Background Adaptation in Children with a History of Mild Retinopathy of Prematurity

Abstract

purpose. In children with a history of mild retinopathy of prematurity (ROP), test the hypothesis that elevation of the parafoveal over peripheral dark-adapted threshold is due to photoreceptor rather than postreceptor dysfunction.

methods. A forced choice procedure was used to measure thresholds for detection of 2^o diameter, 50 msec, blue stimuli presented 10^o (parafoveal) or 30^o (peripheral) eccentric in the dark and in the presence of steady red backgrounds (−4 to +2 log scot td). Four ROP and four control subjects were tested at both eccentricities. A model of the increment threshold function was fit to the data to calculate the eigengrau and dark-adapted threshold.

results. Both ROP subjects with elevated parafoveal thresholds also have elevated parafoveal eigengraus. On the other hand, parafoveal and peripheral eigengraus are equal in ROP subjects without parafoveal threshold elevation. Nevertheless, the dark-adapted thresholds of all ROP subjects are higher than those of any control subject at both sites.

conclusions. The parafoveal threshold elevation is due to rod dysfunction. There is also evidence of peripheral rod photoreceptor involvement in the subjects with ROP.

Retinopathy of prematurity (ROP) onset occurs during preterm ages when the rod outer segments are rapidly developing. ^{1 2} Studies of electroretinogram (ERG) rod phototransduction processes in infants and children with a history of ROP show that significant alterations of rod function persist long after active ROP has resolved. ^{3 4} In a rat model of ROP, ERG abnormalities ⁵ of the same type as in the children and also alterations of the structure and molecular properties of the rod outer segments ⁶ have been demonstrated. 

In addition, rod-mediated vision is affected in some children with a history of mild ROP. ⁷ Specifically, the dark-adapted psychophysical threshold in the parafoveal retina is elevated relative to the threshold in the peripheral retina. ⁷ In control subjects, the parafoveal and peripheral thresholds are equal. Because the course of development of primate parafoveal rod outer segments is delayed compared with that of peripheral rod outer segments, ^{1 8 9} parafoveal rods are less mature than peripheral rod outer segments during the time of active ROP, and possibly ⁷ more susceptible to the outer segment abnormalities that occur in ROP. Short disorganized outer segments were found in a rat model that had ERG abnormalities. ⁶ Assuming similar outer segment changes in ROP subjects, reduced quantum catch and consequent threshold elevation may occur. However, threshold is determined not only by receptoral but also by postreceptoral processes. ^{10 11 12 13}  

Psychophysical studies of scotopic background adaptation in patients with retinal disorders can distinguish between receptoral and postreceptoral sites of the action of the disease. ^{12 14 15 16} For example, a disease acting on the photoreceptors to reduce quantum catch by reducing pigment content, shortening outer segments, or misaligning the receptors, a disease that causes differential loss of receptors at different locations, or a disease that reduces response amplitude will decrease detection of both the test and background stimuli. ¹² In this situation, the dark-adapted threshold and the background at which threshold rises significantly above the dark-adapted level are both elevated. Thus, if the relative elevations of the dark-adapted parafoveal threshold in ROP subjects ⁷ are due to rod dysfunction, the parafoveal increment threshold function, displayed on log-log coordinates, will be shifted up and over from the peripheral increment threshold function. ¹² Changes limited to postreceptoral sites shift the increment threshold function vertically without any horizontal shift. ¹² We obtained parafoveal and peripheral increment threshold functions from ROP and control subjects to test the hypothesis that relative elevation of the parafoveal dark-adapted threshold in ROP is due to rod photoreceptor dysfunction. 

Methods

Rod-mediated thresholds for detection of 50 msec, 2^o diameter, blue (Wratten 47B, λ < 440 nm) spots were obtained using a two alternative forced-choice procedure. Stimuli were presented on a rear projection screen 10^o (parafoveal) or 30^o (peripheral) from the center. A red light-emitting diode (LED) fixation target that subtended 30-minute arc and flickered at 1 Hz was at the center of the screen. The steady red (Wratten 29, λ > 610 nm) background field was circular (109^o diameter) and concentric with the fixation target. Calibrated neutral density filters controlled the intensity of the test and background lights. 

Calculation of the retinal illuminance produced by the test and background lights was based on luminance measurements made with a calibrated photodiode (UDT S-350) placed in the position of the subject’s eyes. The scotopic troland value of the stimuli ¹⁷ was calculated taking each subject’s measured pupillary diameter and media density ¹⁷ into account. The subject’s pupillary diameter was estimated by direct observation with an infrared viewer and comparison to the diameter of the cornea (which is 11 mm). ¹⁸  

After 30 minutes of dark adaptation, the subject sat 50 cm in front of the rear projection screen. As in previous studies of dark-adapted thresholds, ^{7 19 20} subjects viewed stimuli with both eyes. The subject was instructed to fixate on a flickering, red, LED fixation target. Then, a stimulus was presented, and the subject reported stimulus position (right or left). On every trial, the subject received feedback. Threshold was determined with a transformed up-down staircase that estimated the 70.7% correct point of the psychometric function. ²¹ At least five alternations were recorded. 

Subjects were tested first in the dark-adapted (no background) condition and then while adapted to backgrounds that produced retinal illuminances of approximately −4 to approximately +2 log scot td. Half of the subjects were tested first at the parafoveal site and half at the peripheral site. Each subject completed testing in one session. 

A model of the increment threshold function ^{11 12 13} was fit to the threshold data to calculate the values of dark-adapted threshold (T _D ) and eigengrau (A _O ) that minimized the sum of squared deviations from the following equation:  

\[\mathrm{Log\ }\mathit{T}{=}\mathrm{Log\ }\mathit{T}_{\mathit{D}}{+}\mathrm{Log}{[}(\mathit{A}_{\mathit{O}}{+}\mathit{I})/\mathit{A}_{\mathit{O}}{]}\]

where T is the threshold at background intensity I. Disease limited to the receptors elevates both T _D and A _O . Disease limited to a postreceptoral site elevates T _D only. ¹²  

From a sample of ROP subjects previously studied in the dark-adapted condition, ⁷ two with, and two without, relative elevation of the parafoveal over peripheral threshold were recruited for this study of background adaptation. Control subjects were recruited by word of mouth. The characteristics of the subjects are summarized in Table 1 . The study conformed to the tenets of the Declaration of Helsinki and was approved by the Children’s Hospital Committee on Clinical Investigation. Written informed consent was obtained. 

Results

The model ^{11 12} provides a good description of the increment threshold functions in ROP (Fig. 1) and control (Fig. 2) subjects. The values of T _D and A _O for each subject are summarized in Table 2 . 

In the ROP subjects (Nos. 1 and 2) known to have relative elevation of the dark-adapted parafoveal threshold, ⁷ the parafoveal increment threshold function is shifted from the peripheral function (Fig. 2 , left panels). The eigengrau values, A _O , for the parafoveal increment threshold function are shifted toward brighter backgrounds, by 0.86 and 0.35 log units higher than those for the peripheral increment threshold function (Table 2) . The dark-adapted thresholds, T _D , of these two ROP subjects were 0.43 and 0.52 log units higher in the parafoveal than peripheral retinas (Table 2) and agree well with their thresholds (Table 1) measured 2 years earlier. ⁷ Furthermore, their dark-adapted peripheral thresholds at −3.20 and −3.22 log scot td sec are approximately 0.2 log unit above the median dark-adapted threshold of controls (Table 2) . Of interest, their rod photoreceptor sensitivities, ⁴ derived from ERG responses to full-field stimuli, were 0.33 and 0.38 log units below the normal mean (Table 1) . 

In both ROP subjects (Nos. 3 and 4) without elevation of the dark-adapted parafoveal threshold, the parafoveal and peripheral increment threshold functions are superimposed (Fig. 1 , right panels). The values of T _D and A _O are the same at the parafoveal and peripheral sites (Table 2) . The calculated values of T _D are similar (Table 1) to those measured 2 years earlier. ⁷ Even though T _D s in parafoveal and peripheral retinas are equal, they are higher, by 0.23 and 0.25 log units, than the median T _D in control subjects. Their values of A _O (−2.56 to −2.74 log scot td) were significantly lower (Mann–Whitney U = 31; P = 0.004), that is, shifted toward dimmer backgrounds, than those of controls (−2.28 to −2.61 log scot td). Thus, their increment threshold functions differ from those of controls by equal changes in T _D and A _O . In other words, compared with controls, their increment threshold functions are shifted up and over along a diagonal, indicating that receptoral function also determines the threshold differences between these two ROP subjects without relative parafoveal threshold elevation and the control subjects. 

Dark-adapted parafoveal and peripheral thresholds are equal (within 0.04 log unit) in every control subject. The parafoveal and peripheral increment threshold functions of the control subjects, even if myopic, are superimposed (Fig. 2) . Their eigengrau values in the parafoveal (−2.28 to −2.61 log scot td) and peripheral (−2.31 to− 2.55 log scot td) retinas do not differ significantly. 

Discussion

The shift in the parafoveal increment threshold functions of ROP subjects 1 and 2 is evidence that rod dysfunction accounts for the relative elevation of the dark-adapted parafoveal threshold. The calculated values of the parafoveal eigengrau, A _O , for these two ROP subjects are higher than their peripheral values. Furthermore, their peripheral eigengrau values are higher than those of any of the control subjects. This suggests that their rods, both peripheral and parafoveal, have higher intrinsic noise than normal and that the parafoveal rods are noisier than peripheral rods. Possibly these ROP subjects have some disorganization of the rod outer segments, such as that demonstrated in a rat model of ROP. ⁶  

ROP subjects, Nos. 3 and 4, who do not have a shift of their increment threshold functions, have, nevertheless, significant elevations of their dark-adapted thresholds. Their eigengrau values also differ, by a small but significant amount, from those of the controls, being shifted to dimmer backgrounds by approximately 0.2 log unit, an amount equal to the threshold elevation. In these subjects, reduced quantum catch (elevated T _D ) and lower intrinsic noise may be explained by short outer segments, as was found in the rat model, ⁶ and as is also consistent with the low amplitudes of the saturated rod response obtained in ROP subjects. ^{3 4} A difference between the dark-adapted peripheral thresholds of ROP and control subjects was not recognized in the Reisner et al. study. ⁷ The controls in that study were older (range, 10–48 years) than those in the present study, and a correction for media density was not applied. Assuming that media density increases with age, on average, 0.12 log unit per decade, ^{22 23} the median peripheral threshold for controls in the Reisner et al. study ⁷ is −3.43 log scot td sec, or 0.24 log unit more sensitive than the median peripheral threshold of the ROP subjects in the Reisner et al. study. ⁷ In any case, it is well to keep in mind that these are small differences, similar to the intersubject variability reported with this type of testing. ^{7 19}  

Even though all ROP subjects have nearly identical peripheral dark-adapted thresholds that are higher than those of the controls (Table 2) , only the two (subjects 1 and 2) who were high myopes from early childhood have further elevation of the parafoveal threshold. As the results from subjects 5 and 6 attest, myopia alone does not cause parafoveal threshold elevation. ⁷ Subjects 1 and 2 are also known to have mild acuity deficits (20/30 or 20/40 rather than 20/20). ⁷ Thus, both the late-developing parafoveal rods and the foveal cones, which have an even more protracted course of development, ²⁴ may have been disturbed in these ROP subjects. The mechanisms that lead to the association of myopia in mild ROP and parafoveal rod dysfunction remain to be defined. In view of the previous studies of ROP subjects ^{3 4 7} and the rat model, ^{5 6} the present psychophysical results lead us to suspect that those factors that cause persistent disorganization of the photoreceptor outer segments are also involved in the regulation of eye growth in ROP. 

 Supported by National Institutes of Health Grant EY10597.

 Submitted for publication May 20, 1999; revised July 26, 1999; accepted August 27, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Characteristics of Subjects

Subject	Age (years)	ROP*	Refraction Spherical Equivalent (diopters) Right, Left	Psychophysical Threshold, †			Rod ERG, ‡ S (isoms−1 sec−2)
									Parafovea (scot td sec)	Periphery (scot td sec)
1	13	Zone 3, stage 3	−8.30, −8.40	−2.75	−3.06	4.79
2	12	Zone 2, stage 2	−8.90,−8.90	−2.92	−3.25	4.26
3	16	Zone 3, stage 1	0.43, 0.01	−3.21	−3.21	—
4	17	Zones 2–3, stage 2	0.17, 0.04	−3.13	−3.14	—
5	21	None	−4.00,−5.25	—	—	—
6	35	None	−5.00,−4.75	—	—	—
7	14	None	Plano, Plano	—	—	—
8	9	None	Plano, Plano	—	—	—

*  The site (zone) and severity (stage) of the maximum acute phase ROP are designated according to the International Classification of Retinopathy of Prematurity. ²⁵

†  Reisner et al. ⁷ previously reported these dark-adapted thresholds.

‡  In a previous ERG study ⁴ of the activation of phototransduction, these values of S, a sensitivity parameter, were obtained. These values are 0.33 and 0.38 log unit below the normal value of S (mean = 10.2; SD = 1.6 isoms⁻¹ sec⁻²; n = 24 subjects).

Figure 1.

 

The increment threshold functions of ROP subjects. Thresholds at the peripheral site are represented by filled circles, and at the parafoveal site by open circles. The smooth curves plot Equation 1 fit to the data (parafoveal, solid line; peripheral, dashed line). If, in a given adaptation condition, only a filled circle is shown, the parafoveal threshold (open circle) was the same.

Figure 2.

 

Increment threshold functions of control subjects. All features of these graphs are as in Figure 1 .

Table 2.

 

Parameters of Increment Threshold Functions

Subject	Parafoveal Site			Peripheral Site
		T D Dark-Adapted Threshold (log scot td sec)	A O Eigengrau (log td)	T D Dark-Adapted Threshold (log scot td sec)	A O Eigengrau (log td)
ROP
1	−2.77	−1.11	−3.20	−1.97
2	−2.70	−1.46	−3.22	−1.78
3	−3.20	−2.74	−3.20	−2.56
4	−3.18	−2.65	−3.22	−2.7
Controls
5	−3.45	−2.61	−3.45	−2.54
6	−3.45	−2.50	−3.49	−2.55
7	−3.41	−2.41	−3.40	−2.46
8	−3.40	−2.28	−3.40	−2.31