Pediatrics

VOLUME 19 MARCH 1957 Nuxtnui 3

AMERICAN ACADEMY OF PEDIATRICS PROCEEDINGS

THE ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA E. Mead Johnson Award Address

By Arnall Patz, M.D. Departments of and Pediatrics, District of Columbia General Hospital and the Hofiberger Research Laboratories, Sinai Hospital, Baltimore, Maryland

[ In presenting an E. Mead Johnson award problem of retrolental fibroplasia than I)i. Patz. to Dr. Patz, Dr. Bakwin, President of the Acad- His contributions are considered to be a beauti- cmv, made the following remarks: “Dr. Arnall fiil exercise in experimental medicine, the more Patz, Recipient of the Second E. NIead Johnson noteworthy 1)ecause he made them while en- Award, was born in Elberton, Georgia, in 1920. gaging in the private practice of ophthalmol- He attended at and ogy. graduated from the Medical School at that Uni- [“It is for these works that the Academy’s versitv in 1945. Dr. Patz interned at Sinai Committee on Awards has selected 1)r. Patz as Hospital in Baltimore; served in the Medical the recipient of an E. \lead Johnson Award Corps of the Army for 2 years; following which for 1956. This award consists of a handsome he became resident in Ophthalmology at the scroll amid a check for $1000 which we take District of Columbia General Hospital. In 1950 pleasure in iio\v presenting to Dr. Patz. and 1951 Dr. Patz was Research Fellow in [“Dr. Patz will honor us by giving a r-suni’

Ophthalmology at the Armed Forces Institute of his work which has contributed SO miiticli to of Pathology. pe(liatricS. [“Dr. Patz is certified b’ the American Board [ Dii. PTz: “I should like to exjress my of Ophthalmology and serves as an instructor in thanks to the American Academy of Pediatrics

Ophthalmology at Georgetown Umiiversitv and for the honor ou have accorded nie. As an . During the past 9 ophthalmologist it is particularly gratifying to \ears Dr. Patz has published 15 papers, 10 of know that the work (lone in con(ueriI1g retro- them relating to his clinical studies, backed up lental fibroplasia has received recognition from by animal experimentation, which have clearly this distinguished group. demonstrated the adverse effects of oxygen [“Although I have been named as the recipi- administration in relation to retrolental fibro- ent of this avard, it is, in fact, a tribute to all plasia in small infants. No one in this coun- the prominent scientists who have contributed try has been more influential in solving the to the ultimate victors’ against retroleutal fibro-

Presented at the Annual Meeting, October 10, 1956. These studies were aided by grants from the E. Matilda Ziegler Foundation for the Blind, The Na- tional Society for the Prevention of Blindness, and the National Institute of Neurological Diseases and Blindness, United States Public Health Service. ADDRESS: 1212 Eutaw Place, Baltimore 17, Maryland. 504

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 505 plasia. The work of Owens and Owens in iden- oxygen therapy and incidence of cicatricial tifying the true clinical course and of Frieden- RLF. Chart I summarizes these uncon- wald and Reese in clarifying the histopathology trolled data collected between July, 1948, of the disease deserve special recognition. Im- and December, 1950, and shows the con- portant contributions have been made by Boe relation between number of days of oxygen Hellstr#{246}m and Lars Gyllensten of Sweden, Nor- therapy and incidence of retnolental fibro- man Ashton and Mary Crosse of England, Kate plasia (residual membrane cases only). Campbell of Australia, Harry Gordon, Everett In 1949, Kinsey and Zachanias2 reported Kinsev, and Jonathan Lanman of this country, that three changes in pediatric cane, namely, and several others. [“I have been fortunate in having a loyal addition of iron medication, use of water- group join me in these studies. Dr. Leroy Hoeck miscible vitamins and increased use of oxy- has collaborated in the nursery study. Dr. gen were associated with the increased in- Edgar de Ia Cruz and Dr. Thomas Kleh as- cidence of RLF. I think it is fain to state that sisted in the eve examinations. Mrs. Ann East- between 1948 and 1950 most pediatric and ham, Dr. Miriam Reiner and Dr. Don Higgen- ophthalmologic authorities did not consider botham contributed to the animal studies. In that overuse of oxygen was related etio- accepting this award I thank you in the name logically to retrolental fibroplasia in spite of my entire research team and I thank you for of this epidemiologic report. Because of the privilege of addressing this assembly.”] this lack of suspicion of oxygen by leading authorities on the subject, it was extremely O BSERVATIONS on the role of oxygen ad- difficult to convince ourselves of the justifi- ministration in retrolental fibroplasia cation of restricting oxygen for a certain (RLF) have been conducted in the pre- number of infants, as would be necessary in mature nursery of the District of Columbia a controlled randomized study. Hospital with Dr. Leroy Hoeck, Depart- Fortunately for us, Dr. Harry H. Gordon, ment of Pediatrics, since 1948.1 Our atten- then Professor of Pediatrics at the Univer- tion was first focused on oxygen by one of sity of Colorado, spoke before the pediatric our first cases of RLF occurring in an in- staff of the in fant where special measures were utilized November, 1950. He stated that at the Pre- to deliver a high oxygen concentration for mature Center of the Colorado General several days after birth. This incidental Hospital, oxygen therapy routine was cur- observation suggested, what then appeared tailed considerably for several months, and an extremely remote possibility, that oxy- coincidental with this change in policy a gen therapy might in some way be related sharp decrease in RLF cases occurred. Of to the disease. more importance to us, he reported that The infants in the premature nursery during this period the smaller premature were examined at intervals in the nursery infants were apparently thriving with much and afterwards in a special clinic during less added oxygen, as they used to do be- a period between July, 1948, and December, fore efficient methods of giving high con- 1950. However, during this period, there centrations of oxygen were available. This were no rigid controls on oxygen administra- gave us the first reassurance that a rigid tion and indeed the only recording of oxy- curtailment of oxygen might be a safe pro- gen that was available for some premature cedune. In January, 1951, our controlled infants was that on the nursing notes where study was started, placing alternate infants the number of days of oxygen therapy was under 3.5 lb birth weight, into either a recorded; occasionally the flow rate was “high” or “low” oxygen routine. It was still also mentioned. with considerable anxiety on our part that Even from this poorly supervised method the small infants who drew a “low” oxygen of administering and recording oxygen ad- routine were entered into that regimen in ministration, there was noted a strong as- the study. The nursing staff also shared this sociation between the number of days of anxiety and it was not uncommon, early in

Downloaded from www.aappublications.org/news by guest on October 1, 2021 506 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

U)

1500 .. . 4 #{149}..#{149} .#{149} #{149} . x ...: .#{149} . . I .eS#{149} #{149}:.... .#{149}S #{149} . #{149} . .#{149}.#{149} . #{149} X #{149} x ...... z . #{149}.#{149} :. #{149} #{149}#{149}I. #{149}. :. : #{149}.s #{149} S #{149} #{149} #{149} #{149}#{149}#{149}#{149}x S I- 1250 - #{149}5. S #{149} S. x I #{149} #{149}#{149} S #{149}#{149}S #{149}#{149} #{149} #{149} x

S #{149} x#{149}#{149}S w S x S x #{149} S #{149}#{149} S x I x I- S #{149}#{149} #{149} #{149}#{149} cr 1000 - x

S

SS NORMAL

x #{149}RI.?

0 5 10 15 20 25 30 35 40 45 50 55 DAYS IN CONTINUOUS OXYGEN FROM BIRTH

ChAnT I. Showing correlation of retrolental fibroplasia with number of days in continuous oxygen according to birth weight. X represents infant with cicatricial retrolental fibroplasia niembranes and dot represents infant with normal eyes. (Nonrandomized data, July, 1948, through I)ecember, 1950.)

the study, to find that the night nurse had (1,590 gm), entered the study during the turned the oxygen on on increased that al- first year and under :3 lb 5 oz (1,500 gui) ready flowing to an infant in the “low” during the second year. group, just to feel more secure. On an alternate admission basis, the in- The controlled study was carried out in fants were placed into either a high oxygemi the nursery from January, 1951, until May, group on a curtailed oxygen group on ad- 1953. At this latter date, we first observed mission to the nursery. Those infants in high in our laboratory ocular lesions that were oxygen received 60 to 70% oxygen concen- indistinguishable from those of early hu- trations for 28 days on longer; whereas, mn RLF following the administration of those in restricted oxygen received concen- oxygen to young animals. Our controlled trations usually under 40% an(l only for nursery data involving 120 infants under specific clinical indications. The nursery 3.5 lb birth weight implicated oxygen and, routine was otherwise identical in the two with the added incrimination of oxygen groups. from the animal data, we felt that our own Oxygen tensions were controlled by imi- controlled study should be terminated. A cubator samplings at 8-hour intervals with brief r#{233}sum#{233}ofthe nursery study which has a panamagnetic analyzer (accurate within

been previously neportedl 3 is cited. 2% concentration). The infants in each group were examimied PREMATURE NURSERY STUDY ophthalmoscopically at regular intervals With five exceptions all infants included and followed after discharge from the nun- iii the controlled investigation were de- seny in a special clinic until they were 6 livered in the District of Columbia General months of age. Hospital Obstetrical Division. Premature The ocular findings were classified to con- infants with birth weights under 3.5 lb form to the standard classification recom-

-* Downloaded from www.aappublications.org/news by guest on October 1, 2021 MvIERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 507

40 7 4 20 U- 2 D HIGH

30 - LOW -J

=

(I) 20 uJ a:

z 10 LU (-) a: 0 LU 0 5 i ci- -- -. --ri-- ----- UNDER 1001-S 1251- 1501- 1000 1250 1500 1590

BIRTH WEIGHT IN GRAMS

CIIAII1 II. Showing striking difference in incidence of cicatricial retrolental (ibm-

l)lL511t l)etWeen higli_ and low-oxygen casts vitliin each bimth ‘eiglt group. 11111111t5 with birth weights over 1,500 gui were included (luring the first year only. (Randomized contmolle(l (hltl, January, 1951, to January, 195:3.) mended by Reese, Owens, and King.4 Chart fants shown in Chart II shows a chi2 of 1.44 II shows the miuml)er of cases in the con- with a p-’altie of 0.01. trolled study develo1)ing advanced residual During the controlled Stu(ly 21 cases of retrolental niembranes (cicatricial stages C, active disease (active stages 1 and 2) whichi D, amid E) according to birth weight after regressed to normal in the high oxygen dropping the 15 infants oven 1,500 gm in group and 9 cases of active disease (active the first year. From this chart one notes stages 1 and 2) which regressed to normiial that 12 of 60 infants with birth weights in the low oxygen group were noted. under 1,500 gin in the high oxygen group This controlled nursery data is now sup- developed advanced retrolental fibroplasia. ported by the controlled study of Lanman In the curtailed oxygen group, 1 of 60 in- and co-workers5 and the larger co-operative fants (IevelOpe(I the advanced disease. This study of Kinsey and associates.’ There are infant, however, received oxygen for 10 also several excellent nonranclomized clays. studies7ul further incriminating oxygen in

There was fl() indication of any pediatric the etiology of retnolental fibroplasia. The difficulties arising from the rigid curtailment report of Gordon and co-workers’ deserves of oxygen administration. No significant special mention in view of the long -follow- difference iii mortality rate was noted be- up and large number of cases involved. tween the infants in the high and in the EXPERIMENTAL OBSERVATIONS low oxygen groups. A statistical appraisal of the data obtained We were mnost fortunate in the expeni- in the first 24 months of the controlled mental phase of this study in finding that randomized study based on these 120 in- at birth the retinas of mouse, rat, kitten and

Downloaded from www.aappublications.org/news by guest on October 1, 2021 508 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

GESTATIONAL AGE OF FETUS

7 Months 6 Months 5 Months ;hS Retinal V

OPTIC Newborn Kitten NERVE Newborn puppy ORA SERRATA 4 day old rat.

0 Schematic drawing showing chronology of retrool vascuIarzaton in human fetus. The retinal vessels advance from the Disc to the Ora between the 4th and 8th months of gestation. The relationship to retinal vessel devel- opment in the rat, cat, and dog are indicated.

Fic. 1. dog were incompletely vascularized and satisfactory solution for protecting the (hiffenentiated so that we were dealing with mother against pulmonary oxygen poisoning retinas comparable to that of the small pre- evolved. This was accomplished by main- mature infant as depicted in Figure 1. Sev- taming a nursing mother animal with her eral experiments were then instituted in young in room air for every litter in the 1951 to evaluate primarily the ocular, but oxygen chamber. At 24-hour intervals the also the systemic effects of added oxygen on mother animals in oxygen were exchanged these species. The newborn opossum in the with those in room air. This provided a 24- pouch stage was also studied. Due to many hour rest in air and eliminated pulmonary pitfalls and failures it was not until May, oxygen poisoning in the mothers. A healthy 1953, that the animal counterpart of human foster mother animal was then always avail- RLF was successfully produced. able to nurse the young animals which were The greatest obstacle during the early maintained continuously in oxygen. In some experiments was pulmonary oxygen poison- experiments the young were divided be- ing in the nursing mother animals. It is tween the two mothers at the start so that well established that oxygen above 70% con- each mother animal was assigned half hen centration after 48 hours causes irritation own litter and half of her control partner’s of the adult human or animal pulmonary litter. This provided litter mates as controls alveolar epithelium. Pulmonary edema, and insured practically identical nutritional hemorrhages and ultimate asphyxia result factors from maternal lactation in oxygen from the pulmonary engorgement. The treated and control young animals in air alveolar epithelium of the newborn animal (Fig. 2). The opossum in the pouch stage or premature infant is much more resistant was studied in view of the extreme pre- to the pulmonary toxicity of oxygen, and maturity of newborn pouch opossums. The these young subjects can withstand oxygen newborn opossum is known to enter the at 70% to 100% concentrations for much maternal pouch in an extremely premature longer periods without alveolar damage state and incubate there for several weeks. resulting. Several months lapsed before a However, the mother animals showed a

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 509

LE:!Th

‘ - ,‘

Qr#{149}0, o ‘‘2 20% 02

Fie. 2. Schematic representation of method of splitting litters between mother animals in oxygen and their control partners in air. The mother animals are exchanged every 24 hours to prevent oxygen poisoning.

very marked susceptibility to pulmonary to identify the species, one cannot dis- oxygen poisoning and a system of exchiang- tinguish the endothehial proliferating nod- ing foster mothers which worked with other ules of the human disease from those seen species failed. Sixty-nine opossum young experimentally produced by oxygen in the (eight litters) who had received from 3 to 5 animal retina. These experimental lesions days exposure to 70% oxygen showed no are located in exactly the same layers of the significant changes in the eyes on other retina, have identical histochemical response organs when compared with controls.13 to periodic acid fuchsin stain and proceed Longer exposures were impractical due to ii) a similar manner of proliferating capil- pulmonary changes in the mothers. lanes from the retina into the vitreous as is In the mouse, rat, kitten and puppy, the seen in human disease. Hemorrhages are characteristic microscopic changes of early readily observed, but not usually to the human retnolental fibroplasia were pro- degree seen clinically. One minor incon- duced.l These changes (Figs. 3 to 14) con- sistency, however, in these animal expeni- sisted of: ments is that the terminal stage of retinal 1. Endothehial nodules in the nerve-fiber detachment of the human disease has not layer of the retina. been produced except in two puppies in 2. Budding of capillaries from the retina our labonatory. This extremely low mci- into the vitreous with formation of capillary deuce in several hundred animals exposed to tufts. oxygen raises the question of trauma or 3. Retinal edemiia. other undetermined factors being invloved 4. Retinal and intraocular hemorrhages. in these two cases. 5. Vitreous degeneration. It is pertinent, however, that Fnieden- It is important to point out that the cx- wald, Owens and Owens14 described the penimental retinal lesions produced are in- proliferative endothehial retinal nodules and distinguishable histologically from those sprouting of capillaries into the vitreous as seen in the retina of early human RLF. In- the essential characteristic pathology of hu- deed, under high power magnification man retrolental fibroplasia. They considered

, where other landmarks are not available the stage of detachment of the retina as non-

Downloaded from www.aappublications.org/news by guest on October 1, 2021 510 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

Fi;. 3 (Upper). Cross-section of retina of normal 21-day-old mouse. Note relatively acellular character of the nerve-fiber layer. ( llematoxylin-eosin; x400.) Fic. 4 (Lower). Section of retina of 21-day-old mouse raised in 70% oxygen since birth. Arrow points to abnormal proliferating endothelial nodules in the nerve-fiber lay-er. (Hematoxylin-eosin; x 400.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 511

specific resulting from organization of hem- nutrition from the embryonic hyaloid yes- orrhage and vessels in the vitreous which sels in the vitreous and by diffusion from detached the retina from its normal position. the adjacent chonoid.2o Starting at the fourth The close identity of oxygen-induced animal month, the vessels from the hyaloid vas- lesions to the primary pathognomonic lesion cular stalk in the optic nerve extend into seen in human RLF is therefore highly the adjacent retina about the optic nerve significant and should give us confidence in and grow forward toward the anterior cx- applying data learned from animal experi- tremity of the retina (ora serrata). It is not, mentation to problems of human RLF. however, until the eighth month of gesta- It is indeed most remarkable that in tion that the vessels reach the ora serrata species as far removed as the mouse and (Fig. 1). Therefore, a premature infant born human that the identical histologic picture at 7 months gestation has the anterior por- in both is produced by oxygen. Equally tion of his retina still devoid of vessels. astounding is that the animal lesions like This avascular portion then becomes vas- those in infants are limited exclusively to culanized ex-utero. It is the incompletely the eyes, are directly proportional to con- vasculanized retina of the premature infant centration and duration of exposure to oxy- that is susceptible to oxygen damage. Once gen, and depend in both cases on an incom- vascularization is complete, added oxygen pletely vasculanized retina for their produc- exerts no injurious effect on the retina. In- tion. deed the common denominator in all our Representative photomicnographs of the animal experiments-mouse, rat, kitten and human and animal lesions are presented in puppy, has been the uniform susceptibility Figures 3 to 8. Retinal sections of normal of the animal retina to oxygen while in- Premature infants and animals are shown completely vascularized and the resistance for control purposes. Note the location of to oxygen damage once vasculanization is the characteristic endothehial nodules in the complete. The correlation between RLF nerve fiber layer in infant and animal eyes. incidence in our nursery and estimated (Figs. 4 and 6) and the similar appearance growth of retinal vessels is seen in Figure of vessels erupting from retina into the 15. This usceptibi1ity of immature retinal vitreous (Figs. 7 and 8). vessels is comparable to the greater suscep- Independent animal data is now avail- tibility of growing bone to vitamin D. One aI)le from three other laboratories. Ashton might reason that just as rickets and scurvy

and co_\vonkersl5, 1; in England, Hellstr#{246}mn are diseases of growing bones, so retrolental

and 718 in Sweden and Gersh- fibroplasia is a disease of growing retinal man and co-workers’9 in this country, have vessels. contributed important data to further sup- The effects of oxygen on the immature port the conclusions drawn from the expeni- retina can be divided into two phases: 1. A ments cited here. The studies of Ashton on primary or initial vasoconstnictive effect; retinal vasoconstniction and Hellstr#{246}m on during this phase there is also a suppression retinal enzymes deserve special recognition. of vasculanization antenionward. 2. A sec- ondary on late proliferative phase. MECHANISM OF ACTION OF OXYGEN The initial effect of oxygen is one of im- Fundamental to an understanding of the mediate vasoconstriction. In the normal reaction of the retinal vessels to oxygen cx- adult retina or cerebrum, administration of posure is an appreciation of the mode of oxygen at concentrations above 50% causes vascularization of the retina. The retina is a slight (approximately 10 to 20%) decrease

unique in that it is well differentiated before in caliber of the 22 In contrast, having a definitive blood supply of its own. the immaturely vasculanized retinal vessels In the human, the retina is avascular until in the animals studied undergo a rapid the fourth month of gestation, deriving its pronounced constriction when these con-

-- - -

Downloaded from www.aappublications.org/news by guest on October 1, 2021 512 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

., r’) ;> -‘ . ,._

Fic. 5 (Upper). Cross section of retina of normal infant, aged 10 weeks. Note normal

appearance of capillaries in the nerve-fiber layer. (Heniatoxylin-eosin; x 400.) FIG. 6 (Lower). Cross section of retina of li-week-old infant in whom early active retrolental fibroplasia was diagnosed clinically. Arrow points to abnormal endothelial nodule characteristic of early human disease. (Hematoxylin-eosin; x 400.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 513

FIG. 7 (Upper). Twenty-four-day-old kitten raised in 70% oxygen from fourth day of life. Arrow points to capillaries proliferating from retina into the vitreous. Note simi- larity of this lesion to human case shown in Figure 8. (Hematoxylin-eosin; India-ink injected specimen; x400.) FIG. 8 (Lower). Section of retina of 14-week-old premature infant with early active retrolental fibroplasia diagnosed clinically. Arrow points to capillaries budding into the vitreous from the retina. (Hematoxylin-eosin; x400.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 514 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

Fmc. 9 (Upper). India ink-injected retina of normal 21-day-old kitten. Note normal pattern of vessels and vascularization of retina is complete to the ora serrata. The retina is cut radially to permit flat preparation (x 8.) Fic. 10 (Lower). Flat retina preparation from 33-day-old kitten exposed to 70% oxygen for first week of life, then removed to room air. Note abnormal pattern of retinal capillaries with numerous tuft formations. (Unstained preparation; x 20.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 515

,>.

:_ ‘t ...

-

‘1,. )

t

-‘.-e._r %I,?r

4-

- 44

Fic. 11 (Upper). Flat retina preparation of 18-day-old rat raised in 70% oxygen from birth. Arrow points to large retinal hemorrhage. (Unstained preparation; X 100.) FIG. 12 (Lower). Flat retina preparation of 16-day-old rat maintained continuously in 70% oxygen from fourth day of life. Arrow points to abnormal capillary tufts resembling glomeruli. (Unstained preparation; x 100.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 FIG. 13 (Upper). Nornial litter-mate control to animal in Figure 14; 21 days of age, raised in room atmosphere. Note clear honiogeneous pattern of vitreous and abseflce of hem- orrhage. The hyaloid vessels have regressed. (Hematoxylin-eosin; x40.) FIG. 14 (Lower). Twenty-one-day-old mouse eye exposed to 60% oxygen first 6 days of life and removed to room air for 15 days. Note abnormal persistence and dense prolifera- tion of hyaloid vessels (H.V.). The vitreous is grossly disorganized and partially filled with hemorrhage (H). Hematoxylin-eosin; x 40.)

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 517

12 20

15 I- z WLL. o-J 0 ci:cx: z Ui 0. I INCID(NC

I- 0 5 a:: 0

6MOS 7MOS 8MOS TERM 650 GM 1080 GM 1800 GM

FIG. 15. This chart shows the correlation between incidence of retrolental fibro- plasia an(l degree of Iliaturation of the retinal vessels based on gestational age. (Nursery data collected between July, 1948, and I)eceniber, 1950.)

centrations of oxygen are administered. Why response to oxygen occurs in all immature immature retinal vessels undergo total con- vessels. Ashton and co-workers24 studied striction and the mnature retinal vessels only the effect of oxygen on the new vessels in minor constriction on exposure of the sub- granulation tissue of the rabbit and in new ject to oxygen remains a mystery. Several vessels in the rabbit cornea where vasculani-

23 were done in our laboratory zation was stimulated by foreign material in an attempt to throw light on this enigma. injected into the anterior chamber. In The results of these studies are briefly neither case did a significant degree of enumerated: vasoconstniction follow ad ministration of 1. Sympathectomy in puppies prior to oxygen. We must conclude that the marked oxygen exposure; results-no effect on re- vasoconstnictive response of the immature tinal vasoconstriction produced by oxygen. retinal vessels is unique and not simply a 2. Vasodilator drugs-no effect on retinal property of immature vessels. vasoconstniction produced by oxygen. Our experiments25 with an oxygen dcc- 3. Massive doses of vitamin E, an anti- trode placed at the internal surface of the oxidant-no effect on retinal vasoconstnic- retina gives one clue to the vasoconstrictor tion. question. Following severance of the re- 4. Cobalt chloride, based on Dickens tinal artery in the cat the retinal blood flow studies2#{176} that cobalt protected brain slices was totally eliminated. When the animal from oxygen poisoning-no effect on retinal breathed oxygen, the range of oxygen dif- vasoconstniction produced by oxygen. fusion from the choroid was extended to 5. Combining 5% carbon dioxide with reach the internal layers as shown polara- 95% oxygen-no change in retinal vascon- graphically with the electrode. This kept the stniction. Yet in the same animal the carbon internal retinal layers oxygenated even dioxide effect predominated over oxygen though retinal blood flow had ceased. elsewhere and a visible vasodilatation re- Chonoidal diffusion of oxygen apparently suIted in the cerebral vessels as seen furnishes the means of sustaining the vaso- through a cerebral window. One might contniction that occurs in the immature suspect that this exaggerated vasoconstricton retinal vessel on oxygen exposure. If the

Downloaded from www.aappublications.org/news by guest on October 1, 2021 518 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA immature retina is detached from the in the retina. The primary or initial effect choroid while the animal is in oxygen, the of oxygen leads to the later proliferative or constricted vessels in the detached portion secondary stage which appears to be a com- dilate imnmediately whereas in the intact pensatory overgrowth of the remaining re- retina, in its normal position against the tinal vessels to the depleted retina. choroid, the vessels remain in a state of The secondary or later vasoprolifenative vasoconstriction. phase may occur under two cincumnstances. Still unanswered is why the immature First, proliferation can occur while the in- retinal vessel totally constricts on inhalation fant on animal is kept in oxygen continu- of oxygen while the retina that has just ously for a prolonged period. The second- become fully vasculanized fails to show ary or late proliferative phase mnay also this marked vasoconstniction to oxygen. occur where vaso-obliteration has been initi- Animnal experiments3 show that where the ally produced by oxygen and the infant or exposure to oxygen is of relatively short animal removed to room air while sonic of duration, the vasoconstnictive effect of oxy- the retinal vessels are still obliterated. An gen is reversible. However, if vasoconstnic- abnormal proliferation of the remaining tion following exposure to oxygen lasts for capillaries results. The removal to air may approximately 2 days or longer, a true theoretically add an oxygen deficit to time obliteration of some of the smaller vessels retina already depleted by vasoconstniction occurs. Ashton6 has suggested that the op- of its normal blood-borne constituents. A posing endothelium of the vessel walls be- schematic representation of the primary comes adherent on an intravascular clotting vasoconstrictive phase and the secondary occurs with obliteration of the lumen. It is proliferative changes is presented in Figures apparent that during the phase of vasocon- 16 and 17.

striction and obliteration a decrease in net- Hellstr#{246}m2’ 27 has studied the in-vitro inal blood flow results. One cami assume metabolism of the netimia of oxygen exposed that all blood-borne constituents, e.g., glu- rat-hugs and found no change suggesting cose, normally supplied by the retinal a metabolic injury of respiration of the ret- sels are thereby depleted. Also diminished ma in his oxygen treated animals. Like- retinal blood flow favors an accumulation wise lie found in histochemical studies that of metabolic end-products, e.g., lactic acid, there was no evidence of a change in the Effect of Oxygen rt RLF

Primary Secondary

FIG. 16. Note in the primary stage that the retinal vessels are niarkedlv con- stricted. Iii the secondary stage proliferating capillaries in the retina erupt through the retinal surface into the vitreous.

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 519

vasoconstriction

2 \AsoobIiteration \/ DepletkxiRetentn of end-products metabolites (RemoveDeetes to02 air)

soprohferatkn

FI(;. 17. I)iagranmiatic representation of theory of oxygen action on inimature retina. The secondary vasoproliferative stage is depicted to occur while in oxygen. Removal of subject to air when vessels

are in obliterated stage is not necessary to produce vasoproliferation but may influence this reaction.

localization of succinic dehydnogenase ac- tion of niarkedly constricted retinal vessels tivity following oxygen exposure. He has should certainly alert the examiner to the stated that his studies are not adequate to possible later development of proliferative rule out conclusively a direct histotoxic retinopathy. effect of oxygen at the cellular level. Further The hanger vessels near the optic disc are studies are needed along these lines in view usually not completely obliterated by oxy- of the well established histotoxic effect of gen during the initial exposure to otygen oxygen based on studies of Stadie,28 and they undergo a marked dilatation and Dickens,29 Bean3#{176}and hm’ increased tortuosity during the proliferative It is important to translate these animal phase. New vessels are seen in the peniph- observations to clinical findings. The stages ery sprouting from the retina into the of retinal vasoconstniction and obliteration vitreous. Apparently these new vessels and seen in experimental animals is also seen in the existing older vessels are abnormally the eyegrounds of premature infants. How- permeable, as hemorrhages and edema usu- ever, the standard classification of Reese, ally follow. Organization of the hemor- Owens and King4 which is now in wide- rhages and vessels that have invaded the spread usage includes only the secondary vitreous produce traction on the retina with proliferative retinopathy. This is readily cx- detachment resulting. The leakage of serum plained by the origin of this classification on hemorrhage behind the retina augments which is based on the changes that are detachment (Figs. 18 and 19). clinically observed from about the third week of life. However, in our nursery FACTORS INFLUENCING OXYGEN ACTION ON THE PREMATURE RETINA 3 where the eyegnounds were cx-

amined during the first few days of life, a DURATION OF EXPOSURE: Of all factors in- large number of infants showed marked fluencing the amount of oxygen damage to generalized retinal vasoconstriction while the premature retina, duration of exposure receiving oxygen. As retinal vasoconstnic- has been affirmed experimentally to be the tion is fundamental to the development principal one. Recent experiments on new- of netrolental fibroplasia, the observation of born mice22 in our laboratory show that advanced retinal vasoconstriction that per- even where concentrations of 35 to 40% sists while the infant is out of the incubator oxygen (supposed safe levels of oxygen for examination should be considered as a therapy) are given for periods of 10 days, preliminary stage of the classic disease. It a significant number of the animals showed is impractical to add another stage to the ocular hemorrhages and other evidence of five now in use as they have already fur- damage from the oxygen exposure (Table I). nished an excellent standard for mass Yet for shorter exposure these concentra- quantitative study. However, the observa- tions produced no visible effect whatsoever

Downloaded from www.aappublications.org/news by guest on October 1, 2021 520 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

1

11G. 1 S. Cross section of c-se ill terniinal stage of retrolental fibroplasia. Arrow 1)oints to (lisorganized retina vhich is totally (letaclied and against ps(rir -airfaur of lells. This stage woulrl appear clinically as in Figure 19.

on the iiiimnature retina. It is possible that the animal retina and no effect whatsoever.2 1)relnattlre infant is even more sensitive to CONCENTRATION OF OXYGEN: Our data oxygen damage. \Vhene high concentrations, and those of AshtonlG and Hehlstrdm,15 for example 80% oxygen, are administered showed that the severity of oxygen lesions in carefully controlled studies, the differ- produced is directly proportional to the ence in duration of exposure of only a few concentration of oxygen used. For short hours frequently represents the difference periods of exposure of the animal, under 2 l)etween severe damage to the premature to 3 days, concentrations under 40% pro-

FIG. 19. Terminal stage of retrolental fibroplasia with completely detached retinas presenting as white membranes filling the pupillary area. This child is totally blind.

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 521 duced 110 visible effect. Our studies on kit- ized the retina, and a resistance to oxygen tens show that exposure to oxygen at con- damage once the retina is matured or fully centnations above 50% for S on 4 days pro- vascularized.3 duces lesions directly proportional in sever- RATE OF WITHDRAWAL FROM OXYGEN: ity to concentration up to approximately Szewczyk3 and Bedrossian and co-workers3’ 80 There is very little difference ob- have claimed that prolonged gradual with- served on increasing the concentration from drawal from oxygen reduces on prevents 80% to 100% oxygen. the damaging effect of the initial exposure

MATURATION OF THE RETiNA: In general, of the premature infant to oxygen. Recent it has been demonstrated that the more im- controlled studies on mice in our own mature the animal retina, the greaten the labonatory2 and by HellstrOm35 in Sweden susceptibility to oxygen effect. The common have shown that gradual withdrawal from denominator to all the animal experiments oxygen does not prevent ocular lesions de- has been a uniform susceptibility to oxygen veloping in the mouse eye. In our own stud- where the animal retina is immature, that is, ies it was demonstrated that, on the con- where the vessels have not fully vascular- trary, gradual withdrawal frequently in-

Withdrawal From Oxygen OCULAR LESiONS (mice) - Rapid4 =0% l00 ---- GraduaI- =100%

C

01 > 0

air 20- -

0 2’t4 6 8 10 12 14 16 18

Days in #{176}2

FIG. 20. Showing effect of rapid and gradual withdrawal from 3-day stay of new- born mice in 80% oxygen. Note that immediate removal to air resulted in total ocular protection. Gradual reduction in oxygen over B (lays furnished sufficient a(l(litional oxygen to make total stay in oxygen damaging to l00 of animals.

‘I’ABLE I

‘rIlE EFFE(:Ts OF LONG ExposunEs TO OXYGEN OF MOI)EIIATE CONCENTRATIONS

ON TIlE EYE OF THE NEWBORN MOUSE

Number of Number of Number of Days in Mean Oxygen . Ocular . Eyes with Animals Eyes Oxygen Concentra#{252}on Le.,ionx hemorrhages

15 30 5 40% 0 0 9 18 7 4% 0 0 0 40 10 38% 4 10 % U 4 14 40% 5 20. 8%

Downloaded from www.aappublications.org/news by guest on October 1, 2021 522 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA creased both the severity and frequency of velop in a premature infant from this rapid ocular lesions. These mouse experiments increase in oxygen tension, amid could ex- showed that the additional oxygen that plain an occasional case of RLF developing was administered in the period of gradual in a sensitive retina where this relative in- withdrawal from oxygen frequently repre- crease in oxygen tension caused vasocon- sented the added amount of oxygen that was striction and ultimate vasoproliferation. necessary to make an initial nontoxic cx- RETROLENTAL FIBROPLASIA IN FULL posure injurious to the animal eye (Fig. 20). TERM INFANTS These animal data suggest the possibility that this same situation may occur during The occurrence of an occasional case of prolonged gradual withdrawal of a prema- typical retrolental fibroplasia in a full-term tune infant from oxygen. infant challenges the validity of our funda-

I NTERMIrrENT Exposum TO OXYGEN : Al- mental assumption that only the immaturely though we have performed no positively vasculanized retina is susceptible to the dis- controlled study on continuous versus in- ease. A significant variation in stage of termittent oxygen, our studies on duration retinal vessel growth for a given age animal of exposure show that a given concentration of the same litter was noted in our labora- of oxygen is much less injurious when given tory on several occasions.23 It is reasonable intermittently than when the same level of to assume that an occasional full-term infant oxygen is given by continuous exposure. would have a small zone of the anterior retina still incompletely vascularized. This FACTORS OTHER THAN OXYGEN assumption is supported by the actual ob- The development of an occasional case servation of an incompletely vascularized of RLF when infants received little or no retina in cross sections of two full-term, ap- supplementary oxygen caused considerable parently normal eyes in our laboratory.22 skepticism in accepting oxygen as a factor The greaten predilection for retrolental in the disease, early in these studies. In our fibroplasia to involve the temporal quad- own nursery study over a period of 8 years, rants of the retina also supports the basic not a single case in approximately 500 premise of oxygen susceptibility of only premature infants under 1500 gm birth retinal immature vessels, as the anterior weight occurred where a significant amount zone of the temporal retina is the last por- of oxygen was not administered. There are, tion of the human retina to become com- however, well-documented cases3#{176}of an pletely vascularized. exceptionally rare instance of RLF occur- ring without use of supplemental oxygen. RECOMMENDATIONS FOR NURSERIES I would prefer to explain these rare cases The following recommendations would on factors other than oxygen. One may seem to be prerequisite for adequate safe- logically assume that many as yet unnecog- guards in the administration of oxygen in nized factors can alter normal growth of the the premature nursery. immature retinal vasculature which in the 1. A program of orientation on the po- premature infant matures ex-utero in con- tential ocular dangers of oxygen for the tnast to its normal intrauterine environment. premature infant should be frequently prac- However, for those who would prefer ticed in the nursery and all nursery and stretching the oxygen theory to attempt an medical personnel working in the premature explanation of these cases, it is pertinent nursery should be acquainted with the to point out that while oxygen saturation ocular toxic effects of oxygen. of arterial blood is approximately 50% in 2. Every nursery should be equipped utero, after birth arterial saturation rises with an oxygen analyzer. while in room air without added oxygen to 3. A special oxygen chant or some system approximately 90% saturation.37 Retinal of recording on the standard hospital chant vasoconstriction might, theoretically, de- should be provided.

Downloaded from www.aappublications.org/news by guest on October 1, 2021 AMERICAN ACADEMY OF PEDIATRICS - PROCEEDINGS 523

4. Except for emergency use, oxygen tify recommendations for a rigid supervision therapy should require a doctor’s order in of oxygen administration to the premature the nursery and oxygen should be ordered infant to avoid any unnecessary overuse of by concentration in stead of flow-nate. this potentially toxic agent. 5. The concentration of oxygen ordered should be, in general, the minimal concen- ACKNOWLEDGM ENT tration required to give a satisfactory clini- Figures 6, 7, 8, 9, 11, 15 were originally cal response. Attempts should be made to published by the author in the American terminate on reduce the concentration of Journal of Ophthalmology (see references 3 oxygen as soon as practical. It cannot be and 13). Permission has been granted for overly stressed that in a sensitive premature reproduction of these illustrations here. infant exposures for as long as 2 to 3 days at moderate concentrations may be injurious REFERENCES to the premature retina. 1. Patz, A., Hoeck, L. E., and De La Cruz, E. : Studies on the effect of high oxygen SUMMARY AND CONCLUSIONS administration in retrolental fibroplasia; The results of a controlled nursery study, nursery observations. Am. J. Ophth., 35:1248, 1952. supported by observations of others in both 2. Kinsey, V. E., and Zacharias, L. : Retro- uncontrolled and controlled studies, clearly lental fibroplasia; incidence in different established the overuse of oxygen in the localities in recent years and correlation premature nursery as an important and of incidence with treatment given in- probably the principal factor in the develop- fant. J.A.M.A., 139:572, 1949. 3. Patz, A. : Oxygen studies in retrolental ment of retrolental fibroplasia. fibroplasia. IV. Clinical and experimental The clinical data is supported by obsenva- observations. Am. J. Ophth., 38:291, tions in the newborn or young mouse, rat, 1954. kitten, and puppy where retinal lesions, 4. Classification of retrolental fibroplasia. Am. identical to those seen in early human J. Ophth., 36:1333, 1953. retnolental fibroplasia are produced follow- 5. Lanman, J. T., Guy, L. P., and Dancis, J.: Retrolental fibroplasia and oxygen then- ing exposure to an enriched oxygen en- apy. J.A.M.A., 155:223, 1954. vironment. 6. Kinsey, V. E., and Hemphill, F. M. : Etiol- Fundamental to all these experiments, and ogy of retrolental fibroplasia and pre- apparently to the pathogenesis of human liminary report of co-operative study of netrolental fibroplasia, is the uniform ocular retrolental fibroplasia. Tr. Am. Acad. Ophth., 59:15, 1955. susceptibility to oxygen when the retinas 7. Crosse, V. M., and Evans, P. J. : Preven- are immaturely vasculanized, and a resist- tion of retrolental fibroplasia. A.M.A. ance to oxygen damage where vasculaniza- Arch. Ophth., 48:83, 1952. tion is complete. 8. Campbell, K. : Intensive oxygen therapy as The animal lesions produced by oxygen a possible cause of retrolental fibro- plasia: clinical approach. M. J. Austra- are proportional to the concentration of oxy- lia, 2:48, 1951. gen administered and the duration of ex- 9. Locke, J. C. : Retrolental fibroplasia; de- posune, and inversely proportional to the de- finitive role of oxygen administration gree of vascularization of the retina. Dura- in its etiology. A.M.A. Arch. Ophth., lion of exposure to oxygen proved to be the 51:73, 1954. 10. Ryan, H. : Retnolental fibroplasia; clinico- most important single factor in these expeni- pathologic study. Am. J. Ophth., 35:329, ments. 1952. The ocular effects of oxygen on the imma- I 1 . Goldman, H., and Tobler, W. : Etiology of ture retina can l)e divided into a primary retrolental fibroplasia. Schweiz. med. phase of vasoconstniction and obliteration Wchnschr., 82:381, 1952. and a secondary phase of vascular prolifera- 12. Gordon, H. H., Lubchenco, L., and Hix, I.: Observations on the etiology of retro- tion. lental fibroplasia. Bull. Johns Hopkins These clinical and experimental data jus- Hosp., 94:34, 1954.

Downloaded from www.aappublications.org/news by guest on October 1, 2021 524 PATZ - ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA

13. Patz, A., Eastham, A., Higgenbotham, 25. Patz, A. : Oxygen inhalation in retinal D. H. and Kleh, T. : Oxygen studies in arterial occlusion; a preliminary report. retrolental fibroplasia; production of the Am. J. Ophth., 40:789, 1955. microscopic changes of retrolental fibro- 26. Hellstn#{246}m, B. E. : X in vitro metabolism of plasia in experimental animals. Am. J. the retina of oxygen-exposed ratlings. Ophth., 36:1511, 1953. Acta path. et microbiol. scandinav., in 14. Fniedenwald, J. S., Owens, W. C. and press. Owens, E. U. : Retnolental fibroplasia 27. Hellstn#{246}m, B. E. : IX The histochemical in premature infants; pathology of the localization of succinic dehydrogenase disease. Tr. Am. Ophth. Soc., 49:207, in the retina of normal and oxygen- 1952. exposed animals. Acta path. et microbiol. 15. Ashton, N., Ward, B., and Serpell, C.: scandinav. , in press. Role of oxygen in the genesis of retro- 28. Stadie, W. C., Riggs, B. C., and Haugaard, lental fibroplasia; preliminary report. N. : Oxygen poisoning. Am. J. M. Sc., Brit. J. Ophth., 37:513, 1953. 207:84, 1944. 16. Ashton, N., Ward, B., and Serpell, G.: 29. Dickens, F. : Toxic effects of oxygen on Effect of oxygen on developing retinal brain metabolism and on tissue enzymes; vessels with particular reference to the problem of retrolental fibroplasia. Bnit. brain metabolism. Biochem. J., 40:145, 1946. J. Ophth., 38:397, 1954. 17. Gyllensten, L. J., and Hellstr#{246}m, B. E.: 30. Bean, J. W. : Effects of oxygen at increased Retnolental fibroplasia; animal expeni- pressure. Physiol. Rev., 25: 1, 1945. ments-the effect of intermittingly ad- 31. Genschman, R. : Studies on the mechanism ministered oxygen on the postnatal de- of oxygen poisoning, in Report of the velopment of the eyes of fullterm mice; Sixteenth M. & R. Research Confer- a preliminary report, Acta paediat., ence: Retrolental fibroplasia. Columbus, 41:577, 1952. M & R Laboratories, 1955, p. 50. 18. Gyllensten, L. J., and Hellstr#{246}m, B. E.: 32. Patz, A. and Eastham, A. : Effect of rapid Experimental approach to the patho- versus gradual withdrawal from oxygen genesis of retrolental fibroplasia. II. The on the mouse eve. Arch. Ophth., in influence of the developmental maturity press. on oxygen-induced changes in the mouse 33. Szewczyk, T. S. : Retrolental fibroplasia eye. Am. J. Ophth., 39:475, 1955. and related ocular diseases; classifica- 19. Gerschman, R., Nadig, P. W., Snell, A. C., tion, etiology and prophvlaxis. Am. J. and Nye, S. W. : Effect of high oxygen Ophth., 36:1336, 1953. concentrations on eyes of newborn mice. 34. Bedrossian, R. H., Carmichael, P., and Am. J. Physiol., 179:115, 1954. Ritter, J. : Retinopathy of prematurity 20. Mann, I. : Development of the Human Eye. (retrolental fibroplasia) and oxygen: New York, Macmillan, 1928. Clinical Study; further observations on 21. Wolff, H. G., and Lennox, W. G. : Effect disease. Am. J. Ophth., 37:78, 1954. on pial vessels of variations in the ox’- gen and carbon dioxide of the blood. 35. Gyllensten, L. J., and Hellstr#{246}m, B. E.: Arch. Neurol. & Psychiat., 23:1097, Iv. The effects of gradual and of rapid 1930. transfer from concentrated oxygen to 22. Patz, A. : Unpublished data. normal air on the oxygen-induced 23. Patz, A. : Experimental studies. Am. J. changes in the eyes of young mice. Am. Ophth., 40:174, 1955. J. Ophth., 41:619, 1956. 24. Ashton, N., and Cook, C. : Direct observa- 36. Owens, W. C. : Personal communication. tion of the effect of oxygen on develop- 37. Smith, C. A. : The Physiology of the New- ing vessels; preliminary report. Bnit. J. born Infant. Springfield, Thomas, 1951, Ophth., 38:433, 1954. pp. 19, 20, and 82.

Downloaded from www.aappublications.org/news by guest on October 1, 2021 THE ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA: E. Mead Johnson Award Address Arnall Patz Pediatrics 1957;19;504

Updated Information & including high resolution figures, can be found at: Services http://pediatrics.aappublications.org/content/19/3/504 Permissions & Licensing Information about reproducing this article in parts (figures, tables) or in its entirety can be found online at: http://www.aappublications.org/site/misc/Permissions.xhtml Reprints Information about ordering reprints can be found online: http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news by guest on October 1, 2021 THE ROLE OF OXYGEN IN RETROLENTAL FIBROPLASIA: E. Mead Johnson Award Address Arnall Patz Pediatrics 1957;19;504

The online version of this article, along with updated information and services, is located on the World Wide Web at: http://pediatrics.aappublications.org/content/19/3/504

Pediatrics is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. Pediatrics is owned, published, and trademarked by the American Academy of Pediatrics, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 1957 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news by guest on October 1, 2021