Journal of Clinical Medicine Article Assessment of Inner Retinal Layers and Choroidal Thickness in Type 1 Diabetes Mellitus: A Cross-Sectional Study Marc Carbonell 1,2,† , Núria Alonso 3,4,5,† , Esmeralda Castelblanco 4,6,7,† , Jordi Real 4,7 , Anna Ramírez-Morros 3 , Rafael Simó 4,8 , Cristina Hernández 4,8 , Carme Jurjo 9,10 , Alícia Traveset 9,10 , Xavier Valldeperas 1,2, * and Dídac Mauricio 4,5,6,7,10, * 1 Department of Ophthalmology, University Hospital Germans Trias i Pujol, 08916 Badalona, Spain 2 Department of Surgery, Barcelona Autonomous University (UAB), 08035 Barcelona, Spain 3 Department of Endocrinology and Nutrition, University Hospital and Health Science Research Institute Germans Trias i Pujol, 08916 Badalona, Spain 4 Centre for Biomedical Research on Diabetes and Associated Metabolic Diseases (CIBERDEM), 08907 Barcelona, Spain 5 Department of Medicine, Barcelona Autonomous University (UAB), 08035 Barcelona, Spain 6 Department of Endocrinology & Nutrition, Hospital de la Santa Creu i Sant Pau & Institut d’Investigació Biomèdica Sant Pau (IIB Sant Pau), 08041 Barcelona, Spain 7 DAP-Cat Group, Unitat de Suport a la Recerca Barcelona, Fundació Institut Universitari per a la Recerca a l’Atenció Primària de Salut Jordi Gol i Gurina (IDIAPJGol), 08025 Barcelona, Spain 8 Department of Endocrinology and Nutrition, Hospital Universitari Vall d’Hebron, Diabetes and Metabolism Research Unit, Vall d’Hebron Institut de Recerca (VHIR), Universitat Autònoma de Barcelona, 08035 Barcelona, Catalonia, Spain 9 Department of Ophthalmology, University Hospital Arnau de Vilanova, 25198 Lleida, Spain 10 Biomedical Research Institute of Lleida (IRBLleida) & University of Lleida, 25198 Lleida, Spain * Correspondence: xvalldeperas@gmail.com (X.V.); didacmauricio@gmail.com (D.M.); Tel.: +34-93-497-8859 (X.V.); +34-93-556-5661 (D.M.) † Authors contributed equally.  Received: 30 July 2019; Accepted: 4 September 2019; Published: 8 September 2019  Abstract: Recent studies have shown that retinal neurodegeneration may precede visible vascular changes in diabetic retinopathy (DR). In addition, the relationship of choroidal thickness (CT) with DR stage is not well defined. To assess the inner retinal and choroidal structural changes in type 1 diabetic subjects (T1D), a cross-sectional study was conducted in 242 T1D patients and in 69 age-matched, non-diabetic individuals. The nasal retinal nerve fibre layer (RNFL) thickness was lower in T1D patients without DR (p < 0.001), with mild DR (p < 0.001), and with advanced DR (p < 0.001) compared to control subjects. The ganglion cell layer (GCL) thickness was lower in T1D patients with advanced DR compared to those with mild DR (p = 0.003) and without DR (p < 0.001) and compared to the control subjects (p < 0.001). T1D subjects with no DR and mild DR had higher CT than the control subjects, but the CT in T1D patients with advanced DR was lower (p = 0.038) than that in T1D subjects with mild DR and was not significantly different from that of the control subjects. In conclusion, T1D subjects showed a significant thinning of the nasal RNFL in the early stages of the disease, even before any vascular changes in the retina. A decrease in the GCL thickness during advanced DR stages was observed. Choroidal thickness was higher in T1D subjects without DR and in early DR stages but decreased in advanced stages. Keywords: type 1 diabetes; diabetic retinopathy; retinal nerve fibre layer; ganglion cell layer; choroid J. Clin. Med. 2019, 8, 1412; doi:10.3390/jcm8091412 www.mdpi.com/journal/jcm J. Clin. Med. 2019, 8, 1412 2 of 13 1. Introduction Diabetic retinopathy (DR) is one of the leading causes of visual impairment and preventable blindness among the adult working-age population in developed countries. DR is a multifactorial disease with many different risk factors, including glycated haemoglobin (HbA1c), blood pressure, and serum lipid levels [1,2]. Classically, DR has primarily been considered a microvascular disorder in which the initial manifestations include retinal microaneurysms, capillary alterations, and haemorrhaging. These vascular alterations may lead to degeneration of the neuronal and glial structures of the inner retina as the disease progresses. However, data from recent studies using spectral-domain optical coherence tomography (SD-OCT) imaging suggest that morphological changes in retinal neurodegeneration may precede any visible vascular changes [3–7]. In this regard, the earliest neurodegenerative alterations occur in the main layers of the inner retina, as evidenced by the thinning of the retinal nerve fibre layer (RNFL) and the ganglion cell complex (GCC) [4,6–8]. Choroidal alterations also play a key role in the pathophysiology of several retinal diseases, including DR [9–12]. Choroidal thickness (CT) is considered a measure of choroidal blood flow, which supplies the outer retinal layers. Several studies have described the presence of choroidal vasculopathy in diabetic subjects [9,13–16]. However, those studies—most of which were performed in a paediatric population—have reported contradictory findings, likely due to their small sample sizes. A similar controversy exists for the presence of retinal neurodegeneration in subjects who have diabetes without DR. The published data are contradictory, with some authors finding no differences in RNFL or ganglion cell layer (GCL) thickness in patients with type 1 diabetes (T1D) without DR compared to healthy control subjects [17,18], whereas other authors have described differences in these structures between diabetic patients and control subjects [19]. In this context, the aims of the present study were (1) to evaluate the retinal and choroidal structures in a large cohort of patients with T1D using SD-OCT and (2) to determine whether retinal and choroidal alterations are correlated with the degree of diabetic retinopathy. 2. Experimental Section 2.1. Patients This was a cross-sectional study conducted in subjects diagnosed with T1D recruited from outpatient clinics at two university hospitals within the same health care organization (Institut Català de la Salut) in Catalonia, Spain. The inclusion criteria for T1D subjects were as follows: (1) age ≥ 18 years, (2) diagnosis of T1D with disease duration ≥ one year, (3) normal renal function (estimated glomerular filtration rate > 60 mL/min), (4) urine albumin excretion rate < 300 mg/g, and (5) the absence of cardiovascular disease (ischaemic heart disease, cerebrovascular disease, peripheral arterial disease, and heart failure) or diabetic foot disease. The control group consisted of 69 healthy age- and sex-matched volunteers. The selection criteria for the control subjects were the same as those for the T1D subjects, except for the specific microvascular complications (retinopathy and albuminuria). Additional inclusion criteria for the control subjects were: (1) fasting glucose < 100 mg/dL and (2) HbA1c < 5.7%. Ophthalmologic exclusion criteria for both groups included a refractive error of 5 dioptres or more, concomitant retinal diseases, ocular inflammation, glaucoma, macular oedema, previous surgery (during the year before the exam), and laser photocoagulation (during the six months before the exam). The present subject sample was a subset of 340 T1D subjects included in a previously published study aimed at assessing the association of subclinical carotid atherosclerotic disease with the presence and severity of diabetic retinopathy [20]. For the current study, it was also important to exclude patients with previous cardiovascular disease and/or chronic kidney disease, as these two conditions are known to be associated with vascular damage. Of those 340 subjects, 98 were excluded from the present study either because they did not undergo SD-OCT (n = 78) or due to the presence of J. Clin. Med. 2019, 8, 1412 3 of 13 ophthalmologic exclusion criteria (n = 20) as follows: glaucoma (n = 3); laser treatments in the previous 6 months (n = 2); myopia > 5 dioptres (n = 6); and proliferative diabetic retinopathy treated with photocoagulation (n = 9). As a result, a total of 242 subjects were included in the current study. 2.2. Clinical Assessment Serum and urine samples were collected in the fasting state. Subject characteristics (gender, age, and ethnicity), anthropometric parameters (weight, height, body mass index, and waist circumference), other parameters (hypertension, dyslipidaemia, smoking, medication, and disease duration) and laboratory measurements (glucose, HbA1c, triglycerides, total/HDL/LDL cholesterol, creatinine, and albumin/creatinine ratio), assessed using standard laboratory methods previously described, were recorded [21]. 2.3. Ophthalmic Examination A complete ophthalmological evaluation, including retinopathy classification, was performed by two expert ophthalmologists (MC and CJ) in both the patient and control groups. Optical coherence tomography measurements were performed using spectral-domain OCT (Cirrus HD-OCT, model 4000, Carl Zeiss Meditec, Dublin, CA, USA) to measure RNFL, GCL, and CT. Only SD-OCT scans with a signal strength > 8 were used for the analysis. RNFL thickness was measured according to an Optic Disc Cube 200 × 200 protocol, with the data calculated automatically by quadrant (temporal, superior, nasal, inferior). Mean values were also calculated. The GCL thickness was measured (including mean and minimum thicknesses) using the Macula Cube 512 × 128 protocol around the fovea. Choroidal thickness was measured using enhanced-depth imaging (EDI-OCT) with a high-definition 5-line Raster scan pattern. The measurements were made manually with a caliper in three positions (subfoveal, nasal, and temporal to the fovea) at a distance of 2000 µm. Choroidal thickness was calculated as the vertical distance between the first hyperreflective line (Bruch’s membrane) and the second hyperreflective line (the internal surface of the sclera) [22]. DR was classified into five stages according to the Early Treatment Diabetic Retinopathy Study (ETDRS) classification [23], as follows: (1) no apparent retinopathy, (2) mild non-proliferative retinopathy (NPDR), (3) moderate NPDR, (4) severe NPDR, and (5) proliferative diabetic retinopathy (PRD). Diabetic subjects with DR were reclassified into 2 stages (mild [ETDRS stages 1–2] or advanced [ETDRS stages 3–4] DR). Additionally, subjects with SD-OCT choroid measurements of 2 or more standard deviations above the mean value observed in the control group were classified as having a thickening of the choroid. Further, subjects with SD-OCT RNFL measurements less than 2 standard deviations below the mean value observed in the control group were classified as having a thinning of the RNFL. For the analysis, we used the ophthalmological variables obtained for the right eye if both eyes had the same DR grade. Otherwise, the variables from the eye with the highest DR grade were selected. The ethics committees of both participating centres approved the study, which was performed in accordance with the principles of the Declaration of Helsinki. Written informed consent was obtained from all participants. 2.4. Statistical Analysis A descriptive between-group comparison was performed to assess homogeneity in the clinical characteristics of the study population. Student’s t-test was performed to determine the significant differences between two groups (T1D versus control) in the quantitative variables, and a Chi-square test was used to compare qualitative variables between groups. Univariate and multivariate regression models were performed to estimate the crude and adjusted associations. The multivariate results were adjusted for sex, age, hypertension, dyslipidaemia, and body mass index (BMI). The subanalysis in the T1D subjects was performed by examining the association between the DR grade and the ophthalmological variables. Differences between groups were evaluated by analysis of J. Clin. Med. 2019, 8, 1412 4 of 13 variance (ANOVA), and p values were corrected using the post hoc Tukey test. Linear multivariate regression models were performed to assess the associations for each ophthalmological variable. Correction for multiple comparisons was performed using the Benjamini–Hochberg procedure (or false discovery rate). The cut-off for statistical significance was set as p < 0.05. The statistical R package (v. 3.5.1, Free Software Foundation, https://www.r-project.org/) was used for data management and to perform the statistical analysis. 3. Results The clinical characteristics of the study group are summarized in Table 1. Hypertension and dyslipidaemia were more prevalent in the T1D group, who also had higher blood pressure (systolic and diastolic) than the control subjects. There were no significant differences between groups in terms of age, sex, ethnicity, smoking habits, BMI, waist circumference, creatinine levels, or urine/albumin/creatinine ratios. The proportion of microalbuminuria was 3.3% (8/242) in T1D and 2.9% (2/69) in the control group. The mean duration of T1D was 20.6 years. Table 1. Clinical characteristics of the study population. Variables 1 Control Type 1 Diabetes p Value N 69 242 - Age, years 45.1 (11.2) 44.5 (10.7) 0.668 Sex, male 30 (43.5%) 114 (47.1%) 0.692 Ethnicity, Caucasian 69 (100%) 239 (98.8%) 1.000 Current or former smoker 35 (50.7%) 128 (53.2%) 0.387 Antiplatelet agents 0 (0.00%) 67 (27.7%) <0.001 Dyslipidaemia 5 (7.25%) 95 (39.3%) <0.001 Hypertension 7 (10.1%) 58 (24.0%) 0.020 Systolic blood pressure, mmHg 118 (13.2) 126 (17.1) <0.001 Diastolic blood pressure mmHg 71.8 (9.38) 74.4 (10.0) 0.049 Body mass index, kg/m2 24.6 (3.62) 25.6 (4.05) 0.059 Waist circumference, cm 87.2 (11.0) 88.9 (12.6) 0.290 HbA1c, % 5.28 (0.31) 7.60 (1.03) <0.001 Total cholesterol, mg/dL 203 (30.8) 180 (28.9) <0.001 HDL cholesterol, mg/dL 59.9 (12.4) 64.3 (15.4) 0.017 LDL cholesterol, mg/dL 123 (29.3) 102 (24.2) <0.001 Triglycerides, mg/dL 103 (69.4) 74.4 (34.7) 0.002 Creatinine, mg/dL 0.79 (0.13) 0.77 (0.16) 0.151 Urine albumin/creatinine ratio, mg/g 7.65 (10.4) 7.02 (13.6) 0.684 Diabetes duration, years - 20.6 (10.4) - 1All data are given as the mean (standard deviation) or n (%). Student’s t-test was used for quantitative variables and Chi-square test for qualitative variables. HDL, high density lipoprotein; LDL, low density lipoprotein. 3.1. Ophthalmologic Examination Of the 242 T1D subjects, 139 (57.4%) did not have DR, 74 (30.6%) had mild DR, and 29 (12%) presented with advanced DR. The clinical variables of the study group according to DR grade are shown in Table S1. Nasal RNFL was significantly lower in all three T1D subgroups (i.e., those without DR, mild DR, and advanced DR) (p < 0.001) than in the control group, contrary to what was observed in the superior and inferior sectors, in which the RNFL was greater in the T1D subjects than in the control subjects p = 0.004 and p < 0.001, respectively. No significant differences between the T1D subjects and control subjects were observed in the temporal RNFL (p = 0.283) (Figure S1, Table S2). The proportion of subjects with retinal thinning assessed in the RNFL in T1D subjects without DR was 1.5% (2/130) in the superior sector, 0.8% (1/131) in the nasal sector, and 1.5% (2/131) in the inferior sector. The presence of retinal thinning in subjects with advanced DR was 11.1% (1/9) in the superior sector, 0% (0/9) in the nasal sector, and 22.2% (2/9) in the inferior sector. Choroidal thickness was also significantly greater in the subfoveal (p < 0.001), nasal (p < 0.001), and temporal (p = 0.004) sectors in T1D subjects versus those in the control subjects (Figure 2). Up to 8.3% (11/133) of the T1D subjects without DR presented a thickening of the choroid. This proportion of patients was observed in the three sectors analysed (temporal, nasal, and subfoveal). In relation to the group of T1D subjects with mild DR, thickening was observed in 7.1% (5/70) of subjects in the J. Clin. Med. 2019, temporal 8, 1412 sector, 17.1% (12/70) in the nasal sector and 10% (7/70) in the subfoveal sector. Choroidal 5 of 13 thickness was greater in the T1D subjects with mild DR and without DR compared to control subjects (Figure 2), both for the subfoveal region (p = 0.006 and p < 0.001, respectively) and the nasal region (p GCL thickness was significantly lower in T1D subjects with advanced DR versus T1D subjects = 0.001 and p < 0.001, respectively). Temporal CT was greater only in patients with mild DR (p = 0.005) without versus DR (p