Retrolental Fibroplasia: A Modern Parable – Chapter 8

Chapter 8
The Consequences of Oxygen Restriction

Interest in RLF quickly waned after 1955. Except for the engrossment of tort lawyers, very little attention was paid to the condition (Fig. 8-1). Governmental plans for developing expanded facilities for blind children were quietly shelved and there was general agreement that RLF could be chalked up in the “solved” column.

As noted, the under-40%-only policy of oxygen treatment was adopted widely; indeed, it was reinforced by the knowledge that malpractice suits for RLF were burgeoning. Within a short time, however, it became evident that total elimination of RLF could not be realized; some cases continued to occur after 1954-1955. As might have been predicted from the pre-1940 history of the disorder, typical and fully documented instances of RLF occurred in some premature infants who had never received supplemental oxygen. Although this occurred rarely and was first doubted by many, the weight of evidence was finally convincing and the possibility was accepted that 21-percent oxygen in ordinary room air can “produce” RLF. When RLF occurred in infants who were exposed to oxygen very briefly (e.g. only during resuscitation at birth), it created endless arguments about cause and effect. Most of the debates were occurring in front of juries hearing numerous malpractice suits tried in the years after 1954.

The first sign of a backlash effect from the new restrictive practice of oxygen treatment appeared more than 5 years after announcement of the results of the Cooperative Study. Mary Ellen Avery and Ella H. Oppenheimer reviewed autopsies at Johns Hopkins Hospital noting the mortality and frequency of occurrence of the commonest life-threatening complication in premature infants: hyaline membrane disease. (Respiratory distress syndrome is the label applied to a serious breathing problem caused by inadequate stabilizing substance – surfactant – in the lung. In those infants with respiratory distress who die, most have a transparent, membrane-like deposit in the lung which gives the condition its pathologic name: hyaline membrane disease.) The Baltimore investigators reasoned that if the lung condition was the result of incomplete development and unrelated to factors that operate after birth, frequency should be essentially the same from one period of time to the next. Further, they postulated that since there was no definitive treatment, the mortality rate associated with hyaline membrane disease should remain practically constant. However, when they examined the experience of two 5-year periods, they found a definite change; frequency of hyaline membrane disease and mortality were both higher during the years of oxygen sparing (Table 8-1). The authors wisely refrained from applying statistical tests to the retrospective review and cautiously concluded the following:

The increase in the number of deaths from hyaline membrane disease in the period of restricted oxygen use suggests that some infants with respiratory distress may need more oxygen than they have been receiving.

They warned that the observations pertained only to deaths from hyaline membrane disease (but this is the most frequent single cause of death after premature birth). In retrospect, it is curious that the report made no mention of the Gordon letter, which had been published three years earlier and had called attention to the problem of interpreting the conclusions of the Cooperative Study concerning the safety of oxygen restriction (p 50). Even more surprising, the understated, but nonetheless disturbing, observations from Johns Hopkins Hospital evoked relatively little response in neonatal circles. No other publicized review of the association between oxygen policy change and mortality appeared for 14 years (Chapter 9).

Table 8-1

Mortality and Occurrence of Hyaline Membrane Disease in
Two 5-Year Periods (Johns Hopkins Hospital Premature Nursery)

In 1962 Alison D. McDonald reported the results of the follow-up examinations of 1081 English ex-premature children at ages 6 to 8 years. These children who had weighed 1.8 kg (4 lb) or less at birth had been studied earlier in a nationwide investigation of RLF by the Medical Research Council. McDonald sought to discover associations between perinatal events and the occurrence of spastic diplegia, known to occur more often following premature birth than after term delivery. A relatively high rate of diplegia, and low rates of other types of cerebral palsy, were found in this group of ex-premature children. No associations were observed between diplegia and various complications of pregnancy and delivery (e.g., neither asphyxia during delivery nor resuscitation). However, there was a correlation between spastic diplegia and a history of postnatal breathing difficulties associated with cyanotic spells and respiratory distress (as indicated by retractions of the chest wall and grunting respiration). There was also a relationship between the neurologic outcome and the duration of oxygen therapy with an opposing trend as to the status of the eyes, i.e., spastic diplegia frequency fell and RLF rose as the duration of oxygen treatment increased (Fig. 8-2). Doctor McDonald concluded that

There was some indication that prolonged oxygen therapy prevented diplegia in very immature infants with cyanotic attacks. It is suggested that this treatment may be of value if retrolental fibroplasia can be avoided.

A similar follow-up of children who had been enrolled in the American Cooperative Study was planned but never carried out (Chapter 9).

In the early 1960s there was increasing uneasiness with the blanket policy of oxygen restriction for all premature infants. Although the cited reviews played a part in raising these doubts, the tide of opinion/practice was turned by the flood of new information which became available following the technical development making it possible to measure the partial pressure of oxygen in small samples of arterial blood (see chapter notes).

From the outset, when high oxygen had been accepted as the cause of RLF, it seemed unreasonable that oxygen surrounding the infant would be more critical than the oxygen in the blood vessels. When arterial blood oxygen tension in premature infants with respiratory distress syndrome (p 60) was correlated with oxygen in inspired air, it was evident that many of these babies were inadequately oxygenated while breathing 40-percent oxygen (see chapter notes for further details). Studies at Oxford in 1962 demonstrated that most infants with severe respiratory difficulty required very high concentrations of oxygen in inspired air to raise the oxygen level of aortic blood. Further studies suggested that prompt resuscitation and exposure to high concentrations of oxygen (80-90 percent) in premature infants with asphyxia of the newborn (due to inadequate exchange of oxygen and carbon dioxide between the fetus and the mother across the placenta before delivery, or failure of the newborn infant to breathe promptly after birth) might forestall the development of the respiratory distress syndrome. These observations provided the rationale for a liberalization in the mid-1960s of the restrictive policy of oxygen therapy, but there was serious concern that premature infants might be subjected to an increased risk of RLF.

In 1967 the National Society for the Prevention of Blindness convened a meeting of 28 American and Canadian pediatricians, ophthalmologists and biophysicists to consider the problems posed by the new shift in oxygen opinion/practice. The discussion centered on the need for the following:

1. Criteria for supplemental oxygen administration;
2. Accumulating evidence on the association of clinical signs, arterial oxygen measurements, appearance of retina through the ophthalmoscope, and status of brain function in oxygen-treated infants;
3. Improvement of devices for monitoring oxygen concentration in the air breathed by newborn infants;
4. Caution in administering supplemental oxygen when appropriate observations cannot be made;
5. Basic research in factors which control blood vessel function.

Humbled by the fact that only anecdotal clinical evidence was available, the participants wisely refrained from making definite recommendations concerning the treatment of premature infants. However they agreed, tacitly, that a definite change in opinion/practice had once more taken place.

Publications concerning RLF from 1942 through 1962.

Spastic diplegia and RLF according to median duration of oxygen treatment. Among 194 children born before 31 weeks gestation who had increasing duration of oxygen therapy, spastic diplegia occurred less frequently, but RLF turned up more frequently. There was no association between spastic diplegia and duration of oxygen administration in children born after longer periods of gestation.

Return to Contents Page