Selective laser trabeculoplasty versus drops for newly diagnosed ocular hypertension and glaucoma: the LiGHT RCT

Article history

The research reported in this issue of the journal was funded by the HTA programme as project number 09/104/40. The contractual start date was in September 2012. The draft report began editorial review in September 2018 and was accepted for publication in December 2018. The authors have been wholly responsible for all data collection, analysis and interpretation, and for writing up their work. The HTA editors and publisher have tried to ensure the accuracy of the authors’ report and would like to thank the reviewers for their constructive comments on the draft document. However, they do not accept liability for damages or losses arising from material published in this report.

Declared competing interests of authors

Gus Gazzard, David Garway-Heath, Rachael Hunter, Gareth Ambler, Catey Bunce, Richard Wormald, Keith Barton, Gary Rubin and Marta Buszewicz have received a grant from the National Institute for Health Research (NIHR) for the submitted work. Gus Gazzard reports grants from Lumenis (Borehamwood, UK) during the conduct of the study; grants from Ellex Medical Lasers (Adelaide, SA, Australia), Ivantis, Inc. (Irvine, CA, USA) and Thea Pharmaceuticals (Keele, UK) outside the submitted work; and personal fees from Allergan (Dublin, Ireland), Alcon (Fort Worth, TX, USA), Glaukos Corporation (San Clemente, CA, USA), Santen Pharmaceutical Co., Ltd (Osaka, Japan) and Thea Pharmaceuticals outside the submitted work. David Garway-Heath reports personal fees from Aerie Pharmaceuticals, Inc. (Durham, NC, USA), Alcon, Allergan, Bausch + Lomb (Rochester, NY, USA), Quark Pharmaceuticals, Inc. (Ness Ziona, Israel), Quethera Limited and Roche (Basel, Switzerland); grants from the Alcon Research Institute; and grants and personal fees from Pfizer (New York, NY, USA) and Santen Pharmaceutical Co., Ltd, outside the submitted work. In addition, David Garway-Heath was a member of the Health Technology Assessment (HTA) Clinical Trials Board from 2014 to 2017. Keith Barton received a grant from NIHR for the Treatment of Advanced Glaucoma Study during the conduct of the study. In addition, Keith Barton reports grants from Johnson & Johnson Vision (Santa Ana, CA, USA), New World Medical (Rancho Cucamonga, CA, USA), Alcon, Merck & Co. (Kenilworth, NJ, USA), Allergan and Refocus Group (Dallas, TX, USA); that he has had other financial relationships with Alcon, Merck & Co., Allergan, Refocus, AqueSys Inc. (Taipei, Taiwan), Ivantis, Carl Zeiss Meditec AG (Jena, Germany), Kowa Europe GmbH (Düsseldorf, Germany), Santen Pharmaceutical Co., Ltd, Transcend Medical (Scottsboro, AL, USA), Glaukos (San Clemente, CA, USA), Amakem NV (Diepenbeek, Belgium), Thea Pharmaceuticals, Alimera Sciences (Alpharette, GA, USA), Pfizer, Advanced Ophthalmic Implants Pte Ltd (Singapore), Vision Futures (UK) Ltd (London, UK), London Claremont Clinic Ltd (London, UK) and Vision Medical Events Ltd (London, UK), outside the submitted work; and that he has a patent with Ophthalmic Implants (PTE) Ltd. Stephen Morris was a member of NIHR Health Services and Delivery Research (HSDR) Research Funding Board (2014–19), the NIHR HSDR Commissioned Board (2014–16), the NIHR HSDR Evidence Synthesis Sub Board (2016), the NIHR HTA Clinical Evaluation and Trials Board (2007–10), the NIHR HTA Commissioning Board (2009–13), the NIHR Public Health Research Funding Board (2011–17) and the NIHR Programme Grants for Applied Research expert subpanel (2017–present). Marta Buszewicz was a member of the HTA Mental Health Panel from January to May 2018. In September 2018 this panel was amalgamated into the HTA Prioritisation Committee C (mental health, women and children’s health), of which Marta Buszewicz was also a member. Marta Buszewicz has also been a member of the NIHR Research for Patient Benefit, London, funding panel since May 2017.

Copyright statement

© Queen’s Printer and Controller of HMSO 2019. This work was produced by Gazzard et al. under the terms of a commissioning contract issued by the Secretary of State for Health and Social Care. This issue may be freely reproduced for the purposes of private research and study and extracts (or indeed, the full report) may be included in professional journals provided that suitable acknowledgement is made and the reproduction is not associated with any form of advertising. Applications for commercial reproduction should be addressed to: NIHR Journals Library, National Institute for Health Research, Evaluation, Trials and Studies Coordinating Centre, Alpha House, University of Southampton Science Park, Southampton SO16 7NS, UK.

2019 Queen’s Printer and Controller of HMSO

Background

Rationale for research

Research recommendations by NICE17 and a Cochrane systematic review54 have identified the need for robust randomised controlled trials investigating the efficacy and cost-effectiveness of SLT as a first-line treatment for OAG and OHT.

Aims and objectives

Trial design

The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial was designed to evaluate the difference in HRQoL, cost and clinical efficacy between two first-line treatment arms for OAG and OHT. The LiGHT trial is a multicentre, randomised clinical trial, unmasked to treatment allocation, with two treatment arms: initial SLT followed by routine medical treatment (Laser-1st) and routine medical treatment only (Medicine-1st). 55,56

Eligible patients were randomised in a 1 : 1 ratio to receive either SLT (Laser-1st) or medical therapy (Medicine-1st) as the first-line treatment for OAG or OHT. All measurements influencing treatment escalation decisions [VF, Heidelberg retinal tomography (HRT) and IOP] were made by masked observers. Patients were monitored for 3 years and monitoring intervals were guided by a defined protocol to avoid bias in clinical decision-making. A clinical decision algorithm, attempting to capture the complexities of clinical practice, defined triggers for escalation. This was a pragmatic trial aiming to mirror the ‘real-world’ patient experience of treatment as closely as possible and seeking to capture the full effects of laser treatment.

Ethics approval and research governance

The study adhered to the tenets of the Declaration of Helsinki. Ethics approval was granted by the City Road and Hampstead Research and Ethics Committee (former Moorfields and Whittington Research Ethics Committee then East London and The City Research Ethics Committee 1, reference 12/LO/0940) on 20 June 2012. The LiGHT trial is registered as ISRCTN32038223 [the full protocol can be accessed at URL: www.moorfields.nhs.uk/sites/default/files/LiGHT%20Trial%20Protocol%203.0%20-%2020-5-2015_3.pdf (accessed 3 May 2019)].

Patient population

The LiGHT trial aimed to recruit patients with newly diagnosed OAG or OHT in one or both eyes from six collaborating specialist glaucoma clinics at large ophthalmic centres in the UK (see Appendix 1).

Inclusion criteria

Parts of this text have been reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Patients were required to have newly diagnosed OAG or OHT in one or both eyes, which needed treatment. Definitions of OAG and OHT, as well as criteria for initiating treatment, are shown in Appendix 2. The following criteria were also specified:56

• A decision to treat had been made by a glaucoma specialist consultant ophthalmologist.

• Patients were aged > 18 years and were able to provide informed consent.

• Patients were able to complete QoL, disease-specific symptom and cost questionnaires in English (physical help with completion and assistance with reading was permitted, as long as an interpreter was not required).

• It was possible to perform a VF test in the study eye(s) with < 15% false positives.

Exclusion criteria

Parts of this text have been reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Patients were not considered for the study if there was:56

• advanced glaucoma in the potentially eligible eye as determined by Early Manifest Glaucoma Trial (EMGT I)59 criteria (77 VF loss mean deviation (MD) worse than –12 dB in the better eye or –15 dB in the worse eye)

• secondary glaucoma (e.g. pigment dispersion syndrome, rubeosis, trauma, etc.) or any angle closure

• any contraindication to SLT (e.g. unable to sit at the laser-mounted slit-lamp, past history of or active uveitis, neovascular glaucoma, inadequate visualisation of trabecular meshwork)

• an inability to use topical medical therapy because of, for example, physical infirmity and a lack of carers able to administer daily eyedrops

• a previous treatment for OAG or OHT

• congenital or early childhood glaucoma

• a visually significant cataract in symptomatic patients who want to undergo cataract surgery

• any current, active treatment for another ophthalmic condition in the hospital eye service (this applied to both eyes, even if one was not in the trial, as the fellow eye might affect the patient’s visit frequency)

• any history of retinal ischaemia, macular oedema or diabetic retinopathy

• age-related macular degeneration with neovascularisation in either eye or geographic atrophy

• visual acuity (VA) worse than 6/36 in a study eye; non-progressive VA loss better than 6/36 owing to any comorbidity was permitted provided that it did not affect the response to treatment or later surgical choices and that it was not under active follow-up (e.g. an old, isolated retinal scar no longer under review or amblyopia)

• any previous intraocular surgery, except uncomplicated phacoemulsification, at least 1 year before recruitment (this applied to both eyes, even if one was not in the trial, as it could affect the required treatment intensity and visit frequency for any glaucoma in the fellow eye)

• pregnancy at the time of recruitment or intention to become pregnant within the duration of the trial

• medical unsuitability for completion of the trial (e.g. suffering from a terminal illness or too unwell to be able to attend hospital clinic visits)

• recent involvement in another interventional research study (within 3 months) of any topic.

Recruitment

Baseline assessment

At the baseline assessment, and after informed consent was provided, participants underwent VA testing, slit-lamp examination, automated VF testing [Humphrey Field Analyser Mark II (Carl Zeiss Meditec, Dublin, CA, USA) and the Swedish Interactive Threshold Algorithm standard 24-2 programme], HRT optic disc imaging, IOP measurement, gonioscopy, CCT measurement and assessment of the optic discs, maculae and fundi. The patients also completed the following baseline questionnaires: EuroQol-5 Dimensions, five-level version (EQ-5D-5L),60 Glaucoma Utility Index (GUI),61 Glaucoma Symptom Scale (GSS),62 Glaucoma Quality of Life-15 (GQL-15; a visual function, rather than quality-of-life measure)8 and a modified version of the Client Service Receipt Inventory (CSRI) questionnaire. 63

Randomisation and masking

Following the completion of all baseline assessments, eligible patients were randomised to one of two treatment groups: SLT (Laser-1st) or topical medical therapy (Medicine-1st). Randomisation was undertaken online on the same day by the clinical staff who obtained informed consent, using a web-based randomisation service (Sealed Envelope, London, UK) and achieving full allocation concealment. Stratified randomisation with random block sizes was used to randomise in a 1 : 1 ratio at the level of the patient, with the stratification factors of diagnosis (OHT/OAG) and treatment centre. Following randomisation, the details of the treatment and specific arrangements and instructions were communicated to the patients by a member of the trial team. Owing to the pragmatic design of this trial, the patients and clinicians were unmasked to the treatment arm; however, all clinical measurements (IOP, VF, HRT) were carried out by masked observers and treatment decisions were masked by the use of a computerised evidence-based decision support algorithm. 56

Treatment arm allocation

Disease stratification and initiation of treatment

The NICE-recommended thresholds were used for defining disease (OAG or OHT) for entry into the study, as well as in initiating treatment (see Appendix 2). 15 The patients’ clinical evaluation and test outcomes were then entered into the clinical decision algorithm and a disease category and stage were determined. The algorithm used severity criteria from the Canadian target IOP workshop,69 with central-field loss severity criteria defined according to Mills et al. 70 (Table 1). Severity stratification determined the follow-up frequency.

Computerised decision algorithm

The follow-up and treatment escalation protocols were enabled by custom-written clinical decision support software (DSS), which permitted real-time decision-making based on the analysis of multiple clinical measures, including HRT optic disc analysis, VF assessment and IOP measurements. Predefined objective indicators of either disc or field deterioration [change in mean neuroretinal rim area, as determined by HRT, or VF glaucoma progression analysis (GPA)], or IOP above target all triggered earlier follow-up and/or increased treatment intensity.

Setting individual target intraocular pressure

Once the decision to treat was made, a treatment target IOP (target) was set. The target was eye specific and was objectively defined and adjusted by the computerised decision algorithm to avoid bias from unmasked treating clinicians. The lowest permitted target was 8 mmHg for OAG and 18 mmHg for OHT. Although CCT has an effect on IOP measurement and risk of progression, the true magnitude of this interaction is unknown because of complex non-linear interactions between CCT, ‘true’ IOP and corneal material properties; CCT was therefore not used in the algorithm for setting a target IOP. 56 Myopia and family history were also not included in this algorithm, as data on the effect size of these risk factors on progression rates are weak. 71 The target IOP was either an absolute reduction to below a specified level or a percentage reduction from baseline, whichever was lower. The process of setting the IOP target is illustrated in Figure 1. Greater reductions were required for greater disease severity as defined by Canadian glaucoma study criteria. 73

FIGURE 1.

Process for target IOP setting. a, Disease stratification according to Mills et al. 70 POAG, primary open-angle glaucoma. Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018. 72

Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018

Failure to meet target intraocular pressure and target intraocular pressure re-evaluation

Parts of this text have been reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Diurnal fluctuation and measurement error both lead to variation in measured IOP. Kotecha et al. 68 have shown that inter-visit variation may nonetheless be as much as ± 4 mmHg. 68 To prevent an inappropriate escalation to a more intensive treatment, it is therefore important to repeat measurements that deviate only slightly from target. Criteria for failure to meet, and to reassess, target IOP follow those of the Canadian glaucoma study,74 taking into account that inter-visit variation in IOP measurement may be as much as ± 4 mmHg:

• If IOP in an eye was ≥ 2 mmHg but < 4 mmHg above target on two consecutive visits and showed possible or definite progression, then the treatment was intensified and the target remained unchanged.

• If IOP in an eye was ≥ 2 mmHg and < 4 mmHg above target on two consecutive visits and showed no progression (with a minimum of three post-baseline follow-up VF tests required to confirm progression, as per EMGT I),59 then the target was adjusted upwards. In this case the target IOP was revised to the mean of the previous three visits, during which progression did not occur. If VF testing had been carried out at fewer than three follow-up visits, additional visits were required to confirm stability before the target was relaxed.

• If IOP in an eye was ≥ 4 mmHg from target at any visit, then the eye was considered to have failed to reach target and treatment intensity was increased to the next level (unless already at the maximum), irrespective of any progression, unless the clinician identified poor concordance with treatment. In such cases the target remained unchanged. In the presence of poor concordance and in the absence of progression, additional measures to improve concordance before escalation of treatment were permitted, as in usual clinical practice.

• If the IOP of an eye on MMT was ≥ 2 mmHg from target and showed definite progression, then glaucoma drainage surgery was offered to the patient.

• If the IOP of an eye on MMT was ≥ 2 mmHg from target and showed possible progression, then the follow-up frequency was increased until progression was either confirmed or ruled out.

• If the IOP of an eye on MMT was ≥ 2 mmHg from target but below maximum IOP (maximum IOP is that above which surgery may be offered even without progression: OHT, 35 mmHg; mild glaucoma, 24 mmHg; moderate and severe glaucoma, 21 mmHg), and showed no progression (with at least three follow-up VFs), then the target was adjusted (revised to the mean of the previous three visits) with an increase in follow-up frequency. If VF testing had been carried out at fewer than three follow-up visits, additional visits were required to confirm stability.

• A patient with an eye with IOP above the maximum IOP may have been offered surgery without progression at the discretion of the treating surgeon.

• If there was progression and the IOP was at target, then the target IOP was reduced by 20% (according to the Canadian glaucoma study protocol),74 with a lower limit of 8 mmHg, and treatment intensified accordingly.

Failure to meet target can be a result of poor concordance as well as a lack of drug efficacy. As is normal practice, compliance was discussed and patients were counselled at each visit, using validated ‘ask–tell–ask’ techniques. 75–77 Patients were given standard written information from the International Glaucoma Association, face-to-face instruction in eyedrop administration and the offer of further nurse-led support.

Where poor concordance was thought to be the contributing factor, education with written information and repeated face-to-face instruction in eyedrop administration was given. If the decision was made to educate rather than escalate a patient who was not at target, then the reason for an algorithm over-ride was recorded (non-concordance) and the patient recalled after 8 weeks for a repeat IOP check visit. 56

Treatment escalation

To minimise bias for escalating treatment, standardised criteria for any additional intervention were used, in accordance a protocol following international guidelines by the European Glaucoma Society,64 American Academy of Ophthalmology Preferred Practice Pattern65 and the South East Asia Glaucoma Interest Group. 66 Treatment is escalated under the following circumstances:56

• Strong evidence of progression irrespective of IOP (see Defining progression).

• IOP above target by > 4 mmHg at a single visit (irrespective of evidence of progression).

• IOP above target by < 4 mmHg and less strong evidence of progression (see Defining progression). If the IOP is above target by < 4 mmHg with no evidence of progression, then the treatment target IOP is re-evaluated (see Failure to meet target intraocular pressure and target intraocular pressure re-evaluation).

The process for escalating treatment is shown in Figures 2 and 3.

FIGURE 2.

Process for escalating treatment in OHT. a, On two consecutive visits; b, as per protocol; and c, until progression confirmed/refuted. VF progression required three follow-up VF assessments. Maximal IOP, IOP above which surgery was offered even without progression or 35 mmHg for OHT. Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018. 72

Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018

FIGURE 3.

Process for escalating treatment in OAG. a, On two consecutive visits; b, as per protocol; and c, until progression confirmed/refuted. VF progression required three follow-up VF assessments. Maximal IOP, IOP above which surgery was offered even without progression or 35 mmHg for OHT. Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018. 72

Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018

More stringent criteria than those used for laser or medical treatment were applied before being referred for surgery. This reflected the greater risk to a patient’s vision from surgical complications. Strong evidence of progression and/or failure to meet target was usually required in all but the most severe disease. However, extreme elevations of IOP could be an indication for surgery without progression, with lower thresholds in more damaged eyes. Any patient in whom IOP was at or above the maximum was reviewed (in person or remotely) by the PI, who decided whether or not surgery was indicated. In accordance with the principle of patient-centred care, the decision to operate was always a collaboration between clinician and patient. When an IOP-lowering surgical intervention was indicated, cataract surgery was permitted (in the presence of cataract, i.e. not clear lens extraction) when this was the consultant’s usual practice.

Defining progression

Algorithm over-ride

In the following cases the algorithm was over-ridden by the treating consultant if:

• Poor concordance was thought to be the contributing factor to failure to meet IOP target and was followed by patient education and a recall 8 weeks after for an IOP check.

• It was felt that it was in the patient’s best interest to over-ride the algorithm’s decision to either revise the target IOP (upwards or downwards) or to escalate treatment.

The reason for the over-ride was recorded.

Follow-up procedure

Follow-up intervals were set at entry to the study, based on disease severity and lifetime risk of loss of vision, according to NICE guidance,15 and subsequently adjusted on the basis of IOP control, disease progression or adverse reactions. Disease stability, along with all available data, was taken into consideration, but testing for progression did not independently determine follow-up intervals. The routine schedule of appointments for patients who remained at or below the target IOP, without progression or treatment change, and who had no adverse reactions requiring earlier assessment, is shown in Table 2. Additional VF tests were permissible at any visit if clinically necessary to confirm possible progression. Variation in follow-up intervals was permitted to accommodate the clinician’s judgement and/or patient choice. 56

Participants in the Laser-1st arm were reviewed 2 and 8 weeks after SLT application. After the 8-week review in the Laser-1st group and for all treatment changes in the Medicine-1st arm, patients were reviewed at 2 months, following which their treatment was changed (with consequent early assessment of response to second treatment) or they entered a disease severity-tailored routine follow-up schedule. Follow-up of patients with severe OAG was at the discretion of the consultant ophthalmologist. If an eye showed ‘possible progression’, then the follow-up frequency was increased to every 3–4 months until progression was confirmed or ruled out with additional VF testing or HRT. No further tests were conducted at additional visits for IOP check alone. All contacts with medical professionals and optometrists were captured for cost data. Information on contact with health-care providers was collected via the CSRI, a validated method of collecting health-care cost data. 63

Follow-up intervals were planned within the ranges specified by NICE guidance15 and were independently determined on the basis of IOP control or adverse reactions, to minimise bias. The main driver for follow-up frequency was treatment in pursuit of control. Disease stability was considered using all available data, but testing for progression did not independently determine follow-up. Patients who required medication changes or additional laser treatment and patients who suffered AEs or showed progression of glaucoma were seen sooner and reverted to schedule when stable. The worse or more unstable of each patient’s two eyes determined the follow-up interval, whereas treatment was individualised to the needs of each eye.

Follow-up clinical assessments

The schedule of assessments (all assessments were part of routine care) is shown in Table 3. After the full baseline assessment, all patients underwent VF testing and HRT to assess progression at each follow-up visit. The EuroQol-5 Dimensions (EQ-5D-5L) and other HRQoL questionnaires were assessed at baseline and 6-monthly thereafter, with additional questionnaires as outlined in Questionnaires.

Questionnaires

The content of the questionnaires was determined by the use of a number of validated, widely accepted existing questionnaires as follows:

• EQ-5D-5L

• GUI

• GSS

• GQL-15.

Additionally, a modified CSRI was used and two questions regarding concordance. The content of the questionnaires used can be found as supplementary material [see URL: www.journalslibrary.nihr.ac.uk/programmes/hta/0910440/#/documentation (accessed 23 April 2019)]; a sample of each questionnaire completed is presented in Appendix 5. The patient and public involvement group reviewed the final questionnaire for layout and clarity to ensure ease of completion.

Questionnaire delivery and follow-up

The baseline questionnaires were self-completed by participants in a private room, at the time of enrolment, after informed consent had been given but before randomisation. Participants were required to have sufficient English knowledge that translation was not required [practical assistance with the layout (e.g. some questionnaires were printed double-sided, some questions had conditional formatting depending on the patients’ response) and completion of the form were permitted].

Subsequent questionnaires were sent out by post for self-completion at 6-monthly intervals; up to two written reminders followed by one telephone follow-up were implemented in the case of non-response. In the event of a telephone follow-up, if the patient was willing, only the primary outcome measure was collected. Aiming to incentivise LiGHT participants to return the vital final questionnaire, a high street voucher worth £5.00 was sent by post along with the final set of questionnaires to each participant. The central site at MEH managed all questionnaires across all collaborating sites.

Follow-up has been extended beyond the primary study to look additionally at HRQoL outcomes at 6 years; questionnaires are will be posted to participants every 6 months for the duration of the extended period.

Adverse events and serious adverse events

An AE was defined as an unfavourable medical occurrence in a patient that was not necessarily caused by the treatment. GCP guidelines67 were used to determine if AEs should be classified serious [serious adverse events (SAEs)]. AEs and SAEs were reported in accordance with standard operating procedures (SOPs) and GCP guidelines, to achieve standardisation across sites and between treatment arms, with an annual safety report to the Research and Ethics Committee.

Primary outcome measure

The primary outcome measure was HRQoL in patients with OAG or OHT treated with SLT first, compared with HRQoL in patients treated with topical medication first, measured using EQ-5D-5L utility scores at 3 years.

Secondary outcome measures

The secondary outcomes were as follows:56

• Treatment pathway health-care resource use, cost and cost-effectiveness. Health-care resource use was ascertained from the record of treatment episodes and additional health-care contacts using a modified CSRI. 63 The cost components included the cost of SLT, number of visits, number and type of medications and glaucoma surgeries, and clinical tests.

• Glaucoma-specific treatment-related QoL was measured using the GUI, from which quality-adjusted life-years (QALYs) can also be derived.

• Patient-reported disease and treatment-related symptoms using the GSS.

• Patient-reported visual function using the GQL-15.

• Objective measurements of pathway effectiveness for IOP-lowering and visual function preservation (e.g. treatment intensity and time taken to achieve target IOP, the number of target IOP revisions, proportion of patients achieving target after each year of treatment, number of patients with confirmed disease deterioration and rates of ocular surgery).

• Objective safety measures for each pathway.

• Concordance assessed by two questions shown to predict the probability of non-concordance. 80

Reproduced with permission from The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018 with permission from BMJ Publishing Group Ltd. 72

Data collection and management

To standardise data collection and management, researchers were trained to follow specific SOPs for each stage of data handling. Identical electronic and hard-copy case report forms (CRFs) were designed according to a standard CRF template. A web-based database for the Priment Clinical Trials Unit, managed by the company ‘SealedEnvelope’, was used for database entry with direct data entry at the time of patient visit. This included extensive internal consistency and range checking, with a hard-copy backup CRF in case of information technology (IT) failure. Records were identifiable only by unique, confidential trial identification number with no patient-identifiable information included. All data were contemporaneously entered directly into the web-based database CRF.

Data from patient completed questionnaires received by post were scanned on receipt for e-copy back-up and entered onto the database within 1 week of receipt by the trial data management officer. Questionnaire data were from validated, standardised tools (EQ-5D-5L, GUI, GQL-15, GSS and CSRI) (see Table 3). The central site at MEH managed the inputting of data from all questionnaires across all collaborating sites.

Statistical analysis plan

The statistical analysis plan has been published previously. 81 All patients were analysed in the treatment arm to which they were randomised. All analyses were performed in Stata^® version 14 (StataCorp LP, College Station, TX, USA).

Economic evaluation

The aim of the economic evaluation was to calculate the mean incremental cost per QALY of Laser-1st compared with Medicine-1st. Health and social care costs and QALYs were calculated for the within-trial period (36 months). The outputs were:

• mean total patient-level QALYs by trial arm

• mean cost per patient of laser treatment in the Laser-1st arm

• mean cost per patient of eyedrop treatment for glaucoma by trial arm

• mean cost per patient of surgery by trial arm

• mean total health-care cost per patient over 3 years by trial arm

• mean increment cost per QALY of Laser-1st compared with Medicine-1st and 95% CIs

• cost-effectiveness planes

• cost-effectiveness acceptability curves (CEACs).

Patient and public involvement: lay advisory group

Glaucoma patients and relatives from the Cochrane Eyes and Vision Group Consumer Panel formed our independent LAG. Consultation on trial design, choice of outcome measures, recruitment and treatment acceptability took place by e-mail and through online discussions via the Facebook social networking site (Facebook, Inc., Menlo Park, CA, USA; Group: ‘Public Eye – LiGHT Trial Discussion Forum’). All of the suggestions made have been incorporated (e.g. requests to monitor all symptoms in detail, a safety concern about ‘rapid loss of pressure control’ after SLT and more explanation of the relationship between eyedrops and surgical failure). An ‘expert patient’ with treated glaucoma reviewed and commented on the study protocol as a service user member for the Trial Steering Committee (TSC) and another service user representative from the International Glaucoma Association was invited to join. The LAG contributed to the development of tailored information leaflets and consent forms with further consultation with service user groups and via the Friends of Moorfields charity. A survey of 100 new patients attending MEH to assess the acceptability of an invitation to participate in such a trial, before the commencement of the trial, had a 70% positive response.

Patients diagnosed and treated for glaucoma also provided input to a questionnaire sent in 2015 (see Appendix 6), allowing us to design an extension for the main LiGHT trial; the questionnaire looked into the views of these treated patients on current treatment options and their willingness to switch from eyedrops to laser.

As required by the NHS, in line with INVOLVE national guidelines and in accordance with UK Clinical Research Collaboration policy, the results are being communicated to patients, for example via NHS Choices and patient advocate groups (e.g. International Glaucoma Association), and the findings have been published in open access media. 58

Study oversight and management

Recruitment

A total of 16,379 patients were assessed for eligibility; 15,483 were excluded as they did not meet the inclusion criteria. Of the 896 patients who were eligible across the six participating NHS centres, a total of 718 (1235 eyes) were recruited for the study (80.1% participation rate). A recruitment chart for the total recruitment period can be found in Appendix 8. A total of 178 eligible patients declined to participate. Of the patients who declined to participate, 43 did not want to have SLT, 17 did not want to take part in research, nine did not want to use eyedrops, three did not want to receive any treatment, one did not want to travel to the hospital and 105 did not provide an explanation.

Participant flow

A total of 718 patients (1235 eyes) were randomised: 356 patients (613 eyes) were allocated to SLT (Laser-1st pathway) and 362 patients (622 eyes) to medical treatment (Medicine-1st pathway) (Figure 4). Two patients were randomised twice owing to failure, as a result of which the initial randomisation was not visible. Subsequently, a second randomisation was carried out; one of these patients had initially been randomised to medication (non-visible randomisation), but was subsequently randomised to, and received, SLT. The second patient was initially randomised to SLT (non-visible randomisation), but was later randomised to, and received, medication. Four patients who did not meet the eligibility criteria were randomised in error and were subsequently removed from the study (see Appendix 9).

FIGURE 4.

The LiGHT trial Consolidated Standards of Reporting Trials (CONSORT) flow diagram. a, Two patients were randomised twice owing to IT failure, as a result of which the initial randomisation was not visible, and subsequently a second randomisation was carried out. Reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Participant baseline characteristics

Reproduced with permission from The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018 with permission from BMJ Publishing Group Ltd. 72

Participant baseline characteristics are shown in Table 5. Of the 718 patients recruited, approximately 70% were based in London: 52% were recruited by MEH and almost 15% from Guy’s and St Thomas’ Hospital. The average age of the patients was 63.1 years (± 11.8 years), with more male patients recruited than female (55.3% male vs. 44.7% female). In total, 70% of all participants were white, and black was the second largest ethnic group (20%). Thirty per cent of patients reported a family history of glaucoma affecting at least one first-degree relative. Systemic hypertension (defined as the use of prescribed antihypertensive medication) was recorded in 35% of the patients. Use of systemic antihypertensive medication was recorded in these patients; of those, 5% were using beta-blockers, 16% were using calcium channel blockers and 14% were using angiotensin-converting enzyme inhibitors; 27% of all patients were on statins. In total, 11% of the patients were smokers at the time of recruitment. Approximately 30% of the LiGHT patients had a degree or equivalent qualification, 13% had achieved higher education, 12% had achieved Advanced Level or equivalent and 45% did not pursue an education beyond 16 years of age.

A total of 301 patients (41.9%) had bilateral OAG, 161 patients (22.4%) had unilateral OAG (fellow eye healthy), 93 patients (13.0%) had OAG in one eye and OHT in the other eye, 124 patients (17.3%) had bilateral OHT and 39 patients (5.4%) had unilateral OHT (fellow eye healthy). A total of 555 patients (77.2%) were classified as having OAG (if at least one eye was affected by OAG) and 163 patients (22.7%) were classified as having OHT; both eyes were eligible for the trial in 517 patients (72.0%), only the right eye was eligible in 96 patients (13.4%), and only the left eye was eligible in 105 patients (14.6%); 55% of the better eyes were right eyes. 72

The baseline patient characteristics were similar between the two groups in terms of age, sex distribution, ethnicity, general health and family history of glaucoma (see Table 5). There were small differences in the medication and education of the patients. The eye characteristics were also similar between the two treatment arms, with VA, VF MD, HRT optic disc rim area, IOP and CCT comparable between the two treatment arms (Table 6).

The baseline scores for QoL (EQ-5D-5L, GSS, GUI and GQL-15) are shown in Table 7. The two treatment arms had similar average EQ-5D-5L scores (Medicine-1st 0.92 ± 0.13, Laser-1st 0.91 ± 0.13; higher scores indicate better HRQoL), GUI scores (Medicine-1st 0.89 ± 0.11, Laser-1st 0.89 ± 0.12; higher scores indicate better HRQoL) and GQL-15 scores (Medicine-1st 18.7 ± 5.6, Laser-1st 18.9 ± 6.6; higher scores indicate worse HRQoL) at baseline. The Medicine-1st arm showed slightly higher average GSS scores at baseline than the Laser-1st arm (Medicine-1st 83.3 ± 16.6, Laser-1st 81.4 ± 17.2; higher scores indicate better HRQoL).

Primary outcome return rates

A total of 652 patients returned the primary outcome at the trial’s end point at 36 months (overall return rate was 91%: 92% for the SLT arm and 89% for the Medicine-1st arm), and were included in the ITT analysis (with imputation used for missingness). An additional 21 patients supplied 30-month data, which was used to impute their missing 36-month data, such that 673 patients were included in the primary ITT analysis.

Losses to follow-up

A total of 16 patients in the Laser-1st arm and nine patients in the Medicine-1st arm discontinued participation (see Figure 4). In total, two patients were lost to follow-up and were no longer contactable, four patients moved to a different hospital, four patients withdrew from the trial, five patients could not continue participation owing to ill health and 10 patients died.

Quality of life

Primary outcome: EQ-5D-5L

At 36 months, the Laser-1st arm had an average EQ-5D-5L score of 0.90 (SD 0.16), compared with 0.89 (SD 0.18) in the Medicine-1st arm, suggesting little difference between the two treatment arms [adjusted mean difference (Laser-1st – Medicine-1st) 0.01, 95% CI –0.01 to 0.03; p = 0.23] (Table 8 and Figure 5). The results were confirmed in sensitivity analyses (see Appendix 10). Taking into account the outcome data from all time points across 36 months, the two treatment arms had similar EQ-5D-5L scores at 36 months [adjusted mean difference 0.02 (95% CI –0.00 to 0.03) and 0.01 (95% CI –0.01 to 0.02), when using exact times of questionnaire returns).

TABLE 8 - Primary (at 36 months) and secondary (across 36 months) analysis of HRQoL questionnaires

Laser-1st – Medicine-1st. Mean difference is adjusted for baseline score, severity, centre, baseline IOP and number of eyes affected at baseline.

Notes

EQ-5D-5L: higher scores represents a better QoL.

GUI: higher scores represents a higher QoL.

GSS: higher scores represent better outcomes.

GQL-15: higher sores represents poorer glaucoma QoL.

Reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Measure	Medicine-1st		Laser-1st		Adjusted mean differencea	95% CI	p-value
	n	Mean (SD)	n	Mean (SD)
Primary analysis at 36 months
EQ-5D-5L	336	0.89 (0.18)	337	0.90 (0.16)	0.01	–0.01 to 0.03	0.230
GUI	299	0.89 (0.13)	303	0.89 (0.13)	0.01	–0.01 to 0.03
GSS	281	83.3 (17.3)	294	83.1 (17.7)	1.6	–0.8 to 4.0
GQL-15	297	19.8 (7.8)	304	19.8 (7.2)	–0.4	–1.3 to 0.6
QALY	263	2.70 (0.42)	261	2.74 (0.37)	0.025	0.02 to 0.07	0.289
QALY (discounted)	263	2.62 (0.41)	261	2.65 (0.36)	0.024	–0.02 to 0.07	0.286
Repeated measures analysis across 36 months
EQ-5D-5L
Baseline	362	0.92 (0.13)	355	0.91 (0.13)
6 months	332	0.90 (0.15)	330	0.91 (0.13)	0.01	–0.01 to 0.03
12 months	327	0.91 (0.14)	327	0.91 (0.14)	0.01	–0.01 to 0.02
18 months	329	0.90 (0.16)	325	0.90 (0.16)	0.00	–0.02 to 0.02
24 months	326	0.91 (0.14)	326	0.91 (0.14)	–0.00	–0.02 to 0.02
30 months	320	0.90 (0.15)	317	0.90 (0.15)	0.00	–0.01 to 0.02
36 months	323	0.89 (0.18)	329	0.90 (0.16)	0.02	–0.00 to 0.03
GUI
Baseline	361	0.89 (0.11)	355	0.89 (0.12)
6 months	330	0.90 (0.11)	329	0.91 (0.10)	0.01	–0.00 to 0.03
12 months	315	0.89 (0.12)	320	0.91 (0.11)	0.01	–0.00 to 0.03
18 months	305	0.89 (0.12)	303	0.90 (0.13)	0.01	–0.01 to 0.02
24 months	298	0.89 (0.12)	305	0.90 (0.11)	0.02	0.00 to 0.03
30 months	299	0.88 (0.12)	291	0.89 (0.12)	0.02	0.00 to 0.03
36 months	300	0.89 (0.13)	303	0.89 (0.13)	0.01	–0.01 to 0.02
GSS
Baseline	357	83.3 (16.6)	353	81.4 (17.2)
6 months	321	83.0 (16.3)	320	85.6 (14.9)	4.0	2.0 to 6.0
12 months	310	83.0 (17.6)	309	85.2 (15.4)	2.9	0.8 to 4.9
18 months	295	83.1 (16.8)	294	84.6 (15.8)	2.8	0.7 to 4.8
24 months	287	83.3 (16.4)	290	83.3 (16.3)	1.4	–0.7 to 3.5
30 months	288	81.3 (17.6)	276	84.1 (16.7)	3.5	1.5 to 5.6
36 months	282	83.3 (17.3)	296	83.1 (17.7)	2.2	0.1 to 4.2
GQL-15
Baseline	361	18.7 (5.6)	355	18.9 (6.6)
6 months	323	18.8 (5.6)	324	18.3 (5.4)	–0.8	–1.6 to 0.0
12 months	314	19.2 (7.2)	318	18.8 (6.6)	–0.5	–1.4 to 0.3
18 months	302	19.1 (6.4)	298	18.9 (6.5)	–0.6	–1.4 to 0.2
24 months	289	19.5 (7.3)	298	19.2 (6.7)	–0.5	–1.3 to 0.4
30 months	293	19.9 (7.1)	287	19.6 (7.9)	–0.3	–1.1 to 0.5
36 months	298	19.8 (7.8)	304	19.8 (7.2)	–0.4	–1.2 to 0.4

FIGURE 5.

Mean (a) EQ-5D-5L, (b) GUI, (c) GSS and (d) GQL-15 scores at each time point, across 36 months. Time point ‘0’ refers to pre-treatment. a, Higher scores on EQ-5D-5L, GUI and GSS indicate better HRQoL; b, higher scores on GQL-15 indicate worse HRQoL. Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018. 72

Adapted by permission from BMJ Publishing Group Limited. The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics, Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, Ambler G, Bunce C, volume 102, pp. 599–603, 2018

Pathway clinical effectiveness

At 12 months, 606 eyes (98.9%) were available for analysis in the Medicine-1st arm and 608 eyes (97.8%) in the Laser-1st arm. At 24 months, 564 eyes (92.0%) were available for analysis in the Medicine-1st arm and 576 eyes (92.6%) in the Laser-1st arm. At 36 months, 536 eyes (87.7%) of 314 patients in the Laser-1st arm and 536 eyes (86.2%) of 312 patients in the Medicine-1st arm were available for analysis of clinical outcomes.

Concordance/compliance

At baseline, concordance with treatment was lower among patients who were to be treated with Medicine-1st than among those allocated to treatment with Laser-1st (the proportion taking their eyedrops correctly was 75% vs. 92.5%, respectively). By the end of the trial, at 36 months, self-reported concordance had improved and was similar between the two treatment arms (99% of eyedrops used correctly) (Tables 16 and 17).

Cost-effectiveness

Throughout the 36 months of the trial, patients treated with Medicine-1st made 2907 ophthalmology outpatient visits and patients treated with Laser-1st made 3441 visits. The latter includes visits at 2 weeks after the SLT treatment, which served as a safety check. None of these visits revealed a pathology that changed the course of management. A detailed table summarising all of the medical contacts is shown in Appendix 12 (see Table 28).

Incremental cost-effectiveness ratio and cost-effectiveness acceptability curve

For ophthalmology and total costs, Laser-1st dominates Medicine-1st in that it results in more QALYs for a lower cost. For ophthalmology-only costs, the results of the multiple imputation (QALYs imputed only) and bootstrap, accounting for correlation between costs and QALYs using seemingly unrelated regression, are that Laser-1st results in an average cost saving of –£458 per patient, with 95% of iterations falling between –£585 and –£345, and 0.014 additional QALYs, with 95% of bootstrap replications falling between –0.018 and 0.046. If non-eye-related costs are also included, the average cost saving of Laser-1st is –£126 per patient, with 95% of bootstrap replications falling between –£796 and £487.

The CEAC is presented in Figure 6. If ophthalmology-only costs are included, at a £20,000 and £30,000 willingness to pay for a QALY, there is a 97% and 93% probability, respectively, that Laser-1st is cost-effective compared with Medicine-1st over 3 years, discounted and adjusted. For all health-care-related costs, including non-eye-related costs, at both a £20,000 and a £30,000 willingness to pay for a QALY, there is a 68% chance that Laser 1st is cost-effective compared with Medicine-1st, discounted and adjusted, over 3 years.

FIGURE 6.

Cost-effectiveness acceptability curve of Laser-1st compared with Medicine-1st based on bootstrapped, imputed, discounted and adjusted data: ophthalmology costs and all health-care costs. Reproduced from Gazzard et al. 58 © 2019 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).

Protocol deviations and violations

A list of protocol deviations and violations is shown in Appendix 9.

Introduction

We collected costs and monitored QoL over the 3-year time horizon of the trial. However, it is possible that further costs and benefits will accrue beyond the time horizon of the trial.

Aim

The aim of the economic evaluation was to calculate the mean incremental cost per QALY of Laser-1st compared with Medicine-1st for the lifetime of the patient. Health service costs and QALYs were calculated for the lifetime of patients, drawing on data from the trial and the literature where appropriate.

Methods

Costing of the treatment pathway

The second and third Markov models were developed to capture treatment costs, including the costs of eyedrops, surgery and SLT. The model structures for each pathway are shown in Figures 7 and 8. The costs estimated in the model were combined with the health state costs in the disease state model (Figure 9).

FIGURE 7.

Medicine-1st treatment pathway.

FIGURE 8.

Laser-1st treatment pathway.

FIGURE 9.

Markov model structure.

The cost of eyedrops depended on whether the patient was receiving first, second, third or fourth escalation treatment. The model allowed for patients to remain in each eyedrops subcategory after each cycle or progress to the next treatment escalation. Patients could stay in the surgery state for only one cycle, following which they moved into the eyedrops-free window. Patients could also remain in the eyedrops-free state for multiple cycles.

The Laser-1st pathway included a state for patients who received SLT and allowed a number of patients to undergo a second laser treatment, although this transition could be made only once, in line with the trial protocol.

Results

Cost-effectiveness analysis

The average lifetime cost for Laser-1st based on 1000 runs of the PSA was £17,541 per patient, with an average cost per patient of £20,435 for the Medicine-1st arm and a difference of –£2894. Laser-1st resulted in an average QALY of 12.5 over the lifetime time horizon, compared with 12.3 for Medicine-1st (difference of 0.2 QALYs). Laser-1st dominated Medicine-1st in that it was cost saving and resulted in additional QALYs.

There was a 90% probability that Laser-1st is cost-effective compared with medication at a willingness to pay £20,000 for a QALY (Figure 10).

FIGURE 10.

Cost-effectiveness acceptability curve: probability that Laser-1st is cost-effective for a range of values of willingness to pay for a QALY.

Conclusion

When costs, outcomes and surgery rates were projected to a lifetime time horizon, there was a 90% probability that Laser-1st is cost-effective. This is similar to the findings of trial-based analysis, strengthening the finding that Laser-1st is cost saving compared with Medicine-1st.

Summary of findings

This multicentre randomised controlled trial compared initial treatment of OAG or OHT using SLT followed by medication, if required, with the use of IOP-lowering medication alone. Patient HRQoL, clinical efficacy and cost-effectiveness were investigated in six NHS settings in the UK. The study demonstrates that initial SLT is cost-effective, with better clinical outcomes and a trend towards better HRQoL compared with the prescription of IOP-lowering eyedrops from the outset.

Quality of life

The Laser-1st and Medicine-1st treatment arms showed comparable EQ-5D-5L scores at the trial’s end point, at 36 months. The trial’s protocol, whereby each eye was treated to an eye-specific IOP target, led to minimal disease-related differences, such as visual function outcomes. Indeed, VA, VF MD and VF pattern SD were comparable between the two treatment arms at 36 months (see Table 9), and this was reflected in similar EQ-5D-5L scores between the two pathways (see Table 8). The above average baseline HRQoL,82,121–123 weak sensitivity of the EQ-5D-5L to detect glaucoma-specific effects on HRQoL82,123,124 and relatively short duration of this trial, compared with the time for disease progression, may have contributed to the lack of superiority of the EQ-5D-5L in the Laser-1st approach. Recent data on patient self-reported outcome measures (including EQ-5D-5L) now confirm that these may not be sensitive enough to function as primary end points in clinical trials. 125

Glaucoma-specific instruments (e.g. GUI and GQL-15) are better at capturing differences in glaucoma severity than the effect of treatment side effects on patients’ HRQoL. Indeed, the QoL of patients with glaucoma has been related to the extent of VF loss when using the GQL-15. 8,121,126,127 The GUI attributes less weight to local side effects and provides generic health outcome measures and measures of glaucoma severity. 61 The lack of a significant difference in the GUI and GQL-15 in this study was, therefore, somewhat expected; each eye was treated to target and stringent controls over disease progression minimised any substantial differences in disease severity.

The GSS evaluates a visual and an ocular comfort-related domain. The six non-visual ophthalmic symptoms are formed around an ocular comfort domain (burning/smarting/stinging, tearing, dryness, itching, soreness/tiredness and a feeling of something in the eye), and are related to treatment side effects and their measures. The patients were asked to rate their difficulties around blurry/dim vision, hard to see in daylight, hard to see in dark places and haloes around lights, in relation to the four visual ophthalmic symptoms. The GSS has been shown to correlate well with traditional measures of visual function (such as VA and VF),62 which in this study’s end point were comparable between the two treatment arms. Repeated-measures analysis showed worse GSS scores for the Medicine-1st arm at five out of six time points over the course of the 36 months of the trial. Better GSS scores for the Laser-1st arm may represent differences that arise from eyedrop use (64.6% of the patients in the Medication-1st arm were using only one drop per day) (see Table 11), but potentially reflect differences in baseline scores between the two treatment arms (83.3 for the Medicine-1st arm and 81.4 for the Laser 1st arm) (see Table 7).

Clinical efficacy

The treatments were equally effective in lowering IOP and reaching the target IOP for the first time, with 89.6% of the Medicine-1st eyes reaching target at the first planned follow-up visit, compared with 91.0% of the Laser-1st eyes (see Table 10). By 3 years, however, 95% of the Laser-1st eyes were at target IOP, compared with 93.1% of the eyes treated with Medicine-1st. Although Medicine-1st provided more eyes at target IOP at 12 months (96.2% for Medicine-1st compared with 94.7% for Laser-1st) (see Table 10), Laser-1st achieved more eyes at target IOP over the second and the third years of treatment. Interestingly, the percentage of eyes with severe OAG that were at target IOP after the second and third years of treatment was higher in the Medicine-1st arm than in the Laser-1st arm (89.5% compared with 87.5%, respectively, at 24 months and 85.7% compared with 84.6%, respectively at 36 months); the proportions of severe OAG eyes achieving IOP target at 12 months were comparable (91.3% and 91.5%, respectively). A possible explanation may be the limit to which SLT may reduce IOP; severe OAG is likely to have low IOP targets, which may be easier to achieve with more than one medication. A number of studies have reported that SLT reduces IOP by 15–32%,128–134 and another study found a reduction of up to 29.4% at 6 months, up to 30% at 12 months, up to 27.8% at 2 years and up to 25.1% at 3 years. 135

Overall, IOP was at target for 93% of the Laser-1st visits, compared with 91% of the Medicine-1st visits, over the 36-month duration of the trial. IOP control with topical medication (eyedrops) may rely on patient concordance with treatment; indeed, one report94 found that IOP-lowering eyedrops were available to patients for only 69% of the time, whereas concordance has been reported to range between 76% and 86%. More discouraging figures have been reported for patients on multiple eyedrops. 136,137 In this trial, however, self-reported concordance was very high, with patients’ concordance reportedly improving during the 36 months of treatment (see Tables 16 and 17). SLT has also been proposed to provide better diurnal IOP stability, because of its continuous effect on the trabecular meshwork, in contrast to the episodic administration of medication. 138–141 This trial showed a comparable IOP fluctuation between Medicine-1st and Laser-1st (2.5 mmHg and 2.3 mmHg, respectively).

By 36 months, 78.2% (95% CI 74.7% to 81.4%) of the eyes treated with Laser-1st were at target without the need for any topical IOP-lowering medication (eyedrops). In comparison, in 64.6% of eyes treated with Medicine-1st only a single eyedrop was necessary to control IOP. Primary SLT gave eyedrop-free IOP control for at least 36 months to 74.2% of patients (95% CI 69.3% to 78.6%), substantially higher than reported in previous studies that used less stringent success criteria and which used SLT exclusively as the primary treatment. 37,142–144 Prior treatment and more severe disease have been suggested to reduce the magnitude of IOP lowering with SLT,37,38,142–144 possibly explaining the results of this trial in treatment-naive patients. Pre-trial activities with glaucoma patients (LAG) identified eyedrop-free disease control as the most desired outcome, with 90% of a patient focus group feeling that even unilateral eyedrop freedom is beneficial. Concerns about eyedrop use, particularly associated with challenges from cognitive and physical impairment, were rated as a priority by patients in the James Lind Alliance survey of sight loss research questions. 145

By 36 months, rates of disease deterioration were higher in the Medicine-1st arm than in the Laser-1st arm [5.8% (36 eyes) vs. 3.8% (23 eyes), respectively], despite the treat-to-target design, tailoring treatment intensity to disease severity and treatment response. The vast majority of disease progression happened in eyes with OAG (33 out of 36 eyes in the Medicine-1st arm and 21 out of 23 eyes in the Laser-1st arm) (see Table 12). Additionally, there were more treatment escalations over 36 months in the Medicine-1st arm (348, compared with 299 in the Laser-1st arm).

Upwards (22 in the Medicine-1st arm and 26 in the Laser-1st arm) and downwards (16 in the Medicine-1st arm and 15 in the Laser-1st arm) target IOP revisions were overall balanced between the two treatment arms. By 36 months, 11 eyes in the Medicine-1st arm, but none in the Laser-1st arm, had required IOP-lowering surgery.

Safety

This trial demonstrates a greater safety of SLT than previously reported, with low rates of SLT-related AEs. 143 Rates of systemic AEs were balanced between the two treatment arms, indicating a safe treatment pathway for both Medicine-1st and Laser-1st (i.e. no excess cardiac or pulmonary events in the Medicine-1st arm and no systemic AEs as a result of SLT) (see Tables 13 and 14). Differences in the rates of cancer diagnoses and deaths between the two treatment arms (see Table 14) are attributable to chance, with no medical link between the treatments that were administered and the events that took place.

Eyedrop-related systemic and ophthalmic AEs were reported by more patients and more often in the Medicine-1st arm (see Table 13). Ophthalmic AEs led to a change in treatment for 11.3% of the patients treated with Medicine-1st, when IOP was otherwise well controlled.

Selective laser trabeculoplasty resulted in at least one transient AE in 34.4% of the patients treated with Laser-1st. Out of 776 SLT procedures, one patient developed an IOP spike requiring treatment, compared with reported rates of up to 28.8%, possibly because treatment was administered at an earlier stage of disease in this trial. 143 The IOP check conventionally done 2 weeks after SLT did not change management for any of the patients and consequently appears unnecessary.

The ocular comorbidities developed during the trial spanned a range of conditions commonly seen in patients of the age group of this cohort (retinal vascular occlusions, diabetic retinopathy, retinal tears and detachments, macular degeneration, etc.) (see Table 15). The rate of cataract surgery was lower in the Laser-1st arm (13 eyes in the Laser-1st arm, compared with 25 eyes in the Medicine-1st arm) (see Table 15), supporting existing evidence that IOP-lowering eyedrops are associated with a greater incidence of nuclear cataract and earlier need for surgical removal. 12,59,146–148

Economic evaluation

The Laser-1st approach resulted in a significant reduction in the cost of surgery and IOP-lowering medication, with an overall cost saving to the NHS of £451 per patient in specialist ophthalmology costs. The trial-based economic evaluation found that there is a 97% probability that SLT is a cost-effective treatment for OAG and OHT at a £20,000 willingness to pay for a QALY, from an ophthalmology cost perspective. Resource use information was collected from patient files and trial monitoring data and hence is likely to be complete, with limited bias as a result of loss to follow-up or missing data. Including non-eye-related health-care costs alongside ophthalmology costs, the average cost per patient for Laser-1st remained less than that for Medicine-1st, but the differences between the two groups were not significant, with the wide CIs resulting in a 68% probability that Laser-1st is cost-effective compared with Medicine-1st. Non-ocular health-care cost data were, however, based on self-reported health-care resource use and may be unreliable or incomplete. Expensive systemic AEs unrelated to OHT or OAG, such as cancer, may have also skewed the cost results. In the health economic decision model, in which costs and utilities are projected for the lifetime of patients, there is a 90% probability that Laser-1st is cost-effective compared with Medicine-1st at a willingness to pay £20,000 for a QALY health-care cost perspective, with an average cost saving per patient to the NHS of £2894.

Previous economic assessments have attempted to estimate the relative costs of SLT using only economic modelling or estimates of the treatment costs, instead of a direct cost assessment plus modelling. 50,51,53 Compared with mono or multiple drug therapy, and allowing for repetition of SLT within 2–3 years, cost savings have been predicted at 6 years for a Canadian health-care system. 46 The present study, conducted in a NHS setting following pragmatic clinical approaches for the treatment of OAG/OHT, indicates that SLT is cost-effective over a 3-year period and is likely to remain cost-effective over the lifetime of the patients. These findings have important implications for patients and health-care systems. Patients are concerned about the use of IOP-lowering eyedrops,145 and widespread uptake of the Laser-1st strategy would lead to an eyedrop-free interval of at least 36 months for three-quarters of patients, while also providing savings for the NHS. If necessary, additional SLT may further increase the duration of this eyedrop-free window. Patients with OAG have an average life expectancy from diagnosis of 9–13 years,149–151 and so an eyedrop-free period of ≥ 3 years may significant increase the quality of their remaining life. The requirement for intense medical or surgical regimes may be deferred or completely averted by SLT and potentially with improved surgical success rates and still lower cost. 31,33,34

The health economic decision model has some limitations. First, the mortality rate used is not glaucoma specific. Rather, the same all-cause mortality rate was applied across all health states. This is a reasonable assumption given that studies have not found a difference in all-cause mortality rates between patients with and without glaucoma. 149 There is, however, an increased risk of cardiovascular disease in older patients (aged > 75 years), which was not incorporated into the model. This is unlikely to have made a significant impact on the results of the model, given the small difference in OAG progression between the two arms.

The lifetime economic model originally set out to include variables from other studies in the literature, including costs, utilities and disease progression. As noted in other studies that have modelled the progression of OAG, there is heterogeneity in how OHT and OAG are defined, making it difficult to incorporate such values into glaucoma models. 73 Burr et al. 73 conducted a systematic review of studies for glaucoma progression and identified 1-year progression probabilities for mild to moderate glaucoma, moderate to severe glaucoma and severe glaucoma to visually impaired. The progression values for the two health states that overlap with our model for Medicine-1st patients are different from those observed in the LiGHT trial; higher for mild to moderate progression (0.2 in Burr et al. 73 vs. 0.07 in the LiGHT trial) and lower for moderate to severe progression (0.07 in Burr et al. 73 vs. 0.17 in the LiGHT trial). This difference is likely to be a result of differences in the populations used to determine transitions probabilities, as the values for Burr et al. 73 are from observational studies of people at different stages of OAG, whereas the LiGHT trial recruited patients newly diagnosed with OHT or OAG. It may also be as a result of differences in how the various stages are defined, with Burr et al. 73 restricting the definition to mean definition score only, omitting an OHT health state and including an additional visual impairment state. Costs were also more suitably obtained from the trial or NHS reference costs90 given that most studies report costs for medication for OHT and OAG health states, or the costs for eyedrops progression, ophthalmology appointments and trabeculectomy (glaucoma surgery) only. 77 Our model is novel, in that we were able to project medication and trabeculectomy costs over the lifetime of patients while modelling the health-care costs associated with OHT and OAG health states separately, using LiGHT trial data. As discussed, our utility values for health states are similar to those quoted in other studies;73,95,103 modifying these has limited impact on the results, particularly as using published data means that we are unable to use utilities specific to trial arms.

Strengths and limitations

This study mirrored pragmatic clinical practice by tailoring treatment to the patient. Individual IOP targets were based on pre-treatment IOP and disease severity, and adapted both to treatment response and to progression of the disease. Consequently, the concluding findings on disease progression, achievement of target IOP and cost are highly relevant to normal clinical practice. The ‘treat-to-target’ design with computerised decision-supported treatment interventions and follow-up intervals captured the complexity of real-life clinical decision-making, but also allowed objective and unbiased decisions based on clinical observations.

This trial was unmasked. An unmasked design was essential to capture any treatment effects on patients’ perception, which is clinical reality. As an important benefit of SLT is an eyedrop-free treatment window; a study design requiring a treatment arm with placebo eyedrops would have prevented assessment of benefits attributable to IOP control without the use of medication. The impact of treatment on subsequent medication-taking behaviours and concordance was also of interest and will be fully revealed in the extension of this trial. Similarly, treating clinicians had to be unmasked to be able to choose appropriate treatment escalations.

The EQ-5D-5L questionnaire, a generic tool eliciting utility values in multiple settings, may not have been the most sensitive tool to investigate HRQoL in the two treatment arms for a 3-year trial, but is a requirement for a NICE cost–utility analysis. OAG can be asymptomatic, even at levels sufficient to make driving unsafe. 152 Although potentially associated with significant visual impairment over longer periods, only small changes in vision occurred over the 36 months duration of the trial, and this finding may be related to the lack of a difference in the primary outcome at 36 months.

Implications for health care

The results of this randomised controlled trial are widely generalisable, as patients with OHT and both low- and high-pressure OAG were included, from a range of backgrounds and ethnic origins. This trial focused on newly diagnosed, treatment-naive patients with OHT and/or mild or moderate OAG; the results should, therefore, be interpreted with caution in relation to the efficacy of SLT in advanced OAG stages, or in eyes previously treated for high IOP. There are important implications for resource-poor health-care settings, where access to medication is a major barrier to glaucoma treatment. Adequate eyedrop-free IOP control for years after is a promising treatment paradigm for regions of Africa where glaucoma prevalence is high. Longer follow-up, already under way, will permit us to answer further questions regarding the effect of prior SLT on later medication-taking behaviour, treatment intensity and longitudinal HRQoL.

The data support a change in practice; however, clinicians need to consider the perceived necessity of monitoring visits by the patient, in the absence of daily medication, as has been suggested by in the past for other chronic illnesses. 153 Primary SLT is a safe, clinically efficient and cost-effective alternative to eyedrops that can be offered as a first-line treatment to patients with OAG or OHT, in need of IOP reduction.

Recommendations for research

This study addresses the initial 36-month period after initiation of treatment for OAG and/or OHT. OAG and OHT are chronic conditions and treatment is almost always lifelong. Longitudinal research into the clinical efficacy of SLT as a first-line treatment could specify the long-term differences (if any) of disease progression, treatment intensity and ocular surgery rates between a Laser-1st and a Medicine-1st pathway, as well as any effect on subsequent medication-taking behaviours. Longitudinal HRQoL in OAG and OHT will also help clinicians understand the impact medical treatment has on patients over a longer period of time, where more intense medical treatment might become necessary.

Members of the Trial Steering Committee

Independent chairperson: Professor John Sparrow, Honorary Professor, University of Bristol.

Chief investigator: Mr Gus Gazzard, Consultant Ophthalmic Surgeon, MEH.

Trial manager: Dr Amanda Davis, MEH.

Independent clinician with relevant expertise: Professor James Morgan, Professor of Ophthalmology, Cardiff University.

Independent statistician: Dr Marta (Garcia-Finana) Van der Hoek, Reader in Biostatistics, University of Liverpool.

Lead statistician: Dr Gareth Ambler, Senior lecturer in Medical Statistics, University College London.

Patient and public involvement representative: Ms Shelia Page, expert patient.

Independent members of the Data Monitoring Committee

Independent statistician: Professor Chris Rogers, Professor of Medical Statistics and Clinical Trials, University of Bristol.

Independent clinician: Mr John Salmon, Consultant Ophthalmologist, Oxford Eye Hospital.

Independent triallist: Professor Caroline Free, Professor of Primary Care and Epidemiology, London School of Hygiene and Tropical Medicine.

Project management group

Mr Gus Gazzard was the chief investigator of the trial; Dr Amanda Davis was the CTM of the trial (with temporary covers from Ms Ayse Barnes and Ms Danielle Reid); Mr Neil Nathwani was the trial’s CTO; Dr Evgenia Konstantakopoulou was the trial’s CRO; Dr Gareth Ambler was the lead trial statistician; Ms Victoria Vickerstaff was the trial statistician; and Ms Taniya Chowdhury was the trial data officer.

Staff facilitating recruitment and data collection

Moorfields Eye Hospital: Ms Emily Dowse, Ms Karine Girard-Claudon, Ms Seetal Savania-Dholakia and Ms Gurveen Panesar (optometrists); Mr Emerson Tingco, Ms Kanom Bibi, Mr Charles Amoah, Mr Dominic Carrington, Mr Mohamed Illiyas and Mr Yoganand Jeetun (technicians); and Ms Taniya Chowdhury, Ms Varshini Parayoganathan and Ms Nana Fatimah Ogunfemi (data officers).

Guy’s and St’ Thomas’ Hospital: Mr Sheng Lim (local PI); Mr Arij Daas, Ms Panayiota Founti and Mr Pouya Alaghband (research fellows); and Ms Priya Morjaria, Mr Andrew Ho, Ms Jenny Lee, Mr Kishan Patel, Mr Sandip Patel and Ms Seema Rajoria (optometrists).

Royal Victoria Hospital Belfast: Ms Sarah Wilson (local PI); Ms Karen Gillvray and Ms Angela Knox (research fellows); and Ms Tricia Douglas, Ms Katie Graham and Ms Deirdre Burns (optometrists).

Hinchingbrooke Hospital: Professor Rupert Bourne (local PI); and Ms Jane Kean (optometrists).

York Hospital: Mr Timothy Manners (local PI); and Ms Archana Airody (research fellow).

Norfolk and Norwich University Hospital: Mr David Broadway (local PI); and Mr Nuwan Niyadurupola (research fellow).

Contributions of authors

Gus Gazzard (Chief Investigator, Consultant Ophthalmic Surgeon) led the initial conception and design of the trial and writing of the protocol, acquired funding and ethics approval, is the chief investigator, had complete involvement and oversight of the trial, provided clinical expertise and was a major contributor to writing the manuscript.

Evgenia Konstantakopoulou (Research Optometrist) was involved in the day-to-day running of the trial, data collection and data interpretation, wrote the manuscript together with Gus Gazzard and was responsible for the production of the final report.

David Garway-Heath (Professor of Ophthalmology) co-designed the trial and the trial protocol, provided expert advice on clinical and methodological aspects and was involved in the drafting and critical revision of the manuscript.

Anurag Garg (Research Fellow) was involved in the acquisition of the data, data analysis and interpretation, contributed to the writing of Chapter 3 and critically reviewed the manuscript.

Victoria Vickerstaff (Senior Research Fellow) was involved in the design of the statistical analysis plan and the writing of the relevant paper, the data analysis, wrote parts of Chapter 2, provided guidance on the writing of Chapter 3 and critically reviewed the manuscript.

Rachael Hunter (Associate Professor Health Economics) contributed to the design of the outcome measures and data collection as well as the health economics analysis, wrote Chapter 4 and parts of Chapter 3.

Gareth Ambler (Associate Professor) was involved in the design of the statistical analysis plan and the writing of the relevant paper, contributed to the design of the outcome measures, the data to be collected and the data analysis, and was involved in the critical revision of the manuscript.

Catey Bunce (Reader Medical Statistics) contributed to the design of the outcome measures and the data to be collected, provided expert advice on statistical and methodological aspects of the trial and was involved in the critical revision of the manuscript.

Richard Wormald (Consultant Ophthalmologist) was involved in the critical revision of the study design and protocol, provided expert advice on clinical and methodological aspects of the trial and was involved in the critical revision of the manuscript.

Neil Nathwani (Research Optometrist) was involved in the design of the trial, data collection, data interpretation and critical revision of the manuscript. He was responsible for the clinical running of the trial in the central site.

Keith Barton (Consultant Ophthalmologist) was involved in the drafting of the protocol and critical revision of the study design and manuscript, and provided clinical expertise.

Gary Rubin (Professor of Ophthalmology) was involved in the drafting of the study design, provided expert advice on clinical and methodological aspects of the trial, and was involved in the critical revision of the manuscript.

Stephen Morris (Professor Health Economics) contributed to the design of the outcome measures and the data to be collected, and the critical revision of the manuscript.

Marta Buszewicz (Associate Professor in Primary Care) contributed to the original trial design, provided methodological guidance and was involved in the critical revision of the manuscript.

Publications

Vickerstaff V, Ambler G, Bunce C, Xing W, Gazzard G, LiGHT Trial Study Group. Statistical analysis plan for the Laser-1st versus Drops-1st for Glaucoma and Ocular Hypertension Trial (LiGHT): a multi-centre randomised controlled trial. Trials 2015;16:517. https://doi.org/10.1186/s13063-015-1047-9

Gazzard G, Konstantakopoulou E, Garway-Heath D, Barton K, Wormald R, Morris S, et al. Laser in Glaucoma and Ocular Hypertension (LiGHT) trial. A multicentre, randomised controlled trial: design and methodology. Br J Ophthalmol 2018;102:593–8. https://doi.org/10.1136/bjophthalmol-2017-310877

Konstantakopoulou E, Gazzard G, Vickerstaff V, Jiang Y, Nathwani N, Hunter R, et al. The laser in glaucoma and ocular hypertension (LiGHT) trial. A multicentre randomised controlled trial: baseline patient characteristics. Br J Ophthalmol 2018;102:599–603. https://doi.org/10.1136/bjophthalmol-2017-310870

Gazzard G, Konstantakopoulou E, Garway-Heath D, Garg A, Vickerstaff V, Hunter R, et al. Selective laser trabeculoplasty versus eye drops for first-line treatment of ocular hypertension and glaucoma (LiGHT): a multicentre randomised controlled trial. Lancet 2019;393:1505–16.

Data-sharing statement

All available data can be obtained from the corresponding author.

Patient data

This work uses data provided by patients and collected by the NHS as part of their care and support. Using patient data are vital to improve health and care for everyone. There is huge potential to make better use of information from people’s patient records, to understand more about disease, develop new treatments, monitor safety, and plan NHS services. Patient data should be kept safe and secure, to protect everyone’s privacy, and it’s important that there are safeguards to make sure that it is stored and used responsibly. Everyone should be able to find out about how patient data are used. #datasaveslives. More about the background to this citation can be found here: https://understandingpatientdata.org.uk/data-citation.

Disclaimers

This report presents independent research funded by the National Institute for Health Research (NIHR). The views and opinions expressed by authors in this publication are those of the authors and do not necessarily reflect those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care. If there are verbatim quotations included in this publication the views and opinions expressed by the interviewees are those of the interviewees and do not necessarily reflect those of the authors, those of the NHS, the NIHR, NETSCC, the HTA programme or the Department of Health and Social Care.

Diagnosis of open-angle glaucoma and treatment requirements

Primary OAG is defined as an open drainage angle (no iridotrabecular contact on non-indentation gonioscopy in primary position, trabecular meshwork visible over 360°), with no secondary causes (such as trauma):

• and reproducible glaucomatous VF defects as tested by the Swedish Interactive Threshold Algorithm on the Humphrey Visual Field Analyser (i.e. reproducible defect, in at least, of two or more contiguous points with a p-value < 0.01 loss or greater, or three or more contiguous points with a p-value < 0.05 loss or greater, or abnormal Glaucoma Hemifield Test)

• or GON with localised absence of the neuroretinal rim, or cup disc ratio of ≥ 0.7, or asymmetry of cup disc ratio of ≥ 0.2 in similar-sized eyes/optic discs

• and deemed to require treatment in the opinion of the treating (fellowship-trained) glaucoma specialist.

Subjects with pseudo-exfoliation are eligible (as for EMGT I82).

Subjects with GON and IOP in the normal range are therefore eligible. ‘Phasing’ (diurnal IOP pressure measurements) will be performed at the discretion of the treating clinician, and, if performed, the maximum IOP recorded will be used as that day’s measurement.

Diagnosis of ocular hypertension and treatment requirements

Ocular hypertension with IOP > 21 mmHg and requiring treatment as per NICE guidelines. 1 NICE OHT guidelines treat four categories of OHT on the basis of CCT and age (the rest are monitored for 3–5 years).