126: Retinopathy of Prematurity

1. Retinopathy of prematurity (ROP). A disorder of the developing retinal vasculature resulting from interruption of normal progression of newly forming retinal vessels. Vasoconstriction and obliteration of the advancing capillary bed are followed in succession by neovascularization extending into the vitreous, retinal edema, retinal hemorrhages, fibrosis, and traction on, and eventual detachment of, the retina. In most cases, the process is reversed before fibrosis occurs. Advanced stages may lead to blindness.

2. Retrolental fibroplasia (RLF). As originally described, the condition was seen only in its most advanced form, after extensive fibrosis and scarring had already occurred behind the lens. It was, therefore, termed retrolental fibroplasia. It is now understood that recognizable changes occur in the developing vasculature before end-stage fibrosis occurs, making this condition a true retinopathy. Because it is found chiefly in premature infants, it is called retinopathy of prematurity.

3. Cicatricial ROP. The term cicatricial ROP refers to fibrotic disease.

ROP represents ∼20% of blindness in preschool children in the United States. Of particular concern are the increasing numbers of survivors weighing <1000 g, who have the highest incidence of ROP. The U.S. National Institutes of Health (NIH) sponsored the Cryotherapy for Retinopathy of Prematurity (CRYO-ROP) study in 1986–1987, which showed that 65.8% of infants weighing <1251 g developed ROP of any stage. Two percent of infants weighing 1000–1250 g developed threshold stage III+ disease eligible for treatment, whereas 15.5% of infants weighing <750 g did so. Threshold disease occurred at a median postconceptional age of 36–37 weeks, regardless of gestational age at birth or chronologic age. A study of earlier treatment, the ETROP (Early Treatment for Retinopathy of Prematurity) study, carried out in 2002, showed little difference in the overall incidence of the disorder and time of onset from the earlier CRYO-ROP study. The International NO-ROP Group, however, published data in 2005 suggesting that severe ROP in larger infants is an emerging worldwide issue. Care and survival of these premature infants is improving, whereas specialized treatment for ROP is not yet as prevalent.

1. Historical perspective. RLF was first described by Terry in the 1940s and was associated with the use of oxygen in newborn infants by Patz in 1984. The first epidemic, estimated to be responsible for 30% of cases of blindness in preschool children by the end of the 1940s, occurred during a period of relatively liberal oxygen administration. After this association was recognized, oxygen use in nurseries was curtailed. Although the incidence of RLF fell, mortality rates in newborn infants increased. In the 1960s, improved oxygen monitoring techniques made possible the cautious reintroduction of oxygen into the nursery. Despite improved oxygen monitoring, however, a second epidemic of RLF (ROP) appeared in the late 1970s and is associated with the increased survival of very low birthweight infants.

2. Normal embryology of the eye. In the normally developing retina, there are no retinal vessels until about 16 weeks' gestation. Until then, oxygen diffuses from the underlying choroidal circulation. At 16 weeks, in response to a stimulus (experimental evidence suggests relative hypoxia stimulating the release of angiogenic factors as the retina thickens), cells derived from mesenchyme traveling in the nerve fiber layer emanate from the optic nerve head. These cells are the precursors of the retinal vascular system. A fine capillary network advances through the retina to the ora serrata, or retinal edge. More mature vessels form behind this advancing network. Vascularization on the nasal side of the ora serrata is complete at ∼8 months' gestation, whereas that on the temporal side is ordinarily complete at term. Regulation of this process involves various factors including vascular endothelial growth factor (VEGF) and insulin-like growth factor 1 (IGF-1) working in combination. Once the retinal vasculature is completely vascularized, it is no longer susceptible to insults of the type that lead to ROP.

3. Causes

   1. There appear to be 2 phases in the development of ROP. Clinical and experimental evidence lead to the following general concepts.

      1. Early vasoconstriction and obliteration of the capillary network occurs in response to high oxygen concentrations noted experimentally or another vascular insult. Concentrations of IGF-1 are low in the very low birthweight infant in the early postnatal period as maternal levels are no longer available. Experimental evidence in the mouse model suggests that low IGF-1 contributes to the lack of retinal blood vessel formation in early ROP.

      2. Vasoproliferation follows the period of high oxygen exposure or insult, in response to angiogenic factors such as VEGF released by the hypoxic retina. Recent data suggest that VEGF only leads to angiogenesis in the presence of adequate tissue concentrations of IGF-1. When endogenous IGF-1 levels rise in the maturing at-risk premature infant, vasoproliferation is triggered in the presence of VEGF. Considerable evidence has been developed to support this hypothesis. Phelps and Rosenbaum studied kittens made hyperoxic and then allowed to recover in room air (21% oxygen) or 13% oxygen. Those recovering in the hypoxic environment had worse retinopathy than those recovering in room air, suggesting that retinal hypoxia plays a role. VEGF is a known product of the hypoxic retina. Using a mouse model, Smith et al have shown that IGF-1 is an important factor in VEGF action and the development of ROP. Further research will surely elucidate other modulators of this process.

The association of ROP with oxygen alone is not so clear. Transient hyperoxemia alone is not sufficient. Many other factors, such as extreme prematurity, apnea, sepsis, hyper- and hypocapnia, intraventricular hemorrhage, anemia, exchange transfusion, hypoxia, lactic acidosis, and possibly erythropoietin, which is angiogenic, have been implicated. Experimental studies have focused chiefly on the role of oxygen. Oxygen monitoring is an important part of the care of the premature infant, although extreme prematurity is known to be the significant risk factor. Recent data from Hellström et al have shown that postnatal weight gain also significantly affects the development of ROP.

Several methods of classification of ROP have been used. With development of the International Classification of ROP, there is general agreement on the staging of active disease.

1. Stage I. A thin demarcation line develops between the vascularized region of the retina and the avascular zone.

2. Stage II. This line develops into a ridge protruding into the vitreous.

3. Stage III. Extraretinal fibrovascular proliferation occurs with the ridge. Neovascular tufts may be found just posterior to the ridge (Figure 126–1).

4. Stage IV. Fibrosis and scarring occur as the neovascularization extends into the vitreous. Traction occurs on the retina, resulting in partial retinal detachment.

5. Stage V. Complete retinal detachment.

6. Plus disease (eg, stage III+). This may occur when vessels posterior to the ridge become dilated and tortuous. Plus disease has become a primary factor in treatment decisions.

7. Pre-plus disease. Dilation and tortuosity of posterior pole vessels in zone 1; less severe than plus disease.

8. Aggressive posterior ROP (AP-ROP). Rapidly progressive ROP primarily zone I. Requires immediate treatment.

   The retina is divided into circumferential zones I, II, and III to designate how far from the back of the retina (the posterior pole) disease is present. The most severe disease is any stage with plus disease close to the posterior pole, in zone I. The least severe is disease in the peripheral retina, zone III. No treatment is necessary for peripheral zone III disease, as it regresses spontaneously.

Ophthalmoscopic examination by an experienced examiner usually confirms the diagnosis. Binocular indirect ophthalmoscopy (BIO) is generally used. Assessment of newer digital camera technology in the multicenter PhotoROP study demonstrated 100% sensitivity in detecting ROP requiring treatment. However, it does not permit adequate assessment of ROP in the retinal periphery. BIO must be employed to determine when screening can be discontinued. A 2006 joint statement by the American Academy of Pediatrics (AAP), American Association for Pediatric Ophthalmology and Strabismus, and American Academy of Ophthalmology provided updated recommendations for ROP screening examinations in premature infants. These recommendations are evolving and may change as longer-term ROP outcomes are recognized. AAP pain management recommendations from 2006, reaffirmed in 2010, state that retinal examinations are painful, and pain relief measures should be used. They further state that a reasonable approach would be to administer local anesthetic eye drops and oral sucrose.

1. Infants weighing ≤1500 g or ≤30 weeks' gestation and those weighing >1500 g with an unstable clinical course should have dilated eye examinations starting at 4–6 weeks of age or 31–33 weeks' postmenstrual age. Examinations should continue every 2–3 weeks until retinal maturity is reached, if no disease is present.

2. Infants with ROP or very immature vessels should be examined every 1–2 weeks until vessels are mature or the risk of threshold disease has passed. Those at greatest risk should be examined every week.

1. Circumferential cryopexy. Proven to be an effective treatment for progressive (stage III+) disease in an attempt to prevent further progression by destroying cells that may be releasing angiogenic factors. Results of the large collaborative NIH-sponsored trial indicate that cryopexy carried out at stage III+ can reduce the incidence of severe visual impairment by ∼50% if performed within 72 hours of detecting threshold disease. Although myopia is a common feature of ROP, 10-year follow-up shows significant improvement in visual acuity of treated versus control eyes. It is imperative that an ophthalmologist skilled in cryopexy perform the procedure.

2. Laser photocoagulation. Data suggest that this technique is equally effective yet safer than cryopexy, and it has become the treatment of choice. In 1994, the Laser ROP Study Group was formed to carry out a meta-analysis of 4 laser-ROP trials. Treatment was based on the same criteria used in the CRYO-ROP trial. Recognizing the limitations of a meta-analysis, the study group concluded that laser therapy is at least as effective as cryotherapy for ROP, despite a small risk of cataract formation. Ten-year follow-up of a small group of patients suggests better outcomes with laser photocoagulation. More recently, the ETROP study (2002) demonstrated improved outcomes with treatment at any stage when plus disease is present. AAP pain recommendations from 2006, reaffirmed in 2010, state that retinal surgery should be considered major surgery, and effective opiate-based pain relief should be provided while monitoring the infant with an appropriate pain assessment scale.

3. Oxygen for treatment of ROP. In an attempt to reduce angiogenic factors from the hypoxic retina and the progression of ROP from prethreshold to threshold (III+) levels, oxygen therapy was attempted in a large collaborative trial, the Supplemental Therapeutic Oxygen for Prethreshold Retinopathy of Prematurity (STOP-ROP) study. Oxygen saturations were targeted at 96–99% in the treatment group and 89–94% in the conventional group once prethreshold ROP was diagnosed. No significant difference was seen in the rate of progression to threshold disease between the 2 groups, although there was a significant increase in bronchopulmonary dysplasia/chronic lung disease in the high-saturation group. The appropriate saturation ranges remain controversial and under study, although most neonatal intensive care units keep infants <1250 g at saturations <95% when in supplemental oxygen. Multicenter, randomized trials of reduced oxygen saturation therapy are currently being undertaken in several countries. Randomization is to oxygen saturations of 85–89% or 91–95% with similar study designs to allow a planned meta-analysis of the trials when completed.

4. Additional experimental therapies. Therapies that show promise include suppression of VEGF with anti-VEGF therapy and dietary supplementation of omega-3 polyunsaturated fatty acids (PUFA). The balance of omega-3 and omega-6 PUFAs in the retina affects cell survival. Studies in the mouse model have shown a protective effect of omega-3 supplementation. Anti-VEGF therapy has moved from animal studies to human clinical trials. Mintz-Hittner et al recently published their results for the BEAT-ROP collaborative trial in which infants with stage 3+ ROP were randomly assigned to intravitreal bevacizumab monotherapy versus laser therapy. A significant benefit for zone 1 ROP (the most difficult to treat conventionally) was demonstrated. Outcomes were short, and safety could not be assessed due to the small size of the trial. Further study has been recommended.

5. Vitamin E. The administration of pharmacologic doses of vitamin E for ROP has been studied with no proof of clear benefit. Reported side effects include sepsis, necrotizing enterocolitis, and intraventricular hemorrhage. Even so, maintenance of normal serum vitamin E levels is a prudent management objective. (For doses, see Chapter 148.)

6. Decreased lighting intensity. A prospective, randomized, multicenter trial of 409 premature infants weighing <1251 g and 31 weeks' gestation concluded that a reduction in ambient light exposure does not alter the incidence of ROP.

7. Retinal reattachment. Stage IV disease has been treated by attempts at retinal reattachment without significant success to date. Reattachment of late retinal detachments in childhood has met with more success.

8. Vitrectomy. This has not substantially improved the outcome in cicatricial disease.

9. Follow-up eye examinations. Advocated every 1–2 years for infants with fully regressed ROP and every 6–12 months for those with cicatricial ROP. Premature infants are at risk for myopia even in the absence of ROP and should have an eye examination by 6 months of age.

Ninety percent of cases of stage I and stage II disease regress spontaneously. Current information suggests that ∼50% of cases of stage III+ disease regress spontaneously. Of those that do progress to stage III+, the incidence of unfavorable structural outcomes can be reduced by ∼50% and unfavorable visual outcomes by ∼30% if circumferential cryopexy is carried out by a skilled ophthalmologist. Laser photocoagulation appears equally and possibly more effective than cryopexy. Sequelae of regressed disease such as myopia, strabismus, amblyopia, glaucoma, and late detachment require regular follow-up.