Type 1 diabetes mellitus (T1DM) is a chronic disease requiring long-term control and monitoring of chronic complications such as nephropathy, vasculopathy, and retinopathy. Diabetic retinopathy (DR) may lead to blindness resulting from not only vitreous hemorrhage and tractional retinal detachment, but also from macular disease that decreases central visual acuity.3–6 Current methods of DR evaluation involve indirect ophthalmoscopy, color fundus photography, fundus fluorescence angiography, and optical coherence tomography (OCT). OCT is a noninvasive technique with high levels of accuracy and definition in the detection of early and small changes in retinal structure. This technique may therefore be useful when visible changes are absent using a fundus examination.
Recent studies have reported that different mechanisms are responsible for DR, including vascular and neurodegenerative processes. Studies have reported that retinal neurodegeneration may occur in DR before any microcirculatory abnormalities can be detected;7–9 however, all previous studies were performed in adult diabetic patients, so it is unknown whether the neurodegenerative process starts during childhood.
The aims of this study were to determine whether undetectable morphological changes are present in T1DM onset in children without DR, and to determine the relationship between the ganglion cell-inner plexiform layer (GC-IPL) and retinal nerve fiber layer (RNFL) thicknesses and demographic/clinical parameters in children with T1DM without DR.
Patients and Methods
This was a cross-sectional, prospective study. The design of the study adhered to the Declaration of Helsinki and was approved by the institutional ethics committee of Behcet Uz Training and Research Hospital (Izmir, Turkey). Written informed consent was obtained from all participants and their parents before enrollment. In total, 60 eyes of 30 children with T1DM and 30 age- and sex-matched, normal controls were included in the study. The patients were recruited from the outpatient clinic of the Department of Pediatric Endocrinology of Behcet Uz Training and Research Hospital between November 2015 and June 2016. Normal controls who had no ocular or systemic disease were selected from patients at the same hospital. Age, sex, and the duration of DM were recorded. Serum glycosylated hemoglobin (HbA1c) was measured at the time of ophthalmological evaluation. A detailed ophthalmic evaluation, including autorefraction-keratometry; best-corrected visual acuity using a Snellen chart (converted to the logarithm of the minimum angle of resolution); a slit-lamp anterior segment examination; determination of the intraocular pressure (IOP) using Goldmann applanation tonometry; and a fundus examination, including indirect ophthalmoscopy, color fundus photography, and OCT, were performed. The spherical equivalent (SE) was determined using the formula SE = C/2+S, in which C was the cylindrical power and S was the spherical power. The fundus evaluation of all eyes was examined by an expert ophthalmologist (OK) using slit-lamp biomicroscopy with a 90 diopter (D) lens and indirect ophthalmoscopy after dilation of the pupils. Subsequently, using a 45-degree digital camera VISUCAM (Zeiss, Germany), two fundus photographs were taken of each eye: one centered on the fovea and the other on the optic disc. Each eye was classified for DR and maculopathy according to the American Academy of Ophthalmology guidelines.10 The minimum criterion for diagnosis of DR was the presence of at least one microaneurysm. In case there was any doubt as to the DR, fundus fluorescein angiography was performed. Patients who had findings of DR were excluded from the study.
OCT measurements were performed using spectral-domain OCT (SD-OCT) (Cirrus HD-OCT, model 4000, Version 6.5; Carl Zeiss Meditec, Dublin, CA) to measure the GC-IPL and RNFL thicknesses. All SD-OCT scans were performed by the same experienced specialist. Only SD-OCT scans with a signal strength greater than seven were used for analysis; otherwise, they were repeated. A total of 256 specific A-scans aligned in a circle of 3.46 mm diameter centered on the optic disc were extracted to provide average RNFL thickness and data in clock hours and quadrants. This circle was both automatically and manually placed by two independent observers (OK, TK). Each observer was masked to the other observer's measurements. The measurements were performed in a random order and masked fashion. Interobserver reproducibility of the RNFL thickness measurements was assessed by measuring the intraclass correlation coefficient (ICC). The mean of two observer's measurements was used to analyze the agreement between observer and automatic measurements. The mean RNFL thicknesses in four quadrants (superior, nasal, inferior, and temporal) were automatically calculated. The GC-IPL thickness was measured using a Macula Cube 512 × 128 protocol around the fovea, and analyzed in six quadrants (superior, superior-temporal, superior-nasal, inferior, inferior-nasal, and inferior-temporal).
Subjects with amblyopia, strabismus, corneal opacity or dystrophy, cataract, nystagmus, glaucoma, congenital or acquired retinal and optic nerve disorder, or a history of ocular trauma or surgery were excluded. Patients who had any other known systemic disease, including hypertension, anemia, renal disease, and cardiovascular disease, were also excluded. Additionally, subjects with a refractive spherical diopter greater than 5 D or with a high cylinder diopter greater than 3 D were excluded from the study to reduce the effects of refractive error on the SD-OCT measurements.
All statistical data were analyzed using SPSS statistical software for Windows, version 17.0 (SPSS, Chicago, IL). The chi-squared test was used to compare categorical data. The independent t-test was used to compare groups. A P value of less than .05 was considered statistically significant. Correlations between the variables were determined using Pearson's correlation coefficient.
All participants had normal fundoscopy results. Table 1 shows the demographic and clinical characteristics of the groups. The study group included 60 eyes of 30 patients with diabetes (14 males). The mean duration of T1DM and percentage of HbA1c were 5.76 years ± 3.13 years and 9.2% ± 2.2%, respectively. The mean age was 14.2 years ± 2.4 years (range: 10 years to 17 years), the mean SE was 0.02 D ± 0.54 D, and the mean IOP was 14.6 mm Hg ± 3.2 mm Hg. The control group included 60 eyes of 30 subjects (15 males) without diabetes. The mean age in the control group was 13.4 years ± 2.2 years (range: 10 years to 17 years), the mean SE was 0.06 D ± 0.51 D, and the mean IOP was 15.0 mm Hg ± 3.0 mm Hg. There was no significant difference between the groups regarding age, sex, IOP, or SE (P = .07, P = .113, P = .245, and P = .369, respectively). Comparisons of the RNFL and GC-IPL thicknesses in the two groups are shown in Table 2. The diabetic group had a lower mean GC-IPL thickness in all quadrants except the superior-nasal quadrant when compared with control subjects, and the difference between the groups was statistically significant (P < .05). However, the mean RNFL thickness in all quadrants was similar between the groups (P > .05).
Characteristics of Study and Control Groups
Comparison of RNFL and GC-IPL Thickness in All Quadrants
There was a significant inverse correlation between HbA1c and the GC-IPL thickness in the superior-temporal and superior quadrants of the diabetic group (r [Pearson’s correlation coefficient] = −0.272, P = .03; r = −0.269, P = .03, respectively). There was no correlation between the duration of diabetes and RNFL or GCL-IPL thicknesses. The correlations of HbA1c, DM duration, and age with the RNFL and GCL-IPL thicknesses are shown in Table 3.
Correlation of GC-IPL and RNFL Thickness With Age, Duration of Diabetes, and HbA1c
Various biochemical and metabolic abnormalities in the retina, including increased levels of glutamate,11 accumulation of advanced glycation end products,12 and increased oxidative stress,13 have been reported to induce the expression of proapoptotic molecules and cause apoptosis in neuronal cells in patients with diabetes. Previous studies reported that oxidative stress induced by high glucose resulted in the activation of caspase-3 and apoptosis in retinal ganglion cells.14–16 Increased apoptosis of neural retinal cells in rats with experimentally induced diabetes was recently reported; retinal ganglion cells were especially susceptible to change, with a 10% decrease in cell number in diabetic eyes when compared with control eyes.15 Notably, experimental studies reported that the sensory retina from patients with diabetes with or without DR showed immunoreactivity to apoptosis-promoting factors in ganglion cells.16 This study showed that ganglion cells in patients with diabetes expressed proapoptotic factors, resulting in apoptosis that occurred independently of retinopathy. Furthermore, there is increasing evidence that retinal ganglion cell death occurs early in patients with diabetes and that neurodegeneration is an important component of DR.15–16
There have been only two similar studies describing early retinal morphological changes in adult patients with diabetes using SD-OCT. These studies that evaluated GC-IPL and RNFL thicknesses in adult patients with T1DM without DR reported different findings. Van Dijk et al.7 reported no significant changes in the GC-IPL thickness in patients with T1DM without DR when compared with normal control subjects. However, Chen et al.8 reported that the GC-IPL thicknesses in all quadrants in patients with T1DM without DR were significantly thinner than those of control subjects. It should be emphasized that these studies were performed in adult patients with T1DM. The mean ages of the diabetic groups were 30 years ± 11 years7 and 21.8 years ± 5.3 years.8 In contrast to these studies, the mean age of the DM group in the our study was 14.2 years ± 2.4 years. To the best of our knowledge, this is the first study to report retinal morphological changes in early childhood T1DM using SD-OCT, and it is the first study to report the effects of early T1DM on retinal neuronal cells. In contrast to previous studies, our study shows that retinal ganglion cell loss in DM begins in early childhood. Chen et al.8 reported that the RNFL was thinner only at the 9 o'clock position in adult patients with T1DM when compared with the RNFL of the control group. However, RNFL thinning was not seen in the childhood T1DM group in our study, suggesting that RNFL degeneration occurs in the adult onset stage of T1DM. Taken together, these data suggest that only ganglion cells are affected in the early stages of T1DM. Subsequently, secondary thinning of the RNFL in the peripapillar area can result from axonal loss from central ganglion cells during the adult stage of T1DM.
Notably, we detected a negative correlation between HbA1c and GC-IPL thickness, especially in the superior quadrant. However, we did not detect any significant correlation between the duration of DM and GC-IPL thickness. The long-term dysregulation of blood glucose, rather than the duration of DM, may therefore be responsible for retinal neurodegeneration during early childhood. Effective blood glucose control with intensive insulin treatment for T1DM during early childhood may therefore prevent or slow neurodegeneration.
In our study, the interobserver ICCs for the RNFL thickness measurements were 0.921 (95% CI, 0.918–0.944). The ICC was greater than 0.90 for all measurement points. Good agreement was found between the two observers for RNFL thickness measurements (ICC > 0.90). No statistically significant differences were found when comparing the mean measurements of the two observers with the automated determinations of the RNFL. Good correlations (ICC = 0.89 to 0.94) were also found between the two methods. This finding demonstrated that the automatic optic nerve head measurements were in agreement with the results obtained by manual detection of the disc margin in pediatric patients
There are some limitations to our study. First, we investigated a small number of childhood patients with diabetes and control subjects. The small number of participants limited the statistical power to detect small differences between groups. Second, all participants were white, and recent studies have reported that differences in foveal structure might be associated with ethnic differences.17 Third, because we did not expect high ganglion cell loss in our study, visual field test was not performed. Although the visual field test is a useful test for screening for glaucomatous visual field defects, it is shown that in its current form, it does not reliably detect early visual field defects. That is to say, it has been found that before the visual field defect is clinically symptomatic, 30% or more of retinal ganglion cells must be already lost.18,19
In conclusion, SD-OCT can detect earlier retinal structural changes in children with T1DM without retinopathy when compared with ophthalmoscopic examinations. Furthermore, SD-OCT measurements showed ganglion cell loss in children with T1DM when compared with control subjects. We suggest that the decreased GC-IPL thickness results from neurodegenerative effects on the retina that probably occur prior to vasculopathy in the early stages of childhood diabetes. Additional in vitro or in vivo studies including a larger number of participants of different ethnicities are needed to verify our results.
- Moss SE, Klein R, Klein BE. The 14-year incidence of visual loss in a diabetic population. Ophthalmology. 1998;105(6):998–1003. doi:10.1016/S0161-6420(98)96025-0 [CrossRef]
- Klein R, Klein BE, Moss SE, Cruickshanks KJ. The Wisconsin Epidemiologic Study of diabetic retinopathy. XIV. Ten-year incidence and progression of diabetic retinopathy. Arch Ophthalmol. 1994;112(9):1217–1228. doi:10.1001/archopht.1994.01090210105023 [CrossRef]
- Laakso M. Epidemiology of Type 2 Diabetes. In: Goldstein BJ, Mueller-Wieland D, eds. Type 2 Diabetes: Principles and Practice. 2nd ed. New York, NY: Country Informa Healthcare; 2007:1–12.
- American Diabetes Association. Standards of Medical Care in Diabetes – 2011. Diabetes Care. 2011;34(5):11–61. doi:10.2337/dc11-S011 [CrossRef]
- Nathan DM, Zinman B, Diabetes Control and Complications Trial/Epidemiology of Diabetes Interventions and Complications (DCCT/EDIC) Research Group et al. Modern-day clinical course of type 1 diabetes mellitus after 30 years' duration: The diabetes control and complications trial/epidemiology of diabetes interventions and complications and Pittsburgh epidemiology of diabetes complications experience (1983–2005). Arch Intern Med. 2009;169(14):1307–1316. doi:10.1001/archinternmed.2009.193 [CrossRef]
- Klein R, Knudtson MD, Lee KE, Gangnon R, Klein BE. The Wisconsin Epidemiologic Study of Diabetic Retinopathy XXIII: The twenty-five-year incidence of macular edema in persons with type 1 diabetes. Ophthalmology. 2009;116(3):497–503. doi:10.1016/j.ophtha.2008.10.016 [CrossRef]
- van Dijk HW, Verbraak FD, Kok PH, et al. Decreased retinal ganglion cell layer thickness in patients with type 1 diabetes. Invest Ophthalmol Vis Sci. 2010;51(7):3660–3665. doi:10.1167/iovs.09-5041 [CrossRef]
- Chen Y, Li J, Yan Y, Shen X. Diabetic macular morphology changes may occur in the early stage of diabetes. BMC Ophthalmol. 2016;16:12. doi:10.1186/s12886-016-0186-4 [CrossRef]
- Sugimoto M, Sasoh M, Ido M, Wakitani Y, Takahashi C, Uji Y. Detection of early diabetic change with optical coherence tomography in type 2 diabetes mellitus patients without retinopathy. Ophthalmologica. 2005;219(6):379–385. doi:10.1159/000088382 [CrossRef]
- Wilkinson CP, Ferris FL 3rd, Klein RE, et al. Proposed international clinical diabetic retinopathy and diabetic macular edema disease severity scales. Ophthalmology. 2003;110(9):1677–1682. doi:10.1016/S0161-6420(03)00475-5 [CrossRef]
- Kowluru RA, Engerman RL, Case GL, Kern TS. Retinal glutamate in diabetes and effect of antioxidants. Neurochem Int. 2001;38(5):385–390. doi:10.1016/S0197-0186(00)00112-1 [CrossRef]
- Stitt AW. The role of advanced glycation in the pathogenesis of diabetic retinopathy. Exp Mol Pathol. 2003;75(1):95–108. doi:10.1016/S0014-4800(03)00035-2 [CrossRef]
- Kowluru RA. Retinal metabolic abnormalities in diabetic mouse: comparison with diabetic rat. Curr Eye Res. 2002;24(2):123–128. doi:10.1076/ceyr.22.214.171.12458 [CrossRef]
- Stadelmann C, Lassmann H. Detection of apoptosis in tissue sections. Cell Tissue Res. 2000;301(1):19–31. doi:10.1007/s004410000203 [CrossRef]
- Barber AJ, Lieth E, Khin SA, Antonetti DA, Buchanan AG, Gardner TW. Neural apoptosis in the retina during experimental and human diabetes. J Clin Invest. 1998;102(4):783–791. doi:10.1172/JCI2425 [CrossRef]
- Abu El-Asrar AM, Dralands L, Missotten L, Geboes K. Expression of antiapoptotic and proapoptotic molecules in diabetic retinas. Eye (Lond). 2007;21(2):238–245. doi:10.1038/sj.eye.6702225 [CrossRef]
- Patel PJ, Foster PJ, Grossi CM, et al. Spectral-domain optical coherence tomography imaging in 67 321 adults: Associations with macularthickness in the UK Biobank Study. Ophthalmology. 2016;123(4):829–840. doi:10.1016/j.ophtha.2015.11.009 [CrossRef]
- Pan Y, Varma R. Natural history of glaucoma. Indian J Ophthalmol. 2011;59Suppl:S19–S23. doi:10.4103/0301-4738.73682 [CrossRef]
- Mansoori T, Viswanath K, Balakrishna N. Ability of spectral domain optical coherence tomography peripapillary retinal nerve fiber layer thickness measurements to identify early glaucoma. Indian J Ophthalmol. 2011;59(6):455–459. doi:10.4103/0301-4738.86312 [CrossRef]
Characteristics of Study and Control Groups
||14.2 ± 2.4
||13.4 ± 2.2
|Number of Girls, n (%)
|Spherical Equivalent (Diopters)
||0.02 ± 0.54
||0.06 ± 0.51
||9.2 ± 2.2
|Intraocular Pressure (mm Hg)
||14.6 ± 3.2
||15.0 ± 3.0
Comparison of RNFL and GC-IPL Thickness in All Quadrants
|RNFL Thickness (μm)
||102.0 ± 15.5
||102.0 ± 19.4
||128.8 ± 15.2
||126.4 ± 15.4
||76.9 ± 12.8
||78.5 ± 11.5
||131.5 ± 15.1
||130.7 ± 13.7
||70.9 ± 8.2
||71.9 ± 10.6
|GC-IPL Thickness (μm)
||86.8 ± 5.5
||85.6 ± 5.7
||85.0 ± 6.2
||82.4 ± 5.0
||87.9 ± 6.2
||83.0 ± 5.3
||88.2 ± 5.9
||86.4 ± 5.2
||87.4 ± 4.0
||83.1 ± 4.6
||85.4 ± 4.4
||82.6 ± 4.7
||86.7 ± 5.6
||84.2 ± 6.0
Correlation of GC-IPL and RNFL Thickness With Age, Duration of Diabetes, and HbA1c
||Duration of Diabetes