Data are provided for the full analysis set. One patient in the ranibizumab plus laser group had no baseline NEI VFQ-25 data. The 12-month data are reported without last observation carried forward; with the last observation carried forward only 4 patients were lost after baseline.
A, Mean change from baseline to 12 months in NEI VFQ-25 scores. B, NEI VFQ-25 composite score and general vision subscales. C, NEI VFQ-25 near activities and distance activities subscales. Error bars indicate upper 95% confidence limit. aP < .05.bP ≤ .001 vs laser.
A, Baseline visual acuity (VA). B, Better-seeing eye (BSE) or worse-seeing eye (WSE) as the treated eye. C, Baseline VA according to BSE or WSE as the study eye. D, Baseline central retinal (subfield) thickness (CRT). E, Median NEI VFQ-25 composite score. Error bars indicate upper 95% confidence limit. aP < .05.
eFigure 1. Percentage of patients with 5-point gain or more (a) or a 5-point loss or more (b) from baseline in National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) composite score by treatment arm at 3 and 12 months.
eFigure 2. Percentage of patients with 10-point gain or more (a) or a 10-point loss or more (b) from baseline in National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) composite score by treatment arm at 3 and 12 months.
eTable 1. Mean Change in National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Scores From Baseline to Month 3 and From Baseline to 12 Months
eTable 2. Mean Change From Baseline to 12 Months in Composite and Selected National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Scores by Better- or Worse-Seeing Eye as the Treated Eye
eTable 3. Mean Change From Baseline to 12 Months in National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Composite Score for Better- or Worse-Seeing Eye as the Treated Eye and by Baseline Visual Acuity
eTable 4. Mean Change From Baseline to 12 Months in Composite and Selected National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Scores by Baseline Central Retinal Thickness (CRT)a
eTable 5. Mean Change From Baseline to 12 Months in Composite National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Scores by Median Baseline Composite NEI VFQ-25 Score
eTable 6. Regression Results for All Patients for National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Composite Score and Selected Subscale Scores
eTable 7. Regression Results for Patient Subgroups for National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) Composite Score
Mitchell P, Bressler N, Tolley K, et al; RESTORE Study Group. Patient-reported visual function outcomes improve after ranibizumab treatment in patients with vision impairment due to diabetic macular edema: randomized clinical trial. JAMA Ophthalmol. Published online August 22, 2013. doi:10.1001/jamaophthalmol.2013.4592.
Mitchell P, Bressler N, Tolley K, Gallagher M, Petrillo J, Ferreira A, Wood R, Bandello F, for the RESTORE Study Group. Patient-Reported Visual Function Outcomes Improve After Ranibizumab Treatment in Patients With Vision Impairment Due to Diabetic Macular EdemaRandomized Clinical Trial. JAMA Ophthalmol. 2013;131(10):1339-1347. doi:10.1001/jamaophthalmol.2013.4592
Few data are available on relative changes in vision-related function after treatment for diabetic macular edema (DME).
To determine the impact of intravitreal ranibizumab, 0.5 mg, compared with laser on patient-reported visual function.
Phase 3, randomized, double-masked, 12-month study (RESTORE).
Outpatient retina practices in Australia, Canada, and Europe.
Patients 18 years or older with type 1 or 2 diabetes mellitus and visual impairment due to DME.
Patients were randomized to ranibizumab plus sham laser (n = 116), ranibizumab plus laser (n = 118), or sham injections plus laser (n = 111). Ranibizumab and sham injections were given for 3 consecutive months then as needed; laser or sham laser treatment was given at baseline then as needed.
Main Outcomes and Measures
National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25) scores at 0, 3, and 12 months for patients receiving 1 or more study treatments with 1 or more postbaseline NEI VFQ-25 assessments and last observation carried forward for missing data.
Mean baseline NEI VFQ-25 composite scores were 72.8, 73.5, and 74.1 in the ranibizumab, laser, and ranibizumab plus laser groups, respectively. At 12 months, the mean composite scores (95% CIs) improved by 5.0 (ranibizumab vs laser, 2.6 to 7.4; P = .01 vs laser) and 5.4 (ranibizumab plus laser vs laser alone, 3.3 to 7.4; P = .004 vs laser) from baseline in the ranibizumab and ranibizumab plus laser groups, respectively, compared with 0.6 (−1.8 to 3.0) for the laser group. Near activities scores improved by 9.0 (ranibizumab vs laser, 5.0 to 13.0; P = .01) and 9.1 (ranibizumab plus laser vs laser, 5.6 to 12.6; P = .006) compared with 1.1 (−3.0 to 5.2) for the laser group, whereas distance activities scores improved by 5.3 (ranibizumab vs laser, 1.8 to 8.9; P = .04) and 5.6 (ranibizumab plus laser vs laser, 2.3 to 9.0; P = .03) compared with 0.4 (−3.1 to 3.8) for the laser group. Patients with better baseline visual acuity or lower central retinal thickness had greater improvements with ranibizumab treatment compared with laser in composite and some subscale scores compared with patients with worse visual acuity or higher central retinal thickness.
Conclusions and Relevance
These data provide vision-related, patient-reported outcome evidence that mirrors visual acuity outcomes and supports benefits from ranibizumab or ranibizumab plus laser treatment for patients with DME and characteristics similar to those enrolled in this randomized clinical trial.
clinicaltrials.gov Identifier: NCT00687804
Diabetic macular edema (DME) is a major cause of visual impairment in patients with diabetic retinopathy.1- 3 Diabetic macular edema has been reported to negatively affect patients' general health and self-reported visual functioning.4- 7 Effects of treatment on patient-reported visual functioning have been demonstrated for both ranibizumab and pegaptanib in neovascular age-related macular degeneration (AMD).8- 10 However, similar evidence for treatment of DME had been lacking before recent findings from the Ranibizumab Monotherapy or Combined With Laser vs Laser Monotherapy for Diabetic Macular Edema (RESTORE) trial.11
The 12-month, randomized, active-controlled RESTORE trial reported superior visual acuity (VA) outcomes for ranibizumab compared with laser treatment in patients with DME.11 Secondary efficacy end point measures included the National Eye Institute Visual Functioning Questionnaire 25 (NEI VFQ-25), a validated condition-specific instrument that provides a patient-reported measure of visual functioning.12- 16 The NEI VFQ-25 composite score demonstrated a substantial improvement from baseline to 12 months for ranibizumab alone or when combined with laser treatment compared with laser alone.11 This article provides in-depth analyses of the NEI VFQ-25 subscale scores and subgroup analyses by better-seeing eye (BSE) or worse-seeing eye (WSE), baseline VA, central retinal (subfield) thickness (CRT) measures on optical coherence tomography, and baseline NEI VFQ-25 composite scores.
Methods for the RESTORE trial (clinicaltrials.gov Identifier: NCT00687804) have been reported previously.11 In brief, the trial enrolled 345 adult patients across 73 outpatient retina practices in Australia, Canada, and 11 European countries. Patients had type 1 or 2 diabetes mellitus and visual impairment due to DME in one or both eyes, with an Early Treatment Diabetic Retinopathy Study best-corrected visual acuity (BCVA) letter score between 78 and 39 (approximate Snellen equivalent of 20/32 to 20/160). Patients were assigned randomly in a 1:1:1 ratio to treatment with intravitreal ranibizumab, 0.5 mg, and sham laser (ranibizumab monotherapy; n = 116); combined ranibizumab, 0.5 mg, and laser (ranibizumab plus laser; n = 118); or laser treatment and a sham injection (laser alone; n = 111). Ranibizumab or sham injections were administered monthly at baseline and months 1 and 2, with further ranibizumab or sham injections according to protocol-defined retreatment criteria (primarily VA stability) up to month 12.11 Laser treatment and sham were administered on day 1, with subsequent treatments at intervals no shorter than 3 months, if deemed necessary by the investigator.
The NEI VFQ-25 is a validated questionnaire that provides a subjective assessment of visual functioning.12 It has been used widely to assess visual functioning across a range of vision-related conditions, including AMD and DME.13 The NEI VFQ-25 includes 25 questions across 11 vision-related constructs (general vision, ocular pain, near activities, distance activities, social function, mental health, role difficulties, dependency, driving, color vision, and peripheral vision) plus a general health rating and provides a composite score and scores for each subscale. Patients completed the questionnaire at baseline and 3 and 12 months. In this article we focus on the results that relate to the composite score and 3 principal subscales (near activities, distance activities, and general vision) previously reported to correlate well with VA.12 The near and distance activities subscales were prespecified analyses of ranibizumab in vision-related patient-reported outcomes in AMD.9,10,13 Higher scores represent better functioning on a 0- to 100-point scale for the composite and each subscale.12
No published benchmark is available for a clinically meaningful change in NEI VFQ-25 scores in patients with DME. Analysis of AMD trial data suggests that a 4- to 6-point improvement in mean composite NEI VFQ-25 scores represents a clinically meaningful change corresponding to at least a 15-letter change in BCVA.13 However, some studies9 have used a 10-point difference as a marker for clinically meaningful change in AMD. In this study, a 5-point or more gain or loss in NEI VFQ-25 composite score was applied to delineate the proportion of patients within each treatment group who achieved a clinically relevant improvement or deterioration in overall vision-related functioning.
All analyses were performed using the full analysis set, including patients who received at least one study treatment and had at least one postbaseline assessment. The last observation carried forward method was used for missing postbaseline data. Analyses, based on the treated study eye only, were performed for the composite and selected subscales for mean change from baseline in NEI VFQ-25 outcomes. Multiple subgroups were defined.
The first subgroup was defined according to whether the treated eye was the BSE (defined as such when the BCVA letter score at baseline was higher than that for the nonstudy eye), the WSE (defined as such when the BCVA letter score at baseline was worse than that for the nonstudy eye), or neither a BSE nor WSE.8
The second subgroup was defined by whether the BCVA letter score at baseline was greater than 68 (approximate Snellen equivalent >20/50) or 68 or less (corresponding to a threshold for mild or no visual impairment with no impairment of daily activities, such as driving) on the Early Treatment Diabetic Retinopathy Study scale.
The third subgroup was defined by whether baseline CRT measurements on optical coherence tomography were of 400 μm or less or greater than 400 μm. Finally, the fourth subgroup was defined by whether the composite score was greater than the median baseline NEI VFQ-25 composite score.
Multilevel mixed-effects linear regression analyses were also performed to investigate the effect of patient characteristics on the NEI VFQ-25 composite and subscale scores for near activities, distance activities, and general vision. This method accounts for intrapatient correlation arising from multiple observations per patient.17 The NEI VFQ-25 analysis score included time, treatment, and the interaction of time and treatment to indicate whether changes in NEI VFQ-25 scores were associated with ranibizumab treatment (with or without laser) relative to laser alone. The regression analysis evaluated the effect on NEI VFQ-25 outcomes of baseline BCVA; whether the treated eye was the BSE, WSE, or neither; age; sex; glycosylated hemoglobin levels; CRT; and baseline NEI VFQ-25 score. P values were generated using Wald tests. The NEI VFQ-25 regression analyses were exploratory and, therefore, include no adjustments for multiplicity.
Patient demographics and clinical characteristics were comparable across the 3 treatment groups (Table 1). Vision-related baseline characteristics were also similar across the 3 groups, although slightly more patients in the ranibizumab monotherapy group had the BSE as the study eye (Table 1). Patient disposition for the completion of NEI VFQ-25 is shown in Figure 1.
The mean NEI VFQ-25 composite and subscale scores were comparable across the 3 treatment groups at baseline (Table 1). The mean composite score change from baseline was 5.0 points (95% CI, 2.6 to 7.4; P = .01 vs laser) and 5.4 points (95% CI, 3.3 to 7.4; P = .004 vs laser) for ranibizumab monotherapy and ranibizumab plus laser, respectively, at the 12-month follow-up visit, compared with 0.6 point (95% CI, –1.8 to 3.0) for laser alone (eTable 1 in the Supplement and Figure 2A). Greater increases for ranibizumab monotherapy compared with laser alone at the 12-month follow-up visit also were found for the near activities, distance activities, and general vision subscale scores. Of the other NEI VFQ-25 subscales, the score gain at the 12-month follow-up visit was greater for ranibizumab monotherapy compared with laser alone on the role difficulties (P = .05) and social functioning (P = .03) subscales (eTable 1 in the Supplement). As shown for the composite score and the near activities, distance activities, and general vision subscale scores in Figure 2B and C, the greater improvement in the ranibizumab groups compared with laser alone at the 12-month follow-up visit was apparent by the 3-month follow-up visit (eTable 1 in the Supplement). In the laser alone group, short-term improvements in vision-related function observed at 3 months were largely lost by the 12-month follow-up visit.
Overall, 49.1% of patients receiving ranibizumab monotherapy achieved at least a 5-point improvement in NEI VFQ-25 composite score at the 12-month follow-up visit compared with 42.7% of patients assigned to ranibizumab plus laser and 30.6% assigned to laser alone (eFigure 1A in the Supplement). Over time, the proportion of patients in the laser alone group achieving a 5-point or greater gain in NEI VFQ-25 composite score decreased, whereas the proportion increased in the ranibizumab monotherapy and ranibizumab plus laser groups. The converse was seen for the proportion of patients with a 5-point or greater loss (eFigure 1B in the Supplement). A similar pattern was observed for those patients with a gain or loss of 10 points or more (eFigure 2A and B in the Supplement).
The differences in the change in NEI VFQ-25 composite score from baseline to month 12 were greater for the subgroup of patients with a baseline BCVA score of 68 or higher (approximate Snellen equivalent >20/50) compared with less than 68 (worse than 20/50) among patients receiving ranibizumab with or without laser (>6-point change at 12 months) compared with laser alone (−1.6-point change at 12 months) (Table 2 and Figure 3A). Results for the subscales investigated reveal a similar pattern, with greater improvements on the near activities and general vision subscales for the subgroup with better baseline VA (Table 2).
The treated eye was the WSE at baseline in more than 50% of patients in all treatment groups, the BSE in approximately 20% of patients, and neither the BSE nor WSE in the remaining patients (Table 1). Overall, the change from baseline in NEI VFQ-25 results did not differ if the treated eye was the WSE or BSE. Greatest improvements for ranibizumab and ranibizumab plus laser compared with laser alone, however, were for the near activities score when the BSE was the treated eye and for the general vision score when the WSE was the treated eye (eTable 2 in the Supplement). Multivariate regression analyses confirmed this effect even after adjusting for potential confounders, although the reliability of findings for this level of subgroup disaggregation is affected by small sample sizes (eTables 2 through 5 in the Supplement). The subgroup analyses revealed a greater mean change from baseline score for ranibizumab and ranibizumab plus laser compared with laser alone for those patients whose study eye was the BSE with a baseline BCVA letter score of 68 or higher (approximate Snellen equivalent >20/50, eTable 3 in the Supplement and Figure 3C). A smaller improvement in composite scores in the ranibizumab treatment group compared with laser alone group was found when the treated eye was the WSE with a baseline BCVA letter score of 68 or higher (approximate Snellen equivalent >20/50, eTable 3 in the Supplement and Figure 3C). No substantial differences among treatment groups for the BSE or WSE and baseline BCVA letter score less than 68 (approximate Snellen equivalent <20/50) subgroups were identified. There was also a greater likelihood of an improvement in the ranibizumab and ranibizumab plus laser groups for the composite score and for the near activities and general vision subscales for the subgroup with a CRT of 400 μm or less compared with a CRT greater than 400 μm (eTable 4 in the Supplement and Figure 3D). An analysis of the change in composite score for patients in subgroups by baseline composite NEI VFQ-25 score revealed the greatest improvement for both the ranibizumab and ranibizumab plus laser groups over laser alone in the subgroup with a median baseline NEI VFQ-25 composite score of 78 or higher (P = .009 for ranibizumab vs laser and P = .005 for ranibizumab plus laser) (eTable 5 in the Supplement and Figure 3E).
The regression analysis suggests that after controlling for confounding factors, treatment with ranibizumab or ranibizumab plus laser at 12 months appeared to be a greater predictor of higher composite, near activities, distance activities, and general vision NEI VFQ-25 scores compared with laser alone (eTable 6 in the Supplement). For the composite score and selected subscales scores, baseline scores were positively associated with scores at 12 months (P < .001), and for the near activities, distance activities, and general vision subscales, baseline VA was a positive predictor of NEI VFQ-25 outcomes (eTable 6 in the Supplement). After adjusting for the confounders in the regression analyses, treating the WSE appeared to predict better scores relative to the BSE for the near activities and general vision subscales but not for the composite or distance activities subscale (eTable 6 in the Supplement). In addition, ranibizumab monotherapy vs laser alone or ranibizumab plus laser vs laser alone at 12 months appeared to have a greater effect on the composite score in patients with a baseline BCVA letter score of 68 or higher (approximate Snellen equivalent >20/50; P = .006) in contrast with the unadjusted univariate results and in patients with a CRT of 400 μm or less (P = .25) in agreement with unadjusted univariate results, but this was not the case in patients in subgroups with relatively worse BCVA or greater CRT (eTable 7 in the Supplement). Baseline NEI VFQ-25 score was a predictor of the composite score at 12 months in all subgroups. However, baseline VA also appeared to be a predictor of the composite score in the subgroup with a CRT greater than 400 μm (P = .25). Whether the treated eye was BSE, WSE, or neither did not predict outcomes for the subgroups defined by BCVA or CRT (eTable 7 in the Supplement).
A primary finding from the RESTORE trial was a statistically significant improvement in BCVA for patients with DME treated with ranibizumab either as monotherapy or in combination with laser therapy compared with laser therapy alone, as reported previously.11 The trial also included the NEI VFQ-25 questionnaire, one of the most frequently used tools to measure patient-reported, vision-related function in DME and other retinal diseases.13- 16 Detailed analysis of the patient-reported visual functioning data reveals that improvements in the NEI VFQ-25 composite score and in the near activities, distance activities, and general vision subscales were more likely for ranibizumab alone or combined with laser than with laser alone. These outcomes were similar, but of greater magnitude, to an analysis of randomized controlled trial data for pegaptanib treatment.18
In the RESTORE trial, approximately half of the patients achieved at least a 5-point improvement in the NEI VFQ-25 composite score with ranibizumab monotherapy and 43% with ranibizumab plus laser at the 12-month follow-up visit compared with 31% with laser alone. Greater improvements from baseline for ranibizumab monotherapy compared with laser alone were also noted on the subscales for near activities (visual activities, including reading ordinary print in a newspaper and close-up work or hobbies, such as cooking or using hand tools) and general vision (perception of how good their eyesight is). The pattern of relative improvement on the distance activities subscale (visual activities, such as seeing movies or a sports event and reading street signs) was similar to that for the composite score, with a 5-point or greater gain for ranibizumab monotherapy and only a small benefit compared with 12 months for laser alone (0.4 point). Overall, the improvements in vision-related outcomes with ranibizumab monotherapy generally increased from the 3-month to the 12-month follow-up visit. The improvement in the laser alone group at the 3-month follow-up visit was not sustained at the 12-month follow-up visit. The results for a 5-point or greater loss indicate that approximately one-quarter of the patients treated with laser alone experience this outcome at 12 months, whereas the rate is approximately half this for patients receiving ranabizumab plus laser treatment and lower also for those receiving ranibizumab monotherapy.
Other studies (not in DME) have reported a moderate correlation between baseline BCVA and NEI VFQ-25 outcomes.12,19 Patients in the RESTORE trial typically had a baseline BCVA that corresponds to moderate visual impairment, which could affect daily functioning and ability to perform usual activities (eg, the mean BCVA letter score in the trial was 63.5, and as a benchmark the legal limit to drive in the United Kingdom and other European countries is approximately a BCVA letter score of 68). In subgroup analyses, patients with baseline BCVA letter scores of 68 or higher (approximate Snellen equivalent >20/50) had the best NEI VFQ-25 outcomes with ranibizumab therapy, demonstrating a greater than 6-point gain at 12 months on the composite score (compared with a loss of 1.6 points with laser alone), and a greater than 10-point gain on the near activities subscale (compared with −0.9 for laser alone). In contrast, those with worse baseline letter scores did not demonstrate such substantial gains for ranibizumab monotherapy compared with laser alone at 12 months. These subgroup findings suggest that patients with relatively less severe vision impairment at baseline may experience greater treatment benefit on their quality of life, the rationale being that it is possible for these patients to move from a level of BCVA impairment that disrupts daily functioning to a level of BCVA that restores the ability to perform activities, such as reading, cooking, and driving. Alternatively, because eyes with better baseline VA were more likely to be the BSEs, this would lead to eyes with better baseline VA showing greater improvements in patient-reported vision outcomes. Although the separation into the BSE or WSE subgroups for assessment may seem artificial, Rubin et al20 reported that it is the BCVA in the BSE that best predicts binocular BCVA. The association of smaller improvements in patient-reported vision function outcomes in patients with greater CRT thickness (>400 μm vs ≤400 μm) could also potentially be related to the correlation of thicker edema with worse VA at baseline.
Exploratory regression analysis demonstrated that treatment with ranibizumab alone or in combination with laser at 12 months had a substantial effect on NEI VFQ-25 composite and selected subscale outcomes vs laser alone, even after controlling for confounding factors.
A key strength of this analysis is that NEI VFQ-25 data for patients with DME were available from a large, well-designed randomized controlled trial with an active comparator (laser therapy alone) that reflects standard clinical practice in most countries. The NEI VFQ-25 measures improvement in vision-related functioning and, therefore, provides a useful proxy for improvements in health-related quality of life.
A general limitation of the study is that the patient-reported outcomes data were derived from a controlled clinical trial, in which the study population may not be representative of patients with DME in the real-world setting. There were apparent differences in favor of ranibizumab therapy with or without laser across several NEI VFQ-25 subscales, although differences over laser alone were not found for all subscales, partly owing to lack of statistical power. Notably, the dependency subscale, which was previously used for analyses in AMD studies, did not find substantial differences between ranibizumab and laser therapy.9,10 Finally, assessment of changes in visual functioning outcomes over time was restricted to analyses at 3 and 12 months, unlike BCVA outcomes, which were recorded on a monthly basis. Long-term follow-up data from the RESTORE trial are required to assess whether the 12-month visual-functioning benefits found with ranibizumab are sustained during the 3-year period of the trial.
In conclusion, this study provides some of the first evidence, to our knowledge, that ranibizumab therapy increases the likelihood of clinically relevant improvements in patient-reported visual functioning, compared with laser alone, for eyes with vision impairment due to DME during a 12-month follow-up period using NEI VFQ-25 composite scores. Subscale analyses suggest that these benefits are most pronounced for the near and general vision subscales. Additional analyses are ongoing regarding patient responses to individual questions from the NEI VFQ-25. Such analyses may be useful for exploring key aspects of everyday living that are more likely to be improved with ranibizumab therapy than laser.
Submitted for Publication: March 28, 2013; final revision received April 1, 2013; accepted April 1, 2013.
Corresponding Author: Paul Mitchell, MD, PhD, Department of Ophthalmology, The University of Sydney, Eye Clinic, Westmead Hospital, Hawkesbury Road, Westmead, New South Wales 2145, Australia (email@example.com).
Author Contributions:Study concept and design: Mitchell, Bressler, Tolley, Gallagher, Ferreira, Bandello.
Acquisition of data: Mitchell, Tolley.
Analysis and interpretation of data: All authors.
Drafting of the manuscript: Mitchell, Bressler, Tolley, Petrillo, Ferreira, Wood.
Critical revision of the manuscript for important intellectual content: Mitchell, Bressler, Tolley, Gallagher, Petrillo, Ferreira, Bandello.
Statistical analysis: Tolley, Gallagher, Petrillo, Ferreira, Wood.
Obtained funding: Mitchell.
Administrative, technical, and material support: Mitchell, Tolley, Petrillo, Ferreira.
Study supervision: Mitchell, Bandello.
Published Online: August 22, 2013. doi:10.1001/jamaophthalmol.2013.4592.
Conflict of Interest Disclosures: Dr Mitchell reported receiving consultancy fees, lecture fees, and travel support from Novartis Pharma AG, Pfizer, Solvay (Abbott), Bayer, Alcon, and Allergan. Novartis Pharma AG also funds a retina fellowship at Westmead Hospital, Sydney, which Dr Mitchell reported serving as supervisor. Dr Bressler reported serving as principal investigator of grants at The Johns Hopkins University sponsored by the following entities (not including the National Institutes of Health): Bayer, Genentech Inc, Lumenis Inc, National Eye Institute, Notal Vision, Novartis Pharma AG, Optovue, Regeneron Pharmaceuticals Inc, and The EMMES Corporation. Messers Tolley and Wood reported receiving consultant fees by Novartis Pharma AG to analyze these data. Ms Gallagher reports being a former employee of Novartis Pharma AG, Basel, Switzerland, and currently employed by Sanofi, Cambridge, Massachusetts. Drs Petrillo and Ferreira report being employees of Novartis Pharma AG. Dr Bandello reports serving as an advisory board member for Alcon, Inc, Alimera Sciences, Inc, Allergan, Inc, Bausch and Lomb, Bayer Schering Pharma, Farmila-Thea, Genentech, Inc, Hoffmann-LaRoche, Ltd, Pfizer Inc, Novartis Pharmaceuticals Corporation, sanofi-aventis, and Thrombogenics, Inc.
Group Information: The RESTORE study group members include Alfredo Adan Civera, MD, Barcelona, Spain; Cezmi Akkin, MD, Izmir, Turkey; Clare Bailey, MD, Bristol, United Kingdom; Emilio Balestrazzi, MD, Rome, Italy; Nicholas Beare, MD, Liverpool, United Kingdom; Arpad Berecki, PhD, Györ, Hungary; Andras Berta, MD, PhD, Debrecen, Hungary; Shelley Boyd, MD, Toronto, Ontario, Canada; Periklis Brazitikos, MD, PhD, Thessaloniki, Greece; Enrique Cervera, MD, Valencia, Spain; John Chen, MD, Montreal, Quebec, Canada; Catherine Creuzot-Garcher, MD, PhD, Dijon, France; Nicola Delle Noci, MD, Foggia, Italy; Marlene Devriendt, MD, Jette, Belgium; Stavros Dimitrakos, MD, PhD, Thessaloniki, Greece; Bora Eldem, MD, Ankara, Turkey; Karl-Heinz Emmerich, MD, Darmstadt, Germany; Katrin Engelmann, MD, PhD, Chemnitz, Germany; Mehmet Ergin, MD, Izmir, Turkey; Johannes Fleischhauer, MD, Bern, Switzerland; Josep Garcia-Arumi, MD, Barcelona, Spain; Justus Garweg, MD, Bern, Switzerland; Heinrich Gerding, MD, Olten, Switzerland; Mark Gillies, MBBS, PhD, Sydney, Australia; Francisco Gomez Ulla, MD, PhD, Santiago de Compostela, Spain; John Gonder, MD, London, Ontario, Canada, Salvatore Grisanti, MD, Lübeck, Germany; Lutz L. Hansen, MD, PhD, Freiberg, Germany; Berati Hasanreisoglu, MD, Ankara, Turkey; Helmut Hoeh, MD, FEBO, Neubrandenburg, Germany; Frank G. Holz, MD, Bonn, Germany; Vergados Ioaanis, MD, PhD, Athens, Greece; Antonia Joussen, MD, PhD, Duesseldorf, Germany; Stamatina Kabanarou, MD, PhD, Athens, Greece; Anselm Kampik, MD, Muenchen, Germany; Ziya Kapran, MD, Istanbul, Turkey; Ulrich Kellner, MD, Siegburg, Germany; Agnes Kerényi, MD, PhD, Budapest, Hungary; Peter Kertes, MD, FRCS, Toronto, Ontario, Canada; Bernd Kirchhof, MD, Koeln, Germany; B. J. Klevering, MD, PhD, Nijmegen, the Netherlands; Patrik Kloos, MD, FEBO, St Gallen, Switzerland; Vladimir Kozousek, MD, Halifax, Nova Scotia, Canada; Malaika Kurz-Levin, MD, Zurich, Switzerland; Gabriele E. Lang, MD, Ulm, Germany; David Maberley, MD, Vancouver, British Columbia, Canada; J. P. Martinez Ciriano, MD, Rotterdam, the Netherlands; Pascale Massin, MD, PhD, Paris, France; Ugo Menchini, MD, Firenze, Italy; Geeta Menon, Frimley, United Kingdom; Paul Mitchell, MD, PhD, Westmead, Australia; Andreas Mohr, MD, Bremen, Germany; Sebastien Olivier, MD, Montreal, Quebec, Canada; Andras Papp, MD, PhD, Budapest, Hungary; Alfredo Reibaldi, MD, PhD, Catania, Italy; Jean-Paul Romanet, MD, Grenoble, France; Jose M. Ruiz Moreno, MD, PhD, Alicante, Spain; Carlo Sborgia, MD, Bari, Italy; R. O. Schlingemann, MD, PhD, Amsterdam, the Netherlands; Ursula Schmidt-Erfurth, MD, Vienna, Austria; Ulrike Schneider, MD, Basel, Switzerland; Norbert Schrage, MD, Koeln, Germany, Andras I. Seres, MD, Budapest, Hungary; Dianne Sharp, MD, Auckland, New Zealand; Christian Simader, MD, Vienna, Austria; Gungor Sobaci, MD, Etlik/Ankara, Turkey; Georg Spital, MD, Meunster, Germany; Florian Sutter, MD, Zurich, Switzerland; Christoph Tappeiner, MD, Bern, Switzerland; Anna Tarantini, MD, Udine, Italy; Miltiadis Tsilibaris, MD, Heraklion/Crete, Greece; Ramon Torrez Imaz, MD, Madrid, Spain; Joachim Van Calster, MD, Leuven, Belgium; Brendan Vote, MD, MBBS, Launceston, Australia; Michel Weber, MD, PhD, Nantes, France; Peter Wiedemann, MD, Leipzig, Germany; Tien Wong, MD, PhD, Melbourne, Australia.
Additional Contributions: Daniel Shaw, BSc, (formerly of the Adelphi Group) provided statistical support. Polly Field, DPhil (Oxford PharmaGenesis Ltd), Rumjhum Agrawal , and Hareesh Cheela provided technical assistance, and Emma Boulton (Tolley Health Economics) provided administrative support in preparation of the manuscript.
Correction: This article was corrected on November 14, 2013, to fix Table 1, Figures 2 and 3, and the text.