Journal Pre-proof LiGHT trial: 6-year results of primary selective laser trabeculoplasty versus eye drops for the treatment of glaucoma and ocular hypertension

Conflict of Abstract Purpose: The LiGHT trial has shown selective laser trabeculoplasty (SLT) to be clinically and cost-effective 2 as a primary treatment of open-angle glaucoma (OAG) and ocular hypertension (OHT) at 3 years. This paper 3 reports health-related quality of life (HRQL) and clinical effectiveness of initial treatment with SLT compared 4 to intra-ocular pressure (IOP) lowering eye drops, after 6 years of treatment. 5 Design: Prospective multicentre randomized controlled trial. 6 Participants: Treatment-naïve eyes with OAG or OHT, initially treated with SLT or IOP-lowering drops. 7 Methods: Patients were randomly allocated to initial SLT or eye drops. Eye specific target IOP and monitoring 8 intervals were based on international guidelines. After the initial 3 years of the trial, patients in the SLT arm 9 were permitted a 3 rd SLT if necessary; patients in the drops arm were allowed SLT as a treatment switch or 10 escalation. Analysis was by intention to treat. This study is registered at controlled-trials.com 11 (ISRCTN32038223). 12

Selective laser trabeculoplasty (SLT) was endorsed by the United States Food and Drug Administration for the 26 treatment of glaucoma in 2001. SLT has since increasingly been adopted as an alternative to IOP-lowering eye 27 drops, but until recently data on its efficacy as a sole treatment were scarce. 1,2 Recent studies have compared 28 SLT to monotherapy, which does not reflect routine clinical practice where IOP is treated to target. As a result, 29 a Cochrane systematic review called for more research into the efficacy of SLT compared to contemporary 30 medication regimens. 3 31 The Laser in Glaucoma and Ocular Hypertension (LiGHT) trial is a multicentre randomized controlled trial 32 comparing initial treatment with SLT to initial treatment with IOP-lowering eye drops for treatment-naive 33 patients with OAG or OHT, assessing health-related quality of life (HRQL), cost-effectiveness, and clinical 34 efficacy after 3 years. 4 In 2019, the LiGHT trial reported that initial treatment of ocular hypertension (OHT) or 35 open-angle glaucoma (OAG) with SLT is more cost-effective than initial treatment with contemporary IOP-36 lowering eye drops after 3 years, whilst also providing drop freedom to 74.2% of patients, a reduced number of 37 glaucoma surgeries and very low rates of adverse events. 5  Glaucoma is a long-term condition requiring life-long treatment; average life-expectancy at initial diagnosis of 43 glaucoma is 9-13 years 9 and mean life expectancy after trabeculectomy is 7.5 years. 10 While we previously 44 reported that initial treatment with SLT offers drop-freedom to nearly 75% of LiGHT patients for at least 3 45 years, longer term IOP control following initial SLT and additional SLT could further prolong drop-freedom 46 and reduce the requirement for intense medical or surgical treatment over patients' lifetime. Such potential may 47 Symptom Scale (GSS), 24 patient reported visual function using the Glaucoma Quality of , 132 clinical effectiveness and safety of the treatment arms. Adverse events were classified and reported according 133 to local standard operating procedures and good clinical practice guidelines. 25 134

Statistical analysis 135
The statistical analysis plan is described in detail elsewhere. 26 In summary, the primary outcome was analysed 136 using linear regression with terms for randomisation group, baseline EQ-5D, stratification factors (diagnosis 137 and centre), baseline IOP, and number of eyes affected at baseline. The unit of analysis was the patient. If a 138 patient had both eyes in the study, baseline severity and IOP were based on the worse eye, where the worst eye 139 was defined using VF MD at baseline. Several sensitivity analyses were performed to verify the results of this 140 primary analysis (details provided in Appendix 1). In addition, mixed effects models were used to analyse the 141 EQ-5D measurements recorded at all time-points to investigate possible changes in treatment effect over the 72 142 months (using interaction terms between the randomisation group and time) and to estimate the average 143 treatment effect over the 72-month follow-up period. The secondary outcomes were analysed using similar 144 regression methods to those described above. All analyses were performed on an intention-to-treat (ITT) basis 145 with participants analysed according to the group to which they were randomised. Kaplan-Meier plots were 146 used to summarise disease progression, time to glaucoma surgery and phacoemulsification, and the log-rank 147 test was used to compare these outcomes. Eyes were compared with respect to visits at target and number of 148 clinical visits using mixed effects logistic regression and Poisson regression models respectively. Eyes were 149 also compared with respect to the remaining measurement of pathway effectiveness and visual function 150 variables using the t-test for numerical outcomes and the chi-squared test (or Fisher's exact test when numbers 151 were small) for categorical outcomes. The chi-squared test and Fisher's exact test were also used to compare 152 the number of reported adverse and serious adverse events. All analyses were performed in Stata version 17 153 The study was conducted in accordance with good clinical practice guidelines (GCP) and adhered to the tenets 155 of the Declaration of Helsinki. Ethical approval was granted by local boards. All patients provided written 156 informed consent before participation. An independent data and safety monitoring committee was appointed by 157 procedures for the duration of the trial. The LiGHT trial is registered at www.controlled-trials.com 159 (ISRCTN32038223)  and  the  protocol  can  be  accessed  at  160 https://www.journalslibrary.nihr.ac.uk/programmes/hta/0910440/#/. 161

Baseline data 163
Of the 692 patients who completed 3 years of the LiGHT trial, 633 (91.5%) entered the 3-year extension (from 164 36 to 72 months); 313 patients (547 treated eyes) had initially received SLT and 320 patients (549 eyes) had 165 initially commenced treatment with IOP-lowering eye drops ( Figure 1). There were 86 protocol violations or 166 deviations; 30 took place during the first 3 years and 56 during the extension (36 to 72 months), the latter relating 167 to the COVID-19 pandemic. Of the 59 patients not continuing into the extension, 29 came from a single centre 168 that chose not to continue in the study (Appendix 2). A total of 524 patients completed the trial extension (82.8% 169 of those entering the extension phase, 73% of those initially randomised). 170 Presented results refer to the sample of patients who entered the LiGHT trial extension (36 to 72 months); this 171 sample was representative of the original trial participants and maintained the balance of the allocation groups 172 achieved by randomisation. Baseline (month 0) patient and eye characteristics of the patients who participated 173 in the extension phase were similar between the two groups (Table 1, Appendix 2); 493 patients (77.9%) were 174 diagnosed with OAG in at least 1 eye and 140 patients (22.1%) were diagnosed with OHT. The treatment groups 175 had similar average EQ-5D, GUI and GQL-15 scores at baseline (month 0) (Table 2); the medication group had 176 slightly higher average GSS scores at baseline , similarly to the original trial data. 5 At 36 months (start of the 177 extension) the two groups had average EQ-5D, GUI, GSS and GQL-15 scores that continued to be similar to 178 the scores recorded in the first 3 years of the trial. 5 179 Of the 320 patients allocated to medication, 112 (176 eyes, 35% of patients) decided to receive SLT immediately 180 or shortly after the end of the 3-year monitoring period. Of those, 70 patients (115 eyes) had SLT as a treatment 181 switch i.e., to reduce medication load, and 29 patients (35 eyes) had SLT as a treatment escalation due to 182 uncontrolled IOP and/or disease progression. Thirteen patients (26 eyes) had SLT as a treatment escalation in 183 J o u r n a l P r e -p r o o f one eye and as a treatment switch in the other eye. Of the 112 patients who received SLT after 36 months, 94 184 (83.9%) completed the trial extension to 72 months. 185

Health-related quality of life 186
The mean values for the HRQL questionnaires across the 72 months of the trial are shown in in Figure 2. Based 187 on an intention to treat analysis, there was no significant difference in HRQL between the two treatments at 72 188 months for the EQ-5D, GUI and GQL-15 (Table 3); the eye drops group had an average EQ-5D score of 0.89 189 (SD 0.14), compared with 0.90 (SD 0.14) in the SLT group (adjusted mean difference [selective laser 190 trabeculoplasty-eye drops] 0·01, 95% CI −0·01 to 0·04, p=0·18). These results were confirmed in sensitivity 191 analyses (results not shown, see Appendix 1). The average GUI score at 72 months in the SLT group was 0.90 192 (SD 0.14) compared with 0.88 (SD 0.13) for the eye drops group (adjusted mean difference 0·01, 95% CI −0·01 193 to 0·03). Mean GQL-15 scores were also similar between the two groups (20.80 for the SLT group and 20.57 194 eye drops, adjusted mean difference −0.13, 95% CI −1.57 to 1·31). For the GSS, the medication group had 195 worse scores at 72 months with a mean score of 81.3 (SD 17.3) compared to 83.6 (SD 18.1) for the SLT group 196 (adjusted mean difference 3.3, 95% CI 0.54 to 6.0), however this was the only timepoint at which a noticeable 197 difference was observed. Repeated measures analysis for the secondary HRQL outcomes (GUI, GSS,  showed comparable outcomes between the two groups over the course of the trial (Appendix 3). When excluding 199 the eyes that received SLT after the 36-month time point (n=176), mean scores for all HRQL questionnaires 200 were similar between the two groups (Table 3). 201

Measurements of treatment effectiveness and visual function 202
At 72 months, 537 patients (267 in drops arm and 270 in SLT arm) and 930 eyes (460 in the drops arm and 470 203 eyes in the SLT arm) were available for analysis of clinical outcomes (Error! Reference source not found.). 204 Overall, 94.2% of eyes initially treated with SLT were at target at 72 months and target IOP was achieved at 205 92.8% of visits, compared to 94.7% of eyes and 93.2% of visits for eyes initially treated with medication. Fewer 206 eyes initially treated with SLT demonstrated progression from OHT to OAG or deterioration of OAG, compared 207 to eyes initially treated with eye drops (19.6% vs 26.8%, respectively, p=0.006) (Table 4, Figure 3). 208 Drop free IOP control at 72 months, obtained without incisional surgery, was achieved in 69.8% of eyes initially 209 treated with SLT, compared to 18.0% of eyes initially treated with IOP-lowering eye drops. Of the eyes initially 210 J o u r n a l P r e -p r o o f treated with SLT and being drop and surgery free at 6 years, 90% (295 eyes) needed up to 2 SLT treatments in 211 total. Of the eyes initially treated with eye-drops and being drop free at 72 months, 79.5% (66 eyes) had switched 212 to SLT and 20.5% had either cataract surgery alone or cataract surgery and SLT. At 72 months, 61.2% of eyes 213 initially treated with eye-drops were using 1 or 2 medications, compared to 18.5% of eyes initially treated with 214

SLT. 215
Target IOP was revised in 85 eyes initially treated with SLT and in 89 eyes initially treated with IOP-lowering 216 eye drops. Target IOP was revised downwards on 50 occasions in eyes initially treated with SLT and on 65 217 occasions in eyes initially treated with IOP-lowering eye drops and upward on 40 and 31 occasions, respectively. 218 Eyes initially treated with SLT needed fewer trabeculectomies (13 eyes, 2.4%) compared to eyes initially treated 219 with eye drops (32 eyes, 5.8%) (Table 4, Figure 4, p<0.001) and fewer phacoemulsifications (57 compared to 220 95, respectively, p=0.03) (Table 4, Figure 5). Of the 32 eyes that needed a trabeculectomy during trial's 6-year 221 duration, 11 eyes initially treated with drops had a trabeculectomy during the first 3 years of the trial; none of 222 the eyes initially treated with SLT required a trabeculectomy during the initial 3 years of the trial. During the 223 extension of the trial, i.e. from 3 to 6 years, minimally invasive glaucoma surgery (MIGS) was performed in 11 224 eyes of 6 patients initially treated with IOP-lowering eye drops (all were angle procedures; no MIGS was 225 performed in eyes initially randomised to SLT). This may have resulted in fewer trabeculectomy surgeries in 226 the drops arm, but is not expected to have affected the reported statistical and clinical differences in incisional 227 glaucoma surgery between the treatment arms. 228 Eyes initially treated with SLT had higher IOP at 72 months compared to eyes initially treated with IOP-229 lowering eye drops (16.3mmHg vs 15.4mmHg, respectively, p<0.001); however, VF MD loss and visual acuity 230 at 72 months were similar between the two groups (-4.0dB vs -3.9dB, and 0.1 vs 0.1, respectively, both p>0.05) 231 (Table 4, Appendix 4). Patients initially treated with SLT needed a total of 5175 visits over 72 months and 232 patients initially treated with eye-drops needed 4970 visits. Excluding the 2-week post-laser visits resulted in 233 4678 visits for the SLT group compared to 4852 for the eye-drops group. 234

Safety 235
There were no sight-threatening complications of SLT and no clinically identifiable corneal changes throughout 236 the trial (Table 5) were reported in the group initially treated with IOP-lowering eye drops (1470 ocular adverse events were 239 reported by 271 patients) compared to the group initially treated with SLT (897 ocular adverse events by 224 240 patients) (Table 5). Serious adverse events were similar overall between the two groups (180 events in 110 241 patients initially treated with eye-drops; 209 events in 107 patients initially treated with SLT), with pulmonary 242 and cardiac events being balanced between the two groups (Table 5). 243

Discussion 244
In 2019, the LiGHT trial reported that initial treatment with SLT provided newly diagnosed OHT and OAG 245 eyes with predominantly drop free IOP control (78.2% of eyes after 3 years) and a reduced need for glaucoma 246 and cataract surgery, compared to initial treatment with IOP-lowering eye drops 5 . Data from this 3-year trial 247 also indicated that eyes initially treated with SLT may demonstrate less frequent progression to more advanced 248 stages of glaucoma and a further VF analysis indicated that more eyes initially treated with topical medical 249 therapy undergo rapid VF progression compared to eyes initially treated with SLT. 27 250 The LiGHT trial was extended to a total of 6 years to provide longer-term, pragmatic treatment outcome data. 251 Patients within five UK settings, initially treated with IOP-lowering eye drops were permitted to have SLT to 252 reduce medication load, avoid increasing medication load or delay surgery. Patients initially treated with SLT 253 were allowed a 3 rd and final SLT, before escalating to IOP-lowering eye drops. Data after 6 years of treatment 254 indicate statistically significant lower rates of disease progression and reduced need for glaucoma and cataract 255 surgery for eyes initially treated with SLT. Drop free IOP control and safety of SLT as a 1 st line treatment for 256 OHT and OAG are confirmed after 6 years of careful, protocolised monitoring and treatment. 257 SLT allowed successful drop free IOP control in nearly 70% of the eyes after 6 years of treatment. Τhis is only 258 slightly reduced form 78% of eyes not needing topical therapy at 3 years and an important outcome for long-259 term glaucoma and OHT management; of the initial SLT eyes which were drop free, 90% had only one or two 260 SLT treatments. IOP-lowering eye drops come with, sometimes significant, adverse effects, affecting 261 trabeculectomy outcomes, increasing expenditure for healthcare systems and/or patients, 28,29 and often leading 262 to non-adherence. 30 Drop-freedom was achieved in nearly a fifth of eyes initially treated with eye drops, 263 predominantly by switching to SLT (79.5%) alone or after undergoing SLT and/or cataract surgery (20.5%). 264

J o u r n a l P r e -p r o o f
The LiGHT trial reports 70% of eyes being drop free following 6 years of treatment, whereby IOP had to be 265 reduced by a minimum of 20% from pre-treatment IOP (and at least by 30% for moderate and severe OAG) and 266 below 25mmHg for OHT, below 21mmHg for mild OAG, below 18mmHg for moderate OAG and below 267 15mmHg for severe OAG. 4,15 Absolute IOP reduction has been reported elsewhere 31 ; reporting absolute IOP 268 reduction at 6 years has limited usefulness since no washout was preformed and a proportion of eyes were on 269 IOP-lowering topical medical treatment. Success rates for SLT have been published using various definitions. 1,32 270 A large US-based retrospective study has clearly indicated that reported success rates are heavily influenced by 271 disease severity and co-morbidities of the included populations, concluding that SLT can be an effective means 272 of prolonging medication-free IOP-control, 33 but lower SLT success rates have been reported for less carefully 273 selected eyes already on medication 34 . 274 LiGHT used eye-specific target IOPs, which could be revised in the absence of evident deterioration 4 ; this has 275 been suggested to potentially drive the reported outcomes. 35 The European Glaucoma Society Guidelines 276 recommend clinicians consider upward revision of target pressure in stable patients, when the initial target has 277 not been reached. 36 In LiGHT, Target IOP was reassessed using decision support software and applied to both 278 treatment arms, according to pre-set criteria, 37 when VF and disc imaging analysis provided evidence of disease 279 stability accounting for inter-visit IOP measurement variation. 38 A risk-dependent upper limit was set, at which 280 surgery might be offered even in the absence of progressive glaucomatous optic neuropathy. Here we report the 281 number of upward and downward IOP revisions, which are comparable between the two treatment arms and 282 are, therefore, unlikely to affect the reported outcomes. 283 The LiGHT trial has carefully and objectively monitored patients in a pragmatic manner across 5 NHS centres, 284 retaining more than 80% of participants after 6 years of treatment. Data reported by the LiGHT trial are an 285 accurate representation of realistic and complete glaucoma management for newly-diagnosed, previously 286 untreated eyes with OHT/OAG; these data have supported the update of the American, European and UK-NICE 287 glaucoma management guidelines. 6-8 The LiGHT trial population consisted of a large proportion of OHT and 288 mild OAG eyes, for which IOP reduction targets are less stringent than those for more advanced disease. Eyes 289 with advanced OAG will often require more intense treatment, whilst initial intervention might differ from that 290 recommended for early disease. 39 291 J o u r n a l P r e -p r o o f Adding to the evidence from the LiGHT trial, the Glaucoma Intensive Treatment Study (GITS) 40 has reported 292 favourably on the use of SLT as an adjunctive therapy for patients with OAG over 3 years and the West Indies 293 Glaucoma Laser Study (WIGLS) reported that SLT monotherapy safely provides 78% of Afro-Caribbean eyes 294 with at least 20% IOP reduction for 12 months. 41 SLT was also recently shown to be an ideal therapeutic 295 approach in situations where frequent monitoring visits and treatment changes are difficult. 42 With 90% of the 296 drop-free eyes initially treated with SLT needing a maximum of 2 SLT treatments over 6 years and 55.5% 297 requiring only a single SLT treatment, there is great potential for treating patients with SLT in such situations. 298 Data published previously have indicated that initial treatment with SLT might delay progression of OHT and 299 OAG; data from the first 3 years of treatment indicated a 2% difference in eyes progressing and VF analysis 300 suggests more eyes initially treated with IOP-lowering eye drops undergo rapid VF progression compared to 301 eyes first treated with SLT. 5,27 After 6 years of treatment, eyes initially treated with SLT demonstrated reduced 302 objectively defined progression compared to IOP-lowering eye drops; this was achieved despite eyes initially 303 treated with IOP-lowering eye drops achieving lower IOP at 6 years, possibly suggesting other protective roles 304 of SLT. Differences in progression between the two treatment arms also influence the rates of incisional 305 glaucoma surgery. Eyes initially treated with SLT needed fewer trabeculectomies, supporting original trial data. 5 306 For the first three years after initial treatment, no trabeculectomies were needed in eyes receiving initial SLT, 307 whilst at 6 years there were almost three times fewer eyes initially treated with SLT needing a trabeculectomy, 308 compared to eyes initially treated with IOP-lowering eye drops. Excess surgeries in eyes initially treated with 309 eye drops might have led to the slightly lower IOP at 72 months, compared to eyes initially treated with SLT. 310 These data have significant implications for patients and healthcare systems. Trabeculectomy is performed on 311 average 10 years after initial diagnosis and average life expectancy post glaucoma diagnosis is 9-13 years 9,43,44 ; 312 SLT can delay and potentially obviate the need for glaucoma surgery for a proportion of patients. 313 SLT also leads to a reduced need for cataract surgery; at least 50% more eyes initially treated with eye drops 314 needed a cataract surgery during the 6-year course of the LiGHT trial compared to eyes initially treated with 315 SLT, supporting evidence from the Early Manifest Glaucoma Trial on a greater need for surgical cataract 316 removal in eyes treated with IOP lowering eye-drops. 45 317 J o u r n a l P r e -p r o o f LiGHT trial, generic and disease specific HRQL tools indicated that patients using drops had comparable HRQL 319 to those who received initial SLT and these findings are further supported by the LiGHT trial extension to 6 320 years. The single time-point where SLT appeared to lead to better GSS scores was the 72 months and is unlikely 321 to have clinical significance. SLT has also been compared to timolol monotherapy using the WHO/PBD-VF20 322 vision-related quality of life instrument, which also revealed comparable results between the two treatment 323 modalities. 42 Over the recent years the sensitivity of existing QoL tools to capture changes and their suitability 324 as primary outcomes in clinical trials have been questioned. 46 325 The safety profile of SLT remains very good, with no sight threatening complications. IOP rose more than 5 326 mmHg from pre-treatment IOP in only 1% of treated eyes and, of these, only 1 eye needed treatment. Other 327 adverse events were comparable between the two groups. SLT has been shown to be a safe alternative to eye-328 drops in areas where advanced glaucoma is more common and where treatment resources and access to these 329 are limited. 42 The proven safety of SLT in such areas can rapidly transform glaucoma treatment and prevent 330 sight loss. 331

Conclusion 332
After 6 years of treatment and monitoring, SLT safely offers IOP control without the need for medical or surgical 333 treatment in more than 70% of OHT and OAG eyes, whilst demonstrating reduced progression rates and a 334 reduced need for glaucoma and cataract surgery. SLT is now the recommended 1 st line treatment for OAG and 335 OHT by National Institute for Health and Care Excellence (NICE) 8 in the UK and is listed as a 1 st line treatment 336 in the EU and the USA, alongside IOP-lowering eye drops.   randomised twice due to IT failure, where the initial randomisation was not visible and subsequently a second 473 randomisation was carried out. One of these patients was initially randomised to medication but was subsequently 474 randomised to, and received, SLT. The other was initially randomised to SLT but was subsequently randomised to, and 475 received, medication. These patients are included in the diagram according to the second randomisations.

SLT (n=313) Age (years) -Mean (SD)
63.  VF=visual field. MD=mean deviation. ¶Conversion of OHT to OAG required a sign of progression derived from the decision support software and verification by a consultant ophthalmologist; OAG progression OAG required a sign of progression derived from the decision support software; 4 OHT eyes had a single OAG diagnosis during the trial and these were assumed to be errors. See Figure 3 for a full statistical comparison. An analysis of progression by disease severity is available in Appendix 4 ‡ Minimally invasive glaucoma surgery combined phacoemulsification was performed in 11 eyes of 6 patients initially treated with IOPlowering eye drops during the extension of the trial. * See Figure 4 and Figure 5 for a full statistical comparison. **Protocol deviation; 3 eyes of 2 patients. ***Target IOP was reassessed when VF and sequential disc imaging provided evidence of disease stability; IOP was revised following a decision support software recommendation, according to pre-set criteria 21 .
J o u r n a l P r e -p r o o f  Table 1: Adverse events. Adverse events. a: includes excessive lash growth, peri-ocular pigmentation, change in iris colour. b: includes peri-ocular skin rash c: Includes ocular irritation, discomfort, dry eye, retinal haemorrhages, flashes, floater, conjunctivitis, blepharitis, vascular occlusions, diabetic retinopathy, macular J o u r n a l P r e -p r o o f