One-Year Efficacy and Safety of the PAUL Glaucoma Implant Using a Standardized Surgical Protocol

Précis: The PAUL Glaucoma Implant is an effective intraocular pressure (IOP)-lowering device with an acceptable safety profile. Its design as a nonvalved device with a small tube diameter using a vicryl ligation suture around the tube that is posteriorly reabsorbed instead of an ab luminal stent makes it an option in cases selected for glaucoma drainage device implantation. Purpose: The aim was to determine the 1-year efficacy and safety of PAUL Glaucoma Implant using a uniform, standardized surgical procedure. Materials and Methods: Retrospective, cohort study. Patient charts were screened between December 2018 and January 2020, with inclusion requiring a minimum follow-up of 12 months. Primary outcome was IOP-lowering at 12 months, with surgical success defined as ≤18 mm Hg and at least 30% reduction and higher than 5 mm Hg. Absolute success was achieved if no IOP-lowering medication was needed and qualified success if otherwise. Safety outcomes were also analyzed. A standardized protocol was followed in all cases, which included mitomycin C application and 1 vicryl ligation of the tube. Results: A total of 24 eyes from 21 patients fulfilled inclusion criteria. Median patient age at time of surgery was 42 years (range: 1 to 76 y). IOP decreased from 31.4 (10.0) mm Hg in preoperative period to 12.5 (4.3) mm Hg in the last follow-up (P<0.001). Qualified success criteria were fulfilled by 75% of cases, while absolute success was 33%. The mean number of IOP-lowering drugs used before surgery was 3.0 and 0.9 at the 12-month visit (P<0.001). No postoperative hypotony requiring intervention was recorded. Conclusions: PAUL Glaucoma Implant appears to be a safe and effective glaucoma drainage implant for the treatment of moderate and advanced glaucoma.

G laucoma is one of the leading causes of blindness worldwide. Despite extensive research, the only treatment able to modify progression is intraocular pressure (IOP) control. Glaucoma drainage devices (GDDs) have been considered a valuable option for treating advanced glaucoma where medical treatment and/or previous surgeries have failed. Several studies have compared the efficacy and safety of the most commonly used GDDs: the Ahmed Glaucoma Valve FP7 (AGV; New World Medical, Ranch Cucamongo, CA) and the nonvalved Baerveldt Glaucoma Implant (BGI; Abbott Medical Optics, Santa Ana, CAUSA) concluding that the BGI is more effective in long-term IOP-lowering albeit with a higher rate of hypotony-related complications/ failures. [1][2][3] In an effort to minimize hypotony-related complications during the postoperative period, a novel nonvalved device PAUL Glaucoma Implant (PGI; Advanced Ophthalmic Innovations, Singapore, Republic of Singapore) was developed. This striking feature with this GDD is its smaller internal tube diameter (0.127 mm; a third of the BGI's lumen) which acts as an intrinsic flow restrictor while preserving a similar endplate area to the other nonvalve implant (342 mm 2 ). A single study has been published using PGI implantation for refractory glaucoma, which revealed high qualified success rates (93.2%) at 12 months with low postoperative hypotony rate (9.5%). 4 However, despite the large number of patients followed in this study, the lack of a standardized technique among all glaucoma surgeons [such as the use of ligation sutures, ripcord placement or the use of mitomycin C (MMC)] was a limitation.
The present study aims to determine the 1-year efficacy and safety of PGI using a uniform, standardized surgical procedure.

Design
Retrospective cohort study. Patients from 2 surgical centers were included: Hospital Santa Maria (HSM) and Hospital Lusíadas (HLL)-Lisbon, Portugal-between December 2018 and January 2020. Charts were reviewed and patients included if followed up for a minimum of 12 months after surgery. All surgeries were performed by the same glaucoma surgeons (L.A.P. and A.D.B.), using a standardized surgical protocol in both settings.
This study adhered to tenets of the Declaration of Helsinki and all patients provided written informed consent for data retrieval as per General Data Protection Regulation.

Patients
Exclusion criteria for analysis were: pregnancy or breastfeeding, combination with cataract surgery, history of another ocular disease, and lost to follow-up within the first 12 months. DOI: 10.1097/IJG.0000000000001969 Background history and ocular-related characteristics, including baseline IOP (measured by Goldman applanation tonometry), preoperative visual acuity and number of ocular hypotensive medication were retrieved. Safety parameters and IOP data from visits on day 1; at weeks 1 and 4; at months 3, 6, 9, and 12 after surgery were retrieved from charts. The number of postoperative IOPlowering medication and surgical interventions were also examined.

Study Device
The PGI is a nonvalved silicone GDD with a length of 16.1 mm and width of 21.9 mm. The endplate is 0.4 mm thick with an area of 342 mm 2 with an inner tube diameter of 0.1 mm and outer tube diameter of 0.5 mm.

Surgical Technique
Surgeries were performed under sub-Tenon, retrobulbar or general anesthesia according to individual patient health status. The same protocol was performed in all cases in both surgical centers by the same surgical team. A 7-0 silk corneal traction suture was placed at the corneal superotemporal quadrant 1 mm from the limbus. A fornix-based conjunctival flap was fashioned, the sub-Tenon's space was dissected and wet-field cautery of the scleral bed was performed as needed. The PGI endplate was implanted 8 to 10 mm posterior to limbus, tucking the lateral wings under the superior and lateral rectus muscles and sutured to sclera using two interrupted 8-0 nylon sutures. MMC-soaked sponges (0.4 mg/mL) were applied in the sub-Tenon's space over the PGI plate for 90 seconds, followed by abundant saline rinse. A corneal paracentesis was made using a 15 degrees blade and an ocular viscoelastic device (1% sodium hyaluronate) was used to deepen the anterior chamber (AC). A 25-gauge needle was used to track a short scleral tunnel, starting 2 mm from the limbus, to enter the AC parallel to the iris plane. The tube was trimmed to an appropriate length and introduced bevel-up into the AC. A single 7-0 vicryl ligation suture was tied around the tube. No stent or ripcord were used. A patch of bovine pericardium Tutopatch (Tutogen Medical, Bavaria, Germany) or donor corneal graft (if available) was used to cover the exposed tube. The conjunctiva and Tenon's were sutured with two 7-0 vicryl suture at the limbus. Ocular viscoelastic device was washed out the AC and replaced with balanced salt solution. All patients were instructed to suspend hypotensive postoperatively and were started on antibiotic and steroid drops (Tobradex, Novartis, Basel, Switzerland) 4 times a day with slow tapering over 4 to 6 weeks.

Surgical Success and Failure Criteria
Surgical success was defined from as IOP ≤ 18 mm Hg with at least 30% reduction from baseline at 12 months. Furthermore, a minimum IOP of 6 mm Hg was also required, as per the World Glaucoma Association's (WGA) Guidelines on Design and Reporting of Surgical Trials recommendations. 5 Absolute success was considered if the conditions were met with no IOP-lowering medication while success was deemed qualified if medications were needed. 5 Failure criteria was defined as IOP above the predefined threshold or not reduced 30% below baseline, need for an additional glaucoma procedure or significant vision loss (2 or more Snellen chart lines or loss of light perception). Needling was not a criterion for surgical failure at any timepoint.

Outcome Measures
The primary outcome was to analyze the absolute and qualified surgical success rate of the PGI at 12 months. The secondary outcomes included safety parameters (intraoperative and postoperative complications).

Statistical Analysis
Demographics and clinical characteristics of patients were described using the mean (SD) or median (interquartile range: 25th percentile-75th percentile) for quantitative variables and the frequencies (percentages) for categorical variables. To compare IOP values between baseline and 12-month visits, Wilcoxon signed rank test was used. Kaplan-Meier survival analysis was used to assess probability of survival. Univariable analysis was applied to study what variables could be possible associated with surgical outcome. The level of significance P = 0.05 was considered. Data analysis was performed using SPSS software version 25.0 (Chicago, IL).

Baseline Characteristics
A total of 41 PGI implants have been performed since December 2018. At time of chart review (January 2021), 27 had achieved a minimum of 12 months after surgery. As 3 were prematurely lost to follow-up, 24 were included for analysis. Demographic data and overall baseline characteristics are listed in Table 1. As expected from a study involving long drainage devices, a large proportion of patients (n = 10, 41.6%) had undergone previous glaucoma surgeries, 3 of which had a prior GDD implantation.
Outcomes IOP was significantly reduced at the 12-month visit compared with baseline [preoperative: 31.4 (10.0) mm Hg vs. postoperative: 12.5 (4.3) mm Hg; P < 0.001]. At this last timepoint, 18 eyes (75%) had fulfilled the surgical qualified success criteria, 8 (33%) of which did not require any IOPlowering medication (absolute success criteria). Of note, 10 eyes (42%) had a hypertensive phase (defined as IOP > 21 mm Hg during the first 3 months after surgery) and were treated with topical medication. Figure 1 summarizes IOP and medication during follow-up. Medication could not be decreased in these hypertensive patients, with IOP decreasing to success criteria target in half of these patients (ie, 5 patients reached qualified success criteria). The remaining half (n = 5) ended up failing. Failure was overall reported in 6 eyes (25%). The reasons for failure were as follows: IOP above 18 mm Hg (n = 1), not reaching the predetermined 30% IOP's reduction (n = 2) and having required a surgical revision (n = 3). There were no cases of failure because of vision loss or major adverse effects. On univariate analysis no single factor such as age, sex, previous cataract or glaucoma surgeries, baseline IOP and diagnosis was significantly related to success criteria (P > 0.05). Figure 2 depicts the Kaplan-Meier cohort survival curves throughout the 12-month period.

DISCUSSION
Our study assessed the efficacy and safety profile of PGI, a recently developed GDD, during the first year of follow-up using a standardized protocol. Data on PGI use is still lacking in current literature.
Indeed, the single publication about PGI treatment outcomes described high surgical success rate (93.2%) although using a less strict threshold target (IOP target ≤ 21 mm Hg with at least 20% reduction from preoperative baseline). It also revealed a 9.5% rate of hypotony-related complications requiring intervention. 4 While our results report a somehow lower success rates (75%), our criteria for success followed the WGA guidelines report, 5 with a lower IOP target (IOP target ≤ 18 mm Hg) with a more demanding IOP reduction (at least 30% reduction from preoperative baseline), which appears more suitable for the usual candidates for GDD implantation (moderate-toadvanced glaucoma). In fact, both our case series and the published study show a rather similar IOP range (low teen), suggesting a likely similarity in the overall efficacy profile.
Our results were, however, obtained using a standardized technique, mitigating the variability of surgical techniques described in the aforementioned multicentric publication. We made use of the specificities of the reduced internal lumen of the Paul GDD, which is roughly 3 times smaller than the Ahmed or Baerveldt tubes. According to Poiseuille law, inner lumen diameter is one of the most important parameters in flow modulation. This smaller diameter would therefore act as a flow restrictor, potentially precluding early hypotony. How this presumably low flow rate would affect bleb morphology and overall efficacy was part of the purposes of this study, with 1-year data suggesting an acceptable safety profile in terms of hypotony with no significant downsize in IOP-lowering power. Nevertheless, we did use a 7-0 vicryl ligature halfway around the tube as an added flow restrictive measure. MMC can be used to prevent fibroblast proliferation, decrease conjunctival resistance and fibrovascular scarring. While its application on GDD surgery is not fully supported by evidence, as there is insufficient data on the subject, we uniformly use it in these GDD-candidate patients, traditionally with higher risk of surgical failure and a lower target IOP. Combination of these 2 strategies, tube ligation and MMC application, is the basis of our standardized protocol. The PAUL implant's tube, in contrast with standard GDDs, is smaller than the anatomical angle, increasing the chances of a proper placement away from the Descemet's membrane. Moreover, its overall smaller size could hypothetically cause less endothelial damage. While no data exists to support this theoretical assumption, it was a significant reasoning behind the choice of the PGI in a pediatric subcohort, where corneal damage is a major longterm co-morbidity of GDD surgery. 6,7 In fact, a case series recently presented in a pediatric glaucoma meeting appears to support this new trend with good results. 8 Although this was not a comparative study, it is clinically important to understand the current benchmark when analyzing our data. Comparing surgical success of PGI with AGV FP7 and BGI 350 at same follow-up time (12 mo) in ABC study, 2 the qualified success was 77% (AGV) and 64% (BGI) and absolute success was 23% (AGV) and 36% (BGI). These results were achieved under prospective clinical study conditions and considered a less strict success criteria than our real life data report. At the same timepoint (12 mo), a study on the nonvalved Susanna GDD 9 reported a success rate of 86.4%. Again, different criteria were applied. Our study aimed at complying with the WGA Guidelines and the concept that IOP targets for these devices should reflect the population they are usually used upon (advanced and/or progressing glaucoma patients). Regarding complications, the ABC study described similar rates of early complications (43% with AGV and 58% with BGI) and more serious postoperative complications (20% in AGV and 34% in BGI) in comparison with our series (54% of complications but no sight-threatening or adverse effects requiring intervention). Table 4 lists a comparative analysis of the complications described in ABC study and this case series. No early postoperative complications were described in Susanna GDD study, except a late persistent hypotony case that needed reoperation. An overall descriptive analysis of the technical characteristics of the mentioned GDD implants is shown on Table 5.
The current study has some limitations. First, our findings are based on retrospective data and surgeries were performed by the same team. In contrast, as it includes the learning curve and no inclusion/exclusion criteria other than physician discretion of the GDD choice, we feel this may actually raise interest to the reader unfamiliar with the device as it may reflect real world data. Second, the use of MMC in GDD surgical protocol should be addressed. Although the role of antifibrotic agents has been well established for in wound healing modulation to improve surgical outcomes in trabeculectomy, 10,11 the benefits of its use in GDDs remains unproven. 12,13 However, its proven cytotoxic activity on fibroblasts makes its use a tentative option in patients with a higher risk of failure. 14 Accordingly, we routinely use MMC in GDD surgery in our department. Thirdly, this study has a relatively small sample which could have had impacted our ability to detect significant factors associated with success. In addition, no data on preoperative endothelial cell count was available in this retrospective analysis. Despite no recorded case of corneal decompensation has been identified,  endothelial cell loss is of paramount importance when considering long-term implications of GDD surgery, with preoperative cell count encouraged in clinical studies for these types of surgeries. Finally, a relatively short follow-up period of 12 months can limit generalization. Hence, further studies with a long-term prospective, randomized design are needed to adequately compare the PGI with other GDDs.
In conclusion, this standardized protocol of PGI implantation (using MMC coupled with vicryl ligation around tube) appears to be an effective treatment for the management of moderate and severe glaucoma and/or showing signs of disease progression with an acceptable safety profile.