Efficacy and Safety of the Susanna Glaucoma Drainage Device After 1 Year of Follow-up

Precis: This study found a rate of failure of 13.6% of the new Susanna Glaucoma Drainage Device (SGDD) after 1 year of follow-up. Severe complications occurred in 1 patient (4.5%). Purpose: The purpose of this study was to determine the success and complications rates of the SGDD implant in refractory glaucoma. Patients and Methods: This is a retrospective study. Medical records from all consecutive refractory glaucoma patients (failed trabeculectomies with the use of adjunctive mitomycin) followed in a private clinical practice were included in the study. All patients had undergone SGDD implant surgery from September 2016 to July 2019 were included. All surgeries were done by the same surgeon (R.S.). Primary outcome was surgical failure, defined as intraocular pressure (IOP) >18 mm Hg and/or <20% IOP reduction from baseline, IOP<5 mm Hg, reoperation for glaucoma, need of implant removal or loss of light perception vision. Secondary outcomes included mean IOP, use of supplemental medical therapy, and complications. Results: In all, 22 eyes were analyzed. Mean patient age was 63±15 years (range: 27 to 87 y). Mean IOP decreased from 23±7 to 11±8 mm Hg (P<0.001, paired t test) at an average of 18±9 months after the tube implant. The mean SD number of glaucoma medications was reduced from 3.3±1 to 1.5±1.2 at the last postoperative visit (P<0.01, paired t test). No early postoperative complications occurred. There was 1 case of late persistent hypotony in a patient with previous endocyclophotocoagulation that was solved with tube ligature with silk 8-0. Failure because of high IOP occurred in 2 cases. There was no extrusion or erosion of the tube and the plate. Conclusion: The SGDD presented a 13.6% failure rate (86.4% success rate) with very few complications, being an efficient and safe alternative for refractory glaucoma.

R eduction of intraocular pressure (IOP) is the only proven, modifiable method to slow glaucoma progression. 1 Surgical therapy to reduce IOP is usually indicated when clinical or laser therapies fail to provide an adequate IOP control. 2 Trabeculectomy is the most commonly performed glaucoma surgery and aqueous shunts are usually employed in cases were trabeculectomy failed, or in glaucoma subtypes where trabeculectomy is unlikely to succeed. [3][4][5][6][7] In general, aqueous shunts are composed of a long silicone tube typically positioned in the anterior chamber extending to a distal plate, which is sutured to the sclera. Pressure control is related to the rate of aqueous humor drainage through the implant, which is influenced by the resistance to flow through the tube, and resistance of the fibrous capsule that forms around the plate and the plate area. 3 Drainage devices differ in surface area, shape, composition, and presence or absence of flow-restricting valves.
There are 3 available aqueous shunts in Brazil: the Ahmed glaucoma valve (AGV) (New World Medical Inc., Rancho Cucamonga, CA), the Baerveldt Glaucoma Drainage Device (Abbott Medical Optics, Santa Ana, CA), and the UF Susanna Glaucoma Drainage Device (SGDD, Ultra flexible), which is the newest implant approved in Brazil.
The Ahmed Baerveldt Comparison (ABC) 4 Study and the Ahmed Versus Baerveldt (AVB) 5 clinical trials showed better success for Baerveldt than Ahmed, but with higher complications, similarly to the literature. [6][7][8][9][10] A direct comparison between these results and the results of SGDD is not possible, as there is no head to head study comparing these devices. Furthermore, the choice of device depends on many different factors, including surgeon experience, type of glaucoma, technical difficulty, monetary cost, rate of failure, and complications. AGV and Baerveldt glaucoma implant (BGI) in Brazil are expensive, which impose a great economic burden especially for the public health system and for the poor. The new Susanna implant is made in Brazil and costs approximately one third of the price of the AGV and BGI.
One previous study investigating Susanna implant safety showed low rates of serious complications, with 2 cases of conjunctival erosion (3.44%) and 2 cases of late hypotony (3.44%). 11 Although promising results, this study had a relatively short-term follow-up (7.1 ± 3.8 mo) and long-term data is still not available in the literature.
The aim of this study was to evaluate the efficacy and safety of the new UF SGDD in patients with refractory glaucoma after 1 year of follow-up.

PATIENTS AND METHODS
This retrospective study was approved by the Institutional Review Board and was conducted according to the tenets of the Declaration of Helsinki. Medical records from consecutive glaucoma patients followed up in a private clinical practice were included in the study.
All glaucoma patients with previous failed trabeculectomy with adjunctive mitomycin C or extensive superior conjunctival scarring (because of previous incisional surgeries), who had undergone SGDD implant surgery from September 2016 to July 2019 were included. All included patients had at least 1 year of follow-up.
The clinical data collected included age, sex, type of glaucoma, and previous surgeries. Data of IOP and number of medications was collected before the surgery and at 1-year follow-up visit. Primary outcome was surgery failure, defined as IOP > 18 or < 5 mm Hg or < 20% reduction from baseline on 2 consecutive study visits after 3 months, need for any additional surgery for glaucoma, or loss of light perception vision. Secondary outcomes included IOP reduction, use of supplemental medical therapy, and complications.

Device
The SGDD is a nonvalved flexible silicone device with a length of 17.4 mm and a width of 14.3 mm (Fig. 1). The plate has 2 extensions (feet) measuring 4×1 mm, which allows easy fixation to the sclera. The anterior portion is fixed at 5 to 6 mm from the limbus, allowing the plate to be located at 9 to 10 mm, facilitating the implantation and theoretically decreasing the possibility of scleral perforation during implantation or long-term implant extrusion. The presence of fenestrations also reduces motility problems by inducing fibrous scar tissue to form rivets within the bleb, which takes the bleb down in the fenestration sites.
The total thickness of SGDD is 0.5 mm. The area of the Susanna silicone plate is 200 mm 2 , which is in accordance with the suggested optimal surface area. 6,12 The tube has an internal diameter of 230 μm and an external diameter of 530 μm. A thinner tube is expected to reduce the likelihood of exposure and the risk of touching the corneal endothelium, iris, or lens. It is inserted into the anterior chamber through the intrascleral path previously made with a 26.5 G needle, instead of the 23 G used for the insertion of the AGV and BVI tube, which theoretically reduces the likelihood of leakage around the tube. Moreover, with a scleral tunnel of 1.5 to 2.0 mm long made with a 26.5 G needle, there is no flat or shallow anterior chamber that may occur when using 23 G needle to make the intrascleral path to introduce the larger tube of Ahmed or Baerveldt implant. Differences between technical characteristics of glaucoma implants are summarized in Table 1.
Despite being a thinner tube than the Baerveldt tube, the flow rate through the implant is 68.2 µL/min under 10 mm Hg perfusion pressure and 26.3 µL/min under 5 mm Hg perfusion pressure. So, as the aqueous production is from 2.0 to 2.5 µL/min, ligature of the tube is necessary.

Procedure
All surgeries were performed by the same surgeon (R.S.). The surgical procedure was performed under peribulbar anesthesia. A fornix-based conjunctival flap was made, preferentially in the superotemporal quadrant. Dissection was done anteroposteriorly in the sub-Tenon plane. Mitomycin or 5-fluorouracil was not used intraoperatively or postoperatively. The plate was placed 9 to 10 mm behind the corneal limbus and the 2 feet were secured to the sclera with 7-0 silk sutures 5 to 6 mm from the limbus. In patients with previous vitrectomy and scleral buckle, the plate was placed over the scleral buckle without difficulties as the thin plate allows this approach.
The silicone tube was shortened to the desired length before insertion (usually 2 mm length over the cornea) and then the tube was ligated with a 7-0 polyglactin suture near the tube-plate junction, leaving the knot at the back of the tube (scleral face) to facilitate laser suture lysis, if necessary.
Two fenestrations were placed anterior to the ligature using a 7-0 Vicryl suture needle, followed by placement of the silicone tube into the anterior chamber through a 26.5 G needle, insulin needle scleral track 2 to 3 mm in length. The flow through the fenestrations was assessed injecting balanced salt  solution into the anterior chamber. If insufficient, a third fenestration was done anterior to the previous ones. The anterior part of the tube was covered with previously prepared human donor scleral patch graft sutured with Vicryl 8-0. The conjunctiva and the Tenon capsule were closed with 8-0 polyglactin sutures. The eye was inspected for any leaks as the anterior chamber was inflated to a proper pressure using a balanced salt solution. All patients were treated with topical corticosteroid (prednisolone eye drops) every 2 hours and antibiotics (moxifloxacin) 4 times daily during the first week. Topical atropine 1% was also administered 2 times daily for 2 weeks. Prednisolone was then tapered off slowly over 8 to 12 weeks.

Primary and Secondary Outcomes
The primary outcome was surgical failure, defined as IOP > 18 mm Hg and/or < 20% IOP reduction from baseline, IOP < 5 mm Hg on 2 consecutive study visits any time after 3 months, reoperation for glaucoma (including needling), need of implant removal, or loss of light perception vision. Laser suture lysis was not considered in glaucoma reoperation.
IOP, visual acuity (VA), rate of surgical complications, and presence of a hypertensive phase (HP) were secondary outcome measures. Complications recorded in the first month were considered early complications, whereas the ones after that were considered late complications. A serious complication was defined as any complication, early or late, that was associated with major surgery (reoperation in the operating room) to manage the complication or that was associated with a 2-line Snellen acuity decrease. HP was defined in accordance with existing literature as IOP > 21 mm Hg during the first 6 postoperative months after an initial reduction of IOP to < 22 mm Hg during the first postoperative week and tube opening, not caused by tube obstruction, retraction, or valve malfunction.

Statistical Analysis
Descriptive analysis was used to present demographic and clinical data. Preoperative and postoperative IOP and number of antiglaucoma medications were calculated and compared in each group. These continuous data were compared using the paired t test or the Wilcoxon signedrank test, depending on the data distribution analyzed with Shapiro-Wilk test. Kaplan-Meier survival analysis was used to estimate the success rate.
All data are represented as mean, SD, median (range), and frequency (%) values. All statistical analyses and computer simulations were performed using Stata, Version 16 (StataCorp LP, College Station, TX). P-values < 0.05 were considered to be statistically significant.

RESULTS
A total of 22 eyes from 21 patients with a mean age of 63 ± 15 years (median: 64 y and range: 27 to 87 y) were included. Mean follow-up was 20.8 ± 7.4 months (median: 20.25 mo and range: 12.2 to 33.5 mo). Baseline and postoperative characteristics are shown in Table 2. Mean IOP was reduced from 24.6 ± 7.3 mm Hg (median: 22 mm Hg and range: 13 to 38 mm Hg) to 11 ± 6.4 mm Hg (median: 10 mm Hg and range: 2 to 28 mm Hg) at last follow-up visit (P < 0.01, Wilcoxon signed-rank test). IOPs at baseline and postoperative visits are shown in Table 3 and Figure 2. The mean number of antiglaucoma medications used was reduced from 3.3 ± 1 to 1.5 ± 1.2 (P < 0.01, paired t test). Kaplan-Meier survival analysis is represented in Figure 3.
The mean logarithm of the minimum angle of resolution Snellen VA decreased from 0.49 ± 0.55 to 0.51 ± 0.54 at the 1-year follow-up visit; this difference was not statistically significant (P = 0.45, paired t test). Snellen VA was decreased by 2 lines or more from baseline in 2 (9%)  patients. The reason for decreased vision was the development of exudative age-related macular degeneration (4.5%) and 1 case because of recurrent uveitis in a patient with uveitic glaucoma (4.5%). Two patients (9%) presented an HP, 1 in the first month and 1 in the third month postoperatively. IOP was successfully controlled with topical hypotensive glaucoma medications in both patients. No early postoperative complications occurred, including early hypotony. There were no cases of flat anterior chamber, shallow anterior chamber with iridocorneal apposition, wound dehiscence, endophthalmitis, corneal decompensation, choroidal hemorrhage or effusion, aqueous misdirection, cyclodialysis, atrophy/ phthisis, tube, or plate extrusion.
Only 3 patients (13.6%) were considered surgical failures: 2 cases failed due to high IOP, 1 was a primary openangle glaucoma black patient and the other was a case of traumatic glaucoma. IOP at the time of failure was 28 mm Hg in both cases. Micropulse cyclophotocoagulation was performed in each patient in a subsequent visit in 1 month and was able to successfully control IOP. One patient required reoperation due to maculopathy with persistent hypotony, which was solved with 8-0 silk tube ligature performed 12 months postoperatively with no loss of VA. After tube ligature, IOP remained close to 10 mm Hg with no medications.
There was no loss to follow-up, 19 patients were followed at the clinic, and 2 by their referring doctor.

DISCUSSION
This retrospective study evaluated the efficacy and safety of the new SGDD in patients with refractory glaucoma. We found a success rate of 86.4% with at least 1 year of follow-up. Our results support the findings of a previous prospective study, 11 which reported a rate of success of 86% during a 6-month followup period.
Although a direct comparison with other glaucoma drainage device (GDD) was not performed, previous studies with other GDDs have shown similar success rates. [13][14][15][16][17][18][19] To have a valid comparison between different studies, it is important to take into account the study population, methodology, and design.
In the Primary Tube Versus Trabeculectomy (PTVT) Study, the cumulative probability of failure of the BGI at 1 year was 17.3% 20 and in the ABC Study it was 16.4% and 14.0% in the AGV and the BGI groups, respectively. 4 Both studies included patients with similar age and baseline IOP to the present study, which is relevant as a risk factor analysis from pooled data from these studies identified younger age and lower preoperative IOP as significant predictors of tube shunt failure. 21 Furthermore, authors considered as failure IOP > 21 mm Hg; therefore, we could expect equal or higher rates of failure if the threshold was IOP > 18 mm Hg, as in our study.
The randomized AVB Study observed a higher cumulative probability of failure at 1 year; 43% in the Ahmed group and 28% in the Baerveldt group. 5 In contrast, the Tube Versus Trabeculectomy (TVT) Trial found the lowest failure rates (3.9%). 22 This can be partially explained by the study population, which did not include eyes with a high risk of tube failure, previous retinal surgery, or previous cyclodestructive procedures. More recent publications of the TVT Study with 3 and 5 years of follow-up have shown an increased failure rate of 24.0% 23 and 31.8%, 24 respectively.  Indeed, the trend toward an increase in failure rates with time is well established. [25][26][27] In our study, the most common cause for failure was inadequate IOP reduction and there were no early surgical complications. IOPs were controlled with topical medication (< 24 mm Hg) until the spontaneous opening of the tube when tube fenestrations alone were not sufficient to reduce IOP. Previous studies have found significant rates of early complications varying from 20% to 58%. 20,28,29 The absence of early postoperative complications in our study could be partially because of the fact that all surgeries were done by one experienced surgeon in long-tube glaucoma implant surgery. In addition, one previous prospective study with multiple welltrained surgeons investigating the SGDD implant also found low rates of early complications (5.2%). 11 There was only 1 case (4.5%) of late hypotony that resolved with tube ligature (silk 8-0) in a patient that had previously undergone endocyclodesctructive procedure, which could predispose to hypotony. 30,31 In this case, IOP remained low with no medications after tube ligature, probably because of ciliary body failure. Biteli et al 11 observed hypotony in 5 patients (9%) after SGDD. Four of these patients were successfully treated (2 with an early filling of the anterior chamber with viscoelastic and 2 with atropine) and 1 required removal of the implant. This rate was equivalent to the TVT 32 and half of the rate reported by the ABC Study. 4 VA decreased by 2 lines or more of Snellen VA in 2 (9%) of the patients and there was no loss of light perception vision. One patient had uveitic glaucoma and at 1-year follow-up presented with inflammatory activity and the other developed exudative age-related macular degeneration. Both patients had no other complications. These rates of VA loss are lower than the 32% found in the ABC 4 and TVT 22 Study groups at 1 year.
The HP after GDD implantation presents a challenge in management because of the risk of continued optic nerve damage, often in the setting of advanced refractory glaucoma. Its incidence varies and is observed more commonly after insertion of the Ahmed device (40% to 80%) than was the Baerveldt (20% to 30%) or Molteno (20% to 30%). [33][34][35] In our study, 2 patients (9%) presented an HP and they were all successfully managed with topical hypotensive medications.
In this study, no motility disorder was reported. However, Biteli et al 11 found a 9% rate of motility disorders with the SGDD, although they were all transient and did not require further treatment. Because of its size and placement beneath 2 adjacent rectus muscles, 36 motility problems with the Baerveldt 350 shunt have been reported to vary between 3% and 18%. 32,37 In theory, it is reasonable to assume that SGDD should present a lower incidence of motility problem and diplopia than the Baerveldt because of smaller plate size and lack of manipulation of extraocular muscles.
Corneal decompensation is an additional long-term problem with tubes, especially in the pediatric population. The AVB Study found that it was the most common long-term complication. 25 The TVT Study 24 reported a 16% persistent corneal edema rate at 5 years, with a 9% corneal graft rate. Several factors contribute to corneal complications such as tube positioning, motility, length, and size. The rather large tube of the Baerveldt or Ahmed implant may lead to continuous endothelial loss 38 because of a shorter tube-cornea distance. 39 A possible advantage of the current implant is the thinner tube compared with Ahmed and Baerveldt tubes, which may decrease the chances of tube-cornea contact. Endothelial cell loss was not measured; nevertheless, there was no case of corneal edema during the whole period of follow-up. Future research using the SGDD with a longer follow-up are needed to address this parameter.
The main limitations in the current study are the small sample size and retrospective design. We minimized bias by including all consecutive patients and by selecting patients operated by only one surgeon, so there is more uniformity in the surgical technique and clinical management. Another limitation is the lack of a control group, for this reason direct comparison with other GDD is not possible.
In summary, our findings suggest that the SGDD is an effective alternative for managing refractory glaucoma. Additional benefits such as market price and technical challenges should be taken into consideration. Further prospective studies and a straight comparison with other commercially available drainage devices are needed to confirm our findings.