Impact of Aging and Age-Related Maculopathy on Activation of the a-Wave of the Rod-Mediated Electroretinogram

Abstract

purpose. To examine the impact of aging and age-related maculopathy (ARM) on the activation of phototransduction in rod photoreceptors by measuring the a-wave of the flash, full-field electroretinogram (ERG).

methods. Enrollees consisted of older adults (≥60 years of age) in normal retinal health (n = 41) and those with early (n = 39) or late ARM (n = 7), in whom disease presence and severity were defined based on grading of stereoscopic color fundus photographs according to the Wisconsin Age-Related Maculopathy grading system. Young adults (ages 16–30 years; n = 27) were enrolled for comparison purposes. Previously established procedures were used to estimate the ERG response to two families of flash intensities. By computer subtraction of responses, the isolated rod response was identified. Each participant’s ensemble rod responses were fit with the following equation to describe the response (R) as function of flash intensity (I), and time (t): R(I,t) = [1 − exp{−I · S ·(t − t _d)²}] · Rm _P3, where S is sensitivity, t _d is the delay before onset of the a-wave, and Rm _P3 is the maximum amplitude.

results. In analyses of older adults, there was no impact of early ARM presence or severity on log S, Rm _P3, or t _d after adjustment for age and intraocular lens presence. Differences between young and old normal subjects in log S, Rm _P3, and t _d disappeared when analyses were limited to older adults with intraocular lenses.

conclusions. When the light absorption of the aged lens is taken into account and reliable definitions of normal retinal aging and ARM are used, the activation of the a-wave as measured by the rod-mediated full-field ERG is not affected by early ARM, nor is it impacted by normal retinal aging.

Age-related maculopathy (ARM) is the leading cause of new, irreversible vision impairment in older adults in the United States and other countries. ¹ It is a heterogeneous disorder affecting primarily the retinal pigment epithelium (RPE), Bruch’s membrane, and choriocapillaris, and secondarily, the photoreceptors. ^{2 3} Early ARM can be clinically characterized by large drusen and/or pigmentary changes in the macula. Late ARM is characterized by the presence of geographic atrophy and/or choroidal neovascularization. ⁴ Treatment options for ARM are limited and are either directed at the late phases of disease when vision loss is already severe ⁵ or consist of antioxidative nutritional supplements that slow progression to advanced disease for a subset of patients with early disease of intermediate severity. ⁶ To foster the development of preventative measures and more effective treatments, an improved understanding of ARM pathogenesis has become a major public health mission. ^{1 7} One strategy for providing insight into ARM pathogenesis is to focus research on photoreceptor dysfunction because it provides insight into the pathways and mechanisms affected early in ARM pathogenesis before anatomic changes are visible in the fundus. ⁸  

In studies in which rod and cone vision are examined in the same ARM patients, rod-mediated visual sensitivity loss is typically more severe than is cone-mediated sensitivity loss. ^{9 10 11} These results are consistent with findings in histopathological studies on human donor retinas that indicate a predilection for parafoveal loss of rods over cones in early as well as late ARM. ^{12 13} The topography of rod loss and of rod-mediated sensitivity loss are similar, with the greatest loss occurring near the fovea and decreasing as a function of eccentricity. Both ARM-associated scotopic sensitivity loss and slowed dark adaptation extend past the anatomically, clinically, and epidemiologically defined macula, ^{14 15 16} which may indicate that although pathology associated with ARM is greatest in the macula, it is not confined to the macula. 

Rod-mediated sensitivity loss in patients with early ARM cannot be wholly explained by increased lens density, ¹⁷ decreased pupil size, ¹⁸ and slowed dark adaptation. ¹⁹ Rods are exposed to a variety of mechanical ^{20 21} and metabolic insults ^{9 22} that may alter rod function. A candidate mechanism that may underlie rod-mediated sensitivity impairment in ARM is a deficit in the activation of rod phototransduction. Cideciyan and Jacobson ²³ suggest that an accumulation of cholesterol in the rod outer segment may slow the rate of phosphodiesterase activation. Alternatively, rod loss, convolution, deflection, and shortening of the outer segment, and abnormalities of the synaptic terminals of rod outer segments, which are found in aged and some ARM donor eyes, may reduce the maximum amplitude of the rod a-wave. ^{12 20 21}  

To examine the kinetics of phototransduction in the living person, the electroretinogram (ERG) is measured in response to a family of moderately bright light intensities. ²⁴ The a-wave of the rod ERG originates from the activity of the rod photoreceptors, ^{25 26 27 28} representing the rising phase of the rod photocurrent. ^{24 29 30} The behavior of the leading edge of the rod a-wave is similar to that of isolated rod photoreceptors in response to changes in intensity, wavelength, and adaptational state. ^{30 31 32 33} Because of these similarities, computational models describing the activation steps of phototransduction in isolated rod photoreceptors ³⁰ have been applied to the leading edge of the rod a-wave measured in the living person. ³⁴ These models are useful because, when used with care, parameters of rod a-wave activation measured in living humans can be associated with the activation of rod phototransduction. Relevant parameters include the maximum amplitude of the rod response (Rm _P3), a sensitivity parameter that scales the response as a function of flash intensity (S), and a small delay before response onset of the rod response (t _d). Clinically, these models are useful, because the sensitivity and the maximum response of photoreceptors can be dissociated, which can reveal selective deficits, not detectable with the standard ERG. ³⁵ For example, central vein occlusion decreases S, whereas Rm _P3 remains relatively stable. ³⁵  

The impact of ARM on the activation of rod phototransduction has not been addressed, in large part because ARM is considered a highly focal disease involving the macula only. However, the above-mentioned functional studies suggest that some effects of ARM may be more retinotopically diffuse. In addition, drusen and pigmentary changes commonly occur in peripheral retina of older eyes. ^{36 37}  

With respect to the impact of aging on the rod-mediated a-wave, two earlier studies reported an aging-related decline in the a-wave’s sensitivity parameter. ^{23 38} Some investigators have suggested that this aging effect cannot be attributed to the aged eye’s increased lenticular density, which effectively reduces the retinal illuminance of the stimulus flash. This conclusion has been based on estimates of lens density from other study samples that yield a “typical” lens density correction that is then applied to their own data. ²³ However, there is wide variability in lens density in older adults (a three-log-unit range of lens density for persons with IOLs to early cataract ³⁹ ), and thus lens density corrections based on average values do not serve well when interpreting an individual person’s retinal responses. ³⁸ The other approach described in the literature to minimize lens density is to use an achromatic stimulus flash. ³⁸ It is important to emphasize that rod vision is being evaluated in the present context, and rods are maximally sensitive to the shorter wavelengths, also the wavelengths most attenuated by the aged lens. Thus, when interpreting rod-mediated retinal responses, an achromatic flash also has inadequacies in adjusting for reduced retinal illuminance from increased lens density. A more direct method of assessing the impact of aging and ARM on rod responses is the use of older pseudophakic patients where lens density effects are minimized (if not eliminated). 

This study examined to what extent ARM impacts the activation of phototransduction in rod photoreceptor, with a primary focus on early ARM. In addition, we looked at the impact of normal aging on activation in older adults, both for comparison purposes with the ARM patients but also to clarify whether retinal aging itself is associated with changes in activation parameters. Two strengths of our approach, not incorporated in earlier studies, ^{23 38} are the use of fundus photographs and a fundus grading system as the basis of the case definition of both ARM and normal retinal aging, and studying persons with IOLs to control for the impact of the increased light absorption of the aged lens. 

Methods

Subjects

Participants were recruited from the comprehensive ophthalmology and the retina services of the Department of Ophthalmology, University of Alabama at Birmingham, and were among the clinic patients seen over a 15-month period from March 2001 to June 2002. Eligibility criteria for older adults were (1) age at least 60 years and (2) best corrected, distance visual acuity as listed in the medical record of 20/80 or better in at least one eye. Because the primary focus of the study was on rod activation in the early phases of ARM, an acuity cutoff of 20/80 was used; and (3) diagnosis of ARM or normal retinal health based on stereoscopic color 30° fundus photographs taken on the day of ERG testing after dilation of the pupil to at least 6 mm. Photographs were taken with a fundus camera (FF4; Carl Zeiss Meditec, Dublin, CA) on the eye selected for ERG testing, which was the eye with better visual acuity. Photographs were evaluated using the Wisconsin Age-Related Maculopathy Grading System (WARMGS) ¹⁵ at the University of Wisconsin Reading Center by graders masked to the patients’ characteristics, including previous diagnoses. 

For the purposes of this study, retinas were considered not to have signs of early ARM based on the following WARMGS grades: (1) maximum drusen type coded 0 (none), 1 (hard, indistinct drusen), or 2 (hard distinct, drusen); (2) maximum drusen size coded 0 (none), 1 (drusen indistinct or questionable), or 2 (distinct drusen with a diameter <63 μm); (3) increased retinal pigmentation coded 0 (none), or 1 (questionable); and (4) decreased RPE pigmentation coded 0 (none), or 1 (questionable). Early ARM was defined as WARMGS grade of: (1) maximum drusen type coded 3 (soft distinct drusen) or 4 (soft indistinct or reticular drusen); (2) maximum drusen size ≥3 (drusen diameter greater than 63 μm); (3) increased retinal pigmentation ≥2 (presence of increased pigmentation); and (4) decreased RPE pigmentation ≥2 (presence of decreased pigmentation). Late ARM was defined as those with a grade of 2 on the late ARM variable, indicating the presence of geographic atrophy and/or choroidal neovascularization. Eyes with late ARM could also have some of the other characteristics of eyes with early ARM, as described earlier, but their having a late ARM grade of 2 automatically placed them in the late ARM category. 

Patients whose test eye had a grade of 8 (cannot grade) on any of the following WARMGS variables were excluded from the sample: maximum drusen type, maximum drusen size, decreased pigmentation, increased pigmentation, late ARM, geographic atrophy, retinal detachment, subretinal hemorrhage, subretinal scar, and ARM treatment. Patients with evidence of diabetic retinopathy and its associated lesions (i.e., any patient with grades other than 10 [absent], 12 [nondiabetic], or 13 [questionable] for diabetic retinopathy level) were also excluded. 

Young adults were also enrolled in the study, to provide a reference group against which older participants without signs of early ARM could be compared on the outcomes of interest. They were recruited from the comprehensive ophthalmology service as described earlier. Inclusion criteria were (1) age 16 to 30 years; (2) no ophthalmic conditions or signs of maculopathy noted in a dilated comprehensive eye examination performed within the previous 6 months; and (3) best corrected distance acuity of 20/20 or better in each eye according to this examination. Fundus photographs were not taken on young enrollees. 

Patients, regardless of age, were excluded if their medical records or a general health interview indicated that they had any of the following: (1) glaucoma, optic neuropathy, or any ocular conditions other than ARM, refractive error or dry eye; (2) refractive error (spherical equivalent) with an absolute value more than 6 D, (3) neurologic diseases such as Alzheimer’s disease, Parkinson’s disease, or a history of stroke; (4) diabetes; (5) serious frailty or medical conditions expected to lead to mortality or disability within 12 months. 

This study was approved by the Institutional Review Board of the University of Alabama at Birmingham. The research adhered to the Tenets of the Declaration of Helsinki. Informed consent was obtained from all subjects after the nature and possible consequences of the study were explained. 

Procedures

Before ERG testing, best corrected distance visual acuity was measured for each eye on the day of ERG testing using the ETDRS chart ^{40 41} and expressed as logMAR. Contrast sensitivity was assessed for each eye with the Pelli-Robson chart and its standard administration protocol ⁴² and scored according to the letter-by-letter method. ⁴³  

The apparatus used to measure the rod a-wave was a visual electrodiagnostic system (UTAS-E 3000; LKC Technologies, Gaithersburg, MD), a commonly used clinical instrument also suitable for research purposes. Our technique for measuring the rod a-wave was similar to that of previous studies. ^{23 24 34 35 38 44} The tested eye (the eye with better acuity) was dilated to at least 6 mm in diameter using 1% tropicamide and 2.5% phenylephrine hydrochloride. Pupil size was measured before and after the ERG was recorded, to ensure that adequate dilation of the pupil was achieved throughout testing. Each individual’s pupil size was used to calculate the actual flash intensities used to model the ERG responses. Participants were dark adapted for 45 minutes before ERG recording. The cornea was anesthetized with 0.5% proparacaine hydrochloride. Recordings were obtained through the use of a Burian-Allen bipolar electrode placed on the anesthetized cornea. Responses were amplified (band-pass 0.5 to 8000 Hz; four-pole) and digitized at a 7 kHz sampling rate for a duration of 60 msec. Subjects were instructed to look straight ahead and to keep the eyes stable, and visual inspection by the examiner was used to ensure compliance. To estimate the rod response, ERG responses were obtained for two flash wavelengths based on previous approaches. ^{23 24 34 45} The first stimulus was an unattenuated short-wavelength flash (blue; 450 nm; 47A; Eastman Kodak, Rochester, NY) that is predominantly responded to by rods. The second was a long-wavelength flash (Red_Ph; 605 nm; 26; Eastman Kodak) photopically matched to the short-wavelength (450 nm) flash. Cones predominantly drive the response to the long-wavelength flash. The third was produced by an attenuated blue wavelength flash (blue_Sc 450 nm) that is scotopically matched to the red wavelength. The response to this third flash was subtracted from the red response to provide a better estimate of pure cone function. By computer subtraction of the responses to the above wavelengths [blue -(red_Ph -blue_Sc)], the isolated rod response was obtained. Responses were measured to the family of flash intensities (4.28, 3.84, 3.07, and 2.34 log scotopic Td-sec). The interflash interval ranged from 10 to 120 seconds so that complete recovery could occur between measurements. For each flash intensity and wavelength, 2 to 10 responses were averaged. 

Using Igor Pro (Wavemetrics, Inc.; Lake Oswego, OR), each participant’s ensemble rod responses were fit with the following equation to describe the response (R) as function of flash intensity (I), and time (t): R(I,t) = [1 − exp{−I · S ·(t − t _d)²}] · Rm _P3, where S is sensitivity, t _d is the delay before onset of the a-wave, and Rm _P3 is the maximum amplitude. All curve fits met a goodness of fit (R ²) of at least 0.95. Variables S and Rm _P3 were transformed to log values for data analysis. 

To compare demographic and visual function variables across groups of study subjects, analysis of variance (for continuous variables) and χ² tests (for categorical variables) were used. Rod-mediated ERG parameters were compared across study groups with and without adjustment for age and IOL status using analysis of covariance. ERG parameters were also compared according to fundus features using analysis of variance. All statistical analyses were performed on computer (SAS, ver. 8.02; SAS Institute Inc., Cary, NC). 

Results

Table 1 presents descriptive characteristics of the sample for demographics and visual function. There were 39 persons who met the case definition of early ARM and 7 who met that of late ARM. Of the latter, one had geographic atrophy and six had choroidal neovascularization. The average age of the young normal participants was 26 years; each of the three older groups had average ages in the 70s. The old normal group was younger than the early and late ARM groups (P = 0.006 and P = 0.02, respectively). The ages of the early and late ARM groups did not differ significantly. There was a slightly higher proportion of African Americans in the young normal group than in the three older adult groups. However, results from analyses comparing ERG parameters in the young and old normal adults described later are unchanged if analyses are limited to white participants. Visual acuity was highest in young subjects and was lower in the older groups, from better to worse as follows: old-normal, early ARM, and late ARM (all P < 0.01). These findings were true of both the tested and fellow eyes in all patients. A similar pattern of results was also observed for contrast sensitivity with the exception of contrast sensitivity in the tested eye for the old normal and early ARM groups; this difference was not significant. Otherwise, all group comparisons on contrast sensitivity were significant at P < 0.01. 

For the purposes of illustration, model-fitted a-wave families are shown in Figure 1 for a young normal adult (Fig. 1A) , an old normal adult (Fig. 1B) , an old adult with early ARM (Fig. 1C) , and an old adult with late ARM (Fig. 1D) . The data selected for Figure 1 are representative of subjects whose ERG parameters were closest to group averages (age-adjusted). The four solid-line traces, one for each flash intensity, are the rod-isolated ERG responses. The a-wave is the first negative peak of the response, which is followed by the rising phase of the b-wave. To fit the leading edge of the response, the model was applied to the response from 0 ms to the peak of the a-wave. The dashed lines represent the modeled a-wave activation, as if the activation phase was not obscured by the rising b-wave. 

The best-fit estimates of the ERG parameters for each patient group are listed in Table 2 . The first question addressed was what impact normal aging had on rod-mediated ERG a-wave parameters. Old normal subjects on average exhibited lower log S (P < 0.001) and lower log Rm _P3 (P < 0.001) and increased t _d (P = 0.001) compared with young adults. However, the effects of lens density must be taken into account when interpreting these results, because the aging lens preferentially absorbs the short wavelengths ^{17 46} used to isolate the rod response, thus effectively reducing the flash’s retinal illuminance more so in older adults than young. To examine whether older adults’ increased lens density impacts their ERG parameters, we compared the ERG parameters in old normal subjects who had an IOL of the conventional type with an ultraviolet radiation filter (Acrysof; Alcon Laboratories, Fort Worth, TX) in the test eye (n = 9) to those old normal subjects who were phakic (n = 32). Age did not differ between the two groups (P = 0.22) so no age adjustment was necessary in this comparison. There were no IOL versus phakic differences in the old normal group with respect to log Rm _P3 (P = 0.37) and t _d (P = 0.85). However, as illustrated in Figure 2 , old normal subjects with IOLs had a log S (mean, 1.12) approximately 0.3 units higher than that in phakic old normal subjects (mean, 0.8; t = −5.75; P < 0.001). Furthermore, when old normal subjects with IOLs were compared to young adults, the parameters of log S, log Rm _P3, and t _d were not statistically different from those in young adults (P = 0.78, P = 0.29, and P = 0.22, respectively). 

The effect of allowing t _d to vary in the ERG response model was examined in comparison to the more customary approach of holding the parameter constant. ^{23 24 34 35 38 44} The model was reapplied to each subject while holding t _d at 3.5 msec, the mean value of the young group. The unadjusted parameters are shown in Table 3 . For all subjects we found that holding t _d constant reduced the estimate of log S by −0.08 log units (P < 0.001) and slightly reduced the goodness of fit of the model by 0.01 (P < 0.001). The slight differences in parameter estimates did not affect the results of the previously described group analyses. 

The next question was whether the rod-mediated ERG parameters differed among the old normal, early ARM, and late ARM groups. An analysis of covariance was performed on each ERG parameter as the dependent variable examining the main effect of group. In these analyses we also adjusted for the potentially confounding effects of age and IOL, because both of these factors can influence ERG parameters. After age and IOL adjustment, there were no differences in any of the ERG parameters among the older groups (Table 2) . The individual ERG parameters for participants with early ARM are also shown in Figure 2 . 

Are rod-mediated ERG parameters related to ARM characteristics in the older participants as characterized by the WARMGS? The WARMGS characteristics of interest were maximum drusen type, maximum drusen size, drusen area, increased RPE pigmentation, and RPE depigmentation. To ensure adequate power, the levels of each characteristic were concatenated as follows. For maximum drusen type, two levels were created: (1) none or hard drusen and (2) soft drusen. For maximum drusen size, three levels were created: (1) none or questionable drusen, (2) less than a 63-μm diameter, and (3) a 63-μm diameter or more. Three levels were created for maximum drusen area: (1) a less than 63-μm diameter, (2) a less than 250-μm diameter, and (3) a less than 0.5 disc area. Two levels representing the absence or presence of pigmentary changes were created for increased RPE pigmentation and RPE depigmentation. Late ARM characteristics (geographic atrophy and choroidal neovascularization) were not specifically evaluated as part of this analysis, because only a few participants (n = 3) had these fundus features. As displayed in Table 4 , the only significant association between an ERG parameter and a fundus characteristic that emerged was that increased pigmentation of the RPE was related to decreased log S (P = 0.0447). 

Discussion

Our results suggest that patients with early ARM do not have abnormal rod-mediated a-wave responses when compared with their older counterparts who are free of ARM. After adjusting for age differences, there were no differences between ARM and ARM-free older patients in the sensitivity parameter of activation, in the delay of response onset after the stimulus flash, or in the amplitude of the response. These results suggest that the a-wave of the rod full-field ERG is not a useful measure for detecting or monitoring progression of ARM in its earlier phases, for providing clues about the mechanisms underlying rod-mediated visual sensitivity impairment in these patients, or for offering insights into ARM’s early pathogenic mechanisms. The rod ERG data of those with late-stage disease in our sample suggest that even in advanced disease, the surviving rod photoreceptors maintain activation capabilities similar to those of the healthy, older ARM-free retina. However, this result must be considered preliminary because of the small number of persons with advanced disease in the sample. It was further found that specific anatomic characteristics of ARM are in large part unrelated to the a-wave parameters, with the exception of a modest association between increased retinal pigment and the sensitivity of activation. 

The activation phase of rod phototransduction is a G-protein-modulated pathway self-contained in the rod outer segment and not directly dependent on the health of the RPE or Bruch’s membrane. ²⁹ In contrast, the retinoid cycle depends on the health of the RPE and Bruch’s membrane. Because ARM is a disease in which the earliest pathogenic signs appear in the RPE/Bruch’s membrane complex, a better bioassay of disease onset and progression may be disruptions or deficits in the retinoid cycle. As discussed earlier, rod-mediated dark adaptation, which is directly dependent on the retinoid cycle, is dramatically delayed in ARM, even in the very earliest phases of disease when acuity is still 20/25 or better and large and soft drusen are invisible in the fundus. ^{9 22} Future work should address its suitability as a marker for potential disease presence and progression. 

The rod-mediated full-field ERG is a mass electrical response representing the activity of approximately 100 million rod photoreceptors across the retina, with 96% of them outside the macula. ⁴⁷ Our results indicate that in ARM the dysfunction or loss of rods in a rather circumscribed retinal area (i.e., macula) appears to have minimal impact on the mass rod a-wave response (see also Ref. ⁴⁸ ). We cannot rule out the possibility that the activation of the rod photoreceptor response in the macula is indeed abnormal and that measurement of macular rod-mediated ERG a-wave parameters would reveal abnormalities. With the recent development of the multifocal (mf)ERG, ^{49 50} there is now a technique for assessing ERG responses in small localized retinal areas. Most of the prior work in this area has addressed cone-mediated responses. With respect to measuring the rod-mediated mfERG, stray light in the aged eye brings serious challenges in the measurement and interpretation of waveforms from older patients, as discussed previously, ⁵¹ because responses are small and recordings are noisy. Evidence also suggests that the rod mfERG largely reflects rod bipolar activity, ^{52 53} not that of the rod photoreceptor itself. Although there is preliminary evidence that the rod mfMERG may be useful in studying early-onset retinal degenerations, ^{51 54} it remains to be determined whether this technique can be reliably and validly implemented in studying ARM pathogenesis. 

Our results suggest that there is no impact of retinal aging on the activation phase of the rod a-wave, as evidenced by older adults with IOLs in good retinal health who had rod a-wave parameters indistinguishable from those of the young normal subjects. It appears that earlier studies may not have adequately controlled for lens density of their study patients in terms of its impact in reducing the flash magnitude of the ERG stimulus. ^{23 38} It is worthwhile to consider the impact of aging-related increases in lens density on the modeling of the a-wave responses. The flash intensity parameter is a measurement of corneal illumination and is assumed to be attenuated equally in all persons by the optical media before reaching the retina. This assumption is clearly violated when comparing individuals who of widely disparate ages (i.e., young and old adults) or individuals with various levels of cataract. Normal aging of the lens attenuates the transmission of light in a wavelength-dependent manner, even before cataract is considered clinically significant. In the activation model, the parameter S intensity is directly proportional to I, the flash intensity. Thus, an overestimation of flash intensity results in a directly proportional decrease in S. 

Using the lens density model of Pokorny et al. ¹⁷ for the flash and filter characteristics, we estimated that the 70.7-year-old normal adult’s retinal illumination was 0.3 log unit less than that of the average 25.7-year-old normal subject in our sample. Thus, on average, the flash intensities in the model were overstated by 0.3 log unit for our older participants in comparison to the young group. To determine whether this overestimation of the flash intensities in older adults can account for the differences in S between the old normal and young groups, the waveforms of our average old normal adult (t _d = 3.8 ms, S= 7.41, Rm _P3 = −323.60) were refit with the model using flash intensities reduced by 0.3 log unit. The results were that the estimated S increased from 7.41 to 14.79, a value that is quite similar to 13.49 (the mean of young subjects) and 13.18 (mean of old normal adults with IOLs). Similarly, inflating the flash intensities of the average old normal adults with an IOL by 0.3 log unit reduced S from 13.18 to 6.93 which compares to a value of 6.30 for the phakic old adult. Thus, the mean difference in log S of 0.26 between the normal young adults and old normal adults can be accounted for by the 0.3 log unit reduction in retinal illumination of the old normal adult. Both the theoretical computations just described and the empiric observations of old normal adults with IOLs support the hypotheses that the aging-related decrease in S is largely attributable to aging-related changes in lens density. 

The advantage of the ERG for the diagnosis and monitoring of retinal disease lies in its ability to measure objectively and precisely the function of the neuroretinal cell types responsible for visual function. However, our data suggest that the utility of the flash ERG as a diagnostic tool for early ARM is rather limited by its susceptibility to misinterpretation because of the increased lens density of later adulthood, which varies widely among older adults from minimal to cataractous. ⁵⁵ From a practical standpoint, studies to determine the impact of aging on the retinal origins of the ERG response should include pseudophakic subjects, to minimize lens density as a confounding variable. 

In summary, the activation of the a-wave as measured by the rod-mediated full-field ERG is not affected by early ARM, nor is it affected by normal retinal aging, when steps are taken to take the light absorption of the aged lens into account and a reliable definition of normal retinal aging is used. 

 Supported by National Institute on Aging Grant R01-AG04212, National Eye Institute Grant R21-EY14071, a National Institutes of Health Research Supplement for Underrepresented Minorities, the EyeSight Foundation of Alabama, and Research to Prevent Blindness, Inc. CO is a Research to Prevent Blindness Senior Scientific Investigator.

 Submitted for publication January 8, 2004; revised March 16 and May 11, 2004; accepted May 21, 2004.

 Disclosure: G.R. Jackson, None; G. McGwin, Jr, None; J.M. Phillips, None; R. Klein, None; C. Owsley, 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.

 

Demographics and Visual Function in the Patient Groups

	Young Normal (n = 27)	Old Normal (n = 41)	Early ARM (n = 39)	Late ARM (n = 7)
Age, years mean (SD)	25.7 (3.0)	70.7 (5.0)	74.7 (6.0)	76.4 (10.0)
Gender (% male)	35.7	53.5	51.3	57.1
Race (%)
White	85.7	93.0	94.9	100.0
African American	14.3	7.0	5.1	0.0
Visual acuity (logMAR)
Tested eye	−0.05 (0.06)	0.02 (0.08)	0.14 (0.17)	0.27 (0.22)
Fellow eye	0.02 (0.14)	0.14 (0.20)	0.36 (0.34)	0.79 (0.36)
Contrast sensitivity (log)
Tested eye	1.65 (0.14)	1.45 (0.26)	1.37 (0.17)	1.06 (0.40)
Fellow eye	1.62 (0.14)	1.42 (0.27)	1.22 (0.43)	0.79 (0.27)

 Visual acuity and contrast sensitivity are expressed as the mean (SD).

Figure 1.

 

Representative a-wave families of (A) a normal young adult, (B) old normal adult, (C) early ARM patient, and (D) late ARM patient. Dashed lines: best ensemble fit of the equation stated in Methods.

Table 2.

 

Rod-Mediated ERG Parameters for the Young and the Three Older Adult Age Groups

	Young Normal (n = 27)	Old Normal (n = 41)	Early ARM (n = 39)	Late ARM (n = 7)	P
t d, mean (SD)	3.5 (0.31)	3.8 (.30)	3.8 (0.39)	3.6 (0.52)	0.02
Age-adjusted	—	3.8	3.7	3.6	0.43
Age/IOL adjusted	—	3.8	3.7	3.6	0.36
log S	1.13 (0.25)	0.87 (0.27)	0.82 (0.32)	0.77 (0.35)	0.001
Age-adjusted	—	0.89	0.83	0.74	0.43
Age/IOL adjusted	—	0.96	0.92	0.90	0.76
log Rm P3	2.58 (0.05)	2.51 (0.09)	2.49 (0.1)	2.42 (.04)	0.001
Age-adjusted	—	2.50	2.50	2.44	0.24
Age/IOL adjusted	—	2.50	2.50	2.44	0.25

 Young normal subjects are excluded in age-adjusted data.

Figure 2.

 

An individual participant’s estimates (○) of the ERG kinetics plotted for the young normal adults, old normal adults, old normal adults with IOLs, persons with ARM, and persons with ARMs with IOLs (for the latter two categories, early and late ARM are plotted separately). Horizontal line: group mean. (A) Old adults and early ARM patients with IOLs exhibited log S values similar to those of young adults. (B, C) Estimates of t _d and log Rm _P3, respectively, were similar across patient groups and did not appreciably change with IOL status.

Table 3.

 

Effect of Varying and Holding t _d in the a-Wave Model on Estimates of Log S and Log Rm _P3.

	Young Normal (n = 27)			Old Normal (n = 41)			Early ARM (n = 39)			Late ARM (n = 7)
		t d Held	t d Varied	t d Held	t d Varied	t d Held	t d Varied	t d Held	t d Varied
Log S	1.18	1.13	0.88	0.87	0.87	0.87	0.82	0.77
Log Rm P3	2.60	2.58	2.52	2.51	2.50	2.41	2.43	2.44

Table 4.

 

ERG Parameters as a Function of Fundus Features as Defined by the WARMGS

ARM Lesion	F Statistic	P
Maximum drusen size
Log S	0.04	0.96
t d	0.61	0.55
Log Rm P3	0.55	0.78
Maximum drusen type
Log S	0.00	0.99
t d	0.13	0.72
Log Rm P3	0.00	0.95
Drusen area
Log S	0.26	0.78
t d	0.75	0.48
Log Rm P3	0.26	0.77
Increased RPE pigmentation
Log S	4.19	0.04
t d	1.25	0.27
Log Rm P3	2.75	0.10
RPE Depigmentation
Log S	0.94	0.34
t d	0.71	0.40
Log Rm P3	1.32	0.25