Lessons From the E-Ferol Tragedy

(1)

EDITORIAL

Lessons

From

the

E-Ferol

Tragedy

William

F. Balistreri,

MD,

Michael

K. Farrell,

MD, and

Kevin

E. Bove,

MD

From the Division of Pediatric Gastroenterology and Nutrition and Division of Pediatric Pathology, Children’s Hospital Research Foundation, Cincinnati

“Those who cannot remember the past are condemned

to repeat it. “-G. Santayana

Several factors combined to suggest that

supple-mental vitamin E should be administered to low

birth

weight infants. The persistent concern and

controversy, the latter confounded by a paucity of data, have been discussed in recent editorials.”2 At

birth,

tissue stores of the naturally occurring lipid-soluble antioxidant vitamin E (ct-tocopherol) are low. The amount of total tocopherol in the tissue

of premature infants is approximately one half that

of full-term infants.3 Maternal vitamin E supple-mentation seems to have minimal effect on serum

vitamin E levels in the newborn because there is poor placental transfer; maternal blood levels are higher than cord levels.13 Prematurely born infants

are exposed to considerable oxidant stress. A major

biologic function of vitamin E is as an antioxidant

to prevent destruction of unsaturated fatty acids

and cell membranes by uncontrolled free radical

attack. It was, therefore, postulated that pharma-cologic doses of vitamin E might ameliorate some

of the complications of modern neonatal intensive

support. Vitamin E supplements were envisioned as a means to decrease the toxicity of hyperoxia and to possibly reduce the incidence of major prob-lems encountered by the newborn: retinopathy of

prematurity, CNS hemorrhage, bronchopulmonary

dysplasia, and hemolytic anemia.1’2’4 Seemingly the

only issue was to determine the optimal dose and

route of administration. Intramuscular injections were ineffective in view of the decreased muscle mass and poor absorption noted in these infants;

local reactions are not uncommon. The oral route was also ineffective; in the presence of physiologic

steatorrhea, absorption from the intestine was un-reliable. Moreover, there was some suggestion that

enteral administration may be unsafe.5’6

Therefore, the appearance on the market in the fall of 1983 of an intravenous preparation of vita-mm E (E-Ferol) appeared to provide a much needed

approach to vitamin E supplementation. However, this product did not solve the problem; indeed, the

catastrophe that ensued raised new issues and

con-cern. The product was hurried into production,

allegedly bypassing standard regulatory processes because it apparently was viewed as a new form of an old product or as a “nutritional supplement” and not as a drug. No clinical trials regarding safety or efficacy were carried out. The product was imme-diately accepted and used in many medical centers throughout the United States. Subsequent events

have shown that E-Ferol, which contains (per

mu-liliter) 25 USP units of dl-a-tocopheryl acetate solubilized in a mixture of polysorbate 80 (9%) and polysorbate 20 (1%), is toxic in small infants. The product was stocked in 159 hospitals according to

an early survey; 62 of these hospitals never used

E-Ferol and 62 reported “no problem” associated with

its use. However, 35 hospitals reported side effects

in a total of 81 cases. There were 38 deaths reported

from 11 states and 43 other infants sustained

seri-ous effects.7’#{176} Recognition of unusual

complica-tions occurred shortly after the widespread use of

this product. Clinicians in several neonatal inten-sive care units around the country noted that a

(2)

504 E-FEROL

TRAGEDY

hepatomegaly, cholestasis, ascites, and metabolic

acidosis.7’8”#{176} An epidemiologic investigation was

carried out by the Centers for Disease Control

following the initial report from Good Samaritan Hospital in Cincinnati.9 Similar clusters of cases of the unusual illness were subsequently reported.

Af-fected infants shared several epidemiologic and clinical features. All had a low birth weight (<1,500 g) and nutritional support had been maintained via total parenteral nutrition. All had been supple-mented with E-Ferol. In April 1984, the Food and

Drug

Administration requested withdrawal of the product from the market.

We reviewed the data of all infants from one nursery who had received E-Ferol. The syndrome was recognized in eight of the 36 recipients.7’8 Corn-pared with unaffected recipients, infants affected

with the syndrome had a significantly lower birth

weight and had received a higher total dose of E-Ferol for longer periods of time.7’8 Autopsy-derived

tissue was available for 20 infants who had the

clinical features of the E-Ferol syndrome; six were from the index nursery and 14 were from three other collaborating nurseries in the United States. These 20 infants had received E-Ferol in a reported

dose of 25 to 137 U/kg/d for six to 45 days. The

hepatic histology was striking and dissimilar from

that usually noted in infants stated to have

paren-teral nutrition associated cholestasis.”12 The lesion did not include features ascribed to fat overload.

Infants with the E-Ferol syndrome had a striking

vasculocentric hepatotoxic reaction which appeared to be of a progressive nature. This was character-ized, in its early stage, by Kupffer cell exfoliation, centrolobular accumulation of cellular debris, and panlobular congestion, more prominent in central

areas. Infants who had received prolonged

admin-istration of the vitamin E supplement had

intralob-ular cholestasis, inflammation of hepatic venules, and extensive sinusoidal venoocciusion by fibrosis. The progressive nature was further suggested by

the fact that liver dysfunction and abnormal

his-tology persisted in eight patients despite with-drawal of the E-Ferol. This suggests that the intra-hepatic injury was irreversible and that perhaps the hepatotoxic component accumulated with time. The described hepatic histologic changes can well account for important components of the clinical

syndrome. The renal dysfunction may have been

related to poor perfusion; however, the hypotension

noted

in these infants remains unexplained.

There is little question that a strong argument can be made for a specific relationship between the introduction of the intravenous vitamin E prepa-ration, E-Ferol, and the time-clustered outbreaks

of a distinctive clinical and pathologic syndrome in

several geographically dispersed areas throughout

the United States. These features, the near

simul-taneous occurrence and the stereotypic pattern, rule

out a local “accident” such as the inadvertent sub-stitution of salt for sugar in milk formulation’3 or of epinephrine for vitamin E.14 The question re-mains-what is the pathogenesis of the syndrome

and what have we learned from this tragedy? Three possibilities exist: (1) that the vitamin E content

was responsible for the toxicity, (2) that the emul-sifier, polysorbate, initiated the sequence of events,

or (3) that a contaminant of the preparation was

responsible.

There are few data to support the concept that

vitamin E, both tocopherol and tocopheryl acetate, is safe for use in premature infants.’5”6 Large doses

have been given orally to adults, and no adverse effects have been noted.17 However, E-Ferol was

proposed as beneficial and effective with no data

regarding pharmacokinetics or tissue levels. Re-cently published data regarding vitamin E phar-macokinetics are provocative in this regard.’2#{176} It

is clear that the form and the route of administra-tion are crucial in the determination of the ultimate tissue concentration and utilization. When given

orally to newborn rabbits, a-tocopherol in alcohol

and a-tocopheryl acetate result in similar tissue

concentrations. When a-tocopherol is given

intra-venously, high hepatic concentrations are noted.18 Following subcutaneous administration, there is in-creased tocopherol concentration in all tissues, es-pecially the liver; tissue concentrations continue to

increase throughout the dosing interval. However, the intravenous administration of ctt-tocopheryl

ace-tate results in a high pulmonary concentration of

tocopheryl acetate and, after periods as long as six days, only 10% is present as free tocopherol. The

lack of free tocopherol following intravenous

administration of tocopheryl acetate may be due to

a relative lack of esterase activity. These findings

may be related to known features of vitamin E disposition. Specific binding of a-tocopherol occurs in the liver; a hepatic cytosolic binding protein has

been described.21’22 The hepatic uptake of a-to-copherol has been associated with low-density

ii-poprotein receptors.23 Similar hepatic binding and uptake of a-tocopheryl acetate may not occur. The resultant high levels of a-tocopheryl acetate, which is an inefficient antioxidant compared to free

to-copherol, may be deleterious.2426 Hepatic tissue

obtained at autopsy from infants who had received repeated doses of E-Ferol was found to have ex-ceedingly high levels of vitamin E in the range of

4,000 tg/g of tissue.’9 Hepatic accumulation of

a-tocopherol continues even after serum concentra-tions have reached a steady state, suggesting that

at Viet Nam:AAP Sponsored on September 7, 2020

www.aappublications.org/news

(3)

serum concentrations are not predictive of tissue levels.’9’20 A functional assay of vitamin E status, eg, erythrocyte malondialdehyde release, may be

required.27

In the April issue of Pediatrics, Alade et al present evidence that E-Ferol suppresses the re-sponse of human lymphocytes to phytohemagglu-tinin. They have further shown that this

suppres-sion was not due to the vitamin E component but

that the responsible factor was polysorbate 80. Polysorbates represent a series of

polyoxenethy-lated sorbitan esters which have been used

exten-sively as hydrophilic and nonionic surfactants.29

There is insufficient evidence to indict as toxic all preparations that contain polysorbate. It must be

determined what component of polysorbate is

del-eterious, ie, the fatty acid moiety or other constit-uents. It is possible, but not likely, that immuno-suppression induced by polysorbate initiates the

cascade, eventuating in death. Overwhelming

op-portunistic infection was not a feature in these patients. The mechanism of toxicity may, therefore, be a more generalized one. We considered alteration

of membrane fluidity in other cell lines (eg, cells of

vessel walls) leading to changes in structure and function. A mixture of polysorbate 80 and

polysor-bate 20 was present in E-Ferol, and large doses were administered as a component of the

supplementa-tion. Indeed, a high level ofpolysorbate 80 (100 tg/

mL) was found in ascitic fluid obtained from an

affected infant in the index nursery long after dis-continuance of E-Ferol.3#{176} This also suggests that

polysorbate may accumulate as a result of an

alter-ation in the metabolism of this compound by low

birth weight infants.

Provocative data regarding a possible mechanism of polysorbate toxicity have been reported by

Varma et al,31 who showed a dose-dependent hy-potensive effect following intravenous

administra-tion of polysorbate to dogs. Polysorbate 20 may exert a hypotensive effect through release of

endog-enous histamine stores and by increasing capillary

permeability.3234 Chronic oral administration of polysorbate 80 to rats has caused congestion and dilatation of central veins and sinusoids in liver

and possible capillary wall damage.35 Millard et a!36 gave polysorbate 80 to dogs and noted changes in cardiac dynamics and decreased blood pressure. Many years ago, Kellner et a!37 noted that

polysor-bate 80 given intravenously to rabbits caused de-generation and necrosis of the liver; the levels given

were similar to those administered to the low birth

weight infants in whom the E-Ferol syndrome de-veloped. On the basis of the literature related to polysorbate effects in experimental animals, it would have been reasonable to suspect that the

administration of large doses intravenously might have adverse effects.

“The cautious seldom err. “-Confucius

The association of hepatotoxicity, hypotension, and renal failure with the intravenous

administra-tion of an emulsion of vitamin E in polysorbate

should rekindle concern regarding any product

ad-ministered as part of an intravenous nutrition pro-gram to infants. We have recently noted a similar

hepatic lesion in an infant who did not receive

E-Ferol but had received prolonged intravenous

hy-peralimentation that contained a multivitamin so-lution emulsified in smaller amounts of polysorbate.

Given the amply documented history of ill effects in low birth weight infants due to administration

of chloramphenicol,m benzyl alcohol,3#{176} and propyl-ene glycol,4#{176}we believe that no

drug

or substance should be considered for use in premature infants

until the question of safety has been thoroughly

examined.

Much more research into the metabolism,

distri-bution, and pharmacokinetics of vitamin E is needed in low birth weight infants. In addition,

careful scrutiny of the beneficial and possibly toxic effects of the emulsifiers such as polysorbate must be very carefully assessed. If the basic principles of documentation of efficacy or toxicity had been

fol-lowed, the E-Ferol disaster might have been

avoided.

It is easy to appear wise during retrospection. We believe that an initial clear definition of the

prob-lem, documentation of the benefit, and

determina-tion of toxicity in experimental animals, followed

by a controlled clinical trial, might have answered

many of the questions that have subsequently been raised. At present, we are still perplexed and ques-tion “what went wrong?” We should have learned

506 E-FEROL

TRAGEDY

7. Balistreri WF, Bove KE, Kosmetatos N, et a!: Distinctive hepatopathy in low birthweight infants in association with

E-Ferol infusion. Pediatr Res 1985;19:212

8. Bove KE, Kosmetatos N, Wedig KE, et a): Vasculopathic hepatotoxicity associated with E-Ferol syndrome in low-birth-weight infants. JAMA 1985;254:2422-2430

9. Unusual syndrome with fatalities among premature infants: Association with a new intravenous vitamin E product.

MMWR 1984;33:198-199

10. Bodenstein CJ: Intravenous vitamin E and deaths in the

intensive care unit, letter. Pediatrics 1984;73:733 11. Bernstein J, Chang C-H, Brough AJ, et al: Conjugated

hyperbilirubinemia in infancy associated with parenteral hyperalimentation. J Pediatr 1977;90:361-367

12. Cohen C, Olson MM: Pediatric total parenteral nutrition: Liver histopathology. Arch Pathol Lab Med

1981;105:152-156

13. Finberg L, Kiley J, Luttrell CN: Mass accidental salt poi-soning in infancy: A study of a hospital disaster. JAMA 1963;184:187-190

14. Solomon SL, Wallace EM, Ford-Jones EL, et a): Medication

errors with inhalant epinephrine mimicking an epidemic of neonatal sepsis. N Engi J Med 1984;310:166-170

15. Chiswick ML, Johnson M, WOOdhaII C, et al: Protective

effect of vitamin E (dl-a-tocopherol) against

intraventricu-lar haemorrhage in premature babies. Br Med J 1983;287:81-84

16. Sobel S, Gueriguian J, Troendle G, et a): Vitamin E in retrolental fibroplasia. N Engi J Med 1982;306:867

17. Farrell PM, Bieri JG: Megavitamin E supplementation in

man. Am J Clin Nutr 1975;28:1381-1386

18. Knight ME, Roberts RI: Disposition of intravenously

ad-ministered pharmacologic doses of vitamin E in newborn

rabbits. J Pediatr 1986;108:145-153

19. Roberts RI, Knight M, Mortensen M, et al: Vitamin E

content of tissues obtained from human infants given phar-macoloic doses of tocopherol or tocopheryl acetate mtra-venously. Pediatr Res 1985;19:178

20. Knight ME, Roberts RJ: Pharmacologic dosing of vitamin E in newborn rabbits: Influence of dose, dosage form and route of administration on tissue vitamin E levels. Dee Pharmacol Ther 1985;9:96-106

21. Murphy DJ, Mavis RD: Membrane transfer of a-tocopherol:

Influence of soluble a-tocopherol binding factors from the liver, lung, heart, and brain of the rat. J Biol Chem

1981;256:10464-10468

22. Catignani GL, Bieri JG: Rat liver a-tocopherol binding protein. Biochim Biophys Acta 1977;497:349-357

23. Traber MG, Kayden HJ: Vitamin E is delivered to cells via

the high affinity receptor for low density lipoprotein. Am J Clin Nutr 1984;40:747-751

24. McCay PB, King MM: Vitamin E: Its role as a biologic free radical scavenger and its relationship to the microsomal mixed-function oxidase system, in Machlin U (ed): Vitamin E: A Comprehensive Treatise. New York, Marcel Dekker, mc, 1980, pp 289-317

25. Burton GW, Cheeseman KH, Doba T, et al: Biology of

vitamin E: Vitamin E as an antioxidant in vitro and in vivo. Ciba Found Symp 1983;1O1:4-18

26. Diplock AT: Biology of vitamin E: The role of vitamin E in biological membranes. Ciba Found Symp 1983;1O1:45-55

27. Cynamon HA, Isenberg JN, Nguyen CH: Erythrocyte ma-londialdehyde release in vitro: A functional measure of vitamin E status. Clin Chim Acta 1985;151:169-176

28. Alade SL, Brown RE, Paquet A Jr: Polysorbate 80 and E-Ferol toxicity. Pediatrics 1986;77:593-597

29. Moore J: Final report on the safety assessment of

polysor-bates 20, 21, 40, 60, 61, 65, 80, 81, and 85. JAm Coil Toxicol 1984;3:1-82

30. McKean DL, Pesce AJ: Determination of polysorbate in ascites fluid from a premature infant. J Anal Toxicol

1985;9:174-176

31. Varma RK, Kaushal R, Junnarkar AY, et al: Polysorbate

80: A pharmacological study. Arzneimittelforsch 1985;

35:804-808

32. Burnell RH, Maxwell GM: General and coronary hemody-namic effects of Tween 20. Aust J Exp Biol Med Sci

1974;52:151-156

33. Marks LS, Kolmen SN: Tween 20 shock in dogs and related fibrinogen changes. Am J Physiol 1971;220:218-221 34. Masini E, Planchenault J, Pezziardi F, et al:

Histamine-releasing properties of polysorbate 80 in vitro and in vivo-Correlation with its hypotensive action in the dog. Agents Actions 1985;16:470

35. Nityanand S, Kapoor NK: Effect of chronic oral adminis-tration of Tween-80 in Charles Foster rats. Indian J Med Res 1979;69:664-670

36. Millard RW, Baig H, Vatner SF: Cardiovascular effects of radioactive microsphere suspensions and Tween 80 solu-tions. Am J Physiol 1977;232:H331-H334

37. Kellner A, Correll JW, Ladd AT: Sustained hyperlipemia induced in rabbits by means of intravenously injected sur-face-active agents. J Exp Med 1951;93:373-383

38. Sutherland JM: Fatal cardiovascular collapse of infants

receiving large amounts of chloramphenicol. Am J Dis Child 1959;97:761-767

39. Gershanik J, Boecler B, Ensley H, et al: The gasping syn-drome and benzyl alcohol poisoning. N Engi J Med

1982;307:1384-1388

40. Glasgow AM, Boeckx RL, Miller MK, et al: Hyperosmolarity in small infants due to propylene glycol. Pediatrics

1983;72:353-355

(5)

1986;78;503

Pediatrics

William F. Balistreri, Michael K. Farrell and Kevin E. Bove