Article Text

Download PDFPDF

Selective laser trabeculoplasty versus topical medication as initial glaucoma treatment: the glaucoma initial treatment study randomised clinical trial
  1. Ghee Soon Ang1,
  2. Eva K Fenwick2,3,4,
  3. Marios Constantinou3,
  4. Alfred Tau Liang Gan2,
  5. Ryan Eyn Kidd Man2,
  6. Robert J Casson5,
  7. Eric A Finkelstein4,
  8. Ivan Goldberg6,
  9. Paul R Healey6,7,
  10. Konrad Pesudovs8,9,
  11. Sutha Sanmugasundram3,
  12. Jing Xie3,
  13. Rachel McIntosh3,
  14. Jonathan Jackson10,11,
  15. Anthony P Wells12,
  16. Andrew White6,7,13,
  17. Keith Martin1,3,14,15,
  18. Mark J Walland1,
  19. Jonathan G Crowston2,3,4,15,
  20. Ecosse Luc Lamoureux2,3,4
  1. 1 Glaucoma Investigation and Research Unit, Royal Victorian Eye and Ear Hospital, Melbourne, Victoria, Australia
  2. 2 Singapore Eye Research Institute, Singapore National Eye Centre, Singapore
  3. 3 Centre for Eye Research Australia, Royal Victorian Eye and Ear Hospital, Melbourne, Victoria, Australia
  4. 4 Duke-NUS Medical School, Singapore
  5. 5 Ophthalmology and Visual Sciences, South Australian Institute of Ophthalmology, University of Adelaide, Adelaide, South Australia, Australia
  6. 6 Discipline of Ophthalmology, The University of Sydney, Sydney, New South Wales, Australia
  7. 7 Westmead Institute for Medical Research, Sydney, New South Wales, Australia
  8. 8 School of Optometry and Vision Science, University of New South Wales, Sydney, New South Wales, Australia
  9. 9 Vision and Eye Research Institute, Anglia Ruskin University, Cambridge, United Kingdom
  10. 10 National Vision Research Institute, Australian College of Optometry, Melbourne, Victoria, Australia
  11. 11 Royal Victoria Hospital (BHSCT), Belfast, Northern Ireland
  12. 12 Capital Eye Specialists, Wellington, New Zealand
  13. 13 Cambridge NIHR Biomedical Research Centre, University of Cambridge, Cambridge, UK
  14. 14 Ophthalmology, Cambridge University Hospitals NHS Foundation Trust, Cambridge, UK
  15. 15 Ophthalmology, Department of Surgery, University of Melbourne, Melbourne, Victoria, Australia
  1. Correspondence to Dr Ecosse Luc Lamoureux, Ophthalmology, University of Melbourne, Melbourne, VIC 3010, Australia; ecosse{at}


Background/Aims To determine if selective laser trabeculoplasty (SLT) is superior to topical medication as a first-line treatment for glaucoma on quality of life (QoL) and clinical outcomes.

Methods In this international, longitudinal, multisite randomised controlled trial, treatment naïve mild-to-moderate primary open angle or exfoliation glaucoma patients were randomised 1:1 to SLT or topical medication. Glaucoma-specific QoL (primary outcome) was measured using the Glaucoma Outcomes Assessment Tool (GOAT; 342 items, 12 domains). Secondary outcomes included rate of successful intraocular pressure (IOP) reduction (>25% reduction from baseline) and presence of ocular surface disease including conjunctival hyperaemia and eyelid erythema. Our intention-to-treat analysis was performed at months 12 and 24.

Results Of 167 enrolled patients, 83 and 84 were randomised to SLT and topical medication, respectively; and 145 (n=75 SLT, n=70 medication) completed 24-month follow-up. While both treatment arms achieved significant within-group gains in GOAT outcomes at both endpoints, SLT patients reported a greater between-group improvement in ‘social well-being’ compared with medication patients (mean±SE=0.28±0.13; p=0.034) at 24 months. At month 24, the rate of successful IOP reduction was 18.6% (95% CI 3.0% to 34.3%, p=0.022) higher (absolute difference) in the medication compared with SLT group. More individuals in the medication group had conjunctival hyperaemia and eyelid erythema compared with SLT at 24 months.

Conclusion Overall, we did not find evidence that SLT was superior to medication in improving glaucoma-specific QoL. While we found superior IOP reduction in the medication arm, eyelid erythema and conjunctival hyperaemia were more prevalent in these patients compared with the SLT group.

Trial registration ACTRN12611000720910.

  • glaucoma
  • clinical trial

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Globally, glaucoma is the leading cause of irreversible, preventable vision loss, with a significant negative impact on quality of life (QoL), and resultant personal and societal economic burden.1 2 Current treatment targets intraocular pressure (IOP) reduction, which has been shown to delay glaucoma progression.3 Topical medications (eye drops) are often first-line treatment to lower IOP, although the patient burden can be significant in terms of costs and side effects.1 4–6 Other methods to treat glaucoma are available, such as selective laser trabeculoplasty (SLT), which is effective without causing any thermal damage or permanent scarring to the trabecular meshwork (TM). While SLT was initially used as an adjunct treatment to topical medications, studies have reported its suitability, efficacy and repeatability as first-line treatment for IOP reduction.7–12 Two meta-analyses,13 14 comprising results from five clinical trials,15–19 showed no significant difference in IOP reduction between SLT and topical medications as initial treatment for primary open angle glaucoma (POAG).

However, few studies to date have compared the impact of topical medications and SLT as a first-line treatment on patients’ QoL. A recent prospective clinical trial found that glaucoma patients undergoing SLT had significantly more confidence in their therapy, were less concerned about the changed appearance of their eyes, and were less inconvenienced by eye drops post-laser.20 However, this study was not a randomised controlled trial (RCT), had a short follow-up time (12 months) and used non-validated questionnaires. In contrast, the recently published LiGHT (Laser in Glaucoma and ocular HyperTension) RCT21 found no statistically significant differences in health-related QoL (measured using the EuroQoL 5-D (EQ-5D)) or glaucoma-specific QoL scores (measured using the Glaucoma Quality of Life-15 (GQL-15) questionnaire) between initial SLT or initial eye drop treatment groups at 36 months.22 This lack of between-group differences may be because the generic EQ-5D instrument lacks sensitivity in ophthalmology.23 Similarly, the GQL-15 focuses primarily on visual functioning and does not cover QoL domains which are more likely to vary across treatment types (ie, cost, convenience, concerns, social, ocular discomfort). Therefore, RCTs with sensitive, glaucoma-specific, and holistic QoL outcome measures are required to better understand the QoL impact of eye drops compared with SLT.

The Glaucoma Initial Treatment Study (GITS), an international, longitudinal and multicentred RCT, was designed to compare medication and laser trabeculoplasty as a first-line treatment for POAG and exfoliation glaucoma (XFG) in treatment naïve eyes. In this study, the effectiveness of SLT and topical medication on glaucoma-related QoL (primary outcome), health-related QoL, the rate of IOP success (defined as >25% reduction in IOP from baseline), and presence of ocular surface disease (OSD) (secondary outcomes) were compared at 12 and 24 months. We hypothesised that, compared with topical medication, SLT would improve QoL, show comparable IOP reduction, and result in less OSD.

Materials and methods

Design and procedures

The methodology and protocol for GITS have been described in detail previously.24

Treatment-naïve patients aged >35 years with mild-to-moderate POAG and XFG (0 to −12 dB mean deviation at baseline in the study eye on the Humphrey Visual Field analyser) were randomised 1:1 to SLT or medication using computer generated pseudo-random numbers. There were no inclusion/exclusion criteria for IOP. If both eyes met the inclusion criteria, they received the same treatment but the one with higher IOP was selected for the study. If both eyes had the same IOP, the study eye was selected at the investigator’s discretion. Exclusion criteria included previous use of topical or systemic ocular hypotensive medication, previous intraocular surgery (apart from uncomplicated cataract surgery), angle closure and secondary glaucoma, systemic use of medications that may affect IOP, and pregnancy. Patients were randomised after their baseline assessments, which included medical/ophthalmic and medication history taking, best-corrected visual acuity, perimetry, IOP, biomicroscopy and patient-reported outcome measures. Subsequent visits were scheduled for 6 weeks, and 3, 6, 12, 18 and 24 months (online supplementary table 1). If there was a change in medication or performance of SLT, the study patient was reviewed after 6 weeks, outside of scheduled visits. Details about the patient-centred and clinical outcome measures are provided in online supplementary materials and online supplementary table 2.

Sample size

The sample size was based on the primary outcome measure of Rasch-transformed glaucoma-specific QoL scores24 using mean and SD of Glaucoma Outcomes Assessment Tool (GOAT) pilot data and other glaucoma QoL questionnaires.25 26 To detect a 0.156 logit difference (effect size (ES) 0.3) on the Rasch-analysed GOAT at 80% power with a 5% level of significance, 175 patients were needed in each group. Assuming an attrition rate of 15% after 2 years, the target for each group at baseline was 193 patients (total n=386).

Intervention arms

SLT: A single application of pilocarpine 1% and an alpha-adrenergic agonist (either apraclonidine 1% or brimonidine 0.2% or 0.15%) were instilled prior to laser. Laser was delivered via any non-magnifying mirrored gonioscopic prism, with pulse energy ranging from 0.2 to 1.7 mJ with cavitation bubbles as the response end point. Care was taken to avoid excessive bubble formation. The initial treatment was 50–60 spots applied over the inferior 180° of the TM. If repeat SLT was required subsequently, the second SLT treatment of 50–60 spots was applied over the superior 180° of the TM. If the IOP remained insufficiently controlled at subsequent visits despite two 180° SLT treatments, the patient commenced topical medication. The SLT protocol was based on standard common practice of performing SLT in Australia.

Medication: A prostaglandin analogue eye drop (latanoprost 0.005%, travoprost 0.004%, bimatoprost 0.03% or tafluprost 0.0015%) was prescribed with one drop to be instilled in the evening. If the IOP was not sufficiently well-controlled, a stepped regime of topical medications was followed which included the addition of beta-adrenergic antagonists, alpha agonists and carbonic anhydrase inhibitors (figures 1 and 2). The choice of medications was at the discretion of the treating doctor.

Figure 1

SLT treatment regimen: participants follow a stepped treatment regimen, depending on their response to treatment. IOP, intraocular pressure; SLT, selective laser trabeculoplasty.

Figure 2

Topical medication treatment regimen: participants follow a stepped treatment regimen, depending on their response to treatment. IOP, intraocular pressure; SLT, selective laser trabeculoplasty.

Response to IOP-lowering treatment was categorised as: no response (≤10% IOP reduction from baseline), partial response (10%–25% IOP reduction from baseline) and successful response (>25% IOP reduction from baseline). If study patients crossed over both treatment regimens and still did not have adequate IOP control, they exited the trial (figure 2).

Statistical analyses

All analyses were performed using Stata V.15.0. Patients’ characteristics were summarised by mean and SD for continuous variables; and counts and percentages for categorical variables. Our primary analysis was conducted on an intention-to-treat (ITT) basis, and compared outcomes by original treatment assignments, regardless of subsequent crossover to the other arm. Comparisons of continuous and categorical outcomes were performed at each time point using two-sample t-test and χ2 test, respectively. No covariate adjustment was necessary as imbalances in patients’ sociodemographics and outcomes between treatment arms was minimal at baseline, due to randomisation. Within-group changes in QoL outcomes from baseline were assessed with a paired t-test. Ocular surface Disease Index (OSDI) severity was collapsed as a two-level variable (normal vs mild-to-severe) as ≤5 in each arm had either moderate or severe OSDI severity.

Our secondary analysis compared outcomes between treatment arms on a per-protocol (PP) basis. Patients were divided into three groups: topical medication only, SLT only, and those who switched groups. Analysis of variance was used to compare continuous outcomes between the three groups and t-tests to make pairwise group comparisons. The χ2 test was likewise used for PP comparisons of categorical outcomes. Between-group effect estimates were reported as mean differences for continuous outcomes and as absolute percentage differences for categorical outcomes. We reported all effect estimates with SE or 95% CIs, and judged a p value less than 0.05 as statistically significant.


Participant response rate

Of the 215 patients, from 15 sites, who consented to participate in the trial, 48 did not fulfil screening requirements or decided to withdraw for personal reasons (online supplementary table 3; online supplementary figure 1). The remaining 167 patients were randomised to SLT (n=83) or topical medication (n=84), with 145 participants completing the 2-year follow-up (75 and 70 patients in the SLT and medication groups, respectively).

Participants’ characteristics at baseline

The mean age (SD) of the 167 study participants was 63.8 (10.7) years (47.9% (n=80) female) and just over three-quarters were Caucasian (n=116) (table 1). Participants’ mean (SD) IOP was 20.1 mm Hg (4.4), Visual field (VF) index was 91.6% (13.1), and OSDI score was 7.9 (10.5). There were no between-group differences for sociodemographic or clinical variables at baseline (table 1).

Table 1

Participants’ sociodemographic and clinical characteristics at baseline

Treatment allocation crossovers

Seventy-seven (91.7%) patients in the medication group did not receive SLT treatment during the study period. Of these, 40 (51.9%) did not require a change in medication, 20 (26.0%) required the addition of a second medication, 13 (16.9%) required the addition of a third medication, while four (5.2%) required the addition of a fourth medication. At 12 months, 51 (63.8%) of the 80 SLT patients remaining in follow-up maintained IOP reduction without the need for topical medication, with 24 of the 51 (47.1%) requiring a second SLT treatment to do so. At 24 months, 46 (61.3%) of the 75 SLT patients remaining maintained IOP reduction, with 21 of the 46 (45.7%) requiring a second SLT treatment to do so. No patients who received SLT treatment developed any complications (eg, uveitis) attributed to the laser.

ITT analysis: The impact of SLT and medication post-intervention

GOAT (primary outcome)

Patients in the SLT group reported a statistically significant 0.27±0.81 logit mean improvement in ‘social well-being’ scores at 24 months, compared with a −0.01±0.74 mean logit reduction in the medication group. This 0.28 (0.13) mean (SE) between-group difference was statistically significant (p=0.034). No other significant between-group differences on GOAT domains were found, although significant within-group improvements at 12 and 24 months were observed for both treatment groups (table 2).

Table 2

Participants’ health-related and glaucoma-specific QoL scores at baseline, and at 12-month and 24-month follow-up in an intention-to-treat analysis

Health-related quality of life

No statistically significant within-group or between-group differences were observed for the AQoL at 12 or 24 months post-intervention (table 2).

IOP response

At month 12, 62.3% of participants in the medication group recorded a >25% response in IOP reduction compared with 45.5% in the SLT group (table 3), resulting in a significant absolute percentage between-group difference (abbreviated to AD) of 16.9% (95% CI 0.9 to 32.8, p=0.041). Similarly, at month 24, significantly more participants in the medication group (72.1%) responded to treatment compared with the SLT group (53.4%; AD: 18.6% (95% CI 3.0 to 34.3, p=0.022)).

Table 3

Intraocular pressure (IOP) responses at 12 and 24 months in an intention-to-treat analysis

Ocular surface disease

At month 24, significantly more medication group participants had eyelid erythema compared with SLT participants (AD: 12.0% (95% CI 1.5 to 22.5, p=0.027); table 4). At month 12 (AD: 27.6% (95% CI 14.1 to 41.1, p<0.001)) and 24 (AD: 17.9% (95% CI 3.1 to 32.6, p=0.019)), a significantly larger percentage of medication participants had conjunctival hyperaemia compared with SLT participants. There were no significant between-group differences in the risk of other clinical complications.

Table 4

Common side effects at 12-month and 24-month follow-up in an intention-to-treat analysis

PP analysis: Impact of SLT and medication post-intervention

GOAT (primary outcome) and Assessment of Quality of Life (AQoL)

Patients in the crossover group reported a statistically significant 0.68±1.72 and 0.41±1.07 logit mean improvement in GOAT ‘mobility’ and ‘social well-being’ scores at 24 months, respectively, compared with reductions or lesser improvements in the medication-only and SLT-only groups. This between-group difference was driven by the difference in scores between the medication-only and crossover arms (mean (SE) 0.77 (0.30), p=0.011; online supplementary table 4). There were no other between-group differences for the remaining GOAT domains or AQoL-7D scores (online supplementary table 4). For the 12 GOAT QoL domains, there were within-group improvements at 12 and/or 24 months for 7, 10, and 10 QoL domains for the medication-only, SLT-only, and crossover groups, respectively.

IOP response

A significantly higher percentage of participants in the medication-only group recorded a >25% response in IOP reduction, compared with the crossover group at month 12 (AD: 24.5% (95% CI 4.4 to 24.5, p=0.02)), and the SLT-only group at month 24 (AD: 28.7% (95% CI 10.4 to 47.0, p=0.003); online supplementary table 5). The AD in IOP response between the SLT-only and medication-only groups at month 24 was 1.5 times greater than that between the originally assigned groups estimated in the ITT analysis (AD: 18.6% (95% CI 3.0 to 34.3)).


Compared with the medication-only group, no participant in the SLT-only group had eyelid erythema at month 24 (AD: 19.4% (95% CI 9.5 to 29.2, p=0.002)), and a smaller percentage of SLT-only participants had conjunctival hyperaemia at both month 12 (AD: 35.9% (95% CI 22.0 to 49.8, p<0.001)) and month 24 (AD: 17.9% (95% CI 1.3 to 34.5, p=0.045); online supplementary table 6). A significantly higher percentage of participants in the crossover group had conjunctival hyperaemia compared with the medication-only group at month 12 (AD: 22.2% (95% CI 4.2 to 40.1, p=0.026)), and blepharitis compared with the SLT-only group at both month 12 (AD: 23.9% (95% CI 5.9 to 41.8, p=0.008)) and month 24 (AD: 21.7% (95% CI 2.7 to 40.7, p=0.024)). Between-group differences in other OSD parameters were not significant.

Sample size considerations

Our trial proved to be under-powered due to an overestimation of the predicted ES (=0.3) in our initial calculation. With an observed median ES of only 0.15 (range 0.02–0.54) for our QoL outcomes, 1606 patients would have been needed to detect a statistically significant difference in the estimated magnitude of effect between the two treatment groups with 80% power. Indeed, among the 12 GOAT outcomes, only four (with ES ≥0.3) would likely have reached statistical significance at 24 months if the target sample size of 386 was attained. Our 13% attrition rate over 2 years was within expectations.


In this international, longitudinal and multi-site RCT, we found a lack of evidence overall to support our hypothesis that patients undergoing SLT would report improved QoL at follow-up compared with those in the medication group. While we found a significant between-group difference in the GOAT social well-being domain at 2 years in ITT analysis, with patients in the SLT arm showing a significant improvement compared with no changes in the medication arm, no other between-group differences were observed for the remaining QOL domains. The clinical target of 25% IOP reduction was significantly higher in the medication compared with the SLT arm at 12 and 24 months, even after PP analysis. However, the medication group had greater risk of developing eyelid erythema and conjunctival hyperaemia compared with the SLT group at 24 months. On the whole, our results suggest that SLT is not superior to topical medication as a first-line therapy in maintaining QoL outcomes but may be inferior to medication to achieve a 25% clinical target improvement in IOP. Considering both treatments resulted in improved QoL at 24 months compared with baseline, the superiority of medication to reach the target IOP may be a clinical consideration in the treatment of naïve patients with mild/moderate POAG and XFG, notwithstanding a higher risk of developing OSD. Larger trials or meta-analyses are needed to confirm our findings.

Contrary to our main hypothesis, and similar to the recently published LiGHT trial,22 we found that SLT was not superior to topical medication on almost all QoL parameters. This was unexpected, as topical mediations have several local side effects27–29 while SLT does relatively little damage to the TM architecture30 and usually does not require multiple sessions. Moreover, unlike the LiGHT study which focused on generic health-related QoL and glaucoma-specific symptoms and visual functioning alone, our study measured a range of other glaucoma-specific QoL outcomes, including emotional well-being, concerns, social, costs and treatment inconvenience. However, our study was underpowered which may have impeded our ability to demonstrate between-group changes in QoL outcomes. Therefore, while our results suggest that SLT does not offer any advantages over topical medication from the patient’s perspective, larger RCTs or meta-analyses are needed to verify our results.

Contrary to expectations, statistically significant within-group improvements in several GOAT domains were observed for both groups at both timepoints. This finding contrasts with previous reports of no changes in QoL outcomes following treatments for glaucoma, especially in patients with early stage disease.31–33 This discrepancy in findings could be because patients in our study felt that they were proactively managing their glaucoma and feeling confident that they were on a detailed treatment plan. It may also be because the GOAT is a comprehensive instrument comprising >300 items, giving it the content coverage to detect even subtle changes in glaucoma-specific QoL. With a large number of items per domain, the level of test information (defined as the sum of the information contained in each item) in each domain would be far higher compared with brief measures. Indeed, the AQOL-7D, despite having a vision domain, was not sensitive to within-group changes for either treatment arm at both timepoints. This finding suggests that the newly developed GOAT provides comprehensive, precise, and responsive measurement of glaucoma-specific QoL and could be considered for use in future clinical trials involving glaucoma patients, administered using a computerised adaptive testing method to reduce respondent burden.34 Further work is however need to confirm this hypothesis.

Even with our small sample size, we found a significantly higher proportion of patients meeting the clinical target of 25% IOP reduction in the medication compared with the SLT arm. This finding is inconsistent with previous work15–19 and could be because our SLT protocol was limited to two applications only and because the entire 360° of the TM only received SLT once.12 35 Some patients who initially responded to SLT may have achieved 25% IOP reduction over the 2-year study period without the need for additional medications if another 360° of SLT had been applied. Previous trials used different treatment protocols, where initial SLT treatment was 360° (vs initial treatment of 180° in our protocol),15 18 19 and/or the comparison was against latanoprost monotherapy (vs a stepwise escalation of multiple medications in our protocol).16–18

OSD and reduced tear breakup time (TBUT) are common among glaucoma patients, especially with longer usage and larger medication load due to the preservatives and active ingredients.36–39 It is therefore unsurprising that our study demonstrated a greater proportion of patients with eyelid erythema and conjunctival hyperaemia among the medication versus the SLT group. However, there was no significant between-group difference for OSD and TBUT at any timepoint, even after excluding crossover participants. De Keyser and colleagues observed similar findings, where patients with stable POAG who switched from medical therapy to SLT reported less redness of the eye compared with those who remained on medical therapy, despite no difference in TBUT between the two groups.20

The main strength of our study is the comprehensive evaluation of glaucoma-specific QoL using validated and Rasch-transformed questionnaires. The GOAT was specifically developed for glaucoma patients and covers 12 QoL domains.25 Another strength is the recruitment of treatment naïve eyes thereby reducing the confounding effect of previous treatment, the 2-year follow-up duration allowing for comparison of long-term sustainability, the international multisite study design, and assessment of OSD via the OSDI and TBUT.

The main limitation of study was that our target sample size of 386 patients could not be reached, despite extending the recruitment period for 2 years and increasing the number of study sites. Recruitment was difficult because most study sites were tertiary referral centres, where most patients were referred for treatment augmentation rather than initiation, and were therefore not treatment naïve on referral. Future studies could consider engaging primary care partners to help identify treatment naïve patients. Interestingly, a post hoc analysis suggests that even our original target sample size would have been insufficient to detect a between-group difference based on the observed median ES of only 0.15. This finding is supported by a recent study by Jones and colleagues, who also found that patient-reported outcome measures, including glaucoma-specific ones, may not be sensitive enough to function as primary endpoints in clinical trials to detect treatment-induced changes over a relatively short follow-up when participants have newly diagnosed early-stage glaucoma, due to the asymptomatic nature of the disease.31 This is an important consideration for future studies conducted in similar patient populations.

Another limitation was the lack of documentation of the orbital side effects of prostaglandin analogue medications, such as periorbital skin pigmentation and prostaglandin-associated peri-orbitopathy. Prostaglandin analogue medications were the first-line therapy for the medication group; it is likely that some patients would have experienced these side effects. However, the acceptability of these specific side effects was not determined either clinically or through QoL questionnaires. Finally, SLT was performed by different clinicians using different instruments, resulting in a lack of standardisation throughout the various treatment sites. While perhaps not ideal in an RCT, it is probably a more accurate reflection of real-life clinical settings and therefore may make our results more generalisable.

In summary, our patient-centred outcomes were largely comparable across both treatment arms, although medication appears to have a better efficacy to achieve the clinical target reduction in IOP in treatment naïve mild-to-moderate POAG and XFG patients. Further studies that can overcome the substantial recruitment barriers associated with this specific glaucoma population are required to verify our results.


The following Principal Investigators comprise ‘The Glaucoma Initial Treatment Study (GITS) Group’: Dr Lance Liu, Dr Jonathan Ruddle, Dr Mark Walland, Professor Anthony Wells, Dr David Manning, Associate Professor Jamie Craig, Associate Professor Jonathan Jackson, Associate Professor Sushil Vasudevan, Professor Ravi Thomas, Associate Professor Jemaima Che Hamzah, Professor Keith Martin and Dr Andrew White. The authors would like to extend thanks to all study site staff and study coordinators for their coordination efforts in the GITS study.



  • JGC and ELL are joint senior authors.

  • GSA and EKF contributed equally.

  • Contributors GSA and EKF contributed equally as first authors and wrote the manuscript and interpreted the data; MC ran the trial and drafted sections of the paper; ATLG and JX conducted the data analysis, assisted with interpretation of results and drafted sections of the paper; REKM assisted with interpretation of results and drafted sections of the manuscript; RJC, IG, PRH, KP, JJ, APW, AW, KM and MJW contributed to the study protocol; ELL and JGC contributed to the conception of the work and interpretation of the data; all authors revised the manuscript critically for important intellectual content.

  • Funding This Randomised Controlled Trial was funded by a Project Grant from the Australian National Health and Medical Research Council (NHMRC #1009844-EL as PI). CERA receives operational infrastructure support from the Victorian Government. Last and corresponding author ELL was funded by a NHMRC Research Fellowship (#1045280) and co-last author JGC was funded by an NHMRC Practitioner Fellowship (#529915) and by the Dorothy Adele Edols Charitable Trust. Co-main author EKF was funded by an NHMRC Early Career Fellowship (#1072987).

  • Competing interests IG has the following conflicts of interest to declare: Advisory Board member for Allergan, Novartis, Mundipharma; Speaker Bureau for Allergan, Novartis; Travel support from Pfizer. KM has the following conflicts of interest to declare: Consultant to Allergan, Novartis, Quethera, Sensimed.

  • Patient consent for publication Written informed consent was collected from each participant prior to commencement in the study. Recruitment started in 2012 and ceased in 2017.

  • Ethics approval The study was conducted according to the tenets of the Declaration of Helsinki and ethical approval was obtained the Human Research Ethics Committee (HREC) of the Royal Victorian Eye and Ear Hospital (Reference number 11/1024 hour), Westmead Hospital (Reference number 2011/10/4.1 (3391) AU RED HREC/11/WMEAD/229), the Royal Australian and New Zealand College of Ophthalmologists (Reference number 34.11), the Health and Disability Ethics Committees of New Zealand (Reference number 12/STH/13), the Southern Adelaide Clinical Human Research Ethics Committee (Reference number 81.13) and National Research Ethics Services Cambridge East (Reference number 13/EE/0204) the Australian College of Optometry Human Research Ethics Committee (Ref H13 002).

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

  • Data availability statement Data are available upon request.

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost B Jonas