Bremer DL, Palmer EA, Fellows RR, Baker JD, Hardy RJ, Tung B, Rogers GL, for The Cryotherapy for Retinopathy of Prematurity Cooperative Group. Strabismus in Premature Infants in the First Year of Life. Arch Ophthalmol. 1998;116(3):329-333. doi:10.1001/archopht.116.3.329
To present the 3- and 12-month strabismus data from 3030 premature infants with birth weights less than 1251 g enrolled in the Multicenter Trial of Cryotherapy for Retinopathy of Prematurity.
Data from the 3- and 12-month examinations conducted at 23 regional study centers were tabulated for all infants. The main outcome measure, ocular motility, was compared with baseline demographic variables and retinopathy of prematurity severity for the worse eye. Findings at 3 months were compared with the incidence of strabismus at 12 months.
At 3 months, 200 (6.6%) of the 3030 infants were strabismic. In the 2449 infants examined at both time points, 289 (11.8%) were found to have strabismus at 12 months. Retinopathy of prematurity was significant for strabismus at both 3 and 12 months (P<.001). The presence of strabismus at 3 months was found to be a highly significant predictor of strabismus at 12 months. Anisometropia, abnormal fixation, and unfavorable retinal structure also were significant predictors of strabismus at 1 year. The total prevalence of strabismus in the first year of life was 14.7%.
The presence of acute-phase retinopathy of prematurity places the premature infant at increased risk for strabismus.
THE GREATEST fear for parents and physicians who deal with retinopathy of prematurity (ROP) is retinal detachment and blindness. Fortunately, only a small minority of premature infants each year suffer the severe cicatricial complications that lead to visual impairment.1 However, many premature infants experience less severe ROP that resolves spontaneously, and strabismus is a significant concern in these infants.2- 7 Strabismus causes abnormal binocular vision development and can lead to amblyopia. The cosmetic effect of misaligned eyes can have an adverse effect on the psychosocial development of the child.8 Strabismus also creates a potential need for expensive surgical remedies.
The Multicenter Trial of Cryotherapy for Retinopathy of Prematurity (CRYO-ROP) collected motility data prospectively on thousands of premature infants with birth weights less than 1251 g. The purposes of this article are (1) to determine the overall prevalence of strabismus in the first year of life among premature infants enrolled in this study; (2) to analyze demographic factors and characteristics of ROP that predict strabismus at 3 or 12 months of age; and (3) to examine factors present at 3 months (including fixation behavior, retinal structure, anisometropia, neurological development, and strabismus) that predict strabismus at 12 months.
The study population consisted of 4099 infants with birth weights less than 1251 g. The CRYO-ROP study obtained informed consent from parents or guardians and enrolled these infants at 23 participating centers between January 1, 1986, and November 30, 1987. For details concerning the study cohort, the examination and randomization of patients, the International Classification of ROP, and the cryotherapy trial, see previous publications.9- 12 Subsequent to serial examinations for acute-phase ROP, the infants were seen at 3 and 12 months of corrected age. These follow-up examinations included assessment of fixation behavior, ocular motility, anterior segment, refractive error, and fundus findings. The examinations were carried out according to a protocol detailed in the CRYO-ROP Manual of Procedures.13
Fixation behavior was tested binocularly, then with each eye occluded separately. If the infant made eye contact or looked toward a moving face or toy, his or her fixation was judged to be normal. Smooth pursuit movements were not required. If the infant consistently objected to the covering of the eye, the uncovered eye was marked "abnormal." A sleeping or extremely uncooperative infant was judged "unable to assess."
At the 3- and 12-month examinations, strabismus evaluation was summarized only as "normal alignment and motility," "strabismus present," or "unable to assess." Because ocular motility status was not a primary outcome of the study, those infants whose alignment could not be assessed at the regularly scheduled examination were not brought back solely for that purpose. Refractive error was measured at both outcome visits following cycloplegia. Anisometropia was defined as a difference of greater than 2 diopters (spherical equivalent).
After a careful examination of the fundus, the ophthalmologist provided a summary grade of the structural outcome of the retina for each eye. These grades were used to determine favorable or unfavorable outcome. Unfavorable outcome was defined as (1) retinal fold involving the macula; (2) retrolental opacity encroaching on the pupil and/or partial retinal detachment; or (3) total retinal detachment or total pupillary occlusion by fibrovascular membrane.
An abbreviated developmental screening examination was also done. At 3 months, an attempt was made to elicit a response smile, either with direct frontal eye contact or, in the case of a blind infant, by voice alone. The infant's caretaker was also asked whether a smile could usually be elicited at home. If the infant failed both items (a smile could not be elicited at home or during the examination), the infant was considered to have delayed development for the purposes of this study. These developmental screening items were obtained from the Revised Denver Developmental Screening questionnaire.14
For the purposes of this report, each infant's ROP severity was based on the worse eye for that infant. It was determined by the highest stage of ROP and the lowest zone in which ROP occurred. Because there were few infants with ROP in zone I, all stages of ROP in zone I are combined for statistical analysis. Since most ROP in zone III was stage 1, all ROP in zone III was analyzed together.
Of the 4099 infants enrolled in the study, 187 died before age 3 months. Of the remainng infants, 3155 (80.6% of the survivors) were examined at 3 months of corrected age. Motility status could be observed in 3030 infants, of whom 200 (6.6%) were reported to be strabismic. The remaining 125 infants' motility status could not be determined at the 3-month examination.
Table 1 indicates the incidence of strabismus at 3 months by demographic variables (birth weight, gestational age, race, sex, inborn or outborn [born in the hospital or transferred to a study nursery], and single or multiple births) and ROP severity. The highest incidence of strabismus was noted in infants with any ROP in zone I. With higher stages of ROP in zone II, the incidence of strabismus increased accordingly. The incidence of strabismus increased as birth weight or gestational age decreased when analyzed by these variables alone.
Multiple logistic regression analysis was used to determine the significance of these variables for strabismus at 3 months of corrected age (Table 2). When the confounding effects of each of the independent factors on the development of strabismus were adjusted by this analysis, demographic variables (including birth weight and gestational age) were not found to have made a significant contribution to the risk for strabismus at 3 months. Retinopathy of prematurity was found to be a significant predictor of strabismus at 3 months (P<.001) when the group of infants who had any ROP was compared with those who had no ROP. Subgroup analyses showed an increased risk for strabismus with more severe ROP. Although black race has been shown to be a relative protective factor against the development of ROP,14 there was no racial influence on the incidence of strabismus.
Of the 4099 infants enrolled in the study, 2449 (59.7%) had their ocular motility evaluated at both 3 and 12 months after their full-term due dates. Of this group, 11.8% were found to have strabismus at their 12-month examination. Data collected on this cohort were used to compare the incidence of strabismus at 12 months with demographic variables and severity of ROP (Table 3). The incidence of strabismus increased as the zone was more posterior and the stage of ROP worsened.
This cohort of infants evaluated at both 3 and 12 months was used to compare the incidence of strabismus at 12 months with fixation behavior, anisometropia, retinal structure, neurological development, and strabismus status at 3 months. While the incidence of strabismus at 12 months was 9% for infants who had normal fixation in each eye at 3 months, it was 63.9% for infants with abnormal fixation in 1 eye at 3 months (Table 4). The incidence of strabismus at 12 months was 43.1% in patients who had anisometropia at 3 months, but only 9.5% in patients without anisometropia at that age. If the retinal structure at 3 months was favorable in both eyes, the incidence of strabismus at 12 months was 10.3%, while it was 68.2% in infants with unfavorable retinal structure in both eyes at 3 months.
Multiple logistic regression analysis was used to determine the significance of these variables for strabismus at 12 months of corrected age (Table 5). The presence of acute-phase ROP was found to be a significant predictor of strabismus at 12 months (P<.001) when all ROP was compared with no ROP. Subgroup analysis showed all zones and stages of ROP were significantly correlated except for zone I disease.
The presence of strabismus at 3 months was found to be a highly significant predictor of strabismus at 12 months (P<.001). When strabismus was present at 3 months, it was almost 11 times more likely that the infant would be strabismic at 1 year.
Overall, both abnormal fixation behavior at 3 months and unfavorable retinal structure at 3 months were significant predictors of strabismus at 12 months. Subgroup analysis showed that although abnormal fixation in both eyes was a significant predictor of strabismus (P<.005) compared with bilaterally normal-fixating eyes, abnormal fixation in 1 eye and normal fixation in the other eye was not. By contrast, abnormal retinal structure in only 1 eye was significant for strabismus (P<.01) compared with a bilaterally favorable structure, whereas abnormal retinal structure in both eyes was not.
The presence of anisometropia and delayed development at 3 months were also significant for predicting strabismus at 12 months. Anisometropia was a significant contributing factor at the P<.001 level; delayed development was significant at the P<.01 level.
Of the 2449 infants who had their motility evaluated at 3 and 12 months of age, 155 (6.3%) were reported to be strabismic at 3 months. At 12 months of age, 82 of those 155 infants were still reported to be strabismic and had not undergone any therapy for their strabismus, 57 were no longer strabismic without receiving any treatment, and 16 had been treated in the interim. Between 3 and 12 months of age, another 206 infants (8.4%) became strabismic for the first time. Therefore, the total prevalence for strabismus in the first year of life for these 2449 infants was 361 (14.7%).
This report presents prospectively collected ocular alignment data on thousands of low-birth-weight infants from the CRYO-ROP study. When compared with full-term infants,15 premature infants in this study have a higher rate of strabismus. These results confirm what other authors have reported (Table 6).2- 7 This study also shows that the presence of acute-phase ROP is a strong predictor for strabismus in the first year of life (Table 2 and Table 5). As the severity of ROP increases, the incidence of strabismus increases. Birth weight and gestational age were not found to be significant risk factors within this restricted group of premature infants weighing less than 1251 g. Any appearance to the contrary is due to the previously reported association of ROP with these factors.1 Strabismus outcome is dictated by events affecting the retina during the acute phase of ROP, and not by birth weight or gestational age.
Most of the children in this study who reached threshold ROP were treated with cryotherapy to 1 eye per study protocol. The CRYO-ROP study was designed to evaluate the efficacy of cryotherapy in treating threshold ROP, and not the effect of cryotherapy on the development of strabismus. Therefore the role, if any, of cryotherapy in the development of strabismus in this group of children could not be determined.
The CRYO-ROP protocol required examinations at 3 and 12 months of age. Characteristics of the 3-month evaluation that predicted strabismus at 12 months include anisometropia, unfavorable retinal structure, abnormal fixation, poor neurological development, and strabismus. These results support the value of an examination at 3 months of corrected age for infants who had ROP.
The fact that developmental delay or strabismus at 3 months of age predicts strabismus at 12 months could be expected, as it is known that children with neurological handicaps have an increased incidence of strabismus. While the developmental evaluation of study infants at 3 months was rudimentary (presence or absence of response smile), future analysis of the results of later CRYO-ROP evaluations (which included annual eye examinations and developmental screenings at ages 12-66 months) may provide further insight into the relationship between developmental delay and strabismus.
The large cohort, prospective data collection, and good follow-up rate should make the statistics in this report more definitive than previous smaller studies. In this population of very low-birth-weight infants who had their ocular motility evaluated at 3 months of corrected age, 6.6% were reported to be strabismic. The incidence of strabismus for all infants examined at 12 months was 12.8%, as previously reported.8 For infants examined at both 3 and 12 months of age, 6.3% (155/2449) were strabismic at 3 months of age and 8.4% (206/2449) became strabismic for the first time between age 3 and 12 months. The total prevalence of strabismus in the first year was 14.7%.
Results of the 3-month examination may be compared with data from the 12-month examination to consider the natural history of strabismus in these infants. One hundred fifty-five infants who were strabismic at 3 months were also seen for follow-up at 12 months. Of these, 57 (36.8%) had no strabismus at 12 months without receiving any intervention. Given the likelihood for spontaneous improvement in many of these children, a conservative approach to surgery at an early age in the preterm infant appears warranted.
Accepted for publication November 6, 1997.
Corresponding author: Don L. Bremer, MD, Children's Hospital Outpatient Care Center, Suite 4C, 555 S 18th St, Columbus, OH 43205.