Retina Topical Treatment With Brimonidine and Somatostatin Causes Retinal Vascular Dilation in Patients With Early Diabetic Retinopathy From the EUROCONDOR Jakob Grauslund,1–3 Ulrik Frydkjaer-Olsen,1,2 Tunde Peto,2,4 Jimena Fernández-Carneado,5 Berta Ponsati,5 Cristina Hernández,6 José Cunha-Vaz,7 and Rafael Simó6; for the EUROCONDOR 1 Department of Ophthalmology, Odense University Hospital, Odense, Denmark 2 Department of Clinical Research, University of Southern Denmark, Odense, Denmark 3 Steno Diabetes Center Odense, Odense University Hospital, Odense, Denmark 4 Centre for Public Health, Institute of Clinical Sciences, Queen’s University Belfast, Belfast, United Kingdom 5 BCN Peptides, Barcelona, Spain 6 Diabetes and Metabolism Research Unit, Institut de Recerca Hospital Universitari Vall d’Hebron (VHIR), Barcelona, Spain 7 Association for Innovation and Biomedical Research on Light and Image (AIBILI), and University of Coimbra, Coimbra, Portugal Correspondence: Jakob Grauslund, PURPOSE. Structural retinal microvascular changes have been identified as risk markers of Department of Ophthalmology, diabetic retinopathy (DR). In order to estimate the retinal response of neuroprotective eye Odense University Hospital, Sdr. drops, we aimed to evaluate the effect of topical retinal neuroprotection on retinal Boulevard 29, Odense DK-5000, microvascular changes in early DR. Denmark; jakob.grauslund@rsyd.dk. METHODS. Patients with type 2 diabetes with no or early DR were randomized 1:1:1 to topical Please see appendix for members of treatment with placebo, brimonidine, or somatostatin in a 96-week prospective, phase II to the EUROCONDOR. III, European multicenter trial. Retinal vascular calibers were measured semiautomatically in digital fundus images by certified graders at baseline and follow-up and summarized as central Submitted: December 20, 2018 Accepted: April 23, 2019 retinal arteriolar and venular equivalent (CRAE and CRVE). Citation: Grauslund J, Frydkjaer-Olsen RESULTS. Of 449 patients originally included, 297 completed the study with gradable retinal U, Peto T, et al.; for the EUROCON- images. Median age and duration of diabetes was 64.5 and 9.9 years, and 65.7% were male. At DOR. Topical treatment with brimo- baseline, Early Treatment Diabetic Retinopathy Study levels were 10 (no DR, 42.8%), 20 nidine and somatostatin causes retinal (minimal DR, 28.3%), and 35 (mild DR, 29.0%), and CRAE and CRVE did not differ between vascular dilation in patients with early groups. As opposed to patients with no or minimal DR at baseline, patients with mild DR in diabetic retinopathy from the EURO- the active groups developed a larger retinal arteriolar (brimonidine: þ6.2 lm, P ¼ 0.006; CONDOR. Invest Ophthalmol Vis Sci. somatostatin: þ7.2 lm, P ¼ 0.006) and venular (brimonidine: þ13.9 lm, P ¼ 0.01; 2019;60:2257–2262. https://doi.org/ somatostatin: þ14.3 lm, P ¼ 0.0001) caliber in contrast to those in the placebo group. 10.1167/iovs.18-26487 CONCLUSIONS. Topical treatment with brimonidine and somatostatin causes retinal arteriolar and venular dilation in patients with type 2 diabetes and preexisting early DR. Upcoming studies should elaborate on the potential of these findings in arresting early DR. Keywords: diabetic retinopathy, topical, retinal vasculature, brimonidine, somatostatin D iabetic retinopathy (DR) is the most common complication in diabetes1 and among the leading cause of blindness among working age adults in the Western world.2 DR has topically administered brimonidine and somatostatin on retinal neurodysfunction in patients with type 2 diabetes with no or early DR.11 The study reported that in patients with preexisting traditionally been considered a microvascular disease, but retinal neurodysfunction, topical neuroprotection arrested recent evidence has implicated retinal neurodegeneration as further loss of retinal function as measured by multifocal an early event that may even precede vascular dysfunction.3,4 electroretinography implicit time.12 At present, retinal photocoagulation and intravitreal therapy It would also be important to investigate the structural with vascular endothelial growth factor inhibitors are consid- vascular effects of topical neuroprotective treatment. However, ered the gold standard treatment in DR5–7; but given the as demonstrated by the Diabetes Control Complications Trial invasive nature of the treatments, these are only considered in (DCCT), even a strong systemic intervention like strict patients with sight-threatening DR. glycemic regulation would take 3 years to affect the level of Topical treatment with a neuroprotective agent would be an DR.13 Hence, more subtle preclinical endpoints for vascular appealing approach to prevent or arrest DR at an early stage. dysfunction are needed. In this aspect, the retinal vascular This concept has been supported by experimental data calibers, which can be measured noninvasively, have been demonstrating a prevention of retinal neurodegeneration by associated with14,15 and even predictive of sight-threatening topical administration of neuroprotective drugs.8–10 The DR.16 European Consortium for the Early Treatment of Diabetic In order to explore the effect of topical neuroprotection on Retinopathy (EUROCONDOR) aimed to test the effect of the retinal vasculature of patients with type 2 diabetes and early Copyright 2019 The Authors iovs.arvojournals.org j ISSN: 1552-5783 2257 This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License. Downloaded from iovs.arvojournals.org on 02/16/2021 Topical Neuroprotective Treatment in Diabetes Patients IOVS j May 2019 j Vol. 60 j No. 6 j 2258 DR, the aim of the present study was to evaluate the result of 2 up. Of the 153 patients that were not included in the present years of treatment with brimonidine and somatostatin on the study, reasons for noninclusion included dropout during study retinal vascular calibers in a randomized, clinical trial. (n ¼ 109), lack of retinal images at follow-up (n ¼ 17), and nongradable retinal images at baseline (n ¼ 10) or follow-up (n ¼ 17). Among nonparticipants, there was a higher number of METHODS patients allocated to treatment with brimonidine (43.4% vs. 28.6%, P ¼ 0.009), but otherwise participants and nonpartic- Study Subjects ipants did not differ according to age, duration of diabetes, sex, The EUROCONDOR [NCT01726075] was a European multi- glycemic regulation, blood pressure, body mass index, level of center, 96-week prospective, interventional, phase II to III, DR, CRAE, or CRVE (Table 1). randomized controlled clinical trial that aimed to evaluate the effect of topical neuroprotection to arrest or prevent early Statistical Methods retinal neurodegeneration in DR.11 In a study including 11 European centers, a total of 449 patients were recruited and Descriptive statistics were calculated for all parameters. randomized 1:1:1 to topical treatment twice daily (BID) with Continuous data are presented as median (with range), and placebo, brimonidine tartrate 0.2%, or somatostatin 0.1%. categoric data are given as percentage. Development in retinal Criteria of inclusion were: type 2 diabetes with no, minimal, or vascular calibers and level of DR were given as differences mild DR (Early Treatment Diabetic Retinopathy Study [ETDRS] between follow-up and baseline. levels 10, 20, or 35, respectively),17 duration of type 2 diabetes To test for differences in unpaired, continuous data, we for at least 5 years, and age between 45 and 75 years. Exclusion used nonparametric tests including Mann-Whitney U test (two criteria included previous retinal photocoagulation, retinal groups), Kruskal-Wallis test (multiple, nonordered groups) and degenerative diseases (i.e., glaucoma), and refractive error or Cuzick’s test for trend (multiple, ordered groups). Wilcoxon 65 diopters or more, hazy ocular media, inadequate pupil matched-pairs signed-rank test was used to test for differences dilation, renal failure (creatinine >124 lm/L) or HbA1c >10% between baseline and follow-up in matched paired of (86 mmol/mol) in the previous 6 months. observations. Finally, v2 test was used to test for differences For each patient, one eye was identified as the study eye by in categorical data. a central reading center (Coimbra Ophthalmology Reading Statistical analysis was performed by statistical software Centre, Coimbra, Portugal), which also provided grading of the (Stata 15.1; StataCorp LLC, College Station, TX, USA), and P level of DR at baseline and follow-up. values below 0.05 were considered statistically significant. The study was approved and funded by the European Commission Seventh Framework Program (Grant Agreement No. FP7–278040). At all centers, the study was conducted in RESULTS accordance with the tenets of the Declaration of Helsinki with Among the 297 participants of the study, median age and approvals of the local scientific ethnical committees and duration of diabetes at baseline was 64.5 and 9.9 years, and written informed consent from all patients. 65.7% were male. Median levels of HbA1c, blood pressure, and body mass index were 7.0%, 135/78 mm Hg and 30.0 kg/m2, Retinal Vascular Caliber Analysis respectively. Baseline ETDRS levels of DR were 10 (42.8%), 20 Mydriatic cameras (TRC-50IA, -50IX, -50EX, -50DX, -50DX Type (28.3%), or 35 (29.0%), and CRAE and CRVE was 146.6 lm and IA, or NW6S; Topcon Corp., Tokyo, Japan) or a fundus camera 214.8 lm, respectively. (FF450 PLUS IR; Carl Zeiss Meditec, Jena, Germany) were used Number of patients in each treatment group were 108 to capture 308 to 358 modified ETDRS-field 1 images including (36.4%), 85 (28.6%), and 104 (35.0%) for placebo, brimonidine, the optic disc of the study eye. These were subsequently used and somatostatin, respectively. Among treatment groups, for retinal vascular caliber measurements.18 Images were patients did not differ at baseline according to age, sex, analyzed by IVAN (Department of Ophthalmology and Visual duration of diabetes, HbA1c, blood pressure, body mass index, Science, University of Wisconsin, Madison, WI, USA), which is CRAE (146.7 lm, 147.1 lm, and 145.2 lm, P ¼ 0.68) or CRVE a semiautomatic software program that has previously been (215.6 lm, 212.7 lm, and 215.7 lm, P ¼ 0.89; Table 2). validated.19 Likewise, there was no difference in CRAE according to The optic disc is automatically detected by the software, baseline level of DR (level 10, 144.1 lm; level 20, 147.7 lm; and vessel calibers are traced in a zone of 0.5 to 1.0 disc- level 35, 148.2 lm; P ¼ 0.10), but there was a trend toward a diameters from the optic disc margin. After manual adjustment higher CRVE for patients with increasing levels of DR (level 10, by the grader, retinal vascular calibers are summarized into the 213.7 lm; level 20, 214.4 lm; level 35, 216.1 lm; P ¼ 0.056). central retinal arteriolar and venular equivalent (CRAE and During 24 months of topical treatment, there was a median CRVE) according to the ‘‘Big 6-formula’’ that includes the six increase of CRAE and CRVE of 1.0 lm (P ¼ 0.01) and 2.6 lm (P largest arterioles and venules.20 ¼ 0.0002) in the overall population. Changes in retinal vascular At baseline, images were graded by a single, trained grader calibers depended on baseline level of DR and treatment group (UFO), and at follow-up images were graded at a certified (Table 3). Overall, patients with higher levels of DR at baseline retinal image grading center (Yamagata University, Yamagata, developed a higher CRAE (level 10, 1.0 lm; level 20, 1.0 lm; Japan). level 35, þ6.1 lm; P ¼ 0.001; Fig. 1) and CRVE (level 10, þ1.3 lm; level 20, þ1.4 lm; level 35, þ10.3 lm; P ¼ 0.001; Fig. 2) at follow-up. Even though there was no general difference Clinical Examinations between treatment groups in the alteration of CRAE (placebo, Patients were treated in each eye with 1 drop BID (placebo or þ1.8 lm; brimonidine, þ1.9 lm; somatostatin, 0.2 lm; P ¼ somatostatin 0.1% or brimonidine tartrate 0.2%) and followed 0.98) or CRVE (placebo, þ3.7 lm; brimonidine, þ1.3 lm; for 96 weeks with comprehensive ophthalmic examinations somatostatin, þ2.4 lm; P ¼ 0.78), it was demonstrated that the every 24 weeks as previously described.12 development of wider retinal vascular calibers in higher levels Of 449 patients invited at baseline, 297 (66.1%) completed of DR depended on the treatment regimen; patients treated the study with gradable retinal images at baseline and follow- with placebo did not increase in CRAE (P ¼ 0.17) or CRVE (P ¼ Downloaded from iovs.arvojournals.org on 02/16/2021 Topical Neuroprotective Treatment in Diabetes Patients IOVS j May 2019 j Vol. 60 j No. 6 j 2259 TABLE 1. Baseline Characteristics in Patients With and Without Gradable Images* Completed Study Did Not Include With Gradable or With Patient Characteristics Retinal Images Ungradable Images P Value Patients, n 297 153 Age, y 64.5 (45.4–76.0) 63.7 (45.8–75.7) 0.46 Duration of diabetes, y (range) 9.9 (4.9–40.4) 10.7 (4.6–31.7) 0.18 Male sex, % 65.7 67.1 0.76 HbA1c, % (range) 7.0 (5.0-10.0) 7.0 (5.4–10.8) 0.86 Systolic blood pressure, mm Hg (range) 135 (95–183) 135 (98–186) 0.34 Diastolic blood pressure, mm Hg (range) 78 (55–103) 77 (50–115) 0.87 Body mass index, kg/m2 (range) 30.1 (20.1–53.9) 29.7 (19.7–53.7) 0.67 Level of DR, % 0.91 10 42.8 41.3 20 28.3 25.9 35 29.0 32.9 Treatment group, % 0.009† Placebo 36.4 28.0 Brimonidine 28.6 43.4 Somatostatin 35.0 28.7 Central retinal arteriolar equivalent, lm (range) 146.6 (110.4–184.6) 146.9 (111.6–196.0) 0.66 Central retinal venular equivalent, lm (range) 214.8 (160.0–283.9) 217.1 (162.3–290.8) 0.99 Data presented as median (with minimum and maximum) or percent. Level of DR defined according to Early Treatment Diabetic Retinopathy Study scale. * For patients that completed the study with gradable retinal images at baseline and follow-up as compared to those that did not complete the study or had ungradable retinal images at baseline or follow-up. † Statistically significant (P < 0.05). 0.56) with increasing baseline levels of DR. In contrast, there 10 at baseline), respectively. There was no difference in CRAE was a DR-dependent arteriolar widening in patients treated (P ¼ 0.51) or CRVE (P ¼ 0.75) when patients were compared with somatostatin (P ¼ 0.01), and a venular dilation in patients according to development in DR during the study. treated with brimonidine (P ¼ 0.048) and somatostatin (P ¼ 0.01). At follow-up, level of DR improved, remained unchanged DISCUSSION and worsened in 28.0% (n ¼ 83), 65.7% (n ¼ 195), and 6.4% (n In a 96-week, prospective, randomized trial evaluating the ¼ 19) patients, respectively. There were no differences in rates retinal effect of topical neuroprotective treatment, we con- of improvement, unchanged conditions, and DR worsening cluded that in patients with preexisting mild DR, brimonidine between patients in the different treatment groups (placebo, and somatostatin both induced retinal vascular dilation as 28.7% vs. 66.7% vs. 4.6%; brimonidine, 25.9% vs. 63.5% vs. opposed to placebo. 10.6%; somatostatin, 28.9% vs. 66.4% vs. 4.8%, P ¼ 0.47). Of Our results are in contrast with the earlier conception that patients with a potential to improve or worsen more than one topical treatment is inadequate to induce a vascular retinal level of DR from baseline, this was only found in 23.3% (20 of effect given the corneal barrier and the intraocular distance 86 with ETDRS 35 at baseline) and 1.2% (2 of 127 with ETDRS between the anterior and posterior part of the eye. However, TABLE 2. Baseline Characteristics of Study Patients* Patient Characteristics Placebo Brimonidine Somatostatin P Value Patients, n 108 85 104 Age, y (range) 63.9 (46.8–76.0) 64.4 (50.4–74.3) 64.6 (45.4–75.3) 0.73 Duration of diabetes, y (range) 9.8 (5.4–37.3) 9.9 (5.3–40.4) 10.0 (4.9–30.6) 0.78 Male sex, % 68.5 65.9 62.5 0.65 HbA1c, % (range) 7.0 (5.0–10.0) 7.1 (5.4–10.0) 7.0 (5.4–9.7) 0.59 Systolic blood pressure, mm Hg (range) 135 (104–180) 138 (95–164) 150 (108–183) 0.77 Diastolic blood pressure, mm Hg (range) 80 (55–103) 78 (58–103) 75 (55–101) 0.26 Body mass index, kg/m2 (range) 29.8 (20.4–48.5) 29.8 (20.1–53.9) 30.4 (21.6–47.7) 0.87 Level of DR, % 0.70 10 44.4 37.7 45.2 20 28.7 25.9 26.0 35 26.9 31.8 28.9 Central retinal arteriolar equivalent, lm (range) 146.7 (122.7–183.5) 147.1 (114.2–183.5) 145.2 (110.4–184.6) 0.68 Central retinal venular equivalent, lm (range) 215.6 (160.0–276.0) 212.7 (174.1–265.6) 215.7 (162.8–283.9) 0.89 Data presented as median (with minimum and maximum) or per cent. Level of DR defined according to Early Treatment Diabetic Retinopathy Study scale. * For the 297 patients that completed the study with gradable retinal images at baseline and follow-up according to treatment group. Downloaded from iovs.arvojournals.org on 02/16/2021 Topical Neuroprotective Treatment in Diabetes Patients IOVS j May 2019 j Vol. 60 j No. 6 j 2260 þ10.3 (P < 0.0001)† þ1.3 (P ¼ 0 .31) þ1.4 (P ¼ 0.37) (n ¼ 297) 0.001† All þ14.3 (P ¼ 0.0001)† 0.4 (P ¼ 0.82) þ1.5 (P ¼ 0.63) Somatostatin (n ¼ 104) 0.001† Central Retinal Venular Equivalent, lm þ13.9 (P ¼ 0.01)† þ1.5 (P ¼ 0.64) 2.0 (P ¼ 0.24) FIGURE 1. Development in CRAE during 96 weeks of topical treatment Brimonidine according to Early Treatment Diabetic Retinopathy Study levels of DR (n ¼ 85) 0.048† at baseline. There was a higher development with more advanced DR (level 10, 1.0 lm; level 20, 1.0 lm; level 35, þ6.1 lm; P ¼ 0.001). * In 297 patients with type 2 diabetes during 96 weeks of topical treatment according to treatment group and baseline level of DR. this concept has been challenged by animal studies8,9 as well as a 1-week trial by Tilma and Bek21 demonstrating retinal arteriolar narrowing in 22 patients with type 1 diabetes treated þ3.8 (P ¼ 0.046)† þ2.2 (P ¼ 0.23) þ4.0 (P ¼ 0.11) with topical latanoprost. In the present study, we demonstrate (n ¼ 108) Data presented as median, and level of DR defined according to Early Treatment Diabetic Retinopathy Study scale. a long-term effect of the retinal vasculature by two indepen- Placebo ‡ Test for trend of development in central retinal arteriolar and venular equivalent across different levels of DR. 0.56 dent neuroprotective drugs, even though the study period might not have been long enough to arrest or prevent development of DR. As indicated by the DCCT in type 1 diabetes, this would likely take at least 3 years, which was the reason that EUROCONDOR was designed to study surrogate markers of early vascular and neurogenic dysfunction. þ6.1 (P ¼ 0.0001)† 1.0 (P ¼ 0.85) 1.0 (P ¼ 0.71) Retinal vascular calibers have been identified as early (n ¼ 297) biomarkers of retinal structural dysfunction in diabetes.14,22,23 0.001† All In a 16-year prospective study of patients with type 1 diabetes, Broe et al.16 reported that a 10-lm retinal arteriolar narrowing and venular dilation independently predicted long-term prolif- erative DR, nephropathy, and peripheral neuropathy with odds ratios of 1.4 to 3.0. Likewise, the Wisconsin Epidemiological Study of Diabetic Retinopathy found that a 10 lm retinal þ7.2 (P ¼ 0.006)† 1.7 (P ¼ 0.50) 1.8 (P ¼ 0.89) venular (but not arteriolar) dilation in 4 years, associated with a Somatostatin (n ¼ 104) higher risk of various DR-endpoints in the following 6 years.24 Central Retinal Arteriolar 0.01† Equivalent, lm TABLE 3. Development in Retinal Vascular Caliber* þ6.2 (P ¼ 0.006)† þ0.6 (P ¼ 0.78) 2.5 (P ¼ 0.30) Brimonidine (n ¼ 85) 0.07 † Statistically significant (P < 0.05). 0.5 (P ¼ 0.97) þ2.9 (P ¼ 0.20) þ5.2 (P ¼ 0.14) (n ¼ 108) Placebo 0.17 FIGURE 2. Development in CRVE during 96 weeks of topical treatment according to Early Treatment Diabetic Retinopathy Study levels of DR of DR Level at baseline. There was a higher development with more advanced DR P‡ 10 20 35 (level 10, þ1.3 lm; level 20, þ1.4 lm; level 35, þ10.3 lm; P ¼ 0.001). Downloaded from iovs.arvojournals.org on 02/16/2021 Topical Neuroprotective Treatment in Diabetes Patients IOVS j May 2019 j Vol. 60 j No. 6 j 2261 Both of these studies identify retinal vascular dilation as a risk than expected dropout which may limit the generalizability of factor of DR-progression, which might seem in opposition with the results. Third, retinal vascular calibers were measured by our findings of a retinal vascular dilation in response to different graders at baseline and follow-up. neuroprotective treatment. However, comparisons are difficult In conclusion, the present study demonstrated retinal given that most patients in the studies by Broe et al.16 and vascular dilation induced by long-term topical neuroprotection Klein et al.24 had no or minimal DR, whereas our results were in 297 patients with type 2 diabetes and mild DR. The concept found in patients with preexisting mild DR at baseline. In that eye drops may induce retinal vascular changes would be addition, Broe et al.16 only evaluated patients with type 1 appealing in order to prevent or delay DR in the early phases, diabetes as opposed to type 2 diabetes in the EUROCONDOR. and upcoming prospective studies would be needed to In the EUROCONDOR, we have previously reported that translate this into clinical care. there was no association between baseline levels of DR and CRAE, but patients with moderate NPDR had a higher CRVE (ETDRS 10, 214.5 lm; ETDRS 20, 213.2 lm; ETDRS 35, 216.6 Acknowledgments lm; P ¼ 0.04) although this could not be considered clinically Supported by the European Union’s Seventh Framework Pro- significant.18 Noticeably, in the present study only patients gramme for research, technological development and demonstra- with preexisting DR developed arteriolar and venular widen- tion under Grant Agreement No. 278040. U. Frydkjaer-Olsen was ing, and we demonstrate that this was influenced by the supported by Fight for Sight Denmark, The Region of Southern topically administered agent. This is in alignment with the Denmark, and University of Southern Denmark. principal findings of the EUROCONDOR which report that Disclosure: J. Grauslund, None; U. Frydkjaer-Olsen, None; T. neuroprotective treatment was only effective in patients with Peto, None; J. Fernández-Carneado, BCN Peptides (E); B. preexisting retinal neurodysfunction.12 In particular, patients Ponsati, BCN Peptides (E); C. Hernández, None; J. Cunha-Vaz, with mild NPDR in both neuroprotective groups had a retinal None; R. Simó, None venular dilation of more than three times of that in the placebo group. This effect size is likely to be clinically relevant. To illustrate, the retinal venular dilation induced by topical References neuroprotection in our patients with preexisting mild DR 1. Grauslund J, Green A, Sjolie AK. Prevalence and 25 year (þ13.9–14.3 lm) has the same size as the difference in CRVE incidence of proliferative retinopathy among Danish type 1 between an average 50- and 85-year-old person, in which the diabetic patients. Diabetologia. 2009;52:1829–1835. former would be 13 lm wider.25 2. Liew G, Michaelides M, Bunce C. A comparison of the causes There is no clear mechanistic explanation of the topically of blindness certifications in England and Wales in working induced vascular dilation in patients with moderate NPDR, but age adults (16–64 years), 1999–2000 with 2009–2010. BMJ it could potentially be given by compensatory increased blood Open. 2014;4:e004015. flow to prevent peripheral nonperfusion which is strongly 3. Jonsson KB, Frydkjaer-Olsen U, Grauslund J. Vascular changes associated with DR progression26 and poor treatment out- and neurodegeneration in the early stages of diabetic come.27 Another potential explanation was proposed by retinopathy: which comes first? Ophthalmic Res. 2016;56:1– Ludovico et al.28 that demonstrated retinal vasodilation in 9. 50% of patients with early diabetes and suggested that a lack of 4. Simo R, Hernandez C; European Consortium for the Early vasodilation could be a harmful response. While hypoxia Treatment of Diabetic Retinopathy. Neurodegeneration in the induces retinal vasodilation in healthy eyes,29 lack of dilation in diabetic eye: new insights and therapeutic perspectives. diabetes can be caused by basement membrane thickening or Trends Endocrinol Metab. 2014;25:23–33. pericyte loss, which may lead to capillary dropout and closure.30 One of the main reasons of the lack of vasodilation 5. Diabetic Retinopathy Study Research Group. Preliminary in nontreated diabetic eyes can be attributed to the impairment report on effects of photocoagulation therapy. Am J Ophthalmol. 1976;81:383–396. of the neurovascular unit, which prevents vasodilation, thus impeding to adapt to the higher metabolic demands that exist 6. Early Treatment Diabetic Retinopathy Study Research Group. in the diabetic retina. Consequently, the neuroactive drugs Photocoagulation for diabetic macular edema. Early Treat- used in EUROCONDOR may have a favorable effect by ment Diabetic Retinopathy Study report number 1. Arch improving the neurovascular unit function. This effect results Ophthalmol. 1985;103:1796–1806. in vasodilation and, therefore, improved autoregulation and 7. Wells JA, Glassman AR, Ayala AR, et al. Aflibercept, delayed capillary dropout. However, with the data at hand, we bevacizumab, or ranibizumab for diabetic macular edema: cannot conclude if treatment-induced retinal vascular dilation two-year results from a comparative effectiveness randomized in mild DR is a specific consequence of topical neuroprotec- clinical trial. Ophthalmology. 2016;123:1351–1359. tion with brimonidine and somatostatin, or if this would also 8. Hernandez C, Bogdanov P, Corraliza L, et al. Topical be observed with other interventions. administration of GLP-1 receptor agonists prevents retinal It can be difficult to compare retinal vascular calibers neurodegeneration in experimental diabetes. Diabetes. 2016; between studies. For instance, Drobnjak et al.31 reported 65:172–187. median CRAE and CRVE of 163.3 and 251.0 lm, which were 9. Hernandez C, Bogdanov P, Sola-Adell C, et al. Topical substantial higher than the baseline finding of our study (146.6 administration of DPP-IV inhibitors prevents retinal neurode- and 214.8 lm).31 Unbalances might be explained by differenc- generation in experimental diabetes. Diabetologia. 2017;60: es in ages, populations, methodology, and by the fact that 2285–2298. diabetes was only present in a subset of patients in the study by 10. Hernandez C, Garcia-Ramirez M, Corraliza L, et al. Topical Drobnjak et al.31 administration of somatostatin prevents retinal neurodegen- Strengths of the present study include the prospective, eration in experimental diabetes. Diabetes. 2013;62:2569– randomized design and the long follow-up with repeated 2578. measurements over time. On the other hand, limitations should 11. Santos AR, Ribeiro L, Bandello F, et al. Functional and be acknowledged. First, we did not include retinal flow structural findings of neurodegeneration in early stages of measurements which could have provided additional informa- diabetic retinopathy: cross-sectional analyses of baseline data tion regarding retinal perfusion. Second, there was a higher of the EUROCONDOR project. Diabetes. 2017;66:2503–2510. Downloaded from iovs.arvojournals.org on 02/16/2021 Topical Neuroprotective Treatment in Diabetes Patients IOVS j May 2019 j Vol. 60 j No. 6 j 2262 12. Simo R, Hernandez C, Porta M, et al. Effects of topically 26. Silva PS, Dela Cruz AJ, Ledesma MG, et al. Diabetic administered neuroprotective drugs in early stages of diabetic retinopathy severity and peripheral lesions are associated retinopathy: results of the EUROCONDOR clinical trial. with nonperfusion on ultrawide field angiography. Ophthal- Diabetes. 2019;68:457–463. mology. 2015;122:2465–2472. 13. Diabetes Control and Complications Trial Research Group. 27. Torp TL, Kawasaki R, Wong TY, Peto T, Grauslund J. The effect of intensive treatment of diabetes on the Peripheral capillary non-perfusion in treatment-naive prolif- development and progression of long-term complications in erative diabetic retinopathy associates with postoperative insulin-dependent diabetes mellitus. New Eng J Med. 1993; disease activity 6 months after panretinal photocoagulation 329:977–986. [published online ahead August 1, 2018]. Br J Ophthalmol. 14. Grauslund J, Hodgson L, Kawasaki R, Green A, Sjolie AK, https://doi.org/10.1136/bjophthalmol-2018-312195. Wong TY. Retinal vessel calibre and micro- and macrovascular 28. Ludovico J, Bernardes R, Pires I, Figueira J, Lobo C, Cunha-Vaz complications in type 1 diabetes. Diabetologia. 2009;52: J. Alterations of retinal capillary blood flow in preclinical 2213–2217. retinopathy in subjects with type 2 diabetes. Graefes Arch 15. Klein R, Klein BE, Moss SE, Wong TY. Retinal vessel caliber Clin Exp Ophthalmol. 2003;241:181–186. and microvascular and macrovascular disease in type 2 29. Sousa DC, Leal I, Moreira S, Dionisio P, Abegao Pinto L, diabetes: XXI: the Wisconsin Epidemiologic Study of Diabetic Marques-Neves C. Hypoxia challenge test and retinal circula- Retinopathy. Ophthalmology. 2007;114:1884–1892. tion changes - a study using ocular coherence tomography 16. Broe R, Rasmussen ML, Frydkjaer-Olsen U, et al. Retinal vessel angiography. Acta Ophthalmol. 2018;96:e315–e319. calibers predict long-term microvascular complications in 30. Marques IP, Alves D, Santos T, et al. Multimodal imaging of the type 1 diabetes: the Danish Cohort of Pediatric Diabetes 1987 initial stages of diabetic retinopathy. different disease path- (DCPD1987). Diabetes. 2014;63:3906–3914. ways in different patients. Diabetes. 2019–653. 17. Early Treatment Diabetic Retinopathy Study Research Group. 31. Drobnjak D, Munch IC, Glumer C, et al. Retinal vessel Grading diabetic retinopathy from stereoscopic color fundus diameters and their relationship with cardiovascular risk and photographs—an extension of the modified Airlie House all-cause mortality in the Inter99 Eye Study: a 15-year follow- classification. ETDRS report number 10. Ophthalmology. up. J Ophthalmol. 2016;2016:6138659. 1991;98:786–806. 18. Frydkjaer-Olsen U, Soegaard Hansen R, Simo R, et al. Correlation between retinal vessel calibre and neurodegener- APPENDIX ation in patients with type 2 diabetes mellitus in the European Consortium for the Early Treatment of Diabetic Retinopathy (EUROCONDOR). Ophthalmic Res. 2016;56:10– Members of the EUROCONDOR 16. 19. Hubbard LD, Brothers RJ, King WN, et al. Methods for Scientific Institute San Raffaele, Milan, Italy: Francesco evaluation of retinal microvascular abnormalities associated Bandello. with hypertension/sclerosis in the Atherosclerosis Risk in Association for Innovation and Biomedical Research on Communities Study. Ophthalmology. 1999;106:2269–2280. Light and Image, Coimbra, Portugal: José Cunha-Vaz. 20. Knudtson MD, Lee KE, Hubbard LD, Wong TY, Klein R, Klein Moorfields Eye Hospital, London, United Kingdom: Cath- BE. Revised formulas for summarizing retinal vessel diame- erine Egan. ters. Curr Eye Res. 2003;27:143–149. Vall d’Hebron University Hospital, Barcelona, Spain: José 21. Tilma KK, Bek T. Topical treatment for 1 week with Garcı́a-Arumı́. latanoprost but not diclofenac reduces the diameter of dilated University of Aston, Birmingham, United Kingdom: Jona- retinal arterioles in patients with type 1 diabetes mellitus and than Gibson. mild retinopathy. Acta Ophthalmol. 2012;90:750–755. University of Liverpool, Liverpool, United Kingdom: Simon 22. Klein R, Klein BE, Moss SE, et al. Retinal vascular abnormal- Harding. ities in persons with type 1 diabetes: the Wisconsin International Diabetes Federation Europe, Brussels, Bel- Epidemiologic Study of Diabetic Retinopathy: XVIII. Ophthal- gium: Sehnaz Karadeniz. mology. 2003;110:2118–2125. University of Ulm, Ulm, Germany: Gabriele Lang. 23. Cheung N, Rogers SL, Donaghue KC, Jenkins AJ, Tikellis G, Hôpital Lariboisière-APHP, Paris, France: Pascale Massin. Wong TY. Retinal arteriolar dilation predicts retinopathy in University of Padova, Padova, Italy: Edoardo Midena. adolescents with type 1 diabetes. Diabetes Care. 2008;31: BCN Peptides, Barcelona, Spain: Berta Ponsati. 1842–1846. University of Turin, Turn, Italy: Massimo Porta. 24. Klein R, Myers CE, Lee KE, Gangnon R, Klein BE. Changes in Cheltenham General Hospital, Cheltenham, United King- retinal vessel diameter and incidence and progression of dom: Peter Scanlon, Stephen Aldington. diabetic retinopathy. Arch Ophthalmol. 2012;130:749–755. Vall d’Hebron Research Institute, Barcelona, Spain: Rafael 25. Myers CE, Klein R, Knudtson MD, et al. Determinants of Simó, Cristina Hernández. retinal venular diameter: the Beaver Dam Eye Study. University of Southern Denmark, Odense, Denmark: Jakob Ophthalmology. 2012;119:2563–2571. Grauslund, Ulrik Frydkjaer-Olsen. Downloaded from iovs.arvojournals.org on 02/16/2021