Premier article

Article en révision pour « Documenta Ophthalmologica »

Résumé

Nous avons évalué la possibilité d’effectuer un ERG sur les yeux non dilatés en utilisant de fortes intensités lumineuses. Aussi, nous avons examiné les intensités lumineuses standards recommandées par l’ISCEV ( International Society for Clinical Electrophysiology of Vision ) lors de l’enregistrement d’un ERG photopique (cônes) sur les yeux dilatés. Les courbes intensité-réponse aux ERG photopique (N=26, fond lumineux 25.5 cd/m^-2) et scotopique (N=23, lumière verte) ont été obtenu sur un oeil dilaté (OD) et un oeil non-dilaté (OND). Une réduction significative de 14% de l’OND a été observée au niveau du Vmax, en condition photopique (p<0.0001). Un allongement du temps implicite de l’onde b (p<0.0001) ainsi qu’une baisse de la sensibilité rétinienne (logK) de 0.38 unité log (p<0.0001) ont été observés au niveau de l’OND comparativement à l’OD. En augmentant le fond lumineux à 80 cd/m^-2, le temps implicite semble se normaliser à l’OND, alors que l’amplitude du Vmax demeure plus basse pour l’OND. L’amplitude de l’onde b est réduite de 51.5% et 36% dans l’OND pour les intensités lumineuses standards recommandées par la ISCEV à 1.5 et 3.0 cd^.s^.m^-2 respectivement. Le Vmax des bâtonnets est diminué de 7% à l’OND (p<0.05), mais le temps implicite n’est pas affecté. Nous concluons que l’augmentation du fond lumineux en condition photopique de 0.5 unité log semble corriger pour le temps implicite et la forme de l’onde observée à l’ERG pour l’OND. Cependant, une augmentation des intensités lumineuses ne corrige pas complètement l’amplitude de l’onde b. Mentionnons que le Vmax de l’OD est atteint en moyenne à environ 3.9 cd^.s^.m^-2, ce qui suggère que les intensités lumineuses standards de la ISCEV en condition photopique de 1.5 et 3.0 cd^.s^.m^-2 ne sont pas adéquates, spécialement si la dimension de la pupille est réduite et ne peut atteindre une dilatation complète comme c’est le cas chez les personnes âgées.

Assessing the impact of non-dilating the eye on the standard ERG response.

Abstract

We investigated the possibility of performing ERG in non pharmacologically dilated eyes using brighter flashes as well as assessing if the current photopic ISCEV standard flash (SF) range was optimal to record cone ERG in the dilated eye. Photopic (N=26, background 25.5 cd/m^-2) and scotopic ERGs (N=23, green flashes) luminance response functions were obtained simultaneously in a dilated (DE) and non dilated eye (NDE). Compared to the DE, the NDE photopic Vmax b-wave amplitude was reduced by 14% (p<0.0001), implicit time prolonged (p<0.0001) and retinal sensitivity (log K) decreased by 0.38 log units (p<0.0001). Increasing the background to 0.50 log units (80 cd/m^-2) appeared to correct for the implicit time anomaly but not for the amplitude. In the DE, Vmax was observed on average at 3.9 ± 1.0 cd^.s^.m^-2. At the ISCEV SF of 1.5 and 3.0 cd.s.m^-2, the b-wave was reduced by 31% and 13% respectively compared to Vmax. Rod Vmax was decreased by 7% in NDE (p<0.05), but implicit time was not affected. When non dilating the eye pharmacologically, we could not achieve maximal retinal response even in the scotopic ERG. However, our data are showing that such maximal response is not currently achieved even when dilating the eye pharmacologically due to the fact that the current SF intensities are not optimal. The latter could be worsening if pupils’ size do reach full dilation such in elderly patients.

Keywords: dilated eye, electroretinogram, non dilated eye, photopic, scotopic, standard flash, ISCEV

Abbreviations used: DE = dilated eye, ERG = electroretinogram, NDE = non dilated eye, SF = standard flash

Introduction

One of the main goals of the International Society for Clinical Electrophysiology of Vision (ISCEV) is to establish standard methods for full-field electroretinography (ERG) and to update them every 4 years. Clinical standards include light stimulation, electrode types, and pupillary dilation to name a few [1]. Full dilation is recommended for the ERG in order to achieve optimal stimulation of the entire retina. To this end, a standard white flash (SF) in the range of 1.5 to 3.0 cd^.s^.m^-2 is recommended to achieve stable and reproducible photopic cone ERG assessment. However, dilatation of the pupil cannot always be achieved due to allergic reaction. Moreover, the pupil of some elderly patients do not dilate much, which may contribute to the reduction of the ERG amplitude already present due to aging [2]. However, at the present time, there is no data regarding how much reduction should be expected in the non dilated eye (NDE) when using the ISCEV SF. Moreover, if a substantial reduction is observed, could it be possible to compensate for the pupil size by increasing the SF intensity? In fact, technology in recording systems has evolved to a point that very bright flashes intensities can now be achieved.

Our aim was to assess, using the Espion system (Diagnosys LLC, Midleton), the impact of the NDE on the photopic and scotopic luminance response function and then to evaluate what intensity would be needed to achieve the same waveform, implicit times and amplitudes as per the ERG performed on the dilated eye (DE). Taking advantage of our protocol, a secondary aim of our study was to assess if the current ISCEV SF intensity range is indeed optimal to achieve the maximal photopic ERG response.

Participants and Method

Participants

This study was approved by the institutional ethics committee; all patients signed a written informed consent and were paid for their participation.

Photopic ERG recordings (always at 11h00) were performed in 26 participants (6 men, 20 women, aged 21 to 49; mean 29.7 ± 7.2 years) between May 2004 and July 2005. Fifty minutes before the photopic ERG, a scotopic ERG was obtained in all participants, but the data were analysed in only 23 (5 men, 18 women, aged 21 to 49; mean 30.0 ± 8.0 years), due to electrode displacement during recording in the dark - which could be corrected before performing the cone ERG. All participants underwent a formal ophthalmologic exam (ocular history, visual acuity, ophthalmoscopy, slit-lamp examination and intraocular pressure) and were selected only if they had no abnormalities or family history of retinal disease and cataracts along with normal visual acuity of 20/20 or better with corrections of less than ± 5 diopters. Exclusion criterions included irregular menstrual cycles, pregnancy, lactation, past or present substance abuse and night shift workers (to avoid any circadian rhythm disturbances).

Procedure

ERG recording

The eye to be dilated was first anaesthetized by a drop of proparacaine hydrochloride (Alcaine) 0,5% and then fully dilated 1 minute later using tropicamide 1% ( Mydriacyl ^®). Recordings were obtained simultaneously in both eyes (dilated and non-dilated) with DTL electrodes (Shieldex 33/9 Thread, Statex, Bremen, Germany) secured deep into the conjunctival sac. Ground and reference electrodes (Grass gold cup electrodes filled with Grass EC2 electrode cream) were pasted on the forehead and external canthi respectively, as previously described [3, 4].

Flash stimulations were administered with a Ganzfeld (Color Dome; Espion system, DIAGNOSYS LLC, Littleton, MA) to achieve full field retinal stimulation. Participants were first dark adapted for a period of 30 minutes and a scotopic luminance response function was obtained with increasing green light intensities (Color dome standard green LED) ranging from -4.25 to -1.00 log cd^.s^.m^-2with inter-stimuli intervals ranging from 1.5 s (low intensities) to 5 s (high intensities). For each intensity, at least 10 responses or more were averaged in order to achieve an optimal high signal to noise ratio. Participants were then light adapted for 15 minutes with a background light (25.5 cd/m^-2) to prevent the light adaptation effect [5, 6]. A photopic luminance response function was established using 10 intensities ranging from -1.12 to 1.375 log cd^.s^.m^-2with an inter-stimuli interval set at 1.5 second. White flash stimulation was provided by the LEDs of the Diagnosys Color Dome.

Following a visual inspection of the data in the NDE, we observed that the maximal intensity provided by the LEDs (24 cd^.s^.m^-2) did not appear to be sufficient to generate the typical decline in photopic ERG amplitude that usually follows the Vmax. We therefore performed the same protocol with six participants (4 men, 2 women, aged 23 to 34; mean 26.3 ± 3.9 years) using a higher range of intensities (0.076 to 800 cd^.s^.m^-2), provided by the Xenon flash system also part of the Diagnosys Color Dome. The latter protocol allowed us to generate (even in the NDE) the so called Photopic hill as previously described [7, 8].

ERG analysis

DTL placement was assessed at the beginning and at the end of the recording session. If displacement was observed, the recording was excluded for analysis. The latter occurred in 3 participants during the scotopic ERG as indicated in the method section.

A-wave amplitude was measured from baseline to trough, and b-wave, from the trough of the a-wave to the peak of the b-wave. The b-wave amplitudes were plotted against flash intensities in order to generate the photopic luminance response function from which Vmax, log K (intensity necessary to reach one half of the saturation amplitude: ½ Vmax) and slope (n) parameters were calculated according to reported methods [4, 8, 9]. We calculated these parameters with the Origin 7.0 software (OriginLab Corporation, Northhampton, MA), following a sigmoidal curve fitting. By convention, Vmax represented the saturating maximal amplitude of the b-wave and log K, retinal sensitivity [10]. Analysis of the ERG included the b-wave and a-wave amplitudes and implicit times observed at the Vmax in the DE and NDE.

Statistical analysis consisted in paired t-tests performed between DE and NDE for each parameter. We also calculated the percentage difference in amplitude between the two eyes at the proposed ISCEV SF range of 1.5 to 3.0cd^.s^.m^-2.

Results

Photopic condition

Figure 1 presents the mean luminance response function of 26 participants achieved with the flash LED in the DE and NDE. A right shift along the x-axis can be observed in the NDE, suggesting a decreased retinal sensitivity whereas Vmax does not appear to reach the same amplitude compared to the DE.

Insert Figure 1 about here

Figure 2 presents two typical examples of photopic luminance response functions as obtained in the DE and NDE in two women (24 and 34 years old respectively).

Insert 2a and 2b figures about here

Whereas in participant 1 (Fig 2a) it is clear that Vmax was achieved in both the DE and NDE ─ as exemplified by the decrease in amplitude at the end of the luminance response function typical of the Photopic hill [7, 8], ─ in participant 2 (Fig 2b), such decrease in amplitude was not observed. Therefore, in participant 2, we could not be certain that Vmax was achieved. The latter observation occurred in five participants (1 man; 4 women, aged 34 to 44; mean 38.8± 4.55 years). Following the latter finding (during analysis), we decided to test six new participants with the Xenon flash in order to see if in fact Vmax would be achieved at a higher intensity that the one provided by the LED system.

Insert Figure 3 about here

Figure 3 shows the mean luminance response function of six participants obtained in the DE and NDE using the Xenon Flash. Due to the higher flash intensities provided by the Xenon, a complete Photopic hill could now be observed in both eyes. Of interest, as per Figure 1, Vmax in the NDE is still lower than Vmax in the DE. There is also the expected right shift in the Photopic hill , but the typical plateau at the end appears higher in the NDE compared to DE. The latter observation was common to all six participants tested.

Amplitude of b-wave and a-wave

Table 1 presents the intensity at which Vmax was observed in the DE compared to the NDE for the group of 26 participants who were submitted to the flash intensities provided exclusively by LEDs and the second group of six participants that were submitted to brighter flash intensities in order to generate a complete Photopic hill , using Xenon flashes. At the respective values of Vmax, b-wave amplitude was on average 14% lower in the NDE (p<0.0001; 98.7µV) when compared to DE (115.1 µV), but there was no difference in a-wave amplitude between the NDE and the DE (p>0.05).

Insert Table 1 about here

In the Xenon group, Vmax b-wave amplitude was about 12% lower in the NDE compared to the DE (p=0.02). The a-wave amplitude at Vmax was not significantly different between both eyes (p>0.05).

Intensity of Vmax

In the Xenon group, Vmax was observed on average at an intensity of 3.3 ± 0.81 cd^.s^.m^-2. The corresponding number for the NDE was 12.4 ± 1.90 cd^.s^.m^-2, which is a significant increase of 0.57 log unit (p=0.0001) (See Table 1).

Standard flash (SF)

Looking at the low and high range of the photopic ERG ISCEV SF of 1.5 and 3.0 cd^.s^.m^-2, a significant difference could be observed between the DE and NDE (p<0.0001). On average, at 1.5 cd^.s^.m^-2, the b-wave amplitude in the NDE was 51.5% lower than in the DE (See Table 1). At 3.0 cd^.s^.m^-2, the b-wave amplitude in the NDE was on average 36% less than the DE (p<0.0001). Similar differences could be observed in the Xenon group with 52% and 36% reduction in the NDE at intensities 1.5 and 3.0 cd^.s^.m^-2 respectively (p=0.0057). Of interest, b-wave amplitudes observed at 1.5 cd^.s^.m^-2 and 3.0 cd^.s^.m^-2 were on average at 69% and 87% of the Vmax b-wave amplitude respectively.

Log K and slope

As expected, log K values were significantly higher for the NDE (0.27 log unit) when compared to the DE (-0.11 log unit), which indicates a decrease in retinal sensitivity in the NDE in the order of 0.38 log units (p<0.0001). For the Xenon group, a similar result was found (0.34 log unit) (See Table 1).

The slope was also significantly different between both eyes, with the curve of the DE reaching its Vmax more rapidly than the NDE (p<0.0001) (See Table 1). However, in the Xenon group, there was no difference in the slope between both eyes, but a type II error can occur due to the small sample size (N=6).

Implicit time of b-wave and a-wave at respective Vmax

Table 2 presents the implicit time of a-wave and b-wave at the respective Vmax, as observed in the DE and NDE for the LED group of 26 participants and the Xenon group of six participants.

Insert Table 2 about here

The b-wave implicit time was on average 3.3 ms longer at Vmax in the NDE (p<0.0001) which represents an increase of 10% when compared to the DE. A-wave implicit time was also 1.2 ms longer at Vmax in the NDE (p<0.0001), which also represents an increase of 10% when compared to the DE.

The b-wave implicit time was also, on average, 4.5 ms longer at Vmax in the NDE for the Xenon group (p<0.0001), which represents an increase of 14% when compared to the DE. Similarly to the main group (LED group), a-wave implicit time was 1.5 ms longer in NDE (p= 0.0172), representing an increase of 9.8% when compared to the DE.

Background adjustment

Following the observation that we could not achieve the same Vmax amplitude and implicit time in the NDE even after compensating with brighter flash intensities we changed the photopic background to 80 cd/m^-2, which corresponds to an increase of 0.50 log unit. This photopic background was selected based on the fact that the two luminance response curves (DE vs NDE) were shifted approximately by 0.50 log unit. The latter experiment was performed in four participants (2 men, 2 women, aged 24 to 26; mean 25 ± 0.4 years), who have already gone through the Xenon experiment.

Insert Figure 4 about here

For these four participants, on average, the b-wave implicit time at the 25 cd/m^-2 background was 29.3 ms in the DE and 33.3 ms in the NDE. At 80 cd/m^-2, the implicit time went down to 30.75 ms in the NDE and 27 ms in the DE. The average Vmax amplitude in the NDE remained basically the same at 107,8uV and 108,7uV at the 25cd/m^-2and 80 cd/m^-2 background respectively. Therefore, increasing the background intensity appeared to have normalized the timing of the b-wave in NDE but did not improve the amplitude of the Vmax. However, as seen at Figure 4, which shows waveforms obtained with the two backgrounds of 25 cd/m^-2and 80 cd/m^-2, we can see that the waveform was more comparable between the two eyes when using the 80 cd/m^-2 background in NDE. In fact, whereas in the NDE the impact of oscillatory potentials (OPs) is quite obvious in shaping the waveform when using a photopic background of 25 cd/m^-2, the appearance of OPs is less predominant with the 80 cd/m^-2 background as in the DE.

Scotopic condition

Figure 5 presents the mean luminance response function obtained in the DE and NDE in the group of 23 participants. Of interest a right shift of the luminance response curve along the x-axis is also observed although Vmax amplitude seems to be less compromised than in the photopic ERG.

Insert Figure 5 about here

Amplitude of b-wave and a-wave

Table 3 presents the intensity at which Vmax was observed in the DE vs NDE and implicit time for the a-wave and b-wave for the group of 23 participants who were submitted to the flash intensities provided exclusively by the green LED. At the respective Vmax, b-wave amplitude was on average 7% significantly lower in the NDE (p=0.0475; 167.2µV) when compared to DE (178.7 µV), but there was no difference in a-wave amplitude between the NDE and the DE (p>0.05).

Insert Table 3 about here

Intensity of Vmax

In the DE of the 23 participants, Vmax was observed on average at an intensity of -1.5 ± 0.1 log unit. The corresponding values for the NDE were -1.2 ± 0.1 log unit, which represent a significant increase of 0.30 log unit when compared to the DE (p<0.0001) (See Table 3).

Log K and slope

As expected, log K values were significantly higher for NDE (-2.6 log unit) when compared to DE (-2.2 log unit), which indicates a decrease in retinal sensitivity in the NDE in the order of 0.4 log unit (p<0.0001). The slope of the luminance response function was not significantly different between both eyes.

Implicit time of b-wave and a-wave at respective Vmax

B-wave and a-wave implicit times appeared to be prolonged in NDE compared to the DE for the LED group, but this difference was not significant (P>0.05).

Discussion

The main goal of our study was to evaluate, using the Espion system, the impact of the normally constricting eye on the ERG and to assess what intensity could be used to achieve the same standard in terms of waveform, implicit time and amplitude as per the ERG obtained in the pharmacologically DE. Despite the use of higher intensities, results show that in the NDE we could not reach the same amplitude for the Vmax as in the DE. This difference of 14% is modest, but significant. Also, the b-wave implicit time was prolonged significantly at Vmax for the NDE compared to the DE. The b-wave is supposedly controlled for the most part by the bipolar and Müller cells complex [11]. But in the current study, the origin of the difference in amplitude and implicit time in the NDE could be attributed to the fact that the peripheral cone photoreceptors are less solicited by the light due to the natural constriction of the pupil. However, why increasing the light intensity does not simply correct for this physiological constraint? Using a higher photopic background we could normalize for the implicit time but not for the amplitude. This observation is in accordance with the results of Rufiange et al.[8] in photopic ERG with backgrounds varying between 18 to 25 cd/m^-2 showing no difference in b-wave amplitude. Unfortunately, implicit time was not measured in this study. This suggests however, that the cone system may have a voltage limitation as suggested by these authors.

Another possible explanation derived from our results could be explained by the «push-pull model», a theory first introduced by Sieving [12]. According to this model, both depolarizing and hyperpolarizing second-order neurons participate in generating the waveform of the photopic ERG with ON-depolarizing bipolar cells pushing the b-wave to its peak amplitude, and the OFF-hyperpolarizing bipolar pulling back on the b-wave, therefore limiting its rise. In our study, the fact that we are increasing the light intensity in order to achieve the Vmax in the NDE would yield to the central retina being saturated more rapidly than the peripheral area. Therefore, the central photoreceptors might be more at the end of the Photopic hill where more pulling is occurring whereas the peripheral cones are reaching the Vmax where both pulling and pushing forces are balanced. This could explain why Vmax amplitude is lower and also why the plateau at the end of the photopic hill is higher in the NDE; it is due to the fact that the ON-OFF systems are desynchronized between the central and peripheral retina. In other words, when dilating the eye, the dynamic of the push-pull model is the same across the retina since all photoreceptors are stimulated even at low intensities, whereas in the NDE, the peripheral cells become stimulated later on with increasing intensity.

As expected, log K values were significantly higher for NDE than for DEindicating a lower retinal sensitivity of the NDE. In our study, we could calculate that the loss was about 0.50 log unit, a value that was used to adjust accordingly the photopic background. Also, the change in the slope suggests that the DE reached its Vmax more rapidly than the NDE. These latter differences could be also explained by the fact that less light is entering the eye in the NDE.

In scotopic condition, the only difference was a minor decrease of 7% - albeit significant - in the Vmax b-wave amplitude in the NDE compared to the DE. This small difference could be explained by the fact that the eye dilates naturally in the dark but does not reach the same size as the pharmacologically DE. In fact, we tended to observe that the NDE was on average 25% smaller than the pharmacologically DE which could also explain the loss in retinal sensitivity observed in the latter eye.

As per our second objective for this study, we found that there was a great variation in the intensity at which Vmaxwas reached in photopic condition. If we compare the amplitude observed at Vmax in the DE to the amplitude observed at the low and high range of the ISCEV SF intensities, we can see that 69% of the maximal potential is achieved at 1.5 cd^.s^.m^-2whereas 87% of the maximal potential is achieved at 3.0 cd^.s^.m^-2. So, if clinicians are using only the low range intensity at best they are reaching 69% of the full retinal potential (Vmax). Moreover, depending on the patient's disease [13], it is possible that the luminance response function would be shifted to the right along the x-axis, suggesting a decreased in retina sensitivity, and therefore requiring higher intensities to reach the full potential of the b-wave. In our study, an intensity of about 4.0 cd^.s^.m^-2 was necessary to achieved Vmax in the DE. In the light of these results, the recommendation would be that the SF be increased to a range of 3.0 to 5.0 cd^.s^.m^-2 (where most Vmax are expected to be achieved) instead of the current 1.5 and 3.0 cd^.s^.m^-2. In older people (+65 years old), differences between NDE and DE could even be greater than our sample group of relatively young participants, because their pupil size is also smaller even when pharmacologically dilated. All of the above emphases the need to increase the SF in common clinical practices.

Conclusion

Some companies are actually developing an ERG system that can adjust flash intensities with pupil size. However, our data suggest that this new technology could not compensate for the NDE. So, we can anticipate finding ERG responses with an inferior mean of 14% than with DE, with longer a-wave and b-wave implicit time (if keeping the standard background intensity). However, in research practice, outside the clinical testing of patients with retinal dysfunction, it may not be necessary to dilate the eyes albeit a higher photopic background of +0.50 log unit is used. In fact, the 14% loss of the maximal retinal function potential observed in the NDE is not far from the loss of 13% that we are experiencing using the current ISCEV SF.

References

1. Marmor M, Holder G, Seeliger M, Yamamoto S (2004) Standard for clinical electroretinography (2004 update). Doc Ophthalmol 108:107-114.

2. Kergoat H, Kergoat M-J, Justino L (2001) Age-related changes in the flash electroretinogram and oscillatory potentials in individuals age 75 and older. JAGS 49:1212-1217.

3. Lachapelle P (1990) A possible contribution of the optic nerve to the photopic oscillatory potentials. Clin. Vision Sci. 5:421-426.

4. Hébert M, Lachapelle P, Dumont M (1996) Reproducibility of electroretinograms recorded with DTL electrodes. Doc Ophthalmol 91:333-342.

5. Lachapelle P (1987) Analysis of the photopic electroretinogram recorded before and after dark adaptation. Can J Ophthalmol 22:354-361.

6. Peachey N, Alexander K, Fishman G, Derlacki D (1989) Properties of the human cone system electroretinogram during light adaptation. Appl. Optics 28:1145-1150.

7. Wali N, Leguire L (1992) The photopic hill : A new phenomenon of the light adapted electroretinogram. Doc Ophthalmol 80:335-342.

8. Rufiange M, Rousseau S, Dembinska O, Lachapelle P (2002) Cone-dominated ERG luminance-response function: the Photopic Hill revisited. Doc Ophtalmol. 104:231-248.

9. Rufiange M, Dassa J, Dembinska O, Koenekoop R, Little J, Polomeno R, Dumont M, Chemtob S, Lachapelle P (2003) The photopic ERG luminance-response function (photopic hill_: method of analysis and clinical application. Vision Res 43:1405-1412.

10. Fulton A, Hansen R (1987) The relationship of retinal sensitivity and rhodopsin in human infants. Vision Res 27:697-704.

11. Dowling J (1987): Electroretinogram and glial responses, The retina : an approachable part of the brain, Cambridge, Mass. : Belknap Press of Harvard University Press

12. Sieving P, Murayama K, Naarendorp F (1994) Push-pull model of the primate photopic electroretinogram : A role for hyperpolarizing neurons in shaping the b-wave. Visual Neuroscience 11:519-532.

13. Massof R, Finkelstein D, Perry C, Starr S, Jonhson M (1981) Properties of electroretinographic intensity-response function in retinitis pigmentosa. Doc Ophthalmol 57:279-296.