To compare macular and peripapillary retinal nerve fiber layer (RNFL) thickness in children from diabetic compared with nondiabetic pregnancy.
As part of the Sydney Myopia Study, 2367 children from grade 7 (age range 11.1 to 14.4 years) completed detailed ocular examinations during 2004–2005. Examination included determination of best-corrected visual acuity (logarithm of minimal angle of resolution) and autorefraction after cycloplegia. Axial length was measured using noncontact interferometry. Optical coherence tomography (OCT) was performed using Stratus OCT through dilated pupils. Participants and parents completed comprehensive questionnaires including questions on birth parameters and medical history.
There were 1741 and 1687 children with adequate-quality RNFL and macular scans respectively, who had complete examination and questionnaire data. There were 57 children from diabetic pregnancies who had both RNFL and macular scans. Children from diabetic pregnancies had significantly thinner inner (264.9 μm vs 270.2 μm, P = .007) and outer (231.9 μm vs 238.6 μm, P = .0001) macular thickness and macular volume (6.75 mm 3 vs 6.92 mm 3 , P = .0003) compared with children from nondiabetic pregnancies. However, central macular thickness, foveal minimum thickness, and RNFL parameters were not significantly different between the 2 groups.
Diabetes during pregnancy is associated with changes in retinal morphology in the offspring. Thinning of the pericentral macular parameters was evident in Stratus OCT scans of children from diabetic pregnancies. These findings suggest the possibility that maternal diabetes impacts on the development of the retina.
Maternal diabetes occurs in 2% to 8% of pregnancies in Australia and 4% to 14% of pregnancies in the United States. The effects of maternal diabetes on the fetus are numerous. Increased serum glucose concentrations in the mother are transferred to the fetus via the placenta. In response, the fetus increases insulin secretion to compensate. This has a secondary effect of causing excessive growth and a large-for-gestational-age fetus. These larger fetuses are more prone to shoulder dystocia and neonatal asphyxia during vaginal delivery. Postdelivery complications include hypoglycemia, infant respiratory distress syndrome, cardiomyopathy, and polycythemia. During adolescence, children of diabetic pregnancies are more likely to be obese and to develop metabolic syndrome or type 2 diabetes. Maternal type 1 diabetes has been associated with the development of superior segmental optic nerve hypoplasia in offspring. However, we are unaware of any studies that have systematically tested for changes in retinal structure in children whose mothers have had diabetes during pregnancy.
Time-domain optical coherence tomography (OCT) allows imaging of total retinal and nerve fiber layer thickness with a resolution up to 8 μm. It is increasingly used in the diagnosis and monitoring of retinal diseases including age-related macular degeneration, diabetic retinopathy, and glaucoma. The purpose of this study was to determine whether diabetes during pregnancy affects the OCT-measured retinal structure in offspring, by comparing with children born of nondiabetic pregnancy.
As part of the Sydney Myopia Study all eligible year-7 students from 21 high schools were invited to undergo ocular examination. Schools were selected based on socioeconomic stratification of Sydney city. The study methods have been described in detail previously.
The ocular examination included visual acuity (VA) and subjective refraction in children whose presenting VA was <0.02 logMAR (logarithm of minimal angle of resolution) units. Axial length (AL) measurement was performed using noncontact partial coherence laser interferometry (IOLMaster; Carl Zeiss Meditec, Jena, Germany). Cycloplegia and dilation was achieved with cyclopentolate 1% and tropicamide 1% administered twice (5 minutes apart); phenylephrine 2.5% was used if adequate mydriasis (≥6 mm) could not be achieved. Autorefraction (Canon RK-F1; Canon, Tokyo, Japan) was carried out at least 25 minutes after instilling eye drops.
The Stratus OCT (OCT3, software version 4.0.1; Carl Zeiss Meditec, Dublin, California, USA) obtains cross-sectional retinal tomographic scans, which have been found to be highly reproducible. The macular “fast” protocol scans consisted of 6 radial scans each with 128 A-scans over a 6-mm distance. This circular region was divided into 3 concentric areas with boundaries at 0.5 mm, 1.5 mm, and 3 mm from the scan center, termed the central, inner, and outer macula respectively. The peripapillary retinal nerve fiber layer (RNFL) “fast” protocol scans consisted of 256 A-scans arranged as a 3.4-mm-radius circle centered on the optic disc.
A 193-item questionnaire was completed by parents. The questionnaire also specifically asked whether the mother has diabetes or had developed diabetes during pregnancy and whether her child had been diagnosed with diabetes.
Statistical analyses were performed using SAS software (version 9.2; SAS Institute, Cary, North Carolina, USA). Only OCT scans of the right eye that were complete and with signal strength greater than 5 were used in our analyses. Means and 95% confidence intervals of baseline characteristics (age, height, weight, body mass index [BMI], birth weight, AL, spherical equivalent refraction, and visual acuity) were calculated and t tests were used to compare differences in means between gestational diabetes and nondiabetic pregnancy groups. The χ 2 test of proportions was used to compare frequencies of sex and ethnicity between groups. Mixed linear models were used to compare OCT parameters (adjusted for age, sex, height, AL, and ethnicity) between groups, with the school attended as the random effect.
Between January 2004 and December 2005, 2367 high school children were tested, of whom 1741 and 1687 had adequate-quality OCT scans of RNFL and macula respectively, as well as having complete questionnaire and examination data.
Table 1 presents the characteristics of the nondiabetic and diabetic pregnancy groups. 57 of 1741 individuals in the diabetic pregnancy group (3.3%) had a mean age of 12.7 years, of whom 63.1% (36/57) were boys. AL was larger in the diabetic pregnancy group (23.6 mm) compared to the nondiabetic pregnancy group (23.4 mm), with borderline significance ( P = .05). There were no reports of a diabetes diagnosis (either by the child or parent) from any of the children in the diabetic pregnancy group, whereas 2 children in the nondiabetic pregnancy group had a diagnosis of diabetes.
|Nondiabetic Pregnancy N = 1684||Diabetic Pregnancy N = 57||P Value|
|Age, mean years (CI)||12.7 (12.7–12.7)||12.7 (12.6–12.8)||.60|
|Boys, n (%)||873 (51.6)||36 (63.1)||0.11|
|Ethnicity, n (%)|
|Caucasian||1052 (62.2)||29 (50.0)||.06|
|East Asian||250 (14.8)||13 (22.4)||.11|
|South Asian||84 (5.0)||3 (5.2)||.94|
|Middle Eastern||101 (6.0)||3 (5.2)||.80|
|Other||205 (12.1)||10 (17.2)||.24|
|Anthropometry, mean (CI)|
|Height (cm)||156.0 (155.6–156.3)||156.6 (154.5–158.8)||.52|
|Weight (kg)||50.0 (49.4–50.6)||52.6 (48.8–56.3)||.15|
|BMI (kg/m 2 )||20.4 (20.2–20.6)||21.2 (20.0–22.4)||.14|
|Birth weight (g)||3381 (3323–3437)||3411 (3277–3545)||.68|
|Ocular parameters, mean (CI)|
|Visual acuity a||56.4 (56.2–56.7)||55.7 (53.9–57.6)||.47|
|Axial length (mm)||23.4 (23.3–23.4)||23.6 (23.4–23.8)||.05|
|SER (diopters)||0.51 (0.46–0.57)||0.15 (−0.29–0.58)||.10|