medication score decreased to 0.3, and the needling rate was 27%. Thus, despite the high proportion of posttrabeculectomy patients, the results were favorable. Nonetheless, the authors did not provide a subgroup analysis, and it remains unclear whether the posttrabeculectomy eyes had worse outcomes compared with those in the other patients. To address these shortcomings, in this study, we included a control group of patients with initial XEN45 Gel Stent implantation. We found that a prior failed trabeculectomy did not lead to disadvantages with respect to the surgery success rates and complications or the number of revision surgeries. The IOP, medication score, revision surgery rate, and repeat surgery rate, and the success rates as well, according to all criteria, did not differ significantly between the trabeculectomy and the control group.
The limitations of this study are its retrospective nature and the limited follow-up and the small number of patients included in the study as well. The trabeculectomy group did slightly better than the control group, a finding that was not expected. As we performed XEN45 Gel Stent implantation during the study period in all eyes with previous failed trabeculectomy, we can exclude selection bias as an explanation for these findings. The difference in IOP between the trabeculectomy group (13.7 mm Hg) and the control group (14.2 mm Hg) was not statistically significant. This result could have been statistically significant if more patients, for example, > 1000 patients, had been included in the analysis. However, this is not a realistic option for a study with XEN45 Gel Stent implantation after failed trabeculectomy. A similar point can be made with regard to the rate of revision surgery, which was lower in the trabeculectomy group. An explanation for this might be the application of mitomycin C in the previous trabeculectomy. Another explanation might be the younger age of patients in the trabeculectomy group, which was shown to be associated with a favorable outcome after filtering surgery. However, the age difference was not significant in the current study, that is, only about 3 years compared with the ∼30-year age difference in the publication by Hoang et al. 22 From the clinician's point of view, we do not believe that there is a real advantage for eyes after trabeculectomy, but we definitely cannot rule this out. Furthermore, our strategy in patients with failed XEN45 Gel Stent implantation may differ from that of other clinics. First, our aim is that the patients no longer require local IOP-lowering therapy. If the procedure fails, we schedule a revision surgery. As pointed out earlier, we always perform open conjunctival revision, because we do not believe that needling can lower the IOP to the extent needed to diminish the need for supportive therapy. 10 Another advantage of the open revision surgery is the inspection of the stent. We once recognized a damaged stent and were able to exchange it. By performing needling, this would have been overlooked. 23 At the end of the follow-up period, some patients were still under IOP-lowering therapy. This medication was given by their local ophthalmologist or continued by the patients by mistake and was not part of our strategy.
In conclusion, we recommend the XEN45 Gel Stent as an appropriate option after failed trabeculectomy. According to our results, there is no disadvantage compared with initial XEN45 Gel Stent implantation. Therefore, we conclude that there is no need to change the position of stent injection to the lower nasal quadrant after failed trabeculectomy.