Risk factors

Which baby develops ROP, and the visual outcome, is greatly influenced by the standard of neonatal care received. In countries where high-quality neonatal care is available, events such as hyperoxia, hypoxia, hypercarbia, blood transfusions, hyperglycaemia and sepsis can all contribute to the development of ROP, but by far the greatest predictors are indices of immaturity—low gestational age (GA) and low birth weight (BW).7 In countries with high-quality neonatal care, sight-threatening ROP is largely confined to babies with BW <1000 g and is very uncommon in babies >1250 g. In settings where the quality of neonatal care is more variable, the population of babies who develop severe ROP differs, with larger, more mature babies also being affected; indeed, severe ROP is being reported in babies over 34 weeks GA with BWs exceeding 2500 g.8 ,9 This is largely due to prolonged, injudicious use of supplemental oxygen, frequently not blended or adequately monitored, which predominates over all other risk factors. However, many factors compound and contribute. These include—depending on the standard of care—all the aforementioned, suboptimal management of labour, poor nutrition and the absence of pain management.10

ROP screening and treatment

The available evidence suggests that >50% babies with BW <1000 g develop some degree of acute ROP.11 A recent systematic review and meta-analysis estimated that in 2010 around 184 700 babies developed any ROP worldwide with 53 500 of these developing Type I disease (table 1).6 The majority (77%) of these in babies were in low-income and middle-income settings, with 27 000 (15%) being only moderately preterm (born at >32 weeks gestation). These are likely to be underestimates as the study focused on babies receiving neonatal intensive care as data were not available for babies in lower level facilities who may also receive poorly monitored supplemental oxygen.

In a small proportion of babies, ROP does not resolve but progresses to become severe, requiring treatment (box 2). Laser is currently the preferred modality.3 Antivascular endothelial growth factor agents (ranibizumab or bevacizumab), despite a paucity of clinical research evidence, are being increasingly used, either as a primary treatment or as a rescue therapy after laser failure. However, the ocular and systemic safety and long-term efficacy of these agents are not yet known.12 A high proportion of eyes with untreated severe ROP (approximately 60%) and a proportion of treated eyes can develop scarring and distortion of the retina, or retinal detachment, with irreversible vision loss. The role of vitrectomy for the management of Stages 4 and 5 is highly controversial, relating more to patterns of clinical practice than the state of the eye. Some eyes with Stage 4 have limited vision following surgery; the results for Stage 5 are very poor.13 Screening programmes are designed to detect those babies most at risk of progressing to sight-threatening disease. In the UK, for instance, it is recommended that all babies <31 weeks GA or <1251 g BW must be screened, and to catch the important outliers, babies ≥31 and <32 weeks GA or ≥1251 to <1501 g BW should also be screened.14 In low-income or middle-income countries, there is a body of evidence that bigger, more mature babies also develop severe ROP, and in many of these countries the screening criteria are much wider.9

In countries and neonatal units where screening is undertaken, most babies with severe disease are treated. Despite this, around 12% of all babies with severe ROP will be visually impaired, with 5% being severely impaired. In low-income and middle-income settings, the provision of screening and treatment, and consequently visual outcomes, are much more variable. Middle-income countries currently have the largest estimated incidence accounting for three-quarters of all cases of ROP worldwide (table 1). Access to screening and treatment for ROP has lagged behind the development of neonatal intensive care unit (NICU) care, with only an estimated 40% of intensive care neonatal units in middle-income settings having ROP programmes. Much higher rates of visual impairment are therefore seen in these settings compared with high-income settings, with over 80% of visual impairment occurring in middle-income countries (table 1). Access to neonatal intensive care is still very limited in the predominantly low-income settings of South Asia and sub-Saharan Africa, and consequently few babies survive neonatal intensive care to develop ROP (13% of cases of ROP worldwide are from these regions compared with 59% of preterm births).6

Effect of ROP on the eye and its shape

Emmetropisation is the process of refractive development whereby the ocular components grow in a coordinated fashion to create an eye without refractive error. This process, which is fine-tuned by vision, is jeopardised in the preterm baby for two reasons: first, the effect of being born preterm and second, as a consequence of ROP.

Early exteriorisation due to preterm birth shortens the intrauterine period and removes the fetus from a protective environment ideally suited to promote growth and provides the optimal level of stimulation. Unsurprisingly, the visual system of the preterm baby can be affected by removal from this milieu as well as by exposure to a very different biological and physical environment. Fledelius demonstrated over 30 years ago that the eyes of children born prematurely but who did not have ROP did not grow normally.18 He showed that structures at the front of the eye, namely the cornea and lens, both of which play a vital role in focusing light onto the retina, were different: corneas of children born preterm are more curved, and the lens is thicker, both of which increase the focusing power of the eye, leading to short-sightedness (myopia). This myopia, which is not due to ROP, is of low degree and is referred to as ‘myopia of prematurity’.18–21 Studies have shown that all individuals born preterm are more prone than their full-term counterparts to all refractive errors—astigmatism (irregular curvature of the cornea which distorts vision), hypermetropia (long-sightedness) and anisometropia (different refractive state between the two eyes).19 ,20

While mild ROP does not contribute additionally to refractive state, following severe acute ROP refractive development is very different, as all of the aforementioned refractive errors occur, but to a greater degree 19 ,21 ,22 and high myopia (>5 dioptres) only occurs following severe ROP. The prevalence of refractive errors varies across studies, as shown in table 2, being highest in the extremely low BW group (ie, BW <1000 g) and following treatment for severe ROP. Myopia attracts the most attention 23 as high degrees are not rare and even a small amount of myopia affects distance visual acuity which requires correction. In contrast, the impact of hypermetropia on visual performance is not linear and its correction is not clearly defined, although there is no doubt that higher degrees do need to be corrected.24

Long debated, it appears that myopia following severe ROP is related to the disease process itself and is not a complication of ROP treatment by either cryotherapy or laser. Myopia associated with severe ROP progresses during the first 6–9 months and to a lesser extent thereafter, becoming relatively stable by around 3 years of age. This is quite different from the mild myopia associated with prematurity which has a later onset and progresses in severity into the teens. It is possible that myopia is less following bevacizumab treatment compared with laser or cryotherapy, a potentially important finding which needs further study.

To summarise, refractive development is perturbed by preterm birth per se and expreterm children have an increased prevalence of all refractive errors, especially low myopia. Mild ROP does not contribute to the refractive state, over and above that which is attributable to preterm birth. Following severe ROP, there is an increase in all refractive errors, including myopia, and frequently high myopia.

Visual acuity

It has long been known that children who were born preterm, but did not have ROP, may have reduced visual acuity compared with children born at full term.18 ,34 ,35 These deficits are subtle with acuity levels usually lying at the lower border of the normal range. The basis of these acuity deficits has hitherto been unknown, but recent advances in optical coherent tomography show that children born preterm can have subtle changes in the detailed anatomy of the central retina. However, these structural abnormalities do not correlate very well with function, which suggests that other processes are also involved.36

The visual field

In the Multicenter Trial of Cryotherapy for Retinopathy of Prematurity, it was reported 37 that at the age of 10 years the visual field was constricted to a similar extent (ie, by 5° to 10°) whether or not the ROP had been treated by peripheral retinal ablation; this was subsequently confirmed by Larsson and coworkers.38 Laser ablative treatment at the earlier stage of Type I prethreshold not only did not cause further constriction as feared, but counterintuitively resulted in slight preservation of the field compared with eyes treated at threshold.39 The level of visual field constriction attributable to ROP does not have a major functional impact and driving is not precluded.

Colour vision and contrast sensitivity

Subtle colour vision deficits have been described in association with preterm birth but are ROP independent.40 Contrast sensitivity is an important visual function and may be adversely affected by preterm birth, neurological abnormalities and severe ROP.

While the clinical measures of visual function are made in isolation, they frequently coexist. The cumulative impact on functional ability is unknown, in particular in a group of children that may have other physical disabilities, and while normal visual acuity may not be a realistic goal, all treatments aim to maximise visual function.