Three-Dimensional Optical Coherence Tomography

The adoption of optical coherence tomography (OCT) has enabled highly-accurate quantitative assessment of the optic nerve head and surrounding retinal structures to assist in the diagnosis and monitoring of glaucoma. Traditional parameters include assessment of optic disc area, rim area, cup-to-disc ratio, and two-dimensional measurement of the thickness of the peripapillary retinal nerve fiber layer (RNFL) and ganglion cell-inner plexiform layer (GC-IPL). These parameters have been shown to have good reproducibility and the ability to distinguish between glaucomatous and non-glaucomatous eyes [4]. The detection of RNFL damage is helpful in the early detection of glaucoma, frequently preceding the development of visual field loss [5]. However, two-dimensional parameters are susceptible to artifacts that may adversely affect their diagnostic ability [6, 7]. This is especially problematic where there are variations in optic disc size, optic nerve head tilt, peripapillary atrophy, and myopia [8]. These imaging artifacts or inaccuracies may cause erroneous RNFL measurements that may lead to inaccurate assessments [7]. The introduction of three-dimensional volume scans enable high-density sampling of nerve tissue and 3D reconstruction of the neuroretinal rim anatomy which may assist in the diagnosis and monitoring of glaucoma [9].

Three-dimensional scanning of RNFL and ganglion cell layer (GCL) volumes permits the assessment of new parameters such as the minimum distance band (MDB) [10]. The MDB is the shortest distance between the internal limiting membrane (ILM) and the optic disc margin, defined as the termination of the retinal pigment epithelium (RPE)/Bruch’s membrane (BM) [11]. The MDB has several advantages over standard neuroretinal rim parameters. The RPE/BM termination is an objective, consistent, and easily-identifiable anatomic landmark on OCT compared to traditional parameters that define the optic disc margin based on the clinical optic disc margin [9]. Additionally MDB measurements are perpendicular to the course of retinal ganglion cell (RGC) axons, therefore they take into account the variable orientation of RGC axons as they approach the optic nerve head [9]. MDB has been validated as a marker for glaucoma and has been shown to have good diagnostic performance compared to two-dimensional RNFL measurements [9]. Shieh et al. showed that 3D MDB neuroretinal rim thickness measurements had uniformly equal or better diagnostic performance for glaucoma in all quadrants and was significantly better in the nasal region compared to 2D RNFL thickness measurements [9]. Similarly, Tsikata et al. found that 3D MDB had a higher diagnostic capability for glaucoma than RNFL thickness in the inferonasal, superonasal, and nasal sectors as assessed by the area under the receiver operating characteristic (AUROC) curves [12].

Wide-Field Swept-Source Optical Coherence Tomography

Traditionally, multiple scans are required to capture an OCT image. However, a recently introduced technology called swept-source OCT (SS-OCT) uses a swept laser to capture wide-angle, high-quality images of the optic nerve and macula in a single scan [13]. This technology may provide better image quality and be less affected by media opacities [14]. Wide-field scanning has been shown to be effective at discriminating between healthy and glaucomatous eyes with a diagnostic accuracy comparable to spectral-domain OCT (SD-OCT) [15–17]. For the detection of early glaucoma, SS-OCT has demonstrated superior diagnostic ability over conventional criteria analyzing peripapillary RNFL and ganglion cell layers [18]. Like 3D OCT, SS-OCT may also be less susceptible to artifacts and centering errors [15]. In addition, a tunable wavelength of operation enables imaging of deep ocular structures such as the lamina cribrosa [19]. The lamina cribrosa has long been presumed to be the primary site of axonal injury in glaucoma [20]. It is believed that posterior bowing of the lamina cribrosa may cause mechanical and/or an ischaemic insult to RGC axons [21]. Using SS-OCT, Kim et al. showed greater posterior displacement of the lamina cribrosa in eyes with primary open angle glaucoma compared to age-matched healthy eyes [19]. SS-OCT may therefore assist in our understanding of the mechanisms involved in glaucoma pathogenesis.

Optical Coherence Tomography Angiography

Optical coherence tomography angiography (OCT-A) is a new technology that allows non-invasive visualization of the microcirculation of the eye without the use of contrast dye. The technique takes advantage of improvements in OCT image resolution and scanning speed and is gaining popularity in the assessment of retinal vascular diseases. By comparing sequential scans at the same location, OCT-A is able to detect change which are attributed to erythrocyte movement in perfused vessels [22]. The technique offers several advantages over traditional angiography including the ability to simultaneously assess the retinal and choroidal circulations, quantitative assessment of the microcirculation, and three-dimensional assessment of both microvasculature structure and function while avoiding the need for invasive dye injections. Because of the possible role of reduced optic nerve head perfusion and vascular dysregulation in glaucoma [23], OCT-A is being investigated as a tool to help elucidate the pathophysiology of the disease as well as assist clinicians in glaucoma detection and monitoring [24].

Optical coherence tomography angiography provides quantitative information on both blood vessel structure, reported as blood vessel density and foveal avascular zone area, as well as microvasculature function using flow index, a dimensionless parameter between 0 and 1. The measurements have been shown to have high within-visit repeatability and between-visit reproducibility [25, 26]. Differences between healthy and glaucomatous eyes have been observed with respect to vessel density, foveal avascular area, and reduced blood flow index. In patients with ocular hypertension, normal tension glaucoma, and primary open angle glaucoma reductions in vessel density and size of the foveal avascular zone have been reported [27–30]. These changes are associated with a reduction in optic disc flow and correlate with the degree of visual field defect [25]. Jia et al. found a 25% reduction in optic disc flow between healthy and glaucomatous eyes and this reduction correlated strongly with visual field pattern standard deviation (PSD) [25]. Similarly, Liu et al. and Yarmohammadi et al. found significant correlations between flow index and visual field PSD in glaucomatous eyes [26, 31]. The data suggest an association not only with the degree of visual field defect but also the location of the field defect [32]. In a separate study, Yarmohammadi et al. examined vessel density in eyes with visual field defects in a single hemifield and found that vessel density was lowest in the affected hemiretina [32]. Interestingly, reduced vessel density was also noted in the perimetrically intact hemiretina suggesting that microvasculature changes may precede visual field loss [32]. Changes in OCT-A parameters also correlate with the location of RNFL thinning. Mansoori et al. found a reduction in capillary density in eyes with early glaucoma and that capillary density was lowest in areas with focal RNFL defects [33]. This finding is consistent with other studies showing capillary dropout in areas of RNFL thinning [25, 29, 34].

Several studies have investigated the diagnostic performance of OCT-A for glaucoma detection. Using optic disc flow and a cut-off value of 0.1515, Jia et al. reported a sensitivity and specificity of 100% in their study population [25]. The same group then evaluated the ability of peripapillary flow index and vessel density to discriminate between healthy and glaucomatous eyes using AUROC curves and found values of 0.982 and 0.938 respectively [26]. The performance of OCT-A appears to depend on the stage of glaucoma [35]. Wang et al. investigated the correlation between OCT-A parameters and glaucoma severity and found that both vessel density and flow index performed best in advanced glaucoma [35].

There is some emerging data on the effect of IOP reduction on OCT-A parameters [36–38]. In patients with very high IOP who achieved a >50% reduction in IOP with medical therapy there was a significant increase in OCT-A parameters [37]. However, another study found no statistically significant difference in OCT-A parameters in the peripapillary or macular regions following glaucoma filtration surgery, despite an average IOP reduction of 44.2% [36]. Lastly, in patients presenting with acute angle closure, statistically significant changes in OCT-A parameters have been observed following treatment and normalization of IOP [38].

Optical coherence angiography is not without limitations. Currently, there is a lack of comparability between machines and studies due to an absence of standardized measurement protocols. Also, image quality is highly dependent on fixation and patient co-operation. Further longitudinal studies are needed to determine whether OCT-A findings can predict or detect glaucoma progression. Nonetheless, OCT-A remains a promising technology for elucidating the physiology of glaucoma and evaluating structure and function in this disease.

Adaptive Optics

Adaptive optics (AO) is not an imaging modality, but rather a technology used in combination with existing imaging modalities to improve their performance [39]. Initially developed to reduce ocular aberrations from ground-based telescopes, it has been used in conjunction with fundus cameras, scanning laser ophthalmoscopes, and most recently OCT to provide unprecedented resolution and the ability to visualize structures at the cellular level in real time. Representing a major advance in optical technology, AO uses a wavefront sensor that measures aberrations in ocular optics and a deformable mirror or spatial light modulator to compensate for these aberrations in vivo [40].

Because RNFL loss is one of the earliest detectable changes in glaucoma, often preceding changes at the optic nerve head or visual field loss [41], there has been particular interest in using AO to detect RNFL changes allowing for earlier detection, more precise diagnosis, and improved detection of progression in glaucoma [39]. Several groups have used AO to visualize RNFL bundles and the gaps between them [42–45]. Kocaoglu et al. proved that it was possible to measure the dimensions of RNFL bundles in five health subjects [42]. This work was extended by Takayama et al. who demonstrated reduced RNFL bundle dimensions in glaucomatous eyes and that these abnormalities were associated with visual field defects [43]. Showing the promise for early detection, Chen et al. demonstrated changes in RNFL bundles on AO that were difficult, if not impossible, to discern with current OCT technology [44]. Most recently, Hood et al. followed six eyes of five patients with deep glaucomatous visual field defects using adaptive optics scanning light ophthalmoscopy (AO-SLO) and showed progressive changes in RNFL bundles, demonstrating the potential for AO to be used for monitoring glaucoma progression [46].

To date, it has been difficult to visualize individual RGCs with OA . This is because RGCs are nearly transparent, an important attribute to allow light to pass through them and reach photoreceptor cells. In spite of this property, one group has been able to image the individual somas of neurons within RGCs using confocal AO-SLO and showed progressive changes in RNFL bundles, demonstrating the potential for A in both monkeys and humans [47]. This capability to noninvasively image RGC layer neurons in the living eye without fluorescent labels may one day allow for insights into the pathogenesis of glaucoma and a better diagnostic tool [47].

Testing Strategies and Novel Thresholding Algorithms

Currently, 24-2 visual fields are the most commonly used method for investigating visual field defects in glaucoma. However, there is an increasing appreciation that damage at the macula can be detected in even early stages of glaucoma [48]. The macula has the highest density of RGCs [49] and thinning of the ganglion cell complex is seen early in the glaucomatous process [50]. In a recent study of patients with early glaucoma, 16 of 26 eyes (61.5%) classified as normal on 24-2 tests were classified as abnormal on 10-2 visual fields [51]. In patients with ocular hypertension, 28 of 79 eyes (35.4%) classified as normal on 24-2 tests were classified as abnormal on 10-2 visual fields [51]. It is therefore apparent that central visual field damage on the 10-2 test may be missed with the 24-2 strategy alone [51]. These findings suggest that in the future it may be necessary to include 10-2 visual field testing to reliably detect central visual field defects. However, further work is required before this becomes the new standard of care.

In addition, novel thresholding strategies are being investigated that incorporate spatial and structural information to improve the speed and precision of visual field testing. Chong et al. introduced a perimetric algorithm that uses spatial information regarding the location of a field defect to improve the characterization of field loss without increasing testing times [52]. Using a computer simulation, Chong et al. reported improved accuracy and precision in testing regions surrounding scotoma edges [52]. The same group then validated the performance of the new algorithm, called Gradient-Oriented Automated Natural Neighbor Approach; (GOANNA) in humans and found results in agreement with earlier simulation studies [53]. Using an alternative approach, Rubinstein et al. introduced a perimetric algorithm (Spatially Weighted Likelihoods in Zippy Estimation by Sequential Testing; SWeLZ) that uses spatial information on every presentation to alter visual field estimates, to reduce test times without affecting output precision or accuracy [54]. Both of these strategies have the potential for significant time savings in clinical settings but require validation in larger scale clinical trials.

Another approach to improve thresholding procedures is to incorporate structural information into the testing process [55]. An example of this approach is demonstrated by Ganeshrao et al. who developed a perimetric test strategy called Structure Estimation of Minimum Uncertainty (SEMU), that uses structural information to drive stimulus choices [56]. One method of accelerating testing times is to make an estimate of sensitivity at a location before any stimuli are shown, and then carefully test around this estimate [56]. SEMU utilizes this approach and predicts sensitivity at a location based on OCT data. Using a computer simulation, the authors tested the performance of SEMU for three different profiles of patient reliability and found reduced testing times while maintaining accuracy and precision [56]. This and other strategies require formal validation before being introduced into routine clinical practice but show progress toward a patient-tailored approach to improve perimetric procedures.

Impact of Testing Frequency

Detecting visual field progression is a significant challenge in clinical practice. The ability to detect progression depends on many factors including the rate of progression, testing frequency, and level of reliability/measurement variability. It is especially important in the early follow-up period to establish a sufficient baseline to rule out rapid progression [57]. Chauhan et al. calculated that to detect rapid progression (defined as −2 dB/year) the time to detect change with 80% power is 5 years with annual examinations, 2.5 years with two examinations per year, and 1.7 years if examinations are performed three times per year [57]. More recently, Wu et al. examined the impact of testing frequency on the ability to detect progression [58]. Assuming a best-practice scenario with two baselines tests and a requirement to replicate progression on one confirmatory test, they estimated rapid visual progression could be detected with 80% power after 3.3 years, 2.4 years, or 2.1 years when testing was performed once, twice, and three times a year [58]. Based on the diminishing returns from twice to three-times-a-year testing, they concluded that twice yearly testing was a reasonable compromise for achieving sufficient power whilst minimizing treatment burden [58].

Novel Methods of Assessing Visual Fields

Traditional perimetry requires the patient to maintain fixation throughout the duration of the test. Failure to maintain fixation can lead to poor reliability and unreliable fields. A relatively new method of testing visual fields is fundus-tracked perimetry or microperimetry where the fundus is tracked using a retinal imaging system and stimuli are projected at specific retinal locations. Early perimeters evaluated only the central macular region while newer machines now permit testing of the central 30° radius. The performance of microperimetry has been compared with the Humphrey Visual Field Analyzer in eyes with glaucoma and the sensitivities obtained with microperimetry have been found to be repeatable and comparable to conventional perimetry [59–61]. Another advantage is that microperimetry can be combined with retinal imaging to provide stronger structure-function associations.

Recent advances in smartphone and tablet technology have seen significant improvements in display resolution, dynamic range, and accurate calibration [62]. These devices are portable, do not require a continuous power supply, and are relatively inexpensive allowing them to be used for home or community-based visual field testing, even in remote areas [62]. The potential applications for portable perimetry include targeted screening in high-risk populations, especially where access to healthcare is limited, or for home monitoring between office visits in patients with a diagnosis of glaucoma. Johnson et al. have evaluated the use of a free tablet-based perimetry application in Nepal and found the procedure to be portable, fast, and effective for detecting moderate to advanced field loss [63]. The average testing time was just over 3 min however improvements are underway to reduce testing time, improve performance, and add head/eye tracking [63]. The performance of this tablet perimeter has been compared against the Humphrey Field Analyzer and the results show strong correlation as well as comparable test-retest reliability [64]. The system has also been used in a study investigating whether home-based perimetry can increase test frequency and allow for detection of rapid progression more quickly than conventional perimetry [63]. Using a computer simulation, tablet-based perimetry detected rapid visual field loss after 0.9 years with a sensitivity of 80% compared to 2.5 years for 6-monthly clinic-based testing [63]. These results suggest that home-based perimetry may be a viable strategy to increase testing frequency and allow for more timely detection of rapid visual field progression [63].

Rho Kinase Inhibitors

Rho kinase (ROCK) inhibitors are an entirely new class of glaucoma medications. These medications work by relaxing the trabecular meshwork through inhibition of the actin cytoskeleton contractile tone of smooth muscle [69, 70]. This results in increased aqueous outflow, thereby lowering IOP. In addition, animal studies suggest secondary effects which may be beneficial in glaucoma including improved blood flow to the optic nerve [71], neuroprotection of retinal ganglion cells [72], and inhibition of bleb scarring following glaucoma filtration surgery [73].

The most well-studied ROCK inhibitors are ripasudil and netarsudil. Ripasudil has been shown to significantly lower IOP in phase I and II human clinical trials [74, 75]. The medication has a good safety profile with the most common side effect being mild hyperaemia, occurring in approximately 50% of patients [74, 75]. Transient corneal guttae-like findings have been seen and are believed to be due to protrusion formation along intracellular borders caused by the reduction actomyosin contractility in corneal endothelial cells [76]. These are not believed to adversely affect vision [76]. In an open-label study of patients with ocular hypertension or glaucoma, 51 of 388 patients had to discontinue the medication due to blepharitis or allergic conjunctivitis symptoms [77]. Monotherapy with ripasudil 0.4% reduced IOP by an average of 3.7 mmHg at 52 weeks [77]. The medication has also been shown to be effective as an adjunctive agent when combined with either a beta-blocker or prostaglandin analogue [78].

Netarsudil is both a ROCK inhibitor and norepinephrine transporter (NET) inhibitor [79]. This medication is believed to lower IOP by the triple action of reducing aqueous production, increasing trabecular outflow, and decreasing episcleral venous pressure [80]. The medication has been found to be effective and well-tolerated for the treatment of patients with ocular hypertension and open-angle glaucoma in two large randomized, double-masked phase 3 trials (ROCKET-1 and ROCKET-2) [81]. Like other ROCK inhibitors, the most common side effect was conjunctival hyperaemia (occurring in 50–89% of study participants) [79, 81]. In a double-masked, randomized study of netarsudil versus latanoprost in patients with elevated IOP, netarsudil was less effective than latanoprost by approximately 1 mmHg [82]. However, the fixed combination of netarsudil and latanoprost was found to be statistically superior in terms of IOP-lowering than its individual active components at the same concentrations [83].

Latanoprostene Bunod

Latanoprostene bunod is a nitrous oxide-donating prostaglandin agonist that lowers IOP by increasing both trabecular and uveoscleral outflow [84]. The release of nitric oxide relaxes the trabecular meshwork, increasing aqueous outflow [85]. In a randomized, controlled trial comparing latanoprostene bunod and latanoprost 0.005% in patients with ocular hypertension and open-angle glaucoma (VOYAGER study), latanoprostene bunod achieved significantly greater reductions in diurnal IOP while having comparable side effects to latanoprost [86]. The levels of hyperaemia were similar in both treatment arms [86]. In larger subsequent randomized, double-masked, multi-center controlled trials comparing latanoprostene bunod and timolol (APOLLO and LUNAR studies) there was a significantly greater reduction in IOP with latanoprostene bunod than timolol [87, 88].

Alternative Drug Delivery Systems

Recently, a number of alternative drug delivery systems have been developed that aim to reduce the need for daily medical therapy; helping address problems with glaucoma eye drop administration and adherence. These devices aim to provide a slow and controlled release of glaucoma medication to provide effective control of IOP.

Laser Iridotomy

Laser peripheral iridotomy is frequently used to prevent or treat angle closure glaucoma. This procedure is generally safe but may be complicated by visual disturbances or dysphotopsias [96]. In a recent randomized prospective paired eye trial, Vera et al. found that temporal placement of the laser iridotomy was less likely to result in linear dysphotopsia than superior placement [97]. The authors suggested that the ideal location for laser iridotomy was the temporal iris [97]. However, a larger multi-center randomized trial in India found that the incidence of visual dysphotopsia was unaffected by iridotomy location, size, or amount of energy used [98]. In this study, the onset of new dysphotopsia occurred in 8.4% of patients undergoing nasal/temporal iridotomy compared to 9.5% of patients where the iridotomy was placed superiorly [98]. Given data suggesting similar safety with regard to location and dysphotopsia symptoms, it may therefore be advisable to place the iridotomy in a crypt or superiorly which has been shown to cause less pain and discomfort [97].

Clear Lens Extraction for Angle-Closure Glaucoma

An alternative approach to the management of angle-closure glaucoma is surgical lens extraction [99]. A landmark trial has recently evaluated the efficacy, safety, and cost-effectiveness of clear lens extraction compared to laser peripheral iridotomy and topical medical treatment as first-line treatment in people with newly diagnosed primary angle closure (PAC) with an IOP of ≥30 mmHg or primary angle-closure glaucoma (PACG) [100]. It was found that clear lens extraction resulted in greater reduction in IOP, less need for glaucoma medications, and higher quality life scores than laser iridotomy [100]. Furthermore, initial lens extraction was more cost-effective than standard of care with laser iridotomy [100]. Based on these results, the authors suggest that clear lens extraction should be considered as first-line treatment for newly diagnosed PACG or PAC where IOP is 30 mmHg or greater [100].

Primary Trabeculectomy or Tube Surgery for Medically Uncontrolled Glaucoma

In eyes with glaucoma refractory to medical therapy, glaucoma surgery is frequently required. The most commonly performed operations are trabeculectomy and tube shunt surgery [101]. Previously, tube shunt surgery was found to have a higher success rate compared to trabeculectomy at 5 years in eyes with prior trabeculectomy and/or cataract surgery [102]. Recently, a landmark study compared the efficacy and safety of tube shunt surgery and trabeculectomy in eyes without prior ocular surgery [103].

In the Primary Tube Versus Trabeculectomy (PTVT) study patients with medically uncontrolled glaucoma and no previous incisional surgery were randomized to treatment with a 350-mm² Baerveldt glaucoma implant or trabeculectomy with mitomycin C (0.4 mg/mL for 2 min). The trabeculectomy arm was found to have a lower probability of failure (7.9% vs. 17.3%), lower IOP (12.4 ± 4.4 mmHg vs. 13.8 ± 4.1 mmHg), and less need for glaucoma medications (0.9 ± 1.4 vs. 2.1 ± 1.4) compared to tube surgery at 1 year [103]. There was no significant difference in the rates of intraoperative complications [103]. However, the frequency of serious complications producing vision loss or requiring reoperation was lower for tube shunt surgery [103].

Minimally Invasive Glaucoma Surgery

New minimally invasive glaucoma procedures that aim to lower intraocular pressure with a better safety profile and faster recovery than conventional glaucoma surgery are being increasingly used in clinical practice. Recently there have been a number of major developments in this space.

The iStent inject received approval from the Food and Drug Administration (FDA) in 2018 for use in mild to moderate open-angle glaucoma in patients undergoing cataract surgery [104]. The iStent inject trabecular micro-bypass system consists of two titanium stents approximately 0.23 mm × 0.36 mm that are implanted into the trabecular meshwork using a preloaded auto-injection system through a single corneal entry [104]. The FDA approval was based on a pivotal iStent inject US IDE pivotal study, a prospective randomized, multi-center clinical trial including 505 participants who were randomized to receive the iStent inject with cataract surgery or cataract surgery alone [104]. At 2 years, 75.8% of patients in the iStent inject group had a 20% or greater reduction in unmedicated diurnal IOP compared with 61.9% in the cataract surgery-only group [104]. The safety profile was similar between the two arms of the study.

Another development in minimally invasive glaucoma surgery (MIGS) is the Hydrus Microstent. The Hydrus is placed in Schlemm’s canal and helps restore aqueous outflow by bypassing the trabecular meshwork, dilating Schlemm’s canal, and allowing access to a number of collector channels over a 90-degree span. In 2018, the Hydrus also received FDA approval to treat patients with mild to moderate primary open-angle glaucoma in conjunction with cataract surgery [105]. The approval was based on the landmark HORIZON trail which included 556 people with mild to moderate glaucoma undergoing cataract surgery. Patients were randomized to receive cataract surgery with the Hydrus Microstent or cataract surgery alone. In the Hydrus group, 77.2% of patients achieved a 20% of greater reduction in unmedicated IOP compared to 57.8% in the cataract surgery alone group at 2 years [105]. Patients who received the Hydrus were twice as likely to be medication-free compared to those who underwent cataract surgery alone [106]. An international multi-center randomized trial comparing the effectiveness of the Hydrus Microstent to two Glaukos iStents in standalone glaucoma, the COMPARE study, is underway [107].

In 2018 the CyPass Micro-Stent was voluntarily withdrawn from sale by the manufacturer due to safety concerns about endothelial cell loss. This voluntary recall has since been updated to a Class 1 recall by the FDA [108]. The CyPass Micro-Stent is a MIGS device that was implanted into the supraciliary space to increase aqueous outflow via the uveoscleral pathway. It was approved based on the results of the COMPASS trial which showed a significant and sustained 2-year reduction in IOP and glaucoma medication use in mild to moderate open-angle glaucoma when performed with cataract surgery [109]. No safety concerns were identified at 2 years with endothelial cell loss being similar between patients who underwent cataract surgery with the CyPass and those who underwent cataract surgery alone [109]. A subset of patients from the COMPASS trial were followed for an additional 3 years in the COMPASS-XT study and based on these results the CyPass was withdrawn from the market. The COMPASS-XT study found that at 5 years there was a higher rate of endothelial cell loss with cataract surgery and CyPass insertion compared to cataract surgery alone [110]. At 5 years, endothelial cell loss was 20.5% in the CyPass group compared to 10.1% in the cataract surgery arm [110]. The rate of endothelial cell loss was found to relate to the depth of insertion. Where the CyPass was implanted with no retention rings visible on gonioscopy the rate of endothelial cell loss was 1.39% per year, where 1 ring was visible the rate was 2.74% per year, however where 2 or more rings were visible the rate increased to 6.96% per year [110]. Surgeons have been advised to cease implanting the CyPass Micro-Stent and to periodically monitor endothelial cell density using specular microscopy where available [110]. The manufacturer is partnering with the FDA and other regulators to explore labelling changes that would support the reintroduction of the CyPass Micro-Stent in the future [111].

Subconjunctival filtration has traditionally delivered the greatest levels of IOP reduction . Trabeculectomy, while effective in reducing IOP, requires extensive dissection, sclerostomy, and suturing which can lead to unpredictability and complications such as bleb leak, hypotony, suprachoroidal haemorrhage, and a reduction in vision [112, 113]. Two new MIGS devices aim to take advantage of the power of subconjunctival filtration while achieving a good safety profile and short surgical time. The XEN is a flexible gelatin implant, 6-mm long, with a 45 μm internal diameter lumen, which is inserted via an ab interno approach from the anterior chamber to the subconjunctival space [114]. The length and diameter of the implant were chosen based on the Hagen-Poiseuille equation to provide sufficient resistance to aqueous outflow to minimize hypotony [114]. The XEN eliminates the need for conjunctival dissection, cutting a scleral flap, sclerosotomy, and iridectomy. The XEN was approved by the FDA in 2016 based on a pivotal trial in patients with refractory glaucoma. In this study, the XEN reduced IOP from a mean medicated baseline of 25.1 ± 3.7 mmHg to 15.9 ± 5.2 mmHg at 12 months [115]. Glaucoma medication use decreased from a mean of 3.5 ± 1.0 to 1.7 ± 1.5 medications over the same period [115]. The effectiveness and safety of the XEN has been compared with trabeculectomy in a retrospective interventional cohort study [112]. The baseline characteristics were similar in both groups and there was no detectable difference in the risk of failure and safety profiles between standalone XEN insertion and trabeculectomy with MMC [112].

The latest subconjunctival MIGS device to be introduced is the InnFocus MicroShunt. This device is 8.5 mm long with a 70 μm lumen and is inserted via an ab externo approach [116]. Like the XEN, the MicroShunt avoids the need for a scleral flap, sclerostomy, iridectomy, and post-operative suturelysis, resulting in a short surgical time and predictable post-operative recovery [117]. The device is manufactured from an inert biocompatible material called poly(styrene-block-isobutylene-block-styrene) or “SIBS.” This material has been shown to elicit minimal foreign body reaction, inflammation, or capsule formation when implanted in the eye [118]. Similar to the XEN, the dimensions of the MicroShunt are based on the Hagen-Poiseuille equation in an attempt to prevent clinically significant hypotony [116]. In a three-year prospective, non-randomised trial the MicroShunt reduced IOP to the low teens in patients with glaucoma refractory to medical therapy for up to 3 years with only transient adverse events in the first 3 months after surgery [117]. In this study, mean medication IOP was reduced from 23.8 ± 5.3 mmHg to 10.7 ± 3.5 mmHg at 3 years with a reduction in the mean number of medications from 2.4 ± 0.9 to 0.7 ± 1.1 [117]. The most common complications were transient hypotony (13%) and transient choroidal effusions (8.7%), all of which resolved spontaneously [117]. A prospective, randomized controlled trial comparing the MicroShunt to trabeculectomy is underway.