To investigate the morphologic characteristics of macular commotio retinae using spectral-domain optical coherence tomography (SDOCT) and develop a grading system for traumatic photoreceptor damage.
Retrospective, observational case series.
setting : Seoul National University Bundang Hospital. patients : Forty-nine patients with macular commotio retinae, examined by SDOCT within 7 days of the initial traumatic event. observations : A 4-step grading system was based on the morphology revealed by SDOCT. Best-corrected visual acuity (BCVA) and structural integrity of the 3 photoreceptor layers (cone outer segment tips [COST], inner/outer segment [IS-OS] junction, external limiting membrane [ELM]) were evaluated at baseline, at 1 month, and at the final visit. Visual and anatomic outcomes at 1 month and at the final visit were compared among subgroups. main outcome measures : The severity of photoreceptor damage as revealed by SDOCT and BCVA.
The following 4 distinct photoreceptor morphologic features were observed: increase in IS-OS junction reflectivity with the disappearance of the thin hyporeflective optical space (n = 8, grade 1), COST defect only (n = 5, grade 2), COST and IS-OS junction defects (n = 16, grade 3), and COST, IS-OS junction, and ELM defects (n = 20, grade 4). Eyes with higher grades at baseline had significantly worse visual (final BCVA, P = .002) and anatomic outcomes (complete photoreceptor recovery, P < .001).
The number of disrupted photoreceptor layers, as determined using SDOCT images, can be used to grade macular commotio retinae. This system may be useful in documenting the baseline severity of photoreceptor damage and in predicting visual and anatomic outcomes.
The macula is vulnerable to closed-globe injuries, which may be blunt or photic in origin. Commotio retinae —also known as “Berlin edema”—is frequently observed following blunt eyeball trauma. The pathogenesis of commotio retinae has been suggested to be both mechanical and hemodynamic in nature. Direct mechanical forces were thought to damage photoreceptors and blood vessels in the nerve fiber layer (NFL). Histologic examinations have shown that commotio retinae often involves photoreceptor and retinal pigment epithelium (RPE) damage.
Conventional color fundus photography does not allow for the visualization of microstructural abnormalities in the photoreceptor layers. For example, eyes with occult macular dystrophy generally yield normal fundus photographs, despite the presence of characteristic photoreceptor disruption. Optical coherence tomography (OCT) is an imaging technique that has revolutionized our ability to visualize structural abnormalities of the macula. Compared with older time-domain OCT (TDOCT) technologies, spectral-domain OCT (SDOCT) provides significantly better image resolution and detailed images of retinal morphologic features. These features allow for the evaluation of photoreceptors on a microscopic level.
Several previous studies have used TDOCT or SDOCT to investigate the morphologic features of macular commotio retinae. Saleh and associates showed an increase in photoreceptor inner/outer segment (IS-OS) junction reflectivity in eyes with commotio retinae. Souza-Santos and associates demonstrated that both disruption of the IS-OS junction and hyperreflectivity of the overlying retina were associated with a poor visual prognosis. However, the small numbers of eyes included in these studies (n ≤ 20) limited the conclusions that could be drawn with regard to the OCT findings associated with macular commotio retinae and the presence of any association with anatomic/visual outcomes.
Our primary hypothesis was that morphologic characteristics as detected on OCT may be associated with visual and anatomic outcomes in eyes with macular commotio retinae. More specifically, we hypothesized that commotio retinae accompanied by more severe photoreceptor damage, as detected by OCT, leads to worse visual and anatomic outcomes. In this study, we investigated the morphologic characteristics of macular commotio retinae using SDOCT and correlated them with final visual acuity and photoreceptor anatomic status. Also, we proposed and examined a clinical grading scale for macular commotio retinae based on SDOCT findings.
The medical records of 59 eyes from 59 patients who had funduscopic features consistent with macular commotio retinae were reviewed retrospectively. All patients presented at Seoul National University Bundang Hospital for treatment of blunt trauma during the period from January 1, 2009 to October 31, 2012. All patients were followed up for at least 1 month. All 59 patients met the following inclusion criteria: (1) morphologic macular abnormalities on SDOCT (Spectralis OCT; Heidelberg Engineering Inc, Heidelberg, Germany) images of the traumatized eye, (2) visual decline in the traumatized eye, and (3) normal visual acuity and macular morphology, as determined by OCT, in the contralateral eye. Patients with a history of other ocular trauma, macular disease, ocular surgery, or high myopia (>6 diopters) were excluded from all analyses. We also excluded eyes with other manifestations of ocular trauma, including traumatic cataract, massive subretinal or vitreous hemorrhage, choroidal rupture across the fovea, and/or traumatic macular holes. In order to limit our study to cases of acute injury, cases were also excluded from the analyses if the initial visit was >7 days after the traumatic event. The exclusion of 10 patients resulted in the inclusion of 49 eyes from 49 patients. This study was conducted in accordance with the tenets of the Declaration of Helsinki and approved by the Institutional Review Board of Seoul National University Bundang Hospital.
All patients underwent complete ophthalmic examination, including best-corrected visual acuity (BCVA) measurement, slit-lamp biomicroscopy, indirect ophthalmoscopy, fundus photography, and SDOCT. In addition, we recorded subject age at the time of injury and examined causes of the blunt ocular trauma and presence of orbital wall fractures. All measurements of visual acuities were performed by trained examiners in a masked fashion using a standardized Snellen chart. Snellen visual acuities were converted into logarithm of minimal angle of resolution (logMAR) equivalents for statistical analysis. The numeric scores for profound low vision, counting fingers or worse, were substituted for logMAR values as suggested by Lange and associates.
Follow-up scans were automatically performed in the same location by using an eye-tracking system (ART module; Heidelberg Engineering Inc) at each visit, and identical scan position was confirmed by observers when comparing serial images.
In an attempt to investigate the morphologic characteristics of macular commotio retinae, we examined the integrity of the outer retinal layers by comparing our images to magnified images from previous reports. In each scan, 4 distinct high-intensity bands were identified in the outer retinal layers. From innermost to outermost, these represented the external limiting membrane (ELM), photoreceptor IS-OS junctions, cone outer segment tips (COST), and RPE. The integrity of the 3 photoreceptor layers (ELM, IS-OS, COST) was evaluated on each SDOCT image. Two experienced investigators (S.J.A., K.E.K.), who were masked to individual subject information, independently evaluated all SDOCT images. If there were interpretation discrepancies regarding photoreceptor integrity, another investigator (S.J.W.) was consulted for a final decision. For quantitative analyses, we measured the length rather than the area of reduced photoreceptor reflectivity. The length within 3 mm of the macular center was measured on horizontal as well as vertical scans across the fovea. The loss of reflectivity was quantified by averaging these 2 measurements, as done in previous studies. The combined foveal thickness of the photoreceptor inner and outer segments, between the inner borders of the ELM and the RPE, was measured as well. In eyes with unidentifiable foveal ELM and RPE lines, thickness was measured at the point closest to the fovea where ELM and RPE lines were identifiable. If the nearest point was not within 1 mm of the fovea, combined outer and inner segment (IS+OS) thickness was not measured, as the distance between the ELM and RPE at areas >1 mm outside the fovea can be remarkably different from that measured in the fovea.
Baseline OCT images were compared with images obtained at 1 month and at the final visit to examine changes in the photoreceptor layers as the eye healed. The SDOCT images taken at the final visit were also reviewed to evaluate anatomic outcomes and to assess whether or not a complete photoreceptor recovery had occurred. A complete photoreceptor recovery was defined as 3 structurally intact photoreceptor layers, with no defects in the COST, IS-OS junction, or ELM layers.
Descriptive statistics were used for the data pertaining to demographics, ocular trauma causes, and anterior/posterior segment findings. Clinical features, baseline BCVA, and sizes of the photoreceptor layer defects at baseline were compared among grade subgroups.
The Kruskal-Wallis test was used to compare BCVA between groups at baseline, at 1 month, and at the final visit. The percentages of eyes with normal vision (BCVA of 20/20 or better) or severe vision loss (BCVA of 20/200 or less) were also compared between groups using Fisher exact test. Linear regression analyses were used to examine correlations between the length of the reflectivity loss in each photoreceptor layer and final BCVA. The correlations between photoreceptor IS+OS thickness and final BCVA and between morphologic grades and final BCVA were also examined. The percentage of eyes with complete recovery was also compared among groups using Fisher exact test. Odds ratios (OR) for severe vision loss and complete photoreceptor recovery at the final visit were calculated for OCT grade. Continuous data were presented as mean ± standard deviation. All statistical analyses were performed using SPSS statistical software (version 18.0; SPSS, Inc, Chicago, Illinois, USA). Statistical significance was defined as a P value < .05.
Demographic and Ophthalmologic Findings
Subject demographics and ocular examination findings are presented in Table 1 . Mean patient age was 32.0 ± 16.7 years (range, 4-65 years). Thirty-nine patients (79.6%) were men and 10 (20.4%) were women. The right eye was injured in 26 patients (53.1%) and the left eye was injured in 23 (46.9%). The most common cause of blunt trauma was getting hit with a soccer ball (7 subjects, 14.3%).
|Variable||Mean ± SD Value (Range) or Number (%) of Patients|
|Age (y)||32.0 ± 16.7 (4-65)|
|Follow-up periods (mo)||11.7 ± 12.0 (1-38)|
|Cause of ocular trauma|
|BB or paintball guns||2 (4.1)|
|Golf ball||4 (8.2)|
|Soccer ball||7 (14.3)|
|Traffic accident||6 (12.2)|
|Abnormalities in ocular examination|
|Serous subretinal fluid||7 (14.3)|
|Extrafoveal choroidal rupture||2 (4.1)|
|Extrafoveal subretinal hemorrhage||3 (6.1)|
|Mild vitreous hemorrhage a||2 (4.1)|
|Blow-out fracture, present/CT performed||10/30 (33.3)|
|Microhyphema or traumatic iritis b||40 (81.6)|
On slit-lamp examination, microhyphema or traumatic iritis was observed in 40 of 49 eyes (81.6%). The orbital wall was fractured in 10 of 30 patients (33.3%) who submitted to computed tomography imaging of the craniofacial bone. No subject had immediate or delayed optic nerve damage, as no relative afferent pupillary defect was observed at any point throughout the investigation. In each case, the fundus examination revealed retinal whitening or a loss of retinal transparency that obscured the choroid in the macula of the traumatized eye. Extrafoveal choroidal rupture, extrafoveal subretinal hemorrhage, and mild vitreous hemorrhage (grade 1 according to the vitreous hemorrhage density grading scale ) were noted in 2 (4.1%), 3 (6.1%), and 2 eyes (4.1%), respectively. In 7 eyes (14.3%), serous subretinal fluid was observed on SDOCT scans at the time of diagnosis. In each case, this fluid absorbed spontaneously during the follow-up period.
Optical Coherence Tomography Findings and Morphologic Grading Based on the Findings
As presented in Figure 1 , eyes with macular commotio retinae were subdivided according to the level of photoreceptor damage, on the basis of the following SDOCT image features: an increase in reflectivity of the IS-OS junction with disappearance of the thin hyporeflective optical space (grade 1); a COST defect only (grade 2); COST and IS-OS junction defects (grade 3); and COST, IS-OS junction, and ELM defects (grade 4). Most eyes with grade 2, 3, or 4 defects also presented with grade 1 findings in the macula. The clinical features associated with each type of photoreceptor damage are summarized in Table 2 . There were no significant differences in demographic data, follow-up periods, or the prevalence of orbital wall fractures among groups. However, at least 1 of the groups was different from the others in terms of baseline BCVA ( P = .002) or average defect size ( P < .001), according to Kruskal-Wallis test. Among the causes of trauma, hits with a soccer ball, wood, and paintball or BB gun ammunition were more likely to cause severe commotio retinae of grade 3 or 4.
|Grade 1||Grade 2||Grade 3||Grade 4||P|
|Number of eyes||8||5||16||20||N/A|
|Follow-up period (mo)||6.4 ± 12.5||15.2 ± 16.7||11.3 ± 11.1||13.4 ± 11.4||.13|
|Age (y)||33.9 ± 16.9||32.0 ±12.4||34.2 ± 22.1||29.5 ± 12.8||.85|
|Sex (n), male:female||7:1||3:2||12:4||17:3||.56|
|Cause of trauma (n, if >1)||Rubber, fist (2), baseball (2), glass, metal, traffic accident||Golf ball, glass, foot (2), traffic accident||Soccer ball (2), golf ball (2), foot, baseball, paintball gun, metal, traffic accident (3), wood, fist (2), rubber (2)||Soccer ball (5), golf ball, baseball (3), BB gun, metal (3), traffic accident, wood (3), fist (2), rubber||N/A|
|Orbital wall fracture, present/CT performed, n (%)||2/8 (25%)||1/4 (25%)||2/7 (28.6%)||5/11 (45.5%)||.79|
|LogMAR visual acuity (Snellen range)||0.29 ± 0.28 (20/100-20/22)||0.72 ± 0.39 (20/400-20/40)||1.37 ± 0.72 (HM-20/50)||1.39 ± 0.87 (HM-20/25)||.002|
|COST defect size at baseline (μm)||0||496 ± 242||1101 ± 541||2440 ± 1250||<.001|
|IS-OS defect size at baseline (μm)||0||0||406 ± 376||1699 ± 1165||<.001|
|ELM defect size at baseline (μm)||0||0||0||1130 ± 868||<.001|