Refraction and keratometry in premature infants

Abnormalities of emmetropisation, resulting most often in myopia, have long been independently associated by numerous authors from nearly every corner of the globe with increasing prematurity, increasing severity of retinopathy of prematurity (ROP), and retinal ablative therapy. This topic has also been the subject of an editorial in the BJO in 1997, the substance of which remains true today.¹ These refractive error abnormalities of prematurity and of ROP have been found to present early in infancy and persist into adulthood. It is not clear whether the mechanism for the development of myopia of prematurity is the same or different from the mechanism for the development of myopia of ROP.¹ It appears that an alteration in the normal growth patterns of the eye is caused by the prematurity, the retinal ablative therapy, or probably both factors, leading to the refractive error. This disruption in growth results in structural abnormalities of parts of the eye which affect refractive development. These structures might include corneal curvature, axial length, lens thickness, and anterior chamber depth.

Most authors of studies involving myopia of prematurity have had only refractive error data available, which makes it impossible to directly determine which ocular structures have had their development altered before term. Furthermore, most studies have dealt with patients examined well beyond the expected due date, after which other environmental and disease factors could affect the observed outcome.^2,3 There has been until recently few biometric data available for these premature patients to explain the refractive error findings, and essentially no such data during infancy. Donzis and coworkers reported a dramatic flattening in the corneal curvature of six premature patients with very steep corneas between 28–34 weeks gestation and term, which they associated with a change in refractive status.⁴ No comparison data to term infants were available, but these data can be used to support the idea that there is substantial, rapid corneal growth throughout this period. Cook and colleagues found, in premature infants without evidence of ROP (mean post-menstrual age at birth 29.4 weeks), or complex pattern of rapid growth between 30 and 50 weeks post-menstrual age. They noted a significant correlation between corneal curvature growth and refractive status resulting in less hypermetropia when compared to published data from term infants.⁵ Cook and her coworkers did not include comparison data for eyes of children delivered at term in their nursery or information concerning eyes with ROP.

We do not know whether this change in corneal curvature measured at term persists and can be held responsible for myopia of prematurity

Snir and coworkers (p 900), in this issue of the BJO, have provided data that begin to fill two gaps in our understanding of the relation of prematurity to refractive error. They provide information about refraction, corneal curvature, and retinal appearance at 40 weeks post-menstrual age. In their design they studied two patient cohorts, one consisting of babies delivered and examined at term and a second group, delivered an average of 7.1 weeks before term (range 26–35 weeks), but who underwent their ocular evaluation at term. These authors determined refractive error with cycloplegic retinoscopy and corneal curvature with automated keratometry. Unfortunately, their design did not include other biometric measures such as axial length or anterior chamber depth. Funduscopy noted only mild ROP (stages 1 or 2) in 87.9% case, probably because very few infants less than 1000 g were included. Snir and colleagues found significantly less hypermetropia, about 1.1 D on average, in the preterm infants compared to the term infants. They also found the corneas to be 1.5 D steeper on average in the premature group. The similarity in magnitude of these two values when compared to term infants suggests that the alteration in the corneal curvature may be responsible for much or all of the observed change in refractive error state at 40 weeks post-menstrual age.

These data add substantial weight to the hypothesis that exposure to the extrauterine environment for a relatively short time before term significantly affects the growth pattern of the anterior segment. The rapidity of the change lends support to the hypothesis suggested by Fielder and colleagues nearly two decades ago.^1,6 They postulated that the reduced temperature of the extrauterine eye leads to slowing of corneal growth, less flattening of the cornea, and thus less hypermetropia. The role of other factors such as light exposure and patterned visual stimulation for a short period before term on refractive development are not well studied in preterm infants. The similarity of the findings of Snir and colleagues for patients with mild ROP to those findings of Cook and her colleagues for patients without ROP suggests that the alterations in normal corneal curvature development before term may not be greatly affected by ROP.

The applicability of this report is limited by the lack of a wider birthweight distribution of premature infants, especially of infants less than 750 g, and its similarly limited distribution of ROP severity, as almost every patient had stage 1 or 2 disease. These deficiencies make it impossible to generalise to smaller infants with greater ROP severity. Finally, the data from this study cannot confirm that corneal curvature is alone responsible for the epidemic of myopia seen among extreme low birthweight infants and those with severe ROP as, here again, insufficient numbers of such patients were included in this study and other biometric factors were not studied. Anterior chamber depth seems to be an important factor,^5,7 and other factors such as lens thickness and axial elongation remain to be clarified by future investigation. A further limitation is the lack of biometric data on this cohort of patients later in the first year of life. Thus, we do not know whether this change in corneal curvature measured at term persists and can be held responsible for myopia of prematurity.²

This study has confirmed that the normal development of corneal curvature is affected by preterm delivery producing less hypermetropia in the setting of mild ROP. The direct cause for this change is unknown. As for the development of myopia of prematurity and myopia of ROP, it is likely that additional modifying factors following delivery, both in the nursery and in the home environment, has a large role in the refractive error outcome measured at 3 months of age and beyond.