In a series of 194 eyes of 160 consecutive patients who had primary phakic DS, the probability of achieving an IOP of less than 19 mmHg without medications or needle revisions was 85 % at 1 year and 78 % at 3 years (Anand et al. 2011). IOP of less than 13 mmHg was achieved without medications or needle revision in 68 % at 1 year and 60 % at 3 years.

In a randomized controlled trial, DS with mitomycin C yielded complete success (defined as IOP ≤21 mmHg without antiglaucoma medications) in 79 % (15/19) of eyes at 1 year and 53 % (10/19) of eyes at 4 years (Cillino et al. 2005, 2008). Qualified success (defined as IOP ≤21 mmHg with or without antiglaucoma medications) was seen in 100 % of eyes at 1 year and 79 % (15/19) of eyes at 4 years.

In another randomized controlled trial, 43 patients were allocated to DS with reticulated hyaluronic acid (SK-GEL) scleral implant and mitomycin C (Russo et al. 2008). No goniopuncture or bleb needling was performed. At 4 years, 51 % of eyes had complete success (defined as achievement of target IOP without antiglaucoma medications) for a target IOP of <21 mmHg, while 33 % of eyes had complete success for a target IOP of <18 mmHg.

Reviews of the literature differ in their conclusions about whether DS offers an equivalent degree of IOP control to trabeculectomy (Eldaly et al. 2014; Rulli et al. 2013).

There is consensus that, compared to DS, trabeculectomy has a higher risk of complications such as hypotony (relative risk (RR) 2.1), choroidal effusion (RR 3.8), cataract (RR 3.3), and shallow anterior chamber (RR 4.1) (Rulli et al. 2013). Blebitis (1 %) and endophthalmitis (0.5 %) have been observed during long-term follow-up of patients who have had DS (Anand et al. 2011).

In a randomized controlled trial, 35 % of the 25 patients allocated to VC had total success (IOP 6–21 mmHg without medication) at 3 years (Yalvac et al. 2004). Qualified success (IOP 6–21 mmHg with medication) was achieved in 74 % of the patients at 3 years.

Kobayashi et al. (2003) conducted a randomized controlled trial using a paired design in which one eye of each participant received VC whereas the other had trabeculectomy. At 12 months, 64 % of the eyes treated with VC, achieved IOP ≤ 20 mmHg without medication versus 88 % of the eyes treated with trabeculectomy.

Recent systematic reviews and meta-analyses of the available evidence suggest that VC is less effective at controlling IOP than trabeculectomy (Eldaly et al. 2014; Rulli et al. 2013). However, some authors have commented that VC requires a learning curve that may be relevant to outcomes (Eldaly et al. 2014; Mendrinos et al. 2008).

Compared to VC, trabeculectomy has a higher risk of complications such as hypotony (relative risk (RR) 2.6), choroidal effusion (RR 6.0), cataract (RR 3.8), and shallow anterior chamber (RR 5.5) (Rulli et al. 2013).

In a prospective series of 94 patients (Lewis et al. 2011), a suture was successfully placed in 74 (79 %) of them. Mean IOP was reduced from 24.7 mmHg at baseline to 15.3 mmHg at 1 year. This result needs to be interpreted with caution because 1-year outcomes were reported for fewer than two-thirds of the patients.

No randomized controlled trials have been conducted to compare canaloplasty with trabeculectomy. In a retrospective, nonrandomized comparative case series, canaloplasty achieved a 32 % reduction in IOP at 12 months compared to 43 % for the trabeculectomy group (Ayyala et al. 2011).

In the series of Lewis et al. (2011), adverse events were reported in 16 %, with hyphema (3 %) and elevated IOP (3 %) being the commonest.

Precise scleral flap dissection is crucial. The outer or superficial scleral flap should be half to third thickness. If too thin, aqueous will transude through the flap. Also a thin scleral flap is prone to necrosis due to a poor vascular supply. This is particularly relevant if MMC or bevacizumab are used (Fig. 3.5).

The inner or deep sclera flap dissection is crucial. If too shallow, the dissection will pass over the SC with little or no filtration. If the dissection is quite shallow with a “white” scleral bed, a third deep flap may be dissected at the correct depth to deroof the SC. If the dissection is slightly shallow and uneven, tissues forming the SC roof can be grasped and avulsed as shown in Fig. 3.6.

Perforation of the TDM is the most common intraoperative complication. The incidence of perforation is around 30 % for the novice and decreases to 2–3 % in the more experienced surgeon (Karlen et al. 1999; Sanchez et al. 1997). Perforation of the TDM can be classified by size into microperforations or large transverse perforation and by location into anterior or posterior.

TDM microperforations often occur while extending the lateral edges of the flap into the cornea. To avoid perforations, the eye should be made soft by releasing aqueous via a paracentesis. This should be done just before the SC is deroofed. If these microperforations are anterior with minimal shallowing of the anterior chamber, the procedure should be continued as normal. Sometimes the microperforation can be left covered by the corneal–scleral stump. The dissection of the TDM should be meticulous and the surgeon should avoid the temptation to complete the operation quickly after an anterior perforation. The outer sclera flap may be suture tightly and the anterior chamber reformed at the end of the procedure. The use of viscoelastic to maintain the anterior chamber is to be discouraged as residual viscoelastic may cause a postoperative IOP rise. A large transverse tear at the junction between the trabeculum and Descemet’s membrane (corresponds to Schwalbe’s line) may occur spontaneously or on minimal applied pressure with a PVA spear (Fig. 3.7). The incidence has not been reported but may occur in 1–2 % of cases. In both a posterior tear and a transverse tear, the iris will relapse and an iridectomy should be performed (Fig. 3.8). The surgeon can excise the deep flap and perform a punch sclerectomy under the superficial flap, converting the procedure to a trabeculectomy. If the deep flap has already been is higher and tight sutures and good closure is imperative. It is advisable to dissect a half-thickness outer sclera flap during the surgeon’s learning curve to ensure watertight closure in case of perforation. Often the transverse linear perforation is not visible on postoperative gonioscopy (Fig. 3.8).

Intraocular bleeding is rare in NPGS because the IOP is reduced slowly and a peripheral iridotomy is not performed. Blood reflux from Schlemm’s canal may occur when the episcleral venous pressure rises above the IOP. This may be a favorable sign, indicating an intact outflow pathway. It may fill the subscleral lake and may perhaps stimulate subscleral fibrosis postoperatively. Some experts believe that early laser goniopuncture may be indicated in this situation.

Descemet’s membrane detachment (DMD) can occur in all three types of NPGS. The detachment tends to extend from the surgical site in DS and VCT, but tends to occur 180° from the polypropylene suture knot in canaloplasty.

In VCT, the detachment is usually noted at the time of forceful injection of viscoelastic. The management involves removal of the viscoelastic and injection of anterior chamber air bubble tamponade. Rarely, blood can be trapped within the detachment. This blood and can be left if it is peripheral, although it may take months to clear. There is one reported case of blood remnants still visible 2 years after surgery (O’Brart et al. 2004). Hemorrhagic Descemet’s detachment involving the visual axis can lead to long-term decrease in visual acuity (Yalvac et al. 2004). Air tamponade is helpful in this situation but occasionally inert gas (such as perfluropropane (C3F8) or sulfur hexafluoride (SF6) gas), or fixation sutures (descemetopexy) may be needed.

DMD after canaloplasty is not uncommon. In a large case series, DMD was observed 12 out of 124 eyes (7.2 %). Eighty-three percent (10/12) of the DMDs involved the inferior quadrants and measured <3 mm. Hemorrhage within the DMD was seen in 58 %. Two patients had large detachments measuring 5–6 mm extending into the visual axis. DMD resolved completely with or without drainage except for 1 patient who developed corneal decompensation, needing penetrating keratoplasty (Jaramillo et al. 2014).

DMD occurs less frequently after DS. A database search of 1250 DS and combined phacoemulsification and DS procedures done in our institution over 14 years, we identified four cases (0.003 %). Three were observed 2–4 months after DS and one more than 2 years after surgery (Fig. 3.9). All four eyes had IOP over 20 mmHg. Interestingly, the DMD resolved after laser goniopuncture in one case and after needle revision in three cases (Anand N, unpublished data). The implication is that the etiology of DMD after DS differs from that of VCT. DMD after DS may become manifest if the outflow resistance is high and the aqueous passing through the TDM then accumulates between the corneal stroma and Descemet’s membrane. In a case series of nine patients with DMD, four after VCT and five after DS, the authors have emphasized this difference. DMD after VCT is observed immediately after surgery and weeks to month after DS. They performed descemetopexy in four eyes (Ravinet et al. 2002).

Intraoperative adverse events specific to canaloplasty include the inability to cannulate the Schlemm’s canal, trauma to the canal and microcatheter passage into the suprachoroidal space (Grieshaber et al. 2011). Difficulties in cannulating Schlemm’s canal is usually related to the dissection, identification and de-roofing of the canal. Resistance or blockage of the microcatheter may occur due to a tight opening, hitting an open collector channel, an incomplete canal or scarring in the canal. Injection of viscoelastic into the canal can aid penetration, by dilation and lubrication. The surgeon can also try passing the microcatheter in the opposite direction. Excess force may cause a tear in the trabecular meshwork and cause microcatheter penetration into the anterior chamber. The polypropylene suture is passed through the canal and tightened to maintain its patency. If the suture breaks during knot-tying, then it will need to be replaced. In case of unsuccessful circumferential SC catheterization, the procedure may be converted into 180° metal trabeculotomy. If the tension suture cheese wires through the trabecular meshwork after successful complete catheterization, it is converted into 360° trabeculotomy (Alnahrawy et al. 2015).