The Efficacy and Safety of Trabeculectomy after Failed Primary Glaucoma Drainage Devices

Abstract

Introduction

Trabeculectomy is not usually considered for uncontrolled intraocular pressure (IOP) after glaucoma drainage devices (GDD) because of concern that the conjunctiva has been violated and future trabeculectomy surgery is likely to fail due to fibrosis. We examined the clinical outcomes of patients who underwent a trabeculectomy after failed primary GDD.

Methods

This is a cross-sectional study of all patients who had a glaucoma drainage implant that failed or was inadequate in lowering IOP and underwent a trabeculectomy in the same eye from January 2016 to December 2022.

Results

A total of 23 eyes in 22 patients met our criteria. Average IOP [± standard deviation (SD)] prior to trabeculectomy was 21.7 ± 9.3 on 3.2 ± 1.3 medications. The length of follow-up was between 0.3 and 5.0 years with an average follow-up time of 2.2 years. At 1 year (n = 16), IOP was 11 ± 1.9 mm Hg on 1.8 ± 1.5 medications. At 2 years, the average IOP was 11.8 ± 4.6 on 1.9 ± 1.4 medications. At all follow-up points, the decrease in IOP and medication was statistically significant compared with baseline (paired t-tests; p < 0.05). Most postoperative complications self-resolved with medical management (three early wound leaks, two late wound leaks, two instances of hypotony maculopathy, and one instance of cystoid macular edema). One early wound leak required surgical repair. One eye underwent an additional GDD surgery, and three eyes underwent bleb needling. In all, 20 (87%) eyes at final visit were within two lines of their baseline vision prior to trabeculectomy. No eyes progressed to having no light perception visual acuity or had an ocular infection.

Conclusions

This study suggests that trabeculectomy after a GDD is an effective and safe option for IOP control and glaucoma medication reduction. In this small sample of surgical cases, complication and reoperation rates were comparable to published rates.

Key Summary Points

Why carry out this study?

Trabeculectomy is not usually considered after failed glaucoma drainage device (GDD) implantation due to concern of fibrosis following surgery from the violated conjunctiva.

However, a recent study has shown that trabeculectomy could be a viable option to treat uncontrolled intraocular pressure (IOP) after primary GDD.

What was learned from this study?

Our study found that, despite a more stringent criteria for success than previous studies and an older population with a larger proportion of Non-Hispanic Black patients, the success rate 1 and 2 years after trabeculectomy following the failure of a primary GDD is higher than previously described.

This study suggests that trabeculectomy is a safe and effective surgical option for patients with uncontrolled IOP after primary GDD.

Introduction

Trabeculectomy surgery for glaucoma has decreased by 26% among Medicare beneficiaries from 2008 to 2016, while the number of glaucoma drainage device (GDD) implants has increased by 20% [1]. This is possibly due to concerns about trabeculectomy-related complications. GDDs were traditionally reserved for uncontrolled glaucoma following a failed trabeculectomy surgery or when trabeculectomy is contraindicated. However, GDDs have been shown to be an effective primary glaucoma surgery after insufficient intraocular pressure (IOP) control by medications [2]. Additionally, long-term studies have shown that, in patients with prior ocular surgery, trabeculectomy has a higher 5-year failure rate with respect to IOP reduction, at 46.9%, compared with GDD, at 29.8% [3]. The same study found that trabeculectomy also had a higher rate of reoperation compared with GDD.

Risk factors of trabeculectomy failure are previous ocular surgery, Black race, long-term therapy with multiple topical anti-ocular hypertensive medications, younger age, and neovascular or uveitic glaucoma [4]. Mechanisms proposed that may explain the higher failure rate are a stronger healing response in younger patients and higher numbers of conjunctival fibroblasts, macrophages, and lymphocytes in patients with uveitis and those with prior conjunctival incisional surgery [5]. Previous studies have shown that a major cause of GDD failure is capsular fibrosis and encapsulated cyst formation at an incidence between 40% and 80% [6–9]. When primary GDDs are not able to control IOP, other surgical options are needed. Trabeculectomy is not usually considered for uncontrolled IOP after GDD because of concern that the conjunctiva has been violated and future trabeculectomy surgery is likely to fail due to fibrosis.

In this study, we hypothesized that trabeculectomy with mitomycin C (MMC) to control IOP following inadequate IOP control from primary GDD implantation is safe and effective.

Methods

This is a single-center, retrospective study of patients who underwent trabeculectomy after the placement of a primary GDD. Patients were seen between January 2016 and December 2022 at the Duke Eye Center. Epic SlicerDicer was used to identify all patients who had undergone a trabeculectomy after having had a glaucoma drainage implant in the same eye to be included in the study. Epic SlicerDicer is a data analytics tool in the Epic electronic health record (EHR) system with features such as identifying patient cohorts for research purposes or analyzing other data trends. This study was approved by the Institutional Review Board of Duke Health (Pro00112245) and followed the tenets of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act of 1996 (HIPAA). The requirement of informed consent was waived, as this study was retrospective. Surgical decision-making to undergo trabeculectomy surgery after GDD implantation was at the surgeon’s discretion.

The types of primary GDD used in patients included in this study were Baerveldt (BVT) (BVT-350mm² and BVT 250mm²; Abbott Laboratories Inc., Abbott Park, IL) and Ahmed Glaucoma Valve (model FP7; New World Medical, Rancho Cucamonga, CA) devices.

Data collected were demographics (sex, age at trabeculectomy, and race), basic ocular data (lens status, ocular comorbidities, visual field mean deviation, and test grid type), the type and stage of glaucoma, the follow-up duration, and surgical data (indication and location for primary tube, location of trabeculectomy, limbus- versus fornix-based approach, flap size, duration or quantity of MMC, application method for MMC, intraoperative and postoperative complications, laser suture lysis, and any additional surgery performed after trabeculectomy). Race is categorized on the basis of reported race in the electronic health record. IOP, visual acuity, and number of glaucoma medications were recorded preoperatively and at each postoperative visit.

All surgeries were performed by fellowship-trained glaucoma specialists or their fellow under their supervision. The surgical technique was at the discretion of the surgeon.

The primary outcomes were the IOP and the number of glaucoma medications after trabeculectomy. These were compared using three criteria as defined in Table 1. Similar criteria have been used in previous published studies assessing surgical outcomes [10]. Additionally, failure was defined as the need for additional glaucoma surgery or reoperation, ocular infection, and progression to no light perception. IOP and the number of glaucoma medications were recorded 1 day, 1 week, 1 month, 3 months, 6 months, 12 months, 18 months, and 24 months following trabeculectomy.

Table 1.

Criterion used for defining IOP and medication success for trabeculectomy with mitomycin C after failed glaucoma drainage device surgery

Criterion	Definition
Criterion A	IOP at visit ≤ 18 mm Hg and one of the following: ≥ 20% reduction of IOP or a reduction of ≥ 2 medications with IOP ≤ baseline if baseline IOP ≤ 18 mm Hg
Criterion B	IOP at visit ≤ 15 mm Hg and one of the following: ≥ 25% reduction of IOP or a reduction of ≥ 2 medications with IOP at visit ≤ baseline if baseline IOP ≤ 15 mm Hg
Criterion C	IOP at visit ≤ 12 mm Hg and one of the following: ≥ 30% reduction of IOP or a reduction of ≥ 2 medications with IOP at visit ≤ baseline if baseline IOP ≤ 12 mm Hg

IOP intraocular pressure

Results

In all, 23 eyes of 22 patients who underwent trabeculectomy with MMC after a failed primary GDD over the period from January 2016 to December 2022 were analyzed. Fornix-based trabeculectomy surgery was performed in all 23 eyes. The median time between GDD and trabeculectomy surgery was 1.5 (range: 0.2–4.5 ) years. Median follow-up after trabeculectomy was 2.2 (range: 0.3–5.0) years.

Table 2 summarizes the baseline characteristics of patients included in the study. Median age at trabeculectomy was 73 (range: 44–83 ) years. A slight majority of patients were female (54.5%). More than half the patients were Non-Hispanic Black (54.5%), the next largest group was Non-Hispanic White (31.8%), and the remaining patients were Asian (13.6%).

Table 2.

Baseline characteristics for patients undergoing trabeculectomy after GDD

Variables	Values
Number of eyes (patients)	23 (22)
Age at trabeculectomy (years)
Mean (± SD)	67.2 ± 12.09
Range	44–83
Median	73
Sex [N (%)]
Female	12 (54.5)
Male	10 (45.5)
Race/ethnicity [N (%)]
Non-Hispanic White	7 (31.8)
Non-Hispanic Black	12 (54.5)
Asian	3 (13.6)
Laterality [N (%)]
Right	12 (52.2)
Left	11 (47.8)
Diagnosis [N (%)]
POAG	10 (43.5)
Uveitic	2 (8.7)
CACG	5 (21.7)
PXE	3 (13)
Low-tension	1 (4.3)
Other	2 (8.7)
Lens status at time of surgery [N (%)]
Phakic	3 (13)
Pseudophakic	20 (87)
Intraocular pressure (mm Hg)
Mean (± SD)	21.7 ± 9.2
Number of preoperative medications
Mean (± SD)	3.17 ± 1.3

GDD glaucoma drainage device, CACG chronic angle glaucoma, POAG primary open angle glaucoma, PXE pseudoexfoliation glaucoma, SD standard deviation

The most common diagnosis was primary open-angle glaucoma (43.5%), followed by chronic angle-closure glaucoma (21.7%) and pseudoexfoliative glaucoma (13%). The remaining eyes had uveitic (8.7%), low-tension (4.3%), or other (8.7%) glaucoma. In terms of glaucoma severity staging, 18 eyes were severe, and 5 eyes were moderate. A total of 15 eyes had received an Ahmed FP7 GDD, 5 eyes received a BVT 350 GDD, and 3 eyes received a BVT 250 GDD at primary surgery. The choice for primary GDD was based on the surgeon’s preference. The majority of patients were pseudophakic at the time of trabeculectomy (87%). One patient had hand motion visual acuity prior to trabeculectomy. Before trabeculectomy, one patient had three prior GDDs, and two patients had undergone gonioscopy-assisted transluminal trabeculotomy (GATT) after primary GDD.

In all, 21 eyes received MMC by intraoperative injection, and 2 eyes had MMC applied intraoperatively by sponge.

Postoperative interventions included laser suture lysis and bleb needling in the clinic. Laser suture lysis (LSL) was performed in 19 (82%) eyes, with an average of 1.5 (range: 1–4) sutures lysed. Average time to first LSL from trabeculectomy (± SD) was 22.4 ± 19.8 (range: 5–64) days. Three eyes underwent bleb needling. One eye had needling in the operating room (OR) with MMC; one eye had needling in the office with 5-fluorouracil; and one eye had needling in the office without MMC or 5-fluorouracil followed by needling in the OR with MMC.

The mean pre-trabeculectomy IOP (± SD) was 21.7 ± 9.2 mm Hg. The postoperative mean IOP was 11.2 ± 4.0 mm Hg at 3 months, 11.9 ± 4.3 mm Hg at 6 months, 11 ± 1.9 mm Hg at 1 year, and 11.8 ± 4.6 mm Hg at 2 years (p < 0.05 for all visits; Fig. 1).

Fig. 1.

Bar graph illustrates the average intraocular pressure (IOP) at each time period with a 95% confidence interval. The change in IOP after trabeculectomy from baseline was statistically significant at all time points (paired t-test; ^*p < 0.05; ^**p < 0.005)

The mean pre-trabeculectomy number of medications (± SD) was 3.2 ± 1.3. The postoperative mean number of glaucoma medications was 1.1 ± 1.2 at 3 months, 1.5 ± 1.3 at 6 months, 1.8 ± 1.5 at 1 year, and 1.9 ± 1.4 at 2 years (p < 0.05 for all visits; Fig. 2).

Fig. 2.

Bar graph illustrates the average number of glaucoma medications at each time period with a 95% confidence interval. The change in glaucoma medications after trabeculectomy from baseline was statistically significant at all time points (paired t-test; ^*p < 0.05; ^**p < 0.005)

For the one patient with three GDD prior to trabeculectomy, mean pre-trabeculectomy IOP was 14 mm Hg. The postoperative mean IOP was 5 mm Hg at 3 months, 7 mm Hg at 6 months, 12 mm Hg at 1 year, and 12 mm Hg at 2 years. The mean pre-trabeculectomy number of medications was three. The mean postoperative mean number of glaucoma medications was zero at 3 months, two at 6 months, zero at 1 year, and zero at 2 years.

Most postoperative complications self-resolved with medical management (three early wound leaks, two late wound leaks, two instances of hypotony maculopathy, and one instance of cystoid macular edema). One eye developed an early bleb leak that required surgical repair. No patients progressed to having no light perception, and no cases of bleb infection occurred. One eye required further glaucoma surgery with the placement of an additional BVT 250 GDD after the trabeculectomy.

The success rate for Criterion A and B (Table 1) after 1 year was 81% (n = 16) and 58% (n = 12) after 2 years. Criterion C had 63% success after 1 year and 50% success after 2 years. Figure 3 presents the proportion of patients who meet each success criterion over several time points across 2 years.

Fig. 3.

This line graph illustrates the proportion of patients meeting criteria A, B, and C as defined in Table 1

In total, 20 eyes (87%) at final visit were within two lines of their baseline vision. Three (13%) patients lost more than two lines of vision.

Discussion

Trabeculectomy surgery following failure of a primary GDD is often not considered due to concern that the conjunctiva is significantly violated, potentially increasing the likelihood of failure. However, given the increasing prevalence of GDD use as the primary surgery to control IOP in glaucoma, more research is needed to determine which treatment options are suitable following GDD failure. In this study, we reviewed the use of trabeculectomy after GDD failure. In our sample of 23 eyes from 22 patients receiving trabeculectomy with MMC following a failed primary GDD, the postoperative mean IOP and number of glaucoma medications were both significantly lower than their respective pre-trabeculectomy values. The success rate for patients meeting the most stringent criteria (Criterion C; Table 1) for IOP and medication control was 50% after 2 years. These findings are consistent with our hypothesis and confirm the published literature.

We hypothesize that these favorable findings are possibly due to the location of the secondary surgery and synergistic effects of the primary GDD. Patients in our study received conjunctival incisions for GDD 5–6 mm posterior to the limbus while the trabeculectomy conjunctival incisions were all at the limbus. Besides the location of the incision being at different locations for both surgeries, each of these surgeries were done in different quadrants thus avoiding possible fibrosis due to the prior GDD. In addition, while dense scars can form around the trabeculectomy site due to increased cytokines in the aqueous humor, prior partially functioning GDDs could evacuate some of the pro-inflammatory aqueous fluid, decreasing the cytokine concentration at the site of trabeculectomy [11].

A previous study has shown similar safety and efficacy of trabeculectomy and MMC following failed primary GDD as we have demonstrated in this paper [10]. In a study on 20 eyes, Alizadeh et al. reported a success rate of Criterion C of 49.1% (± 10.8%) at 1 year of follow-up and 32.7% (± 12%) between 2 and 5 years of follow-up. In their study, the mean postoperative IOP was 9.8 ± 2.2 mm Hg at 1 year, 8.8 ± 3.2 mm Hg at 3 years, and 8.4 ± 1.5 mm Hg at 5 years. The postoperative mean number of glaucoma medications was 0.4 ± 0.2 at 1 year, 0.8 ± 0.3 at 3 years, and 1.1 ± 0.4 at 5 years. The authors of this study found a higher success rate with trabeculectomy following failed GDD compared with studies evaluating repeat trabeculectomies, a second GDD, and endoscopic cyclophotocoagulation (ECP). Some differences in success rate between Alizadeh et al. and our study may be due to differences in how criteria were met. Alizadeh et al. established failure when a given success criterion was not met at two consecutive visits. Our study had a more stringent criteria for success, as patients were required to have IOP and medication number that met the criterion at each follow-up time point measured compared with their baseline to be considered successful. Despite having more stringent criteria, we found a higher success rate for the most stringent criterion at 1 and 2 years. In terms of demographics, our study sample had a higher median age (73 years versus 64 years), included a majority non-Hispanic Black patients (54.5% versus none), and had a larger proportion of patients who were pseudophakic (87% versus 50%).

Other studies have investigated outcomes for a second GDD after initial GDD failure. Nilforushan et al., Burgoyne et al., Shah et al., and Anand et al. found that patients receiving an additional tube had a statistically significant decrease in mean IOP in the range of 33–53.2% [12–15]. In our study, we report a similar mean IOP decrease by 45.6% at 2 years from 21.7 ± 9.2 to 11.8 ± 4.6 mm Hg showing a similar efficacy of trabeculectomy and additional GDD after prior GDD. This similar efficacy could be related to both options utilizing a new surgical site, which may somewhat decrease the tendency for excessive fibrosis despite prior GDD failure [13].

Studies evaluating other surgical options following failed GDD show lower efficacy on mean IOP reduction compared with trabeculectomy. Wang et al. and Schaefer et al. compared trans-scleral cyclophotocoagulation (CPC) to a second GDD following initial GDD failure, showing that patients receiving a second GDD had greater IOP reduction compared with those who received CPC (38.8% versus 28.3% and 30.5% versus 27%, respectively) [16, 17]. Mosaed et al. evaluated patients treated with Trabectome surgery following failed GDD and reported a decrease of mean IOP by 34.6% (p = 0.05) [18]. These reported rates of mean IOP reduction for CPC and Trabectome surgery are less than what we have found in this study for trabeculectomy following failed primary GDD. The lower efficacy of such trabecular bypass procedures in patients with prior failed trabeculectomy and/or GDD is hypothesized to be the result of strong wound-healing responses [18].

Some limitations of our retrospective study include the fact that some patients were lost to follow-up and a lack of longer-term data. By 24 months of follow-up, the number of patients remaining was 12. In addition, there were limited data beyond 24 months for patients included in this study, limiting our ability to discuss long-term complications or success rates. Selection bias is possible in this study, as patients were not randomly assigned to receive primary GDD or secondary trabeculectomy and had these procedures because of the judgment of their surgeon, who may have predicted an acceptable postoperative prognosis. This study was conducted on patients who received treatment at one tertiary care center, so the results may not be generalizable to other care settings.

Conclusion

The data from this study support previous literature that suggests trabeculectomy with MMC is a reasonable surgical option following primary GDD failure, with results comparable to or better than other surgical options. The small sample of patients who underwent trabeculectomy after primary GDD failure had significantly lower postoperative IOP and number of medications, possibly due to trabeculectomy being done at a new surgical site and the synergistic effects of a partially functioning primary GDD.

Author Contributions

Daniel Wang, Zhuangjun Si, Sanjay G. Asrani, Joanne C. Wen, and Divakar Gupta contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Daniel Wang with the supervision of Zhuangjun Si and Divakar Gupta. The first draft of the manuscript was written by Daniel Wang and Zhuangjun Si; Sanjay G. Asrani, Joanne C. Wen, and Divakar Gupta commented on previous versions of the manuscript. Daniel Wang, Zhuangjun Si, Sanjay G. Asrani, Joanne C. Wen, and Divakar Gupta read and approved the final manuscript.

Funding

No funding or sponsorship was received for this study or publication of this article.

Data Availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Daniel Wang, Zhuangjun Si, Joanne C. Wen, and Divakar Gupta declare that they have no competing interests. Sanjay G. Asrani is a Section Editor of Ophthalmology and Therapy. Sanjay G. Asrani was not involved in the selection of peer reviewers for the manuscript nor any of the subsequent editorial decisions.

Ethical Approval

This study was approved by the Institutional Review Board of Duke Health (Pro00112245) and followed the tenets of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act of 1996 (HIPAA). The requirement of informed consent was waived, as this study was retrospective. Patients were not involved in the study trial design or dissemination of results.