Comparison of visual acuity results obtained using Early Treatment for Diabetic Retinopathy Study charts at the 10-year examination with contrast sensitivity results obtained at the same examination using the Pelli-Robson letter charts. A, Results for all treated (T) and control (C) eyes of patients in the randomized portion of CRYO-ROP trial. B, Results for paired eyes of patients in the randomized trial who had 1 treated eye that was sighted and 1 control eye that was sighted.
Cryotherapy for Retinopathy of Prematurity Cooperative Group. Contrast Sensitivity at Age 10 Years in Children Who Had Threshold Retinopathy of Prematurity. Arch Ophthalmol. 2001;119(8):1129-1133. doi:10.1001/archopht.119.8.1129
To evaluate monocular contrast sensitivity (CS) at age 10 years in eyes that underwent cryotherapy and in control eyes of participants in the randomized trial of cryotherapy for retinopathy of prematurity (CRYO-ROP), and in a comparison group of 10-year-old low-birth-weight children who did not develop ROP in the neonatal period.
Eligible participants were the 255 survivors of the group of 291 infants with severe (threshold) ROP in 1 or both eyes who were enrolled in the randomized CRYO-ROP trial as neonates, as well as 104 children in the CRYO-ROP study who did not develop ROP. All had birth weights less than 1251 g. Contrast sensitivity was measured using Pelli-Robson charts at a test distance of 1 m with luminance at or higher than 64 candelas (cd)/m2. Contrast sensitivity was estimated using the total number of letters identified.
Unfavorable CS outcomes (<27 letters identified) were found in a greater proportion of control eyes (59.3%) than of treated eyes (39.7%) (P<.001). In patients with bilateral threshold ROP who had quantifiable CS in both eyes, CS results were similar in treated and control eyes, suggesting no detrimental effect of cryotherapy. In the "no ROP" group, 96.9% of eyes showed CS in the normal range, compared with 48.1% of treated eyes and 34.6% of control eyes.
The results further confirm the beneficial effect of cryotherapy on visual function, and show no evidence of adverse effects of cryotherapy on CS. With or without cryotherapy, eyes with severe ROP showed substantially poorer CS than did eyes of preterm children who did not develop ROP.
THE STANDARD clinical measure of visual function is visual acuity (VA), which provides an indication of a patient's ability to discriminate fine detail(high spatial frequencies). However, some serious ocular disorders, for example, optic neuritis1,2 and cataract,3 can have profound effects on a patient's ability to detect low-range and mid-range spatial frequencies, while leaving high-contrast VA relatively intact.
A primary goal of the multicenter study of Cryotherapy for Retinopathy of Prematurity (CRYO-ROP) was to evaluate the functional outcome of eyes that received cryotherapy for severe ROP compared with eyes with severe ROP that served as controls. Functional outcome was evaluated by assessment of VA when children were 1, 2, 3½, 4½, and 5½ years of age, as well as by measurement of visual field extent and color vision at age 5½ years. However, to date, no measurement of sensitivity to low-range or mid-range spatial frequencies has been reported for eyes with a history of severe ROP or cryotherapy.
Sensitivity to a range of spatial frequencies can be measured using sinusoidal gratings that vary in contrast and spatial frequency (eg, the Vistech Contrast Sensitivity [CS] Chart4) or a chart with letters of various levels of contrast (eg, the Pelli-Robson Letter Sensitivity Chart5). At the 10-year CRYO-ROP study examination, treated and control eyes were tested with Pelli-Robson letter sensitivity charts after completion of distance and near monocular acuity assessment. The results provide an additional comparison of visual function in eyes that underwent cryotherapy vs eyes with severe ROP that served as controls. Further comparison is also made with CS results from a subset of CRYO-ROP study participants who did not develop ROP during the neonatal period.
Subjects in the randomized cohort of the CRYO-ROP study were 291 children with birth weights of less than 1251 g who developed severe (threshold) ROP in 1 or both eyes during the neonatal period (threshold ROP was defined as 5 contiguous or 8 cumulative clock hours of stage 3+ ROP in zone 1 or zone 2).6 All were born between January 1, 1986, and November 30, 1987. In the 240 children who were found to have threshold ROP in both eyes at the same examination (bilateral threshold ROP), one eye was randomly assigned to undergo cryotherapy and the other was assigned to serve as a control. In the remaining 51 children, threshold ROP developed in only one eye, and that eye was randomly assigned to receive cryotherapy or to serve as a control. More complete descriptions of the study population have been published elsewhere.7- 11 Among the 255 children who survived until age 10 years, 247 (97%) completed the 10-year study examination, and 8 (3%) failed to return for this follow-up examination.
A comparison group of 104 CRYO-ROP study participants from one selected Participating Center (Philadelphia, Pa) was also eligible for follow-up to age 10 years, and 102 of these children completed the 10-year study examination. Children in this comparison ("no-ROP") group, who also had birth weights of less than 1251 g, were participants in the natural history portion of the CRYO-ROP study but did not develop ROP.12
Informed consent was obtained from parents prior to study entry, at randomization, and prior to both the 5½-year and 10-year follow-up phases of the study.
Monocular CS testing was conducted during the CRYO-ROP 10-year study examination by a tester who was masked to the treatment status of each child's eyes. Testing occurred after the child had completed monocular distance and near VA testing, but prior to dilated eye examination and visual field testing. Refractive error was corrected during testing.11
Contrast sensitivity was measured using 2 Pelli-Robson CS charts (Clement Clark Inc, Columbus, Ohio), 1 for the right eye, which was tested first, and another for the left eye.5 Each Pelli-Robson chart contains 8 lines of 4.9 × 4.9-cm letters, with 6 letters (2 triplets) per line. Within each triplet, letters have the same contrast, but contrast decreases in 0.15–log unit (lu) steps between successive triplets. The charts were illuminated uniformly to a luminance at or greater than 64 candelas(cd)/m2. Subjects were tested at a distance of 1 meter, and were asked to identify each letter on the chart, starting with the high-contrast letters in the upper left corner of the chart and reading horizontally across each line. Only 1 attempt per letter was permitted, and subjects were asked to guess even when they indicated that they thought that the letters were invisible. Contrast sensitivity was estimated by the total number of letters identified by the subject.13
Data were analyzed separately for treated eyes and control eyes of patients in the randomized group. For the no ROP comparison group, one eye per patient was selected at random for inclusion in the data analysis.
As in previously published analyses of CRYO-ROP VA outcomes,8- 10 CS outcomes were divided into 4 categories: normal, below normal, poor, and blind. "Normal" CS was defined as correct identification of 33 or more letters on the Pelli-Robson chart or roughly equivalent to a CS of 1.50 lu. This cut-off for normal was based on data from Elliott and Whitaker,14 showing an average CS of 1.7 lu in normal subjects between 10 and 19 years of age, and recent CS data presented by Myers et al15 from a group of 106 full-term 10-year-old children without ocular pathologic features, indicating that 96% of the children detected 33 or more letters on the Pelli-Robson chart. "Below normal" CS was defined as detection of 27 to 32 letters (roughly equivalent to a CS between 1.20 and 1.50 lu). This lower cut-off was based on data from Myers et al15 indicating that 99% of normal, full-term 10-year-old children detected 27 or more letters. Contrast sensitivity outcomes in the normal and below normal categories were classified as "favorable" outcomes.
Eyes that did not meet the criteria for classification as favorable CS outcomes were classified as "unfavorable." Eyes in the unfavorable category were divided into 2 subgroups: (1) poor (ie, those with quantifiable CS scores of ≤26 letters) and (2) blind to CS targets (ie, those eyes that were unable to detect any letters on the CS chart). This included eyes with measurable recognition acuity that could not detect letters on the CS chart, eyes without measurable recognition acuity, and eyes of children who were exempted from recognition acuity and CS testing because both the parents and the physician agreed that the child's VA was no better than light perception.
As presented in Table 1, 85 treated eyes (39.7%) and 125 control eyes (59.3%) had unfavorable CS outcomes. Thus, there was a reduction of 33.1% in unfavorable CS outcomes in treated vs control eyes (P<.001). Within the favorable outcome category, there were more treated eyes than control eyes in both the normal and below normal subcategories. On the other hand, within the unfavorable category, there were similar numbers of treated eyes and control eyes in the"poor" subcategory, and more control eyes than treated eyes categorized as blind.
Table 1 also provides data from eyes of study patients who did not develop ROP during the neonatal period. Among eyes in this no ROP group, nearly all (96.9%) showed normal CS. One patient was unable to complete CS testing because of neurodevelopmental delay, and 3 additional patients underwent their 10-year study examination before the CS charts were available.
Among the sighted eyes in Table 1 (ie, those eyes in the normal, below normal, and poor subcategories), treated eyes showed better CS than did control eyes. However, some patients contributed data from 2 sighted eyes to Table 1, while others contributed data from only 1 sighted eye, leading to a potential selection bias in the comparison of results from sighted treated vs sighted control eyes. To avoid this potential bias, we examined data from the subset of 82 patients who had bilateral threshold ROP (1 treated and 1 control eye) and who had quantifiable CS in both eyes. As presented in Table 2, the proportion of eyes in the normal, below normal, and poor categories was similar in the treated and control groups, indicating no difference in CS between treated vs control eyes, in which sight was preserved.
At the 10-year examination, all CRYO-ROP study participants had visual acuity tested with Early Treatment for Diabetic Retinopathy Study (ETDRS) logMAR letter charts,16 in accordance with study protocol.10,11Figure 1 and Table 3 provide a comparison of VA results obtained using ETDRS charts vs CS results obtained using the Pelli-Robson letter charts. Visual acuity results are classified as favorable or unfavorable, and subdivided as normal (20/40 or better), below normal (worse than 20/40 to better than 20/200), poor (quantifiable acuity of 20/200 or worse), and blind (minimal pattern vision, light perception, or no light perception).11 Data from all treated and all control eyes of patients in the randomized portion of CRYO-ROP trial are presented in part A of Figure 1 and in data columns 2 and 3 of Table 3. Data from paired eyes of patients in the randomized trial who had both a treated eye and a control eye with quantifiable vision are presented in part B of Figure 1and in the 2 right data columns of Table 3. Data from eyes of patients who did not develop ROP are also presented in Table 3. In the no ROP group, nearly all eyes showed normal CS and VA, whereas in eyes with severe ROP, there were many eyes that showed better CS than VA.
The CRYO-ROP study was designed to determine the efficacy and safety of peripheral retinal cryoablation for the treatment of eyes with severe (threshold) ROP. During infancy and early childhood, visual function in treated eyes and control eyes was compared using a grating acuity measure.8,9 As follow-up continued and children were able to perform more complex tasks, other measures were added to the study examination to provide a more comprehensive assessment of visual function.9,10,17,18 Overall, the results of the various assessment procedures have shown a benefit of cryotherapy in preserving both visual function and retinal structure, and have failed to reveal any clear indications of adverse consequences.
At the 10-year study examination, CS testing using Pelli-Robson charts was added to provide a more complete evaluation of the effect of cryotherapy on pattern vision. The results of CS testing confirm the clear benefit of cryotherapy on preservation of pattern vision. Among treated eyes, 39.7% showed unfavorable CS outcomes, whereas 59.3% of control eyes showed unfavorable CS outcomes. These results are similar to VA outcomes at the 5½-year examination, at which 47.1% of treated eyes showed unfavorable VA outcomes and 61.7% of control eyes had unfavorable outcomes.10 The results are also similar to the 10-year examination VA values of 44.4% unfavorable VA outcomes in treated eyes and 62.1% unfavorable outcomes in control eyes.11 Thus, the results of CS testing provide further support for the efficacy of cryotherapy in threshold ROP.
To evaluate the safety of cryotherapy, treated eye and control eye CS outcomes were compared in the subset of patients with bilateral threshold ROP who had quantifiable CS results in both eyes (Table 2). The results showed no difference in the distribution of CS outcomes between treated and control eyes, suggesting that cryotherapy has no adverse effect on CS in eyes with severe ROP that retain vision.
The CS results support both the efficacy and safety of cryotherapy in preserving sensitivity to low-range and mid-range spatial frequencies. Similarly, the results of VA testing presented elsewhere10,11 support the efficacy and safety of cryotherapy in preserving sensitivity to high spatial frequencies. To compare the relative benefits of cryotherapy in preserving vision for higher vs lower spatial frequencies, VA results from the 10-year examination11 were compared with CS results in treated and control eyes. As in previous reports of the CRYO-ROP study, acuity results were categorized as normal, below normal, poor, and blind, with the first 2 categories being "favorable" and the latter 2 "unfavorable."8- 11 As presented in Figure 1 and Table 3, the proportion of eyes that showed better CS than VA was substantially greater than the proportion of eyes that showed better VA than CS, for both treated eyes and control eyes. This suggests that severe ROP is more detrimental to sensitivity to high spatial frequencies than to sensitivity to low-range and mid-range spatial frequencies. Examination of the results from patients with bilateral threshold ROP in whom sight was preserved in both eyes (part B of Figure 1 and fourth and fifth data columns of Table 3) indicated no difference in the proportion of treated eyes vs control eyes in which CS was better than VA. This suggests that the relation between CS and VA in these eyes with severe ROP in the neonatal period was not influenced by whether or not peripheral retinal ablation with cryotherapy was performed.
One problem with categorical comparison of 2 continuous measures of visual function is that the cut-off points for the categories are somewhat arbitrary, and the results of the comparison will depend on the where the cut-off points are placed. In the data presented in Figure 1, better agreement between VA and CS results would be obtained if the cut-off points of VA were lowered or if the cut-off points for CS were raised. However, such a change is not supported by normative data for VA and CS for children in this age range. Data from Myers et al15 based on 106 normal full-term children, showed that 99% had VA that would be classified as normal (20/40 or better) in the CRYO-ROP study and that 96% had CS who would be classified as normal by the criterion used in the present paper. Thus, widening the range of "normal" VA to below 20/40 or raising the"normal" range of CS would not seem reasonable.
In addition to the outcomes of treated and control eyes, this report also includes data from a comparison group of children who did not develop ROP, who had birth weights less than 1251 g, and who were recruited into the CRYO-ROP study during the same time as the subjects who participated in the randomized trial. Contrast sensitivity outcomes in those eyes that did not develop ROP were considerably better than those of treated or control eyes that developed severe ROP during the neonatal period (Table 1). Among the no ROP eyes, 96.9% showed normal CS, while only 48.1% of treated eyes and 34.6% of control eyes showed CS results that were within the normal range. The finding of poorer outcomes in treated and control eyes than in no ROP eyes is apparently related to the severe ROP that affected both treated and control eyes during and after the neonatal period. However, other demographic characteristics of the population may also have contributed to the poorer-than-normal CS outcomes observed in many treated and control eyes. For example, patients who had severe ROP had a lower mean birth weight(800 g; SD, 167 g) than patients in the no ROP group (1061 g; SD, 141 g). Also, patients in the severe ROP group were enrolled at the 23 study centers located throughout the United States, whereas all patients in the no ROP group were at the Philadelphia participating center.
In summary, results of the present study of CS performed at age 10 years confirm the beneficial effect of cryotherapy on visual function of eyes that developed severe ROP during the neonatal period. Furthermore, among eyes in which sight was preserved, there was no evidence of a detrimental effect of cryotherapy on CS. Comparison of CS results with VA results obtained at the same study examination suggested that sensitivity to low-range and mid-range spatial frequencies may be preserved better than sensitivity to high spatial frequencies in both treated eyes and control eyes. Regardless of treatment status, CS is substantially poorer in eyes with a history of severe ROP in the neonatal period than in eyes of preterm children who did not develop ROP, indicating that eyes that have severe ROP in the neonatal period are at risk for reduced pattern vision, as well as for reduced VA.
Accepted for publication June 22, 2000.
The CRYO-ROP study is supported by cooperative agreement U10 EY05874 from the National Eye Institute, National Institutes of Health, US Department of Health and Human Services, Bethesda, Md.
We thank the following members of the writing committee for this article: Velma Dobson, PhD (Chair), Graham E. Quinn, MD, Robert J. Hardy, PhD, Dale L. Phelps, MD, Earl A. Palmer, MD, C. Gail Summers, MD, and Betty Tung, MS.
Reprints: Earl A. Palmer, MD, CRYO-ROP Headquarters, Casey Eye Institute, Oregon Health Sciences University, 3375 SW Terwilliger Blvd, Portland, OR 97201-4197.