The influence of congenital corneal opacity on ERGs obtained using an abbreviated protocol

ABSTRACT Background Visual electrophysiology may be used to assess visual potential in infants with congenital corneal opacities (CCO). It is essential to recognize confounding effects from these opacities on the flash electroretinogram (ERG). Methods ERGs were recorded in awake children employing skin electrodes placed at the lower eyelid crease, both referred to a midfrontal electrode (Fz). A hand-held stimulator was used to present a mixed rod-cone and a dim white stimulus. Recordings were carried out before and after penetrating keratoplasty (PK), when performed. Results Five infants under the age of 12 months with visually significant CCO were evaluated. In all cases, initial ERGs employing the mixed rod-cone stimulus showed well-defined a-wave with reduced amplitude b-wave. Reduction of stimulus intensity resulted in an increase in the b-wave and normalization of the b:a ratio from 1.1 (range 0.7 to 1.3) to 2.8 (range 1.5 to 4.3). In three cases who underwent PK, the postoperative ERGs recorded with a mixed rod-cone stimulus were normal in waveform shape with a mean b:a ratio of 2.0 (range 1.7 to 3.0). Conclusion Selective reduction of the scotopic bright flash ERG b-wave is typically caused by retinal dysfunction that is post-phototransduction or inner retinal. In infants with CCO, scotopic ERGs to bright flashes can show a reduced b:a ratio that improves or normalizes either after PK or stimulus intensity reduction. The study highlights that media opacity can contribute to the generation of an ERG with reduced b-wave in the absence of inner retinal dysfunction.


Introduction
Full thickness congenital corneal opacities (CCO) in children can be a difficult management dilemma, given the reported poor surgical results of penetrating keratoplasty in infants and toddlers (1,2). Measuring vision or assessment of retinal or post-retinal visual pathways in this age group is notoriously challenging due to the reliance on behavioral responses of the infant (3,4). Any method that introduces a degree of objective visual potential assessment is crucial to help navigate the decision tree for the management of these rare cases.
Electrophysiological measures can provide valuable functional information of the visual system during the pre-surgical assessment and have been described in adults undergoing corneal transplants (1,5,6). In adults, it is known that a dense corneal opacity and/or cataract can result in either a decrease in the aand b-wave amplitude of the flash ERG (due to decreased light entering the eye) or increased b-wave flash ERGs due to light scatter (the 'frosted glass' effect) (5,6). The flash ERG a-wave and flash Visual Evoked Potential (VEP) P2 component measures have been shown to be reliable predictors of visual prognosis prior to corneal transplantation (7)(8)(9).
In this study, we describe five infants below 12 months of age with visually significant CCO associated with a reduced bright flash ERG b-wave. The aim of the study is to examine the possible cause of ERG b-wave reduction, by recording ERG with different types of stimuli preoperatively and by recording post-operative ERGs in those that underwent corneal surgery.

Materials and methods
This study was conducted retrospectively from data obtained for clinical purposes. The study was approved and registered by the University of Pittsburgh Institutional Review Board (STUDY21070059). Electroretinogram evaluation of five infants with corneal opacity was undertaken at our tertiary care centre (UPMC, Children's Hospital of Pittsburgh, PA, USA).
Flash ERG responses were recorded using a pediatric ERG protocol with both eyes open, as previously published (10,11). Recordings were performed preoperatively and postoperatively when penetrating keratoplasty was performed. All patients were awake, unsedated and pupils were not dilated for recording. Skin ERG electrodes were placed on the inferior orbital rim close to the lower eyelid. All electrodes were referred to Fz,

Results
Five infants aged less than 12 months with CCO were referred for visual electrophysiology evaluation. At clinical examination, one had bilateral silver appearing corneas with primary congenital aphakia (S1), two had bilateral corneal opacities with variable degrees of irido-corneal and kerato-lenticular adhesions (S2 and S3), one had a failed penetrating keratoplasty in the right eye and a faint unilateral corneal opacity in the left eye (S4), and the last had bilateral posterior polymorphous corneal dystrophy (S5). Patients' characteristics and figures are reported in Table 1 and Figure 1, respectively, while full clinical and treatment histories are given in the supplemental information.
Electrophysiology testing was achieved in all infants. Flash ERGs were evident in each eye. Given the small and localized corneal opacity of the left eye of S4, we excluded these measures from the analysis. In the remaining nine eyes, the ERG responses to the HHS GR4 were atypical in terms of waveform shape with well-defined a-waves and reduced amplitude b-waves resulting in a mean b:a wave ratio of 1.13 (range 0.73 to 1.33) (Figures 1 and 2). All ERGs had a reduced b:a ratio compared to lab normal (mean 3.3, range 2.2 to 6.5) (Figure 1). Reducing the intensity of the GR4 stimulus to GR1 resulted in a normal ERG waveform shape in all nine eyes (Figure 1). There were significant differences in the ERG responses to GR4 and GR1 stimuli.
There was a significant decrease of the a-wave amplitude (Z = −2.40, p = 0.02) and an increase of the b-wave amplitude (Z = −2.66, p = 0.01) of GR1 ERGs compared to GR4 (Figure 2a). There was also a decrease in the a-wave implicit time (Z = −2.20, p = 0.03) and an increase in the b-wave implicit time (Z = −2.53, p = 0.01) (Figure 2b). The combination of increased b-wave amplitude and decreased a-wave amplitude resulted in a group mean b:a ratio of 2.8 (range 1.5 to 4.3).
The extent of the reduced b-wave was also associated with the extent of opacity. In S2 and S3, the eye with the greater corneal opacity had a greater reduced b-wave ( Figure 1). Penetrating keratoplasty was performed in three cases resulting in a normalization of the b:a ratio of 2.0 (range 1.7 to 3.0) to a GR4 stimulus postoperatively ( Figure 2).

Discussion
Visual electrophysiology is a well-described technique for assessing both retinal disorders and disorders of the postretinal visual pathway in neonates, young infants, and children or in those unable to comply with subjective behavioral tests of visual function (14)(15)(16). The a-wave of the full field flash ERG reflects photoreceptor function and cone Off bipolar cell activity, while the b-wave reflects inner retinal layer function. In all eyes of our series, retinal responses to a mixed rodcone stimulus showed atypical waveform shape with a reduced b-wave. These responses would be clinically suggestive of inner retinal layer dysfunction. In the pediatric population, the most common causes of negative or negative approaching ERGs are congenital stationary night blindness and X-linked retinoschisis (17). None of our patients had clinical features or family history of either condition. However, negative ERGs are also observed in many other conditions such as Duchenne muscular dystrophy, central retinal artery occlusion, birdshot In particular, ERGs from the right (RE) and left (LE) eyes in response to a mixed rod/cone (MRC) stimulus and a dim white stimulus (predominantly rod-driven-PRD) in infant subjects (S1-5). Responses from a control age-matched patient are shown for comparison (C). Note the normal ERG waveform shape in the LE of S4, with both GR4 and GR1 stimuli. For S5 we provided the Anterior Segment Optical Coherence Tomography images to highlight corneal thickness and scatter; these features may not be evident from the images provided, since they are taken in the operative room with a collimator to enhance the red reflex. The ERGs in the right side of Figure 1   mixed rod-cone (MRC) a-and b-wave components (fine black grating) and predominantly rod (PRD, gray) at baseline, and mixed rod-cone a-and b-waves in cases s1-3, before (prMRC, coarse grating) and after surgery (poMRC, coarse grating/grey background). *= significant differences; bars represent the standard deviation. chorioretinopathy, melanoma-associated retinopathy, and glycosylation disorders (12,18,19).
In our cases, we demonstrated that the reduced b-wave amplitude to a mixed rod-cone stimulus was likely a result of increased light scatter with the subsequent so-called "photopic hill." The photopic hill describes how the cone b-wave reaches maximal amplitude at an intermediate stimulus intensity, and paradoxically decreases in amplitude with further increases in stimulus strength. Corneal opacities in our patients likely caused adequate scatter to increase the relative stimulus intensity (20). In addition, the short inter-stimulus interval and noncomplete scotopic conditions balanced the ERGs evoked towards cones. Reducing the intensity of our standard mixed rod-cone stimulus resulted in an increased amplitude of the b-wave and in normalization of its waveform shape, thus showing that there was no true inner retinal dysfunction. Further support of increased light scatter as the cause of the reduced b-wave is provided by S2 and S3 that had inter-ocular differences in corneal opacity. The eyes with the least opacity had the least reduced b-wave amplitude (Figure 1).
Despite corneal electrodes evoking superior signal-to-noise ratio responses (21), they may not be tolerated by most infants and young children without sedation or anesthesia. In this study, we adopted a previously published protocol, which has a high diagnostic accuracy compared to the ISCEV-ERG, employing stimuli demonstrated to be equivalent to those employed in ISCEV-ERG recordings (11,13). The protocol is less invasive with responses obtained in a shorter time, specifically with children in mind. Considering the high number of evaluations under general anesthesia that are necessary for infants with CCO treated with penetrating keratoplasty, it may be advisable to avoid unnecessary anesthesia during the preoperative evaluation. Moreover, general anesthesia itself can influence b-wave waveform shape and amplitude (22,23). For these reasons, skin electrodes have been included in the most recent standard update (13,21).
This study highlights an ERG abnormality secondary to the effects of CCO, potentially relevant to other ERG protocols, including international standard methods, as the ERG response to our mixed rod-cone stimulus (GR4) can be misleading and suggestive of inner retinal layer dysfunction. Although experienced electrophysiologists would further investigate a significantly reduced b-wave employing an intensity series to determine real inner retinal pathology or the effect of the photopic hill, as ERG equipment becomes more available and cheaper, recordings are being obtained and interpreted by nonelectrophysiologists.
These findings may influence the pre-surgical decisionmaking process, especially as CCO may, rarely, be associated with retinal and cortical developmental abnormalities (24,25). Distinguishing this artifact caused by the isolated corneal opacity from systemic conditions with associated retinal involvement is important to appropriately manage these patients at an early age. A further confirmation is derived from the normalization of the b:a ratio after surgery (Figure 2).
A limitation to this study is the small sample size, but CCO are rare conditions. Moreover, we used a non-standard non-Ganzfeld method, and although there may be advantages in testing young children, they are unlikely to provide the same preferential stimulation of rods. This may also explain the magnitude of the "photopic hill phenomenon" under scotopic conditions, and short dark-adaptation and non-standard methods should be acknowledged as possible contributory factors.
Visual electrophysiological assessments are of benefit in infants with corneal opacities and even in the most severe cases can provide information regarding retinal and visual pathway function. When ERG responses in eyes with clinically significant CCO have reduced b-wave amplitude, the photopic hill related to the "frosted glass" effect should be considered as a cause, and further investigated employing reduced intensity stimuli. If despite this adaptation the reduced b-wave amplitude persists, there may be a co-retinal morbidity which would affect management decisions.

Disclosure statement
No potential conflict of interest was reported by the author(s).

Funding
P30 CORE award EY08098 from the National Eye Institute, NIH, and unrestricted supporting funds from The Research to Prevent Blindness Inc, NY, the Eye and Ear Foundation of Pittsburgh and the Children's Foundation.

Data availability statement
The data that support the findings of this study are not publicly available because they contain information that could compromise the privacy of research participants but are available from K.K.N. upon reasonable request.

Statement of ethics
The study was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. This study protocol was reviewed and approved by the University of Pittsburgh Institutional Review Board (approval number STUDY21070059).