Vitamin E and Necrotizing Enterocolitis

(1)

Vitamin

E and Necrotizing

Enterocolitis

Neil N. Finer,

MD, FRCP(C),

Kathrine

L. Peters,

RN, MN,

Zamir

Hayek,

MD, and Cheryl

L. Merkel,

RN, BScN

From the Department of Newborn Medicine, Royal Alexandra Hospital, and Department of

Pediatrics, University of Alberta, Edmonton, Alberta, Canada

ABSTRACT. Although vitamin E has been shown to reduce the incidence of severe sequelae from retrolental fibroplasia, there have been recent suggestions that its use may be associated with an increased incidence of necrotizing enterocolitis (NEC). A review was made of experience with vitamin E, both intramuscular and oral, and NEC over a 4#{189}-yearperiod. Of 418 infants of birth weight less than 1,500 g admitted during this period, 28/ 209 infants who had received vitamin E had definite NEC (13.4%) compared with 12/209 who had not received vitamin E (5.74%, x2 = 7.07, P = .008). For infants of birth weight less than 1,250 g, 16/103 infants who re-ceived vitamin E developed NEC v 1/159 who had not

(x’ 21.1, P < .001); the incidence of NEC was not significantly different between the two groups for infants with birth weight between 1,250 to 1,500 g. The early mortality (< seven days) for infants with birth weight of 1,500 g or less was significantly greater for those who had not received vitamin E (43.5% v 13.8%, x2 = 44.9, P <

.001), most probably a reflection of the omission of this drug for the most critically ill infants in this retrospective review. The incidence of NEC was not different for infants with birth weight of 1,500 g or less who received intramuscular vitamin E compared with control infants from the same period. For those infants for whom serum tocopherol levels were available, no infant who developed NEC and who had received only oral vitamin E had a serum tocopherol level of greater than 3.5 mg/100 mL. The increased incidence of NEC appeared to be associ-ated with an oral dose of 200 mg of vitamin E of the currently available hyperosmolar preparation. In an ef-fort to reduce serious eye sequelae from retrolental fibro-plasia, infants with birth weight less than 1,250 g who require oxygen should receive either parenteral vitamin E or a low osmolar oral preparation within 12 hours of birth. Pediatrics 1984;73:387-393; necrotizing

enterocoli-tis, vitamin E, low-birth-weight infant.

(RLF) nearly a quarter of a century ago, interest in this form of therapy was overshadowed by the

sub-sequent discovery of the central role of oxygen in

this disorder.2’3 There has recently been renewed interest in the use of vitamin E to reduce serious ocular sequelae from RLF. Vitamin E has been

reported to lower the incidence of severe RLF, both orally4’5 and intramuscularly in the low-birth-weight infant.6’7 Vitamin E has been considered a safe medication with no clinical adverse effects noted. Recently, however, Sobel et al8 reported that there may be an increased incidence of necrotizing enterocolitis (NEC) in the low-birth-weight neo-nate associated with elevated vitamin E levels (>3.5 mg/100 mL), and Johnson et al,9 in a preliminary observation, reported that infants of less than 1,500 g (with prolonged high vitamin E levels (4.48 mg/ 100 mL) had a higher incidence of NEC and sepsis after 7 days of age than infants with low serum levels. In our initial controlled evaluation of pan-enteral vitamin E, we did not observe an increased incidence of NEC in infants receiving vitamin E.6 Since completion of the trial, we have, within our own nursery, recommended the administration of oral vitamin E to all infants weighing less than

1,500 g. In view of the concern of an increased

incidence of NEC associated with vitamin E ther-apy, we retrospectively reviewed our experience with this medication relative to the incidence and severity of NEC, over a 4#{189}-yearperiod.

MATERIALS

AND

METHODS

Although Owens and Owens’ first described the use of oral vitamin E to treat netnolental fibroplasia

Received for publication July 13, 1983; accepted Sept 19, 1983. Reprint requests to (N.N.F.) Department of Newborn Medicine, Royal Alexandra Children’s Pavilion, 10240 Kingsway Aye, Ed-monton, Alberta T5H 3V9, Canada.

(2)

1981, it was recommended that all infants weighing

less than 1,500 g receive oral vitamin E, initially in an undiluted dose of 100 mg/d beginning within the first 12 hours of life. In September 1981, we began using the current dose of 200 mg/d, with a diluted preparation (dilution 1:2). Not all infants did, in fact, receive this medication, usually due to an initial oversight, or the preference of the individual attending neonatologist.

Infants weighing more than 1,500 g had a detailed

chart review only if the diagnosis of NEC or

sus-pected NEC appeared in the NICU’s admission book, NICU monthly statistics, and/or in the in-fant’s discharge diagnosis. All infants for whom the diagnosis of NEC was confirmed or suspected dun-ing the same interval had a detailed review of their medical record and the noentgenograms of infants with at least one or more of the following clinical findings were reviewed: significant abdominal dis-tension, abdominal tenderness, and/or gross blood in the stools. The roentgenograms of all infants with the above symptoms were then reviewed by a pediatric radiologist who was unaware of the in-fants’ vitamin E status or clinical diagnoses. Defi-nite NEC was diagnosed only in infants with at least one or more of the following radiologic signs: pneumatosis intestinalis, portal venous gas, pen-toneal fluid, or the presence of unchanging or per-sistent bowel loops with a thickened bowel wall on sequential abdominal radiographs.’#{176} For each in-fant with definite NEC, a complete chant review was carried out and numerical data for 103 clinical and laboratory variables were coded into a comput-enized data form (available from the authors upon request).

The infants with definite NEC were placed in the

vitamin E category if they had received supplemen-tary vitamin E in any form (intramuscularly on orally) before the onset of NEC.

In each case of definite NEC, therapy consisted of nasogastnic decompression, the administration of fresh frozen plasma, 10 mL/kg every eight to 12 hours for at least the first two to four days of therapy, early surgical consultation, a complete sep-tic work-up (blood, urine, and CSF cultures) fol-lowed by broad-spectrum parenteral antibiotics and total parenteral nutrition continued for a minimum of ten days. Oral feedings were started no sooner than ten days after diagnosis, when the infant’s condition was stable and the abdominal roentgen-ogram findings were normal. Feedings were started slowly with diluted formula or breast milk (1 to 2

mL/h) and progressed to full volume over four to

seven days.

The indications for surgery were perforation

(proven on suspect) and progressive clinical deteni-oration. No infant in the current series required

surgery for acute obstruction on gastrointestinal hemorrhage.

Statistical analysis was performed using

x2

anal-ysis, and nonpaired Student’s t test with P .05 was considered significant.

RESULTS

For the time interval studied, there were 418

infants admitted to the NICU who weighed 1,500 g. During this period there were 176 deaths, a mortality of 42%, and 40 infants developed definite NEC. There was a significant increase in incidence of NEC (number of cases of NEC in infants weigh-ing 1,500 g/number of infants weighing 1,500 g admitted) from 1979 to 1982, the four complete years included in the study (3.3%, 6.97%, 12.9%, 15.2%, x2 = 6.57, P = .037). Of209 infants weighing

less than 1,500 g who received vitamin E, 28 had proven NEC (13.4%), compared with 12/209 in-fants who had not received vitamin E (5.74%), and this difference was statistically significant

(x2

= 7.07, P = .008) (Fig. 1). For infants weighing less than 1,250 g, 16/103 infants who received vitamin E developed proven NEC for an incidence of 14.6%, compared with only 1/159 infants who did not

Fig 1. Schematic representation indicating develop-ment of necrotizing enterocolitis for infants weighing 1,500 g and infants weighing >1,500 g relative to ad-ministration of vitamin E. Infants were placed in vitamin

E group if they had received any vitamin E prior to

(3)

TABLE. Data for Infants Weighing 1,500 Grams with Necrotizing Enterocolitis (NEC): Vitamin E v No Vitamin E*

Vitamin E Oral (n=21)

1,184.1 ± 234 29.4 ± 2.1 20.6 ± 13.2

16

8

4

5.2 4.6 4.9

7.6 6.5 7.6

6.8 ± 5.5 6.8 ± 4.4 6.9 ± 6.8

158.3 ± 66.3 146.4 ± 45.8 178.9 ± 94.2 receive vitamin E for an incidence of 0.63%, and

this difference was significant

(x2

= 21.1, P < .0001). The incidence of NEC was not significantly increased for infants weighing between 1,250 and 1,500 g who received vitamin E (12/106, 11.3%)

compared with infants in this weight group who had not (11/50, 22.0%,

x2

= 2.89, P = .09).

For the same period of time, there were 43 cases of NEC in infants who weighed more than 1,500 g. Of these infants, only three had received vitamin E. The incidence of NEC pen year (cases of NEC in infants weighing >1,500 g/number of admissions of infants weighing >1,500 g) was 1.4%, 0.79%, 1.25%, and 0.02% and this reflects no significant change for the four complete years of the study (1979 through 1982).

Early NEC (7 days of age) was diagnosed in 27/ 83 infants with NEC (32.5%). Of these, seven (25.9%) weighed less than 1,500 g and 20 (74.1%)

weighed more than 1,500 g at birth. There were significantly more infants with NEC in the first week of life among infants of birth weight more than 1,500 g than for the smaller infants (20/43 v

7/40,

x2

= 7.95, P = .005). The age of onset of NEC in this early group was not different between the infants weighing 1,500 g and those weighing more than 1,500 g (4.2 days v 4.9 days). However, the age of onset of NEC for the entire group of infants was significantly different with the larger infants hay-ing a significantly earlier onset of NEC, (1,500 g

= 21.0 days, v >1,500 g = 10.9 days, t = 3.90, P < .0001).

Overall, the mortality directly attributable to NEC for all 83 infants with NEC was 13.3% (11/ 83), and 27 (33.5%) of these 83 infants required

surgery for NEC. The incidences of surgical inter-vention or death from NEC were not significantly different between infants who received vitamin E compared with those who did not, although there were no deaths secondary to NEC observed in in-fants weighing 1,500 g on less who had not received vitamin E (0/12 v 7/28,

x2

+ Yates = 2.11, P = .137).

There were ten infants who developed NEC who had never received any oral formula or breast milk, and three of these ten did not receive oral vitamin E: three infants weighed less than 1,000 g, one weighed between 1,000 and 1,250 g, five weighed between 1,250 and 1,500 g, and one infant weighed more than 2,000 g. Comparison of infants weighing less than 1,500 g with those weighing more than 1,500 g at birth revealed that the smaller infants required a greaten duration of mechanical ventila-tion (19.9 v 4.03 days, P = .015). There were no other differences noted when age of onset of feed-ing, maximum volume of feeding prior to NEC, need for surgery, and mortality were compared.

The incidence of NEC was similar for infants receiving intramuscular vitamin E (7/62, 11.3%) compared with the control infants in our original trial (6/64, 94%)6 Infants with NEC weighing 1,500 g or less who received vitamin E either intra-muscularly on orally were compared with infants in this weight group who did not receive vitamin E, and there were no significant differences noted between these groups for any of the measured van-ables (selected variables shown in the Table). When the 126 infants weighing 1,500 g or less enrolled in the original panenteral trial of vitamin E were ne-moved from the current analysis, for the remaining

Birth weight (g)t Gestational age (wk)t

Age at diagnosis of NEC (d)t No. of infants requiring

mechani-cal ventilation

Lowest blood pressure-systolic (mm Hg)t

No. of infants with patent ductus arteriosus

No. of infants with umbilical ar-terial catheter

Apgar 1 mm 5 mm

Age at initiation of enteral feed-ings (d)t

Maximum volume of feeding prior to NEC (mL/kg)t

* None of the differences were significant. t Values are means ± 1 SD.

Vitamin E IM (n

=

7)

1,226.0 ± 206 31.0± 1.0 26.2 ± 7.3

4

40.6 ± 11.4 39.6 ± 3.7

9 4

15 5

No Vitamin E (n

=

12)

1,296.7 ± 137 30.9 ± 2.7 19.0 ± 16.7

10

(4)

infants weighing 1,500 g or less, there were 21/147 (14.3%) infants who received oral vitamin E with

definite NEC compared with 6/145 (4.1%) infants who did not receive vitamin E, and this difference was significant

(x2

= 8.95, P = .003) (Fig 1).

For infants weighing less than 1,250 g in the controlled trial of parentenal vitamin E, 1/35 in-fants (2.9%) v 1/34 (2.9%) infants developed defi-nite NEC in the vitamin E and control groups. Removing these 69 infants from analysis, 15/68 infants (22.1%) receiving oral vitamin E compared with none of 125 infants who did not receive oral vitamin E developed NEC

(x2

= 29.9, P < .0001).

The serum tocophenol level prior to the develop-ment of NEC was available for 9/28 infants weigh-ing less than 1,500 g who received vitamin E; the mean level was 3.1 mg/100 mL. Only one infant had a level of greater than 3.5 mg/100 mL (8.7 mg/ 100 mL), and this infant received intramuscular followed by oral vitamin E as part of our original controlled trial.

To reduce the effect of changing neonatal mor-tality over the five years of this review, we reana-lyzed our data, removing all infants weighing less than 1,500 g who died in the first week of life (none of whom died of NEC). In the infants weighing 1,500 g, the early mortality was 29/209 or 13.8%

in the group receiving vitamin E and 91/209 or

43.5% in the group that did not receive vitamin E.

Similarly, for infants ofweighing 1,250 g, the early mortality was 26% and 50%

(x2

= 44.93, P < .0001). The incidence of NEC was no longer significantly increased for infants weighing 1,500 g on less ne-ceiving vitamin E v those who did not [28/180

(15.6%) v 12/118 (10.2%),

x2

= 1.8, P = .17)]. There was, however, a statistically significant difference for infants weighing less than 1,250 g with 16/76 (21.0%) infants receiving vitamin E developing

NEC compared with 1/79 infants not receiving vitamin E developing NEC (1.27%,

x2

= 15.53, P <

.0001). In addition, there was a significantly

in-creased incidence of NEC found in infants weighing between 1,251 and 1,500 g, who did not receive vitamin E (11/39) when compared with the mci-dence of those who did (12/104),

(x2

= 5.837, P =

.016).

After removal of the 126 infants in the original intramuscular trial together with the early neonatal

deaths, there remained a statistically significant increase in the incidence of NEC for infants weigh-ing 1,500 g or less who received oral vitamin E compared with those who did not receive oral

vi-tamin E (21/127 (16.5%) v 6/113 (5.38%),

x2

= 7.55, P = .006). The incidence of NEC for infants weigh-ing less than 1,250 g who received oral vitamin E remained significantly increased with NEC diag-nosed for 15/49 infants (30.6%) compared with no

diagnosed NEC for the 96 infants who did not receive vitamin E (x2 = 32.77, P < .0001). Once again, the incidence of NEC was greaten for infants not receiving oral vitamin E among the infants weighing between 1,250 and 1,499 g (6/17 v 6/78,

x2 + Yates = 7.297, P = .007).

DISCUSSION

Our diagnostic criteria for NEC were adapted from Bell et al” and are sufficiently strict to remove all suspect cases of NEC. We have not observed significant gastrointestinal sequelae in our nursery in infants who had only “suspect” NEC. From our retrospective review, we have determined that the incidence of NEC has increased in association with our increased use of oral vitamin E in a dose of 200 mg/d in babies weighing 1,500 g on less and that the increase is most significant for infants weighing less than 1,250 g. The figure of 1,250 g was chosen as a dividing point because in our past experiences, infants whose birth weight was less than 1,250 g were at the highest risk for serious ocular seque-lae.6’7 We had originally increased our oral dose of vitamin E from 100 mg/d to 200 mg/d in an effort

to produce serum tocopherol levels similar to those found in infants who received intramuscular vi-tamin E.6’7 In earlier studies, infants receiving oral vitamin E had significantly greaten levels than those who received no vitamin E (P = .001) and significantly lower levels than infants who received intramuscular vitamin E (P = .00i). It is of interest that between July 1 and Sept 1, 1981, only 1/13 infants weighing less than 1,500 g who received 100 mg of vitamin E per day orally developed definite

NEC, compared with 2/8 infants in the same weight group, admitted during the same interval, who did not receive any form of vitamin E. For infants weighing between 1,250 and 1,500 g who survived the first week of life, the incidence of NEC was

greater in those who did not receive oral vitamin E,

suggesting that the toxic effects of the oral vitamin E preparation may be more pronounced in the more immature infants, or that the sickest infants in this weight group did not receive vitamin E due to an initial oversight. Similarly, the mortality was sig-nificantly greater for infants weighing less than 1,500 g who did not receive vitamin E, and this may be due to the fact that in the earlier years of the study period, ordering vitamin E for extremely ill babies was not a priority for treatment.

It has been our policy not to enterally feed infants weighing less than 1,000 g until these infants reach a postnatal weight of 1,000 g and we begin panen-teral nutrition within the first five to seven days of life for such infants. As a result, it is not surprising that 4/17 infants with NEC (28.5%) with a birth

(5)

NEC with birth weight less than 1,500 g (22.5%) had never received oral formula or breast milk prior to the development of definite NEC. Other studies have noted that a few neonates have developed NEC without having received previous enteral feed-ing; Kliegman et al’2 reported that 3/123 infants and Wilson et al’3 reported that 11/86 infants weighing 1,500 g on less developed NEC who had never received entenal feeding prior to their disease. In view of the fact that three of the infants who developed NEC without prior enteral feeding had also not received any oral medication, it would appear that oral medication on feeding alone are insufficient causes for NEC.

There was no difference between the group given vitamin E and the group not given vitamin E

rela-tive to their receiving other oral medications (vi-tamin D, theophylline, and indomethacin). White and Harkavy’4 have suggested that oral medica-tions, especially those that are hyperosmolar, may be associated with the development of NEC, and they reported NEC in an infant following the

ad-dition of calcium glubionate to the formula

feed-ings. Vitamin E is one such hyperosmolar medica-tion and we have witnessed an increased incidence of NEC associated with oral vitamin E in spite of the use of a twofold dilution of this medication (measured osmolality of the diluted preparation = 2,025 mmol/kg) and the administration of this drug in small doses (25 mg/dose). Book et al’5 reported a significantly increased incidence of NEC in in-fants weighing less than 1,200 g who were fed a hypenosmolar elemental formula.’5 It has been pre-viously suggested that hypertonic solutions may produce microscopic damage to mucosal lining cells and the intestine.’6 In addition, such solutions may reduce intestinal perfusion, perhaps by acutely de-creasing blood volume secondary to the movement of fluid into the gastrointestinal tract and the sub-sequent lowering of gastrointestinal blood flow.

Similar to the observations of Wilson et al,’7 we observed that NEC occurred earlier in larger in-fants, with infants weighing more than 1,500 g having their disease onset at 10.9 days compared with 21.0 days for the smaller infants (P < .0001).

In the current study, we have observed NEC in infants who received both oral and intramuscular vitamin E. Our use of intramuscular vitamin E was for a limited period, and we did not note an in-creased incidence of NEC for the infants weighing less than 1,500 g who received this medication compared with control infants from the same period (11.3% v 94%)6 Our subsequent experience with oral vitamin E for infants weighing less than 1,500 g indicates that our increased incidence of NEC in this weight group has paralleled our increased use of oral vitamin E and is due to the higher incidence

of NEC in infants weighing less than 1,250 g. Re-moval of the infants in our controlled trial of pan-enteral vitamin E confirms that the increased risk of NEC is associated with the oral form of this medication.

Since our nursery’s use of oral vitamin E for infants weighing less than 1,500 g, we have not seen a single case of cicatnicial RLF of grade III or greater in any infant who has received oral vitamin E within the first 12 hours of birth. One must therefore balance the benefits of giving oral vitamin E early (within 12 hours of birth) to low-birth-weight infants with the potential risks of this then-apy which now must include the development of NEC.

For infants weighing less than 1,250 g over the period of the present study, only 1/53 infants who received vitamin E (oral or intramuscular) within 12 hours of birth developed cicatnicial RLF of any grade, compared with 7/32 (21.8%) infants who did not receive vitamin E

(x2

= 7.15, P = ptOO7). More importantly, cicatnicial RLF of grade III or greater, always associated with visual sequelae,’8 was seen in 5/32 infants (15.6%) who did not receive vitamin E and this degree of cicatnicial RLF was not seen in any of the 53 infants who received early vitamin E

(x2

6.203, P = .013) (Fig 2). Of the five infants with RLF of grade III or greater, four were bilat-erally blind.

RLF DATA

<- i250 GM

ALL ADMISSIONS

VITAMIN E NO VITAMIN

n.i03 n-159

VITAMIN E >i2H VITAMIN E < i2H

n-8 n.95

SONy VED DEAD SURVIVED DEAD

n=58 n-37 .35 n-i24

EYEFM

A

I NO EYE EYE NO EYE EYE

F/UP F/UP FIIJP F/liP

INSIGNANTOR’53SIGNIFICT INSIG:IToRSIGNIFICAN1

NOEYEDISEASDISEASE EYE DISEASE NO EYEDISEASE EYE DISEASE

INSIGNIFICANT OR SIGNIFICANT

NOEYE DISEASE EYE DISEASE

n.6 n-2

(6)

The incidence of NEC in the infants weighing less than 1,250 g receiving oral vitamin E was 22.1%, the need for surgery was 4.8%, and the

mortality from NEC was 5.8% of this entire popu-lation. No significant sequelae were observed in infants with NEC who did not require surgery. Thus, in our experience, for infants weighing less than 1,250 g, the risk of developing NEC is almost identical with the risk of developing cicatnicial RLF (22.1% v 21.8%).

Although visual handicap and blindness are both serious and unfortunate complications of prema-tunity, RLF does not lead to death. In view of our results, serious reconsideration must be given to the use of hyperosmolar oral vitamin E in the low-birth-weight infant. The measured hyperosmolality of the current preparation of oral vitamin E (Aqua-sol E, USV Canada, Mississauga, Ontario) is in large part related to the propylene glycol contained in the preparation. Each 100 mg of vitamin E is constituted with 400 mg of propylene glycol, 400 mg of sorbitol, 500 mg of Polysorbate-80, and 2 mL of water. Propylene glycol, although generally con-sidered to be nontoxic, may be toxic in these doses to low-birth-weight neonates19 and we are currently studying this issue in greater detail. We are now using a less osmolar oral preparation (105 mmol/ kg) containing no propylene glycol which has been produced and used by H. M. Hittner (personal communication, 1983), and this formulation may alleviate concerns regarding the hyperosmolality and possible toxicity of propylene glycol.

It is possible that vitamin E itself, apart from the hyperosmolality of the preparation, may be toxic to the gastrointestinal tract as shown by the study of Johnson et al9 using an intravenous preparation of vitamin E. In this study, very high serum levels of vitamin E were seen, and at such levels there may well be a direct toxic effect on the gastrointestinal tract. Only one infant with NEC in the current study had a serum tocopherol level of greater than 3.5 mg/100 mL, the level considered potentially

toxic,8 although levels were not available for the majority of infants. More recently, since the corn-pletion of this review we have obtained serum to-copherol levels for five infants receiving oral vi-tamin E on the day that their NEC was diagnosed, and the average level was 1.8 mg/100 mL with no infant having a level of 3.5 mg/100 mL or greater. The fact that we have seen a significant incidence of NEC in infants with levels of vitamin E below that reported by Sobel et al8 lends further support to the argument that the hyperosmolality or an-other constituent of the formulation is the culprit. The association of oral vitamin E with NEC further strengthens previous arguments that hyperosmolar

feedings may be a causative factor in NEC and should be avoided in the low-birth-weight prema-tune infant.

Consideration should be given to the use of a parenteral form of vitamin E for the critically ill low-birth-weight neonate in the first few days of life until oral administration may be safely insti-tuted. The use of a low osmolar preparation in a dose of 100 mg/kg/d should be effective in reducing the incidence of cicatnicial RLF, while not increas-ing the incidence of NEC, although further pro-spective comparisons will be necessary to confirm this speculation.

Phelps2#{176} has discussed the possible risks of vi-tamin E therapy, including NEC, and the findings in the current study, which are retrospective in nature, support her concerns and the need for care-ful ongoing prospective evaluation of newer forms of therapy for the low-birth-weight neonate.

REFERENCES

1. Owens WC, Owens EU: Retrolental fibroplasia in premature infants: II. Studies on the prophylaxis of the disease: The use of alpha tocopheryl acetate. Am J Ophthalmol 1949: 32:1631-1637

2. Ashton N, Ward B, Serpell G: Role of oxygen in the genesis of retrolental fibroplasia: Preliminary report. Br J Ophthal-mol 1953;37:513-520

3. Kinsey VE: Retrolental fibroplasia: Cooperative study of retrolental fibroplasia and the use of oxygen. Arch Ophthal-mol 1956;56:481-543

4. Hittner HM, Godio LB, Rudolph AJ? et al: Retrolental fibroplasia: Efficacy of vitamin E in a double-blind clinical study ofpreterm infants.NEnglJMed 1981;305:1365-1371 5. Hittner HM, Godio LB, Speer ME, et al: Retrolental fibro-plasia: Further clinical evidence and ultrastructural support for efficacy of vitamin E in the preterm infant. Pediatrics

1983;71:423-432

6. Finer NN, Schindler RF, Grant G, et al: Effect of intramus-cular vitamin E on frequency and severity of retrolental fibroplasia: A controlled trial. Lancet 1982;1:1087-1091 7. Finer NN, Schindler RF, Peters KL, et al: Vitamin E and

retrolental fibroplasia: Improved visual outcome with early vitamin E. Ophthalmology 1983;90:428-435

8. Sobel 5, Gueriguian J, Troendle G, et al: Vitamin E in retrolental fibroplasia lett. N EngI J Med 1982;306:867 9. Johnson L, Bowen F, Herrman N, et al: The relationship of

prolonged elevation of serum vitamin E levels to neonatal bacterial sepsis (SEP) and necrotizing enterocolitis (NEC).

Pediatr Res 1983;17:319A

10. Kogutt MS: Necrotizing enterocolitis of infancy: Early roentgen patterns as a guide to prompt diagnosis. Radiology

1979;130:367-370

11. Bell MJ, Ternberg JL, Feigin RD, et al: Neonatal necrotizing enterocolitis: Therapeutic decisions based upon clinical staging. Ann Surg 1978;187:1-7

12. Kliegman RM, Fanaroff AA: Neonatal necrotizing entero-colitis: A nine year experience. I. Epidemiology and uncom-mon observations. Am J Dis Child 1981;134:603-607

13. Wilson R, Kanto WP, McCarthy BJ, et al: Age at onset of necrotizing enterocolitis: Risk factors in small infants. Am

J Dis Child 1982;136:814-816

(7)

enter-1984;73;387

Pediatrics

Neil N. Finer, Kathrine L. Peters, Zamir Hayek and Cheryl L. Merkel