Efficacy of Trabeculectomy Combined With Limited Deep Sclerectomy Versus Trabeculectomy Alone: A Randomized-controlled Trial

Précis: The addition of limited deep sclerectomy (LDS) to conventional trabeculectomy increases the success rate, especially in pseudophakic eyes. Purpose: This study aimed to evaluate the efficacy of LDS in enhancing the intraocular pressure (IOP)-lowering outcomes of trabeculectomy. Design: This was a parallel-arm, single-masked, randomized-controlled trial. Methods: A total of 68 patients (68 eyes) with moderate to advanced primary open-angle glaucoma or primary angle-closure glaucoma with pseudophakia were recruited at a tertiary eye care center in Northern India and randomized into 2 groups. Thirty-five eyes in group 1 were subjected to trabeculectomy alone and 33 eyes in group 2 were subjected to trabeculectomy with LDS. All cases were supplemented with low-dose subconjunctival mitomycin-C (0.1 mg/mL for 1 min) in both groups. The primary outcome measure was IOP, and success rates were calculated for IOP ≤18/15/12 and ≥5 mm Hg. The secondary outcome measures included evaluation of bleb morphology, presence of an intrascleral aqueous lake, and supraciliary flow on ultrasound biomicroscopy. Results: The mean postoperative IOP at 12 months was 13.4±1.83 mm Hg in group 1 and 12.5±1.67 mm Hg in group 2 (P=0.04). For IOP cutoff ≤15 mm Hg, absolute success was noted in 11 (31.4%) and 22 eyes (66.6%) in groups 1 and 2, respectively (P=0.004). At 12 months, the intrascleral lake was detected in 2 (6%) eyes in group 1 and in 9 (29%) eyes in group 2 (P=0.02). Supraciliary flow was detected in 2 eyes (6.2%) in group 2 and none in group 1 at 12 months (P=0.02). In the Cox proportional hazard model, trabeculectomy with LDS was associated with a lower rate of failure (hazard ratio: 0.32, 95% confidence interval: 0.13-0.75, P=0.009). Conclusion: The addition of LDS to trabeculectomy led to lower mean IOP and higher success rates at 12 months, compared with trabeculectomy alone.

G laucoma is the second leading cause of blindness worldwide after cataract. The global prevalence of glaucoma is expected to increase from 76 million in 2020 to 111.8 million by 2040. 1 Intraocular pressure (IOP) is the only modifiable risk factor in the management of glaucoma. 2 Therefore, all modalities to treat glaucoma aim to regulate the IOP, viz., ocular hypotensive drugs, laser-assisted modalities, and filtration surgery.
Trabeculectomy is the most commonly performed glaucoma filtration surgery that lowers IOP by creating a fistula, protected by a superficial scleral flap to allow guarded aqueous outflow from the anterior chamber to the sub-Tenon space. 3 Nonetheless, fibrosis of the subconjunctival tissue continues to be a significant obstacle for the long-term success of trabeculectomy. Numerous adjuncts like antimetabolites [mitomycin-C (MMC), 5-fluorouracil], anti-vascular endothelial growth factor agents (bevacizumab), biodegradable collagen implants (Ologen), amniotic membrane, expanded polytetrafluoroethylene membrane, gel film, etc. 4,5 have been used to improve the outcome of this surgery. Antimetabolites like MMC and 5-fluorouracil are the most frequently used adjuncts due to their proven efficacy in reducing reactive scarring in the subconjunctival space. The use of low-dose MMC has now decreased the associated risk of bleb-related complications (thin avascular bleb, hypotonic maculopathy, blebitis, and endophthalmitis). 6 Although other modifications and adjuncts have shown promise, the low quality of evidence limits their applicability. 3 We have previously described a novel modification of trabeculectomy, that is, trabeculectomy combined with limited deep sclerectomy (LDS). 7 This modification presumably worked by enhancing filtration by forming an aqueous lake in the intrascleral space, thus decreasing fibrosis between the scleral bed and flap, promoting new aqueous vein formation in the intrascleral space, and promoting supraciliary/suprachoroidal filtration. 8 Our research question was whether the addition of an LDS with the creation of an intrascleral aqueous lake can improve the success of conventional trabeculectomy. To answer this question, we conducted a randomized trial to compare the IOP-lowering outcomes of these 2 techniques.

Inclusion Criteria
Patients older than 40 years of age, with moderate to severe POAG (phakic only) or PACG with pseudophakia (Hodapp, Parish, and Anderson criteria), 9 with uncontrolled IOP on maximal tolerable topical ocular hypotensive medications, and willing to provide written informed consent were recruited. Only patients with pseudophakia who underwent phacoemulsification (clear corneal incisions with the posterior chamber intraocular lens in the bag) ≥ 1 year ago were included.

Exclusion Criteria
Patients with no light perception, conjunctival, or scleral pathology (causing scarring in the superior quadrant), history of previous glaucoma surgery or laser trabeculoplasty, anterior segment pathology that could have potentially affected the IOP measurement (corneal ectasia, corneal opacity), unwilling to stop contact lens use after surgery, systemic comorbidities (diabetes, stroke, or coronary artery disease), pregnancy, and inability to attend follow-up were excluded from the study.

Study Protocol
For all the patients recruited in the study, demographic details such as age, sex, highest baseline IOP, central corneal thickness (CCT), topical hypotensive medications, history of previous ocular surgeries, and history of drug allergy were recorded at baseline. Patients were evaluated preoperatively and postoperatively on day 1, 1 week, 1, 3, 6, 9, and 12 months (Fig. 1).
Each preoperative and postoperative examination included best-corrected visual acuity (BCVA), comprehensive anterior segment evaluation, IOP, and dilated fundus evaluation. BCVA was measured using Snellen visual acuity chart at 20 feet and was converted into logarithm of the minimum angle of resolution (logMAR) notation by the standard conversion table for analysis; IOP was measured using a Goldmann applanation tonometer and the average of 3 readings was noted. CCT was measured on Lenstar (LS 900; Haag-Streit AG, Switzerland). The presence of intrascleral aqueous lake (intrascleral hypoechoic cavity) and suprachoroidal flow (suprachoroidal hypoechoic area) on ultrasound biomicroscopy (UBM) (Vumax; Sonomed Escalon 1.5.0.0, Lake Success, NY) was noted at 6, 9, and 12 months. 10 Bleb morphology was evaluated using anterior segment optical coherence tomography (ASOCT) (Casia SS-1000 OCT; Tomey, Nagoya, Japan) 11 and on slit lamp using The Indiana Bleb Appearance Grading Scale (IBAGS) 12 at 1, 3, 6, 9, and 12 months. Bleb wall thickness was measured on ASOCT as the distance between the first reflective signals of the conjunctiva to the top of the subconjunctival fluid space. 11 The need for continued follow-up and timely visits was emphasized to the patients while obtaining informed consent and then subsequently throughout the study. Each patient received a preplanned follow-up schedule, and patients missing an appointment for 4 weeks or more were considered as missed follow-up, whereas total attrition was considered as lost to follow-up.

Surgical Technique
A single surgeon (T.D.) performed all surgeries. Under peribulbar anesthesia, the eye was cleaned and draped. Corneal traction suture was placed with polyglactin 6-0 vicryl (Ethicon; Johnson & Johnson, Aurangabad, Maharashtra, India) and a fornix-based conjunctival flap was created. Sub-Tenon dissection was performed, and hemostasis was achieved as needed. Multiple sponges soaked in a 0.1 mg/mL solution of MMC were applied for 1 minute in the diffuse area under the superior bulbar conjunctiva. Subsequently, the area was irrigated thoroughly with the balanced salt solution. A partial-thickness (one third) 4×4 mm scleral flap was created in the superior limbal area. Anterior chamber paracentesis was performed and air was injected into the chamber. A 1×2 mm ostium was created, and a peripheral iridectomy was performed through the ostium using Vannas scissors. The scleral flap was secured with two 10-0 monofilament nylon sutures (Ethicon; Johnson & Johnson). The conjunctiva was closed by 8-0 polyglactin vicryl sutures (Ethicon; Johnson & Johnson). In case of trabeculectomy combined with LDS, all steps were identical, except that after creation of the partial-thickness scleral flap, a 3×3 mm deep scleral flap was raised below the initial flap and then excised, leaving behind a crater in the scleral bed (Fig. 1).

Outcome Assessment
The primary outcome measure was IOP. In addition, the Kaplan-Meier success rate was calculated for 3 IOP criteria: (A) ≤ 18 mm Hg, (B) ≤ 15 mm Hg, and (C) IOP ≤ 12 mm Hg. IOP above the cutoff criteria or hypotony (IOP < 5 mm Hg with irreversible loss of BCVA > 2 lines on the Snellen chart) at 2 consecutive follow-ups was considered as failure. Needling and need for anterior chamber reformation were not considered as a failure.
The secondary outcome measures were bleb morphology on slit lamp (IBAGS), bleb height and bleb wall thickness on ASOCT, presence of intrascleral aqueous lake, and suprachoroidal flow on UBM, 10 identifying potential risk factors for failure (age, PACG with pseudophakia, CCT, highest baseline IOP, presence of intrascleral lake, and suprachoroidal flow).

Randomization and Masking
The randomization scheme was simple randomization with allocation concealment. The participants were randomly assigned into 2 groups with the allocation ratio of 1:1. After obtaining written informed consent from the patients, their details were enclosed in an opaque brown envelope, and every envelope was allotted random numbers from 0 to 9 by an optometrist unrelated to the study. Numbers 0 to 4 were taken up in group 1 and 5 to 9 in group 2. The randomization scheme was generated using computer-generated random numbers. The participants were masked to the intervention; however, the surgeon or UBM investigator could not be masked (the latter because of the deep scleral pocket being visible on UBM). A single person (A.S.) performed all ASOCT-based and UBM-based evaluations, who was masked for IOP and IBAGS. The person recording IOP and IBAGS (D.A.) was masked to the intervention.

Statistics
For 90% power and type I error of 5%, considering a mean difference and SD of 15.3 ± 5.8 and 11.0 ± 4.0 mm Hg of IOP in 2 groups, 13 a sample size of 28 was calculated for each group (56 in total). A total of 68 eyes were subjected to followup to compensate for possible attrition. The Kaplan-Meir survival analyses were carried out to assess the success on the basis of the 3 IOP criteria described above. The log-rank test was used to compare survival curves between the 2 groups. The Cox proportional hazard regression analysis was used to evaluate potential risk factors for failure. Independent Student t test/Mann-Whitney U test was used to compare the means. The χ 2 /Fisher exact test was used for categorical variables. A P-value ≤ 0.05 was considered statistically significant.

Demography
Out of 86 eyes assessed for eligibility, 68 eyes of 68 patients with moderate to severe glaucoma were recruited and randomized to undergo either trabeculectomy with MMC (group 1, n = 35) or trabeculectomy with MMC and LDS (group 2, n = 33). All patients were South Asians from the northern part of India. Patients were recruited from April 4, 2018 to December 1, 2018 and followed up for 12 months postintervention. Baseline characteristics of both groups were statistically comparable (P > 0.05) and are summarized in Table 1. At the end of 1 year, 2 patients each in groups 1 and 2 were lost to follow-up (Fig. 2).

POAG and PACG With Pseudophakia
Success rates of POAG and pseudophakic PACG eyes were separately analyzed. For POAG eyes, no significant difference was observed in the absolute and cumulative success rates between groups 1 and 2.
For pseudophakic PACG eyes, a significant difference was noted in the absolute success rates between groups 1 and 2. For criteria A and B, absolute success was noted in 5 of 22 eyes (22.7%) in group 1 and 9 of 16 eyes in group 2 (56.2%) (P = 0.03). For criterion C, success was observed in 2 of 22 eyes (8.3%) and 7 of 16 eyes (43.7%) (P = 0.02). In terms of cumulative success, no significant difference was noted.

IBAGS and ASOCT Bleb Grading
Both the groups were similar in terms of bleb morphology according to the IBAGS. There was no statistically significant difference in bleb height and bleb wall thickness on ASOCT in both groups (Table 2).

Surgical Complications
No intraoperative complications were noted in any group. Three eyes (8.5%) in group 1 and 6 eyes (18.1%) in group 2 (P = 0.29) required anterior chamber reformation within the first week of surgery. Needling was performed for 4 eyes (11.4%) in group 1 and none in group 2 (P = 0.11), till 6 months postoperatively. No patient developed a serous choroidal detachment at any point of evaluation. There was no incidence of vision loss of > 2 lines on the Snellen chart or hypotonic maculopathy.

DISCUSSION
Trabeculectomy is a high-yield, high-risk, and highmaintenance surgery. Our goal is to increase the yield and decrease the risk and maintenance required with a simple   modification in surgical technique that can be easily adopted by ophthalmic surgeons across the globe. Scarring under and over the scleral flap in the subconjunctival space is the most common event leading to filtration failure after trabeculectomy. 14 Asian and African eyes are known to have a more aggressive scarring profile as compared with whites. 14 We carried out the current study on Asian eyes to ascertain whether a combination of trabeculectomy with LDS could achieve a better outcome than trabeculectomy alone, particularly when performed with a lower concentration of MMC.
We observed significantly lower IOP in group 2 as compared with group 1 at each follow-up (13.45 ± 1.83 vs. 12.54 ± 1.67 mm Hg, P = 0.04 at 12 mo). For IOP, ≤ 18 mm Hg absolute success rates were twice as high in group 2 compared with group 1 (33.3% in group 1 and 66.7% in group 2, P = 0.004). The cumulative success rates for the same IOP cutoff were 91.4% in group 1 and 100% in group 2 (P = 0.24).
Kayikcioglu et al 13 reported complete and qualified success rates of 83.3% and 100% for combined trabeculectomy and deep sclerectomy, compared with 63.6% and 90.9% for conventional trabeculectomy (at 12 mo and IOP < 22 mm Hg). The higher success rate reported in this study could be attributed to the smaller sample size, only POAG eyes, and higher IOP cutoff.
Feusier and colleagues combined trabeculectomy with deep sclerectomy in pediatric cases with glaucoma. After 9 years, the reported complete and qualified success rates were 52.3% and 70.6%, respectively, for IOP ≤ 18 mm Hg. 15 Recently, Sangtam and colleagues published the outcomes of combined trabeculectomy with deep sclerectomy with much broader inclusion criteria. They reported complete and qualified success rates of 43.2% and 56.8% at a median follow-up of 40 months for IOP ≤ 18 mm Hg. 16 It is noteworthy here that deep sclerectomy as mentioned in aforecited studies was the conventional technique requiring deroofing of Schlemm canal, which can be timeconsuming, requires additional surgical training, instrumentation, and a high level of surgical expertise, thereby limiting its universal applicability. This is unlike our technique of creation of a limited deep scleral pocket described here, where no anterior dissection or deroofing of Schlemm canal is required and it is thus much easier to adopt. This enhanced IOP lowering by combining these 2 techniques can be attributed to increased aqueous egress compared with stand-alone deep sclerectomy or trabeculectomy. 17 The surgically induced scleral thinning in the area of LDS increases the permeability of sclera 18 and acts as an additional outlet over and above the ostium. Once the scleral dissection is deep enough, the pressure gradient between the anterior chamber and the supraciliary space that is generated both by the colloid osmotic absorption of the choroidal vessels [19][20][21] and the outflow through the sclera and its emissaries creates a perfect condition for a pressure-dependent outflow route.
Adding trabeculectomy to modified/LDS removes the main obstacle to aqueous outflow at the level of the trabecular meshwork. Three routes of aqueous drainage are utilized by this modified surgical technique: subconjunctival, intrascleral, and suprachoroidal. 8,17 With aqueous pooling both above and below the scleral flap, the 2 most important causes of failure of filtration surgery, episcleral and intrascleral fibrosis, are dealt with. This combination technique results in an increased height of the intrascleral lake. We believe that this intrascleral lake acts as a natural spacer to prevent a secondary collapse of the flap and prevent fibrosis at the edges and below the scleral flap. In addition, pooling of aqueous in the intrascleral lake combined with deep dissection on the floor of the sclerectomy decreases the reliance on subconjunctival filtration by redirecting a part of aqueous to the suprachoroidal space and promoting supraciliary outflow. Altogether, this modified technique attempts to shift the focus from a precarious conjunctival bleb to a much safer intrascleral bleb.
It is prudent to note here that both POAG and pseudophakic PACG eyes of group 2 achieved higher success rates compared with group 1, but statistical significance at the final follow-up was seen only in pseudophakic PACG eyes, perhaps due to proportionately fewer POAG eyes recruited in the study. This difference can also be attributed to a unique immune response in PACG eyes resulting in increased levels of cytokines such as monocyte chemotactic protein-3, macrophage-derived chemokine, vascular endothelial growth factor, and reduced levels of monocyte chemotactic protein-1 and macrophage inflammatory protein-1β levels. 22 Other factors such as the short height of the trabecular meshwork in PACG eyes 23 and higher levels of inflammatory markers in pseudophakic eyes 24 are also responsible for the relatively poor outcomes with conventional trabeculectomy in PACG eyes. Therefore, when the conventional trabeculectomy was augmented with LDS, significantly better outcomes were observed. Phakic POAG eyes lacked these additional risk factors and therefore performed fairly well even with conventional trabeculectomy, although better IOP outcomes were observed at 6 months with addition of LDS. A larger sample size and longer follow-up would be required for a final conclusion on safety and efficacy in POAG eyes.
For a lower IOP cutoff value of 12 mm Hg, absolute success rates were nearly 3 times higher in group 2 versus group 1 (48.8% vs. 14.22%, P = 0.002). These results suggest that this modified surgical technique may be the choice of surgery when dealing with pseudophakic eyes, advanced glaucoma, or in other conditions where chances of failure are high.
The success rates of conventional trabeculectomy with MMC were relatively lower in this study as compared with the Primary Tube Versus Trabeculectomy (PTVT) Study: 33% versus 59%. This can be attributed to the lower IOP cutoff of 18 mm Hg in this study versus 21 mm Hg in PTVT and lowdose MMC, 0.01%, for 1 minute versus 0.04% for 2 minutes in PTVT. The rationale for using a lower dose of MMC was 2-fold. First, it would add to the safety of surgery by reducing the incidence of long-term complications and second, the efficacy of LDS would be highlighted, thereby reducing the need for MMC-augmented trabeculectomy. In future, trials can be designed to compare conventional trabeculectomy with MMC and trabeculectomy with LDS without MMC.
Intrascleral lake and suprachoroidal flow were detected using UBM and were visible as hypoechoic areas. 10 However, the presence of the intrascleral lake and suprachoroidal flow at 6 months did not significantly reduce the risk of failure. The intrascleral lake was detected in 17 (50%) and 22 (75.8%) eyes at 6 months compared with 2 (6%) and 9 (29%) eyes at 12 months in groups 1 and 2, respectively (P = 0.03 and 0.01). The decline in the visibility of these hypoechoic areas was seen in both groups with time, perhaps due to the limited resolution of UBM (40 µm for 50 MHz probe) 25 or the inadvertent fibrosis leading to complete disappearance of hypoechoic areas. The presence of suprachoroidal space in the early postoperative period has been reported to be associated with lower IOP in nonpenetrating deep sclerectomy by Choiu et al 26 and Kazakova et al 27 ; however, this has been questioned by other studies, perhaps owing to the differences of techniques and late postop evaluation. 10,28 In terms of risk factors, we excluded patients with diabetes 29 due to the risk of poor wound healing, and patients with hypertension, coronary artery disease, or cerebrovascular accident to mitigate the risk of hemorrhagic complications. From the list of remaining risk factors, viz., highest baseline IOP, age, CCT, and PACG with pseudophakia, none were associated with significantly increased risk of failure. Takihara et al 30 observed significantly poor outcomes of trabeculectomy in pseudophakic eyes at 1 year. However, the difference between the phakic and pseudophakic groups in their study was not significant in terms of postoperative antiglaucoma medications, number of laser suture lysis procedures, and postoperative complications.
In terms of complications, 3 eyes (8.5%) in group 1 and 6 eyes (18.1%) in group 2 developed a shallow anterior chamber and required anterior chamber reformation in the operation theater during the first week of surgery. Needling was performed in 4 eyes (11.4%) in group 1 and none in group 2. Kayikcioglu et al 13 observed hypotony in 2 of 12 eyes (16.6%) in the combined deep sclerectomy and trabeculectomy group, but these were managed conservatively. The use of additional sutures on the scleral flap, varying from 2 to 6, probably mitigated the risk of a shallow anterior chamber requiring surgical intervention in these eyes. In agreement with our results, none of the patients in the trabeculectomy with deep sclerectomy group required needling, whereas 3 eyes had nonfunctional blebs and 3 eyes had a cystic bleb morphology in the control group requiring needling. Early complications reported by Sangtam et al 16 included hypotony requiring intervention in 6 of 44 eyes (13.6%), hypotonic maculopathy in 2 eyes (4.4%), hyphema in 3 eyes (6.8%), and bleb leak in 2 eyes (4.5%). Needling was needed in 22 eyes (50%), of which 16 eyes were subjected to needling in the first 6 months. The higher rate of complications noted in this study was probably due to the higher concentration and duration of MMC usage along with the intrascleral spacer.
There are limitations in our study. The analysis of the intrascleral lake could not be carried out as a scale variable due to its disappearance in subsequent follow-ups. Although we attempted to minimize the detection bias by blinding the investigator (A.S.) to the IOP, blinding for the intervention was not feasible as the deep scleral pocket is frequently visible on UBM. The follow-up period was only 1 year. Since this was a single-surgeon and single-center intervention trial, these results cannot be generalized. Further studies with different surgeons and with patients of varied ethnicities are recommended to establish the efficacy and safety of this technique. Finally, a larger sample size would be advantageous to evaluate differences in subgroups of POAG and PACG.
In conclusion, the results of this randomized-controlled trial suggest a higher likelihood of achieving target IOP with trabeculectomy enhanced with LDS compared to trabeculectomy alone, particularly in pseudophakic PACG eyes. The current technique is a step towards augmenting the IOP-lowering capability of trabeculectomy and reduce the chances of failure without adding any extra expense to the procedure. In addition, the use of low-dose MMC as done in the present study can reduce the risk of associated long-term complications (especially bleb-related infections and hypotony). 6 Long-term follow-up studies could focus on evaluating and comparing these 2 procedures with additional imaging modalities such as aqueous angiography.