Article Text

Download PDFPDF

One-year outcomes of stand-alone ab externo SIBS microshunt implantation in refractory glaucoma
  1. Georges M Durr1,2,
  2. Matthew B Schlenker3,
  3. Saba Samet3,
  4. Iqbal Ike K Ahmed3
  1. 1 Ophthalmology, University of Montreal, Montreal, Canada
  2. 2 Ophthalmology, Centre Hospitalier Universitaire de Montréal (CHUM), Montréal, Canada
  3. 3 Ophthalmology and Vision Sciences, University of Toronto, Toronto, Canada
  1. Correspondence to Georges M Durr, Department of Ophthalmology, Centre Hospitalier Universitaire de Montréal (CHUM), 1051 Rue Sanguinet, Montreal H2X 3E4, Canada; georgesdurr{at}


Background/Aims Determine the efficacy of stand-alone implantation of the ab externo SIBS or poly(styrene-block-isobutylene-block-styrene) microshunt with mitomycin C in glaucomatous eyes, refractory to previous subconjunctival filtering surgery, over 1 year of follow-up.

Methods Consecutive retrospective cohort of patients with intraocular pressure (IOP) above target and previous subconjunctival filtering surgery, who received the microshunt between July 2015 and April 2019. Primary outcome was a complete success, with failure defined as IOP <6 mm Hg with vision loss, >17 mm Hg or <20% reduction in IOP without medications. Secondary outcomes included thresholds of 6 to 14 mm Hg and 6 to 21 mm Hg for both complete (no medications) and qualified (with medications) success as well as qualified success for thresholds of 6 to 17 mm Hg. Risk factors for failure, IOP, medications and complications were also assessed.

Results 85 eyes of 79 patients with a preoperative median IOP of 22.0 mm Hg (IQR 18.0–29.0) on four (IQR 3–4) medications were included. Postoperative median IOP was 13.0 mm Hg (IQR 10.0–17.0) on zero (IQR 0–2) medication at 1 year. 61.0% achieved complete success and 79.7% achieved qualified success. Mild-to-moderate disease was associated with failure (adjusted HR 2.37; 95% CI 1.23 to 4.59). Needling was performed in 11.8%, and 8.2% underwent anterior chamber reformation. Complications were transient, consisting of hyphaema, choroidal detachment and hypotony maculopathy, with 7.1% of patients undergoing reoperation.

Conclusions In a group of high-risk eyes that had already failed at least one subconjunctival filtering surgery, the SIBS microshunt demonstrates reasonable surgical success over 1-year follow-up, with relatively few complications.

  • Glaucoma

Statistics from

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.


Intraocular pressure (IOP) control has been the mainstay of glaucoma treatment to decrease risk of progression and further visual field loss.1 Topical and laser treatments are at the forefront of early treatment. Ultimately, when conservative treatment fails, patients move on to subconjunctival filtering procedures. A proportion of these filters will fail over time. This subset of patients with a failed subconjunctival filtering surgery are more challenging to obtain success with a subsequent surgery.2–6 The same factors, including propensity for scarring, ethnicity and low-grade inflammation, that lead to failure for the initial surgery predispose to failure for subsequent surgery. Furthermore, scarring from previous incisions and mitomycin C (MMC) use has often affected conjunctival integrity which also influences the next potential procedure.2–6

In the Tube versus Trabeculectomy study which reviewed surgical success in refractory glaucoma, tube shunts had lower failure rates than trabeculectomy with a cumulative probability of failure (defined as IOP >21 mm Hg or less than 20% reduction below baseline IOP) of 29.8% in the tube group and 46.9% in the trabeculectomy group at 5 years. The trabeculectomy group had higher rates of hypotony and reoperations than the tube group; however, both groups had significant decreases in visual acuity (VA) at 5 years (decrease in VA of >2 lines: 46% in tube group vs 43% in trabeculectomy group).

Newer, potentially less invasive options for filtering procedures have led to a shift in treatment options. The ab interno gel stent (Xen gel stent (Allergan, Irvine, California, USA)) and the ab externo poly(styrene-block-isobutylene-block-styrene) or ‘SIBS’ microshunt (PreserFlo Microshunt (Santen, Miami, Florida, USA)) are new potential alternatives.7 8 Refractory glaucoma to previous filtering surgery poses a unique challenge to glaucoma surgeons. Newer novel surgeries have not been evaluated much in these challenging refractory eyes.9 10

The ab externo SIBS microshunt is composed of an inert biocompatible biomaterial.11 The device is 8.5 mm in length with a fin positioned 4.5 mm from the proximal tip of the microshunt bevel and has an internal lumen of 70 μm. The fin is designed to prevent migration of the implant and blocks any peritubular flow. The posterior drainage and bleb creation suggest that this implant could be considered for inferior placement, as we sometimes do with tube shunts. However, inferior blebs have been notoriously associated with higher infection rates.12–14

Little has been published on the microshunt thus far. Batlle et al 8 and Schlenker et al 15 have shown promising results in virgin eyes. However, there are no published results for eyes with previous subconjunctival filtering surgery. The purpose of this study was to determine efficacy, adverse events and risk factors for failure of a stand-alone novel ab externo microshunt with MMC in refractory glaucomatous eyes.


This study reports 1-year outcomes of an investigator-initiated, single-centre, retrospective, series of consecutive patients implanted with an ab externo SIBS microshunt with MMC between 1 July 2015 and 1 April 2019, by several surgeons at an academic hospital in Mississauga, Ontario, Canada. The protocol adhered to the tenets of the declaration of Helsinki and institutional review board approval was obtained from Trillium Health Partners. Written informed consent was obtained from all patients prior to surgery. Preoperative baseline data including demographics (age, gender, ethnicity and diabetes status) as well as ocular characteristics and history, preoperative IOP (decision IOP), glaucoma medications, glaucoma diagnosis and severity, cup-to-disc ratio, best-corrected visual acuity (BCVA), mean deviation and history of previous laser peripheral iridotomy, laser trabeculoplasty, cataract surgery or glaucoma surgery were collected.

Inclusion and exclusion criteria

Patients with any type of glaucoma having IOP above target despite maximal medical treatment and at least one previous subconjunctival filtering surgery (trabeculectomy, Ahmed glaucoma valve (AGV; New World Medical, Rancho Cucamonga, California, USA), Baerveldt glaucoma implant (BGI; Johnson & Johnson Vision, Santa Ana, California, USA), Ex-Press Mini Glaucoma Shunt Device (Alcon Laboratories, Fort Worth, Texas, USA), deep sclerectomy, non-penetrating deep sclerectomy with a collagen implant (Aquaflow Collagen Glaucoma Drainage Device; Staar Surgical Company, Monrovia, California, USA) and Xen gel stent (Allergan)) were included.

Any patient receiving a combined procedure with cataract surgery or having a history of prior retinal or corneal graft surgery was excluded. Those with less than 1 month of follow-up were also excluded.

Ab externo microshunt procedure

Our surgical technique has been previously described.15 In short, after a traction suture was placed in the cornea, the superior and inferior conjunctiva were inspected to plan the placement of the device. Areas of conjunctival scarring or antimetabolite toxicity (thin conjunctiva, large ropy vessels and difficult mobilisation of the tissue) were avoided. The position of the implant (superonasal, superotemporal, inferonasal or inferotemporal) was chosen depending on the quality of the conjunctiva, areas of previous surgery and extension of their blebs, eyelid position and presence of peripheral anterior synechiae (PAS). Preference was always given to place the implant in one of the superior quadrants. Inferior placement was only made if none of the superior quadrants were available due to previous filtration surgery or poor conjunctival health. A conjunctival peritomy was then performed with Tenon’s disinsertion in the quadrant chosen. Light cautery was applied. MMC was then placed on three merocele sponges (two behind the rectus muscle bellies and one anterior under Tenon’s layer) with concentrations from 0.2 mg/mL to 0.5 mg/mL for a total of 2 min, carefully avoiding any contact with the limbus. Throughout the study period, MMC dosing was increased with increasing comfort with bleb morphology, starting with 0.2 mg/mL for the first patients, followed by 0.4 mg/mL and 0.5 mg/mL. The microshunt blebs were found to be posterior and diffuse with minimal avascularity leading to an increase in MMC concentration throughout the study period in order to increase the likelihood of surgical success. After removal of the sponges, balanced salt solution (BSS) was used to copiously irrigate the surface of the eye. A scleral tunnel was created 3 mm from the limbus with a 1 mm microknife in one motion entering the anterior chamber (AC) just above the trabecular meshwork. A side-port was then created. The implant was positioned in the AC and primed with a BSS injection in the AC. Typically, implants are placed in between the two rectus muscles. Tenon’s and conjunctiva were then closed in separate layers ensuring the implant is straight and unobstructed at the end of the case with no leaks and a bleb starting to form.

Postoperative management

Patients were instructed to stop all glaucoma medication (topical and oral) and start 1% dexamethasone drops every 2 hours for the first week and then four times per day for 6 weeks followed by a slow taper. Topical moxifloxacin drops were used three times per day for a week and a non-steroidal anti-inflammatory drop for 1 month. During the postoperative period, reintroduction of glaucoma medications, needling, digital ocular compression, AC reformation, surgical revisions and reoperations were performed as per surgeon discretion.15 A Seidel test was performed routinely during the first 2 weeks postoperatively. A dilated fundus exam was performed during the first postoperative week and repeated as needed if there were any signs of clinical hypotony on the initial dilated fundus exam (ie, IOP <6 mm Hg with choroidal detachment, hypotony maculopathy, shallow AC or corneal folds).

Needling was performed at the slit-lamp using a 25G trocar as previously described.15 A total of 0.1 mL of MMC (0.5 mg/mL) was injected 20 min prior to the start of the needling. Topical antibiotics were then used for 1 week and 1% dexamethasone drops were used every 2 hours for a week and then slowly tapered.

Surgical revision

If a revision was decided on, the conjunctiva was reopened, and the overlying scar tissue with Tenon’s capsule was excised. The device was then carefully inspected and replaced if any irregularity was found such as intraluminal obstruction or poor flow. Additional MMC was placed on three LASIK merocele sponges of 0.5 mg/mL for a total of 2–3 min. The distal end of the implant was loosely sutured to the sclera. Conjunctival closure was then performed with 9-0 vicryl ensuring a watertight seal.

Outcome measures

The primary outcome was percentage of patients who attained complete success at 1 year, accounting for loss to follow-up. With consideration of the World Glaucoma Association’s Guidelines on Design and Reporting of Glaucoma Surgical Trials as well as key publications,4–8 15–19 failure was defined as any of the following: (1) IOP less than 6 mm Hg and vision loss of more than 2 lines from baseline on two consecutive visits, (2) IOP of more than 17 mm Hg on two consecutive visits, (3) less than 20% reduction from decision IOP on two consecutive visits, (4) glaucoma medication use, (5) surgical revisions or reoperations or (6) no light perception (NLP) vision. IOP or medications outside of the range within the first postoperative month were not considered failure. In-clinic interventions (including needling) were permitted. Surgical revisions, reoperations (cyclophotocoagulation (CPC) or any further glaucoma surgery) or NLP vision were immediate failures regardless of time of occurrence.

Secondary outcomes included percentage of patients who attained qualified success (with glaucoma medication or trabeculoplasty) at 1 year for IOP of 6–17 mm Hg, as well as complete and qualified success at IOP thresholds of 6–14 mm Hg and 6–21 mm Hg. Additional outcomes were median IOP and medication number, BCVA at last follow-up, revisions and reoperations.

HR of failure was assessed for various characteristics potentially associated with failure, on both univariate basis and multivariable basis: superior versus inferior microshunt placement, age (≥70 years of age vs <70 years of age), gender, ethnicity (Caucasian vs non-Caucasian), eye (right vs left), lens status (phakic vs pseudophakic), preoperative vision (0.4 logarithm of the minimum angle of resolution (logMAR) (20/50) or worse vs better than 0.4 logMAR), type of glaucoma (primary open-angle glaucoma (POAG) vs other), disease severity (mild to moderate vs advanced), MMC dose (<0.4 vs ≥0.4 mg/mL), decision IOP (≥21 mm Hg vs <21 mm Hg), previous trabeculoplasty and diabetes.

Postoperative interventions (such as needling, AC reformation or AC paracentesis), complications (including leak or dehiscence, blocked or exposed microshunt, corneal decompensation, prolonged inflammation, hyphaema, iris incarceration, vitreous haemorrhage, choroidal detachments, macular folds, macular oedema or diplopia) and more serious complications (including angle closure, retinal detachment, suprachoroidal haemorrhage, malignant glaucoma, blebitis, endophthalmitis or NLP vision) were also recorded.

Statistical analysis

Snellen VA was converted to logMAR units. Proportions of categorical variables were calculated and compared with Fisher’s exact test. Following assessment for normality, continuous variables were reported using mean (SD) or median (IQR) and compared with two-sided t-tests and Mann-Whitney U tests, as appropriate.

To account for loss to follow-up, survival analysis for complete and qualified treatment success for each range of IOP was calculated and displayed with Kaplan-Meier curves. Cox proportional-hazards models were used to evaluate results based on risk factors for failure, with both univariable (crude) and multivariable (adjusted) HRs reported. The Efron method was utilised to approximate data with equivalent failure intervals. The proportional-hazards assumption was tested with the martingale residual method. Collinearity was investigated by monitoring for significant inflation of the SE. Following model determination, a random statement was included in the model for intrapatient (between eyes) correlation. A sensitivity analysis was performed where second eyes were excluded. For secondary categorical and continuous outcomes, univariate and multivariable general linear models accounting for intrapatient correlation were utilised. Statistical significance was set at p≤0.05 (two-sided) and all analyses were performed using SAS University Edition software (SAS Institute, Cary, North Carolina, USA).


Study population and baseline characteristics

Of 356 who underwent stand-alone SIBS microshunt implantation during the study period, 98 were refractory eyes of which 96 had no prior retinal or corneal surgery. Eleven of these eyes were excluded for having less than 1-month follow-up (out-of-town patients without available follow-up). As such, 85 eyes of 79 patients, consisting of 51 superior and 34 inferior implantations, were included.

Baseline demographics are provided in table 1. Baseline preoperative IOP was 22.0 (IQR 18.0–29.0) and number of medication classes was 4 (IQR 3–4). Median baseline visual field mean deviation was −15.4 (IQR −22.8 to −5.7), with two-thirds (67.1%) of patients meeting visual field criteria for advanced disease. With regard to previous ocular laser and surgery, 49.4% of patients had a single previous trabeculectomy, 4.7% had two previous trabeculectomies, 16.5% had both a prior trabeculectomy and tube shunt surgery (AGV/BGI) and 3.5% had two prior tube shunt surgeries (AGV/BGI). Overall, 24.7% of eyes had two or more prior failed subconjunctival surgeries.

Table 1

Baseline characteristics

Primary and secondary outcomes: surgical success at 1 year

Considering success criteria of IOP 6–17 mm Hg (inclusive) with 20% reduction from decision IOP, without clinical hypotony or deviation from these parameters on two consecutive visits, 61.0% of patients achieved complete success, while 79.7% achieved qualified success at 1 year. One patient with a superiorly implanted microshunt failed due to hypotony, with an overfiltering bleb and hypotony maculopathy requiring bleb revision, compression sutures and cautery. Subsequent to revision at 3.8 months, macular folds and hypotony resolved in this patient with an unmedicated IOP of 8 mm Hg and VA of 20/30 by the first 1-year postoperative visit. A second superiorly implanted patient failed due to consecutive IOPs of 4 mm Hg and reduced vision; however, subsequent to receiving compression sutures at 6.7 months, unmedicated IOP improved to 6 mm Hg with VA of 20/40.

For the 6–14 mm Hg (inclusive) IOP range, complete success was 53.8%, while 71.3% achieved qualified success. Likewise, for the 6–21 mm Hg (inclusive) IOP range, 61.0% achieved complete success and 82.2% achieved qualified success. All criteria include 20% reduction from decision IOP. Kaplan-Meier survival curves for each success criteria are shown in figure 1 and reasons for treatment failure based on the primary outcome are shown in table 2. Assessing only cases who underwent surgery in their first eye as a sensitivity analysis, 58.6% had complete success at 1 year for the 6–17 mm Hg IOP criteria.

Table 2

Reasons for treatment failure for primary outcome

Figure 1

Kaplan-Meier survival curves for complete and qualified success. Complete (without medication) and qualified (with or without medication) success: 20% reduction in IOP and IOP within the following range: (A,D) 6–14 mm Hg, (B,E) 6–17 mm Hg and (C,F) 6–21 mm Hg.

Cox proportional-hazards models: factors associated with failure

To identify risk factors for failure with higher statistical power (ie, highest event rate), complete success in the 6–14 mm Hg IOP range was used for the Cox proportional-hazards regression analyses. Mild-to-moderate disease severity, defined as visual field mean deviation better than −12 dB (HR 1.98, 95% CI 1.11 to 3.53), and right surgical eye (HR 1.90, 95% CI 1.04 to 3.47) were significant factors for failure based on univariable analysis. Mild-to-moderate disease (HR 2.37, 95% CI 1.23 to 4.59) was at higher risk of failure on multivariable analysis. Other demographic and surgical factors were found to have insufficient evidence for association with failure, including age, gender, ethnicity, history of diabetes, preoperative vision, POAG diagnosis, previous cataract surgery or trabeculoplasty, high decision IOP (≥21 mm Hg), superior vs inferior placement of the microshunt and MMC dose <0.4 mg/mL (figure 2).

Figure 2

Graph displaying both the crude (left) and adjusted (right) HRs for various risk factors, with 95% CIs. 1Mild-to-moderate glaucoma as defined by a mean deviation on visual fields better than −12 dB. 2Intraocular pressure ≥21 mm Hg at the time surgery was decided. 3Preoperative best-corrected visual acuity better than or equal to 0.4 logarithm of the minimum angle of resolution (20/50). IOP, intraocular pressure; MMC, mitomycin C; POAG, primary open-angle glaucoma.

Intraocular pressure, medication use, visual acuity and number of visits

Censoring for reoperations, the median IOP at postoperative year 1 was 13.0 mm Hg (IQR 10.0–17.0) with zero (IQR 0–2) medication class (figure 3). Using last observation carried forward analysis for reoperations, median IOP was 13.0 mm Hg (IQR 10.0–17.0) with zero (IQR 0–2) medication. Figure 4 shows a scatter plot of preoperative IOP versus IOP by the first reported 1-year postoperative visit, classified by the number of classes of medications at this timepoint for all patients. As demonstrated in figure 5, 64.5% of implanted patients were medication-free by the first 1-year postoperative visit. Looking at patients who achieved qualified success in the 6–17 mm Hg IOP range with 20% reduction from decision IOP, 88.3% of patients were on less medication, 8.3% on the same number and 3.3% were on more medications by the first year.

Figure 3

Box and whisker plot of preoperative to postoperative IOP. Censored for reoperations (n=6 eyes). The middle lines indicate the median; the lower and upper ends of boxes indicate the lower and upper quartiles; the whisker ends span 1.5 times the IQR. IOP, intraocular pressure; Meds, medications; POD, postoperative day; POM, postoperative month; POW, postoperative week; preop, preoperative.

Figure 4

Scatter plot of IOP at time of surgical decision versus IOP at first reported postoperative month 12 visit. Censored for reoperations (n=6 eyes). Plotted according to number of medication classes at postoperative month 12. IOP, intraocular pressure; Meds, medications; POM, postoperative month.

Figure 5

Glaucoma medication burden at time of surgical decision versus first reported postoperative month 12 visit. Censored for reoperations (n=6 eyes). Meds, medications; POM, postoperative month.

Median BCVA was 0.4 (IQR 0.2–1.0) at 1 year postoperatively. Assessing the difference in BCVA between preoperative and postoperative measures, there was no significant difference between the two timepoints (p=0.34). A 28.2% of the total group lost 2 or more lines of vision. The only factor associated with worse visual recovery was poor preoperative vision, defined as less than 20/50 acuity (HR 1.35, 95% CI 1.13 to 1.61) on univariable and multivariable analyses (HR 1.46, 95% CI 1.18 to 1.80). The specific eye, gender, age, ethnicity, history of diabetes, POAG diagnosis, disease severity, previous history of trabeculoplasty or cataract surgery, high decision IOP, MMC dose and implant location were not significantly associated.


Table 3 outlines early and late complications. A 12.9% of implanted patients had early choroidal effusions or detachments (within 3 months postoperatively). No patients experienced late choroidal detachments. Of the 11 patients who had choroidals, 5 patients had resolution with conservative management including topical steroid reduction and atropine 1.0%, and 1 patient had resolution with observation alone. One patient received a sub-Tenon’s triamcinolone acetonide injection to help raise IOP (0.2 mL of 40 mg/mL), two patients had AC reformation with air fill for shallow AC, one patient had AC reformation with ophthalmic viscosurgical device (OVD) along with oral steroids and one patient had choroidal drainage. All cases of choroidal detachments resolved within the first 2 months postoperatively.

Table 3

Patients with postoperative complications

Hypotony maculopathy developed in five patients. Two cases were mild and resolved without management, one patient had resolution following AC reformation with OVD and one patient had resolution subsequent to bleb revision with compression sutures. One case experienced loss of central vision (from 20/25 to hand motion vision) and ciliary body effusion in the early postoperative period with persistent hypotony maculopathy; a bleb revision was done at 9 months for overfiltration and VA improved to 20/200 by 1 year. Although the patient continued to have chronic macular folds, IOP improved from <6 mm Hg to 7–9 mm Hg postrevision and the patient declined any further intervention.

One patient developed new-onset corneal oedema. One superiorly implanted patient with a history of trabeculectomy, BGI, phacoemulsification and tube revision had borderline corneal function prior to microshunt surgery. One year postoperatively, the patient developed corneal oedema and decompensation requiring a subsequent Descemet’s stripping automated endothelial keratoplasty.

One patient in our cohort developed an exposed microshunt inferiorly at postoperative month 6 which required a revision. The patient had a history of cataract surgery, previous trabeculectomy and tube shunt surgery.

None of the patients experienced NLP vision, retinal detachment or malignant glaucoma. There was one case of endophthalmitis at postoperative week 3 in an inferiorly implanted patient; however, the endophthalmitis seemed to originate from a previous thin avascular trabeculectomy bleb. The superior trabeculectomy bleb was found to be Seidel positive with an infiltrate in the bleb, likely aggravated by the surgical event and/or postoperative medications. The microshunt was quiet, with no signs of blebitis.

Postoperative interventions

The most common postoperative intervention was needling, with 5.9% undergoing the intervention prior to 3 months and 5.9% receiving late needling more than 3 months postoperatively (table 4). The second most common intervention was AC reformation, consisting of approximately equal proportions receiving air or OVD injections within the first 3 postoperative months in the majority of cases.

Table 4

Patients with postoperative interventions

Median time to needling was 3.7 (IQR 1.1–9.5) months, with an IOP of 21.0 mm Hg (IQR 18.0–26.0) and zero (IQR 0–1) medication. Female gender was statistically predictive for needling (HR 4.35, 95% CI 1.13 to 16.85) as well as intraoperative MMC dose of less than 0.4 mg/mL (HR 4.39, 95% CI 1.27 to 15.15), on a univariable basis. With multivariable analysis, however, only intraoperative MMC dose of less than 0.4 mg/mL approached significance (HR 8.06, 95% CI 0.98 to 66.67). Of the 10 patients who underwent needling, none received second needling or reoperations and four were medication-free at last follow-up. Fifty percent achieved qualified success and 40% achieved complete success by 1 year for the success criteria of 6–17 mm Hg with 20% IOP reduction. Among these patients, 1-year postoperative median IOP was 18.5 mm Hg (IQR 14.0–21.0) with two (IQR 0–4) medications.

Revisions and reoperations

Ab externo bleb revision was performed in four patients after postoperative month 3 (table 5). Two patients received bleb revision for elevated IOP at postoperative months 7.4 and 9.5, with the first patient subsequently receiving CPC at postoperative month 10.3. These two eyes had pressures of 15 mm Hg on three medications and one medication at last follow-up, respectively. One patient received CPC at 8 months postoperatively for elevated IOP. Goniosynechialysis was performed in one patient due to chronic history of PAS since having a trabeculectomy prior to microshunt implantation.

Table 5

Patients with postoperative revisions and reoperations


This interventional, retrospective, single-centre series reviewed surgical success rates, risk factors for failure and adverse events of implantation of an ab externo SIBS microshunt in refractory glaucoma, over a 1-year period. This subset of refractory eyes to at least one failed subconjunctival glaucoma surgery are challenging eyes to achieve success and have increased complication risks.2–6 Previous glaucoma surgery is a known risk factor for failure for further filtering procedures.3 5 20 21 Overall, there are few studies examining success rates of subconjunctival surgery, even less so with novel subconjunctival implants, in eyes with a previous failed filtering procedure as they are intrinsically at higher risk of failure and complications. Our main outcome, complete surgical success, was achieved in 61.0% of eyes, and qualified success was achieved in 79.7% of eyes.

The success rates seen in this study appear to be at or better than reported for tube shunt or trabeculectomy surgery results in these high-risk refractory eyes. The microshunt appears to have qualities both of trabeculectomy in terms of efficacy and of tube shunts with hypotony prevention and safety.

Refractory glaucomas are often managed with tube shunts as supported by the results of the TVT trial.4 5 In that study, eyes with previous cataract surgery and/or trabeculectomy were considered refractory and showed complete (IOP threshold 6–21 mmHg and 20% reduction of IOP with no medications) success rate of 34% in the tube group and 63% in the trabeculectomy group at 1 year (p<0.001); for qualified success (same IOP criteria with or without medication), success rates increased to 96% in the tube group and 87% in the trabeculectomy group.4 The TVT study had less stringent success IOP criteria (21 mmHg) versus this study (17 mmHg) and broader definition of refractory glaucoma, including a significant number of patients with previous cataract surgery. Nevertheless, the results were also stratified, with 0% of those having had previous trabeculectomy with MMC achieving complete success (46% qualified success) in the tube group and 29% achieving complete success in the trabeculectomy group (29% qualified success).

In the Ahmed versus Baerveldt study, comparing the AGV (New World Medical) with a Baerveldt-350 implant (Abbott Medical Optics, Santa Ana, California, USA), 1-year outcomes for refractory glaucoma eyes defined as having previous medicinal, laser or surgical therapy (37% had previous incisional glaucoma surgery) or atypical glaucoma showed complete success (IOP threshold 5–18 mm Hg and 20% reduction of IOP with no medication) of 12% in the total cohort and qualified success (with or without medication) in 53%.6 The 5-year pooled analysis of the AVB study and Ahmed Baerveldt Comparison study of tube shunts yielded success rates of 51% and 63% for the Ahmed and Baerveldt groups, respectively, given success criteria of 6–18 mm Hg with 20% IOP reduction.17

The ab interno gel microstent has similar theoretical advantages to the ab externo SIBS microshunt and has also been used in refractory glaucoma.22 In a prospective, multicentre study, refractory glaucoma was defined as patients with a prior glaucoma procedure (only 63% had previous incisional glaucoma surgery) and/or uncontrolled IOP on maximally tolerated medical therapy. Of the 65 patients implanted, success rate was 76.3% (based on the imputation model) as defined by an IOP decrease of 20% from baseline on the same or fewer medications. Although the success criteria used in the Grover study are different from the one commonly used in the literature, the success rates are not as high as the ones obtained in this paper. Despite discrepancies in how refractory glaucoma is defined, as well as significant differences in baseline characteristics from study to study, the comparison with the current literature and different filtering procedures suggests that this study’s results achieve similar or better results compared with other standard-of-care options.

As with any filtering surgery and even more so in complex refractory eyes, many factors may influence the success or failure of the procedure. In our univariable models, right eye and mild-to-moderate disease were associated with higher failure rates, with mild-to-moderate disease maintaining a significant association on multivariable analysis. These are somewhat unexpected outcomes. The mild-to-moderate patient population would intuitively seem to be at lower risk of failure than the advanced disease patients. Yet, there are no studies to support that mild-to-moderate disease have worse outcomes compared with advanced disease eyes.

A superior approach was our preferred implant placement due to well-known concerns of infection risk, as previously described for trabeculectomy.12–14 However, in patients with poor superior tissue due to previous surgeries and antimetabolite exposure, the superior quadrants are not useable. Thus, in these eyes, inferior placement is the only viable location for subsequent subconjunctival surgery. Inferior bleb placement is not preferred and has had a history of increased infection risk in trabeculectomy blebs. When needed, tube shunts have been used inferiorly in these cases. Our experience with bleb morphology with the microshunt has been typical low-lying diffuse posterior blebs with rare avascular or cystic components which we feel lower the risk for infection. Due to this impression, and the limited options for these patients who need a low IOP, we have cautiously used the microshunt in the inferior quadrant with success and will continue to carefully watch for the risk of infection in particular.

Patients receiving 0.2 mg/mL of MMC had higher needling rates than patients with higher concentrations of MMC. A lower dose of MMC at the time of surgery was not found to increase failure rates in our study; however, there is some evidence for this in multivariable modelling in a previous study by our group in patients with stand-alone surgery.15

Overall, few complications were found in our study considering these were high-risk eyes. Although the set lumen and length of the device theoretically protect against hypotony assuming normal aqueous production, these diseased eyes with previous filtering surgery in place are likely more susceptible to hypotony. The AVB study had rates of choroidal detachments in 13% of eyes at 1 year.6 In the TVT study, persistent hypotony was found in 23% of the trabeculectomy group and none of the tube group at 1 year, which increased to 31% and 13%, respectively, at 5 years.4 5 In our patients, the majority of those who developed signs of clinical hypotony were transient and improved over time with conservative treatment or minor interventions. Choroidal drainage was required in one patient, bleb revision for overfiltration in two others and no intervention for a case of mild macular folds with good vision. AC reformation was occasionally needed, and when performed, the choice of agent was left to the physician. In general, air was used as a first option and then viscoelastic as a second line.

Postoperative interventions consisted mainly of needling (11.8%) and AC reformation (8.2%) with air or viscoelastic. The needling rates were similar compared with our previous report of eyes without previous subconjunctival surgery (8.5%).15 In the ab interno gel microstent refractory study, needling was performed in 32% of eyes, with 11% of eyes requiring more than one needling,22 and in non-refractory studies, the needling rates have approached 40%.7 23 Needling rates vary widely in the literature depending on the procedure, patient factors and surgeon preference. In our study, needling was the preferred next step in the escalation of therapy prior to adding topical medications, unless the bleb was significantly hyperencapsulated or avascular. In certain cases, when needling procedure had failed or the bleb was not amenable to needling, a surgical revision of the microshunt was performed. Given our overall success with surgical revision, this has become our emerging preferred approach to treating bleb fibrosis or device obstruction and re-establishing flow rather than proceeding with another filtering surgery altogether or cycloablative therapy. It allows for the best visualisation of the cause of failure and debulking of the fibrosis over the implant. In some cases, if the conjunctiva is significantly scarred, a tube shunt can be performed in its place.

Reoperation rates were low (7.1%)—the main procedures undertaken were bleb revision and CPC. In the TVT study, reoperations with an additional glaucoma surgery were performed in one patient in the tube group (1%) and in five patients in the trabeculectomy group (5%) at 1 year.4 As discussed previously, our preferred approach was revision of the microshunt, but in cases where the conjunctiva was not amenable to further surgery or patient preference, CPC was performed.

Some studies have suggested an increased risk of infection rates with multiple filtering surgeries.24 In our sample, one case of endophthalmitis was seen; however, this was not from the microshunt bleb. Another consideration for this implant and others is the risk of long-term erosion. Only one patient developed corneal oedema postoperatively (in an eye with an already compromised endothelium). As with any implant in the AC, corneal endothelial cell loss will be something to watch for. Long-term data will help us better understand the infection, erosion and corneal endothelial cell loss rates.

Some limitations of this paper are inherent to the study design and patient population. These results represent our learning curve and evolution in technique and postoperative management (most notably our increase in the MMC dose and utilisation of the inferior conjunctiva). Although the surgeries were performed by one surgeon and fellows in a multicultural setting, there may be limitations with the generalisability of results. Additionally, as is the case for many glaucoma studies evaluating surgical outcomes, IOP is only one piece of the puzzle in assessing for treatment success. Reduction of structural and functional disease progression is the main goal of any glaucoma procedure. IOP has been the best surrogate measure to quantify surgical success, as we know that IOP control is key in slowing progression. Future randomised studies comparing different surgeries are essential in better tailoring a particular procedure to each patient’s needs.

Our 1-year results show that the ab externo SIBS microshunt in refractory eyes to previous failed subconjunctival surgery demonstrated reasonable rates of qualified and complete success, decreased drop use, with relatively few complications or further interventions. Long-term surveillance will be particularly important in inferior implants to assess for erosion, infection and endothelial cell loss rates, although these complications were uncommon at 1 year in our series. These encouraging results make for a promising alternative procedure in complex eyes with failed filtering surgeries.



  • Contributors Conceptualisation: GMD, MBS and IKA; data curation: GMD, MBS and SS; formal analysis, visualisation, writing—original draft and writing—review and editing: GMD, MBS, SS and IKA.

  • Funding This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

  • Competing interests GMD: honoraria from, and consultant to, Alcon (Fort Worth, Texas, USA); honoraria from, and consultant to, Allergan (Irvine, California, USA), honoraria from, and consultant to Bausch and Lomb (Rochester, New York, USA), honoraria from, and consultant to, Glaukos (San Clemente, California, USA); honoraria from Novartis (Basel, Switzerland); honoraria from Thea-Labtician (Oakville, Ontario, Canada); honoraria from MicroSurgical Technology (Redmond, Washington, USA); honoraria from, and consultant to Santen Pharmaceutical Co, (Kita-ku, Osaka, Japan); honoraria from Sight Sciences (Menlo Park, California, USA). MBS: Honoraria from, and consultant to, Alcon (Fort Worth, Texas, USA); honoraria from, and consultant to, Allergan (Irvine, California, USA); consultant to Light Matter Interaction (Toronto, Ontario, Canada); consultant to Santen Pharmaceutical Co (Kita-ku, Osaka, Japan); honoraria from Aequus (Vancouver, British Columbia, Canada); honoraria from Johnson & Johnson (Jacksonville, Florida, USA); honoraria from Thea-Labtician (Oakville, Ontario, Canada). SS: No financial disclosures. IKA: speaker’s honoraria and research support from, and consultant to, Abbott Medical Optics (Abbott Park, Illinois, USA); consultant to Acucela (Seattle, Washington, USA); research support from, and consultant to, Aerie Pharmaceuticals (Durham, North Carolina, USA); speaker’s honoraria and research grant support from, and consultant to, Alcon (Fort Worth, Texas, USA); speaker’s honoraria and research support from, and consultant to, Allergan (Irvine, California, USA); consultant to ArcScan (Golden, Colorado, USA); consultant to Bausch and Lomb (Rochester, New York, USA); speaker’s honoraria and research support from, and consultant to, Carl Zeiss Meditec (Jena, Germany); consultant to Centervue (Padova, Italy); consultant to Clarity Medical Systems (Pleasanton, California, USA); consultant to ElutiMed (New Orleans, Louisiana, USA); consultant to Envisia Therapeutics (Morrisville, North Carolina, USA); consultant to Equinox (Newport Beach, California, USA); consultant to Eyelight (Funo, Italy); consultant to ForSight Labs (Menlo Park, California, USA); research support from, and consultant to, Glaukos (San Clemente, California, USA); consultant to Gore (Newark, Delaware, USA); consultant to InjectSense (Emeryville, California, USA); consultant to Iridex (Mountain View, California, USA); consultant to iStar (Wavre, Belgium); research support from, and consultant to, Ivantis (Irvine, California, USA); consultant to KeLoTec (Orange County, California, USA); consultant to LayerBio (Medford, Massachusetts, USA); consultant to Leica Microsystems (Wetzlar, Germany); research support from, and consultant to, New World Medical (Rancho Cucamonga, California, USA); consultant to Omega Ophthalmics (Lexington, Kentucky, USA); consultant to Ono Pharma (Chuo-ku, Osaka, Japan) consultant to PolyActiva (Melbourne, VIC, Australia); consultant to Sanoculis (Kiryat Ono, Israel); consultant to Santen Pharmaceutical Co (Kita-ku, Osaka, Japan); consultant to Science Based Health (Spring, Texas, USA); consultant to SOLX (Waltham, Massachusetts, USA); consultant to Stroma (Irvine, California, USA); consultant to TrueVision (Santa Barbara, California, USA).

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

  • Data availability statement Data are available upon reasonable request.

Linked Articles

  • At a glance
    Frank Larkin