Some studies of ROP

Neonatologists who say that the nursery lights do preemies no harm base this assertion on a small scale trial published in 1952 which claimed to have ruled out a connection between exposure to light and ROP. The authors reported that just as many preemies had developed the disease when their eyes were patched with gauze as when they were not. However, in that study the babies' eyes were patched, not immediately, but within up to 24 h after birth (103). That is more than enough time for the fluorescent light to overdose their fragile retinae with damaging blue radiation.

A year after this misleading study, another flawed but very influential study asserted that oxygen was the major cause of ROP. In that study, 18 nurseries withheld oxygen from some of the preemies and found fewer cases of ROP among the survivors of that group (104). This result was acclaimed as a victory over ROP and led to severe oxygen rationing for most preemies that has endured to this day.

However, the unacknowledged reason for the apparent reduction in the incidence of ROP was that fewer preemies, with immature lungs and eyes, survived long enough to display the symptoms of ROP. In fact, the lack of sufficient oxygen killed most of the babies whom ROP would have blinded, plus many more whom ROP would have spared (105).

A British researcher estimated two decades later that each case of ROP avoided by withholding oxygen "may have cost some 16 deaths" (106).

Preemie deaths from early breathing problems increased the overall preemie mortality in Baltimore, Maryland, from 8% when all low birth-weight babies were freely given oxygen, to 13% when oxygen was rationed. A retrospective analysis of all preemie deaths in the U.S. and in England and Wales found that, with oxygen rationing, the day-of-birth death rates of preemies either staved stable or increased, instead of decreasing as they had done for decades (107).

In addition, the lack of oxygen causes much brain damage among the survivors. For instance, a British study of more than 1000 ex-preemies found in 1962 that denying supplementary oxygen during the first few days had quadrupled the incidence of spastic diplegia among preemies from 5% for those with 10 or more days in oxygen to 20% for those with fewer or none (106). However, because the oxygen restrictions reduced the apparent incidence of ROP, these well-documented side-effects of the prevention strategy are mostly disregarded.

By the mid and late 1960s, researchers studying the safety of laser light for industrial applications discovered that light could damage eyes not just by burning, the retina with heat as in welding accidents or Sun-staring -- but also through a slower, non-thermal process which they found to be photochemical (108-110).

This discovery, together with the then recent but already widespread use of intense light to treat jaundice in newborn babies, led some physicians to re-open the question about the effect of light on preemies.

In August 1970, a team of physicians from Boston and Philadelphia described in their paper "Retinal Changes produced by Phototherapy" (67) how they had placed newborn piglets under phototherapy lamps with a total irradiance of 300 ftc "to determine if retinal damage does in fact occur during phototherapy of the newborn infant".

They had picked piglets because "the newborn piglet eye is developmentally of close approximation to the human of comparable age, and since the piglet is a diurnal animal whose eyes are similarly pigmented".

One of the piglets lost the patch over its control eye which had not been dilated, and which remained relatively protected by its heavy eyelids, covered with hair, and thick eyelashes. Although that eye was exposed to the lights for less than 12 h, the next day that piglet had become totally blind. When its retinae were examined 3 weeks later under an electron microscope, both showed virtually the same "marked damage" as the other exposed piglets' retinae.

These researchers concluded that preemie eyes exposed to phototherapy lights should be shielded with the utmost care, and warned:

"The intensities of light used for illumination in nurseries may need to be evaluated for their short- and long-term detrimental effect on vision in the newborn infant."

No evaluation of short-term effects from nursery lighting was published for another 15 years. Meanwhile, a few studies searched for long-term effects, but they were grossly flawed and misleading. In all these follow-up studies, only those ex-preemies were included who successfully completed a series of eye tests (111-113).

The authors of one paper even state that in selecting patients for their study they had specifically excluded children with ROP or other eye problems, and that they had picked only those with no history of visual or other neurologic dysfunction (114).

A logical equivalent would be to choose only white marbles from a basket containing, marbles of various colors, and then to conclude that all the marbles in the basket must be white. These papers contained many other major errors and yet, they are still frequently cited as support for the assertion that light is safe for immature eyes.

The next study about the short-term effects of nursery lighting, was published 15 years after the piglet-blinding paper. Penny Glass and her co-authors found that reducing the light on the baby from 60 ftc to 25 ftc reduced the number of ROP victims in the two nurseries studied from 86% among the unprotected preemies with birthweights below 1000 grams to 54% among the less exposed ones in the same highest-risk group. The authors concluded that "This research raises serious questions regarding the levels of light appropriate for pre-term infants" (69).

Advocates of bright nursery lights dismiss the Glass et al. study because the groups of preemies exposed to 60 ftc and 25 ftc had not been studied simultaneously, but serially. These critics say that such sequential study designs are not valid (115-118).

Ironically, the 1953 study on which is based the mainstream neonatologist belief about oxygen as the alleged cause of ROP had a similar sequential design. In that study, which dominates the treatment of preemies today, the infants were assigned to the "routine oxygen" group only during the first 3 months of the investigation. The results from this very small group were then compared with those from the much larger "curtailed oxygen" group most of whose members were enrolled during the remaining 9 months of the study. The study description stated:

It was recognized that this method of assignment would not provide for ideal temporal control. However, in the absence of any seasonal variation of retrolental fibroplasia, the advantages of this design were thought to greatly outweigh the disadvantages (119).

According to the logic of those who dismiss the Glass et al. study, a sequential study design is "not scientifically credible" (116) and is "weakened by methodological shortcomings" (117) when its results support the light theory, but the same sequential study design is acceptable to these critics when it supports the oxygen theory (115, 116).

(Addendum in 2003:  A more recent oxygen withholding study which was published in the February 2003 issue of Pediatrics used again that same before- and- after format and found that the blinding rate diminished during the years of more severe oxygen rationing.  The authors compared no other parameters, such as changes in nursery lighting or isolette shading, but suggested anyway that their rationing practices may have been responsible for the reduction in blinding.  Here again, that format is acceptable when it confirms the official doctrine but not credible when it highlights problems. (119a))

The activities based on the initial sequential study, itemized as Ventilation and Oxygen Administration, currently account for a third of all the costs in the more than $2.6 billion-a-year intensive care nursery industry in the U.S. (120, 121).

Three months after Glass et al. had published their results, obtained with permanently mounted gray filter sheets over the top and down the back of each isolette, three nurses in the Newborn Special Care Unit of the Yale/New Haven Hospital began to lay receiving blankets over the top half of the isolettes (which is not where the preemie's eyes are directed; they all lie with their head to the side) and to measure light levels at the face of each baby.

Two years later, they sequentially compared the incidence of ROP among those babies with that during the preceding 2 years when they had not shielded the babies and had not measured the light intensity. They found no significant difference in the incidences of ROP (122).

This Yale study had repeated the error of the 1952 study of eyepatching. The authors covered the isolettes of the babies only after the babies' "first few days", although they acknowledge that

"the most crucial time for reduction of light intensity for the sick premature infant who requires oxygen and ventilatory support may be immediately after birth".

Instead of addressing this blatant contradiction, they concluded that decreasing the light exposure had not prevented the blinding.

Even more misleadingly, one of the authors co-wrote elsewhere about that trial: "Focusing our attention on background nursery lighting is not likely to provide the answer" (123). This logic is the equivalent of applying sunscreen to skin already covered with sunburn blisters and concluding that applying sunscreen is not likely to prevent sunburn.

Continue