Purpose
To assess retinal morphology changes in patients born at different stages of prematurity, accounting for the presence or absence of retinopathy of prematurity (ROP) and comorbidity, using spectral-domain optical coherence tomography (SD OCT).
Design
Retrospective cohort study.
Methods
Preterm and term infants underwent an ophthalmologic assessment (best-corrected visual acuity, stereoacuity, cycloplegic refraction and funduscopy). Retinal layers were imaged, segmented, and measured by SD-OCT. In total 114 full term controls and 60 preterm children, classified as late preterm (32–36 weeks gestational age), early preterm (<32 weeks of gestational age) without ROP, and early preterm with previously treated ROP, were included in the study.
Results
No retinal structure differences were observed in preterm infants with no treated ROP compared to term infants. Early preterm infants with previous treated ROP had decreased retinal nerve fiber layer (RNFL) thickness in the superior and nasal quadrants, increased RNFL in the temporal quadrant, and a thinner ganglion cell and inner plexiform layer complex (GCL-IPL). Low birthweight percentile was associated with increased foveal thickness and ganglion cell damage (RNFL and GCL-IPL) independent of gestational age. Among all the coexisting events, inflammation and hypoxia were correlated with more severe detrimental effects.
Conclusions
In the absence of treated ROP, prematurity was not associated with disturbed retinal structure. Severe ROP and low birthweight were related to neuronal and axonal damage in the inner retinal layers. Detailed comorbidity should be reviewed when evaluating preterm infants.
Premature birth interrupts developmental processes, including neuronal differentiation and cell migration in all neurological tissues. Premature birth modifies the developmental environment, depriving the fetus of certain critical inputs for neurodevelopment. Prematurity also predisposes an infant to pathologic events that may directly damage neuronal tissues.
As part of the central nervous system, the retina may also be affected by prematurity. The structural consequences of prematurity, retinopathy of prematurity (ROP), have been characterized prevously. However, significant advances in the field have been made owing to improved digital imaging technology. The latest imaging device, spectral-domain optical coherence tomography (SD OCT), presents a unique opportunity to accurately assess the retinal morphology in preterm infants. It provides images with an axial resolution of approximately 5 μm and allows retinal layer segmentation.
Previous studies of preterm infants reported inconsistent results for the retinal nerve fiber layer and macular thicknesses. The inner retina appears to have a higher risk for structural abnormalities. However, why certain preterm children exhibit retinal abnormalities while others do not remains unclear.
Prematurity is the final consequence of various intrauterine pathologies that affect fetal development. In order to accurately assess prematurity’s effects on retinal structure, many prenatal and postnatal noxious events should be accounted for in addition to gestational age. Most available studies report results from heterogeneous study groups with wide gestational age ranges per group. These studies did not account for important perinatal outcomes such as intrauterine growth restriction, adverse perinatal events, or concomitant diseases as confounding variables.
This study aimed to assess morphologic retina changes in different prematurity stages taking into account the presence or absence of ROP and comorbidity in the preterm infants. To this end, we used SD OCT to determine the measurements and segmentation values.
Materials and Methods
All procedures adhered to the tenets of the Declaration of Helsinki. The study protocol was prospectively approved by the local ethics committee (Comité ético de Investigación clínica de Aragón; CEICA). Written informed consent was obtained from the parents or guardians of each child.
The study involved 2 cohorts of children consecutively recruited in the Pediatric Ophthalmology Unit of the Miguel Servet University Hospital, Zaragoza (Spain). The first were less than 37 weeks of gestational age at birth (preterm group) and the second were between 37 and 42 weeks of gestational age at birth (term group). Preterm children were classified according to their gestational age at birth, either late preterm (from 32 to 36 weeks) or early preterm (less than 32 weeks). Early preterm children were further divided into 2 groups, depending on whether the children were previously treated for ROP with diode laser photocoagulation. Small-for-gestational-age neonates were defined as those children whose birthweight was below the 10th percentile according to local standards. Prenatal, neonatal, and infancy data were collected from medical records. Examiners were masked to perinatal data and the study group of the studied patients.
All early preterm children received ROP screening starting at 4 weeks of age, which was repeated every 1–3 weeks until the retina was fully vascularized based on funduscopy results. Children with threshold and prethreshold type 1 ROP were all treated with diode laser photocoagulation.
Subjects with a history of ocular diseases other than refractive errors or ROP, genetic or chromosomal defects, congenital malformations, or significant refractive errors (greater than 5 diopters spherical equivalent refraction or 3 diopters astigmatism) were excluded from the study.
All children underwent full ophthalmologic assessments including best-corrected visual acuity (using the logMAR scale), stereoacuity (TNO test), cycloplegic refraction, funduscopy, and retinal assessment with SD OCT. SD OCT scans were performed by 1 of 2 experienced examiners (I.A. or G.G.) using the Cirrus HD-OCT (Carl Zeiss Meditec Inc, Dublin, California, USA).
The retina was imaged using the Optic disc cube 200 × 200 and the Macular cube 200 × 200 protocols. The Optic disc cube protocol analyzes a 6 mm 2 grid of 200 horizontal scan lines. The software algorithm automatically detects the optic disc center from this volume scan, positions a 3.46-mm-diameter calculation circle over this point, and calculates the thickness. The Macular cube 200 × 200 protocol performs 6 consecutive macular radial scans, 6 mm in length, centered on the fovea. The images were analyzed using the OCT3 mapping software. Retinal thickness was defined as the distance between 2 interfaces: the retinal pigment epithelium (RPE)–choriocapillaris interface and the vitreoretinal interface. Foveal thickness was defined as the mean thickness at the intersection point of the 6 radial scans including the distance between the inner retina and the RPE. The ganglion cell analysis algorithm segmented a 14.13 mm 2 elliptical annulus area centered on the fovea. The complex ganglion cell and inner plexiform layers (GCL-IPL), excluding the retinal nerve fiber layer (RNFL), were measured.
The following OCT measurements were analyzed: average RNFL thickness, thickness in the 4 quadrants (superior, temporal, inferior, and nasal), average macular thickness, foveal thickness, average GCL-IPL, and minimum GCL-IPL.
An internal fixation target was used to maximize reproducibility. Images with a quality score lower than 7 were rejected and repeated. No patient was excluded owing to poor image quality. Only the patients’ right eyes were included in the analysis.
Statistical analyses were performed using SPSS 21.0 (SPSS Inc, Chicago, Illinois, USA) statistical software. ANOVA tests and Pearson χ 2 tests or Fisher exact tests were used to compare the quantitative and qualitative data, respectively. Perinatal factors potentially influencing retinal measurements were analyzed by forward stepwise multiple linear regression analysis. The included factors were gestational age at birth, birthweight, birthweight percentile, maternal smoking, and adverse perinatal events including maternal pre-eclampsia, cerebral hypoxic events, inflammatory events, anemia, bronchopulmonary dysplasia, apneas, intraventricular hemorrhage, and treated ROP. Cerebral hypoxic events were defined as any condition that deprives of an adequate oxygen supply to the brain. Any situation, other than ROP and its treatment, resulting in an inflammatory response, clinically and analytically, was considered as inflammatory event.
Results
The study included 60 preterm children and 114 controls born at term. Among the preterm children, 33 were classified as late preterm (gestational age from 32 to 36 weeks) and 27 as early preterm (gestational age less than 32 weeks). Of the early preterm patients, 17 had not undergone previous ROP treatment (without treated ROP) and 10 had been treated previously (with treated ROP). The age at examination ranged from 4 to 14 years in the term group and from 4 to 13 years in the preterm group.
The study groups’ descriptive characteristics, including gestational age, weight at birth, age at study inclusion, sex, visual acuity, refractive error (presented as the spherical equivalent), and stereoacuity, are summarized in Table 1 . No difference in visual function was observed in children without ROP compared to the term group. Children with ROP showed significantly worse visual acuity and stereoacuity and had higher refractive errors on average.
| At Term | Late Preterm | Early Preterm Without ROP | Early Preterm With ROP | P | |
|---|---|---|---|---|---|
| Gestational age at birth, wk | 39.39 (1.20) | 34.94 (1.46) | 29.71 (1.26) | 27.80 (1.23) | <.001 a |
| Birth weight, g | 3302.12 (515.67) | 2308.48 (584.96) | 1307.50 (404.35) | 1070 (283.78) | <.001 a |
| Birth weight, percentile | 51.00 (32.08) | 35.03 (35.34) | 29.59 (36.02) | 23.80 (29.23) | <.001 a |
| Age at examination, y | 8.11 (2.25) | 8.65 (1.82) | 7.65 (2.51) | 8.64 (2.73) | .409 |
| Sex, male:female | 56:57 | 14:17 | 4:13 | 6:4 | .203 |
| VA, logMAR | 0.01 (0.13) | −0.03 (0.08) | 0.06 (0.11) | 0.52 (1.04) | <.001 a |
| Refractive error, spherical equivalent | 0.25 (1.28) | 0.65 (1.86) | 1.14 (2.39) | −1.56 (5.30) | .006 a |
| Stereocuity, degrees of arc | 77.33 (61.51) | 101.43 (216.96) | 73.33 (26.46) | 415 (552,18) | <.001 a |
| N | 114 | 33 | 17 | 10 |
Table 2 shows retinal OCT measurements. Using the OCT Cirrus software, the retina was segmented into 3 structures: the RNFL, the GCL-IPL complex, and the outer retina (reported as foveal thickness). In the absence of treated ROP, no differences were found in the average RNFL, GCL-IPL, fovea, or macula measurements among term-born, late preterm, or early preterm children. Early preterm children with previous treated ROP had flatter RNFL profiles with decreased thickness in superior and nasal quadrants. In addition, ROP treatment resulted in a thicker temporal quadrant and a thinner average and minimum GCL-IPL complex. No structural differences in the macula or fovea were observed among the study groups.
| At Term | Late Preterm | Early Preterm Without ROP | Early Preterm With ROP | P | |
|---|---|---|---|---|---|
| Average RNFL thickness, μm | 96.96 (10.79) | 97.03 (10.69) | 94.56 (12.89) | 90.60 (15.04) | .328 a |
| Superior RNFL thickness, μm | 121.42 (18.31) | 123.70 (17.72) | 118.19 (31.16) | 95.00 (21.91) | .001 a |
| Temporal RNFL thickness, μm | 68.04 (12.56) | 65.27 (10.21) | 66.19 (13.80) | 99.00 (30.91) | <.001 a |
| Inferior RNFL thickness, μm | 127.83 (19.14) | 128.70 (23.19) | 120.13 (16.12) | 111.60 (36.46) | .064 |
| Nasal RNFL thickness, μm | 70.58 (14.52) | 70.61 (12.58) | 73.38 (11.32) | 56.50 (31.29) | .036 a |
| Foveal thickness, μm | 99.24 (2.98) | 99.95 (1.97) | 97.14 (7.71) | 99.63 (1.70) | .243 |
| Macular thickness, μm | 279.51 (17.00) | 274.21 (16.73) | 280.53 (12.09) | 282.60 (16.22) | .333 |
| Minimum GCL-IPL, μm | 79.24 (12.04) | 77.87 (16.03) | 70.31 (14.26) | 36.75 (24.97) | <.001 a |
| Average GCL-IPL, μm | 83.35 (8.57) | 83.57 (4.98) | 82.35 (4.96) | 65.30 (18.52) | <.001 a |
| N | 114 | 33 | 17 | 10 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
