Does Cryotherapy Affect Refractive Error?

Does Cryotherapy Affect Refractive Error? Results from Treated versus Control in The Cryotherapy for Retinopathy of Prematurity Trial Graham E. Quinn, MD, 1 Velma Dobson, PhD, 2 R. Michael Siatkowski, MD, 3 Robert J. Hardy, PhD, 4 Jane Kivlin, MD, 5 Earl A. Palmer, MD, 6 Dale L. Phelps, MD, 7 Michael X. Repka, MD, 8 C. Gail Summers, MD, 9 Betty Tung, MS, 4 Wenyaw Chan, PhD, 4 for the Cryotherapy for Retinopathy of Prematurity Cooperative Group Purpose: To evaluate the effect of cryotherapy on refractive error status between ages 3 months and 10 years in children with birth weights of less than 1251 g in whom severe retinopathy of prematurity (ROP) developed in one or both eyes during the neonatal period. Design: Randomized clinical trial. Participants: Two hundred ninety-one children in whom severe ROP developed during the neonatal period. Intervention: Cryotherapy for ROP. Main Outcome Measures: Cycloplegic Refraction Methods: The children underwent repeated follow-up eye examinations, including cycloplegic retinoscopy, between 3 months and 10 years after term due date. Refractive error data from all eyes that were randomized to cryotherapy were compared with data from all eyes that were randomized to serve as controls. Refractive error data were also compared for a subset of children who had both a treated and a control eye that could be refracted. Results: At all ages, the proportion of treated eyes that were unable to be refracted because of retinal detachment, media opacity, or pupillary miosis was approximately half the proportion of the control eyes that were unable to be refracted. When data from all eyes that could be refracted were considered, the distribution of refractive errors between fewer than 8 diopters (D) of myopia and more than 8Dofhyperopia was similar for treated and control eyes at all ages. The proportion of eyes with 8 D or more of myopia was much higher in treated than in control eyes at all ages after 3 months. In the subset of children who had a treated eye and a control eye that could be refracted, distributions of refractive errors in treated versus control eyes were similar at most ages. Conclusions: In both treated and control eyes, there was an increase in the prevalence of high myopia between 3 and 12 months of age. Between 12 months and 10 years of age, there was little change in distribution of refractive error in treated or control eyes. The higher prevalence of myopia of 8Dormoreintreated eyes, as compared with control eyes, may be the result of cryotherapy s preservation of retinal structure in eyes that, in the absence of cryotherapy, would have progressed to retinal detachment. Ophthalmology 2001;108:343 347 2001 by the American Academy of Ophthalmology. Originally received: December 30, 1999. Accepted: September 12, 2000. Manuscript no. 99855. 1 Division of Pediatric Ophthalmology, The Children s Hospital of Philadelphia and Scheie Eye Institute, University of Pennsylvania Health System, Philadelphia, Pennsylvania. 2 Departments of Ophthalmology and Psychology, University of Arizona, Tucson, Arizona. 3 Department of Ophthalmology, University of Oklahoma, Oklahoma City, Oklahoma. 4 University of Texas Health Science Center at Houston, School of Public Health, Houston, Texas. 5 Department of Ophthalmology, Medical College of Wisconsin, Milwaukee, Wisconsin. 6 Casey Eye Institute, Oregon Health Sciences University, Portland, Oregon. 7 Departments of Pediatrics and Ophthalmology, University of Rochester, Rochester, New York. 8 Wilmer Eye Institute, The Johns Hopkins University School of Medicine, Baltimore, Maryland. 9 Departments of Ophthalmology and Pediatrics, University of Minnesota, Minneapolis, Minnesota. Presented in part at the annual meeting of The Association for Research in Vision and Ophthalmology, May 1998, Fort Lauderdale, Florida. Supported by the National Eye Institute, Bethesda, Maryland (cooperative agreement no. U10 EY05874). A listing of cooperative group participants can be found in Archives of Ophthalmology 1996;114:417 24. Correspondence to Graham E. Quinn, MD, Division of Pediatric Ophthalmology, The Children s Hospital of Philadelphia, One Children s Center, Philadelphia, PA 19104. E-mail: quinn@email.chop.edu. Reprint requests to CRYO-ROP Study Headquarters, Oregon Health Sciences University, Department of Ophthalmology, L467, 3375 SW Terwilliger Boulevard, Portland, OR 97201-4197 (Earl A. Palmer, MD). 2001 by the American Academy of Ophthalmology ISSN 0161-6420/00/$ see front matter Published by Elsevier Science Inc. PII S0161-6420(00)00527-3 343

Ophthalmology Volume 108, Number 2, February 2001 Table 1. Spherical Equivalent Refractive Error at 1 and 3.5 Years after Randomization 11,13 1-year Examination Treated Control 3.5-year Examination Treated Control Hyperopia 6 D 7 (3.1) 9 (4.1) 4 (1.9) 5 (2.3) 2 Dto 6 D 13 (5.8) 5 (2.3) 13 (6.0) 9 (4.2) Myopia 2 Dto hyperopia 2 D 37 (16.4) 36 (16.4) 29 (13.5) 32 (15.0) 2 Dto 6 D 51 (22.7) 47 (21.4) 44 (20.5) 33 (15.5) 6 D 75 (33.3) 50 (22.7) 81 (37.7) 58 (27.2) Unable to refract 42 (18.7) 73 (33.2) 44 (20.5) 76 (35.7) Total 225 (100) 220 (100) 215 (100) 213 (100) D diopters cryotherapy results in an increase in the prevalence of high myopia; however, this conclusion would ignore the fact that there were almost twice as many control eyes as treated eyes that were unable to be refracted because of retinal detachment. It is reasonable, therefore, to hypothesize that many of the eyes in the control group would have been highly myopic if ocular structure had been preserved as it was in many eyes in the cryotherapy group. The purpose of this report was to explore this hypothesis by comparing the changes in refractive error, including spherical equivalent and astigmatic error, between ages 3 months and 10 years in eyes in the CRYO-ROP trial that underwent cryotherapy for severe ROP and eyes with severe ROP that served as controls. An important feature of this report is that it provides the distribution of refractive error at the various test ages, whereas previous reports have only provided mean and range of refractive error. Previous reports, including those from the multicenter Cryotherapy for Retinopathy of Prematurity (CRYO-ROP) study, have shown that increasing severity of retinopathy of prematurity (ROP) is associated with an increased prevalence of myopia. 1 4 Recently, it has been suggested that use of cryotherapy for severe ROP further increases the severity of myopia in these eyes. 4 9 One explanation for the high myopia seen in eyes after cryotherapy is that the surgical intervention results in a change in the pattern of refractive error development in eyes with severe ROP. Another explanation is that cryotherapy preserves retinal structure in potentially highly myopic eyes that, in the absence of surgical intervention, would have progressed to retinal detachment. To examine these two possible explanations, a population that includes both treated and untreated eyes with severe ROP is needed. Such a population is available in the CRYO-ROP trial, which included eyes with severe ROP that were randomly assigned to receive cryotherapy or no cryotherapy during the neonatal period. 10 14 In the CRYO-ROP trial, eyes of 291 infants with severe ROP were randomly assigned to receive cryotherapy or to serve as controls. 10 Approximately 80% of the infants had bilateral severe (threshold) ROP, and one eye received cryotherapy within 72 hours of the diagnosis. In the remaining 20% of infants, severe ROP developed in only one eye, and that eye was randomly assigned to cryotherapy or control status. All study participants received follow-up examinations, including cycloplegic refractions, at 1- to 1.5-year intervals through age 10 years. Results from treated and control eyes at the 1-year 11 and 3.5-year 13 examinations indicated that the percentage of eyes with spherical equivalent refractive errors between 2 diopters (D) and high hyperopia ( 6 D) was approximately equal in treated and control eyes (25.3% vs. 22.8% at 1 year; 21.4% vs. 21.5% at 3.5 years; see Table 1). Slightly more treated than control eyes had myopia between 2 D and 6 D (22.7% vs. 21.4% at 1 year; 20.5% vs. 15.5% at 3.5 years), and substantially more treated than control eyes had myopia of 6 D or more (33.3% vs. 22.7% at 1 year; 37.7% vs. 27.2% at 3.5 years). One may be tempted to conclude from these data that Methods Patients The study population consisted of the 291 children who participated in the randomized trial of cryotherapy for ROP. All were born between January 1, 1986, and November 30, 1987, with birth weights less than 1251 g. During the neonatal period, 240 of the infants developed bilateral severe (threshold) ROP, defined as a minimum of five contiguous or eight cumulative clock hours of stage 3 ROP in zone 1 or zone 2. In these infants, one eye was randomly assigned to undergo cryotherapy and the other eye assigned to serve as a control. The remaining 51 infants developed severe ROP in only one eye, and that eye was randomly assigned to receive cryotherapy or to serve as a control. Informed consent was obtained before entry into the study, prior to randomization if the child developed threshold ROP, and prior to entry into the long-term follow-up phase of the study. Extensive descriptions of the study population have been published elsewhere. 10 15 Eye Examinations Follow-up eye examinations, including cycloplegic retinoscopy, were conducted at 3 months, and at 1, 2, 3.5, 4.5, 5.5, 7, 8, 9, and 10 years after randomization (which coincided approximately with the child s term due date). The cycloplegic agent used was 0.5% cyclopentolate at the 3-month examination and 1% cyclopentolate at later examinations. When cyclopentolate was medically contraindicated in the opinion of the treating ophthalmologist, 1% tropicamide was used. Eye examinations were performed by ophthalmologists who had undergone a training and certification procedure that included quality control comparisons of retinoscopy results. Astigmatic errors were recorded using a plus cylinder convention. Examiners also noted when refraction was not possible (unable to grade) because of retinal detachment, media opacity, or pupillary miosis. Data Analysis Data were examined in terms of spherical equivalent refractive error. Refractive error categories are slightly different from the two previous publications on this topic from the CRYO-ROP study, 11,13 and, therefore, data previously presented for the 1- and 344

Quinn et al Does Cryotherapy Affect Refractive Error? Figure 1. Distribution of spherical equivalent refractive error for all treated and control eyes at 3 months (248 treated, 248 control eyes), 1 year (199 treated, 184 control eyes), 2 years (181 treated, 171 control eyes), 3.5 years (182 treated, 175 control eyes), 4.5 years (170 treated, 160 control eyes), 5.5 years (170 treated, 165 control eyes), 7 years (160 treated, 150 control eyes), 8 years (160 treated, 151 control eyes), 9 years (157 treated, 144 control eyes), 10 years (167 treated, 140 control eyes). UG unable to be refracted. 3.5-year follow-up ages are again presented in this paper. The figures presented in this manuscript reflect this new grouping for all years. For eyes that underwent vitrectomy, scleral buckling, iridectomy, glaucoma procedure, or cataract surgery, refractive error results were excluded from all examinations subsequent to the time of the operative procedure. Statistical analyses were conducted on data from the subset of patients who had both a treated eye and a control eye that could be refracted. For each patient, the difference in refractive errors for treated and control eyes was calculated for each time at which the patient was examined. A mixed linear model was used to test whether the means of these differences were different from zero. This technique takes into account the correlated nature of outcomes observed on the same subjects at successive examinations. The correlation of outcomes between two time points was assumed to depend on the time between examinations. Results Spherical Equivalent Refractive Errors Figure 1 shows the distribution of spherical equivalent refractive errors for all treated (dark bars) and all control (hatched bars) eyes Figure 2. Distribution of plus cylinder astigmatic error for all treated and control eyes at 3 months, 1, 2, 3.5, 4.5, 5.5, 7, 8, 9, and 10 years. (Number of eyes as in Figure 1.) UG unable to be refracted. at the 10 test ages. At all ages, the proportion of eyes that could not be refracted is approximately twice as great in the control group as the treated group, as reported previously. 11,13 Among eyes that could be refracted at 3 months, the distribution of refractive errors in treated eyes was similar to that for control eyes. At all other test ages, the distribution of refractive errors from 8 D of myopia to high hyperopia in treated eyes was similar to that in control eyes. However, the proportion of eyes with 8 D or more of myopia was much higher in treated eyes than control eyes at all test ages after 3 months. At all ages, the proportion of eyes that were unable to be refracted because of retinal detachment, media opacity, or pupillary miosis is much higher in eyes that did not undergo cryotherapy than in treated eyes. Astigmatic Refractive Errors Figure 2 shows the distribution of astigmatic errors for all treated (dark bars) and all control (hatched bars) eyes at the 10 examination ages. There is a tendency for a greater proportion of treated eyes rather than of control eyes to have astigmatic errors of 1Dor more. Within-subjects Comparison of Treated and Control : Spherical Equivalent and Astigmatic Refractive Errors To this point, we have presented data from all eyes in which threshold ROP developed and that were randomized to cryotherapy or to serve as a control. Because there is a high correlation 345

Ophthalmology Volume 108, Number 2, February 2001 treated and control eyes of patients who could be refracted in both eyes at the 10 examination ages (figure not provided). As previously shown in Figure 2, there was a tendency at older test ages for treated eyes to show more astigmatic errors than control eyes. Discussion Figure 3. Distribution of spherical equivalent refractive error for treated and control eyes of patients who could be refracted in each eye at 3 months (n 125), 1 year (n 117), 2 years (n 110), 3.5 years (n 109), 4.5 years (n 103), 5.5 years (n 103), 7 years (n 89), 8 years (n 88), 9 years (n 89), and 10 years (n 90). between refractive error of fellow eyes, 16 these results may be confounded by the inclusion of two eyes from some patients and one eye from other patients. Therefore, we examined the data from the subset of patients who had bilateral threshold ROP (one treated eye and one control eye) and in whom both eyes provided refractive error data. Figure 3 shows the distribution of spherical equivalent refractive errors at the 10 examination ages for treated (dark bars) and control (hatched bars) eyes for patients in whom both eyes could be refracted. The number of patients providing data at each age (see Fig 3 legend) was substantially less than the total number of patients with bilateral threshold ROP (n 240) because of the large number of eyes that had retinal detachment and could not be refracted. The distribution of refractive error at 3 months and 1 year is similar in treated versus control eyes. At later test ages, there was a tendency for more treated eyes than control eyes to be in the 8 D of myopia category. Comparison of the treated and control refractive error distributions across the serial examinations was conducted using a linear mixed model, which takes into account the correlation of measurements made in the same patients at different ages. The mean differences in refractive errors between treated and control eyes were not significant (P 0.09; 95% confidence interval, 1.17, 0.09). In addition, we conducted an analysis to determine if there was a trend across age in differences in spherical equivalent refractive errors between treated and control eyes. The result was not significant; that is, the slope of the differences over time was not significantly different from zero (P 0.08). The distribution of astigmatic errors was examined for the When evaluating any therapeutic method, the clinician must consider possible adverse results of the treatment. It has been suggested that cryotherapy increases the likelihood of high degrees of myopia in eyes with severe ROP. 4 The comparison of refractive errors in all eyes that underwent cryotherapy and all eyes that served as controls is presented in this paper. Figure 1 indicates that, at all 10 follow-up ages, the distribution of refractive error in the range from less than 8 D of myopia to more than 8 D of hyperopia was similar in treated and control eyes. This suggests that the development of refractive error within this range is unaffected by cryotherapy. In contrast, at all ages after 3 months, there was a higher percentage of treated than control eyes with at least 8 D of myopia (e.g., 38.9% [65/167] of treated eyes vs. 29.3% [41/140] control eyes at age 10 years). Because the distributions of refractive error for treated and control eyes were similar in refractive error categories between less than 8 D of myopia and more than 8 D of hyperopia, the greater proportion of treated eyes in the 8 D of myopia category is likely the result of a reduced proportion of treated than control eyes in the remaining category (unable to grade). This suggests that in at least some eyes in which cryotherapy preserves ocular structure, high myopia develops instead of retinal detachment and unsuitability for refracting. We also examined the possibility that high degrees of astigmatism in the group of control eyes could account for the lower prevalence of very high myopia in control eyes. For instance, because myopia was reported in terms of spherical equivalent, different amounts of plus cylinder astigmatism in the treated and control groups could lead to an apparent difference in the amount of myopia between groups, even though the amount of myopia in the most myopic meridian may be equivalent between groups. As shown in Figure 2, this possibility is not supported by the data, because there was a tendency for the astigmatism to be greater in treated eyes than in control eyes. This would reduce, not increase, any difference in the prevalence of myopia between treated and control eyes. To evaluate the possibility that differences in patient population produced the refractive error differences observed between treated and control eyes in Figure 1, we compared the distribution of refractive error in treated versus control eyes in the subset of patients with bilateral severe ROP in whom both eyes (one treated, one control) could be refracted. The results (Fig 3) showed that there were no significant differences in the distribution of refractive error between treated and control eyes across the 10 examination ages and there was no trend for the distributions to change with time. This suggests that cryotherapy does not influence refractive error status in eyes with severe ROP. 346

Quinn et al Does Cryotherapy Affect Refractive Error? Previous analyses of data from the CRYO-ROP study population indicate that the higher the severity of acutephase ROP and the higher the severity of retinal residua, the more likely that high myopia will develop in the eye. 1,3 The data presented in Figure 1 provide an additional piece of information concerning which infants are at risk for the development of high myopia. That is, some of the infants in whom high myopia developed were those whose ROP was so severe that, without cryotherapy, retinal detachment occurred (i.e., control eyes that would presumably have been highly myopic if retinal detachment had not occurred). Unfortunately, data on other systemic factors that may contribute to the development of high myopia, for example, duration of oxygen exposure or occurrence of central nervous system insult, were not collected in the CRYO-ROP study. Furthermore, detailed biometric measurements that may yield information concerning the ocular components that may contribute to the development of high myopia in eyes with severe acute-phase ROP were not possible in this multicenter study. Recent case series and randomized studies with small numbers of subjects have reported a lower incidence of high myopia in eyes with severe ROP treated with laser photocoagulation than in cryotherapy-treated eyes. 1 5 Laws et al 8 and Holmström et al 4 suggested that the higher incidence of high myopia in cryotherapy-treated eyes compared with laser-treated eyes was the result of a detrimental effect of cryosurgery. The results of the present paper cannot address the comparison of eyes that underwent cryotherapy versus eyes that underwent laser treatment for severe ROP. This would require a randomized trial of cryotherapy versus laser photocoagulation in severe ROP. However, the results of the present paper do suggest that cryotherapy does not change refractive error status in eyes with severe ROP. Rather, cryotherapy prevents retinal detachment and allows refractive error assessment in some eyes, and it is these eyes in which high myopia is likely to be manifest. References 1. Quinn GE, Dobson V, Repka MX, et al. Development of myopia in infants with birth weights of less than 1251 g. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Ophthalmology 1992;99:329 40. 2. Fledelius HC. Pre-term delivery and subsequent ocular development. A 7-10 year follow-up of children screened 1982 84 for ROP. 3) Refraction. Myopia of prematurity. Acta Ophthalmol Scand 1996;74:297 300. 3. Quinn GE, Dobson V, Kivlin J, et al. Prevalence of myopia between 3 months and 5-1/2 years in preterm infants with and without retinopathy of prematurity. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Ophthalmology 1998;105:1292 300. 4. Holmström M, el Azazi M, Kugelberg U. Ophthalmological long term follow up of preterm infants: a population based, prospective study of the refraction and its development. Br J Ophthalmol 1998;82:1265 71. 5. Knight-Nanan DM, O Keefe M. Refractive outcome in eyes with retinopathy of prematurity treated with cryotherapy or diode laser: 3 year follow up. Br J Ophthalmol 1996;80:998 1001. 6. Algawi K, Goggin M, O Keefe M. Refractive outcome following diode laser versus cryotherapy for eyes with retinopathy of prematurity. Br J Ophthalmol 1994;78:612 4. 7. White JE, Repka MX. Randomized comparison of diode laser photocoagulation versus cryotherapy for threshold retinopathy of prematurity: 3-year outcome. J Pediatr Ophthalmol Strabismus 1997;34:83 7. 8. Laws F, Laws D, Clark D. Cryotherapy and laser treatment for acute retinopathy of prematurity: refractive outcomes, a longitudinal study. Br J Ophthalmol 1997;81:12 5. 9. Connolly BP, McNamara JA, Sharma S, et al. A comparison of laser photocoagulation with trans-scleral cryotherapy in treatment of threshold retinopathy of prematurity. Ophthalmology 1998;105:1628 31. 10. Multicenter trial of cryotherapy for retinopathy of prematurity. Preliminary results. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1988;106:471 9. 11. Multicenter trial of cryotherapy for retinopathy of prematurity. One-year outcome structure and function. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1990;108:1408 16. 12. Multicenter trial of cryotherapy for retinopathy of prematurity. Three-month outcome. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1990;108: 195 204. 13. Multicenter trial of cryotherapy for retinopathy of prematurity. 3 1/2-year outcome structure and function. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1993;111:339 44. 14. Multicenter trial of cryotherapy for retinopathy of prematurity. Snellen visual acuity and structural outcome at 5 1/2 years after randomization. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1996;114: 417 24. 15. The natural ocular outcome of premature birth and retinopathy. Status at 1 year. The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Arch Ophthalmol 1994;112: 903 12. 16. Gallo JE, Lennerstrand G. A population based study of ocular abnormalities in premature children aged 5 10 years [review]. Am J Ophthalmol 1991;111:539 47. 347