Lens extraction for chronic angle‐closure glaucoma

Abstract

Background

Primary angle‐closure glaucoma (PACG) is characterized by a rise in intraocular pressure (IOP) secondary to aqueous outflow obstruction, with relative pupillary block being the most common underlying mechanism. There is increasing evidence that lens extraction may relieve pupillary block and thereby improve IOP control. As such, comparing the effectiveness of lens extraction against other commonly used treatment modalities can help inform the decision‐making process.

Objectives

To assess the effectiveness of lens extraction compared with other interventions in the treatment of chronic PACG in people without previous acute angle‐closure attacks.

Search methods

We searched CENTRAL, MEDLINE, Embase, one other database, and two trials registers (December 2019). We also screened the reference lists of included studies and the Science Citation Index database. We had no date or language restrictions.

Selection criteria

We included randomized controlled trials (RCTs) comparing lens extraction with other treatment modalities for chronic PACG.

Data collection and analysis

We followed standard Cochrane methodology.

Main results

We identified eight RCTs with 914 eyes. We obtained data for participants meeting our inclusion criteria for these studies (PACG only, no previous acute angle‐closure attacks), resulting in 513 eyes included in this review. The participants were recruited from a diverse range of countries. We were unable to conduct meta‐analyses due to different follow‐up periods and insufficient data.

One study compared phacoemulsification with laser peripheral iridotomy (LPI) as standard care. Participants in the phacoemulsification group were less likely to experience progression of visual field loss (odds ratio [OR] 0.35, 95% confidence interval [CI] 0.13 to 0.91; 216 eyes; moderate certainty evidence), and required fewer IOP‐lowering medications (mean difference [MD] ‐0.70, 95% CI ‐0.89 to ‐0.51; 263 eyes; moderate certainty evidence) compared with standard care at 12 months. Moderate certainty evidence also suggested that phacoemulsification improved gonioscopic findings at 12 months or later (MD ‐84.93, 95% CI ‐131.25 to ‐38.61; 106 eyes). There was little to no difference in health‐related quality of life measures (MD 0.04, 95% CI ‐0.16 to 0.24; 254 eyes; moderate certainty evidence), and visual acuity (VA) (MD 2.03 ETDRS letter, 95% CI ‐0.77 to 4.84; 242 eyes) at 12 months, and no observable difference in mean IOP (MD ‐0.03mmHg, 95% CI ‐2.34 to 2.32; 257 eyes; moderate certainty evidence) compared to standard care. Irreversible loss of vision was observed in one participant in the phacoemulsification group, and three participants in standard care at 36 months (moderate‐certainty evidence).

One study (91 eyes) compared phacoemulsification with phaco‐viscogonioplasty (phaco‐VGP). Low‐certainty evidence suggested that fewer IOP‐lowering medications were needed at 12 months with phacoemulsification (MD ‐0.30, 95% CI ‐0.55 to ‐0.05). Low‐certainty evidence also suggested that phacoemulsification may have improved gonioscopic findings at 12 months or later compared to phaco‐VGP (angle grading MD ‐0.60, 95% CI ‐0.91 to ‐0.29; TISA500 MD ‐0.03, 95% CI ‐0.06 to ‐0.01; TISA750 MD ‐0.03, 95% CI ‐0.06 to ‐0.01; 91 eyes). Phacoemulsification may result in little to no difference in best corrected VA at 12 months (MD ‐0.01 log MAR units, 95% CI ‐0.10 to 0.08; low certainty evidence), and the evidence is very uncertain about its effect on IOP at 12 months (MD 0.50 mmHg, 95% CI ‐2.64 to 3.64; very low certainty evidence). Postoperative fibrin reaction was observed in two participants in the phacoemulsification group and four in the phaco‐VGP group. Three participants in the phaco‐VGP group experienced hyphema. No data were available for progression of visual field loss and quality of life measurements at 12 months.

Two studies compared phacoemulsification with phaco‐goniosynechialysis (phaco‐GSL). Low‐certainty evidence suggested that there may be little to no difference in mean IOP at 12 months (MD ‐0.12 mmHg, 95% CI ‐4.72 to 4.48; 1 study, 32 eyes) between the interventions. Phacoemulsification did not reduce the number of IOP‐lowering medications compared to phaco‐GSL at 12 months (MD ‐0.38, 95% CI ‐1.23 to 0.47; 1 study, 32 eyes; moderate certainty evidence). Three eyes in the phaco‐GSL group developed hyphemas. No data were available at 12 months for progression of visual field loss, gonioscopic findings, visual acuity, and quality of life measures.

Three studies compared phacoemulsification with combined phaco‐trabeculectomy, but the data were only available for one study (63 eyes). In this study, low‐certainty evidence suggested that there was little to no difference between groups in mean change in IOP from baseline (MD ‐0.60 mmHg, 95% CI ‐1.99 to 0.79), number of IOP‐lowering medications at 12 months (MD 0.00, 95% CI ‐0.42 to 0.42), and VA measured by the Snellen chart (MD ‐0.03, 95% CI ‐0.18 to 0.12). Participants in the phacoemulsification group had fewer complications (risk ratio [RR] 0.59, 95% CI 0.34 to 1.04), and the phaco‐trabeculectomy group required more IOP‐lowering procedures (RR 5.81, 95% CI 1.41 to 23.88), but the evidence was very uncertain. No data were available for other outcomes.

Authors' conclusions

Moderate certainty evidence showed that lens extraction has an advantage over LPI in treating chronic PACG with clear crystalline lenses over three years of follow‐up; ultimately, the decision for intervention should be part of a shared decision‐making process between the clinician and the patient. For people with chronic PACG and visually significant cataracts, low certainty evidence suggested that combining phacoemulsification with either viscogonioplasty or goniosynechialysis does not confer any additional benefit over phacoemulsification alone. There was insufficient evidence to draw any meaningful conclusions regarding phacoemulsification versus trabeculectomy. Low certainty evidence suggested that combining phacoemulsification with trabeculectomy does not confer any additional benefit over phacoemulsification alone, and may cause more complications instead. These conclusions only apply to short‐ to medium‐term outcomes; studies with longer follow‐up periods can help assess whether these effects persist in the long term.

Plain language summary

What are the benefits and risks of lens extraction in treating chronic primary angle‐closure glaucoma?

Why is this question important?
Primary angle‐closure glaucoma (PACG) is a type of glaucoma which is one of the leading causes of blindness worldwide. It occurs when there are problems with fluid drainage from the eye because the iris (the colored part of the eye) has blocked the drainage channels. A blockage can happen suddenly (acute PACG) or gradually (chronic PACG), causing a build‐up of fluid and raising the pressure inside the eye, which can damage the optic nerve and lead to vision loss. Treatment options include eye drops, laser treatment, and surgery. 'Lens extraction' is a type of surgery where the natural lens is replaced with an artificial lens. This also has the effect of opening up the drainage channels and may help to treat PACG. We reviewed the research evidence to find out how lens extraction compares to other treatments for chronic PACG.

How did we identify and evaluate the evidence?
We searched the medical literature for studies that compared lens extraction with other treatments for chronic PACG, compared the results and summarized the evidence from all the studies, and rated our confidence in the evidence.

What did we find?
We found eight studies comprising 513 eyes with chronic PACG that met our inclusion criteria. The studies followed participants for six to 69 months, and compared lens extraction against:

‐ laser therapy;
‐ lens extraction plus an injection of thick liquid to break iris adhesions with the aim of encouraging fluid outflow (viscogonioplasty, VGP);
‐ lens extraction plus breaking the iris adhesions mechanically, known as goniosynechialysis (GSL);
‐ trabeculectomy (creating a flap to facilitate fluid drainage); and
‐ lens extraction plus trabeculectomy.

These are the main findings of our review, focusing on results one year after treatment (unless otherwise stated).

1. Lens extraction compared with laser therapy (1 study)

When compared against laser therapy, the evidence suggests that lens extraction probably:

‐ limits loss of visual field (the area that can be seen when the eye is looking straight ahead);
‐ reduces the number of pressure‐lowering medicines needed;
‐ open the drainage angle more; and
‐ makes little or no difference to quality of life, vision clarity, or eye pressure.

One person treated with lens extraction, and three people treated with laser therapy, experienced irreversible loss of 10 or more EDTRS letters of vision in the three years after treatment.

2. Lens extraction compared with lens extraction plus VGP (1 study)

When compared against lens extraction plus VGP, the evidence suggests that lens extraction may:

‐ reduce the number of pressure‐lowering medications needed;
‐ open the drainage angle more; and
‐ make little or no difference to clarity of vision.

There is uncertain evidence as to whether the two treatments have different effects on eye pressure. The study did not investigate the effects on visual field loss and quality of life.

Eye inflammation occurred in two people treated with lens extraction, and in four people treated with lens extraction plus VGP. Three people treated with lens extraction plus VGP experienced bleeding in the front of the eye.

3. Lens extraction compared with lens extraction plus GSL (2 studies)

When compared against lens extraction plus GSL, the evidence suggests that lens extraction:

‐ probably does not reduce the number of pressure‐lowering medications needed; and
‐ may make little or no difference to eye pressure.

The studies did not investigate the effects on visual field loss, eye drainage, vision clarity, and quality of life.

Bleeding in the front of the eye occurred in three eyes treated with lens extraction plus GSL.

4. Lens extraction compared with lens extraction plus trabeculectomy (3 studies)

When compared against lens extraction plus trabeculectomy, the evidence from one study suggests that lens extraction may make little or no difference to:

‐ eye pressure;
‐ the number of pressure‐lowering medications needed; and
‐ vision clarity.

There is uncertain evidence as to whether one treatment leads to more unwanted effects than the other. The studies did not investigate the effects on visual field loss, eye drainage, and quality of life.

What does this mean?
The evidence suggests that:

‐ lens extraction is probably a better treatment than laser therapy for chronic PACG;
‐ combining lens extraction with VGP or GSL may not work better than lens extraction alone; and
‐ there is uncertain evidence as to whether combining lens extraction with trabeculectomy makes a difference.

How‐up‐to date is this review?
The evidence in this Cochrane Review is current to 13 December 2019.

Summary of findings

Summary of findings 1. Phacoemulsification compared with laser peripheral iridotomy for primary angle‐closure glaucoma.

Phacoemulsification compared with laser peripheral iridotomy for primary angle‐closure glaucoma
Patient or population: participants with primary angle‐closure glaucoma Settings: 30 hospital eye services in Australia (1), mainland China (1), Hong Kong (2), Malaysia (2), Singapore (2), and the UK (22) Intervention: phacoemulsification Comparison: laser peripheral iridotomy (LPI)
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of eyes (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	laser peripheral iridotomy	phacoemulsification
Progression of visual field loss (worsening of one or more stages, according to the Glaucoma Staging System‐2) (at 12 months)	165 per 1000	65 per 1000 (25 to 152 per 1000)	OR 0.35 (0.13 to 0.91)	216 (one study)	⊕⊕⊕⊝ moderatea	‐
Mean IOP change from baseline to 12 months (mmHg)	The mean change in IOP in the control group was ‐6.48 mmHg	The mean change in IOP in the intervention group was 0.03 mmHg higher (95% CI ‐2.34 mmHg to 2.32 mmHg)	‐	257 (one study)	⊕⊕⊕⊝ moderatea	‐
Mean number of medications to control IOP (at 12 months)	On average, the number of medications in the control group was 0.98	On average, the number of medications in the intervention group was 0.70 lower (95% CI ‐0.89 to ‐0.51)	‐	263 (one study)	⊕⊕⊕⊝ moderatea	‐
Gonioscopic findings (Degree of angle closure) (at 12 months or later)	The mean angle closure in the control group was 203°	The mean angle closure in the intervention group was 84.93° less (95% CI 38.61° to 131.25°)		106 (one study)	⊕⊕⊕⊝ moderatea	‐
Visual acuity (ETDRS letter chart) (at 12 months)	The mean visual acuity in the control group was 77.4	The mean visual acuity in the intervention group was 2.03 letters greater (95% CI ‐0.77 to 4.84)	‐	242 (one study)	⊕⊕⊕⊝ moderatea	‐
Adverse effects	No data available		‐	‐	‐	‐
Quality of life measures (measured on the EQ‐5D; higher = better) (at 12 months)	The average score on the EQ‐5D in the control group was 2.88	The average score on the EQ‐5D in the intervention group was 0.04 higher (95% CI ‐0.16 to 0.24)	‐	254 (one study)	⊕⊕⊕⊝ moderatea	‐
*The assumed risk is based on the estimate in the control group. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; IOP: intraocular pressure; MD: mean difference; OR: odds ratio; EQ‐5D: European Quality of Life‐5 Dimension
GRADE Working Group grades of evidence High certainty. Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty. Further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate. Low certainty. Further research is very likely to have an important impact on our confidence in the estimate of effect, and is likely to change the estimate. Very low certainty. We are very uncertain about the estimate.

^aDowngraded one level for imprecision because sample size was not adequately powered (original study was powered to investigate participants with both primary angle closure (PAC) and primary angle‐closure glaucoma (PACG), whereas only participants with PACG were included in this analysis).

Summary of findings 2. Phacoemulsification versus phacoemulsification plus viscogonioplasty for primary angle‐closure glaucoma.

Phacoemulsification versus phacoemulsification plus viscogonioplasty for primary angle‐closure glaucoma
Patient or population: participants with primary angle‐closure glaucoma Settings: university hospital in Iran Intervention: phacoemulsification Comparison: phacoemulsification plus viscogonioplasty (phaco‐VGP)
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	phacoemulsification plus viscogonioplasty	phacoemulsification
Progression of visual field loss	No data available for this outcome		‐	‐	‐	‐
Mean IOP change from baseline to 12 months (mmHg)	The mean change in IOP in the control group was ‐8.8 mmHg	The mean change in IOP in the intervention group was ‐8.3 mmHg; on average, 0.50 mmHg higher (95% CI ‐2.64 to 3.64)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Mean number of medications to control IOP (at 12 months)	The mean number of medications to control IOP in the control group was 0.4	The mean number of medications to control IOP in the intervention groups was 0.1; on average, 0.30 fewer (95% CI ‐0.55 to ‐0.05)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	at 6 months: MD ‐0.30 (95% CI ‐0.56 to ‐0.04; 1 study, 91 eyes)
Gonioscopic findings (Spaeth grading system) (at 12 months or later)	The mean change of angle grading in the control group was 2.0	The mean change of angle grading in the intervention group was 1.4; on average, 0.60 less (95% CI ‐0.91 to ‐0.29)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Gonioscopic findings (TISA500) (at 12 months)	The mean TISA500 in the control group was 0.054	The mean TISA500 in the intervention group was 0.020; on average, 0.03 less (95% CI ‐0.06 to ‐0.01)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Gonioscopic findings (TISA750) (at 12 months)	The mean TISA750 in the control group was 0.119	The mean TISA750 in the intervention group was 0.084; on average, 0.03 less (95% CI ‐0.06 to ‐0.01)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Visual acuity (log MAR units) (postoperatively)	The mean best corrected visual acuity in the control group was 0.28	The mean best corrected visual acuity in the intervention group was 0.27; on average, 0.01 less (95% CI ‐0.10 to 0.08)	‐	91 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Adverse effects (at 12 months)	hyphema (3 eyes), postoperative fibrin reaction (4 eyes)	postoperative fibrin reaction (2 eyes)	‐	‐	‐
Quality of life measures	No data available for this outcome		‐	‐	‐	‐
*The assumed risk is based on the estimate in the control group. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; IOP: Intraocular pressure; TISA: trabecular iris space area
GRADE Working Group grades of evidence High certainty. Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty. Further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate. Low certainty. Further research is very likely to have an important impact on our confidence in the estimate of effect, and is likely to change the estimate. Very low certainty. We are very uncertain about the estimate.

^aDowngraded for high risk of attrition bias
^bDowngraded one level for imprecision due to wide confidence intervals

Summary of findings 3. Phacoemulsification versus phacoemulsification plus goniosynechialysis for primary angle‐closure glaucoma.

Phacoemulsification versus phacoemulsification plus goniosynechialysis for primary angle‐closure glaucoma
Patient or population: participants with primary angle closure glaucoma Settings: tertiary eye care center and university hospital in Vietnam, Thailand, and Hong Kong Intervention: phacoemulsification Comparison: phacoemulsification plus goniosynechialysis (phaco‐GSL)
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	phacoemulsification plus goniosynechialysis	phacoemulsification
Progression of visual field loss	No data available for this outcome		‐	‐	‐	‐
Mean IOP change from baseline (mmHg) (at 12 months)	The mean change in IOP in the control group was ‐6.47 mmHg	The mean change in IOP in the intervention group was ‐6.59; on average, 0.12 lower (95% CI ‐4.72 to 4.48)	‐	32 eyes (one study)	⊕⊕⊕⊝ moderatea	at 6 months: MD ‐0.04 mmHg (95% CI ‐0.93 to 0.85; 1 study, 64 eyes)
Mean number of medications to control IOP (at 12 months)	The mean number of medications to control IOP in the control group was 0.94	The mean number of medications to control IOP in the intervention groups was 0.56; on average, 0.38 fewer (95% CI ‐1.23 to 0.47)	‐	32 eyes (one study)	⊕⊕⊕⊝ moderatea	at 6 months: MD ‐0.35 (95% CI ‐0.63 to ‐0.07; 1 study, 64 eyes)
Gonioscopic findings (Anterior segment optical coherence tomography parameters) (at 6 months)	See comments	See comments	‐	64 eyes (one study)	‐	AOD500: MD ‐0.04° (95% CI ‐0.27 to 0.19) AOD750: MD 0.01° (95% CI ‐0.27 to 0.29) TISA500: MD ‐0.02° (95% CI ‐0.06 to 0.02) TISA750: MD ‐0.03° (95% CI ‐0.17 to 0.11) SSA: MD ‐1.59° (95% CI ‐6.75 to 3.57)
Visual acuity (log MAR units)	No data available for this outcome		‐	‐	‐	‐
Adverse effects (at 6 months)	hyphema (3 eyes)	none reported	‐	‐	‐
Quality of life measures	No data available for this outcome		‐	‐	‐	‐
*The assumed risk is based on the estimate in the control group. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). AOD: angle opening distance; CI: confidence interval; IOP: intraocular pressure; MD: mean difference; SSA: scleral spur angle; TISA: trabecular iris space area
GRADE Working Group grades of evidence High certainty. Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty. Further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate. Low certainty. Further research is very likely to have an important impact on our confidence in the estimate of effect, and is likely to change the estimate. Very low certainty. We are very uncertain about the estimate.

^aDowngraded one level for imprecision due to wide confidence intervals

Summary of findings 4. Phacoemulsification versus phacoemulsification plus trabeculectomy for primary angle‐closure glaucoma.

Phacoemulsification versus phacoemulsification plus trabeculectomy for primary angle‐closure glaucoma
Patient or population: participants with primary angle‐closure glaucoma Settings: university hospital in Egypt Intervention: phacoemulsification Comparison: phacoemulsification plus trabeculectomy
Outcomes	Illustrative comparative risks* (95% CI)		Relative effect (95% CI)	No of Participants (studies)	Certainty of the evidence (GRADE)	Comments
	Assumed risk	Corresponding risk
	Phacoemulsification combined with trabeculectomy	Phacoemulsification
Progression of visual field loss (at 12 months)	No data available for this outcome		‐	‐	‐
Mean IOP change from baseline to 12 months (mmHg)	The mean IOP in the control group was 13.2 mmHg	The mean IOP in the intervention group was 12.6 mmHg; on average, 0.60 mmHg lower (95% CI ‐1.99 to 0.79)	‐	63 eyes (one study)	⊕⊕⊝⊝ lowa,b
Mean number of medications to control IOP (at 12 months)	The mean number of medications to control IOP in the control group was 0.5	The mean number of medications to control IOP in the intervention group was 0.5; on average, there was no difference (95% CI ‐0.42 to 0.42)	‐	63 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Gonioscopic findings	No data available for this outcome		‐	‐	‐	‐
Mean best corrected visual acuity (on Snellen chart) (at final follow‐up)	The mean best corrected visual acuity in the control group was 0.38	The mean best corrected visual acuity in the intervention group was 0.35: on average, 0.03 lower (95% CI ‐0.18 to 0.12)		63 eyes (one study)	⊕⊕⊝⊝ lowa,b	‐
Adverse effects (up to 12 months)	Intraoperative and postoperative complications		RR 0.59 (0.34 to 1.04)	63 eyes (one study)	⊕⊕⊝⊝ lowa,b
	580.6 per 1000	343.8 per 1000
	Additional IOP‐lowering procedures required		RR 5.81 (1.41 to 23.88)
	64.5 per 1000	375 per 1000
Quality of life measures (at 12 months)	No data available for this outcome		‐	‐	‐	‐
*The assumed risk (e.g. the median control group risk across studies) is based on the estimate in the control group. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: confidence interval; IOP: intraocular pressure; RR: risk ratio
GRADE Working Group grades of evidence High certainty. Further research is very unlikely to change our confidence in the estimate of effect. Moderate certainty. Further research is likely to have an important impact on our confidence in the estimate of effect, and may change the estimate. Low certainty. Further research is very likely to have an important impact on our confidence in the estimate of effect, and is likely to change the estimate. Very low certainty. We are very uncertain about the estimate.

^aDowngraded for high risk of attrition bias
^bDowngraded one level for imprecision due to wide confidence intervals

Background

Description of the condition

Glaucoma is an umbrella term for a group of progressive optic neuropathies that are typically associated with elevated intraocular pressure (IOP), and result in characteristic patterns of visual loss and optic disc cupping. Primary angle‐closure glaucoma (PACG) is a subtype characterized by anatomical iridotrabecular contact, with consequent elevated IOP and glaucomatous optic neuropathy (Casson 2011).

Epidemiology

Glaucoma is the third most common cause of visual impairment, and the most common irreversible cause of blindness worldwide (Stevens 2013). Over 60 million people are living with glaucoma, and the numbers are expected to continue rising (Tham 2014). While primary open‐angle glaucoma is the most common subtype, PACG results in a disproportionately higher prevalence of blindness, accounting for almost half the cases of bilateral blindness due to glaucoma (Baskaran 2015; Dandona 2000; Foster 2001; Quigley 2006).

The prevalence of PACG varies across the globe. It is most common in Asian populations, who comprise up to 86.5% of PACG diagnoses worldwide (Quigley 2006; Tham 2014). Other risk factors include female gender, increasing age, and family history (Amerasinghe 2008; Amerasinghe 2011; Kavitha 2014; Qu 2011; Quigley 2006). Anatomical factors that increase the risk of PACG include hypermetropia, reduced axial length, increased lens thickness, and shallow anterior chamber depth (Devereux 2000; Lowe 1970; Nongpiur 2011; Salmon 1999; Sng 2012). Several potentially responsible candidate genes have been identified, but the relationship is complex, and the mechanisms are as yet unclear (Shastry 2013).

Clinical Presentation and Diagnosis

Chronic PACG presents insidiously – over time, patients may develop visual field defects characteristic of glaucoma if the longstanding pressure damage to the optic nerve is not corrected, which means that it is often asymptomatic until the late stages. Signs include optic disc cupping, elevated IOP, and a narrow or closed iridocorneal angle on gonioscopy. Ultrasound biomicroscopy and anterior segment optical coherence tomography may also be used to assess angle
configuration.

Description of the intervention

PACG treatment aims to lower IOP and thereby arrest optic nerve damage. This can be achieved by a stepped approach, using medical treatment (ocular hypotensive agents), laser (laser peripheral iridotomy, argon laser peripheral iridoplasty, selective laser trabeculoplasty), surgical means (lens extraction, trabeculectomy, goniosynechialysis), or any combination of these treatments.

Lens extraction is increasingly viewed as a viable treatment for PACG. This is typically achieved with phacoemulsification and intraocular lens (IOL) implantation, although older techniques such as extracapsular cataract extraction may uncommonly be used complex cases where phacoemulsification proves difficult, and a variation thereof (manual small incision cataract surgery) is commonly used in developing countries.

How the intervention might work

Relative pupillary block and angle crowding are the two main mechanisms underlying the pathogenesis of PACG (Wright 2015).

In relative pupillary block, there is obstruction of the free flow of aqueous humor from the posterior to the anterior chamber of the eye. The increased pressure gradient across the pupil forces the iris to bow forwards, narrowing the iridocorneal angle, and reducing access to the trabecular meshwork through the apposition of lens and iris at the pupil. Prolonged iridotrabecular contact can also lead to the formation of peripheral anterior synechiae (PAS), which can exacerbate the pupillary block.

Another mechanism is angle crowding, which often coexists with pupillary block, but may also cause PACG by itself. Angle crowding may be due to plateau iris, an anatomical configuration wherein the peripheral iris root is compressed against the trabecular meshwork by more anteriorly positioned ciliary processes, thus predisposing to iridocorneal contact (Ritch 2003; Shukla 2008; Wand 1977). It may also result from other abnormal iris configurations, such as a thick peripheral iris roll (Shabana 2012).

More recent studies have established that exaggerated lens vault plays an important role in the pathogenesis of PACG. Compared to normal eyes, eyes with PACG were found to have thickened crystalline lens with a greater lens vault that displaces the iris anteriorly, thereby reducing the anterior chamber depth (Ozaki 2012; Nongpiur 2011; Shabana 2012). As such, lens extraction can plausibly relieve pupillary block and angle crowding, thus aiding in IOP control. Indeed, there is evidence that replacing the thickened cataractous lens with a synthetic IOL can deepen the anterior chamber depth by 1.0 mm or more (Hayashi 2000; Yang 1997), and early non‐randomized trials showed significant improvement in IOP in PACG patients following cataract extraction (Hayashi 2001; Kubota 2003).

Why it is important to do this review

There is some evidence that lens extraction may help slow the progression of PACG. However, its effectiveness compared to other conventional interventions has not been systematically evaluated. In addition, there is no consensus as to whether lens extraction is sufficiently effective that it should be performed on PACG patients without coexisting cataract (i.e., clear lens extraction), which is important because the surgery is not risk‐free.

The original version of this review did not find any randomized controlled trials on the topic, and the two eligible non‐randomized studies did not provide proof of the effectiveness of lens extraction over other interventions for chronic PACG (Friedman 2006). There has been increasing interest in this subject since then. With the rising disease burden, it is timely for a systematic examination of the evidence on the utility of lens extraction in the treatment of
chronic PACG.

Objectives

To assess the effectiveness of lens extraction compared to other interventions in the treatment of chronic primary angle‐closure glaucoma in people without previous acute angle‐closure attacks.

Methods

Criteria for considering studies for this review

Types of studies

We included only randomized controlled trials (RCT) in the updated review. We excluded the non‐randomized comparative studies from the original version of this review. We did not place any restrictions relating to language, date of publication, or number of participants.

Types of participants

We included studies that enrolled participants diagnosed with chronic primary angle‐closure glaucoma (PACG), defined as gonioscopic evidence of angle closure in association with either glaucomatous optic neuropathy with or without visual field defects, or elevated intraocular pressure (IOP).

We excluded participants with a known history of acute angle‐closure attacks. This population may develop chronic PACG, but the disease may differ substantially from those with asymptomatic disease (Ang 2004; Chen 2012; Tham 2009a). Where studies included both participants with chronic PACG with known symptomatic attacks in the past and those without, we attempted to procure data on the latter group alone. We placed no restrictions on age, gender, ethnicity, comorbidities, use of adjunctive medications, or number of participants.

Types of interventions

We included studies that compared lens extraction with other treatment modalities for chronic PACG, including but not limited to laser iridotomy, medications, and laser iridoplasty.

Types of outcome measures

Primary outcomes

(1) Proportion of participants with evidence of progression of visual field loss at different follow‐up time points. The main primary outcome was at one year of follow‐up. We adopted the criteria in the included studies to define progression of visual field loss as measured using a validated method.

(2) Mean change in IOP from baseline to one year, measured by any method. We also planned to report the mean change in IOP from baseline to different follow‐up time points. If data on mean change from baseline and its standard deviation were not reported in the manuscript and could not be obtained from the authors, we planned to analyze the final mean IOPs at different follow‐up time points.

Secondary outcomes

(1) Mean change in depth of the anterior chamber from baseline in millimeters, measured by any method

(2) Number of medications used to control IOP at six months, one year, and at different follow‐up time points, as reported in the included studies

(3) Gonioscopic findings – we planned to summarize the available information on the examination of the angle, including angle width, from the included studies.

(4) Visual acuity as reported in the included studies.

Adverse effects

We summarized adverse effects related to lens extraction reported in the included studies.

Quality of life measures

We planned to summarize quality of life outcomes reported in the included studies.

Follow‐up

There were no restrictions based on length of follow‐up.

Search methods for identification of studies

Electronic searches

The Cochrane Eyes and Vision Information specialist conducted systematic searches in the following databases for RCTs and controlled clinical trials. We placed no restrictions on language or year of publication. The date of the search was 13 December 2019.

• Cochrane Central Register of Controlled Trials (CENTRAL; 2019, Issue 12 [which contains the Cochrane Eyes and Vision Trials Register]) in the Cochrane Library (searched 13 December 2019; Appendix 1);

• MEDLINE Ovid (1946 to 13 December 2019; Appendix 2);

• Embase Ovid (1980 to 13 December 2019; Appendix 3);

• LILACS (Latin American and Caribbean Health Sciences Literature Database; 1982 to 13 December 2019; Appendix 4);

• US National Institutes of Health Ongoing Trials Register ClinicalTrials.gov (www.clinicaltrials.gov; searched 13 December 2019; Appendix 5);

• World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP; www.who.int/ictrp; searched 13 December 2019; Appendix 6).

Searching other resources

We searched the reference lists of the included studies. We planned to contact the primary investigators of identified RCTs for details of any additional trials not found by our electronic and manual searches. We used the Science Citation Index to search for references that cited the included studies.

Data collection and analysis

Selection of studies

Two review authors independently assessed the titles and abstracts of all reports identified by the electronic and manual searches. We labeled each report as (a) definitely exclude, (b) unsure, or (c) definitely include. We excluded studies labeled as 'definitely exclude'. We assessed studies labeled as 'definitely include' for risk of bias. We assessed the full text of abstracts labelled as 'unsure' according to the inclusion criteria for this review. We recorded the studies excluded after the full‐text review as well as the reasons for exclusion. We resolved any discrepancies between our assessments through discussion.

Data extraction and management

Two review authors independently recorded primary and secondary outcome data in Covidence (Covidence). We resolved discrepancies through discussion. One review author entered the data into Review Manager 5; a second review author verified the accuracy (Review Manager 2014). We contacted study authors to clarify any points of doubt.

Assessment of risk of bias in included studies

For each included trial, two review authors independently assessed the methodological quality and risk of bias according to the guidelines in Chapter 8 of the Cochrane Handbook for Systematic Review of Interventions (Higgins 2011). We assessed the following domains:

1. Sequence generation (selection bias): randomization methods such as computer‐generated random numbers were considered to be at low risk of bias; methods such as alternation were considered to be at high risk of bias; methods using medical record numbers, social security numbers, etc. were considered quasi‐randomized, and were consequently deemed to be at high risk of bias as well.

2. Allocation concealment (selection bias): methods such as central randomization and the use of sequentially numbered sealed envelopes were considered to be at low risk of bias. If participants or investigators were able to foresee upcoming assignments (e.g. using open random allocation schedule, date of birth, alternation etc.), the study was judged as having high risk of bias.

3. Masking (performance bias and detection bias): we assessed masking of outcome assessors where masking was feasible. Masking of participants and physicians was not feasible due to the nature of the intervention studied. If outcome assessors were adequately masked to the assignments, we judged the study as having low risk of bias. If masking was incomplete and outcome measurements were likely to be influenced by lack of masking, the study was judged as having high risk of bias for this domain.

4. Incomplete outcome data (attrition bias): we assessed methods for handling loss to follow‐up and missing data. We judged the study as having low risk of bias if there were no missing outcome data, or where intention‐to‐treat analysis was followed. We judged the study as having high risk of bias if intention‐to‐treat analysis was not followed, or reasons for missing outcome data were likely to be related to true outcome. We attempted to contact trial authors for missing data. Where analyses involved imputing data on participants who dropped out or were lost to follow‐up, we extracted information on methods used for such imputation.

5. Selective outcome reporting (reporting bias): we attempted to assess for evidence of selective reporting of outcomes or analyses on selective populations by comparing the published results to the original protocol. We judged the study as having low risk of reporting bias if the study protocol was published and all pre‐specified outcomes were adequately reported. If not all pre‐specified outcomes or outcomes prescribed in Methods section were reported or outcomes were reported incompletely, we judged the study as having high risk of bias for this domain.

6. Other sources of bias: we evaluated evidence of other sources of bias. We judged high risk of bias for this domain if the study had a potential source of bias due to the specific study design, and judged as low risk of bias if none of such bias was identified. We judged as having unclear risk of bias if there was insufficient information to allow judgement.

Measures of treatment effect

We calculated summary risk ratios (RR) with 95% confidence intervals (CIs) for dichotomous outcomes. In this review, dichotomous outcomes included the primary outcome ‘proportion of patients with evidence of progression of visual field loss.’ For one trial, we presented odds ratio (OR) ‐ one author (SSV) analyzed the individual participant data acquired from the original trial, following its prespecified statistical analysis plan to account for correlation within participants (Azuara‐Blanco 2016). This analysis was performed using Stata (Stata).

We calculated the mean difference (MD) with 95% CIs for normally distributed continuous outcomes. We did not use the standardized mean difference (SMD) because all included studies reported outcomes on the same scale. In this review, continuous outcomes included the primary outcome ‘mean change in IOP’, and the secondary outcomes ‘mean change in anterior chamber depth’, ‘mean change in number of medications used to control IOP’, ‘gonioscopic findings’, and ‘mean change in visual acuity’.

Unit of analysis issues

In one study, the participant was used as the unit of randomization, while the eye was used as the unit of analysis. However, the analysis did not take into account correlation between eyes of the same participant (Moghimi 2015). Another study used the eye as a unit of analysis, without explicitly stating whether one or both eyes of a participant were included (Angmo 2019).

Dealing with missing data

We contacted trial investigators to obtain additional information, individual participant data, or both. We used the published data available when we did not receive a response.

Assessment of heterogeneity

We quantified the proportion of intra‐study variability that could be explained by heterogeneity using the I² statistic (Higgins 2002). We examined the inter‐study variance using Tau². When we observed substantial statistical heterogeneity (defined here as I² > 50%) in a particular outcome, we did not combine the study results in a meta‐analysis, but reported them in a qualitative (narrative) summary instead.

Assessment of reporting biases

For selective outcome reporting, we compared the prespecified outcomes in protocols or trial registries with outcomes reported in published papers. For the studies for which the protocol was not publicly available, we compared the outcomes presented in the Methods section with those reported in the Results section of the published papers. We did not assess for publication bias as we did not include 10 or more studies in meta‐analysis.

Data synthesis

Data analysis followed the guidelines set out in Chapter 9 of the Cochrane Handbook for Systematic Review of Interventions (Deeks 2011).

We planned to conduct a meta‐analysis if there was no substantial heterogeneity. We planned to use the fixed‐effects model for outcomes with fewer than three studies, and the random‐effects model for those containing three or more studies.

Subgroup analysis and investigation of heterogeneity

Where heterogeneity was high (I² > 50%), we planned subgroup analyses to explore the underlying causes. This was conditional on having sufficient data to carry this out.

Sensitivity analysis

Where appropriate, we planned sensitivity analyses to determine the impact of excluding randomized studies of higher risk of bias, quasi‐randomized studies, unpublished studies, or industry‐funded studies from the meta‐analysis. This was ultimately unnecessary because of the limited number of eligible studies.

Summary of findings and assessment of the certainty of the evidence

We created 'Summary of findings' tables for the main comparisons. Two review authors independently graded the quality of the body of evidence for each outcome as one of four levels (i.e. high, moderate, low, or very low) using the GRADE approach (GRADEpro GDT). We resolved disagreement by discussion and obtaining consensus within the review team. We downgraded the certainty of the body of evidence by assessing the following parameters:

1. High risk of bias among included studies;

2. Indirectness of evidence;

3. Unexplained heterogeneity or inconsistency of results;

4. Imprecision of results (i.e. wide confidence intervals);

5. High probability of publication bias.

Results

Description of studies

Results of the search

The electronic searches for the original review were conducted in July 2005 and April 2006; the details were described previously (Friedman 2006). In brief, no relevant randomized controlled trials (RCT) were identified from 395 records at the time. An updated search on 13 December 2019 yielded 2381 records. Five additional records were identified after searching through reference lists. After screening titles and abstracts, we retrieved 52 full‐text reports for further evaluation. After screening the full texts, we included eight relevant studies (17 records); identified one ongoing study (one record), six studies (six records) await classification, and we excluded 28 studies (28 records). In total, we have excluded 45 studies listed in the 'Characteristics of excluded studies' table. We listed six studies in the 'Characteristics of studies awaiting classification' section; we were unable to retrieve two articles, or clarify the eligibility of participants for the remaining four studies (one of which was identified from a manual search). We listed the protocol of an ongoing RCT which commenced in April 2019 under Characteristics of ongoing studies (ChiCTR1900022198). We attempted to contact the relevant authors. We will update the review with relevant information if and when they become available. The flow diagram of the search results is shown in Figure 1.

1.

Study flow diagram

Included studies

Types of studies

The previous version of this review comprised only non‐randomized comparative studies as no eligible RCTs were identified at the time. For this update, we focused solely on RCTs.

We included a total of eight RCTs in this update. Three studies were initially eligible for inclusion in the review (Angmo 2019, El Sayed 2019, Moghimi 2015). Another included participants with both primary angle‐closure glaucoma (PACG) and primary angle closure (PAC), but subsequently provided data for participants with PACG, such that we could perform individual patient data analysis for inclusion in the review (Azuara‐Blanco 2016). Four studies did not meet the eligibility criteria initially, but we obtained data on participants with PACG without a previous history of symptomatic acute angle closure attacks (as per our eligibility criteria) from the trial authors, which we have included below (Husain 2019; Tham 2008; Tham 2009; Tham 2013).

Types of participants

The eight eligible RCTs included comprised 914 eyes in total. Of these, 513 eyes met our inclusion criteria of having PACG without a previous history of acute angle closure attacks. This number does not include three studies, for which the number of eligible eyes was unclear (Tham 2008; Tham 2009; Tham 2013).

Azuara‐Blanco 2016 was a multicenter trial that recruited from across the United Kingdom, Australia, and Southeast Asia (China, Hong Kong, Malaysia, and Singapore). Husain 2019 recruited participants from Southeast Asia (Hong Kong, Singapore, Thailand, Vietnam), as did Tham 2008, Tham 2009, and Tham 2013, all of which recruited from Hong Kong and China. Angmo 2019 took place in India, Moghimi 2015 in Iran, and El Sayed 2019 in Egypt.

PACG was defined as occludable angles (at least 180 degrees of iridotrabecular contact in all studies, apart from Moghimi 2015, which specified at least 270 degrees); glaucomatous visual field loss or glaucomatous optic neuropathy, or both; and elevated IOP. Azuara‐Blanco 2016 enrolled participants who were newly diagnosed with PACG, as well as those with PAC with an intraocular pressure (IOP) of 30 mmHg or higher. Notably, study participants did not have symptomatic cataracts. This was also the case for Tham 2013, which only evaluated participants with medically uncontrolled PACG without cataracts. Conversely, all participants in Angmo 2019, Husain 2019; Moghimi 2015, Tham 2008; Tham 2009 had PACG as well as visually significant cataracts in the study eye. The final study comprised a mix of eyes with PACG with and without cataracts (El Sayed 2019).

Four studies included participants with a previous history of acute angle closure attacks (Husain 2019; Tham 2008; Tham 2009; Tham 2013). We obtained data for those with PACG without such a history through private correspondence with the authors of these studies.

All eight studies had two study arms, and all the baseline characteristics were equivalent in both arms, apart from Husain 2019, which reported a discrepancy in the mean central corneal thickness across study arms.

Types of interventions

Azuara‐Blanco 2016 compared phacoemulsification (clear lens extraction) and standard care (laser peripheral iridotomy). For the phacoemulsification group, synechialysis was allowed according to local practice. For the standard care group, laser iridoplasty was subsequently allowed if angle closure persisted, and participants could undergo lens extraction only if they subsequently developed a visually significant cataract or if their ophthalmologist recommended this to help with IOP control.

Moghimi 2015 compared phacoemulsification and phacoemulsification combined with viscogonioplasty (phaco‐VGP), a technique for breaking the PAS using a cohesive viscoelastic. This was injected near the iridocorneal angle twice for 360°, without touching the angle with the cannula, i.e. no surgical instruments were used to mechanically break the PAS, in contrast to phaco‐GSL.

Both Angmo 2019 and Husain 2019 compared phacoemulsification and phacoemulsification combined with goniosynechialysis (phaco‐GSL). The surgeons in Angmo 2019 used a viscoelastic cannula to mechanically break the PAS along approximately 270 degrees of the iridocorneal angle. In Husain 2019, GSL was performed using an iris repositor or similar.

Three studies compared phacoemulsification versus combined phaco‐trabeculectomy (El Sayed 2019; Tham 2008; Tham 2009).

One compared phacoemulsification versus trabeculectomy with adjunctive mitomycin C (MMC) (Tham 2013).

Types of outcomes

Primary outcomes

Two studies evaluated the proportion of participants with evidence of progression of visual field loss at 24 months (Tham 2008; Tham 2009). None of the eight studies assessed this parameter at 12 months postoperatively, although Azuara‐Blanco 2016 reported progression in visual field loss at 12, 24, and 36 months.

Two studies reported the mean change in IOP from baseline to 12 months (Husain 2019; Moghimi 2015). The remaining six reported the final mean IOP at various time points: Angmo 2019 up to six months; El Sayed 2019 up to 12 months; three up to 24 months, including at 12 months (Tham 2008; Tham 2009; Tham 2013); Azuara‐Blanco 2016 up to 36 months (including at 12 months). We also obtained the raw data from the authors of Azuara‐Blanco 2016 to calculate the mean change in IOP from baseline to 12 months.

Secondary outcomes

None of the eight studies reported the mean change in anterior chamber depth (ACD) from baseline.

All eight studies evaluated the mean number of IOP‐lowering medications required at various follow‐up intervals: Angmo 2019 at six months; Azuara‐Blanco 2016 at 6,12, 24, and 36 months; El Sayed 2019 and Moghimi 2015 at 1, 3, 6, and 12 months; Husain 2019 at 12 months; and Tham 2008, Tham 2009, and Tham 2013 at 3‐monthly intervals up to 24 months.

Two studies reported the mean change in gonioscopic grading from baseline: one at 12 months postoperatively (Husain 2019), and the other at the last follow‐up visit, which appears to have varied for each participant (Moghimi 2015). Azuara‐Blanco 2016 reported the degree of appositional and synechial angle closure at 36 months postoperatively. Two studies also measured changes in the iridocorneal angle in terms of quantitative anterior segment optical coherence tomography (AS‐OCT) parameters (Angmo 2019; Moghimi 2015).

Five studies described the best corrected visual acuity (BCVA) at various time points: Azuara‐Blanco 2016 at 12 and 36 months; El Sayed 2019 at the final follow‐up (varying intervals after 12 months for each individual); and Tham 2008, Tham 2009, and Tham 2013 at 12 and 24 months postoperatively.

Adverse effects

Seven studies reported both intraoperative and postoperative complications in both treatment arms (Azuara‐Blanco 2016; El Sayed 2019; Husain 2019; Moghimi 2015; Tham 2008; Tham 2009; Tham 2013). One described only postoperative complications, and it was unclear whether intraoperative complications were documented (Angmo 2019).

Quality of life measures

Only one study included self‐reported quality of life as an outcome. Azuara‐Blanco 2016 measured health status using the European Quality of Life‐5 Dimensions (EQ‐5D), as well as vision‐related quality of life, using the National Eye Institute Visual Function Questionnaire‐25 and the Glaucoma Utility Index.

Excluded studies

The 45 studies excluded after full text review are briefly described in the 'Characteristics of excluded studies' section. In summary, we excluded 29 studies due to incorrect study design (not RCTs), ten studies due to ineligible participants (not participants with chronic PACG), and six studies due to interventions.

Risk of bias in included studies

We summarized the risk of bias assessments in Figure 2.

2.

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included study

Allocation

All eight studies reported adequate methods of generating the randomization schedule. Five studies used random number tables (Angmo 2019; El Sayed 2019; Tham 2008; Tham 2009; Tham 2013) ; one used computer‐generated random blocks (Moghimi 2015); one randomized participants in a 1:1 ratio by way of a random‐number generator (Husain 2019); while one created a randomization schedule with the aid of a web‐based application that used sex, center, ethnic origin, diagnosis, and one or both eyes suitable for treatment as minimization covariates (Azuara‐Blanco 2016).

Two studies used appropriate allocation concealment methods (Azuara‐Blanco 2016; Moghimi 2015). It was unclear from the text whether allocation concealment took place in the remaining six studies.

Blinding

Three studies were at low risk of detection bias as they blinded both participants and outcome assessors to the intervention performed (Angmo 2019; Husain 2019; Moghimi 2015).

For the remaining five studies, the inherent differences in the interventions compared meant that it was difficult to mask participants and study personnel, particularly the surgeons performing the intervention. Azuara‐Blanco 2016 took steps to counteract this, undertaking a rigorous standardized masking protocol to minimize detection bias for IOP assessment. One study reported that the study personnel were not masked (Tham 2009), while this was not mentioned in three studies (El Sayed 2019; Tham 2008; Tham 2013).

Incomplete outcome data

We judged three studies to be at low risk of bias because intention‐to‐treat analyses were performed as planned in their study protocol (Azuara‐Blanco 2016; Husain 2019; Tham 2013). This was presumed to be the case for a further two that had no apparent losses to follow‐up (Tham 2008; Tham 2009).

We judged three studies to be at high risk of attrition bias (Angmo 2019; El Sayed 2019; Moghimi 2015). Only data from participants who completed follow‐up were analyzed and presented; missing participants were excluded from the outcomes and baseline data completely, and their reasons for dropping out were not elucidated.

Selective reporting

We deemed three studies to be at low risk of reporting bias as the results of all outcomes described in their protocols were duly reported (Angmo 2019; Azuara‐Blanco 2016; Husain 2019).

One study was at risk of reporting bias ‐ the results of several prespecified outcome measures described in the protocol were not reported (Tham 2013).

The risk was unclear for four studies because we had no protocol to compare the results against (El Sayed 2019; Moghimi 2015; Tham 2008; Tham 2009).

Other potential sources of bias

One study was at unclear risk because it is still ongoing and only the interim report was available (Angmo 2019). We judged the remaining studies to be at low risk of bias.

Effects of interventions

See: Table 1; Table 2; Table 3; Table 4

Meta‐analysis was not possible due to the absence of equivalent intervention groups, relevant data, and follow‐up periods. We therefore presented the results narratively. We show the summary of results for each comparison in Table 5.

1. Summary table.

Intervention	Phacoemulsification	Phacoemulsification	Phacoemulsification	Phacoemulsification	Phacoemulsification
Comparison	Laser peripheral iridotomy (LPI)	Phacoemulsification plus viscogonioplasty	Phacoemulsification plus goniosynechialysis (phaco‐GSL)	Phacoemulsification plus trabeculectomy	Trabeculectomy
Study ID	Azuara‐Blanco 2016	Moghimi 2015	Angmo 2019; Husain 2019	El Sayed 2019 (Data not presented for Tham 2008; Tham 2009)	Tham 2013
Progression of visual field loss	OR 0.49 (95% CI 0.24 to 0.92) at 6 months; OR 0.35 (95% CI 0.13 to 0.91) at 12 months; OR 0.37 (95% CI 0.17 to 0.82) at 24 months; OR 0.42 (95% CI 0.20 to 0.87) at 36 months	NR	NR	Unable to complete analysis to compare between groups	Unable to complete analysis to compare between groups
Mean change in IOP (in mmHg)	MD ‐0.65 mmHg (95% CI ‐3.12 to 1.82) at 6 months; MD ‐0.03 mmHg (95% CI ‐2.34 to 2.32) at 12 months; MD 0.74 mmHg (95% CI ‐1.66 to 3.13) at 24 months; MD 1.04 mmHg (95% CI ‐1.22 to 3.31) at 36 months	MD 0.50 mmHg (95% CI ‐2.64 to 3.64; 91 eyes) at 12 months	MD ‐0.04 mmHg (95% CI ‐0.93 to 0.85; 1 study, 64 eyes) at 6 months; MD ‐0.12 mmHg (95% CI ‐4.72 to 4.48; 1 study, 32 eyes) at 12 months	MD 0.50 mmHg (95% CI ‐1.95 to 2.95) at 1 month; MD 0.00 mmHg (95% CI ‐2.36 to 2.36) at 3 months; MD ‐0.30 mmHg (95% CI ‐1.89 to 1.29) at 6 months; MD ‐0.60 mmHg (95% CI ‐1.99 to 0.79) at 12 months	Unable to complete analysis to compare between groups
Mean change in anterior chamber depth	NR	NR	NR	NR	NR
Number of medications used to control IOP	MD ‐0.65 (95% CI ‐0.85 to ‐0.45) at 6 months; MD ‐0.70 (95% CI ‐0.89 to ‐0.51) at 12 months; MD ‐0.73 (95% CI ‐0.94 to ‐0.52) at 24 months; MD ‐0.65 (95% CI ‐0.87 to ‐0.43) at 36 months	MD ‐0.30 (95% CI ‐0.56 to ‐0.04; 91 eyes) at 6 months; MD ‐0.30 (95% CI ‐0.55 to ‐0.05; 91 eyes) at 12 months	MD ‐0.35 (95% CI ‐0.63 to ‐0.07; 1 study, 64 eyes) at 6 months; MD ‐0.38 (95% CI ‐1.23 to 0.47; 1 study, 32 eyes) at 12 months	MD 0.30 (95% CI 0.01 to 0.59) at 1 month; MD 0.40 (95% CI 0.06 to 0.74) at 3 months; MD 0.20 (95% CI ‐0.15 to 0.55) at 6 months; MD 0.00 (95% CI ‐0.42 to 0.42) at 12 months	Unable to complete analysis to compare between groups
Gonioscopic findings (measured in degrees)	Appositional angle closure MD ‐84.93° (95% CI ‐131.25 to ‐38.61) at 12 months; MD ‐74.16° (95% CI ‐121.34 to ‐26.98) at 36 months Synechial angle closure MD ‐69.63° (95% CI, ‐104.34 to ‐16.92) at 12 months; MD 1.12 (95% CI, ‐42.37 to 44.62) at 36 months	Angle width MD ‐0.60 (95% CI ‐0.91 to ‐0.29; 91 eyes) at 12 months or later; TISA500 MD ‐0.03 (95% CI ‐0.06 to ‐0.01; 91 eyes) at 12 months; TISA750 MD ‐0.03 (95% CI ‐0.06 to ‐0.01; 91 eyes) at 12 months	AOD500 MD ‐0.04 (95% CI ‐0.27 to 0.19; 1 study, 64 eyes); at 6 months; AOD750 MD 0.01 (95% CI ‐0.27 to 0.29; 1 study, 64 eyes) at 6 months; TISA500 MD ‐0.02 (95% CI ‐0.06 to 0.02; 1 study, 64 eyes) at 6 months; TISA750 MD ‐0.03 (95% CI ‐0.17 to 0.11; 1 study, 64 eyes) at 6 months; SSA MD ‐1.59 (95% CI ‐6.75 to 3.57; 1 study, 64 eyes) at 6 months	Unable to complete analysis to compare between groups	Unable to complete analysis to compare between groups
Best corrected visual acuity	MD 2.03 (95% CI ‐0.77 to 4.84) at 12 months; MD 4.30 (95% CI 0.74 to 7.87) at 36 months	MD ‐0.01 (95% CI ‐0.10 to 0.08; 91 eyes) postoperatively	NR	MD ‐0.03 (95% CI ‐0.18 to 0.12; 1 study, 63 eyes)	Unable to complete analysis to compare between groups
Adverse effects	Phacoemulsification group: malignant glaucoma (1 eye); irreversible loss of vision of more than 10 ETDRS letters (1 eye); other related adverse events (1 eye). LPI group: lost ten or more ETDRS letters irreversibly (3 eyes); other related adverse events (10 eyes).		Phaco‐GSL group hyphema (3 eyes)	Less risk of: Intra‐ and post‐operative complications (RR 0.59, 95% CI 0.34 to 1.04); additional IOP‐lowering procedures (RR 5.81, 95% CI 1.41 to 23.88) in the phacoemulsification group	none
Quality of life measures (European Quality of Life‐5 Dimension)	MD 0.03 (95% CI ‐0.16 to 0.22) at 6 months; MD 0.04 (95% CI ‐0.16 to 0.24) at 12 months; MD 0.005 (95% CI ‐0.19 to 0.20) at 24 months; MD ‐0.04 (95% CI ‐0.24 to 0.16) at 36 months	NR	NR	NR	NR

AOD: angle opening distance; CI: confidence interval; IOP: Intraocular pressure; NR: not reported; MD: mean difference; OR: odds ratio; RR: risk ratio; SSA: scleral spur angle; TISA: trabecular iris space area

Phacoemulsification versus laser peripheral iridotomy (LPI)

One study compared phacoemulsification with standard care (LPI) in participants with both PAC and PACG without cataracts (i.e. clear lens extraction) (Azuara‐Blanco 2016). We obtained the raw study data and repeated the analysis for participants with PACG only, using the statistical analysis plan published and provided by the trial investigators. We include a caveat here: the following results should be interpreted with caution because the study was not powered to investigate participants with PACG alone, and we have adjusted the GRADE score accordingly. Of those with PACG, 127 eyes underwent phacoemulsification and 136 had LPI. See Table 1.

Primary outcomes

Progression of visual field loss

The proportion of participants with evidence of progression of visual field loss was lower in the phacoemulsification group than the LPI group at six months (odds ratio [OR] 0.49, 95% confidence interval [CI] 0.24 to 0.92; 216 eyes), 12 months (OR 0.35, 95% CI 0.13 to 0.91; 216 eyes), 24 months (OR 0.37, 95% CI 0.17 to 0.82; 216 eyes), and 36 months (OR 0.42, 95% CI 0.20 to 0.87; 216 eyes). We graded the certainty of the evidence as moderate, having downgraded one level for imprecision (‐1) because the study cohort was not adequately powered to evaluate participants with PACG alone.

Mean change in IOP from baseline to one year

There was no observable difference in change from baseline IOP between the intervention groups at any time point. The mean difference between the groups was ‐0.65 mmHg (95% CI ‐3.12 to 1.82; 257 eyes) at 6 months; ‐0.03 mmHg (95% CI ‐2.34 to 2.32; 257 eyes) at 12 months; 0.74 mmHg (95% CI ‐1.66 to 3.13; 257 eyes) at 24 months; and 1.04 mmHg (95% CI ‐1.22 to 3.31; 257 eyes) at 36 months. We graded the certainty of evidence as moderate for this outcome, downgrading one level for imprecision (‐1) because the study cohort was not adequately powered to evaluate participants with PACG alone.

Secondary outcomes

Mean change in depth of the anterior chamber

Not reported.

Number of medications used to control IOP

The phacoemulsification group required fewer IOP‐lowering medications than the LPI group at 6 months (mean difference [MD] ‐0.65, 95% CI ‐0.85 to ‐0.45; 263 eyes); 12 months (MD ‐0.70, 95% CI ‐0.89 to ‐0.51; 263 eyes); 24 months (MD ‐0.73, 95% CI ‐0.94 to ‐0.52; 263 eyes); and 36 months (MD ‐0.65, 95% CI ‐0.87 to ‐0.43; 263 eyes). We graded the certainty of evidence as moderate for this outcome, downgrading one level for imprecision (‐1) because the study cohort was not adequately powered to evaluate participants with PACG alone.

Gonioscopic findings

Gonioscopic findings favored the phacoemulsification group, which had less appositional angle closure at 12 months (MD ‐84.93°, 95% CI ‐131.25° to ‐38.61°; 106 eyes) and 36 months (MD ‐74.16°, 95% CI ‐121.34° to ‐26.98°; 101 eyes) compared to the LPI group. The phacoemulsification group also had less synechial angle closure (extent of PAS) at 12 months (MD ‐69.63°, 95% CI ‐104.34° to ‐16.92°, 51 eyes) compared to the LPI group, with similar outcomes at 36 months (MD 1.12°, 95% CI ‐42.37° to 44.62°, 52 eyes). We graded the certainty of evidence as moderate for this outcome, downgrading one level for imprecision (‐1) because the study cohort was not adequately powered to evaluate participants with PACG alone.

Visual acuity

Best corrected visual acuity, measured using the Early Treatment Diabetic Retinopathy Study (EDTRS) letter chart, was comparable between the groups at 12 months (MD 2.03, 95% CI ‐0.77 to 4.84; 242 eyes) but favored the LPI group at 36 months (MD 4.30, 95% CI 0.74 to 7.87; 242 eyes). We graded the certainty of evidence as low for this outcome, downgrading by one level because of imprecision (‐1) because the study cohort was not adequately powered to evaluate participants with PACG alone.

Adverse effects

At six months, malignant glaucoma was reported in one eye of one participant in each of the phacoemulsification and LPI groups; irreversible loss of vision of more than ten ETDRS letters was reported in one eye of one participant in the phacoemulsification group; and other related adverse events were reported in five eyes in the phacoemulsification group, and three eyes in the LPI group. The analysis included 118 eyes of 118 participants in the phacoemulsification group, and 132 eyes of 132 participants in the LPI group for all outcomes except malignant glaucoma (117 eyes in 117 participants in the phacoemulsification group and 131 eyes in 131 participants in the LPI group), and other related adverse events (81 eyes in 81 participants in the phacoemulsification group, and 92 eyes in 92 participants in the LPI group).

At 36 months, malignant glaucoma was reported in one eye of one participant in the phacoemulsification group, irreversible loss of vision of more than ten ETDRS letters was reported in one eye of one participant in the phacoemulsification group and three eyes of three participants in the LPI group, and other related adverse events were reported in one eye in the phacoemulsification group and ten eyes in the LPI group. The analysis included 112 eyes of 112 participants in the phacoemulsification group and 121 eyes of 121 participants in the LPI group for all outcomes except other related adverse events (85 eyes of 85 participants in the phacoemulsification group and 88 eyes in 88 participants in the LPI group).

We graded the certainty of evidence as moderate for this outcome, downgrading by one level because of imprecision (‐1).

Quality of life measures

The study used the European Quality of Life‐5 Dimension (EQ‐5D), with three rating scales in five dimensions of health, to assess health status; higher score means better quality of life. There was no evidence of difference between the two groups at 6 months (MD 0.03, 95% CI ‐0.16 to 0.22; 254 eyes), 12 months (MD 0.04, 95% CI ‐0.16 to 0.24; 254 eyes), 24 months (MD 0.005, 95% CI ‐0.19 to 0.20; 254 eyes), or 36 months (MD ‐0.04, 95% CI ‐0.24 to 0.16; 254 eyes). We graded the certainty of evidence as moderate, downgraded by one level because of imprecision (‐1).

Phacoemulsification versus phacoemulsification with viscogonioplasty (phaco‐VGP)

Moghimi 2015 compared phacoemulsification and phaco‐VGP in eyes with PACG and co‐existing cataracts, reporting the results for 46 eyes in the phacoemulsification group and 45 in the phaco‐VGP group for participants who had completed at least 12 months of follow‐up. See Table 2.

Primary outcomes

Progression of visual field loss

Not reported.

Mean change in IOP from baseline to one year

Reduction in IOP was observed in both the phacoemulsification (‐8.3 mmHg ± 6.8 mmHg) and phaco‐VGP (‐8.8 mmHg ± 8.4 mmHg) groups at one year; there was no evidence of difference between groups (MD 0.50 mmHg, 95% CI ‐2.64 to 3.64; 91 eyes; Analysis 1.1). We graded the certainty of evidence as low for this outcome, downgrading by two levels because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

1.1. Analysis.

Comparison 1: Phacoemulsification versus phaco‐viscogonioplasty (phaco‐VGP), Outcome 1: Mean IOP change from baseline

Secondary outcomes

Mean change in depth of the anterior chamber

Not reported.

Number of medications used to control IOP

Fewer IOP‐lowering medications were used in the phacoemulsification group (0.1 ± 0.4) compared to the phaco‐VGP group (0.4 ± 0.8) at 6 months (MD ‐0.30, 95% CI ‐0.56 to ‐0.04; 91 eyes; Analysis 1.2), and 12 months (0.1 ± 0.3 versus 0.4 ± 0.8; MD ‐0.30, 95% CI ‐0.55 to ‐0.05; 91 eyes; Analysis 1.2). We graded the certainty of evidence as low for this outcome, downgrading by two levels because of high risk of attrition bias (‐1) and wide confidence intervals (‐1).

1.2. Analysis.

Comparison 1: Phacoemulsification versus phaco‐viscogonioplasty (phaco‐VGP), Outcome 2: Mean number of medications to control IOP

Gonioscopic findings

There was increased angle width in both the phacoemulsification and phaco‐VGP groups, which was higher in the phaco‐VGP group (1.4 ± 0.8 versus 2.0 ± 0.7; MD ‐0.60, 95% CI ‐0.91 to ‐0.29; 91 eyes; Analysis 1.3). This was measured at the last follow‐up visit (at 12 months or later) which varied for each individual.

1.3. Analysis.

Comparison 1: Phacoemulsification versus phaco‐viscogonioplasty (phaco‐VGP), Outcome 3: Gonioscopic findings

Changes in anterior segment optical coherence tomography (AS‐OCT) parameters were also described: both phacoemulsification and phaco‐VGP groups had changes in the trabecular iris space area (TISA500 [0.020 ± 0.084 versus 0.054 ± 0.033]) and TISA750 (0.084 ± 0.057 versus 0.119 ± 0.061) at 12 months; this increase was higher in the phaco‐VGP group (MD ‐0.03, 95% CI ‐0.06 to ‐0.01 in TISA500, and MD ‐0.03, 95% CI ‐0.06 to ‐0.01 in TISA750; 91 eyes; Analysis 1.3). We graded the certainty of evidence as low for this outcome, downgrading by two levels because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

Visual acuity

The mean logMAR best‐corrected visual acuity (BCVA) was comparable at baseline (0.49 ± 0.28 in the phacoemulsification group versus 0.49 ± 0.35 in the phaco‐VGP group; MD 0.00, 95% CI ‐0.13 to 0.13; 91 eyes), and postoperatively (0.27 ± 0.2 in the phacoemulsification group versus 0.28 ± 0.23 in the phaco‐VGP group; MD ‐0.01, 95% CI ‐0.10 to 0.08; 91 eyes; Analysis 1.4). We graded the certainty of evidence as moderate for this outcome, downgrading by two levels due to high risk of attrition bias (‐1) and imprecision from wide confidence intervals (‐1).

1.4. Analysis.

Comparison 1: Phacoemulsification versus phaco‐viscogonioplasty (phaco‐VGP), Outcome 4: Mean visual acuity

Adverse effects

Three participants in the phaco‐VGP group developed postoperative hyphema that was successfully managed with viscotamponade. Four participants in the phaco‐VGP group and two in the phacoemulsification alone group developed a postoperative fibrin reaction that resolved after a few days. No other adverse effects were reported.

Quality of life measures

Not measured.

Phacoemulsification versus phacoemulsification with goniosynechialysis (phaco‐GSL)

Two studies compared phacoemulsification alone with combined phaco‐GSL (Angmo 2019; Husain 2019). Angmo 2019 comprised 30 eyes in the phacoemulsification group and 34 in the phaco‐GSL group. Husain 2019 included participants with both PAC and PACG with and without a previous history of acute angle closure attacks. We obtained the raw study data from the lead author and repeated the analysis for only participants with PACG without previous acute angle closure attacks. This gave us 16 eyes in the phacoemulsification group and 16 in the phaco‐GSL group. We were unable to conduct any meta‐analyses because the outcomes were measured at different time points in both studies. See Table 3.

Primary outcomes

Progression of visual field loss

Not reported by either study.

Mean change in IOP from baseline to one year

Angmo 2019 did not report the mean change in IOP from baseline. However, there was no evidence of difference in mean IOP at six months between the phacoemulsification and phaco‐GSL groups (13.17 mmHg ± 1.66 mmHg versus 13.21 mmHg ± 1.97 mmHg; MD ‐0.04 mmHg, 95% CI ‐0.93 to 0.85; 64 eyes; Analysis 2.1). We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

2.1. Analysis.

Comparison 2: Phacoemulsification versus phacoemulsification with goniosynechialysis (phaco‐GSL), Outcome 1: Mean IOP change from baseline

In Husain 2019, the mean change in IOP from baseline to 12 months was comparable between the phacoemulsification and phaco‐GSL groups (‐6.59 mmHg ± 6.72 mmHg versus ‐6.47 mmHg ± 6.57 mmHg; MD ‐0.12 mmHg, 95% CI ‐4.72 to 4.48; 32 eyes; Analysis 2.1). We graded the certainty of evidence as moderate for this outcome, downgrading by one level because of imprecision due to wide confidence intervals (‐1).

Secondary outcomes

Mean change in depth of the anterior chamber

Not reported by either study.

Number of medications used to control IOP

In Angmo 2019, the phacoemulsification group required fewer IOP‐lowering medications at six months compared to the phaco‐GSL group (1.70 ± 0.66 versus 2.05 ± 0.46; MD ‐0.35, 95% CI ‐0.63 to ‐0.07; 64 eyes; Analysis 2.2). We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to small sample size (‐1).

2.2. Analysis.

Comparison 2: Phacoemulsification versus phacoemulsification with goniosynechialysis (phaco‐GSL), Outcome 2: Mean number of medications to control IOP

In Husain 2019, the phacoemulsification and phaco‐GSL groups required a comparable number of IOP‐lowering medications at 12 months (0.56 ± 1.09 versus 0.94 ± 1.34; MD ‐0.38, 95% CI ‐1.23 to 0.47; 32 eyes; Analysis 2.2). We graded the certainty of evidence as moderate for this outcome, downgrading by one level because of imprecision due to wide confidence intervals (‐1).

Gonioscopic findings

Angmo 2019 described the iridocorneal angle changes in terms of quantitative anterior segment optical coherence tomography (AS‐OCT) findings at six months postoperatively. While there was an improvement in angle opening for both groups, there was no evidence of difference for the five parameters, measured between the phacoemulsification and phaco‐GSL groups: angle opening distance (AOD)500 (0.48 ± 0.21 versus 0.52 ± 0.65; MD ‐0.04, 95% CI ‐0.27 to 0.19; 64 eyes); AOD750 (0.72 ± 0.3 versus 0.71 ± 0.77; MD 0.01, 95% CI ‐0.27 to 0.29; 64 eyes); trabecular iris space area (TISA)500 (0.15 ± 0.06 versus 0.17 ± 0.10; MD ‐0.02, 95% CI ‐0.06 to 0.02; 64 eyes); TISA750 (0.30 ± 0.12 versus 0.33 ± 0.40; MD ‐0.03, 95% CI ‐0.17 to 0.11; 64 eyes); scleral spur angle (SSA) (31.47 ± 10.84 versus 33.06 ± 10.11; MD ‐1.59, 95% CI ‐6.75 to 3.57; 64 eyes; Analysis 2.3).

2.3. Analysis.

Comparison 2: Phacoemulsification versus phacoemulsification with goniosynechialysis (phaco‐GSL), Outcome 3: Gonioscopic findings

Visual acuity

Not reported by either study.

Adverse effects

Angmo 2019 reported three cases of postoperative hyphema in the phaco‐GSL group. These data were not available for the subgroup in Husain 2019.

Quality of life measures

Not reported by either study.

Phacoemulsification versus combined phaco‐trabeculectomy

Three studies compared phacoemulsification alone with phaco‐trabeculectomy in 63 eyes of 63 participants with chronic PACG (El Sayed 2019), 72 eyes of 72 participants with medically controlled chronic PACG with coexisting cataract (Tham 2008), and 51 eyes of 51 participants with medically uncontrolled chronic PACG with coexisting cataract (Tham 2009). The latter two studies also included participants with previous acute angle‐closure attacks. We requested the data for the subgroup of participants who did not have a history of acute angle‐closure attacks from the trial investigators, but could not incorporate them, because they only provided the data for the phacoemulsification group. This comprised 26 eyes (Tham 2008), and 20 eyes (Tham 2009) in the phacoemulsification groups that did not have a previous acute angle‐closure attack. Participants who did not complete the 12‐month follow‐up were excluded from analysis (number unknown). See Table 4.

Primary outcomes

Progression of visual field loss

One (3.8%) participant (Tham 2008), and 10 (50%) participants (Tham 2009) in the phacoemulsification groups had progression of visual field loss at 24 months. Visual field loss was not reported by El Sayed 2019.

Mean change in IOP from baseline to one year

The mean IOP at 12 months in the phacoemulsification groups was 14.2 mmHg ± 2.8 mmHg in Tham 2008, and 15.8 mmHg ± 3.4 mmHg in Tham 2009.

El Sayed 2019 did not report the mean change in IOP from baseline. The mean IOP between the phacoemulsification and phaco‐trabeculectomy groups were statistically similar at baseline (21.6 mmHg ± 9.2 mmHg in phacoemulsification group versus 25.6 mmHg ± 11.1 mmHg in phaco‐trabeculectomy group; MD ‐4.00 mmHg, 95% CI ‐9.04 to 1.04; 63 eyes). The mean IOP remained statistically comparable at all time points measured postoperatively: at 1 month (14.4 mmHg ± 3.6 mmHg versus 13.9 mmHg ± 6.0 mmHg; MD 0.50 mmHg, 95% CI ‐1.95 to 2.95; 63 eyes), 3 months (13.0 mmHg ± 5.4 mmHg versus 13 mmHg ± 4.1 mmHg; MD 0.00 mmHg, 95% CI ‐2.36 to 2.36; 63 eyes), 6 months (12.4 mmHg ± 2.5 mmHg versus 12.7 mmHg ± 3.8 mmHg; MD ‐0.30 mmHg, 95% CI ‐1.89 to 1.29; 63 eyes), and 12 months (12.6 mmHg ± 2.6 mmHg versus 13.2 mmHg ± 3.0 mmHg; MD ‐0.60 mmHg, 95% CI ‐1.99 to 0.79; 63 eyes; Analysis 3.1). We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

3.1. Analysis.

Comparison 3: Phacoemulsification versus combined phaco‐trabeculectomy, Outcome 1: Mean IOP

Mean change in depth of the anterior chamber (ACD)

The ACD data was not available for the subgroup of participants without previous acute angle‐closure attacks in Tham 2008 and Tham 2009, and was not reported by El Sayed 2019.

Secondary outcomes

Number of medications used to control IOP

In El Sayed 2019, the mean number of IOP‐lowering medications was 2.3 (SD 0.9) in the phacoemulsification group and 2.8 (SD 1) in the phaco‐trabeculectomy group preoperatively. This was lower in the phaco‐trabeculectomy group than the phacoemulsification group at 1 month (0.1 ± 0.2 versus 0.4 ± 0.8; MD 0.30, 95% CI 0.01 to 0.59; 63 eyes), and at 3 months (0.1 ± 0.4 versus 0.5 ± 0.9; MD 0.40, 95% CI 0.06 to 0.74; 63 eyes); but comparable between groups at 6 months (0.5 ± 0.8 in the phacoemulsification group versus 0.3 ± 0.6 in the phaco‐trabeculectomy group; MD 0.20, 95% CI ‐0.15 to 0.55; 63 eyes), and 12 months (0.5 ± 0.8 versus 0.5 ± 0.9; MD 0.00, 95% CI ‐0.42 to 0.42; 63 eyes; Analysis 3.2).

3.2. Analysis.

Comparison 3: Phacoemulsification versus combined phaco‐trabeculectomy, Outcome 2: Mean number of medications to control IOP

The number of IOP‐lowering medications was not available for the subgroup of participants without previous acute angle‐closure attacks in Tham 2008 or Tham 2009.

We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

Gonioscopic findings

Gonioscopic findings were not available for the subgroup of participants without previous acute attacks of angle closure in Tham 2008 and Tham 2009, and were not reported by El Sayed 2019.

Visual acuity

For El Sayed 2019, the best corrected visual acuity was comparable between the phacoemulsification and phaco‐trabeculectomy groups at baseline (0.29 ± 0.3 versus 0.27 ± 0.25; MD 0.02, 95% CI ‐0.12 to 0.16; 63 eyes), and at the final follow‐up visit (variable time scale for each participant) (0.35 ± 0.31 versus 0.38 ± 0.3; MD ‐0.03, 95% CI ‐0.18 to 0.12; 63 eyes; Analysis 3.3).

3.3. Analysis.

Comparison 3: Phacoemulsification versus combined phaco‐trabeculectomy, Outcome 3: Mean visual acuity

Visual acuity data were not available for the subgroup of participants without previous acute angle‐closure attacks in Tham 2008 or Tham 2009.

We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

Adverse effects

El Sayed 2019 reported that the phacoemulsification group experienced fewer complications than the phaco‐trabeculectomy group overall (risk ratio [RR] 0.59, 95% CI 0.34 to 1.04; 63 eyes; Analysis 3.4). Phacoemulsification resulted in one case of intraoperative posterior capsular tear with vitreous loss, nine of corneal edema, and one of pupillary membrane; whereas, phaco‐trabeculectomy resulted in 18 complications, including intraoperative hyphema (1), corneal edema (6), corneal ulcer (1), hypotony (3), choroidal detachment (1), conjunctival retraction (1), shallow anterior chamber (2), pupillary membrane (2), and wipe‐out (1). In addition, more eyes in the phaco‐trabeculectomy group required further IOP‐lowering procedures (eleven needed one or more needle bleb revisions, and one underwent cyclodiode laser) than the phacoemulsification group (two eyes required trabeculectomy) (RR 5.81, 95% CI 1.41 to 23.88; 63 eyes; Analysis 3.4).

3.4. Analysis.

Comparison 3: Phacoemulsification versus combined phaco‐trabeculectomy, Outcome 4: Adverse effects

Two of 26 eyes in Tham 2008 and two of 20 eyes in Tham 2009 had intraoperative complications in the phacoemulsification group. They did not observe any postoperative complications.

We graded the certainty of evidence as low for this outcome, downgrading by one level each because of high risk of attrition bias (‐1) and imprecision due to wide confidence intervals (‐1).

Quality of life measures

Not reported by any study.

Phacoemulsification versus trabeculectomy

Tham 2013 compared phacoemulsification with trabeculectomy in participants with medically uncontrolled chronic angle‐closure glaucoma without cataract. We obtained the data for a subgroup of participants without history of acute angle‐closure attacks from the trial investigators. Of 26 participants originally randomized to phacoemulsification, 16 did not have previous acute attacks. The data in the trabeculectomy group were not provided, and as such, we were unable to analyze the between‐group differences.

Discussion

Summary of main results

We identified eight randomized controlled trials (RCT) comprising 513 eyes (total number of participants unknown) that addressed the utility of lens extraction in treating chronic angle‐closure glaucoma. The participants were recruited from a diverse range of countries and were followed up for a minimum of 12 months.

One RCT compared phacoemulsification with laser peripheral iridotomy (LPI) in chronic primary angle‐closure glaucoma (PACG) with clear crystalline lenses. A smaller proportion of participants who had phacoemulsification experienced progression of visual field loss than those who had LPI at 12 months. There was greater widening of the iridocorneal angle in participants in the phacoemulsification group compared with those in the LPI group, but no evidence of difference in intraocular pressure (IOP) reduction at all time points up to 3 years. However, the phacoemulsification group required fewer IOP‐lowering medications to achieve this reduction. There was no evidence of difference between phacoemulsification and LPI for health‐related quality of life, or for adverse events in the hands of experienced surgeons. All these findings, together with moderate certainty of evidence, suggest a clinical benefit to phacoemulsification over LPI. It is also important to note that this study was not powered to look only at those with PACG and this is a sub‐analysis; lack of significance for some findings could be perhaps be attributed to the small sample size. In addition, these results are applicable over a three‐year time period, and longer term comparisons are not yet known.

Only one study compared phacoemulsification with trabeculectomy, but we did not have the data to analyze and draw conclusions.

Three studies compared phacoemulsification with phacoemulsification and trabeculectomy. There were limited data on progression of visual field loss, change in ACD, or iridocorneal angle within these studies. There was no evidence of difference between groups for IOP, or the number of medications used to control IOP. However, there was evidence that phacoemulsification was associated with a reduced risk of complications compared with phacoemulsification plus trabeculectomy in one trial (El Sayed 2019) (low certainty of evidence).

The remaining three studies compared phacoemulsification alone versus phacoemulsification (phaco) plus supplemental interventions (viscogonioplasty [VGP] or goniosynechialysis [GSL]). All these trials included 6‐month or 12‐month follow‐up. There were no data on progression of visual field loss in these trials. Moderate to low certainty evidence suggest that there may be little or no evidence of additional benefit with the supplemental procedures compared with phacoemulsification alone for reducing IOP or medications to control IOP, increasing width of the iridocorneal angle, or visual acuity. While more complications were reported among participants in the phacoemulsification plus supplemental procedure groups than in the phacoemulsification groups, there were insufficient data to draw conclusions with certainty. Until further evidence becomes available, these supplemental procedures may be left to the discretion of the surgeon or local protocols.

Overall completeness and applicability of evidence

Overall, the review possesses good external validity. The definition of chronic PACG used in each study was generally similar across the board. The overall study population appropriately reflected the real‐world distribution of PACG. In addition, participants included those with medically controlled and uncontrolled PACG, which is similarly a representative microcosm of the wider population with PACG. The eight studies reported on chronic PACG patients with cataractous as well as clear lenses, providing a broad perspective on the utility of lens extraction in PACG. The studies encompassed a wide range of clinically relevant interventions, including one that compared lens extraction against standard care, enabling the evaluation of lens extraction's true effectiveness as a treatment modality. The majority of the included studies assessed both IOP measurements and IOP‐lowering medications, which are clinically important in terms of arresting glaucoma progression. However, the ultimate goal of PACG treatment is also to prevent additional damage to the optic nerve in order to slow or arrest visual field loss, and IOP alone may not be sufficient to assess this. Even so, few studies included progression of visual field loss as an outcome measure. In addition, the impact of the interventions on participants' health‐related quality of life was not always sought. Future studies will want to consider assessing the impact of interventions on both visual field and quality of life.

The potential differences between RCT and real world outcomes is also worth noting. Compared to a routine phacoemulsification case, eyes with PACG tend to have anatomical differences including shallower anterior chambers, shorter axial lengths, and thicker and more anteriorly positioned crystalline lenses, among others. Phacoemulsification in PACG eyes may therefore be a more challenging endeavour, thereby increasing the innate risks of intraoperative complications. While it is reassuring that the complication rates in the included studies were rates were modest, an important caveat to note is that these surgeries were invariably performed by highly experienced surgeons, whereas this may not necessarily hold true in a real world setting.

Finally, where outcomes of interest were inadequately reported, we wrote to the study authors to seek further clarification, and thereby, ensure that the review was as complete as possible. However, we were regrettably unable to obtain comparative data for three studies, and complete safety data for a few others (Tham 2008; Tham 2009; Tham 2013).

Quality of the evidence

The quality of evidence varied among the included studies. We graded the certainty of evidence to be moderate or low due to risk of bias and imprecision related to the small number of participants included in each comparison of interventions. Although randomization was adequate and there were minimal conflicts of interest across the board, the remaining risk of bias domains were problematic, primarily due to poor adherence to the CONSORT guidelines (Moher 2010). Allocation concealment and masking were generally poorly reported. In addition, loss to follow‐up was inadequately documented, and three studies lacked intention‐to‐treat analyses, instead excluding missing patients entirely. For the comparison of phacoemulsification and phaco‐VGP or phaco‐GSL in particular, variability in outcome reporting also hampered the synthesis of information. All these reduce confidence in the validity of the findings, which should therefore be interpreted with caution.

Potential biases in the review process

The review team used standard Cochrane methodology, including a comprehensive search strategy designed by experienced information specialists. Two review authors, including at least one clinician and one methodologist, independently assessed the search results for eligibility and extracted data. In addition, we reached out to the authors to seek clarification on the eligibility of the study for inclusion, as well as points of contention. As there were no language restrictions, multiple publications available only in Chinese were carefully reviewed in the screening process, which was important because a fair proportion of people with PACG reside in Chinese‐speaking regions. As such, we can confidently conclude that we have identified most, if not all, pertinent results in this review. However, it is important to note that the results should be interpreted with caution, as we obtained data for participants meeting our inclusion criteria only (chronic PACG, no previous acute angle‐closure attacks) to conduct several analyses, meaning that the data might not be adequately powered to answer the clinical question posed.

One of the review authors (DSF) co‐authored one study included in this review (Azuara‐Blanco 2016). However, two other review authors (AYO and SMN) independently assessed the study eligibility, extracted the data, and judged risk of bias following Cochrane standard methodology, and a third (VSS) performed the statistical analyses for this study. We consider the risk of bias induced by this association to be minimal.

Agreements and disagreements with other studies or reviews

This review demonstrates evidence for the benefits of lens extraction in terms of improving IOP control, increasing ACD, and reducing the number of IOP‐lowering medications required in people with chronic PACG. This is consistent with findings from previous non‐comparative studies, which demonstrated similar IOP reduction and increased ACD after lens extraction in participants with PACG, primary angle closure (PAC), or even normal non‐glaucomatous eyes (Atalay 2017; Kashiwagi 2006; Siak 2016). Masis 2017 published a systematic meta‐analysis of the final change in IOP following phacoemulsification in 597 eyes with PACG. They reported an absolute change of ‐6.4 mmHg (95% confidence interval [CI] ‐9.4 to ‐3.4; IV = 99%) at the final follow‐up visit. While synthesizing these results to determine the absolute effect of phacoemulsification was out of the scope of this review, it is worth mentioning that Masis 2017 included a mix of retrospective and prospective non‐randomized studies in addition to RCTs. Further, participants with previous acute angle‐closure attacks were eligible for inclusion in the review, and IOP was measured at the final follow‐up visit, which would vary between participants, as well as across each study. While this lends credence to the hypothesis that the lens may indeed play an important role in the pathophysiology of chronic PACG, it is possible that the true effect may differ from that described.

The role of goniosynechialysis is also worth discussing further. This review suggests that there may be little or no evidence of additional benefit with of supplemental goniosynechialysis or viscogonioplasty compared with phacoemulsification alone for reducing IOP or medications to control IOP, increasing width of the iridocorneal angle, or visual acuity. Ahmed 2019 reviewed a range of studies, including both comparative and non‐comparative studies, as well as participants with both PAC and PACG. Ahmed 2019 highlights the confounding effect that phacoemulsification may have in itself releasing peripheral anterior synechiae (PAS), the fact that a successful outcome depends not simply on performing the procedure, but also on the amount of residual PAS, and finally, the lack of sufficient intraoperative gonioscopy in establishing an accurate baseline and outcome in terms of PAS. These are all important considerations that should be considered in future studies, before dismissing goniosynechialysis or viscogonioplasty out of hand.

In their systematic review and meta‐analysis comparing phacoemulsification, trabeculectomy, and phaco‐trabeculectomy in the treatment of 1495 eyes with PACG, Deng 2011 concluded that phacotrabeculectomy was superior to both trabeculectomy and phacoemulsification in reducing IOP. Wang 2016 published a separate review and meta‐analysis comparing phacoemulsification and phaco‐trabeculectomy in 468 eyes, which concluded that phacoemulsification alone had a better effect on IOP reduction than phaco‐trabeculectomy. Interestingly, the studies included in Wang 2016 overlapped with that of Deng 2011. Our findings differ from the conclusions from both reviews – while we were unable to draw any meaningful conclusions regarding phacoemulsification
and trabeculectomy, we found some evidence to suggest that there was no significant benefit for phaco‐trabeculectomy over phacoemulsification alone, with the former more likely to produce more complications and result in further interventions. There are several methodological differences worth noting here: both Deng 2011 and Wang 2016 included non‐randomized retrospective studies, thereby increasing the potential for bias, and result in a less robust estimate of effect (Reeves 2019). Finally, neither review downgraded for risk of bias in the included studies.

Authors' conclusions

Implications for practice.

Moderate quality evidence suggests that lens extraction is preferable to standard care in treating patients with chronic primary angle closure glaucoma (PACG) and clear crystalline lenses. These findings are potentially practice‐changing, but should ultimately be part of a shared decision‐making process between the clinician and the patient, taking into account their preferences, health status, and visual prognosis.

For patients with visually significant cataracts, low quality evidence suggests that combining phacoemulsification with either goniosynechialysis (GSL) or viscogonioplasty (VGP) does not appear to confer any additional benefit over phacoemulsification alone.

We found insufficient evidence to draw any meaningful conclusions regarding phacoemulsification versus trabeculectomy. In addition, low‐quality evidence suggests that combining phacoemulsification with trabeculectomy does not confer any additional clinical benefit over phacoemulsification alone, and may perhaps lead to more complications. This suggests that combined phaco‐trabeculectomy may not be the best course of action for chronic PACG in the first instance. Whether a two‐step procedure is beneficial and when to undertake this may be of interest, but is not within the scope of the review.

Implications for research.

Studies with longer follow‐up periods (over five years) will help to determine whether lens extraction confers a significant benefit in the long term. Since those with a history of acute angle‐closure attacks may be substantially different from those with asymptomatic disease, we suggest that stratification by this variable should be adopted in studies including both types of participants. Though most forms of therapy for glaucoma are aimed at controlling intraocular pressure, adequate data on visual field changes using validated methods and validated definitions, visual acuity, vision‐related quality of life, and economic outcomes will help to evaluate different interventions for this condition comprehensively.

What's new

Date	Event	Description
11 December 2019	New citation required and conclusions have changed	Updated search: eight new studies (Angmo 2019; Azuara‐Blanco 2016; El Sayed 2019; Husain 2019; Moghimi 2015; Tham 2008; Tham 2009; Tham 2013) included
11 December 2019	New search has been performed	Ariel Y Ong and Sueko M Ng were added as authors

History

Protocol first published: Issue 4, 2005
Review first published: Issue 3, 2006

Date	Event	Description
17 October 2008	Amended	Converted to new review format.
1 March 2006	New citation required and conclusions have changed	Substantive amendment

Appendices

Appendix 1. CENTRAL search strategy

#1    MeSH descriptor: [Glaucoma, Angle‐Closure] this term only
#2    angle* near/3 (occlud* or narrow* or width or close* or closure)
#3    glaucoma* near/3 (occlud* or narrow* or width or close* or closure)
#4    PAC or PACS or PACG or ACG
#5    #1 or #2 or #3 or #4
#6    MeSH descriptor: [Cataract] explode all trees
#7    MeSH descriptor: [Cataract Extraction] explode all trees
#8    MeSH descriptor: [Capsulorhexis] this term only
#9    MeSH descriptor: [Phacoemulsification] this term only
#10    pha?oemulsif*
#11    MeSH descriptor: [Aphakia, Postcataract] this term only
#12    MeSH descriptor: [Lenses, Intraocular] explode all trees
#13    MeSH descriptor: [Lens Implantation, Intraocular] this term only
#14    intraocular len* or IOL*
#15    lensectom* or capsulotom*
#16    cataract* or lens or lenses
#17    #6 or #7 or #8 o r#9 or #10 or #11 or #12 or #13 or #14 or #15 or #16
#18    #5 and #17

Appendix 2. MEDLINE Ovid search strategy

1. randomized controlled trial.pt.     
2. (randomized or randomised).ab,ti.     
3. placebo.ab,ti.     
4. dt.fs.     
5. randomly.ab,ti.     
6. trial.ab,ti.     
7. groups.ab,ti.     
8. or/1‐7     
9. exp animals/     
10. exp humans/     
11. 9 not (9 and 10)     
12. 8 not 11     
13. Glaucoma, Angle‐Closure/     
14. (angle$ adj3 (occlud$ or narrow$ or width or close$ or closure)).tw.     
15. (glaucoma$ adj3 (occlud$ or narrow$ or width or close$ or closure)).tw.     
16. (PAC or PACS or PACG or ACG).tw.     
17. or/13‐16     
18. exp cataract/     
19. exp cataract extraction/     
20. capsulorhexis/     
21. capsulorhexis.tw.     
22. phacoemulsification/     
23. pha?oemulsif$.tw.     
24. aphakia, postcataract/     
25. exp lenses intraocular/     
26. lens implantation intraocular/     
27. (intraocular lens or IOL$).tw.     
28. (lensectom$ or capsulotom$).tw.     
29. (cataract$ or lens or lense).tw.     
30. or/18‐29     
31. 17 and 30     
32. 12 and 31

The search filter for trials at the beginning of the MEDLINE strategy is from the published paper by Glanville  2006.

Appendix 3. Embase Ovid search strategy

1. exp randomized controlled trial/     
2. exp randomization/     
3. exp double blind procedure/     
4. exp single blind procedure/     
5. random$.tw.     
6. or/1‐5     
7. (animal or animal experiment).sh.     
8. human.sh.     
9. 7 and 8     
10. 7 not 9     
11. 6 not 10     
12. exp clinical trial/     
13. (clin$ adj3 trial$).tw.     
14. ((singl$ or doubl$ or trebl$ or tripl$) adj3 (blind$ or mask$)).tw.     
15. exp placebo/     
16. placebo$.tw.     
17. random$.tw.     
18. exp experimental design/     
19. exp crossover procedure/     
20. exp control group/     
21. exp latin square design/     
22. or/12‐21     
23. 22 not 10     
24. 23 not 11     
25. exp comparative study/     
26. exp evaluation/     
27. exp prospective study/     
28. (control$ or prospectiv$ or volunteer$).tw.     
29. or/25‐28     
30. 29 not 10     
31. 30 not (11 or 23)     
32. 11 or 24 or 31     
33. closed angle glaucoma/     
34. glaucomatous optic neuropathy/     
35. neovascular glaucoma/     
36. secondary glaucoma/     
37. (angle$ adj3 (occlud$ or narrow$ or width or close$ or closure)).tw.     
38. (glaucoma$ adj3 (occlud$ or narrow$ or width or close$ or closure)).tw.     
39. (PAC or PACS or PACG or ACG).tw.     
40. or/33‐39     
41. exp cataract/     
42. exp cataract extraction/     
43. capsulorhexis/     
44. capsulorhexis.tw.     
45. phacoemulsification/     
46. pha?oemulsif$.tw.     
47. aphakia/     
48. lens implant/     
49. exp lens implantation/     
50. (intraocular lens or IOL$).tw.     
51. (lensectom$ or capsulotom$).tw.     
52. (cataract$ or lens or lenses).tw.     
53. or/41‐52     
54. 40 and 53     
55. 32 and 54

Appendix 4. LILACS search strategy

glaucoma OR angle [Words] and occlud$ OR narrow$ OR width OR close$ OR closure [Words] and cataract OR lens OR lenses OR capsulorhexis OR phacoemulsification OR intraocular lens OR IOL OR lensectomy OR capsulotomy [Words]

Appendix 5. ClinicalTrials.gov search strategy

 (cataract OR lens OR IOL OR phacoemulsification OR phacotrabeculectomy) | Glaucoma, Angle‐Closure

Appendix 6. WHO ICTRP search strategy

CONDITION = angle closure glaucoma AND INTERVENTION = cataract OR lens OR IOL OR phacoemulsification OR phacotrabeculectomy

Data and analyses

Comparison 1. Phacoemulsification versus phaco‐viscogonioplasty (phaco‐VGP).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1.1 Mean IOP change from baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.1.1 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.2 Mean number of medications to control IOP	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.2.1 at 6 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.2.2 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.3 Gonioscopic findings	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.3.1 Change of angle grading at 12 months or later	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.3.2 TISA500	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.3.3 TISA750	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.4 Mean visual acuity	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.4.1 at baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
1.4.2 postoperatively	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Comparison 2. Phacoemulsification versus phacoemulsification with goniosynechialysis (phaco‐GSL).

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
2.1 Mean IOP change from baseline	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1.1 at 6 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.1.2 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2 Mean number of medications to control IOP	2		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2.1 at 6 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.2.2 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3 Gonioscopic findings	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.1 AOD500	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.2 AOD750	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.3 TISA500	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.4 TISA750	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
2.3.5 SSA	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected

Comparison 3. Phacoemulsification versus combined phaco‐trabeculectomy.

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
3.1 Mean IOP	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1.1 at baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1.2 at 1 month	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1.3 at 3 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1.4 at 6 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.1.5 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2 Mean number of medications to control IOP	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2.1 at baseline	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2.2 at 1 month	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2.3 at 3 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2.4 at 6 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.2.5 at 12 months	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.3 Mean visual acuity	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.3.1 preoperatively	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.3.2 at the final follow‐up	1		Mean Difference (IV, Fixed, 95% CI)	Totals not selected
3.4 Adverse effects	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
3.4.1 Intra‐ and post‐operative complications	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected
3.4.2 Additional IOP‐lowering procedures	1		Risk Ratio (M‐H, Fixed, 95% CI)	Totals not selected

Characteristics of studies

Characteristics of included studies [ordered by study ID]

Angmo 2019.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): 80 eyes in total; 40 eyes to phaco, 40 eyes to phaco‐GSL Exclusions after randomization: not reported Number analyzed (total and per group): 30 eyes of 30 participants in phaco group and 34 eyes of 34 participants in phaco‐GSL group Unit of analysis: eyes Losses to follow up: 10 in phaco group and 6 in phaco‐GSL group (numbers lost to follow‐up or excluded not distinguished) How was missing data handled?: N/A Reported power calculation: Y (power 80%) "By assuming that lens extraction with goniosynechialysis reduces IOP further by 2 mmHg at 1‐year follow‐up, 65 eyes are required per group to detect the difference in the reduction of IOP between the 2 groups to be statistically significant in the Student t test with a 5% alpha error, 80% power, and a common standard deviation of 4 mmHg. By allowing 20% for losses in the follow‐up, 80 eyes per group are needed." Unusual study design?: Unclear whether single or both eyes of participant was/were included
Participants	Country: India Age: phaco group 58.77 (± 8.14) years; phaco‐GSL 56.5 (±9.17) years Gender (percent): not reported Inclusion criteria: phakic eyes of people who were aged more than 45 years with a visually significant cataract (visual acuity < 6/12), who had mild‐moderate PACG, who received laser peripheral iridotomy at least 8 weeks before, and who had IOP > 21 mmHg on maximally tolerable medical therapy. Primary angle‐closure glaucoma was defined as occludable angles with PAS extending 180 degrees, along with IOP > 21 mmHg on > 3 occasions in association with optic nerve cupping > 0.6 and corresponding glaucomatous visual field changes. Exclusion criteria: patients with advanced glaucoma in the potentially eligible eye as determined by (1) visual field loss (mean deviation worse than12 decibels) or (2) cup‐disc ratio 0.9. Patients with an acute attack of angle‐closure glaucoma, defined as a sudden elevation in the IOP with associated decreased vision, pain, corneal edema, and a mid‐dilated pupil; primary angle‐closure suspects; 1‐eyed patients; patients with corneal opacity precluding anterior chamber evaluation; patients who had any prior ocular/glaucoma surgery; patients who were not receiving regular follow‐up; patients with unreliable visual fields (fixation loss > 20%, false‐positive, and false‐negative >15%) Equivalence of baseline characteristics: none; baseline characteristics comparable (p > 0.05 for all reported characteristics)
Interventions	Intervention 1: phacoemulsification (phaco) alone Intervention 2: phacoemulsification combined with goniosynechialysis (phaco‐GSL) Length of follow‐up: Planned: 2 years Actual: 6 months (interim report)
Outcomes	Primary outcome, as defined in study reports: IOP reduction at each follow‐up time point Secondary outcomes, as defined in study reports: anterior chamber angle parameters, namely, AOD 500 and 750, TISA 500 and 750, SSA, visual acuity, and number of glaucoma medications postoperatively Adverse events reported: hyphema in the early postoperative period occurred in 3 participants after phaco‐GSL, which spontaneously resolved. No other complications were noted in either group Intervals at which outcomes assessed: 1 week, 1 month, 3 months, and 6 months
Notes	Type of study: published Funding sources: Council of Scientific and Industrial Research, India, Ministry of Science and Technology, Government of India Disclosures of interest: none Study period: June 2015 to June 2017 Reported subgroup analyses: no Trial registration: CTRI/2015/05/005742
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Eighty eyes were assigned randomly into 2 parallel groups and underwent phaco alone in group 1 or phaco‐GSL in group 2, using a computer‐generated random number table using block randomization with variable block size."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) Outcome assessors	Low risk	Performance bias: unable to mask participants and personnel due to nature of interventions Detection bias: "All participants were blinded, and the investigators measuring IOP and anterior segment OCT were also blinded."
Incomplete outcome data (attrition bias) All outcomes	High risk	10/40 (25%) in phaco group and 6/40 (15%) in phaco‐GSL group were excluded from analysis due to loss to follow‐up or incomplete records rather than using intention‐to‐treat analysis.
Selective reporting (reporting bias)	High risk	Protocol stated that gonioscopy would be performed to assess degree of angle closure ‐ not reported. Number of IOP medications reported was not one of the original outcomes in the protocol either. Note that this is an interim analysis of the trial, which is still ongoing
Other bias	Unclear risk	Interim analysis ‐ trial still ongoing; early reporting ‐ potentially selective reporting

Azuara‐Blanco 2016.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): 263 eyes of 263 participants with PACG: 136 to clear lens extraction, 127 to laser peripheral iridotomy Exclusions after randomization: not reported Number analyzed (total and per group)(intention‐to‐treat analysis): 408 eyes of 408 participants in total; 204 in clear‐lens extraction, 205 in laser peripheral iridotomy Unit of analysis: participants Losses to follow up: data not available for 4 participants in clear‐lens extraction group, and 6 in laser peripheral iridotomy group How was missing data handled?: intention to treat analysis was followed for all the main outcomes, and participants with at least one observation were included in the analyses; a score was generated if there were missing questions in the NEI‐VFQ‐25 Reported power calculation: Y (power 90%) "We calculated that with 170 participants in each group, the study would have 90% power at 5% significance level to detect a difference in mean EQ‐5D score of 0.35 SD, which represents an absolute change in score of 0.05 and is likely to be clinically important." Unusual study design?: none
Participants	Country: Australia; China (Mainland China and Hong Kong); Malaysia; Singapore; UK Age: 67.0 (61.0 to 73.0) years in clear‐lens extraction group, and 67.0 (61.0 to 73.0) in laser peripheral iridotomy group Gender (percent): 122 (59%) female in clear‐lens extraction group, and 121 (57%) in laser peripheral iridotomy group Inclusion criteria: phakic aged 50 years or older (to limit the effect of loss of accommodation associated with clear‐lens extraction) with newly diagnosed primary angle closure with intraocular pressure 30 mmHg or greater or primary angle‐closure glaucoma. Primary angle closure was defined as iridotrabecular contact, either appositional or synechial, of at least 180° on gonioscopy, and primary angle‐closure glaucoma as reproducible glaucomatous visual field defects, glaucomatous optic neuropathy, or both, and intraocular pressure greater than 21 mmHg on at least one occasion. Exclusion criteria: patients with symptomatic cataract, advanced glaucoma (mean deviation worse than –15 dB or cup‐to‐disc ratio ≥ 0.9), or previous acute angle‐closure attack, or who had undergone previous laser or ocular surgery Equivalence of baseline characteristics: "Baseline characteristics were similar in the two treatment groups (table 1)" ‐ but no statistical analysis comparing baseline characteristics."
Interventions	Intervention 1: clear‐lens extraction Intervention 2: laser peripheral iridotomy Length of follow‐up: 36 months
Outcomes	Primary outcome, as defined in study reports: health status (using the EQ‐5D); IOP at 3 year final assessment; incremental cost per quality adjusted life year (QALY) gained with QALYs based on the responses to the EQ‐5D Secondary outcomes, as defined in study reports: 1. Patient‐centered: GPI, a glaucoma‐specific utility instrument; NEI‐VFQ‐25, a vision‐specific health profile measure; 2. Clinical: need for glaucoma surgery (e.g. trabeculectomy); best corrected visual acuity (ETDRS); progressive visual field loss; extension of angle closure (degrees of appositional and synechial angle closure) determined clinically; escalation of therapy; opening of anterior chamber angle; number of anti‐glaucoma medications; intolerance of medications; annual incidence of acute attacks of angle closure; 3. Economic: costs to the NHS and participants; cost‐utility analysis; incremental cost per QALY based on the response to the GPI; cost‐effectiveness analysis (incremental cost per case of glaucoma surgery avoided); 4. Safety: complications during or after cataract surgery; loss of best corrected visual acuity > 10 ETDRS letters Adverse events reported: "There were no serious adverse events. 75 participants (25 in the clear‐lens extraction group and 50 in the standard care group) had at least one complication. Posterior capsule rupture was seen after two clear‐lens extractions (1%). No severe complications were reported as a direct consequence of laser iridotomy. Irreversible loss of vision of more than ten ETDRS letters was seen in one participant in the clear‐lens extraction group and three in the standard care group. Intolerance of medications was reported less frequently in the clear‐lens extraction group than in the standard group (three vs ten participants, difference 3.3%, 95% CI 0.004 to 6.6; P = 0.049). Further intraocular surgery was needed to manage complications of the primary or additional interventions in three participants (zonulo‐hyaloido‐vitrectomy for malignant glaucoma, repositioning of a subluxated intraocular lens, and injection of antibody against VEGF for macular edema) in the clear‐lens extraction group and one participant (pars plana vitrectomy for dislocated lens) in the standard care group. Also in the latter group, 12 (6%) participants underwent surgery for clinically relevant cataracts. One participant in the clear‐lens extraction group developed transient corneal edema and another suffered malignant glaucoma. Central corneal thickness did not differ between groups." Intervals at which outcomes assessed: baseline, 6 months, 12 months, 24 months, and 36 months
Notes	Type of study: published Funding sources: Medical Research Council (ref G0701604) Disclosures of interest: "We declare no competing interests." Study period: January 2009 to December 2011 Reported subgroup analyses: Yes Trial registration: ISRCTN44464607
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The randomisation schedule was created with a web‐based application (CHaRT, http://viis.abdn.ac.uk/HSRU/chart) at the Center for Healthcare Randomized Trials, University of Aberdeen, Aberdeen, UK. The randomisation algorithm used sex, center, ethnic origin (Chinese or non‐Chinese), diagnosis, and one or both eyes suitable for treatment as minimisation covariates."
Allocation concealment (selection bias)	Low risk	See above
Blinding (performance bias and detection bias) Outcome assessors	Low risk	Detection bias: "Intraocular pressure was taken to be the average of two readings by Goldmann tonometry. Two observers at each site, following a masking protocol, were involved in the measurements. One observer randomly set the starting force and recorded the pressure values obtained by the other observer, who interacted directly with the patient but did not look at the results on the measurement dial." Performance bias: unable to blind participants to results due to nature of interventions.
Incomplete outcome data (attrition bias) All outcomes	Low risk	"All the main analyses were based on intention‐to‐treat (ITT) and were done at the end of the trial."; 4 (1.9%) in clear‐lens extraction and 6 (2.8%) in peripheral iridotomy who were randomized were not included in intention‐to‐treat analysis"
Selective reporting (reporting bias)	Low risk	All prespecified outcomes in the protocol were reported in the final report
Other bias	Low risk	None identified.

El Sayed 2019.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): 63 in total: 32 to phacoemulsification, 31 to phaco‐trabeculectomy Exclusions after randomization: participants who did not complete 12‐month follow‐up were excluded, but the number of those participants was not reported Number analyzed (total and per group): 63 in total: 32 in phacoemulsification, 31 in phaco‐trabeculectomy Unit of analysis: participants (one eye per participant was enrolled) Losses to follow up: not reported How was missing data handled?: excluded from analysis Reported power calculation: yes (power of 80%) "Estimation of sample size was performed considering a study power of at least 0.8 with an α error of 0.05 aiming to detect a difference of 4 mmHg in mean IOP in the postoperative month 12 between the 2 groups, assuming a SD of 5 mmHg. On the basis of this estimation, a total of 52 eyes were found to be adequate and considering a total of 20% dropout during the follow‐up, recruitment of at least 62 study subjects was targeted." Unusual study design? participants with incomplete follow‐up (defined as less than 12 months) were excluded
Participants	Country: Egypt Age: phacoemulsification 58.8 (± 8.4) years; phaco‐trabeculectomy 57.3 (± 8.3) years Gender (percent): 16 (50%) female in the phacoemulsification group; 11(35.5%) female in the phaco‐trabeculectomy group Included criteria: PACG diagnosed when the following criteria were present: (1) pre‐treatment IOP > 21 mmHg, (2) glaucomatous optic neuropathy, and (3) gonioscopy revealing >180° of iridotrabecular contact Excluded criteria: (1) eyes with retinal pathologies, (2) corneal opacities impeding visualization during phacoemulsification, (3) previous intraocular surgery except for laser iridotomy, (4) previous acute attack of angle closure, and (5) participants who did not complete 12 months of follow‐up Equivalence of baseline characteristics: Yes. No statistically significant difference between the two treatment groups in terms of age, sex, diabetic status, lens clarity, laterality, vertical CDR, automated perimetry MD, extent of PAS, preoperative BCVA, number of medications, or pre‐operative IOP.
Interventions	Intervention1: phacoemulsification Inventention 2: phacoemulsification with trabeculectomy Length of follow‐up: 12 months
Outcomes	Primary outcome, as defined in study reports: success rates, defined as IOP of 6 to 21 mmHg with no other signs of glaucoma progression (defined as structural progression where the vertical CDR increases by 0.1 or more or functional progression wherein visual field testing shows worsening of 2 or more points within or adjacent to a previously existing scotoma by 10 dB or more); Complications Secondary outcomes, as defined in study reports: mean IOP, mean BCVA, and mean number of glaucoma medications at various follow‐up time points Adverse events reported: "In the phaco‐trabeculectomy group, 25 eyes (81%) required removal of 1 or more releasable sutures to titrate the IOP postoperatively. In total, 11 eyes (35%) required 1 or more needle bleb revisions, with 5‐FU. One eye required diode laser cyclophotocoagulation to reduce the IOP. Three eyes (13%) in the phaco‐trabeculectomy group developed hypotony (IOP < 6 mmHg over 2 consecutive visits). One was caused by conjunctival retraction and required conjunctival repair and 2 were caused by overfiltration after removal of releasable sutures, of which 1 case resolved spontaneously and 1 required anterior chamber reformation by healon. One eye developed a corneal ulcer after needling with 5‐FU, which resolved with conservative treatment. In 1 participant with advanced glaucoma, the BCVA deteriorated from 0.01 to hand motion 3 months postoperatively although the IOP was < 14 mmHg at all follow‐up visits. In the phacoemulsification group, 2 eyes (6%) required trabeculectomy to control the IOP, after 4 and 9 months." Intervals at which outcomes assessed: postoperative day 1, months 1, 3, 6, 9, and 12
Notes	Type of study: published Funding sources: not reported Disclosures of interest: "The authors declare no conflict of interest" Study period: not reported Reported subgroup analyses: no Trial registration: not identified
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Randomization was done using a random number table."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) Outcome assessors	Unclear risk	Masking of participants and personnel was not described. It is difficult to blind personnel due to inherent difference in procedures and post‐operative medication regimes.
Incomplete outcome data (attrition bias) All outcomes	High risk	"We excluded eyes.... patients who did not complete 12 months of follow‐up."
Selective reporting (reporting bias)	Unclear risk	Protocol not available
Other bias	Low risk	None identified

Husain 2019.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): 78 eyes of 78 participants in total: 40 to phacoemulsification, 38 to phaco‐GSL Exclusions after randomization: 2 in phacoemulsification, 3 in phaco‐GSL Number analyzed (total and per group): 78 eyes of 78 participants in total: 40 in phacoemulsification, 38 in phaco‐GSL Unit of analysis: participants (one eye per participant was enrolled) Losses to follow up: 3 in phacoemulsification, 2 in phaco‐GSL How was missing data handled?: last‐observation‐carried‐forward method Reported power calculation: Yes (power 88%) "Taking a conservative estimate of a difference of 4 mmHg and an SD from the previously used sample populations of 5.5 mmHg, 35 participants were needed in each arm for the study to have an 88% power to detect a difference between the null (no difference in mean IOPs postoperatively) and alternative (4 mmHg difference between the 2 postoperative means) hypotheses at the α significance level of 0.05. Assuming a dropout rate of 10%, 78 participants were needed to be recruited into this study." Unusual study design? none
Participants	Country: Hong Kong, Singapore, Thailand, Vietnam Age: phacoemulsification 67.3 (± 8.6) years; phaco‐GSL 68.1 (± 9.2) years Gender (percent): 29 (72.5%) female in the phacoemulsification group; 25 (65.8%) female in the phaco‐GSL group Inclusion criteria: Age ≥ 30 years Diagnosis of PACD More than or equal to 90 degrees of PAS (not necessarily contiguous) IOP > 21 mmHg if untreated or ≤ 21 mmHg on topical medication Lens opacity deemed sufficient to be causing decreased vision in the opinion of the operating surgeon Written informed consent Exclusion criteria: Previous intraocular surgery (except laser iridotomy was allowed) Previous eye trauma For patients on warfarin, INR > 3.0 on day of surgery Evidence of moderate non‐proliferative diabetic retinopathy, neovascularization, or rubeosis iridis Chronic use of topical or systemic steroids Only eye (visual acuity worse than 6/60 Snellen in non‐study eye) Advanced glaucoma with severe visual field (VF) defect, defined as visual field deficit threatening fixation Equivalence of baseline characteristics: similar except for central corneal thickness (531.5 μm in phacoemulsification, 557.1 μm in phaco‐GSL). Mean age, gender, mean preoperative BCVA, IOP, number of medications, Shaffer grade (gonioscopic grading), axial length, ACD, and endothelial count were comparable between the two groups.
Interventions	Intervention1: phacoemulsification Inventention 2: phacoemulsification combined with goniosynechialysis (phaco‐GSL) Length of follow‐up: 12 months
Outcomes	Primary outcome, as defined in study reports: IOP at 12 months Secondary outcomes, as defined in study reports: intraoperative or postoperative complications; degree of angle opening; amount of peripheral anterior synechiae from baseline to 12 months (clock hours) as measured by gonioscopy; change in the degree of angle opening as measured by gonioscopy Adverse events reported: "There was 1 intraoperative complication; mild zonulysis was noted during PEI‐GSL in 1 patient and necessitated the insertion of a capsular tension ring. In terms of postoperative complications, there were 3 in the PEI group (7.5%) and 3 in the PEI‐GSL group (7.9%). The most common postoperative complication was an IOP spike (IOP ≥ 30 mmHg) during week 1, which occurred in 3 patients (7.5%) — all in the PEI group (P = 0.24). The other complications, all in the PEI‐GSL group, were excessive anterior chamber inflammation (1 patient [2.6%]) in the early postoperative period and posterior capsule opacification (2 of 38 [5.3%])." Intervals at which outcomes assessed: postoperative month 12
Notes	Type of study: published Funding sources: This study was supported by grant Council NMRC/NIG/1042/2011 from the Singapore National Medical Research Council. Disclosures of interest: none reported Study period: 29 September 2011 to 16 March 2015 Reported subgroup analyses: yes (participants with previous acute PAC or not) Trial registration: NCT02376725
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomized (1:1) using a random number generator"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) Outcome assessors	Low risk	"All personnel performing study procedures were masked to the randomization of patients. The masking code was broken after the study had been completed and analysis had taken place" "The investigators measuring the main outcomes (RH, TD, JL, NK, SAP, and CLH) were masked to the treatment"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Intention‐to‐treat analysis was conducted as planned
Selective reporting (reporting bias)	Low risk	All outcomes specified in clinical trial registry were reported in the final paper
Other bias	Low risk	None identified

Moghimi 2015.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): 101 eyes (number of participants unknown) in total: 50 to phacoemulsification, 51 to phacoemulsification with viscogonioplasty Exclusions after randomization: excluded due to loss to follow‐up (see below) Number analyzed (total and per group)(intention‐to‐treat analysis): in total ‐ 46 in phacoemulsification, 45 in phacoemulsification with viscogonioplasty Unit of analysis: eyes Losses to follow up: 10 in total; 4 in phacoemulsification, 6 in phacoemulsification with viscogonioplasty How was missing data handled? excluded from analysis Reported power calculation: Yes (power 80%) "For an SD of 3 mmHg, we needed 37 patients in each group to achieve an 80% power to detect a difference of 2 mmHg in final IOP between the 2 groups with a P value of 0.05." Unusual study design?: none
Participants	Country: Iran Age: phacoemulsification with viscogonioplasty 63.2 (± 6.9) years; phacoemulsification 63.2 (± 6.9) years Gender (percent): 26 (58%) female in phacoemulsification with viscogonioplasty; 27 (59%) female in phacoemulsification Included criteria: (1) PACG defined by at least 270° of iridotrabecular contact (appositional orsynechial on gonioscopy) and glaucomatous optic neuropathy determined by optic disc cupping and glaucomatous visual field loss; (2) synechial closure of at least 90°; (3) visually significant cataract with best corrected visual acuity (VA) of worse than 20/30; (4) IOP 35 mmHg; and (5) patent peripheral iridotomy Excluded criteria: (1) eyes with a history of an acute or subacute attack of angle‐closure glaucoma; (2) presence of any cause of secondary glaucoma including uveitic glaucoma, neovascular glaucoma, exfoliative glaucoma, and phacomorphic glaucoma; and (3) history of any previous incisional eye surgery. Equivalence of baseline characteristics: yes; mean age, gender, mean preoperative BCVA, mean ACD, and mean lens vault were comparable between the two groups
Interventions	Intervention1: phacoemulsification with viscogonioplasty Intervention 2: phacoemulsification Length of follow‐up: 12 months
Outcomes	Primary outcome, as defined in study reports: IOP after 1 year; number of IOP‐lowering medications after 1 year Secondary outcomes, as defined in study reports: angle and anterior segment parameters, amount of synechial closure, and surgical complications Adverse events reported: None of the cases developed vitreous loss or endophthalmitis. Four participants in the phaco‐VGP and 2 in the phaco alone group developed fibrin reaction postoperatively that resolved after a few days. No prolonged uveitic episodes or significant iris trauma documented postoperatively. Three participants in the phaco‐VGP group developed hyphema intraoperatively that subsided within seconds with viscotamponade. Intervals at which outcomes assessed: postoperative day 1, week 1, months 1, 3, 6, and 12
Notes	Type of study: published Funding sources: not reported Disclosures of interest: "The authors declare no conflict of interest" Study period: October 2009 to February 2012 Reported subgroup analyses: no Trial registration: registered in Iranian Registry of Clinical Trials, but the identification number was not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Patients were randomized into two groups using computer‐generated random blocks to receive either cataract extraction alone (phaco alone group) or cataract extraction with VGP (phaco‐VGP group)"
Allocation concealment (selection bias)	Low risk	"An investigator with no clinical involvement in the trial consecutively assigned each patient enrolled by the clinical investigators to one of the two groups according to the random blocks"
Blinding (performance bias and detection bias) Outcome assessors	Low risk	"Patients and clinical investigators who performed gonioscopy, AS‐OCT, and IOP measurement were masked to the treatment arm"
Incomplete outcome data (attrition bias) All outcomes	High risk	No intention‐to‐treat analysis (participants who were lost to follow‐up were excluded from the analysis)
Selective reporting (reporting bias)	Unclear risk	Although the study was registered with the Iranian Registry of Clinical Trial (IRCT), there was no available published protocol to which we could compare
Other bias	Low risk	none identified

Tham 2008.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): Total: 72 eyes of 72 participants Per group: phacoemulsification alone = 35 eyes; combined phacotrabeculectomy with adjunctive mitomycin C = 37 eyes Exclusions after randomization: none reported Number analyzed (total and per group): Total: 72 eyes of 72 participants Per group: phacoemulsification alone = 35 eyes; combined phacotrabeculectomy with adjunctive mitomycin C = 37 eyes Unit of analysis: individual, one eye per participant Losses to follow‐up: not reported, only participants with 24 months of follow‐up are included in this report How was missing data handled? not reported Reported power calculation: No. The authors stated that power calculation was performed, but this was not elaborated upon in the published text beyond the statement "this study was powered to study the differences in IOP control between the 2 treatment groups." In addition, the authors stated that "this study may not have sufficient power (sample size) and follow‐up duration to examine outcomes, such as glaucomatous progression." Unusual study design? none
Participants	Country: Hong Kong, China Age: phacoemulsification alone: mean 71.9 ± 6.7 years (range 51 to 85); combined phacotrabeculectomy with adjunctive mitomycin C: mean 71.4 ± 6.6 years (range 57 to 83) Gender (percent): phacoemulsification alone 63% female; combined phacotrabeculectomy with adjunctive mitomycin C 41% female Inclusion criteria: patients with medically controlled chronic angle‐closure glaucoma (CACG) and coexisting cataract; patients able and willing to give informed consent to phacoemulsification or combined phacotrabeculectomy before randomization Diagnostic criteria for CACG: At least 180° of angle closure obliterating pigmented part of trabecular meshwork, whether synechial or appositional, segmented or continuous, in the presence of a patent peripheral iridotomy Requiring IOP‐lowering medications or IOP > 21 mmHg without IOP‐lowering medications Visual field loss compatible with glaucoma, or glaucomatous optic disc changes, or both Medically controlled CACG was defined as IOP ≤ 21 mmHg, with 3 or fewer topical drugs (combination drugs counted as 2 drugs) Diagnostic criteria for cataract: presence of nucleus sclerosis, cortical cataract, or subcapsular cataract VA of 20/50 or worse, and affecting activities of daily living Exclusion criteria: patients with a single functional eye, or refusing surgery for either indication, and prior intraocular surgery except prior laser peripheral iridotomy and argon laser peripheral iridoplasty Equivalence of baseline characteristics: yes; mean age, gender, laterality, preoperative IOP, IOP‐lowering drugs, visual acuity, vertical cup to disc ratio, visual field measured by automated perimetry were similar between the two groups
Interventions	Intervention 1: phacoemulsification alone. Topical prednisolone and chloramphenicol were given, as appropriate, postoperatively. One dose of oral acetazolamide (250 mg) given to prevent intraocular pressure (IOP) spike. Intervention 2: combined phacoemulsification with trabeculectomy with adjunctive mitomycin C. Mitomycin C (0.4 mg/mL) was applied to the scleral surface for two minutes. Topical prednisolone and chloramphenicol were given as appropriate postoperatively. Additional interventions: 16 of 35 eyes in phacoemulsification group and 17 of 37 eyes in combined phacoemulsification group required posterior synechialysis or pupil stretching Length of follow‐up: Planned: minimum of 2 years postoperative follow‐up Actual: mean ± SD follow‐up durations were 30.9 ± 8.4 months in the phacoemulsification group (range 24 to 46.9 months) and 30.6 ± 5.9 in the combined phacotrabeculectomy with adjunctive mytomycin C group (range 24 to 45 months)
Outcomes	Primary outcomes, as defined in study reports: IOP and number of IOP‐lowering drugs Secondary outcomes as defined in the study: best corrected visual acuity (BCVA), changes in optic nerve morphology, visual field changes, surgical complications, and need for additional surgical interventions Adverse events reported: no Intervals at which outcomes assessed: postoperatively at 1 month, 3 months, and then every 3 months for a minimum of 2 years
Notes	Type of study: published Funding source(s): direct grant for research from the Chinese University of Hong Kong 2004‐2005 Disclosures of interest: "The authors have no proprietary or commercial interest in any materials discussed in this article" Study period: enrollment between September 2002 and September 2005; end of follow‐up not clearly reported Reported subgroup analyses: no Trial registration: protocol registered in the Center for Clinical Trials of the Chinese University of Hong Kong, but the identification number unknown
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The study eye was then randomized into 1 of 2 treatment groups by using a random number table"
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) Outcome assessors	Unclear risk	Masking was not reported
Incomplete outcome data (attrition bias) All outcomes	Low risk	No apparent loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	Although the study protocol was registered with the Center for Clinical Trials of the Chinese University of Hong Kong, there was no published protocol available to which we could compare the results
Other bias	Low risk	None identified.

Tham 2009.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): Total: 51 eyes of 51 participants Per group: phacoemulsification alone = 27 eyes of 27 participants; combined phacotrabeculectomy = 24 eyes of 24 participants Exclusions after randomization: none Number analyzed (total and per group): Total: 51 eyes of 51 participants Per group: phacoemulsification alone = 27 eyes of 27 participants; combined phacotrabeculectomy = 24 eyes of 24 participants Unit of analysis (individuals vs. eyes): individual, one eye per participant Losses to follow up: none reported How was missing data handled?: not reported Reported power calculation: Yes (power of 80%) "The mean preoperative IOP (± standard deviation) was 27.4 (± 8.5) mmHg in the medically uncontrolled CACG eyes with cataract in the pilot study. The aim of the surgery was to achieve a “normal” untreated IOP, which was defined as 21 mmHg or less. Thus, the required reduction in IOP was 6.4 mmHg. On the basis of these data, and assuming an alpha of 0.05 and a power of 80%, to prove that the 2 interventions were equivalent in decreasing the IOP to the normal level, a minimum of 23 patients would be required in each arm." Unusual study design? none
Participants	Country: Hong Kong Age: mean: phacoemulsification alone 70.3 ± 7.4 years (range 52 to 83); combined phacotrabeculectomy 70.4 ± 9.0 years (range 55 to 91) Gender (percent): phacoemulsification alone 66% female; combined phacotrabeculectomy with adjunctive mitomycin C 75% female Inclusion criteria: patients with medically uncontrolled CACG and cataract Diagnostic criteria for CACG At least 180° of angle closure obliterating pigmented part of trabecular meshwork, whether synechial or appositional, segmented or continuous, in the presence of a patent peripheral iridotomy Requiring IOP‐lowering medications or IOP > 21 mmHg without IOP‐lowering medications Visual field loss compatible with glaucoma, or glaucomatous optic disc changes, or both Medically uncontrolled CACG was defined as IOP greater than 21 mmHg despite maximally tolerated medications, or requiring more than 3 topical drugs for IOP control Diagnostic criteria for cataract: Presence of nuclear sclerosis, cortical cataract, or subcapsular cataract VA of 20/50 or worse, and affecting activities of daily living Exclusion criteria: single functional eye; patient refusing either cataract extraction or trabeculectomy; previous intraocular surgery, with the exception of laser peripheral iridotomy and argon laser peripheral iridoplasty Equivalence of baseline characteristics: yes; mean age, gender, laterality, preoperative IOP, IOP‐lowering drugs, visual acuity, vertical cup to disc ratio, visual field measured by automated perimetry were similar between the two groups.
Interventions	Intervention 1: phacoemulsification alone Intervention 2: combined phacotrabeculectomy with adjunctive mitomycin C Length of follow‐up: Planned: 2 years Actual: mean follow‐up duration 33.4 ± 10.5 months (range 24.0 to 53.5 months) in the phacoemulsification group; 37.6 ± 10.3 months (range 24.0 to 69.4 months) in the combination group
Outcomes	Primary outcome, as defined in study reports: IOP and number of IOP‐lowering drugs Secondary outcomes, as defined in study reports: BCVA, changes in optic nerve morphology and visual field, surgical complications, need for additional surgical interventions Adverse events reported: yes Intervals at which outcomes assessed: baseline, 1 day, 1 week, 1 month, 3 months, and then every 3 months for a minimum of 2 years
Notes	Type of study: published Funding sources: supported by a direct grant for research from the Chinese University of Hong Kong 2004–2005. The funding organization had no role in the design or conduct of this research. Disclosures of interest: none reported Study period: recruitment from March 2002 to October 2007 Reported subgroup analyses: no Trial registration: protocol registered in the Center for Clinical Trials of the Chinese University of Hong Kong (identification number unknown)
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"The study eye was then randomized into 1 of 2 treatment groups using a random number table."
Allocation concealment (selection bias)	Unclear risk	Not described
Blinding (performance bias and detection bias) Outcome assessors	High risk	"Observers were not masked from the surgery performed."
Incomplete outcome data (attrition bias) All outcomes	Low risk	No apparent loss to follow‐up
Selective reporting (reporting bias)	Unclear risk	Although the study protocol was registered with the Center for Clinical Trials of the Chinese University of Hong Kong, there was no published protocol available to which we could compare the results
Other bias	Low risk	None identified

Tham 2013.

Study characteristics
Methods	Study design: randomized, parallel‐group, controlled trial Number randomized (total and per group): Total: 50 eyes of 50 participants Per group: phacoemulsification ‐ 26 eyes of 26 participants; trabeculectomy with mitomycin C ‐ 24 eyes of 24 participants Exclusions after randomization: none Number analyzed (total and per group): Total: 50 eyes of 50 participants Per group: phacoemulsification ‐ 26 eyes of 26 participants; trabeculectomy with mitomycin C ‐ 24 eyes of 24 participants Unit of analysis (individuals vs. eyes): individual, one eye per participant Losses to follow up: none reported How was missing data handled?: N/A Reported power calculation (Y/N): N Unusual study design? none
Participants	Country: Hong Kong Age: phacoemulsification: mean 66.4 ± 8.1 years (range 47 to 74); trabeculectomy: mean 65.1 ± 12.2 years (range 40 to 87) Gender (percent): phacoemulsification 62% female; trabeculectomy with adjunctive mitomycin C 79% female Inclusion criteria: age 50 years and older, eyes with medically uncontrolled CACG and no cataract, patients able and willing to give informed consent Diagnostic criteria for medically uncontrolled CACG: At least 180° of angle closure obliterating pigmented part of trabecular meshwork, whether synechial or appositional, segmented or continuous, in the presence of a patent peripheral iridotomy Requiring IOP‐lowering medications, or IOP of more than 21 mmHg without IOP‐lowering medications Visual field loss compatible with glaucoma, glaucomatous optic disc changes, or both Definition of medically uncontrolled: IOP of more than 21 mmHg despite maximally tolerated medications or requiring more than 3 topical drugs for IOP control (combination drugs counted as 2 drugs Diagnostic criteria for no cataract: best corrected visual acuity of 20/40 or better, and not affecting activities of daily living Exclusion criteria: single functional eye, patient declining either lens extraction or trabeculectomy, previous intraocular surgery, with the exception of laser peripheral iridotomy and laser peripheral iridoplasty Equivalence of baseline characteristics: yes; mean age, gender, laterality, preoperative IOP, IOP‐lowering drugs, visual acuity, vertical cup to disc ratio, visual field measured by automated perimetry were similar between the two groups
Interventions	Intervention 1: phacoemulsification with primary intraocular lens implantation Intervention 2: trabeculectomy with mitomycin C Length of follow‐up: planned: at least 2 years; actual: 2 years
Outcomes	Primary outcome, as defined in study reports: IOP and number of IOP‐lowering drugs Secondary outcomes, as defined in study reports: logMAR BCVA, surgical complications, and need for additional surgical interventions Adverse events reported: yes Intervals at which outcomes assessed: baseline, 1 day, 1 week, 1 month, 3 months, and then every 3 months for 2 years
Notes	Type of study: published Funding sources: supported by a direct grant for research from the Chinese University of Hong Kong 2008–2009, Hong Kong SAR, People’s Republic of China. The funding organization had no role in the design or conduct of this research. Disclosures of interest: none Study period: recruitment period from April 2005 to April 2009 Reported subgroup analyses: not reported Trial registration: protocol is available from: www2.ccrb.cuhk.edu.hk/registry/public/24
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“The study eye then was randomized into 1 of 2 treatment groups, using a random number table: lens extraction by phacoemulsification with primary intraocular lens implantation or trabeculectomy with adjunctive mitomycin C chemotherapy.”
Allocation concealment (selection bias)	Unclear risk	Allocation concealment was not reported.
Blinding (performance bias and detection bias) Outcome assessors	Unclear risk	The surgeon was not masked to the intervention group: "All surgical procedures were performed by a single experienced glaucoma surgeon." No other information was provided about masking of participants.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No apparent loss to follow‐up.
Selective reporting (reporting bias)	High risk	Several prespecified outcomes were not reported as part of the published results. This included the anesthetic method required, anesthetic complications, and subjective pain score by visual analog scale; operative time; angle status by indentation gonioscopy, and ultrasound biomicroscopy (UBM); change in visual field by Humphrey automated perimetry; change in vertical cup‐to‐disc ratio of the optic nerve head; rate of aqueous outflow by pneumatonometer measurements; quality of life measures; and cost‐effective analysis, including any additional procedures required to control IOP and cost of glaucoma drugs.
Other bias	Low risk	None identified

BCVA: best corrected visual acuity 
EQ‐5D: European Quality of Life‐5 dimensions
ETDRS: Early Treatment Diabetic Retinopathy Study
IOP: intraocular pressure
NEI‐VFQ‐25: National Eye Institute Visual Function Questionnaire‐25
PACG: primary angle‐closure glaucoma
QALY: quality adjusted life year

Characteristics of excluded studies [ordered by study ID]

Study	Reason for exclusion
Acton 1997	Non‐comparative study
Cai 2009	Acute angle closure was studied
Casson 2011	Editorial, no original data presented
Chen 2016	Studied acute angle‐closure glaucoma rather than chronic PACG
Chen 2019	Studied acute angle‐closure glaucoma rather than chronic PACG
Fang 2012	Studied acute angle‐closure glaucoma rather than chronic PACG
Ge 2000	Non‐comparative study
Greve 1988	Non‐comparative study
Gunning 1991	Non‐comparative study
Gunning 1998	Not a RCT
Guo 2018	Studied acute angle‐closure glaucoma rather than chronic PACG
Hayashi 2000	Non‐comparative study
Hayashi 2001	Non‐comparative study
Hou 2014	Incorrect comparison (one‐site vs two‐site phacotrabeculectomy) and participant group
Huang 2011	Non‐comparative study
IRCT20181017041367N	Incorrect intervention ‐ goniotomy rather than lens extraction
Jacobi 2001	Non‐comparative study
Jacobi 2002	Studied acute angle‐closure glaucoma rather than chronic PACG
Jamil 2011	Non‐comparative study
Ko 2004	Non‐comparative study
Kubota 2003	Not a RCT
Lee 2009	Non‐comparative study
Leng 2013	Incorrect interventions
Li 2011	Non‐comparative study
Li 2012	Not a RCT
Li 2015	Studied acute angle‐closure glaucoma rather than chronic PACG
Li 2018	Studied acute angle‐closure glaucoma rather than chronic PACG
Mori 1993	Non‐comparative study
Reibaldi 1992	Non‐comparative study
Satoh 2003	Non‐comparative study
Sheng 2006	Retrospective study
Shin 2010	Non‐comparative study
Shokoohi‐Rad 2019	Incorrect intervention
Steuhl 1992	Non‐comparative study
Su 2011	Non‐comparative study; acute angle closure was studied
Su 2013	Studied acute angle‐closure glaucoma rather than chronic PACG
Varma 2010	Included participants with previous acute angle‐closure attacks. Based on private correspondence with authors, who did not keep data on this, and were thus unable to provide subgroup data for analysis.
Wu 2008	Non‐comparative study
Yamagami 1994	Non‐comparative study
Yang 1997	Non‐comparative study
Yu 2012	Non‐comparative study
Zhang 2007	Non‐comparative study
Zhang 2016	Incorrect intervention
Zhang 2019	Incorrect intervention
Zhou 2009	Non‐comparative study

Characteristics of studies awaiting classification [ordered by study ID]

Feng 2007.

Methods	Methods unclear from information available in abstract ‐ description suggestive of a prospective cohort study
Participants	Participants with PACG and cataract with different degrees of synechial angle closure
Interventions	1. phacoemulsification plus intraocular lens implantation 2. phacoemulsification plus intraocular lens implantation and goniosynechialysis
Outcomes	1. Intraocular pressure; 2. Anti‐glaucoma medication use; 3. Visual acuity.
Notes	Unable to retrieve full text to clarify eligibility ‐ description suggestive of a prospective cohort study rather than a RCT

Lee 2015.

Methods	Randomized, parallel‐group, controlled trial
Participants	Particpants with medically well controlled chronic angle‐closure glaucoma with cataract
Interventions	1. phacoemulsification alone 2. combined phacoemulsification and goniosynechialysis
Outcomes	1. Best‐corrected visual acuity; 2. Intraocular pressure; 3. Visual field; 4. Axial length; 5. Anterior chamber depth; 6. Angle structure by goniolens; 7. Number of glaucoma drugs used
Notes	Awaiting author response regarding PACG participants without previous history of acute angle‐closure attacks

Lin 2017.

Methods	Randomized, parallel‐group, controlled trial
Participants	PACG with cataract
Interventions	1. phacoemulsification combined with intraocular lens implantation 2. peripheral iris resection surgery
Outcomes	1. Intraocular pressure; 2. Visual acuity; 3. Anterior chamber depth; 4. Complication
Notes	Awaiting author response to obtain data on participants without previous history of acute angle‐closure attacks

Liu 2011.

Methods	Randomized, parallel‐group, controlled trial
Participants	Participants with PACG with coexisting cataract
Interventions	1. phacoemulsification with goniosynechialysis 2. phacoemulsification alone
Outcomes	1. Intraocular pressure; 2. Visual acuity
Notes	Unable to retrieve full text to clarify eligibility

Paul 2014.

Methods	Randomized, parallel‐group, controlled trial
Participants	Participants with medically controlled and medically uncontrolled chronic angle‐closure glaucoma with coexisting cataract
Interventions	1. phacoemulsification alone 2. combined phacotrabeculectomy with adjunctive mitomycin C
Outcomes	1. Surgical complications; 2. Intraocular pressure 3. Progressions of glaucomatous optic nerve head morphology; 4. Cup to disc ratio 5. Progressions of glaucomatous visual field loss
Notes	Awaiting author response to obtain data on participants without previous history of acute angle‐closure attacks

Shao 2015.

Methods	Randomized, parallel‐group, controlled trial
Participants	Participants with PACG
Interventions	1. phacoemulsification with goniosynechialysis 2. phacoemulsification alone
Outcomes	1. Visual acuity; 2. Intraocular pressure; 3. Number of antiglaucoma medications; 4. Anterior chamber depth; 5. Cornea simK 6. Central corneal thickness
Notes	Awaiting author response regarding PACG participants without previous history of acute angle‐closure attacks

PACG: Primary angle closure glaucoma

Characteristics of ongoing studies [ordered by study ID]

ChiCTR1900022198.

Study name	A multicenter randomized controlled study for lens extraction combined with intraocular lens implantation for primary angle‐closure glaucoma
Methods	Randomized, parallel‐group, controlled trial
Participants	Participants with PACG
Interventions	Early PACG 1: Lens extraction; Early PACG 2: Laser peripheral iriotomy; Late PACG 1: Trabeculectomy; Late PACG 2: Lens extraction; Poor intraocular pressure control after laser peripheral iriotomy with PACG 1: trabeculectomy; Poor intraocular pressure control after laser peripheral iriotomy with PACG 2: Lens extraction
Outcomes	1. Intraocular pressure; 2. Degree of angle closure
Starting date	Enrollement started in April, 2019
Contact information	apps.who.int/trialsearch/Trial2.aspx?TrialID=ChiCTR1900022198
Notes	Recruitment commenced in April 2019 and is still ongoing

Differences between protocol and review

We excluded non‐randomized studies in this update because we identified relevant RCTs. We added 'Assessment of reporting biases' and 'Summary of findings' in the Methods for this update. We used odds ratio instead of risk ratio because individual participant data were analyzed following the prespecified statistical analysis plan in the original study to account for correlation within participants (Azuara‐Blanco 2016). The protocol specified assessment of methodological quality per previous versions of the Cochrane Handbook for Systematic Reviews of Interventions. In the review, we adopted methods, including assessment of risk of bias, as described in Higgins 2011. 

Contributions of authors

Coordinating the update of this review: AYO, SMN
Undertaking manual searches: AYO, SMN
Screening search results: AYO, SMN
Organizing retrieval of papers: AYO, SMN
Screening retrieved papers against inclusion criteria: AYO, SMN
Assessing risk of bias: AYO, SMN
Abstracting data from papers: AYO, SMN
Writing to authors of papers for additional information: AYO, DSF
Providing additional data about papers: AYO, DSF
Data management for the review: AYO, SMN
Entering data into RevMan 5: AYO, SMN, SSV
Analysis of data: AYO, SMN, SSV
Interpretation of data: AYO, SMN, SSV, DSF
Writing the review: AYO, SMN, SSV, DSF
Performing previous work that was the foundation of current review: SSV, DSF

Sources of support

Internal sources

• No sources of support supplied

External sources

• Cochrane Eyes and Vision US Project, USA

  Supported by grant 1 U01 EY020522, National Eye Institute, National Institutes of Health

• National Institute for Health Research (NIHR), UK

  • Richard Wormald, Co‐ordinating Editor for the Cochrane Eyes and Vision (CEV) received financial support for his CEV research sessions from the Department of Health through the award made by the National Institute for Health Research to Moorfields Eye Hospital NHS Foundation Trust and UCL Institute of Ophthalmology for a Specialist Biomedical Research Centre for Ophthalmology

  • The NIHR also funds the CEV Editorial Base in the UK

  The views expressed in this publication are those of the authors and not necessarily those of the NIHR, NHS, or the Department of Health.

Declarations of interest

DF: has received payment for consulting from Novartis, W.L. Gore, and Life Biosciences, Bausch and Lomb.

SMN, AYO, and SSV: None.

New search for studies and content updated (conclusions changed)