Photopic ERGs in Patients with Optic Neuropathies: Comparison with Primate ERGs after Pharmacologic Blockade of Inner Retina

Abstract

purpose. To determine whether anterior ischemic optic neuropathy and compressive optic neuropathy in humans alter the photopic flash ERG and to investigate the cellular origins of the waves that are affected by pharmacologic agents in primates.

methods. Photopic flash ERGs were recorded differentially, with DTL electrodes, between the two eyes of 22 patients with diagnosed optic neuropathy (n = 17, anterior ischemic optic neuropathy [AION]; n = 5, compressive optic neuropathy) and 25 age-matched control subjects and in 17 eyes of 13 monkeys (Macaca mulatta). The stimulus consisted of brief (<5 ms) red (λ_max = 660 nm) Ganzfeld flashes (energy range, 0.5–2.0 log td-s) delivered on a rod-saturating blue background of 3.7 log sc td (λ_max = 460 nm). An eye of the patient with ischemic changes at the disc was classified as symptomatic if it showed visual field defects with a mean deviation (MD) of P < 2%. Recordings in macaque monkeys were made before and after inner retinal blockade with tetrodotoxin (TTX) (1.2–2.1 μM; n = 7), TTX+N-methyl-d-aspartate (NMDA; 1.4–6.4 mM; n = 7), and cis-2, 3 piperidine dicarboxylic acid (PDA; 3.3–3.8 mM; n = 3).

results. The PhNR amplitude was significantly reduced in both symptomatic (P = 3.4 × 10⁻⁸) and asymptomatic (P = 0.036) eyes of patients with AION or compressive optic neuropathy (P = 0.0054) compared with control subjects. The PhNR amplitude in the symptomatic eye showed a moderate correlation with field defects (P < 0.05) similar to previous findings in open-angle glaucoma. The a-wave also was reduced significantly in the symptomatic eye (P = 0.0002) of patients with AION. The i-wave, a positive wave on the trailing edge of the b-wave peaking around 50 ms, became more prominent in eyes in which the PhNR was significantly reduced. In monkeys, the PhNR was eliminated by TTX. The a-wave at the peak and later times was reduced by TTX, further reduced by NMDA, and eliminated after PDA in response to the red stimuli. PDA also eliminated the i-wave.

conclusions. PhNR amplitude is significantly reduced in eyes with open-angle glaucoma, AION, and compressive optic neuropathy. Experiments in primates indicate that this reduction reflects loss of a spike-driven contribution to the photopic ERG. There also are small spike-driven contributions to the a-wave elicited by full-field red stimuli. The i-wave, which becomes more prominent when the PhNR is reduced, has origins in the off-pathway distal to the ganglion cells.

The electroretinogram (ERG) is a mass potential recorded at the surface of the eye that represents the summed activity of all retinal cells. The photopic flash ERG typically is recorded in response to brief (<10 ms) or longer-duration (150–200 ms) white flashes on a rod-saturating background. Pharmacologic dissection experiments performed in monkeys, whose retinas are very similar to those of humans, have contributed greatly to our understanding of the origins of the various components of the photopic flash ERG. These studies have shown that the a-wave originates in part from the cone photoreceptor and in part from second-order hyperpolarizing cells (i.e., off-bipolar, or perhaps horizontal cells ¹ ), with possible contributions from inner retinal neurons. ² The b-wave arises from depolarizing (on-) bipolar cells, shaped by second-order hyperpolarizing cells, ^{3 4} and the d-wave, which can be seen in response to the offset of long-duration flashes, has contributions from off-bipolar cells, as well as the offset of on-bipolar cells and cone photoreceptors. ^{3 5}  

Until recently, it was generally believed that the full-field flash ERG does not include measurable contributions from retinal ganglion cells and hence it was considered to be of little use in diagnosis of optic nerve diseases. Instead the pattern ERG (PERG) has been used in evaluation of optic neuropathies (for review see Ref. ⁶ ). The most sensitive component of the transient PERG to the presence of optic neuropathies is a late negative wave called the N95, for the timing of its trough. ^{6 7 8}  

There is now evidence that a component of the photopic ERG called the photopic negative response (PhNR) may arise from the same generator as the N95 wave of the transient PERG. ⁹ The PhNR, a negative-going wave that follows the b-wave and again after the d-wave in response to a long flash is easily seen in response to red flashes on blue backgrounds. Using this stimulus, it can be as large as 40 μV when measured from baseline, which is 5 to 10 times larger than the typical N95 of the PERG. 

The origins of the PhNR were first determined by comparing photopic ERGs in pharmacologic blockade studies in monkeys and cats ^{9 10 11} with ERGs in monkeys with experimental glaucoma. Viswanathan et al. ¹¹ proposed that the PhNR originates from ganglion cells (or their axons) because blockade of the sodium-dependent spiking activity of inner retinal neurons (ganglion and amacrine cells) with tetrodotoxin (TTX) eliminated the PhNR, and experimental glaucoma, which kills ganglion cells, reduced or, depending on its severity, eliminated the response. Subsequent studies in humans with primary open-angle glaucoma (POAG) and a few with suspected ocular hypertension showed that the PhNR amplitude was reduced in these human eyes as well. ^{12 13 14} Both in the studies of monkeys with experimental glaucoma and humans with POAG, the PhNR amplitude was significantly lower than in control eyes, even when visual field defects were mild. 

The studies cited indicate that the PhNR amplitude holds promise in early detection of the optic nerve damage that occurs as a consequence of elevated intraocular pressure (IOP). An important question is whether the reduction in PhNR amplitude, as has been shown for the N95 of the PERG, ⁸ is an indicator of any type of optic neuropathy, or is specific to neuropathies that occur as a consequence of elevated IOP. We know that it is not the elevated pressure per se that produces the reduction in PhNR amplitude on the day of recording. The ERG in most of the patients in the POAG study of Viswanathan et al. ¹³ was medically controlled, and Viswanathan et al. ¹¹ found no correlation between IOP and PhNR amplitude in monkeys with experimental glaucoma. A way to determine the generality of the effects on the PhNR amplitude would be to record photopic ERGs in patients with other diseases of the optic nerve not associated with increased IOP. Gotoh et al. ¹⁵ in a recent study observed that the PhNR in response to a bright-white flash on a white background is reduced in patients with optic nerve atrophy as a result of a variety of optic nerve disorders including inflammatory, compressive, and traumatic optic neuropathy. 

In the present study, we recorded photopic ERGs from patients with anterior ischemic optic neuropathy (AION) or compressive optic neuropathy, by using stimulus parameters similar to those used by Viswanathan et al., ¹³ and found a significant reduction in the PhNR amplitude in eyes with AION and in eyes with profound field loss due to compressive optic neuropathy. Because there also was a small reduction in the a-wave amplitude in patients with AION, we investigated further in monkeys the cellular origins of the response to the stimulus conditions used to elicit the PhNR and discovered significant negative-going contributions from spiking inner retinal neurons late in the leading edge of the a-wave as well as in the PhNR. Some of these results have been reported in abstracts ¹⁶ (Rangaswamy NV, et al. IOVS 2004;45:ARVO E-Abstract 2143). 

Methods

ERG Recordings in Humans

Photopic ERGs were recorded in 17 patients with nonarteritic AION (13 with monocular clinical signs and 4 with binocular signs; mean age, 55.2 ± 11.3 years; range, 34–74), in 25 age-matched control subjects, and in 5 patients with compressive optic neuropathy (age range, 44–69 years). All patients were seen at the University Eye Institute of the University of Houston. In the four patients with AION with clinical signs in both eyes, the more affected eye was selected for analysis (described later). The patients with AION or compressive optic neuropathy underwent a thorough clinical examination at the University Eye Institute that included refraction, pupil and afferent pupillary defect, slit lamp biomicroscopy and ophthalmoscopic examination, and stereoscopic fundus photographs. Visual field measurements were obtained with using a field analyzer (Humphrey Field Analyzer; Carl Zeiss Meditec, Dublin, CA), followed by optic nerve head imaging using confocal scanning laser ophthalmoscopy (Heidelberg Retinal Tomograph [HRT]; Heidelberg Engineering, Heidelberg, Germany) and scanning laser polarimetry (GDx Nerve Fiber Analyzer; Laser Diagnostics, San Diego, CA). Fundus fluorescein angiography was performed in some patients. 

An eye was classified as having AION (or being symptomatic) if it satisfied the following four criteria: (1) signs of an ischemic event to the optic disc with optic disc edema at acute stages and atrophy at later stages; (2) a visual field defect having a mean deviation (MD) with a probability of less than 2%; (3) IOP within normal limits; and (4) no other detectable retinopathies. We did not include any patients with arteritic AION in this study, because those with giant-cell arteritis (temporal arteritis) may have other retinopathies (retinal and/or choroidal ischemia) and may be receiving steroid treatment. We wanted to exclude any conditions such as these that could interfere with our results. 

The visual acuity in eyes with AION ranged from 20/20 to counting fingers at 2 ft and in eyes with a compressive lesion ranged from 20/20 to no light perception, whereas in our age-matched control subjects acuity was 20/20 or better (Table 1) . All procedures were approved by the University of Houston Committee for Protection of Human Subjects and adhered to the Declaration of Helsinki. Informed consent was obtained from all subjects after the procedures were completely explained. 

ERGs were recorded differentially between DTL ¹⁷ fiber electrodes moistened with carboxymethyl cellulose 1% and placed in the lower cul-de-sac of each eye. Each DTL fiber was anchored with a dab of petroleum jelly near the inner canthus and electrically connected by a clip lead at the outer canthus. The ground electrode was an adhesive silver/silver chloride electrocardiograph (EKG) electrode that was placed on the forehead. Pupils were fully dilated to approximately 9 mm with tropicamide (1%) and phenylephrine hydrochloride (2.5%). Responses were averaged over 60 to 80 stimulus presentations. 

To assess test–retest variability we compared results from five normal control subjects. The differences in a- and b-wave and PhNR amplitude (measured at their respective peaks) between the two recordings fell within a range in which 95% (mean ± 1.96 SD) of the differences are expected to fall. ¹⁸  

ERG Recording in Monkeys

Photopic flash ERGs were recorded in 17 eyes of 13 monkeys (Macaca mulatta). All experimental and animal care procedures adhered to the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research and were approved by the Institutional Animal Care Committee of the University of Houston. 

Preparation for Recording.

Animals were anesthetized intramuscularly with ketamine (20–25 mg/kg per hour) and xylazine (0.8–0.9 mg/kg per hour) and were treated with atropine sulfate (0.04 mg/kg, injected subcutaneously). The depth of the anesthesia was maintained at a level sufficient to prevent the animals from blinking or moving. 

Pupils were dilated fully to 8.5 to 9 mm in diameter with topical tropicamide (1%) or atropine (0.5%) and phenylephrine hydrochloride (2.5%). The nonstimulated eye was covered. Heart rate and blood oxygen were monitored with a pulse oximeter (model 44021; Heska Corp., Waukesha, WI) and body temperature maintained between 36.5°C and 38°C with a water-circulating heating pad (Heska Corp.) or in later experiments, a thermostatically controlled blanket (TC1000-Temperature Controller; CWE, Ardmore, PA). 

ERGs were recorded differentially between DTL electrodes that were placed across the center of the cornea and under a corneal contact lens of each eye. A needle inserted under the scalp served as ground electrode. The recording sessions lasted 3 to 8 hours, and over that time, control ERG waveforms for the monkeys did not change systematically, always closely resembling those in control eyes of human subjects. This suggests that cumulative effects of the anesthesia on the ERG were minimal. 

Intravitreal Injections.

Intravitreal injections of 40 to 50 μL were made nasally and temporally in the globe with a sterile 30-gauge needle inserted through the pars plana behind the limbus into the vitreal cavity. Intravitreal concentrations of the pharmacologic agents were estimated by assuming the vitreal volume to be 2.1 mL. The following drugs and concentrations, all in sterile balanced salt solution, were used: TTX (1.2–2.1 μM) to block sodium-dependent spiking activity generated by retinal ganglion cells, amacrine cells, and interplexiform cells. ^{19 20} N-methyl-d-aspartate (NMDA; 1.4–6.4 mM), to suppress light responses of retinal cells that possess functional NMDA receptors (i.e., the amacrine and retinal ganglion cells) cis-2,3 piperidine dicarboxylic acid (PDA: 3.3–3.8 mM) to block ionotropic glutamatergic transmission to cells that contain α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid/kainic acid (AMPA/KA) receptors (i.e., off bipolar, horizontal, amacrine and ganglion cells), ²¹ as well as to act as an NMDA receptor agonist (Sigma-Aldrich, St. Louis, MO). The concentrations that we chose were sufficient in other experiments on monkeys to separate pharmacologically the components of full-field flash ERG. ^{3 11 23} Recordings were made before and at least 1 hour after injections when effects had stabilized. 

Although the anesthesia that we used, ketamine, is a noncompetitive NMDA receptor antagonist, in the present and previous studies NMDA injections were effective in altering inner retina components of the macaque ERG that were not affected by a prior injection of TTX. ²² We cannot say whether the changes produced by NMDA would have been greater had the ketamine not been used. 

Visual Stimulation

Full-field stimulation was produced with a Ganzfeld stimulator (ESPION; Diagnosys Inc., Littleton, MA) positioned very close to one eye. Human subjects maintained fixation on a red dot in the center of the Ganzfeld. The nontested eye was covered. Stimuli were red LED flashes (λ_max = 660 nm) of brief duration (<5 ms) on a rod-saturating blue LED background (λ_max= 460 nm). Interstimulus intervals were of adequate duration to avoid adaptive effects. Scotopic trolands (scot td) for the blue background (3.7 log scot td), and troland seconds (td-s) for the brief flashes (0.5–2.0 log td-s) were calculated for a pupil diameter of 9 mm without correction for the Stiles-Crawford effect. These stimuli were selected because they were similar to those used in the previous study by Viswanathan et al. ¹³ on patients with POAG and monkeys. ¹¹ In some monkeys, an LED stimulus from Viswanathan et al. ^{11 13} was used (red stimulus λ_max = 630 nm; blue background λ_max = 450 nm). 

Statistical Analysis

The statistical analyses in this study were performed with the independent t-test to compare ERG amplitudes between the control subjects and the symptomatic eyes and between the control subjects and the asymptomatic eyes of the patients (Table 2 ; see Fig. 3 ). For comparison of the amplitudes before and after pharmacologic agents were injected individually into monkey eyes, we used the paired t-test (Table 3) . 

We understand that while performing multiple comparisons (between control subjects and the two eyes of the each patient, e.g., Table 2 ) to avoid a type I error, a correction (e.g., Bonferroni) would be necessary if AION was known to be a bilateral condition. However, our interest in this study was to evaluate independently the condition of the PhNR in the asymptomatic eyes versus control eyes; therefore, we did not use any correction. 

Results

Figure 1 shows the photopic flash ERGs in response to brief flashes obtained from age matched control subjects (left) and the eyes of two different patients with AION in whom one eye was symptomatic according to the four criteria. The ERGs were recorded in response to the four standard stimulus flash energies that were used for the human ERGs described in this study. The typical ERG responses of the age-matched control subjects to these stimuli were composed of three major waves: an initial negative-going a-wave, a positive-going b-wave, and a slow negative-going potential, the PhNR, after the b-wave. With increasing stimulus energy from the bottom up to the third flash of 1.7 log td-s, there was an increase in the amplitude of all three potentials. In response to the highest flash energy (2.0 log td-s; and higher energies that were tested in a few cases but are not shown), the b-wave and PhNR amplitudes had saturated, and responses generally were not very different for the two strongest stimuli (1.7 and 2.0 log td-s). For the quantitative analyses, responses to the 1.7 log td-s flash were selected. 

Photopic Flash ERG in AION

The middle and righthand columns of Figure 1A show the photopic ERGs obtained from the eyes, symptomatic and asymptomatic, of a 53-year-old man (FL). This patient had an inferior altitudinal defect in his symptomatic eye with a mean deviation (MD) of −8 dB and an MD of 0.4 dB in his asymptomatic eye. It is evident that the PhNR amplitude was lower in Figure 1A for the symptomatic eye than for the age-matched control. The a-wave amplitude is also slightly lower. The righthand column shows that the PhNR amplitude of the asymptomatic eye also was lower than in the age-matched control eyes. 

In Figure 1B , the middle and righthand columns show the ERGs of a 34-year-old man with AION (SB). This patient had an inferior altitudinal defect in his symptomatic eye (MD = −23.4 dB), whereas the optic disc in the asymptomatic eye did not show signs of an ischemic episode. Again PhNR amplitude was lower in both the symptomatic eye and the asymptomatic eye than in the control eye, and the a-wave amplitude in the symptomatic eye was slightly lower than in the control eye. 

The ERGs shown in Figures 1A and 1B were selected to illustrate typical effects of AION in both older (1A) and younger (1B) eyes, and also to point out differences that occurred in a small positive wave, called the i-wave, that appeared late on the trailing edge of the b-wave (peak time ∼50 ms). The i-wave was present to various extents in the control subjects’ ERGs, but almost always was more prominent in the patients’ recordings. An increased prominence of the i-wave was observed previously in patients with POAG ¹³ or optic atrophy ¹⁵ and after intravitreal injection of TTX in monkeys. ^{11 23} Origins of this wave are addressed in a later section where results of using pharmacologic blockade in monkey eyes are reported. 

To compare quantitatively the PhNR and a- and b- wave amplitudes (responses to stimulus intensity of 1.7 log td-s) between age-matched control subjects and patients, amplitudes of these components were measured at fixed times after the stimulus. We measured the PhNR amplitude from the baseline rather than from the b-wave peak, ²³ because in previous work using the same stimulus conditions as were used in the present study, the PhNR amplitude was no more reliable when the amplitude was measured from the b-wave peak. ¹³  

To determine the average peak time of the PhNR, we measured its amplitude in 5-ms steps from 55 to 85 ms after the stimulus in the control subjects, and found that it was largest, on average 70 ms, but not very different from amplitudes measured at 65 ms (data not shown). Table 2A shows the peak amplitudes of the PhNR, a- and b-wave in control eyes and in both eyes of patients with AION. At 70 ms (and 65 ms), the PhNR amplitude was significantly lower in both the symptomatic (P < 0.01) and asymptomatic (P < 0.05) eyes of the patients, compared with the age-matched control eyes. The a-wave amplitude measured at its peak time determined in control subjects at 15 ms after the flash, also was significantly lower in the symptomatic eye (P < 0.01). The small average amplitude reduction in the asymptomatic eye was not statistically significant. The b-wave amplitude measured in control subjects from the a-wave trough (15 ms) to the average peak time of the b-wave of 31 ms was slightly larger on average in both eyes, than in the age-matched control eyes, but these differences were not statistically significant. 

Figure 2 shows a plot of PhNR (absolute) amplitude (measured for a stimulus intensity of 1.7 log td-s at 70 ms) as a function of age in control subjects (open circles) and in the symptomatic (filled circles) and asymptomatic (filled squares) eyes of patients. Results are plotted versus age, because ERG amplitudes, including PhNR amplitudes are known to decline with age. ^{13 24 25} In the present study, control PhNR amplitude also was found to correlate significantly with age. There was little overlap in the PhNR amplitude between the control eyes and symptomatic patient eyes; and, as noted, the PhNR amplitude of the symptomatic eye was significantly lower than the control amplitude (P < 0.01). The PhNR amplitude of the asymptomatic eye was also significantly lower than in control eyes and the amplitudes for this eye fell between those in the control and symptomatic eyes. In most symptomatic eyes, the PhNR amplitude was so small that it was difficult to measure its implicit time. Therefore, we did not consider differences in PhNR implicit time between patients and control subjects. There was no clear relationship between the amplitude and implicit times of either the a- or the b-wave with age in the control subjects or the patients in this study (data not shown). 

Figure 3 shows plots of the average ERG amplitude at fixed times after the flash between 7 and 31 ms in symptomatic eyes (filled symbols in A) vs. control eyes (open symbols), and for asymptomatic versus control eyes (Fig. 3B) . Fig. 3A shows that in the symptomatic eyes, the a-wave amplitude was not significantly less than that of control subjects at times earlier than 15 ms, but the response was significantly less (i.e., less negative than) the age-matched control eyes at 15 ms as well and at all later times up to 29 ms, along the leading edge of the b-wave (Fig. 3A , **P < 0.01). Thus, although the positive amplitude was not significantly greater for the b-wave at the peak time of 31 ms, it was significantly greater at earlier times. This larger b-wave was probably due to the loss of a negative potential in the ERG. This possibility was investigated in monkey eyes, and the results will be described in a later section. In the asymptomatic eye, the mean amplitude was significantly less than that of the a-wave at 17 ms and was significantly greater than the control amplitude at 27 and 29 ms (Fig. 3B , *P < 0.05). 

Viswanathan et al. ¹³ did not observe a significant effect of POAG on either the a- and b-wave amplitude or implicit time. However, they measured the a- and b-wave amplitudes at their respective implicit times rather than at a fixed time. To compare the present results with those of the POAG study, we also measured implicit times of the a- and b-waves and measured amplitudes at those times (Table 2B) . As predicted by plots in Figure 3A , the a-wave implicit time was significantly earlier in the symptomatic eye (13.9 ± 0.3 ms) of patients with AION than in the age-matched control subjects (15.0 ± 0.3 ms; P < 0.05), but the amplitude was not significantly different from that in control eyes. The b-wave implicit time in the symptomatic eyes was earlier than in the control eyes, but the time difference was not significant. The b-wave amplitude when measured at the slightly earlier time was significantly greater in the symptomatic eye (81.4 ± 5.3 μV) than in the control eye (66.6 ± 4.1 μV; P < 0.05), as would be expected from looking at the plot in Figure 3 . In the asymptomatic eye, the a- and b-wave implicit times and amplitudes were not statistically different from those in age-matched control eyes. 

Severity of AION and PhNR

In this study, we used the presence of significant visual field defects (P < 2%) as one of the criteria for classifying an eye as symptomatic. Figure 4 shows a- and b-wave and PhNR amplitudes (measured at fixed peak times as in Table 2A ) plotted versus MD in symptomatic and asymptomatic eyes. The plots show that as the visual field sensitivity declined, the amplitudes of the PhNR and a-wave decreased. The correlation between the amplitude of the potentials and MD was significant for PhNR (r = 0.56; P < 0.05) and the a-wave (r = 0.5; P < 0.05), but not for the b-wave (r = 0.4; P = 0.13). At times after the peak of the a-wave, when the amplitude was significantly lower in the symptomatic eye, the values were not significantly correlated with the MD. The mean and 2SD of control amplitudes are represented by solid and dashed horizontal lines, respectively, in these plots. 

Sensitivity and Specificity of PhNR in AION

Receiver operating characteristic (ROC) curves were used to evaluate the effectiveness of the PhNR amplitude in distinguishing between normal and AION eyes. Figure 5 shows the ROC curves for the PhNR and a-wave amplitudes. b-Wave amplitude was not included because it was not significantly different between AION and control eyes. These curves were generated by plotting sensitivity versus 1 − specificity calculated for different criterion values. Sensitivity shows how well the PhNR amplitude can perform as a test for detecting AION and specificity shows how well the PhNR amplitude can identify those subjects who do not have AION. The PhNR amplitudes for the 1.7 log td-s stimulus intensity were pooled together from the control and AION symptomatic eyes, and criterion values were decremented in 0.5-μV units for the whole range of amplitudes. Figure 5 also shows the 1:1 line; curves farther from the 1:1 line have larger area under the curve (AUC) and perform well in distinguishing between normal subjects and patients. The AUC was 0.96 for PhNR amplitude and was 0.80 for the a-wave. The optimal cutoff amplitude of the PhNR determined by the criterion value associated with the lowest general error rate (GER: total percentage of false positives and false negatives) was 15.0 μV. Our sample in this study consisted of patients and control subjects from a wide range of age groups (35–75 years). To determine the effect of age on the AUC of PhNR, we divided our data into two age groups (35–50 years, and 50–75 years). The AUC for the PhNR amplitude was unaffected when we compared control subjects and patients aged between 35 and 50 years, in whom the relationship between PhNR amplitude and age was weak (Fig. 2) . However, the AUC decreased to 0.91 when we included patients and control subjects aged between 50 and 75 years, in whom the relationship with age was stronger. 

PhNR in Compressive Optic Neuropathy

We further investigated whether the reduction in PhNR amplitude is present in another type of optic neuropathy by recording photopic ERGs from five patients with a diagnosis of optic neuropathy due to compressive lesions. PhNR and a- and b-wave amplitude from these patients were plotted as a function of MD (Fig. 4 , filled diamonds to the right in each plot), along with the AION data. The figure shows that the PhNR amplitude for four of these five patients, who had poor visual performance (no visual field could be measured) are well below the control amplitudes and the PhNR amplitudes are as affected as the severely affected eyes of patients with AION. Data from one patient (AB: shown with an arrow) who had a compressive lesion overlaps with the control amplitude. This patient had undergone treatment (before the ERG recording session) that shrunk the tumor and the patient had a normal visual field (MD = −0.7 dB). 

Removal of Inner Retinal Responses in AION Patients Compared with Inner Retinal Pharmacologic Blockade in Primates

In the previous sections, we have examined the effects of AION on specific time points in the ERG. To assess the effect of AION on the entire waveform of the ERG, Figure 6A shows the average ERG in the symptomatic eyes (dashed line) and the age-matched control subjects (solid line). The dotted line in the lower part of the plot shows the difference between average symptomatic AION eyes and control eyes, thereby showing the waveform of the ERG that was eliminated by AION. As described earlier, in addition to the PhNR amplitude being smaller than that in age-matched control subjects, the earlier waveform starting at the a-wave trough also was lower than in control eyes Hence, there is a negative wave at earlier times in the difference record. 

Figure 6B shows the same average a-wave on an expanded time scale as in Figure 6A . The elevation of the waveform at the a-wave trough and later times (15 ms and later) is clearly visible in this figure as is the difference between symptomatic eye and age-matched control. 

Previous studies of PhNR in monkeys have indicated that it originates from spiking activity of ganglion cell axons (perhaps mediated by K⁺ currents in glial cells ^{11 23} ). To determine whether the earlier negative potential at the peak of the a-wave and later is spike related in monkeys, we recorded the photopic ERG under similar or identical stimulus conditions, before and after TTX to block sodium-dependent spiking activity. Figures 6C and 6D shows a representative waveform from one monkey eye, before and after TTX and the difference record for the whole ERG (Fig. 6C) and the a-wave (Fig. 6D) . The difference record (Fig. 6C , dotted line) confirms that TTX removes the PhNR and it also removes a negative potential at earlier times. The change in PhNR was significant after TTX in the seven animals studied (Table 3) . Similar to the findings in optic neuropathy, TTX also reduced the negative a-wave amplitude, and tended to elevate the b-wave peak. Although injection of TTX alone did not significantly reduce the a-wave, addition of NMDA resulted in a significant reduction (P < 0.05) and PDA eliminated it. The b-wave amplitude was not significantly altered by any pharmacologic treatment. 

As noted earlier, the i-wave became prominent in optic neuropathy, and Figure 6 shows that this also occurred after pharmacologic blockade of the inner retina in monkeys which suggested that the wave is generated by cells distal to retinal ganglion cells. The i-wave was eliminated with addition of PDA to block ionotropic glutamatergic transmission after inner retinal blockade (Fig. 6D , inset), strongly suggesting that the i-wave originates from the off-pathway. 

Figure 6 shows that the i-wave was more prominent in human control ERG than in monkey control ERG; and, in these human subjects but not in the monkeys, we identified a trough and a peak for the wave. We compared the peak-to-trough amplitude in control subjects and patients and found the difference was not statistically significant (P = 0.07), which suggests that optic neuropathy hardly affects the i-wave (peak-to-trough) amplitude. It mainly raised the whole ERG waveform above the baseline. In support of this we found that the amplitude at a fixed time (50 ms, corresponding to the peak of the i-wave) was significantly greater (more positive) in both humans (optic neuropathy; P < 0.01) and monkeys after pharmacologic inner retinal blockade (P < 0.01) than in control subjects (Table 4) . 

Discussion

PhNR in Optic Neuropathy

This study shows that the PhNR amplitude is also sensitive to optic nerve damage due to causes other than elevated IOP. The two optic neuropathies examined in this study were nonarteritic AION, in which the damage is due to ischemic injury to the optic nerve, and compressive optic neuropathy, in which the insult to the optic nerve is due to compression by a tumor. None of the patients examined in the present study had elevated IOP, suggesting that the reduction in PhNR amplitude compared with control eyes in these patients was primarily due to optic nerve damage. Gotoh et al. ¹⁵ found that PhNR was also reduced in optic neuropathies due to trauma and inflammation, confirming that the PhNR amplitude is sensitive to optic nerve damage of different causes, similar to the N95 component of the PERG. 

PhNR in the Asymptomatic Eye

An interesting outcome in this study was that the PhNR amplitude was lower in both the asymptomatic and symptomatic eyes of patients with AION than in control eyes. It has been reported that there is nearly a 6% chance of having another AION episode in the same eye ²⁶ that has had a previous episode and a 25% chance of development of AION in the fellow eye. ²⁷ The asymptomatic eye of the 13 patients examined in this study had normal visual fields and did not present with any optic disc swelling or atrophy. However, the presence of a small cup or no cup in the optic disc in the asymptomatic eye in 8 of the 13 patients was observed. A small cup has been noted as one of the ocular risk factors that makes an eye prone to an ischemic attack. ²⁸ The reduced PhNR amplitude that we found in the asymptomatic eye of the patients with AION indicates that PhNR amplitude is sensitive to disease in the optic nerve, even before most standard clinical signs appear. 

PhNR and Optic Nerve Head Measurement

There is a growing body of literature on optic nerve head topography in patients with glaucoma, reporting studies in which a variety of techniques have been used to determine which parameters are most sensitive to early glaucomatous damage. Two recent studies of the PhNR in patients with POAG ^{12 13} showed a correlation between the PhNR amplitude and at least one measure of the cup-to-disc ratio. This suggests that future work correlating PhNR changes with sensitive parameters could be useful. Currently, there is only limited literature available on optic disc topography in AION. Gotoh et al. ¹⁵ recently found a good correlation between the PhNR amplitude and retinal nerve fiber layer (NFL) thickness measured with optical coherence tomography (OCT) in patients with optic atrophy of several causes. In the present study, our results with the GDx (Laser Diagnostics) and HRT (Heidelberg Engineering) were mainly negative, except for a moderate correlation of the PhNR amplitude with the height variation contour (mean thickness of the retina around the optic disc), but these analyses are currently tuned for use in patients with glaucoma. Future studies comparing ERG measurements with optic nerve head parameters should use appropriate approaches such as OCT ¹⁵ or other techniques/parameters that detect changes occurring in AION. 

Pharmacologic Blockade and the Photopic Full-Field ERG

Apart from the reduction in the PhNR amplitude in the symptomatic eye of the patients with AION compared with control subjects, there also was a reduction in the a-wave amplitude, an elevation in the leading edge of the b-wave, and an increased prominence of the i-wave. 

Viswanathan et al. ^{11 13} did not report changes in the a- and b-wave amplitude and implicit time in monkeys with experimental glaucoma and patients with POAG. These studies did not examine the leading edge of the b-wave to see whether it was elevated, as in our Figure 3 . However, it may have occurred in some subjects (e.g., see Fig. 3 in Ref. ¹³ ). At present, given the small samples, it probably is not appropriate to conclude that the a-wave is altered in AION and not in glaucoma. Caution is necessary in interpretation of b-wave results showing elevation of the leading edge and even peak amplitude in asymptomatic patients. Figure 4 , middle, shows that in some patients with AION and compressive optic neuropathy, the b-wave amplitude is actually lower than in control eyes. These findings suggest that during the early course of the condition, the b-wave amplitude is elevated because of the removal of a negative potential, but at later stages the b-wave amplitude itself may be reduced. The changes in the a-wave in patients with AION and in monkeys after TTX suggest that the a-wave of the photopic ERG produced by low to moderate intensities has very little contribution from the photoreceptors and is mostly postreceptoral. 

The i-wave became prominent in monkeys after intravitreal injection of TTX or any agent that blocked inner retinal activity in the present study, confirming previous results of TTX injection, ^{11 23} TTX/NMDA injection, ²² or laser-induced experimental glaucoma in monkey eyes. ¹¹ The negative potential removed by pharmacologic blockade of the inner retina around the time of the i-wave (50 ms) was larger than that removed at the peak of the a-wave. The removal of these negative potentials resulted in the whole waveform of the photopic ERG’s being elevated above the baseline. This suggests that the i-wave does not have an origin in the inner retina. In contrast, intravitreal injection of PDA after inner retinal blockade to block the off-bipolar cells and horizontal cells ¹ eliminated the i-wave, suggesting that the i-wave has an origin distal to the retinal ganglion cells, probably in the off-pathway, as previously suggested, ^{29 30 31} although the possibility of a horizontal cell contribution cannot be excluded. 

Stimulus Conditions Used in ERG Recordings

In this study, we used a red flash on a rod-saturating blue background to ensure photopic conditions similar to that used by Viswanathan et al. ^{11 13} Other investigators have used other stimuli to elicit PhNR and also have observed that the PhNR is reduced in patients with POAG ^{12 14} and patients with optic atrophy. ¹⁵ Colotto et al. ¹² used a 12° × 12° field of 2-Hz modulation on a computer monitor (mean luminance of 78 cd/m²), whereas Drasdo et al. ¹⁴ used a silent substitution technique to isolate the S-cone ERG by suppressing the L-cone response with a 650-nm light adaptation and silencing the M-cone by substituting between 450 and 535 nm. The PhNR amplitude recorded in normal subjects by these studies was smaller (∼2 μV in Colotto et al. ¹² and <10 μV in Drasdo et al. ¹⁴ ) than that recorded by Viswanathan et al. ^{11 13} and the present study of greater than 25 μV, on average. Monochromatic full-field stimuli may produce a larger PhNR than a broadband stimulus because the monochromatic stimuli produce less inhibitory center-surround interactions in the spectrally opponent retinal ganglion cells than the broadband stimulus. When Gotoh et al. ¹⁵ recorded photopic ERGs with a bright-white flash on a white background, the PhNR amplitudes were similar to those recorded in the present study (although the i-wave in control subjects was even more prominent). This suggests that broadband stimuli of much higher intensities are necessary to obtain PhNR amplitudes similar to those recorded by Viswanathan et al. ^{11 13} and us. 

In conclusion, the findings in this study in optic neuropathy and pharmacologic inner retinal blockade suggest that ganglion cell contribution to the photopic ERG for a red flash on a blue background is present as early as at the peak of the a-wave (15 ms). The profound effect of AION, compressive optic neuropathy, and POAG on the PhNR amplitude and the ease with which this response can be recorded suggests that measurement of the PhNR can be of great clinical value in the assessment of these optic neuropathies. 

 Supported by National Eye Institute Grants R01-EY06671 and P30-EY07751.

 Submitted for publication April 20, 2004; revised June 25, 2004; accepted June 29, 2004.

 Disclosure: N.V. Rangaswamy, None; L.J. Frishman; None; E.U. Dorotheo, None; J.S. Schiffman, None; H.M. Bahrani, None; R.A. Tang, None

 The publication costs of this article were defrayed in part by page charge payment. This article must therefore be marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

 

Table 1.

 

Patient Details

No.	Subject	Age/Sex	VA			MD dB
						OD	OS	OD	OS
AION
1	FL	54/M	20/30+	20/25+	−8.0	0.4
2	RS	55/M	20/20−	20/20−	−11.0	−2.2
3	RJ	74/M	20/30	20/30	−15.3	−21.0
4	CA	50/F	20/20	20/20	−7.2	0.8
5	JM	62/F	20/25−	20/20	−13.1	−1.0
6	SB	34/M	20/20−	20/20	−1.9	−23.4
7	JG	70/M	20/20+	20/20+	−11.6	−3.6
8	CM	66/F	20/25+	20/60−	−11.0	−2.7
9	PR	55/F	20/15−	20/15−	−7.1	−3.5
10	RM	55/F	20/20	20/20	−0.1	−9.3
11	MG	34/M	20/20	20/20	−0.3	−7.2
12	CG	58/M	20/25+	20/20	−7.0	2.6
13	RH	62/M	CF	20/20	−27.4	1.1
14	WD	59/F	20/20	20/30−	−0.7	−7.7
15	LR	44/M	20/15	20/40	−0.6	−10.9
16	DP	44/F	20/30	20/20	−1.4	−25.9
17	JT	63/M	20/50+	20/20	−2.3	−12.0
Compressive Optic Neuropathy
1	AB	57/F	20/20	20/20	0.1	−0.7
2	PP	44/F	LP	20/30	Not recorded	VF with stim V
3	KP	51/F	NLP	20/15−	Not recorded	−2.1
4	SW	50/F	NLP	20/20	Not recorded	−1.4
5	IP	69/F	20/20−	HM	−0.2	Not recorded

Table 2.

 

ERG Parameters in Patients with AION

	Control	Symptomatic Eye	Asymptomatic Eye
A. Peak Amplitude
PhNR @ 70 ms	−26.7 ± 2.0	−7.5 ± 1.8	−20.9 ± 2.5
		(P = 3.44 × 10−8), †	(P = 0.036)*
95% CI	(−30.8, −22.6)	(−11.4, −3.7)	(−26.2, −15.3)
a-Wave @ 15 ms	−15.2 ± 0.9	−10.5 ± 0.7	−13.0 ± 1.4
		(P = 0.0002), †	(P = 0.13)
95% CI	(−17.0, −13.3)	(−11.9, −8.6)	(−15.9, −10.0)
b-Wave @ 31 ms	59.4 ± 3.6	64.4 ± 4.6	64.9 ± 5.0
		(P = 0.36)	(P = 0.69)
95% CI	(52.1, 66.8)	(54.3, 74.0)	(54.0, 75.8)
B. Amplitude and Implicit Time of a- and b-Wave
a-Wave	15.0 ± 0.3	13.9 ± 0.3	14.5 ± 0.2
Implicit time (ms)		(P = 0.008), †	(P = 0.74)
b-Wave	32.1 ± 0.3	31.1 ± 0.5	30.7 ± 0.4
Implicit time (ms)		(P = 0.09)	(P = 0.07)
a-Wave	17.6 ± 1.0	17.2 ± 1.1	18.2 ± 1.6
Amplitude (μV)		(P = 0.8)	(P = 0.9)
b-Wave	66.6 ± 4.1	81.4 ± 5.3	76.1 ± 6.7
Amplitude (μV)		(P = 0.03)*	(P = 0.5)

 Data are expressed as the mean ± SD. CI, confidence interval.

*   P < 0.05.

†   P < 0.01.

Table 3.

 

ERG Peak Amplitude after Pharmacologic Blockade

	TTX (n = 7)			TTX+NMDA (n = 7)			Inner Retinal Blockade+PDA (n = 3)
		Before	After	Before	After	Before	After
PhNR (70 ms)	−12.9 ± 3.7	5.7 ± 4.3	−20.4 ± 4.7	10.6 ± 2.3	−14.0 ± 9.2	0.5 ± 7.5
		(P = 0.004), †		(P = 0.002), †		(P = 0.04)*
a-Wave (15 ms)	−11.0 ± 1.7	−9.7 ± 2.1	−11.8 ± 1.0	−3.2 ± 2.3	−10.4 ± 1.3	0.9 ± 0.5
		(P = 0.24)		(P = 0.006), †		(P = 0.02)*
b-Wave (31 ms)	66.6 ± 5.9	68.5 ± 4.7	57.2 ± 5.1	52.2 ± 6.1	48.6 ± 5.6	56.2 ± 8.8
		(P = 0.64)		(P = 0.32)		(P = 0.15)

 Data are expressed as the mean ± SD.

*   P < 0.05.

†   P < 0.01.

Figure 1.

 

Photopic ERGs in response to brief red flashes on a blue rod-saturating background in two patients: 53-year-old patient FL (A) and 34-year-old patient SB (B) for the four stimulus energies used in the study.

Figure 2.

 

PhNR amplitude measured for a flash intensity of 1.7 log td-s from the baseline plotted as a function of age in control eyes (○) and the symptomatic (•) and asymptomatic (▪) eyes of patients with AION. Shown are linear regressions for the PhNR amplitude in the symptomatic (solid line) and asymptomatic (dotted line) AION eyes and control eyes (dashed line).

Figure 3.

 

Amplitude measurements for a flash intensity of 1.7 log td-s made at fixed times from the baseline from 7 to 31 ms in symptomatic (A) and asymptomatic (B) eyes, along with the age-matched control eyes. Data are shown as mean amplitudes in (○) control and (•) symptomatic (A) and asymptomatic (B) eyes. Error bars, ±SEM. *P < 0.05; **P < 0.01, by t-test.

Figure 4.

 

Amplitudes for a flash intensity of 1.7 log td-s as a function of MD obtained from patients’ visual fields. The a-wave (top) and the PhNR (bottom) amplitude were measured from the baseline, and the b-wave amplitude (middle) was measured from the trough of the a-wave. Filled symbols represent the symptomatic eyes and symbols with crosses the asymptomatic eyes of patients with AION. Filled diamonds: the five patients with compressive optic neuropathy. Arrow, bottom: PhNR amplitude in patient AB who had normal visual fields after treatment of a tumor.

Figure 5.

 

ROC curves for the PhNR and a-wave in response to a flash intensity of 1.7 log td-s. The AUC was 0.96 for the PhNR, 0.80 for the a-wave, and 0.58 for the b-wave (data not shown).

Figure 6.

 

Comparison of photopic ERG in optic neuropathy with that in monkeys after pharmacologic blockade of inner retinal activity. (A) Average photopic ERG from control eyes from symptomatic eyes, with the average difference shown between control and symptomatic ERGs. (B) The same waveforms as in (A) are represented with a shorter time scale, to illustrate the effect of AION on the a-wave. The difference trace illustrates the shape of the waveform eliminated by AION. (C) Photopic ERG obtained from one eye of a monkey before and after intravitreal injection of TTX, with the difference trace between control and after TTX. (D) The a-wave in the same eye as in (C) before and after intravitreal injection of different pharmacologic agents to block inner retinal activity at stimulus intensity of 1.7 log td-s. Inset: whole photopic ERG from one eye of a monkey before and after intravitreal injection of IR+PDA for flash intensity of 1.7 log td-s. Arrow: timing of the i-wave. In this figure, the human ERG traces are averages of results in control subjects and patients with AION, whereas the monkey ERGs are from an individual monkey, in which paired comparison in the same eye was shown before and after drug injection. Averaging was not appropriate because there were fewer monkeys, and the stimulus used in the monkeys was not always exactly the same one.

Table 4.

 

i-Wave (time slice at 50 ms)

	Control	Symptomatic/after Pharmacologic Agents
AION	−7.5 ± 1.2	7.5 ± 1.8
		(P = 2.7 × 10−6)*
TTX	−3.8 ± 2.5	15.4 ± 2.6
		(P = 0.0012)*
TTX+ NMDA	−8.8 ± 4.5	15.0 ± 4.4
		(P = 0.0012)*
IR blockade+ PDA	−6.9 ± 11.6	18.6 ± 7.3
		(P = 0.12)

 Data are expressed as the mean ± SD.

*   P < 0.01.