12 Retinopathy of Prematurity
First edition chapter originally authored by J. Arch McNamara and Brian P. Connolly
Retinopathy of prematurity (ROP) is a potentially blinding retinal vascular disorder that affects the retina of premature infants. ROP has become more frequent, as advances in neonatal medicine have allowed increased survival of premature infants. Well over half of infants with a birth weight of 700 g 1 and those born at 24 weeks of gestational age (GA) 2 are now able to survive. Neonates who are born very premature and with an extremely low birth weight are at greatest risk for the development of ROP. For example, in the Early Treatment for ROP (ETROP) study, 6,998 infants with birth weight < 1,251 g were screened from 2000 to 2002, and any degree of ROP was found in 68% of these babies. 3 However, ROP developed in 93% of patients with birth weight < 750 g compared to 44% of those with birth weights 1,000 to 1,250 g and in 89% of patients = 27 weeks of GA compared to 14% of those = 32 weeks of GA. 3
12.2 Initial Findings and the Role of Oxygen
The first description of ROP as a disease entity related to prematurity was made by Terry in 1942, and he coined the term retrolental fibroplasia to refer to the late presentation of total retinal detachment with overlying fibrovascular membranes behind the lens. 4 , 5 Within a decade of his initial reports, ROP was the leading reported cause of childhood blindness in industrialized nations. 6 Oxygen exposure was noted as an association with ROP in the early 1950s, 7 and subsequent studies illustrated the role of supplemental oxygen treatment in the development of ROP. 8 Health care providers in neonatal nurseries were encouraged to curtail oxygen therapy, and the incidence of ROP rapidly declined. However, in the 1960s, oxygen deprivation was reported to be associated with increased neonatal mortality from neurological and pulmonary disorders, and this led to more liberal use of oxygen and the resurgence of ROP. 9 The advent of continuous transcutaneous oxygen monitoring allowed health care providers to set goals for oxygenation saturation in an attempt to balance the risks of ROP and mortality, but the target range for oxygenation saturation remains controversial. 10 , 11 , 12 Furthermore, it is now recognized that ROP is a complex and multifactorial disease, rather than a retinopathy due solely to supplemental oxygen exposure.
The overall prevalence of ROP in a given nation is related to socioeconomic status and neonatal intensive care unit (NICU) capacities. Low-income nations have low rates of ROP because the mortality rates for premature infants are high. More severe forms of ROP are seen in high-income nations due to advances in neonatal care that allow the youngest and sickest infants to survive, but often with worse ocular and systemic sequelae. The ETROP study took place in the United States in the 2000s, 3 and while the overall incidence of ROP was similar to the Cryotherapy for Retinopathy of Prematurity (CRYO-ROP) study from the 1980s, there were more patients with posterior disease in the ETROP study, which occurred most often in the youngest and sickest infants.
Middle-income nations are currently experiencing a “third epidemic” of ROP due to increased numbers of NICUs, but many such NICUs lack sophisticated oxygen-monitoring devices. 13 Based on 1993 epidemiological data, the World Health Organization estimated that more than 50,000 children were blinded by ROP throughout the world. 14 Most of these children lived in Latin America and former socialist economies.
12.4 Clinical Features
It is fortunate that the natural history of acute ROP favors spontaneous regression in the majority of cases. Among infants whose ROP regressed spontaneously, Flynn et al 15 reported that ROP lasted for an average of 15 weeks. The earliest sign of regression is vascularization of the retina beyond the fibrovascular ridge. As the arterioles and venules extend into the avascular retina, the caliber and tortuosity of the vessels posterior to the shunt and in the posterior pole diminish.
Those infants who progress to advanced stages of ROP represent a small minority. The most important risk factors are low birth weight and young GA. In the CRYO-ROP study, of 4,099 infants who were screened, 291 (7%) reached threshold disease that required randomization. 16 The mean birth weight of these infants was 800 ± 165 g and the mean GA was 26.3 ± 1.8 weeks.
In the United States, 83% of infants with birth weights less than 1,000 g are at greatest risk for the development of ROP. 3
Retinal vessels normally reach the nasal ora serrata by 36 weeks of gestation and temporally by 40 weeks, although complete remodeled vascularization is achieved 2 to 3 months after term birth. 17 Premature infants are born prior to full vascularization of the retina, and the relatively hyperoxic extrauterine environment and supplemental oxygen cause downregulation of growth factors such as vascular endothelial growth factor (VEGF), which slows down vascular development. 18 Existing vasculature becomes attenuated and may undergo apoptosis. 19 The vaso-obliteration and the increasing metabolic demands of the maturing peripheral retina make the retina hypoxic, resulting in upregulation of growth factors. This leads to pathologic fibrovascular proliferation. The attenuated vessels merge with surrounding mesenchymal cells to form arteriovenous shunts, represented as the demarcation line in stage 1 ROP as described in the next section. 20 The shunt regresses in most cases and peripheral retinal vascularization occurs. However, the neovascular drive may persist in high-risk infants, causing the primitive endothelial cells of the shunt to proliferate through the internal limiting membrane and into the vitreous. 21 This fibrovascular tissue exerts vitreoretinal contractile forces, which can lead to tractional retinal detachment.
12.6 Diagnosis and Classification
In 1984, a classification system for acute ROP, devised by 23 ophthalmologists from 11 countries, was published and widely accepted. 22 The International Classification for Retinopathy of Prematurity (ICROP) systematically defined the location of the disease in the retina as well as the extent of the developing vasculature that was involved. Moreover, this classification system laid the groundwork for controlled randomized clinical trials. Three parameters are used to characterize the amount of disease that is present: zone, stage, and (the presence or absence of) “plus” disease. The extent of disease is further specified as number of clock hours (Fig. 12-1).
To define the anatomic location of disease in the retina, the retina is divided into three zones centered around the optic disc. Zone I is a circle with a radius defined as twice the distance between the center of the optic nerve and the foveal center. Zone II extends from the anterior edge of zone I to a circle whose radius is equal to the distance between the center of the optic disc and the nasal ora serrata. Finally, zone III represents the temporal crescent of the retina anterior to zone II.
Fortunately, acute ROP regresses in most eyes without treatment.
When an abnormal vascular response is present, it is the basis for the second parameter, the stage of disease. When there is no demarcation line between vascularized and nonvascularized retina, immature vasculature (but not acute ROP) is said to be present (Fig. 12-2a). Four stages were included in the original classification, and they are described as follows. Stage 1 (demarcation line) is defined by the presence of a thin but definite structure that separates the avascular retina anteriorly from vascularized retina posteriorly (Fig. 12-2b). Abnormal branching or arcading of vessels is seen leading up to the line. The line is flat and white and is in the plane of the retina. Stage 2 (ridge) is present when the line of stage 1 has grown, has height and width, and extends out of the plane of the retina (Fig. 12-2c). The ridge may be either pink or white in color. Vessels may leave the plane of the retina to enter the ridge. Small tufts of new vessels may be noted on the surface of the retina posterior to the ridge. These vessels do not constitute fibrovascular growth, which is a necessary condition for stage 3 ROP. Stage 3 (ridge with extraretinal fibrovascular proliferation) exists when extraretinal neovascularization is added to the ridge of stage 2 ROP (Fig. 12-2d). Stage 4 (retinal detachment) is the addition of (partial) tractional retinal detachment to stage 3 findings (Fig. 12-3a,b), and stage 5 is total retinal detachment, often with retrolenticular tissue (Fig. 12-4). Stage 5 retinal detachments are total. Stage 5 is subdivided based on the shape of the funnel. The funnel is divided into anterior and posterior parts, allowing for four subdivisions depending on whether the funnel is open or narrow in both its parts.
Progressive vascular incompetence, occurring in association with changes at the edge of the abnormally developing retinal vasculature, is confirmed by increasing dilation and tortuosity of the peripheral retinal vessels, iris vascular engorgement, pupillary rigidity, and vitreous haze. When vascular changes are so marked that the posterior veins are enlarged and the arterioles are tortuous, a plus sign is added to the ROP stage number (“plus” disease) (Fig. 12-5a). At least two quadrants of plus disease are required for the diagnosis of plus disease. Although it is not part of the classification, the committee recognized that regression is the most common outcome of ROP. The myriad patterns of regression were thought to be too numerous to classify, but they were listed as shown in Table 12-1.
In 2005, the ICROP report was updated using wide-angle digital fundus photographs, and included the following revisions 23 : (1) The classification of “preplus” disease was introduced. It represents posterior pole vascular changes that do not meet criteria for plus disease (Fig. 12-5b). (2) A severe form of ROP called “aggressive posterior ROP” (AP-ROP) was also recognized (Fig. 12-5c). This occurs in the youngest infants, and is characterized by posterior plus disease, often with flat neovascularization and rapid progression that may skip the classic ROP stages.
AP-ROP is a rapidly progressive form of ROP found in zone I or posterior zone II of very premature infants, and may skip the traditional ROP stages. AP-ROP needs to be treated immediately.
12.6.1 Flourescein Angiography and Optical Coherence Tomography
Wide-field fluorescein angiography (FA) can be accomplished in patients with ROP using digital imaging systems, for example, RetCam (Clarity Medical Systems). 25 , 26 , 27 , 28 , 29 , 30 It allows precise delineation of the vascular–avascular junction, clear visualization of the fibrovascular proliferation, and identification of areas of capillary nonperfusion. Flat stage 3 disease often seen in posterior ROP, and popcorn lesions (small isolated tufts of regressing stage 3), may also be more easily appreciated on angiography, compared to clinical or digital fundus images alone. Other features such as focal capillary dilatations, arteriovenous shunting, and irregular branching of peripheral vessels are also observed on angiography. Finally, FA can be used to image areas of residual avascularity after treatment to determine appropriate re-retreatments or timing of follow-up (Fig. 12-6). It may be useful but at this time is not considered essential in routine ROP management.
Optical coherence tomography (OCT) has provided new insights into ROP pathophysiology and foveal development (Fig. 12-7a). 31 Initial techniques required prone positioning of neonates under general anesthesia to set the infants’ chins onto the OCT interfaces, 32 , 33 but hand-held systems have been developed to allow easier and safer scanning in the supine position. 34 , 35 , 36 With OCT imaging, it was observed that approximately half of premature infants develop cystoid macular changes (Fig. 12-7b). 37 , 38 This appears to be independent of clinical ROP staging. 38 , 39 Many patients with a history of ROP may have poor visual outcomes despite good anatomic results. Previously, the poor vision was commonly attributed to cerebral dysfunction, but the maculopathy appreciated on OCT may also be contributory. 33 Furthermore, OCT is being investigated as a method for analysis of choroidal 40 and optic nerve 41 development.
12.6.2 Screening Examination Schedule
The key elements of the recently updated recommendations for the initial ophthalmologic examination of premature infants are as follows 44 :
All infants with a birth weight = 1,500 g or GA of 30 weeks or less should be examined. Additionally, infants with a birth weight 1,500 to 2,000 g or GA > 30 weeks who have experienced an unstable clinical course may also be screened at the discretion of the treating neonatologist or pediatrician.
The first examination should take place either at 31 weeks of postmenstrual age (PMA) or 4 weeks after delivery.
After the initial examination, follow-up is determined by the following guideline, but will be determined by the practitioner:
1 week or less: zone I stage 1 or 2, zone II stage 3, or suspicion for AP-ROP
1 to 2 weeks: posterior zone II immature, or zone II stage 2
2 weeks: zone II immature, stage 1, or regressing.
2 to 3 weeks: zone III stage 1, 2, or regressing.
Criteria to help the clinician decide to conclude acute ROP screenings include the following:
Vascularization to zone III in eyes with no previous ROP.
Full vascularization to the ora serrata for.
PMA 50 weeks with no prethreshold disease or worse.
Regression of ROP, with no abnormal vascular tissues that may reactivate.
In eyes treated with intravitreal bevacizumab, follow-up examinations are recommended until there is full vascularization of the retina.
It is recommended that infants with high-risk prethreshold, or Type 1 ROP (which has a >15% risk of retinal detachment), be treated with laser within 72 hours. Waiting until threshold disease is no longer recommended. Type 1 ROP includes the following:
Zone I: any stage with plus.
Zone I: stage 3 without plus.
Zone II: stage 2 or 3 with plus.
It is recommended that infants with Type 2 ROP (which has a <15% risk of retinal detachment) be followed closely. Type 2 ROP includes the following:
Zone I: stage 1 or 2 without plus.
Zone II: stage 3 without plus.
Examinations usually take place in the intensive care nursery with monitoring of the infant by the nursery staff. To decrease the risks of emesis and aspiration, examinations should take place no sooner than 1 hour after feeding. Although various other regimens of mydriatic agents can be used, we use three drops of a combination of cyclopentolate 0.2% and phenylephrine 1% (Cyclomydril, Alcon). Topical anesthesia with 0.5% proparacaine or 0.5% tetracaine is administered, and a lid speculum is inserted. Anterior segment examination followed by binocular indirect ophthalmoscopy with a 28-diopter lens is then performed.
The degree of pupillary dilation should be noted on anterior segment examination. Poor dilation may reflect iris vascular engorgement, implying very active ROP. This is often mistaken for iris neovascularization. The presence or absence of “plus” disease should be noted before scleral depression. Scleral depression can modify the amount of vascular engorgement that is apparent. On indirect ophthalmoscopy, the zone should be graded and the stage noted for each clock hour.
If vitreous hemorrhage or stage 5 ROP is present, B-scan ultrasonography may be needed. Ultrasonography can help determine whether a retinal detachment exists behind vitreous hemorrhage that precludes adequate visualization of the retina. In advanced stage 5 ROP, the configuration of the entire funnel-shaped retinal detachment can be evaluated. The anterior and posterior parts of the funnel can be graded as open or closed. This is helpful for providing prognosis and in planning surgery.
12.6.3 Telemedical ROP Screening
The incidence of severe ROP is on the rise, 45 , 46 and it is becoming increasingly difficult for ophthalmologists to keep up with the demand for ROP screening services. 47 To improve access to care, telemedical ROP screening may provide a viable alternative or adjunct to conventional bedside screening. Currently, the telemedicine model that appears to be best suited for ROP screening is for trained personnel to acquire bedside fundus images using wide-field digital fundus cameras. The images are then uploaded for remote review by expert readers in a “store-and-forward” manner. 48 , 49 The concepts of “referral-warranted” 50 and “clinically significant” 51 ROP were introduced to identify patients who would benefit from bedside examinations with indirect ophthalmoscopy by the treating ophthalmologist. Studies have shown that remote digital fundus imaging has high sensitivity and specificity, 52 , 53 , 54 especially in older infants. 54 , 55 Telemedical screening programs have demonstrated the reliability of identifying patients who require treatment. 56 , 57
In 2012, the American Academy of Ophthalmology published an Ophthalmic Technology Assessment of detecting clinically significant ROP with wide-angle digital photography, and classified seven studies with Level I evidence. 58 Benefits of the technology included objective documentation, being able to compare photographs longitudinally and send them for outside consultations, and the opportunities for education and research. Challenges included personnel training, medicolegal issues, HIPAA compliance, coordination with the NICU, and clear designation of responsibilities.
12.7 Cryotherapy and the CRYO-ROP Study
For infants with progressive ROP, various interventions can stabilize or cause regression of the disease. As early as 1967, Nagata et al 59 noted regression of ROP after xenon arc photocoagulation. Nonetheless, this method was cumbersome and was supplanted by cryotherapy, which was first reported to induce regression by Yamashita 60 in 1972. Several pilot studies of cryotherapy paved the way for the landmark multicenter trial of CRYO-ROP. 16 , 61 , 62 This study demonstrated the effectiveness of cryotherapy for reducing the likelihood of an “unfavorable outcome” in infants with threshold disease. In the CRYO-ROP study, an unfavorable outcome was defined as the presence of retinal folds involving the macula, retinal detachment involving zone I, or retrolental tissue or “mass.” Threshold disease was defined as 5 contiguous or 8 cumulative clock hours of stage 3 ROP with plus disease. 22 Treating threshold ROP resulted in reduction in unfavorable outcomes from 51% in the observation group to 31% in the cryotherapy treatment group (a risk reduction of about 40%). 62 This translated into a lower risk of blindness in the treatment group. 63
An infant’s systemic status may preclude treatment. In the CRYO-ROP study, 62 9% of infants suffered bradycardia, arrhythmia, or significant apnea during administration of anesthesia and cryotherapy. Brown et al reported three instances of respiratory arrest and one of cardiopulmonary arrest in 80 consecutive infants treated with cryotherapy for ROP. 64 These severe systemic complications underscore the need to have infants monitored carefully by a neonatologist or anesthesiologist during treatment. If the neonatologist believes that cryotherapy may be too stressful for the infant, then treatment should be deferred.