When oxygen began to be seriously considered as a cause of RLF, further confusion arose because there was suggestive evidence supporting each of three conflicting views: (1) exposure to added oxygen was either a primary or a contributory cause of RLF (Crosse and Evans, 1952; Campbell, 1951); (2) exposure to added oxygen did not influence the incidence of disease (Lelong et al., 1952; Exline and Harrington, 1951); (3) lack of oxygen was the cause of RLF. The latter view had two divisions: (1) the hypoxia which many premature infants spontaneously developed was itself the cause of RLF (Szewczyk, 1952), and (2) infants exposed to added oxygen became relatively anoxic at the time of their often abrupt removal from an oxygen-enriched environment (Szewczyk, 1952; Jefferson, 1952). It was believed that gradual reduction of supplemental oxygen lessened this effect and that reinstitution of oxygen therapy was beneficial if signs of RLF appeared.
With regard to the apparent diametrically opposed views on too much versus too little oxygen, it was not until 1953 that Ashton and coworkers demonstrated in kittens that a high oxygen tension caused constriction of the retinal arteries which ultimately led to retinal ischemia. This work, described below in greater detail, revealed and explained for the first time the paradox of too much oxygen leading ultimately to retinal damage from too little oxygen. Until this paradox was understood, it was difficult to resolve the conflicting and often erroneous claims from clinical observations in human infants.
In evaluating the claims that oxygen therapy was unrelated to RLF, it was recognized that some centers employed routine oxygen therapy but no RLF was noted and there were presumably reliable, albeit usually verbal, reports of the disease having appeared in infants never exposed to added oxygen.