A Multicountry Comparison of Real-World Management and Outcomes of Polypoidal Choroidal Vasculopathy

A Multicountry Comparison of Real-World Management and Outcomes of Polypoidal Choroidal Vasculopathy

Accepted Manuscript A multi-country comparison of real world management and outcomes of polypoidal choroidal vasculopathy – Fight Retinal Blindness! C...

1MB Sizes 0 Downloads 5 Views

Accepted Manuscript A multi-country comparison of real world management and outcomes of polypoidal choroidal vasculopathy – Fight Retinal Blindness! Cohort Kelvin Yi Chong Teo, David M. Squirrell, Vuong Nguyen, Gayatri Banerjee, Amy Cohn, Daniel Barthelmes, Chui Ming Gemmy Cheung, Mark Gillies PII:

S2468-6530(18)30314-2

DOI:

https://doi.org/10.1016/j.oret.2018.11.003

Reference:

ORET 427

To appear in:

Ophthalmology Retina

Received Date: 20 April 2018 Revised Date:

31 October 2018

Accepted Date: 1 November 2018

Please cite this article as: Chong Teo K.Y., Squirrell D.M, Nguyen V., Banerjee G., Cohn A., Barthelmes D., Gemmy Cheung C.M. & Gillies M., A multi-country comparison of real world management and outcomes of polypoidal choroidal vasculopathy – Fight Retinal Blindness! Cohort, Ophthalmology Retina (2018), doi: https://doi.org/10.1016/j.oret.2018.11.003. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT 1

Title: A multi-country comparison of real world management and outcomes of polypoidal

2

choroidal vasculopathy – Fight Retinal Blindness! Cohort

3 4

Running head: Comparison of PCV subtype in the FRB cohort

RI PT

5

Authors and affiliations: Kelvin Yi Chong Teo1,2,3, David M Squirrell4, Vuong Nguyen5,

7

Gayatri Banerjee5,6, Amy Cohn7, Daniel Barthelmes8, Chui Ming Gemmy Cheung1,2,9.10,

8

Mark Gillies4

9

SC

6

1 Singapore National Eye Centre, Singapore

11

2 Singapore Eye Research Institute, Singapore

12

3 Sydney Eye Hospital Foundation, Sydney Eye Hospital, Sydney, Australia

13

4 Department of Ophthalmology, University of Auckland, Auckland, New Zealand

14

5 The University of Sydney, Sydney, Australia

15

6 Nepean Valley Eye Surgeons, Australia

16

7 Centre For Eye Research Australia, Royal Victorian Eye and Ear Hospital

17

8 University Hospital Zurich, University of Zurich, Zurich, Switzerland

18

9 Department of Ophthalmology, Yong Loo Lin School of Medicine, National University of

19

Singapore, Singapore

20

10 Ophthalmology Academic Clinical Program, Duke-NUS Graduate Medical School,

21

Singapore

22 23 24 25

AC C

EP

TE D

M AN U

10

ACCEPTED MANUSCRIPT 26 27 28 29 30 31 32

Financial Support: The FRB! project has been supported by grants from the Royal Australian NZ College of Ophthalmologists’ Eye Foundation (2007 -2009), the National Health and Medical Research Council, Australia (NHMRC 2010-2012), the Macular Diseases Foundation, Novartis and Bayer. These supporting organizations had no role in the design or conduct of the research.

33

data for this analysis

34

Correspondence to:

35

Dr. Kelvin Yi Chong Teo

36

The Save Sight Institute,

37

Sydney Medical School, The University of Sydney, Sydney, Australia.

38

Singapore Eye Research Institute

39

Singapore National Eye Centre

40

11 Third Hospital Avenue, Singapore 168751

41

Phone: (65) 69881460, Fax: (65) 62263995

42

Email: [email protected]

47 48 49 50 51 52 53

EP

46

SC

Original Manuscript not previously presented

Conflict of Interest: no conflicting relationship exists for any author

AC C

45

M AN U

TE D

43 44

RI PT

Conflict of interest: Gillies and Barthelmes are inventors of the software used to collect the

Financial Support: None

ACCEPTED MANUSCRIPT 54 55 56 57

RI PT

58 59 Abstract

61

Objective:

62

To compare the 12-month real-world visual and disease activity outcomes of eyes with

63

polypoidal choroidal vasculopathy (PCV), treated with a combination of photodynamic

64

therapy and anti-vascular endothelial growth factor (anti-VEGF) injections [The combination

65

group] versus those eyes treated with anti-VEGF monotherapy alone with rescue PDT being

66

used as required [The monotherapy group].

67

Design: Database comparative observational study

68

Participants: Eyes with PCV as graded in Fight Retinal Blindness! database from Australia,

69

New Zealand, Singapore and Switzerland.

70

Methods:

71

Clinical information from a multi-site, international registry of nAMD was analysed with

72

intention-to-treat approach.

73

Results: There were 41 and 152 eyes that received combination therapy and anti-VEGF

74

monotherapy respectively. All anti-VEGF agents were pooled, and bevacizumab comprised

75

of 66.1% of injections administered. The adjusted mean (CI) change in visual acuity between

76

the combination group and monotherapy group at 12 months was +16.9 (10.6 - 23.3) letters

77

and +8.2 (5.2 – 11.3) letters respectively (p=0.02). Proportion of inactive lesions and mean

78

(CI) time to inactivity was 85.3% and 80.7 (62.8- 98.5) days in the combination group

AC C

EP

TE D

M AN U

SC

60

ACCEPTED MANUSCRIPT compared to 76.8% and 150.4 (132.8 – 168.0) days in the monotherapy group (p=0.01). The

80

mean number of injections (CI) of anti-VEGF between the combination and monotherapy

81

group was 4.3 (3.6 – 5.2) and 6.4 (5.9 – 6.9) respectively (p=0.01).

82

Conclusions:

83

The real-world outcomes for treatment of PCV showed larger gains in vision, higher

84

proportion of inactive lesions, quicker time to inactivity and fewer injections administered in

85

the combination group compared to the monotherapy group. These findings are consistent

86

with current evidence reporting the advantages of combination therapy for PCV.

SC

RI PT

79

87

AC C

EP

TE D

M AN U

88

ACCEPTED MANUSCRIPT 89 Introduction

91

Age related macular degeneration (AMD) is one of the leading causes of blindness

92

worldwide.1 In its exudative or wet form, the presence of a choroidal neovascular network

93

(CNV) results in an exudative maculopathy and rapid vision loss. Intravitreal injections of

94

anti-vascular endothelial growth factor (anti-VEGF) agents have become the mainstay

95

treatment for CNV in AMD and can produce good outcomes in all CNV subtypes.1-3

96

Polypoidal choroidal neovascularization (PCV) is a particular subtype of neovascular AMD

97

characterized by polypoidal lesions arising from terminal ends of branching vascular

98

networks which are best diagnosed on indocyanine green angiography (ICGA).4 50-80% of

99

large serosanguinous maculopathy in neovascular AMD can be attributed to PCV5,6, which

M AN U

SC

RI PT

90

may also be accompanied by extensive subretinal or sub-retinal pigment epithelium (sub-

101

RPE) haemorrhage, often resulting in acute and severe vision loss. Compared to the typical

102

subtypes of CNV in AMD, the PCV subtype has a predilection for pigmented races,

103

occurring in 22-62% of AMD amongst Asians7-11 compared to 5-20% in Caucasians12,13 and

104

also tends to afflict younger patients.

EP

105

TE D

100

There is currently no consensus on the best treatment for PCV. Early studies reported that

107

despite being a subtype of CNV in AMD, PCV did not seem to respond as well to anti-VEGF

108

monotherapy14-16, while results from recent randomised controlled trials (RCT) reported an

109

improvement in visual outcomes with anti-VEGF therapy.10,17 The combination of a vaso-

110

occlusive treatment like photodynamic therapy with verteporfin (PDT) in combination with

111

an anti-VEGF injection in PCV is an attractive treatment strategy as it potentially targets both

112

the polypoidal lesion as well as the branching vascular network components of the PCV. The

113

EVEREST II trial reported that the use of PDT in combination with ranibizumab treatment

AC C

106

ACCEPTED MANUSCRIPT was more effective than ranibizumab monotherapy in treating PCV. It reported superior

115

visual acuity gains (8.3 vs 5.1 letters, p=0.013) and polyp closure rates (69.3% vs 34.7%,

116

p<0.001) in patients treated with combination therapy as compared to ranibizumab

117

monotherapy.18 The PLANET study, on the other hand, demonstrated that aflibercept

118

monotherapy resulted in visual gains of 10.7 letters, with no additional benefit from rescue

119

PDT.19

RI PT

114

120

These two randomized controlled trials currently represent the best evidence for the treatment

122

of PCV. In our study, we aimed to use multi-site observational data from the FRB! registry,

123

to compare the 12-month real-world treatment outcomes of combination PDT and anti-VEGF

124

therapy with anti-VEGF monotherapy in treatment naive eyes with PCV.

M AN U

SC

121

125

AC C

EP

TE D

126

ACCEPTED MANUSCRIPT Methods

128

We performed an observational study of the outcomes of treatment-naïve eyes diagnosed

129

with PCV in routine clinical practice tracked in the Fight Retinal Blindness! (FRB!)

130

database.20 Briefly, the FRB! system collects data from each clinical visit including the

131

visual acuity (VA), lesion subtype (predominantly classic (PC), minimally classic (MC),

132

occult (OC) and PCV), greatest linear diameter of the lesion, lesion activity, treatment

133

administered and any ocular adverse events. Visual acuity recorded was whichever reading

134

was best: uncorrected, corrected or pinhole. Visual acuity scores were expressed as the

135

number of letters read on a logarithm of the minimum angle of resolution (logMAR) VA

136

chart. Treating physicians determined all management decisions including frequency of

137

visits and treatment modality, thereby reflecting real-world practice. Institutional ethics

138

approval was obtained from the Human Research Ethics Committees of the University of

139

Sydney, the Royal Victorian Eye and Ear Hospital, the Royal Australian and New Zealand

140

College of Ophthalmologists, the University Hospital, Zurich, and Singhealth, Singapore.

141

The research described adhered to the tenets of the Declaration of Helsinki. Practitioners

142

located in Australia, New Zealand, Switzerland and Singapore contributed the data for the

143

study. This paper followed the STROBE checklist items for reporting observational study

144

data.21

SC

M AN U

TE D

EP

AC C

145

RI PT

127

146

Patient Selection and Groups

147

Patients included in the analysis were treatment naïve for neovascular AMD and had lesion

148

type graded in the FRB! system as “PCV”. Patients entered in the database after the July 2016

149

was excluded from the study to allow for at least 12 months of potential follow-up. PCV was

150

diagnosed by each individual treating physician based on ICGA.

151

ACCEPTED MANUSCRIPT 152 We compared the outcomes of eyes that received initial PDT as performed by the treating

154

physician (within the first 90 days) in combination with anti-VEGF therapy, the “combination

155

group”, with those that received anti-VEGF monotherapy, the “monotherapy” group. The

156

criteria and scheduling of treatment reflected real-world practice where, due to clinical or

157

logistical constraints, initial PDT treatment may have been administered within the first 3

158

monthly anti-VEGF injections rather than at the first visit. Within the cohorts studied,

159

bevacizumab, ranibizumab and aflibercept were used to treat PCV and the different anti-

160

VEGF agents were combined for the analysis. The monotherapy group also included eyes

161

that received rescue PDT (from 90 days after the first visit), since it was likely that the

162

practitioner had originally planned to treat these eyes with anti-VEGF monotherapy. The

163

observations in these eyes in which PDT was deferred were censored immediately before the

164

application of rescue PDT and carried forward as part of the monotherapy group.

165

Eyes were “non-completers” if they did not complete 12 months of follow-up. The primary

166

analysis was performed based on the principle of intention-to-treat (ITT). All eyes were

167

considered for final analysis with the last observation carried forward (LOCF) method used

168

for non-completers.

SC

M AN U

TE D

EP

AC C

169

RI PT

153

170

Study Measurements

171

Patient age (years), sex, VA in logMAR letters and lesion size (micrometres) were recorded

172

at the time of first treatment. All treatments were recorded, along with VA, lesion activity,

173

and ocular adverse events at each visit. Lesion activity status was graded by the treating

174

physician based on findings from clinical examination, optical coherence tomography, or dye

175

angiography, alone or in combination, at each visit.

176

ACCEPTED MANUSCRIPT Outcome Measurements

178

The primary study outcome was the mean change in VA of each group over 12 months after

179

initiating treatment. Secondary visual outcomes were the final vision at 12 months, the

180

proportion of eyes with 15 letter loss or gain and the proportion of eyes with VA better than

181

6/12. Secondary outcomes include the proportion of eyes that achieved CNV inactivity and

182

time to inactivity of the CNV lesion. Additional secondary outcomes included the mean

183

number of treatments administered over 12 months . In two sub group analyses within the

184

monotherapy group, the outcomes of individual anti-VEGF agents were performed and we

185

also described the outcomes of these eyes that underwent rescue PDT in combination with

186

anti-VEGF therapy.

M AN U

SC

RI PT

177

187 Statistical Analysis

189

Descriptive data are presented as mean (SD), median (25th and 75th percentiles, Q1 and Q3)

190

or number (percentage). Statistical test such as Student’s t-test, Wilcoxon rank sum, and

191

Fisher’s tests was used where appropriate to compare baseline characteristics between the

192

anti-VEGF monotherapy and initial PDT in combination with anti-VEGF therapy groups.

193

Locally weighted scatterplot smoothing (LOESS) curves were used to analyse VA throughout

194

the follow-up for different treatment groups and eyes with varying starting vision. ITT

195

analysis was performed and mean change in VA between treatment groups at 12 months were

196

assessed by mixed-effects regression models with the therapy group as the main predictor

197

variable. Adjusted means were used to assess the change in VA from baseline to 12 months

198

after adjusting for age, baseline VA, lesion size, and country. A logistic regression model

199

adjusted for age, baseline VA, and lesion size and country was used to compare the

200

proportion of eyes with 15 letter loss, 15 letter gain and VA better than 6/12 at 12 months

201

between groups.

AC C

EP

TE D

188

ACCEPTED MANUSCRIPT 202 The number of injections was compared by a Poisson regression model adjusted for age,

204

baseline VA, lesion size, and country, with log days of follow-up included as an offset

205

variable. Cox proportional hazards regression analysis adjusted for age, baseline VA, lesion

206

size and country was used to compare the proportion of eyes that became inactive and the

207

mean time to inactivity. Kaplan-Meier curve analysis was used to display the corresponding

208

results. Appropriate models were applied to the subset analysis comparing anti-VEGF agents

209

in the monotherapy group using Holm-Bonferroni correction for multiple comparisons. All

210

analyses were calculated using R22 with the lme4 package for mixed-effects regression

211

analysis and the survival package for Kaplan-Meier analyses.

M AN U

SC

RI PT

203

212

AC C

EP

TE D

213

ACCEPTED MANUSCRIPT Results

215

Study Participants

216

We identified 207 treatment- naive eyes diagnosed with PCV from 5082 neovascular AMD

217

patients captured on the FRB! database between December 1, 2010, and July 1, 2016; 152

218

eyes underwent anti-VEGF monotherapy; 41 eyes underwent initial PDT in combination with

219

anti-VEGF therapy and 14 eyes underwent focal laser in combination with anti-VEGF

220

therapy. The eyes that underwent focal laser were excluded and final statistical analyses

221

comparing the monotherapy (n=152) and combination groups (n=41) was based on 193 eyes.

222

Baseline characteristics between treatment groups are summarised in Table 1. The median

223

baseline VA (Q1, Q3) was significantly lower in the initial PDT combination group (46 [20,

224

67.3] letters) compared to the monotherapy group (60 [45.0, 70.0] letters, p=0.03). Other

225

baseline characteristics such as age, sex, and lesion size were not significantly different

226

between groups. A large proportion of eyes (patients) were non-Caucasian (60%) of which

227

most (97%) were Asian. Different ethnic distributions between treatment groups also

228

reflected the treatment patterns that varied from country to country. There was much higher

229

rate of combination therapy in the non-Caucasian group (most from Singapore where PDT

230

was used between 2009-2013) as compared to the Caucasian group, where PDT was not the

231

physicians’ initial treatment of choice and used only in cases with sub optimal response to

232

anti VEGF therapy or was presumably less accessible. In the monotherapy group,

233

Caucasians consisted of 40.8% while 55.2% were Asian. The baseline median VA (Q1, Q3)

234

was higher in Caucasians compared to Asians (64.0[54.75, 70.5] versus 54.0[39.5, 68],

235

p=0.03)

AC C

EP

TE D

M AN U

SC

RI PT

214

236 237

Vision Outcomes

ACCEPTED MANUSCRIPT The mean VA change for all eyes, including noncompleters, grouped by treatment type

239

during 12 months is shown in Figure 1. Despite the difference in baseline VA, the VA

240

improved significantly at month 12 and was similar between the combination and

241

monotherapy groups, (70 vs 67.5 letters, p=0.44). The mean VA change between

242

combination therapy and monotherapy (including the LOCF at the point of rescue PDT

243

treatment) is shown in Table 2. Mean (CI) VA increased by 14.3 (5.5 - 23.0) letters in the

244

combination group compared with 8.4 letters (5.4 – 11.3) in the monotherapy group. The

245

difference in crude VA change between treatment groups was statistically significant

246

(p=0.03), as was the difference in adjusted VA change of 8.8 letters (p = 0.03) favouring the

247

combination group. There was a higher proportion of eyes that gained 15 letters or more in

248

the combination group as compared to the monotherapy group, although the difference was

249

not statistically significant (51.2% versus 32.0%, p =0.09). The proportions of eyes that lost

250

15 or more letters in the combination and monotherapy groups were similar (8.8% versus

251

5.0% respectively, p = 0.87). There was no difference in the proportion of eyes that achieved

252

vision better than 6/12 between the combination and monotherapy groups (46.8% versus

253

55.8%, p=0.21 respectively).

SC

M AN U

TE D

EP

254

RI PT

238

The overall mean VA change over 12 months and comparison between treatment groups

256

during the 12 months stratified by baseline line vision is shown in Figures 2 and 3. Eyes

257

were stratified into 3 groups according to their presenting baseline VA: (a) “good baseline

258

vision” ≥70 letters (49 eyes); (b) “moderate baseline vision”, between 36 and 69 letters (97

259

eyes); and (c) “low baseline vision”, ≤ 35 letters (42 eyes). The largest gains in vision (mean

260

[CI]) were seen in eyes that started in the low baseline vision band (20.9 [12.6 – 29.2]

261

letters). Eyes in the combination group gained 28.6 [11.7 - 45.4] letters versus 17.0 [7.4 –

262

26.6] letters in the monotherapy group (p = 0.22) in this band. In the good vision band, there

AC C

255

ACCEPTED MANUSCRIPT appeared to be a slight loss in vision in the monotherapy group compared to the combination

264

group, where there was a modest gain over 12 months (-0.3 [-4.7 – 4.2] letters versus 4.2 [1.1

265

– 7.3] letters respectively, p=0.09). In the moderate vision band, the gains in vision were

266

similar between groups (11.4 [0.3 – 22.6] letters versus 10.1 [6.8 – 13.3] letters respectively,

267

p = 0.81).

RI PT

263

268 269 Lesion Activity

271

A Kaplan-Meier survival curve representing the proportion of eyes achieving inactivity

272

between groups is shown in Figure 4. More eyes were graded as inactive over a shorter

273

period in the combination group than in the monotherapy group. The proportion of eyes

274

graded as inactive and time to inactivity between groups during the 12-month period is

275

shown in Table 2. In the combination group, 85.3% of eyes were eventually graded inactive

276

compared to 76.8% in the monotherapy group (p = 0.01). Eyes in the combination group also

277

became inactive more quickly than those in the monotherapy group (mean [CI] 80.7 [62.8-

278

98.5] days versus 150.4 [132.8 – 168.0] days to first grading of inactivity, p = 0.01).

M AN U

TE D

281

Injection Outcomes and Treatment Requirements

282

Over the first 12 months, the combination group received fewer injections mean [CI] (4.3

283

[3.6-5.2]) than the monotherapy group (6.4 [5.9 – 6.9], p = 0.01). The total number of anti-

284

VEGF injections administered was 161 in the combination group (n=41) versus 698 in the

285

monotherapy group (n=152) over 12 months. Most anti-VEGF injections administered were

286

bevacizumab ([70.8% [114 injections] in the combination group and 65.0% [454 injections]

287

in the monotherapy group). Ranibizumab was given in 8.6% (14 injections) of the

AC C

280

EP

279

SC

270

ACCEPTED MANUSCRIPT combination group and 18.3% (128 injections) of the monotherapy group while aflibercept

289

was used in 20.4% (33 injections) in the combination group and 16.7% (116 injections) in the

290

monotherapy. A subset analysis of outcomes of the different agents in the monotherapy

291

group showed no significant difference in visual acuity outcomes, however the number of

292

eyes treated exclusively with ranibizumab (n=16) and aflibercept (n=13) were small. (Table

293

4).

RI PT

288

294

Outcomes of Eyes in Monotherapy Group requiring Rescue PDT

296

Twelve eyes among the 152 patients in the monotherapy group received PDT after at least 90

297

days of treatment with anti-VEGF therapy. Based on ITT, these 12 eyes were included in the

298

monotherapy arm for the main analysis. Since the timing and criteria for adding PDT after

299

initial monotherapy were not standardized, this subgroup was considered to be receiving anti-

300

VEGF monotherapy with rescue PDT. The mean [CI] VA of this subgroup increased from

301

54.0 [42.0 – 65.8] letters to 58.6 [46.3 – 70.0] letters (+4.3 [-10.9 – 19.5] letters, p=0.57) at

302

12 months. The mean [CI] time to treatment with PDT was 225.2 [178.2 – 272.3] days and

303

the average number of anti-VEGF injections until PDT was performed was 5.2 [4.5 – 6.1].

304

The mean [CI] time to the first grading of lesion inactivity was 281.7 [216.0 – 347.5] days.

305

There was no significant difference between the VA at time of PDT compared to final VA at

306

12 months (7.2 [-6.1 – 20.5] letters, p= 0.26).

M AN U

TE D

EP

AC C

307

SC

295

308

Non-completion Rate

309

The proportion of eyes that did not complete the 12-month study was similar for the

310

combination therapy and monotherapy groups (17.0% versus 11.8%, p = 0.37), as was the

311

mean time to last visit for non-completers between the 2 groups (184.9 [103.5 – 266.2] days

312

versus 202.8 [172.1 – 233.7) days, p= 0.63). The mean VA change from baseline to the

ACCEPTED MANUSCRIPT LOCF for non-completers also did not differ significantly between the combination therapy

314

and monotherapy groups (22.6 [3.5-41.6] letters versus 12.7 [2.8 – 22.5] letters, p = 0.43).

315

Reasons were given for discontinuation and loss of follow up in only 6 of the 22 dropouts.

316

These included changing of practitioner (3), patient declining further treatment (2) and

317

further treatment deemed futile (1).

RI PT

313

318 319

SC

320

AC C

EP

TE D

M AN U

321

ACCEPTED MANUSCRIPT Discussion

323

This study compared the real-world outcomes over 12 months for the treatment of PCV with

324

either initial PDT in combination with anti-VEGF therapy or anti-VEGF monotherapy,

325

including a small number that received “rescue” treatment with PDT, with observations from

326

the FRB! database. After adjusting for baseline characteristics, including visual acuity, the

327

combination group gained more letters (+14.3) at 12 months, the primary outcome, than those

328

receiving monotherapy (+8.4). These findings are consistent with a recent meta-analysis, a

329

RCT18 and several smaller non-randomised comparative studies23-26 which have all reported

330

better vision outcomes using PDT in combination with anti-VEGF therapy.

SC

RI PT

322

M AN U

331

The baseline VA was significantly lower in the combination group in the present study,

333

possibly because clinicians used PDT in cases deemed more advanced. Other reasons for this

334

difference may be that Asian patients present later and since the proportion of patients

335

receiving PDT was predominately Asian, this resulted in an overall lower baseline VA within

336

the combination group. The Asian patients within the monotherapy group also had lower

337

baseline visual acuity than the Caucasians. It is generally acknowledged that eyes with worse

338

vision gain more vision after starting treatment27,28, however the difference in visual acuity

339

outcomes between the 2 groups in the present study was still statistically significant after

340

adjusting for baseline VA. Overall we also found much larger visual acuity gains in both

341

groups in this real-world study than in the EVEREST-II study despite the use of bevacizumab

342

in most of our cohort rather than ranibizumab. This is likely due to the much lower starting

343

vision in our study than in the EVEREST-II rather than different efficacy of the 2 agents.18

AC C

EP

TE D

332

344 345

The anatomical response to combination therapy appeared to be better than monotherapy,

346

with a higher proportion of CNV lesions being graded as inactive within a shorter period.

ACCEPTED MANUSCRIPT Several other studies have also reported better anatomical outcomes with combination

348

therapy for PCV29,30. PCV-specific trials like the EVEREST-II and PLANET studies used

349

several metrics for the measure of anatomical outcomes including ICGA confirmed polyp

350

regression/closure (a PCV specific outcome) and disease inactivity based on presence of fluid

351

on OCT and clinical exam (a more generic outcome).31,32

352

As the assessment of PCV regression is a disease-specific measure requiring the use of

353

ICGA, these data were not captured by the FRB! system. Instead, our study used the grading

354

of inactivity as a marker of anatomical response. Disease acitivty in the FRB! system is

355

defined as the presence of intraretinal or subretinal fluid or haemorrhage that is attributable to

356

activity of the neovascular lesion as determined by the treating ophthalmologist. This could

357

be based on clinical examination or imaging (optical coherence tomography, or dye

358

angiography, alone or in combination). This definition is similar to the disease activity

359

crtieria in the EVEREST II and PLANET studies. The proportion graded as inactive in the

360

combination group (85.3% at month 12) of our study was similar to the EVEREST-II (79.5%

361

at month 11) but different when comparing the monotherapy group (76.8% inactive at 12

362

months in the FRB monotherapy group versus 50.0% inactive in the EVEREST-II

363

monotherapy group). This proportion of inactivity in the FRB monotherapy group was more

364

similar to the proportion of inactive eyes (defined by the absence of fluid on OCT) in the

365

PLANET study (81.7% in the aflibercept monotherapy group and 88.9% in the aflibercept

366

plus active PDT group. Our reported disease activity outcome is broadly consistent with

367

PCV RCTs and reflects the real-world nature of the study where treatment may be guided by

368

lesion activity as noted clinically or on OCT rather than on dye angiography which may be

369

difficult to obtain regularly.

370

AC C

EP

TE D

M AN U

SC

RI PT

347

ACCEPTED MANUSCRIPT An increasing concern in the era of anti-VEGF therapy is the burden of repeated injections.

372

We found that in eyes in the combination group received fewer anti-VEGF injections than in

373

the monotherapy group. This has also been observed in several other studies33,34, including

374

the EVEREST II study.18 Direct closure of the lesion with PDT may result in reduced VEGF

375

levels, thereby reducing the overall anti-VEGF treatment burden. In the present study, about

376

2 injections were saved with the administration of PDT with a mean number of 1.23 PDT

377

treatments applied over 12 months. While savings of anti-VEGF injections are noted, this

378

result must take into account cost considerations specific to different healthcare jurisdictions

379

and remuneration models and are largely dependent on subsidies and funding for each

380

treatment type.

M AN U

SC

RI PT

371

381

Analysis of these observational data was carried out using the principle of ITT and including

383

data from patients that discontinued their treatment. Noncompleters gained more letters than

384

completers at their final visit, however this was not statistically significant. This could be

385

because there was only a small number of noncompleters and a large variability in the letter

386

change in vision within groups. Nonetheless, this result suggests that the reasons for

387

discontinuation from FRB! were likely to be independent of treatment regimen or visual

388

outcome. (Table 3)

390

EP

AC C

389

TE D

382

Results of eyes within the monotherapy group that received rescue PDT should be

391

interpreted with caution, as the indication and timing for rescue PDT was highly variable.

392

The concept of ITT was also applied in the analysis of these patients. In this small group, the

393

physicians’ initial management intention was for monotherapy, hence their outcomes just

394

prior to rescue PDT were carried forward into the monotherapy group. The reasons for

395

applying rescue PDT were not captured by the FRB! system, however presumably it reflected

ACCEPTED MANUSCRIPT a suboptimal response of anti-VEGF monotherapy in terms of visual outcome or treatment

397

burden. We found no difference in the mean visual acuity prior to rescue PDT and the 12-

398

month vision of monotherapy eyes that did not receive rescue PDT (61.7 [52.3 – 70.0] letters

399

versus 63.3 [59.3 – 67.3] letters, p=0.63). This suggests that both explanations are plausible

400

in the cohort of patients we studied.

RI PT

396

401 402

We acknowledge the lower internal validity of observational studies, where there is a lack of standardised treatment and study protocols, compared with RCTs. One potential

404

weakness of the present study is the inter-observer variability of the grading of lesion

405

inactivity, which was based on decisions made by unmasked individual treating physicians.

406

However, this effect is likely to have been consistent across the multiple sites and physicians

407

were included in this study. Whilst care should be taken when comparing our results with

408

those of RCTs, large observational studies such as this may provide a truer reflection of real

409

world outcomes.

M AN U

TE D

410

SC

403

As the present analysis was a comparison of 2 treatment strategies, we have attempted to mitigate some of the shortcomings inherent in observational studies by correcting for

412

baseline characteristics and inter-practice effects in the regression model to minimize bias.

413

Despite analysing a large cohort of AMD patients, there was a relatively low prevalence of

414

PCV in the FRB! database, especially in the Caucasian population. A possible explanation

415

for this could be a high false negative rate in diagnosing PCV, especially in centres that do

416

not have access to ICGA or where ICGA is not part of the standard clinical assessment and is

417

only performed if PCV is suspected. While no meaningful prevalence data can be derived

418

from this study, we still believe that cases entered as PCV are highly likely to have had the

419

condition. When polled, all clinicians involved at the very least indicated that an ICGA

420

would be performed if there was clinical or imaging suspicion of PCV, while in some

AC C

EP

411

ACCEPTED MANUSCRIPT 421

practices dye angiography including ICGA was performed as standard in all cases of

422

neovascular AMD to rule out PCV.

423 While our results suggest that early PDT in combination with anti-VEGF therapy is

425

more efficacious than anti-VEGF monotherapy, it is important to note that most eyes were

426

treated with bevacizumab. This again reflects real world practice, where on-label anti-VEGF

427

agents may not be accessible due to cost constrains. There have been suggestions that

428

outcomes for PCV may differ when treated with different anti-VEGF agents, however there

429

has been no large head to head comparison to date.10,35-37 In our study, a subset analysis

430

based on agent type found no significant difference in visual outcomes between the 3 anti-

431

VEGF agents, however sample sizes were small for the ranibizumab and aflibercept

432

monotherapy groups. In addition, the majority of data collected was prior to 2016 where

433

aflibercept was yet to be widely available in many regions. The inactivation rates in this

434

study were much lower than reported in the recent randomized clinical trials. At 12 weeks,

435

the EVEREST-II study of ranibizumab reported disease inactivity in 73.6% and 39.3% in the

436

combination arm and the monotherapy arm respectively.18 The PLANET study of aflibercept

437

reported inactivity in 77.1-77.4% at the similar time point.38 Our study reported a lower

438

proportion of inactivity in both groups at 12 weeks (63.4% of the combination group and

439

25.1% of the monotherapy group). This lower rate of inactivation might be due to a weaker

440

effect of bevacizumab which was the main anti-VEGF agent used in our cohort.

442

SC

M AN U

TE D

EP

AC C

441

RI PT

424

Despite some of the shortcomings, this study nonetheless showcases the opportunity

443

of a system such as the FRB! registry to compare treatment patterns for different subtypes of

444

nAMD across different international practices. Unfortunately, no meaningful comparisons of

ACCEPTED MANUSCRIPT 445

treatment strategies could be performed between ethnicities as only a small number of

446

Caucasian patients underwent combination therapy.

447 448

The main finding from this study is that combination therapy appears to give better outcomes than anti-VEGF monotherapy for the management of PCV over 12 months in

450

routine clinical practice. While all anti-VEGF agents were theoretically available,

451

bevacizumab was the agent that was most commonly used. Although this study lacks the

452

robust criteria of a RCT, the results reflect real world practice and outcomes. It is

453

encouraging to note that the results reported here are consistent with the findings from the

454

current landmark RCTs, further strengthening the evidence for the use of combination

455

therapy in the treatment of PCV.

456

M AN U

SC

RI PT

449

Acknowledgements

458

The Fight Retinal Blindness! investigators: Auckland District Health Board, New Zealand

459

(Dr D Squirrell); Cairns Eye Surgery, Queensland (Dr A Field); Canberra Hospital,

460

Australian Capital Territory (Dr C Dayajeewa, Dr J Wells); Centre for Eye Research

461

Australia, Victoria (Professor R Guymer); Eye Associates, New South Wales (Dr M Gillies);

462

Eyemedics, South Australia (Dr N Saha, Dr S Lake); Marsden Eye Specialists, New South

463

Wales (Dr J Arnold, Dr D Chan); Nepean Valley Eye Surgeons, New South Wales (Dr G

464

Banerjee); Retina Associates, New South Wales (Dr S Fraser-Bell); Retina Specialists, New

465

Zealand (Dr R Barnes); Singapore National Eye Centre , Singapore (Dr G Cheung);

466

Specialist Eye Group, Victoria (Dr A Cohn, Dr L Chow); Strathfield Retina Clinic, New

AC C

EP

TE D

457

ACCEPTED MANUSCRIPT 467

South Wales (Dr C Lim); University Hospital Zurich, Switzerland (Dr D Barthelmes);

468

Victorian Eye Surgeons, Victoria (Dr A Cohn)

469

RI PT

470 471 472

SC

473 474

476

References

477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503

1.

5.

6.

7.

8.

9.

10.

TE D

4.

EP

3.

Lim LS, Mitchell P, Seddon JM, Holz FG, Wong TY. Age-related macular degeneration. The Lancet. 2012;379(9827):1728-1738. Bressler NM. Age-related macular degeneration is the leading cause of blindness. JAMA. 2004;291(15):1900-1901. Comparison of Age-related Macular Degeneration Treatments Trials Research G, Martin DF, Maguire MG, et al. Ranibizumab and bevacizumab for treatment of neovascular age-related macular degeneration: two-year results. Ophthalmology. 2012;119(7):1388-1398. Yannuzzi LA, Sorenson J, Spaide RF, Lipson B. Idiopathic polypoidal choroidal vasculopathy (IPCV). Retina. 1990;10(1):1-8. Ahuja RM, Stanga PE, Vingerling JR, Reck AC, Bird AC. Polypoidal choroidal vasculopathy in exudative and haemorrhagic pigment epithelial detachments. Br J Ophthalmol. 2000;84(5):479-484. Cackett P, Htoon H, Wong D, Yeo I. Haemorrhagic pigment epithelial detachment as a predictive feature of polypoidal choroidal vasculopathy in a Chinese population. Eye (Lond). 2010;24(5):789-792. Kuo JZ, Wong TY, Ong FS. Genetic risk, ethnic variations and pharmacogenetic biomarkers in AMD and polypoidal choroidal vasculopathy. Expert review of ophthalmology. 2014. Laude A, Cackett PD, Vithana EN, et al. Polypoidal choroidal vasculopathy and neovascular age-related macular degeneration: same or different disease? Progress in retinal and eye research. 2010;29(1):19-29. Lim LS, Cheung CMG, Wong TY. Asian age-related macular degeneration: current concepts and gaps in knowledge. The Asia-Pacific Journal of Ophthalmology. 2013;2(1):32-41. Wong CW, Wong TY, Cheung CMG. Polypoidal choroidal vasculopathy in Asians. Journal of clinical medicine. 2015;4(5):782-821.

AC C

2.

M AN U

475

ACCEPTED MANUSCRIPT

16.

17. 18.

19.

20.

21.

22. 23.

24.

25.

26.

RI PT

15.

SC

14.

M AN U

13.

TE D

12.

Wong CW, Yanagi Y, Lee W-K, et al. Age-related macular degeneration and polypoidal choroidal vasculopathy in Asians. Progress in retinal and eye research. 2016. Coppens G, Spielberg L, Leys A. Polypoidal choroidal vasculopathy, diagnosis and management. Bull Soc Belge Ophtalmol. 2011(317):39-44. Yadav S, Parry DG, Beare NAV, Pearce IA. Polypoidal choroidal vasculopathy: a common type of neovascular age-related macular degeneration in Caucasians. Br J Ophthalmol. 2017;101(10):1377-1380. Lai TY, Lee GK, Luk FO, Lam DS. Intravitreal ranibizumab with or without photodynamic therapy for the treatment of symptomatic polypoidal choroidal vasculopathy. Retina. 2011;31(8):1581-1588. Gomi F, Sawa M, Sakaguchi H, et al. Efficacy of intravitreal bevacizumab for polypoidal choroidal vasculopathy. Br J Ophthalmol. 2008;92(1):70-73. Kang HM, Koh HJ. Long-term visual outcome and prognostic factors after intravitreal ranibizumab injections for polypoidal choroidal vasculopathy. Am J Ophthalmol. 2013;156(4):652-660. Cheung CM, Wong TY. Ranibizumab and bevacizumab for AMD. N Engl J Med. 2011;365(23):2237; author reply 2237. Koh A, Lai TYY, Takahashi K, et al. Efficacy and Safety of Ranibizumab With or Without Verteporfin Photodynamic Therapy for Polypoidal Choroidal Vasculopathy: A Randomized Clinical Trial. JAMA Ophthalmol. 2017;135(11):1206-1213. Penterman EJ, Nijman HL, Saalmink K, Rasing S, van der Staak CP. [Assessing aggressive behaviour at the psychiatric emergency service with a checklist: a replication study]. Tijdschr Psychiatr. 2013;55(2):93-100. Gillies MC, Walton R, Liong J, et al. Efficient capture of high-quality data on outcomes of treatment for macular diseases: the fight retinal blindness! Project. Retina. 2014;34(1):188-195. von Elm E, Altman DG, Egger M, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. Lancet. 2007;370(9596):1453-1457. R: A language and environment for statistical computing [computer program]. Vienna, Austria: R Foundation for Statistical Computing; 2017. Sakurai M, Baba T, Kitahashi M, et al. One-year results of intravitreal ranibizumab combined with reduced-fluence photodynamic therapy for polypoidal choroidal vasculopathy. Clin Ophthalmol. 2014;8:235-241. Lim JY, Lee SY, Kim JG, Lee JY, Chung H, Yoon YH. Intravitreal bevacizumab alone versus in combination with photodynamic therapy for the treatment of neovascular maculopathy in patients aged 50 years or older: 1-year results of a prospective clinical study. Acta Ophthalmol. 2012;90(1):61-67. Kang HM, Koh HJ. Two-year outcome after combination therapy for polypoidal choroidal vasculopathy: comparison with photodynamic monotherapy and antivascular endothelial growth factor monotherapy. Ophthalmologica. 2014;231(2):8693. Song MH, Ryu HW, Roh YJ. One-year results of intravitreal ranibizumab with or without photodynamic therapy for polypoidal choroidal vasculopathy. Ophthalmologica. 2011;226(3):119-126.

EP

11.

AC C

504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549

ACCEPTED MANUSCRIPT

585 586 587 588 589 590

32. 33.

34.

35.

36.

37. 38.

RI PT

31.

SC

30.

M AN U

29.

TE D

28.

Gillies MC, Campain A, Barthelmes D, et al. Long-Term Outcomes of Treatment of Neovascular Age-Related Macular Degeneration: Data from an Observational Study. Ophthalmology. 2015;122(9):1837-1845. Writing Committee for the UKA-RMDEMRUG. The neovascular age-related macular degeneration database: multicenter study of 92 976 ranibizumab injections: report 1: visual acuity. Ophthalmology. 2014;121(5):1092-1101. Wang W, He M, Zhang X. Combined intravitreal anti-VEGF and photodynamic therapy versus photodynamic monotherapy for polypoidal choroidal vasculopathy: a systematic review and meta-analysis of comparative studies. PLoS One. 2014;9(10):e110667. Tang K, Si JK, Guo DD, et al. Ranibizumab alone or in combination with photodynamic therapy vs photodynamic therapy for polypoidal choroidal vasculopathy: a systematic review and Meta-analysis. Int J Ophthalmol. 2015;8(5):1056-1066. Cheung GCM, Lai TYY, Gomi F, Ruamviboonsuk P, Koh A, Lee WK. Anti-VEGF Therapy for Neovascular AMD and Polypoidal Choroidal Vasculopathy. Asia Pac J Ophthalmol (Phila). 2017;6(6):527-534. Cheung CMG, Lai TYY, Ruamviboonsuk P, et al. Polypoidal Choroidal Vasculopathy: Definition, Pathogenesis, Diagnosis, and Management. Ophthalmology. 2018. Gomi F, Oshima Y, Mori R, et al. INITIAL VERSUS DELAYED PHOTODYNAMIC THERAPY IN COMBINATION WITH RANIBIZUMAB FOR TREATMENT OF POLYPOIDAL CHOROIDAL VASCULOPATHY: The Fujisan Study. Retina. 2015;35(8):1569-1576. Akaza E, Yuzawa M, Matsumoto Y, Kashiwakura S, Fujita K, Mori R. Role of photodynamic therapy in polypoidal choroidal vasculopathy. Japanese journal of ophthalmology. 2007;51(4):270-277. Saito M, Kano M, Itagaki K, Oguchi Y, Sekiryu T. Switching to intravitreal aflibercept injection for polypoidal choroidal vasculopathy refractory to ranibizumab. Retina. 2014;34(11):2192-2201. Hata M, Oishi A, Tsujikawa A, et al. Efficacy of intravitreal injection of aflibercept in neovascular age-related macular degeneration with or without choroidal vascular hyperpermeability. Invest Ophthalmol Vis Sci. 2014;55(12):7874-7880. Cho HJ, Kim KM, Kim HS, et al. Intravitreal Aflibercept and Ranibizumab Injections for Polypoidal Choroidal Vasculopathy. Am J Ophthalmol. 2016;165:1-6. Iida T. Results of the PLANET study. 10th Asia-Pacific Vitreo-retina Society Congress; December 9, 2016, 2016; Bangkok, Thailand.

EP

27.

AC C

550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584

ACCEPTED MANUSCRIPT 591 592 593 594

RI PT

595 596 597

SC

598 599

M AN U

600 601 602 603

605

TE D

604

Legends

607

Figure 1: Loess regression curves over 12 months stratified by baseline vision. Most eyes

608

stayed within their own bands, except for the eyes with poor baseline vision (35 letters of

609

worse) which managed to gain enough letters to cross into the moderate vision band at the

610

end of 1 year. VA: visual acuity

AC C

611

EP

606

612 613

Figure 2: Loess regression curves over 12 months in the 2 treatment groups.

614

A larger increase in visual acuity was noted in the initial PDT in combination with anti-

615

VEGF therapy over 12 months and appeared to continue to increase throughout the 12-month

ACCEPTED MANUSCRIPT 616

time period as compared to Vision in the monotherapy with anti-VEGF group, where the

617

increase was less over 12 months and vision appeared to plateau. VEGF: vascular

618

endothelial growth factor, PDT: photodynamic therapy

619

RI PT

620 621

Figure 3: Loess regression curves over 12 months for monotherapy with anti-VEGF and

623

initial PDT in combination with anti-VEGF therapy stratified by baseline vision. Vision

624

gains were evident in all bands for initial PDT in combination with anti-VEGF therapy group

625

with largest gains in the low vision band crossing into the moderate vision band.

626

Monotherapy with anti-VEGF group appeared to have gains in all bands except for those in

627

the good vision band where vision appeared to be lost over 12 months. VEGF: vascular

628

endothelial growth factor, PDT: photodynamic therapy

M AN U

SC

622

630

TE D

629

Figure 4. Kaplan-Meier curve representing the proportion to inactivity over time between the

632

2 treatment groups, monotherapy with anti-VEGF group and initial PDT in combination with

633

anti-VEGF therapy group. A larger proportion of eyes over a shorter time achieved inactivity

634

in the initial PDT in combination with anti-VEGF therapy compared to the monotherapy with

635

anti-VEGF group. VEGF: vascular endothelial growth factor, PDT: photodynamic therapy

AC C

636

EP

631

ACCEPTED MANUSCRIPT Table 1. Comparison of index visit characteristics of eyes with PCV between different initial treatment modalities. Anti-VEGF monotherapy

Initial PDT + antiVEGF

No. of eyes, n (%)

152 (73.4)

41 (19.8)

Mean age, years (CI)

70.7 (68.9 – 72.4)

69.0 (65.3 – 72.6)

0.16§

P value*

Overall

193

RI PT

70.6 (69.1 – 72.1)

69 (61.5)

18 (43.9)

0.21+

87 (45.0)

Median Baseline VA, letters (Q1, Q3) VA better than 6/12 at baseline, n (%) Median GLD, µm (Q1, Q3)

60 (45.0, 70.0)

46 (20, 67.3)

0.03++

60 (40, 70)

41 (26.9)

9 (22.0)

0.63+

58 (25.9)

3000 (1825, 4500)

2442 (1648, 5173)

0.54++

3000 (1710, 4518)

Caucasian, n (%)

62 (40.8)

3 (7.3)

-

65 (33.7)

Non-Caucasian, n (%)

84 (55.2)

31(75.6)

-

Undisclosed, n (%)

6 (3.9)

7 (17.1)

SC

Sex, male (%)

M AN U

115 (59.6)

-

13 (6.7)

*P-value – Anti-VEGF monotherapy versus Initial PDT in combination with anti-VEGF therapy; t-test§, Fishers exact test+ and Wilcoxon rank sum test++ VEGF: vascular endothelial growth factor, PDT: photodynamic therapy, VA: visual acuity,

AC C

EP

TE D

GLD: greatest linear diameter

ACCEPTED MANUSCRIPT Table 2. Comparison of outcomes between initial treatment modalities: anti-VEGF monotherapy and initial PDT in combination with anti-VEGF therapy.

Anti-VEGF monotherapy

Initial PDT + antiVEGF

No. of eyes, n (%)

152 (73.4)

41 (17.0)

Median VA at 1 year or LOCF, letters (Q1, Q3)

70.0 (60.0, 78.0)

67.5 (49.0, 75.0)

0.44+

Crude mean change in VA, letters (CI)

8.4 (5.4-11.3)

14.3 (5.5- 23.0)

0.03+

Adjusted mean change in VA, letters (CI)

8.2 (5.2 – 11.2)

16.9 (10.6 – 23.3)

0.02+++

15 letter loss, n (%)

10 (5.0)

15 letter gain, n (%)

47 (32.0)

Time to inactivity of lesion, days (CI)

Mean number of injections, n (CI) Mean number of PDT treatments, n (CI)

RI PT

SC

0.87++

21 (51.2)

0.09++

82(55.8)

22(46.8)

0.21++

150.4 (132.8 – 168.0)

80.7 (62.8 – 98.5)

0.01§

6.4 (5.9-6.9)

4.3 (3.6 -5.2)

0.01§§

-

1.23 (1.13 – 1.36)

-

112 (76.8)

35 (85.3)

0.01§

TE D

Eyes that were inactive at month 12, n (%)

3 (8.8)

M AN U

VA better than 6/12, n (%)

P-value*

P-value* – Significance for anti-VEGF monotherapy versus Initial PDT in combination with anti-VEGF therapy; calculated using Wilcoxon rank sum+, logistic++, linear+++, cox

EP

proportional hazards§ or Poisson§§ regression model correcting for age, baseline visual acuity, practice and baseline greatest linear dimension of lesion.

AC C

VEGF: vascular endothelia growth factor, PDT: photodynamic therapy, VA: visual acuity, GLD: greatest linear diameter, LOCF: last observation carried forwards, CI: confidence interval

ACCEPTED MANUSCRIPT Table 3. Demographics and outcomes for non-completers Non-completers

Mean age, years (CI) Median baseline VA, letters (Q1, Q3) Median GLD at baseline, µm (Q1, Q3) Median VA at last observation, letters (Q1, Q3) Mean change in VA, letters (CI)

Initial PDT + antiVEGF

Overall 25

18 (11.8)

7 (17.0)

168

68.3 (63.8-72.8) 50 (35, 66) 3057.5 (2296, 3575)

67.2 (61.6-79.8) 50 (35.0, 57.5) 3157.5 (2974.5, 4316)

70.8 (62.0-79.8) 59 (35, 69) 2500 (1300, 3275.5)

70.6 (69.0 -72.2) 60 (40, 70) 2948 (1744, 4733)

73(60, 75)

65 (57.5, 73.5)

75 (73, 75)

15(4.3-25.7)

12.1 (-1.8-26.2)

24.2 (-2.3-50.7)

P value*

0.91§ 0.60+

overall

RI PT

No. of eyes, n (%)

completers

Anti-VEGF monotherapy

0.06+

P value*

0.52§ 0.65+ 0.86+

0.31+

70 (57, 78)

0.49+

0.40§

9.3(6.3 -12.3)

0.16++

SC

P-value* – Significance of anti-VEGF monotherapy versus Initial PDT in combination with anti-VEGF therapy; t-test§, Wilcoxon rank sum+ and linear++ regression model correcting

M AN U

for age, baseline visual acuity, country and baseline greatest linear dimension of lesion VEGF: vascular endothelia growth factor, PDT: photodynamic therapy, VA: visual acuity,

AC C

EP

TE D

GLD: greatest linear diameter, CI: confidence interval

ACCEPTED MANUSCRIPT Table 4. Subset analysis of anti-VEGF treatment by single agent within the monotherapy group P value*

82 (73.8)

16 (14.4)

13 (11.7)

8 (0, 15.0)

6.5 (0, 13.0)

5 (0, 15.0)

0.22

60 (40.0, 60.0)

70 (58.5, 75.0)

68 (51.0, 70.0)

1.00

70.0 (60.0, 78.0)

75.0 (65.5, 80.0)

73.0 (67.0, 80.0)

0.17

5(6.1)

1(6.2)

0(0)

1.00

21(25.6)

1(6.2)

4(30.1)

1.00

42 (51.2)

8 (50.0)

8 (61.5)

1.00

5 (3, 6)

5(3, 7)

Ranibizumab vs. Aflibercept

5(3, 7)

1.00

Bevacizumab vs. Aflibercept

RI PT

Aflibercept

1.00

0.16

1.00

0.16

0.42

0.18

1.00

1.00

0.60

1.00

1.00

1.00

1.00

1.00

SC

Ranibizumab

M AN U

No. of eyes, n (%) Median change in VA, letters (CI) Median Baseline VA, letters (Q1, Q3) Median VA at 1 year or LOCF, letters (Q1, Q3) 15 letter loss, n (%) 15 letter gain, n (%) VA better than 6/12, n, (%) Median number of injections, n (Q1, Q3)

Bevacizumab vs. Ranibizumab & Aflibercept

Bevacizumab

*Holms corrected P-value for multiple comparisons

VEGF: vascular endothelia growth factor, VA: visual acuity, LOCF: last observation carried

AC C

EP

TE D

forwards, CI: confidence interval

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT Title: A multi-country comparison of real world management and outcomes of polypoidal choroidal vasculopathy – Fight Retinal Blindness! Cohort Precis:

AC C

EP

TE D

M AN U

SC

RI PT

Photodynamic therapy in combination with anti-vascular endothelial growth factor (VEGF) results in better vision, higher proportion of inactivity and a lower anti-VEGF treatment burden compared to anti VEGF monotherapy in treating polypoidal choroidal vasculopathy.