Article Text

Surgical outcomes of a new affordable non-valved glaucoma drainage device and Ahmed glaucoma valve: comparison in the first year
1. Vanita Pathak Ray1,
2. Divya P Rao2
1. 1 Department of Glaucoma, Centre for Sight, Banjara Hills, Hyderabad, India
1. Correspondence to Dr Vanita Pathak Ray, Department of Glaucoma, Centre for Sight, Banjara Hills, Hyderabad 500034, India; vpathakray{at}gmail.com

## Abstract

Background A new low-cost, indigenously manufactured, non-valved glaucoma drainage device (GDD) has been introduced and its design is based on the Baerveldt Glaucoma Device. We aim to demonstrate the efficacy and safety of this GDD (Aurolab aqueous drainage implant, AADI) vis-à-vis the valved device, Ahmed glaucoma valve (AGV), in the management of refractory glaucomas.

Design Retrospective, comparative, interventional.

Participants Case review of consecutive patients who underwent GDD surgery by a single fellowship-trained surgeon at a Tertiary Centre between January 2014 and November 2016.

Primary outcome measure Intraocular pressure (IOP).

Secondary Antiglaucoma medication (AGM), LogMAR best-corrected visual acuity (BCVA), complications.

Results A total of 88 eyes of 83 patients were included; 36 eyes received AGV and 52 AADI. Preoperative parameters were similar between groups. Median follow-up was 13 and 12 months for AADI and AGV, respectively. Overall success rate was higher in AADI (92.3%) vs AGV (80.5%) (p<0.001). The median IOP in mm Hg (Quartiles; IQR) (AADI 14 (10,15;5) vs AGV 16 (14,20;6)) and AGM (AADI: 0 (0,1;1) vs AGV 2 (1,2.75;1.75)) was significantly lower in the AADI group at last follow-up (p<0.001). LogMAR BCVA improved in both groups; complication rates (AADI 44.2% vs AGV 52.7%) were comparable (p=0.59).

Conclusions Both procedures were effective in reduction of IOP and need for AGM. Nevertheless, overall success rate was higher in the AADI group and IOP and number of AGM required was significantly lower in the AADI group; this affordable GDD could have a tremendous impact in the management of refractory glaucomas in low-income to middle-income countries.

• glaucoma drainage device
• GDD
• valved GDD
• Ahmed glaucoma valve
• non-valved GDD
• Aurolab Aqueous Drainage Implant

## Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

## Introduction

Glaucoma drainage devices (GDD) have now assumed an important role in the treatment of refractory glaucomas. They are used either as a primary surgical modality or a secondary procedure where trabeculectomy with or without adjunctive antimetabolite therapy has either failed or is known to yield poor results. The paradigm shift in the use of GDDs has been validated by the results of Tube Versus Trabeculectomy (TVT) Study,1 which demonstrated that patients with prior trabeculectomy had a higher success rate with GDD surgery when compared with trabeculectomy with mitomycin-C, especially when there was a history of previous incisional surgery.

Worldwide, Ahmed glaucoma valve (AGV, New World Medical, Rancho Cucamonga, California, USA) and Baerveldt Glaucoma Device (BGD, Advanced Medical Optics, Santa Ana, California, USA) are widely used.

In India, BGD is unavailable and AGV is available at an approximate price-tag of US$250; this is at a substantial cost to the patient. A prototype of the Baerveldt 350 design, the Aurolab Aqueous Drainage Implant (AADI, Aurolabs, Madurai, India), has been made available only recently and is being manufactured indigenously at a relatively lower price of US$50. Being a low-cost device, it has the potential to break the cost barrier in developing countries like ours in the quest to prevent blindness in glaucomas that are difficult to treat by conventional techniques like trabeculectomy.

There is paucity of data with regard to the efficacy and safety of AADI and no studies thus far have compared the efficacy and safety of these two devices in refractory glaucomas in Indian eyes. We have previously reported the results of AADI implantation in a cohort of 54 eyes.2 Current study reports a comparison, within the same population, the results at 1 year of follow-up of the AADI cohort with AGV implantation. Therefore, the purpose of this study was to compare the efficacy and safety of AADI and AGV in refractory glaucomas in the Indian population in the first year of implantation.

## Methods

### Study design

A retrospective, comparative, interventional study where a review of charts of consecutive patients who underwent GDD surgery from January 2014 to November 2016 by a single fellowship-trained surgeon and who had at least 3 months of documented postop follow-up, was undertaken. Those subjects requiring lower target IOP in the long term or those eyes which were able to withstand an initial high IOP for first 6 weeks were chosen to receive AADI.

Ethical clearance was obtained from the Ethics Committee; the study adhered to the principles as laid down by the Declaration of Helsinki. Informed consent for surgery was obtained from all the eligible participants.

#### Inclusion criteria

Consecutive AGV/AADI surgery with a minimum of 3 months follow-up.

#### Exclusion criteria

GDD surgery in those eyes where Goldmann Applanation Tonometry (GAT) was either not possible or compliance was poor (GDD in Kerato-prosthesis and paediatric eyes).

### Surgical methods

A fornix-based conjunctival opening is created, commonly in the superotemporal or inferotemporal quadrant. The AADI tube is checked for patency and is ligated with a 6–0 vicryl (Braided coated polyglactin 910 violet; Ethicon, Johnson & Johnson, Himachal Pradesh, India) suture and occlusion is tested. Venting incisions are made anterior to the ligated tube, approximately 3–4 pairs and vents are checked for patency. Adjacent recti are identified and hooked and the underbelly cleaned, prior to placement of the wings of the AADI underneath them. The implant plate is anchored with two interrupted 10–0 prolene suture (monofilament polypropylene blue; Ethicon, Johnson & Johnson, Himachal Pradesh, India), with the anterior edge of the plate approximately 10 mm posterior to the limbus. The suture knots are rotated into the fixation eyelets. The tube length is shortened to approximately 3 mm with a bevelled tip opening towards the cornea. A 23-gauge needle is used to create a track 2 mm behind the limbus through which the tube is inserted into the anterior chamber just anterior and parallel to the iris for anterior chamber placement and behind the iris for a sulcus placement. The tube is inserted through the needle track and secured to the sclera with a figure-of-eight 10–0 nylon suture (monofilament polyamide black, Ethilon; Ethicon, Johnson & Johnson, Himachal Pradesh, India). Almost the entire length of the tube is covered with a corneal patch graft, prepared a priori and secured over the tube with fibrin glue. The conjunctiva and Tenon are brought forward and secured back into position with fibrin glue and 8–0 vicryl (Braided coated polyglactin 910 violet; Ethicon, Johnson & Johnson, Himachal Pradesh, India) wing and continuous sutures. At the end of the procedure, a subconjunctival injection of steroid is given.

#### AGV implantation

Standard routine procedure was followed.

Patients were instructed to use topical antibiotics four times daily for 1 week with topical steroids for 6–8 weeks in tapering dose and topical cycloplegics for 1–2 weeks, as per requirement.

### Outcome criteria

Primary outcome measure was intraocular pressure (IOP) and secondary outcome measures were number of antiglaucoma medication (AGM), LogMAR best-corrected visual acuity (BCVA) and complications. Complete success was defined as IOP≥5 mm Hg and ≤21 mm Hg or reduction of IOP by ≥20% from baseline without the use of AGM. Qualified success was defined as reaching the above IOP criteria with the use of AGM. Failure was defined as the inability to meet IOP criteria, loss of perception of light, explantation of device or any additional glaucoma surgery to reduce IOP.

Hypertensive phase (HTP) was defined by a tense cystic bleb around the plate with much increased height accompanied with IOP>21 mm Hg from the third week onwards requiring AGM for IOP reduction. Subsequent reduction in bleb height, with step-down of AGM, or discontinuation, was defined as resolving or resolved HTP.

### Statistical methods

Descriptive statistics were performed to compare baseline demographic and ocular characteristics of the treatment groups. Data are presented as median (Quartiles; IQR). Snellen visual acuity was converted to LogMAR acuity for statistical analysis. Continuous and quantitative variables were analysed using Mann–Whitney U test, and discrete and qualitative variables were analysed using Pearson χ² test.

Kaplan–Meier survival analysis was performed to compare times to failure. A multivariate Cox-regression model with stepwise elimination using Akaike information criteria was performed to assess the association between study factors and time to failure.

All statistical tests were 2-sided, and statistical significance was defined as p<0.05. Statistical analyses were performed using the R software (V.2.12). R is available as Free Software under the terms of the Free Software Foundation’s GNU General Public License in source code form.

## Results

A total of 88 eyes underwent GDD surgery by a single fellowship-trained surgeon between January 2014 and November 2016 for refractory glaucoma. Fifty-two eyes underwent AADI surgery and 36 eyes underwent AGV surgery (p=0.5, χ² test).

### Baseline characteristics

Both the groups were comparable (p>0.05). Median follow-up in the AADI group was 13 months (IQR 11); median follow-up in the AGV group was longer (17 months) and hence was compared with AADI in the first year only (table 1).

Table 1

Baseline demographic and ocular characteristics between the two groups, AADI and AGV

Aetiology of glaucoma (figure 1): In both groups, secondary glaucomas were found to be the most common type of glaucoma.

Figure 1

Aetiology of glaucoma in AADI and AGV groups. AADI, Aurolab aqueous drainageimplant; AGV, Ahmed glaucoma valve.

### Intraocular pressure and antiglaucoma medication

In the AADI group, median IOP (IQR) decreased from 36.5 mm Hg (18.25) at baseline to 14 (5) mm Hg at last follow-up (p<0.001); percentage reduction was 65.7%. Similarly, in the AGV group, IOP (median, IQR) decreased from 31 mm Hg (15) at baseline to 16 mm Hg (6) at last follow-up (p<0.001) and the percentage reduction was 48.4%. The AGV group had a significantly lower median IOP than the AADI group at 1 day and 1 week follow-up visit (both p<0.001). However, the median IOP in the AADI group was significantly lower than the AGV group at every postoperative visit thereafter  (p<0.001) (table 2 and figure 2).

Table 2

Baseline and follow-up IOP and AGM for the two groups, AADI and AGV

Figure 2

Median IOP and *AGM in AADI and AGV groups at baseline and at every visit in the follow-up period. AADI, Aurolab aqueous drainageimplant; AGM, antiglaucoma medications; AGV, Ahmed glaucoma valve; IOP, intraocular pressures.

Similarly, there was a significant reduction in the need for medical therapy in both treatment groups (p<0.001) with a significantly lower need for AGM in the AADI group at every postoperative visit after 6 weeks.

Early (<3 months) postoperative complications were encountered in 18 eyes in AADI group (34.6%) and 14 eyes in AGV group (38.8%%). Late (>3 months) postoperative complications were encountered in five eyes (9.6%) in the AADI group and five eyes (13.9%) in the AGV group. The total number of complications encountered were 23 (44.2%) in the AADI group and 19 (52.7%) in the AGV group (p=0.59), although one eye may have had more than one complication. Any eye which underwent a reprocedure or which had a reduction in BCVA of >2 lines was considered a serious complication. Twenty (38.4%) such eyes in the AADI group and 9 (25%) eyes in the AGV group were found to have serious complications but this difference was not statistically significant (p=0.18) (table 3).

Table 3

Postoperative complications—early, late and serious

### Reprocedures for complications

Sixteen eyes in AADI group (30.7%) and seven eyes in AGV group (19.4%) required interventions for these complications; this difference was not statistically significant (p=0.23) (tables 3 and 4).

Table 4

Reprocedures for early and late complications in both AADI and AGV groups

## Visual acuity

Almost a similar proportion of eyes in both groups (n=39, 75% in AADI group and n=26,72.1% in AGV group; p=0.79) had either no change or an improvement in BCVA. None developed loss of perception. Eleven (21.2%) and two eyes (3.8%) in AADI and eight (22.2%) and two eyes (5.0%) in AGV group had decline of BCVA by 1–2 lines (p=0.9) and >2 lines (p=0.7), respectively. Change in median LogMAR BCVA before (1.25 (0.575,2;1.425)) and after (1.0 (0.6, 1.925;1.325)) surgery at 1-year follow-up was found to be statistically significant in the AGV group (p=0.03); this was not the case in the AADI group (presurgery 1.25 (0.6,2;1.4) vs postsurgery 0.95 (0.5,2;1.5) (p=0.39). However, the change in visual acuity between the two groups was not statistically significant (p=0.86).

## Hypertensive phase

HTP was seen more frequently in the AGV group (n=21, 58.3%) (p<0.001, χ² test), where it was detected most commonly at 6 weeks and only two eyes (5.5%) showed resolution. On the other hand, 11 eyes (21.1%), in the AADI group had HTP, and more than half of them either resolved or were resolving (6 eyes, 11.5%) (p=0.001).

## Outcomes

Complete success was seen in 34 eyes (65.4%) in AADI group with another 14 eyes (26.9%) as qualified success (total success 48 eyes, 92.3%). Complete success was seen in 8 eyes (22.2%) in AGV group and an additional 21 eyes (58.3%) as qualified success (total success 29 eyes, 80.5%). Thus, the overall success was significantly higher in the AADI group at last follow-up (p<0.001). Four eyes failed in AADI group (7.7%) and seven eyes (19.5%) in AGV group (p<0.001). All seven eyes in the AGV group failed on IOP criterion; device explantation was commoner in AADI. None of the eyes in either group failed on VA criterion.

The cumulative complete success (figure 3A) probability between AADI and AGV was compared; probability of complete success in AADI group was 66% at 12 months and in AGV group it was 22% at 12 months. When cumulative complete success probability was scrutinised in terms of the model of AGV implant, at the end of 1 year, FP7 model (n=28) survival was approximately 22% while that of FP8 model (n=8) was 12% (figure 3B).

Figure 3

(A) Kaplan-Meier survival curve of AADI and AGV: Complete success. (B) Kaplan Meier survival curve of AADI and FP7 and FP8 models of AGV: Complete success. AADI, Aurolab aqueousdrainage implant; AGV, Ahmed glaucoma valve.

In univariate analysis, AGV, male gender, phakic patients, HTP and secondary glaucoma post vitreo-retinal surgery were associated with higher failure rates. Only AGV, male gender and HTP remained significant in the multivariate analysis (table 5).

Table 5

Cox proportional hazard model to evaluate the risk factors for failure of glaucoma drainage device surgery

## Discussion

The two commonly used GDDs worldwide are AGV and BGD. The principal difference between the two devices is that the AGV has a built-in, venturi valve which regulates flow and serves to prevent hypotony in the early postoperative period. BGD on the other hand is a non-valved implant and requires the tube to be temporarily ligated to prevent early postoperative filtration, until there is adequate encapsulation around the endplate to regulate flow post autolysis of ligature (approximately 5–6 weeks). Larger surface area of BGD results in lower target IOP3; our experience has also been similar. Although there is no literature available comparing BGD 350 and AADI (also has a surface area of 350 mm2), it is presumed that the two GDDs would be equivalent, as the former has been the design inspiration for the latter.4

Our study found that both devices lowered IOP and medication use significantly from baseline. However, it was the non-valved GDD, AADI, which produced a significantly lower IOP and required fewer AGM at every visit beyond 6 and 12 weeks postoperatively respectively, when compared with AGV in the first year of follow-up. Furthermore, AADI group had a significantly higher overall success rate. Even though the overall incidence of serious complications and interventions for these complications were higher in the AADI group, they did not reach statistical significance.

There are only three non-comparative studies which have reported the efficacy and safety of AADI in Indian eyes. A prospective study in 34 eyes of paediatric refractory glaucomas with an average follow-up of 18.3±6.9 months reported the cumulative probability of success of 91.18% at 6 months and 81.7% at 18–24 months with a significantly lower IOP and AGM requirement in the postoperative period.4 Our own cohort of 54 AADI implantations2 had very similar success rate (overall success was 92.6%) at 12.1±6.3 months. The third, as yet unpublished study by Puthuran et al,5 reports the efficacy of AADI in 30 eyes with a success rate of 97% in the intermediate term. This comparative study had a cohort of 52 patients with AADI implantation; it shows comparable cumulative total success probability of 92.3% at median 13 months.

There are several studies originating in India, which have reported outcomes of AGV in Indian eyes. Das et al 6 reported approximately 85% success rate at 1 year and 88% was reported by Parihar et al 7 which are marginally higher than the success rate of AGV in our study (80.5%). This could be attributed to the FP8 models that were used in a small proportion of eyes in our study which fared worse than the FP7 models at the end of 1 year.

Several studies have compared clinical outcomes of AGV and Baerveldt implants; this has been presented in a recent meta-analysis.8 Our study was comparable to one of the larger retrospective studies conducted by Goulet et al 9 where they found that BGD has a higher success rate than AGV—92% vs 73% 1 year.

The are two large RCTs which have compared AGV and BGD; similar to BGD in the Ahmed Baerveldt Comparison (ABC) study10 and the Ahmed Versus Baerveldt (AVB) study11 at the end of 1 year, the AADI group in our study achieved a significantly lower median IOP at every visit after the early postoperative period (6 weeks). Failure rate in our study was significantly lower in the AADI group (7.7%) when compared with the AGV group (19.5%), akin to the AVB study where the cumulative probability of failure a 1 year was 43% in AGV and 28% in BGD (p=0.02).

The pooled data from the ABC and AVB studies12 at the end of 5 years also showed that BGD group produced a lower mean IOP (p=0.001) on fewer medications (p=0.007), had lower failure rates (p=0.007) and required lesser glaucoma reoperations ((p=0.006) than the AGV group. However, the BGD group carried a higher risk of hypotony (p=0.002), which was not seen in our study.

Serious complications and reoperation rates between groups were not found to be statistically significant in our study. These findings are contrary to the ABC study which showed a higher rate of serious complications in the BGD group (p=0.014) and the AVB study showed a higher reprocedure rate in the BGD group (p=0.009). We hypothesise that meticulous technique followed painstakingly by a single surgeon may have helped to minimise complications.

There was mild improvement of BCVA in both the groups; although this reached statistical significance only in the AGV group, there appeared to be no difference between the groups. This was unlike the ABC study which showed a decline in BCVA in both groups at the end of 1 year whereas the AVB study showed no difference after 1 year of follow-up (p=0.66).

We surmise that in our study, the improvement in BCVA was mostly attributable to better control of IOP. Reduction in BCVA of 2 or more lines in the AGV group occurred in two eyes, both due to corneal oedema; in the AADI group such a reduction also occurred in two eyes. None of the eyes in either group lost perception of light.

A serious complication that occurred in both groups was corneal oedema (n=4 in each group); in the AGV group, three of these eyes had previous grafts. However, two eyes developed progressive peripheral anterior synechiae (PAS) which eventually lead to failure of graft; PAS is known to be an independent risk factor for graft failure.13

In the AADI group, tube repositioning was required in one eye even though clinically there was no tube-endothelial touch demonstrable.

We encountered greater incidence of conjunctival retraction in the AADI group occurring in a cluster, in the early postoperative period; the root of this complication was traced to inadequate preparation of the corneal patch graft and this was rectified.

Plate exposure occurred more commonly in the AADI group and this ensued in eyes which had previous multiple incisional surgery, VR surgery being the most common. All four eyes had reprocedures (conjunctival autograft or choroidal drainage or both) and eventually AADI had to be explanted in two eyes.

One pseudophakic eye in the AGV group developed aqueous misdirection (AM) within 1 week of GDD surgery. It did not respond to conservative management and surgery was undertaken in the form of Irido-Zonulo-Hyaloido-Vitrectomy.14 This resulted in prompt reversal of the AM process with deepening of AC and visual salvage with IOP control.

A unique feature of GDDs is the occurrence of HTP, analogous to the encysted bleb that occurs post-trabeculectomy. It is reported to be around 40%–80% with AGV and 20%–30% with Baerveldt.15–17 Such a difference in the rate of HTP was demonstrable in our study too and was statistically significantly higher in the AGV group and was found to be a significant factor for failure in multivariate analysis. Furthermore, the rate of resolution was also lower in this group. This difference is hypothesised to be due to the inherent characteristics in the design of these devices. The venturi valve in AGV is restrictive but nonetheless permits immediate flow of aqueous before capsule formation around the plate; this allows inflammatory factors from AC to stimulate a fibrotic response in the subconjunctival space. On the other hand, deferred flow due to planned non-permanent ligature in a non-valved implant may elicit a less aggressive fibrous reaction. Predictably, HTP appeared much earlier in the AGV group; onset was much delayed in the AADI group.

The second important design difference is the size of the drainage plate, which is 184 mm2 for AGV FP7 model while AADI has a surface area almost double (350 mm2). The size of the AGV FP8 model is even lower (96 mm2) and therefore has the least impressive IOP reduction compared with larger plate sizes.3 In multivariate analysis, AGV was found to be a significant risk factor for failure, as was the occurrence of HTP, which was more frequent in AGV. Therefore, if HTP occurs in AGV, failure is a distinct probability.

There are several limitations of our study—a retrospective study design inherently has limitations. Lack of randomisation between the two GDD and limited number of patients in each group, as only those patients with at least a 3-month documented follow-up were included, may have introduced a bias, offset somewhat by inclusion of consecutive cases in the study period. However, the follow-up period too is only short-term. Furthermore, we have included both models of AGV (FP7 with the larger plate and FP8 with the smaller plate) and this may have influenced the results. Nevertheless, we have demonstrated that IOP in AGV, the costlier of the two devices, is more likely to be controlled by AGM (with lifelong cost implications), more likely to fail with greater incidence of HTP when compared with AADI.

Therefore, in conclusion, when compared with AGV, indigenously manufactured low cost device, AADI, is capable of achieving significantly lower IOP, with significantly lesser number of medications and this is sustained over the follow-up period, with comparable safety. Further follow-up is required to assess its sustainability over the longer term, as this affordable GDD will have a tremendous impact in the quest to prevent blindness due to refractory glaucomas in developing countries like ours and in other low-income to middle-income countries worldwide.

## Acknowledgments

For part data acquisition: Dr. Kiranmaye Turaga, L V Prasad Eye Institute, Vishakhapatnam, India. For assistance in statistics: Dr. Nikhil S Choudhari, L V Prasad Eye Institute, Hyderabad, India.

## Footnotes

• Contributors VPR: concept, data acquisition, data analysis and interpretation, critical review of manuscript, final approval of manuscript. DPR: data acquisition, data analysis and interpretation, drafting and critical review of manuscript, final approval of manuscript.

• Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

• Competing interests None declared.

• Patient consent Not required.

• Ethics approval Institutional Review Board/Ethics Committee at L V Prasad Eye Institute, Hyderabad, India.

• Provenance and peer review Not commissioned; externally peer reviewed.