Randomized Clinical Trials in the Treatment of Retinopathy of Prematurity
By Michael X. Repka, MD
Clinical trials for retinopathy of prematurity (ROP) are not new. Clinical research into the disease -- at least its ocular manifestations -- dates back more than 50 years. Research has led to enhanced prevention of the disease as well as improved outcomes of treatment. Improvements in the neonatal intensive care unit (NICU), in the training of NICU staff, and developments in the fields of neonatology and nutrition -- all of which have evolved over the last two decades -- have boosted prevention. And improved treatment includes all examinations and the latest interventions performed by an ophthalmologist.
ROP research methods, like those of any disease, include observational studies, in which its natural history is monitored; and case series, in which an assessment is made of the impact of changes in nursery care, the impact of a nutritional change or the impact of an intervention. These studies utilize patient populations drawn from two different time periods, and cannot account for other changes in care. Thus, such studies cannot prove the beneficial or deleterious effect of other changes. Randomized clinical trials are the most effective in proving the value of an intervention. Here, we examineclinical trials that have been part of research efforts to cure ROP.
The mean birthweight of patients with threshold ROP in North America has gone from over 800 grams (~1.76 pounds) 10 years ago to just over 600 grams (~1.32 pounds) today. Smaller babies are surviving premature births and developing ROP, while larger babies are now being spared. Thus, ROP still is responsible for much of childhood blindness in the developed world. One estimate of blindness in the U.S. caused by ROP is about 1,200 children annually. In addition to those who experience blindness or severe visual impairment due to ROP, there is a larger number of children with lesser degrees of vision impairment.
Precursors to Modern Studies
An observational trial in several nurseries from the 1950s in Melbourne, Australia, suggested a relationship between ROP and oxygen exposure. Oxygen given to premature babies was expensive; when curtailed, the incidence of scarring ROP declined. A clinical strategy adopted for cost containment was thus serendipitously found to be helpful in reducing the disease, though it did not cause ROP to disappear.
As early as 1949, vitamin E supplementation had been suggested, and studies to test this hypothesis were conducted. Owens and Owens published a trial suggesting a benefit of oral vitamin E. The nutrient was felt to be a means of preventing ROP by ameliorating the toxic impact of oxygen and its effects upon the developing retinal blood vessel. In a randomized trial in Philadelphia during the early 1980s that looked at babies treated with vitamin E, there was a suggestion of less ROP overall; there was no actual change in the incidence of severe ROP. Another study from UCLA found no benefit for ROP, but because other health factors were being monitored, an untoward effect was detected: increased intraventricular hemorrhage.
Studies in this field that have been, or are currently, supported by the National Eye Institute of the National Institutes of Health are: the CRYO-ROP Study (1988); the STOP-ROP Study (1994); the Light-ROP Study (1995); and a current project examining early treatment for the prevention of retinopathy of prematurity (ETROP), which began enrollment in fall 2000.
The CRYO-ROP Study
This acronym stands for cryotherapy for retinopathy of prematurity, a controversial and aggressive treatment at the time (mid-late 1980s). The historical basis for the study was that ablative treatments, like cryotherapy, damage a portion of the peripheral retina, halting the growth of abnormal blood vessels. This type of therapy was previously used to treat diseases such as diabetes and sickle cell disease, suggesting that it might help treat ROP. As used for ROP, 40-50 percent of the retina was destroyed in an effort to save the remainder -- the most important portion for preservation of functional vision. No physician knew what 80 years of life with a cryotherapy treated eye was going to be like for the children who participated in this study; it was unknown whether there would be a significant adverse impact of this treatment on an infant's eye. The motivating concept behind the study was the potential benefit of an attached, functional retina.
Nearly 10,000 children were screened nationally at 23 centers involving more than 50 hospitals across the US. The data came predominantly from a group of randomized children: 291 babies who received cryotherapy to one eye and no therapy to the other eye.
The ophthalmologist froze a portion of the eyeball from the outside in using a cryoprobe, a hollow instrument filled with a super-cooled refrigerant. This caused small freeze burns, similar to frostbite, in the most forward and undeveloped portion of the retina. Cryotherapy was stressful on the baby, on the neonatologist and on the ophthalmologist.
Most children did well in both eyes. Of the randomized group three months after treatment, 53 percent of untreated eyes did well and 74 percent of treated eyes did well. The difference in improvement from cryotherapy was significant to both the statisticians and the clinicians; an improvement of 20 percent in outcome translates into many more "good" eyes and many fewer babies who are blind.
These children reached their tenth birthdays in the late 1990s. Forty-four percent of untreated eyes had favorable visual acuity (better than 20/200), but 62 percent who were treated had favorable acuity outcomes; once again generating an outcome that is significant both clinically as well as statistically. These data confirm that cryotherapy works well, though not in every case.
Equally important, there were no significant adverse effects noted in the children as a result of the therapy. In the ten-year outcome data, one-fourth of children in each group had acuity of 20/40 or better. Children who were treated were much less likely to be blind in the treated eye than the untreated eye. However, it is easy to see from these data that this treatment did not make more perfect seeing eyes but substantially improved the number of children with partial vision. The treatment is imperfect, with most of the benefit at the lower end of the vision spectrum, reducing the number of children with severe vision loss. During 2001 and 2002, 15-year outcome data has been gathered, largely to look for any changes in acuity and the functional impact of therapy and ROP, prior to legal driving age.
Secondary Outcomes in the Cryo Study
The CRYO-ROP study has reported that myopia is overwhelmingly the most common secondary eye problem from this disease. It is a problem affecting all premature children, but more so those children who develop significant ROP. These children are at risk for acute angle closure glaucoma in their second or third decades of life. Also they are at a small risk for late retinal detachment, perhaps three percent. So, these children -- even those whose vision is good -- will have to remain under an eye doctor's care throughout their lifetimes, unlike the typical teenager or twenty-something.
The second most common abnormality is an eye movement disturbance, some of which will be due to faulty vision. Others will be motor-related, due to damage to the brainstem or other portions of the brain. And for a few children, motor problems may result from trauma to the eye due to the therapy itself.
When babies develop severe threshold retinopathy of prematurity, they have a high risk of other problems in the brain. Some children will have loss of vision due to brain injury, called cortical visual impairment, which may be partial or complete. Whatever the level of damage, this loss of function will compound the child's visual disability. The CRYO-ROP Study evaluated hydrocephalus and the need for a CSF shunt as a marker of future neurological impairment. Eleven percent of treated children had hydrocephalus to the point that they needed shunting.
Severe ROP was also a marker for future ophthalmological problems. These children had much higher rates of requiring additional eye surgery -- often for glaucoma or strabismus.
Amblyopia was also found to be very common. Amblyopia is the component of visual loss not associated with retinal or brain damage, but rather with visual loss due to lack of stimulation of the eye. When amblyopia was diagnosed, the conventional patching treatment was prescribed. However, few patients improved with such therapy, perhaps because their retinas and brains were not completely normal.
It's important to report a cautionary note about results from the CRYO-ROP Study: Children were enrolled in 1986, and they may no longer be the prototype of children with ROP in the US.
The Light-ROP Study
The historical basis for a study of light reduction and ROP (Light-ROP) dates to the 1940s, when light was identified as a toxic agent to the retina. Though it was known that exposure to a laser could damage vision, investigators learned that prolonged exposure, even to sunlight, leads to cataracts and retinal destruction. An observational study by Dr. Penny Glass and her colleagues found that bright nursery lights seemed to be associated with the development of ROP in two cohorts managed in the Children's National Medical Center in Washington, DC. The Light-ROP Study set out to confirm or refute that finding and to determine if manipulation of ambient light could modify the disease course.
The study did not treat ROP, but tried to prevent the occurrence or reduce its severity. Modified sunglasses or goggles were placed on children within the first 24 hours after birth and worn for a little more than a month. These children were randomized at three different sites; 188 babies wore goggles to block light and 173 experienced normal nursery lighting. The ROP incidence with light-reduction goggles was identical to that in controls. Therefore, researchers concluded with a high degree of certainty that a reduction in ambient-light exposure does not alter the incidence of ROP; children were still developing ROP at about the same rates, even when their eyes were kept in the dark for a month.
The STOP-ROP Study
The supplemental therapeutic oxygen for pre-threshold ROP (STOP-ROP) Study was a huge undertaking by neonatologists and ophthalmologists. The idea for this study was based on the observation that an oxygen deficit causes the development of abnormal blood vessels in animal models of ROP. The concept of adding oxygen was turning the tables on 50 years of neonatal management aimed at reducing the rate of blindness from ROP. Infants who were mildly hypoxic from a large number of national centers were randomized into this study; some had their oxygen saturation kept in the low 90s (maximum = 100), while others were supplemented with oxygen up to 96-100 percent. A large number of children were entered; 324 were supplemented and 325 served as controls.
As would be expected, some of the children had good retinal outcomes, while others had poor outcomes. Overall, there was neither a beneficial nor an adverse impact of this treatment on severe ROP. The authors also evaluated other health issues potentially impacted by use of oxygen. The oxygen-supplemented group had an increased risk of pulmonary disease, making this treatment possibly undesirable as provided in the study. Future research is needed to find the right balance of oxygen.
Another secondary outcome evaluated by the study group was the value of the quality of vision as a marker for future disability. Children with favorable acuities were much more likely to do well on developmental assessment of motor, social, verbal and cognitive scales; and children with unfavorable acuities were much less likely to do well.
Early Treatment of ROP (ETROP) Study
In the ongoing ETROP Study, we're using computer analysis of historical data of a child with ROP that predicts what the outcome for that patient will be three months after term if he/she had been treated at the normal time. If a child is predicted to have a poor outcome, the researcher will suggest treating one eye early. Parents have a tough decision: They're actually deciding to have an intervention because there is a >15 percent risk of blindness, but there is an equivalent risk that the child may need no intervention at all. This sounds like a tough decision, but this balanced possibility is the best way to establish whether early treatment is actually better. The conclusion made by this study will save many children either unnecessary treatment or blindness -- a good outcome no matter which way it turns out.
The plan is to randomize 370 patients so that each child will be treated early in one eye and at the conventional time in the fellow eye with a laser. Like the cryoprobe, laser treatment results in some swelling, scarring and pigmentation of a part of the retina not used for central vision. The enrollment in this study began in early 2001 and has continued; results will not be available until mid- to late-2003.
These ROP treatment trials have produced -- and will continue to lead to -- improvement in the management of ROP. Such steps will reduce the total number of children with blindness or partial sight. In addition, research to prevent ROP may also translate into better overall function and less disability for these children. This certainly has to be considered a major secondary goal after the preservation of vision.
Therapy of ROP today is imperfect. To have a therapy that fails 25 percent of the time is insufficient. Twelve hundred children still become blind from ROP in the United States each year, but that number is dwarfed by the large numbers of children who are surviving and suffering from this disease worldwide. It is especially true in the second world where neonatology has been exported successfully but ophthalmology lags far behind. There literally is no one trained to diagnose and treat the disease. The visual loss experienced in the developing world is staggering compared to what is seen in North America. In countries without neonatal centers, ROP is not yet an important issue, simply because children who would be at risk do not survive.
By Michael X. Repka, MD, Professor of Ophthalmology and Associate Professor of Pediatrics at the Wilmer Ophthalmological Institute, The Johns Hopkins School of Medicine, Baltimore, MD
From Lighthouse International's EnVision newsletter (Summer 2002 issue)