Beneficial effects of early hypocaloric

enteral feeding on neonatal

gastrointestinal function: Preliminary
report of a randomized trial
L. Dunn, MD, S. Hulman, MD, J. W e i n e r , MD, a n d R. K l i e g m a n , MD
From the Department of Pediatrics at Case Western Reserve University, Rainbow Babies and
Chitdrens Hospital, Cleveland, Ohio

In a prospective randomized trial, we studied the effects of early hypocaloric

enteral feedings (PO) begun at 48 hours of a g e in 49 infants c o m p a r e d with 20
infants who received no enteral feedings (NPO) for at least the first 9 days of
life. Both groups initially received the majority of their calories by parenteral
alimentation. The groups were similar with respect to birth weight, gestational
age, sex, Apgar score, and major neonatal diagnoses. The early enteral feeds
proved to be significantly beneficial without an increased incidence of
complications. The PO group reached full enteral feedings faster than the NPO
group (31.2 vs 47.3 days). The PO group had a greater decline in serum bilirubin
concentration over the first 2 weeks of life and spent less time under photother-
apy (6.8 vs 9.5 days). Less cholestasis was observed a m o n g the PO infants (6.7%
vs 33%), and peak direct bilirubin levels were also lower (0.7 vs 2.5 mg/dL).
Osteopenia of prematurity, manifested by significantly lower alkaline phos-
phatase activity, was also decreased in the PO group, perhaps because of
greater calcium intake during the first month a m o n g PO infants (1.3 vs 0.8 g).
C o m p a r e d with complete bowel rest, early onset of hypocaloric enteral
feedings has beneficial effects on indirect hyperbilirubinemia, cholestatic
jaundice, and metabolic bone disease of very low birth weight infants. (J PEDIATR
1988;112:622-9)

The use of parenteral nutrition and a prolonged delay in nutrition may ameliorate these problems. Because these
starting enteral feeding has become standard care in many oral feedings do not contain sufficient calories to sustain
neonatal intensive care units. 18 However, intravenous
nutrition in the absence of oral feeding may be associated NICU Neonatal intensive care unit
with a variety of serious complications in the low birth NPO Nothing by mouth
PO Enteral feeding
weight infant, such as cholestatic jaundice, metabolic bone
GGT 3,-Glutamyl transferase
disease, and atrophy of the intestinal mucosa. 9"16 SGOT Serum glutamic-oxaloacetic transaminase
Evidence is mounting that even small volumes of enteral SGPT Serum glutamic-pyruvic transaminase
NEC Necrotizing enterocolitis

Supported by funds from Mead Johnson and a grant from the

Rainbow Board of Trustees. somatic growth, this approach has been termed "hypo-
Presented as an abstract at the 1987 Society for Pediatric caloric feeding." In adults, hypocaloric feedings reduced
Research Meetings, Anaheim, California. the incidence of parenteral alimentation-induced cholesta-
Submitted for publication July 1, 1987; accepted Nov. 14, 1987. sis.9 In animals, small quantities of oral nutrients prevented
Reprint requests: R. Kliegman, MD, Rainbow Babies and Chil- the development of both intestinal mucosal atrophy and
dren Hospital, 2101 Adelbert Rd., Cleveland, OH 44106. villous flattening.17,18

622
Volume 112 Early hypocaloric enteral feeding 623
Number 4

Our study was performed to test the hypothesis that Infants in the PO group received parenteral alimenta-
early hypocaloric enteral feeding would improve gastroin- tion according to the same schedule as the NPO group. In
testinal function and lessen the complications of parenteral addition, they received hypocaloric feedings of half-
alimentation among sick, very low birth weight infants. strength Enfamil Premature Formula (Mead Johnson &
Co., Evansville, Ind.), which were initiated at 48 hours.
METHODS The volume fed was 15 to 20 m l / k g / d . Feeding intervals
This study was conducted at Rainbow Babies and were every hour if the infant's weight was <1000 g and
Childrens Hospital, Cleveland, from September 1984 until every 2 hours if the infant's weight was > 1000 g. Therefore
January 1986. The protocol was approved by the hospital's actual volumes of individual feeding varied between 0.5
Institutional Review Board on Human Investigation. and 2.0 ml/per feeding, depending on the infant's birth
Signed parental consent was obtained for each infant in weight. All feedings were by intermittent gavage. Feedings
this investigation. All infants weighing 1500 g or less with were stopped if feeding intolerance developed, as defined
respiratory distress who required mechanical ventilation by emesis, ileus, marked abdominal distension, heme-
and placement of an umbilical arterial catheter were positive stools, or bile-stained residual. Small (<2 ml)
eligible for inclusion in the study. Criteria for exclusion nonbilious gastric residuals were refed and were not
were serious congenital anomalies and postnatal age >48 considered to be an indication for fasting. If feedings were
hours. tolerated for 48 hours, full-strength formula was intro-
Infants were enrolled in the study after being stratified duced. If the infant's clinical condition required fluid
to one of four weight categories: <750 g, 750 to 999 g, restriction, such as for treatment of a patent ductus
1000 to 1249 g, or 1250 to 1500 g. Within each category, arteriosus, only the intravenous fluid volume was reduced
the infants were randomly assigned by cards in paired accordingly. The volume of feedings in this group
envelopes to either the control group (standard nutritional remained constant for 7 days, or until the ninth day of
management, specifically, parenteral alimentation alone life.
[NPO]) or the experimental group (parenteral alimenta- At the end of the trial period (ninth day of life) the
tion plus hypocaloric enteral feeds [PO]). volume of feedings could be advanced for the infants in the
Infants in the NPO group received no enteral nutrition PO group, and enteral feedings could be introduced in the
until after 9 days of life. At age 48 hours of life, parenteral infants in the N P O group. Decisions regarding nutritional
nutrition was begun and was constituted as follows: Dex- management at this time were made by the clinicians,
trose was begun as 10% glucose in water (g/dL) if the except with respect to rapidity of feeding advancement.
infant was glucose tolerant. (Glucose intolerance was Infants weighing <1000 g would not reach full-volume
defined as serum glucose concentration over 175 m g / d L (150 ml/kg/d), full-strength enteral feedings in <10 days,
[9.7 mmol/L] and >- 2+ glucosuria by reagent test [Clin- and infants weighing <1500 g but >1000 g could not reach
itest].) If glucose intolerance was present, the dextrose full-volume, full-strength enteral feedings in less than 7
concentration was reduced until glucosuria resolved. If the days. After that time, decisions concerning response to
infant was glucose tolerant, the dextrose concentration was "feeding intolerance" or gastric residuals were not influ-
advanced to 12.5%. The protein concentration was begun enced by the investigators, but were recorded.
at approximately 1.5 g / k g / d and advanced to 2.5 g / k g / d Blood was drawn for liver function tests (total serum
if the blood urea nitrogen level was less than 20 m g / d L protein, albumin, SGOT, SGPT, alkaline phosphatase,
(7.1 mmol/L). Intravenous fat emulsion (Intralipid) was total and direct bilirubin, 3,-glutamyl transferase) at
withheld until the total bilirubin level was less than 6 enrollment and twice during the trial week. Liver function
m g / d L (102 t~mol/L). It was then begun at 0.5 g / k g / d was assessed weekly thereafter, until discharge.
and advanced to a maximum of 2.0 g / k g / d . Fluid volume Data on total fluid intake (both enteral and parenteral)
and electrolytes were given at the discretion of the house and daily weight and data pertinent to standard neonatal
staff and the attending physician on the basis of a stan- management were prospectively recorded for the duration
dard protocol and modified-as required by clinical con- of the hospital stay. Necrotizing entercolitis was defined by
dition. In our institution, fluid intake on day 1 of life is the presence of pneumatosis intestinalis, portal venous gas,
generally 80 m l / k g / d and volume is advanced by 20 or autopsy criteria in a patient with abdominal distension
ml/kg/d. The actual volume is modified on the basis of and hematochezia. Guidelines for the use of phototherapy
clinical evaluation, urine output, daily weight measure- in our institution are uniformly used and depend on levels
ments, and serum electrolyte values. By the end of the first of total bilirubin and the infant's birth weight. Photother-
week, the infant will generally receive between 150 and apy is used in an infant weighing _<1000 g if the serum
180 m l / k g / d . indirect bilirubin level is >--5 mg/dL, and in an infant
624 Dunn et al. The Journal of Pediatrics
April 1988

Table I. Demographics Table III. Effect of hypocaloric feeding on nutrition and

weight gain
NPO PO
(n = 20) (n = 19) NPO PO
( n = 45) (n = t5)
Weight (g) 931 +_ 51 989 ___ 57
GA (wk) 27.1 _+ 2.5 26.8 _+ 2.8 Days to full feeding 47.3 _+ 26.7 31.2 _+ 9.4*
Apgar score Days on hyperalimentation 34.3 _+ 24.0 23.2 _+ 10.7
1 Min 3.8 ___.6 3.3 + .6 Days to regain birth weight 24.4 ___ 8.5 19.9 +_ 6.2
5 Min 6.6 _+ .4 5.8 _+ .5 Days NPOI" 10.2 _ 11.3 6.4 _+ 8.4
F/M 9/11 13/7
B/W 8/12 10/8/1 Asian *P <0.05.
Inborn/transfer 11/9 11/8 "~Twelveof 24 hours NPO after test period and before full feedings.
Twins 5 3
GA, gestational age.
Table IV. Effect of hypocaloric feeding on indirect
hyperbilirubinemia
Table II. Major diagnoses NPO PO
(n = 15) (n = 15)
NPO PO
(n = 20) (n = ,19) Days under 9.5 + 4.0 6.8 _+ 2.4*
phototherapy
IVH
Bilirubin (mg/dl)
None 9 6
Start 5.99 + 1.32 6.57 _+ 1.78
Gr I-II 8 7
2 wk 5.27 _+ 2.67 4.35 _+ 2.01I"
Gr III-IV 3 6
Highest total bilirubin 8.77 _+ 1.67 7.94 _ 1.68
RDS
(mg/dl)
I-Ill (mild) 7 7
II-III (moderate) 5 5 *P <0.05.
Ill-IV (severe) 8 7 1P <0.01.
PDA treatment
Fluid-restriction 4 2
Indomethacin 1 3
Ligation 7 7__.~* of cholestasis were compared with the Fisher exact test.
Total 12 11 Results are recorded as the mean + SD.
BPD
02 >30 days 12 9 RESULTS
O2 >60 days 6 6 Thirty-nine infants were enrolled in the study, 20 in the
Exchange transfusion 0 2
Sepsis 4 6 N P O group and 19 in the PO group (Table I). There were
no statistically significant differences in birth weight,
IVH, intraventricular hemorrhage; RDS. respiratory distress syndrome; gestational age, Apgar scores, sex, or major complidations
PDA, patent ductus arteriosus; BPD, bronchopulmonary dysplasia.
*One after failure of indomethacin. of prematurity (Table II).
Five infants in the N P O group and four infants in the
PO group did not complete the study. All five infants in the
weighing >--1000 g if the bilirubin level is 0.5% of birth N P O group died: three from respiratory insufficiency, one
weight (e.g., >--7 m g / d L in an infant weighing 1400 g). with massive air embolism, and one with N E C and
In a small subgroup of these infants, an oral glucose perforation. In the PO group, one infant died from
tolerance test was performed at 9 days of age. Oral respiratory insufficiency. The other three infants were
glucose, 0.5 g / k g of 10% dextrose in water, was given by removed from the study because of a previously unrecog-
nasogastric tube. Blood samples for glucose determination nized aortic coarctation, overwhelming systemic candidia-
were drawn through an indwelling arterial catheter at 0, sis complicated by shock, and an ileus that precluded
15, 30, 60, and 120 minutes. feeding at 48 hours. This last patient was being treated
Statistical analysis. Clinical measurements (birth weight, with pancuronium and was also hemodynamically compro-
gestational age, calories per kilogram per day, etc.) were mised by a patent ductus arteriosus.
compared between the groups with the two-tailed unpaired The PO group attained full enteric feedings (tolerating
Student t test. Changes in liver function tests within a at least 150 m l / k g / d oral fluids) at a mean of 16.1 days
given group were evaluated by either a paired Student t (P < 0.05) earlier than the N P O group (Table III). The
test or chi-square analysis. The incidence of N E C and that number of days on parenteral atimentation and the number
Volume 112 Early hypocaloric enteral feeding 625
Number 4

t-
-r 1.8 NPO
hi PO
-r 1.6
n,- /I f
rn
IJ-
o 1.4
I--
z
hi
0 1.2
LtJ
(2.
O0
i.O
t'--
3::
C9
Ld I I I I I I
0.8
0 20 40 60

DAYS OF LIFE
Fig. I. Average weight as percent of birth weight. Initial weight loss and subsequent weight gains were similar for both
experimental (NPO) and control (PO) groups.

of days with feeding intolerance requiring periods of being all occurring after week 2 of life. This is suggestive of a
on an N P O regimen for at least 12 hours per day did not diagnosis Of cholestatic jaundice. Only 6.7% of the PO
differ significantly, nor was the time to regain birth weight infants demonstrated this problem (P <0.01). Additional-
(Table III). Furthermore, there were no significant growth ly, the highest direct bilirubin value in the NPO gi'oup at
or weight differences between the two groups throughout any time during their hospital stay was significantly
the entire study (Fig. 1). The percent of calories by the greater than thai of the PO group (2.5 _+ 3.1 vs 0.7 ___ 0.8
enteral route during the first week was 20.7% in the PO mg/dL; P < 0.05).
group and, by definition, 0% in the NPO group. Thereafter Serum SGOT values did not differ significantly (data
the percent of total calories taken by mouth increased in not reported). None of the infants in the PO group had an
both groups during the second phase of the study. None- elevated SGOT value (normal defined as up tO 67 U / L ) ,
theless, the total caloric intake of the two groups did not including the infant who had direct hyperbiiirubinemia.
differ significantly at any time during the study or during The only infants in the NPO group who had elevated levels
the remainder of the infants' hospital stay. of SGOT were the five infants who had elevated serum
The number of days in the N I C U (68 vs 57), in the levels Of direct bilirubin. Elevated serum GGT levels were
hospital (102 vs 98), with an endotracheal tube (34 vs 30), not associated with the presence of direct hyperbilirubi-
and with an umbilical arterial catheter (13 vs 10) did not hernia (data not reported). No differences in the GGT
differ significantly between the NPO and the PO groups, levels between the two study groups could be demonstrat-
respectively. ed. Among all infants With eholestasis, only one had an
The incidence of NEC did not differ between the NPO elevated level of GGT. However, six infants (three in the
and the PO groups (5% vs 16%), although the number of NPO group, and three in the PO group) had elevated GGT
patients is too small to avoid a type II error. These infants levels in the presence of normal direct bilirubin values. All
demonstrate the pattern of N E C that is common among of these infants were clinically well and receiving full
babies who weigh <1000 g in our hospital: onset late in the enteric feedings. In these well infants, the elevated serum
hospital course (day 18 to 90) and after the study values of GGT subsequently reverted to normal.
period. At no time during the hospital course were the serum
The PO group had an average of 2.7 fewer days under total protein and albumin concentrations different between
phototherapy than did the NPO group (Table IV), perhaps the two groups (data not reported). Alkaline phosphatase
because of a greater decline in the bilirubin concentration levels did, however, demonstrate a pronounced and signif-
during the first 2 weeks of life in the PO group. Of the 30 icant rise that began 6 weeks after birth in the NPO
infants who completed the study, 33.3% of the N P O infants (Fig. 2). In contrast, there were no consistent
infants had direct bilirubin measurements of >2.0 mg/dL, excessive elevations (levels >500 U / L ) of serum alkaline
626 Dunn et al. The Journal of Pediatrics
April 1988

NPO
8oor 9 p<o.o
600

t **p<~176 i * ' '

I 2 3 4 5 6 7 8 9 I0 It 12 13 14

WEEKS AFTER BIRTH

Fig. 2. Serial alkaline phosphatase levels. Infants in control group (PO) did not have progressive increase in serum
alkaline phosphatase levels observed in experimental group (NPO). Differences in alkaline phosphatase values between
the two groups reached statistical significancein weeks 6 to 11 and 13.

20 ** p <.02
Z NPO N=5
o
O3
n-" I0
(..)
X
bJ
IJJ
O3
0 \
(J
:3 -I0 \
.J \
t.q \
x _ _ - e PO N=3

-20 0 15 30 60 120 180

M I NUTES
Fig. 3. Glucose tolerance test results. The y-axis plots glucose excursion (in mg/dL) with baseline defined as value at
time 0. The x-axis illustrates minutes after bolus of glucosewas given. NPO group exhibited much greater excursion above
baseline, compared with PO group. At 30 and 60 minutes, differences between the two populations reached statistical
significanceof P < 0.05 and P < 0.02, respectively.

phosphatase activity among PO infants. (Normal alkaline for the elevated alkaline phosphatase levels, we examined
phosphatase values in our laboratory are 50 to 380 U/L; calcium, phosphorus, and vitamin D intake. Calculations
levels over 500 U / L warrant intervention.)Elevations of of intake were based on the recorded compositions of the
the enzyme occurred in 66% (10/15) of the NPO infants commercial formuia and parenteral alimentation fluid.
but in only 20% (3/15) of the PO group (P <0.01). In most The PO group received more calcium than the NPO
of the infants with elevated serum alkaline phosphatase infants during the first week (129 ___27 vs 66 + 2 4 mg;
Values, other liver function tests were normal. In three of P < 0.01) and cumulatively during,the first month of life
the patients in whom increased levels of alkaline phospha- (1.3 vs 0.8 g; P <0.05). However, after week 4, the calcium
tase were noted, the elevated enzyme fraction was from intakes were equivalent in the two groups, and at no time
bone. Two of these infants had radiographic evidence of were the phosphate intakes different. In addition, vitamin
pathologic fractures. D intake was calculated to be greater for the PO infants
To identify nutritional variables that may be responsible during the first (64 _+ 19 vs 30 ___ 12 IU; P < 0.01) and
Volume 112 Early hypocaloric enteral feeding 627
Number 4

second (114 + 30 vs 63 + 21 IU; P < 0.01) weeks of life. enteral feedings. Traditionally, these very low birth weight
Thereafter vitamin D intakes were similar. infants would not have received enteral alimentation dur'
For determination of the effects of enteral alimentation ing the acute phas e of their medical illnesses.1 The provi-
on glucose homeostasis, a pilot study of oral glucose sion of hypocaloric feeding was not associated with feeding
tolerance was performed on the ninth day of life in eight intolerance, aspiration pneumonia, ileus, or an increased
randomly selected infants. The NPO group had signifi- incidence of NEC. Moreover, hypocaloric enteral feeds
cantly greater increases of plasma glucose concentrations had beneficial effects on neonatal jaundice. The PO group
from the baseline at 30 and 60 minutes after an oral received fewer days of phototherapy than the NPO group.
glucose challenge than did the PO group (Fig. 3). This effect may be the result of enhanced intestinal
motility and excretion of bilirubin in the stool of infants
DISCUSSION receiving hypocaloric feedings.Z4 The incidence of cholesta-
The philosophies for providing nutritional support to sis was also significantly lower in the PO group. The
premature infants have evolved over the past 50 years?. 4-s incidence of choiestasis in our control group is similar to
Initially, sma!l preterm infants were fasted for various that previously reported by Beale et al. 13(23% of 62 infants
periods for fear of aspiration pneumon!a, edema forma- who weighed <2000 g, and 50% of 14 infants <1000 g).
tion, and abdominal distension. 4'6 The next phase of Similar observations that enteral feeding protect against
neonatal nutritional practices was the result of investiga- hyperalimentation-induced cholestasis have been reported
tions that compared early versus late enteral or intravenous by Pallares et al. 9 In adult patients receiving parenteral
feedings.4 These studies demonstrated that infants receiv- alimentation, the addition of hyp0ca!oric oral feedings
ing fluid and calories (either enteral or parenteral) had less prevented the typically observed elevations of direct biliru-
hypoglycemia, dehydration, hyperbilirubinernia, azotemia, bin, alkaline phoSphatase, and alanine aminotransferase,
and fever than did totally fasted infants? With the advent which were noted to occur in the patients who received
of parenteral alimentation, many NICUs began to provide parenteral alimentation alone. The mechanisms responsi-
all fluid, mineral, caloric, and vitamin needs by the ble for the reduced incidence of cholestatic jaundice are
intravenous route. 7'8 unknown. O n e proposed mechanism relates to nutrient-
Parenteral atimentation has many potential complica- induced local intestinal hormone secretion, with concomi-
tions that are related to catheter placement, the nutrients tant stimulation of hepatic bile flow. Another possible
provided, and other undetermined variables. 7'9A2-15 Rager mechanism relates to the reduction of endotoxin produc-
and Finegold 19have hypothesized that fasting, rather than tion by the intestinal flora.9 Nonetheless, the present
the solutions given, contributes to the cholestatic jaundice investigation among newborn infants supports Rager and
associated with parenteral alimentation. Finegold's original hypothesis a9 that the cholestasis associ-
Fasting in adult mammals has also been associated with ated with parenteral alimentati0n may be due to fasting
intestinal flattening of villi and mucosal atrophy. 17-~s-:~ rather than to the toxic effects of intravenous nutrients.
Animals can achieve positive nitrogen balance and grow We obtained GGT levels because of evidence25 that this
while receiving parenteral alimentation; however, intesti- enzyme is a very sensitive indicator of liver dysfunction
nal mucosal cell proliferation is reduced in animals receiv- during parenteral alimentation. In our study, elevations of
ing only parenteral nutrition. TM Enteral infusion of small GGT were not predictably related to direct hyperbilirubi-
quantities of nutrients can reverse the mucosal atrophy nemia but had a more sporadic occurrence, most typically
noted in animals during total parenteral alimentation. ~s2~ long after parenteral nutrition had ceased. Black et al. 26
lntraluminal nutrition stimulates the release of local intes- suggested that canalicular damage, as evidenced by an
tinal hormones, which may enhance enterocyte growth and elevated GGT level, occurred early in the course of
stimulate bile flow.2~ parenteral alimentation-induced cholestasis. Such a pat-
Experience among adult patients with sepsis or chronic tern was not seen in this study. GGT levels were neither
medical conditions has demonstrated a beneficial effect of sensitive nor specific for cholestasis in our population.
enteral alimentation. 22'23This method is not suitable for all In the majority of infants with a marked elevation of
patients. Nonetheless, enteral alimentation for sick adult serum alkaline phosphatase, results of liver function tests
patients has been found to be safe, efficacious, and were normal, suggesting a nonhepatic source of increased
economical compared with parenteral nutrition. 22'23 There alkaline phosphatase activity among those infants. In the
is little information about early hypocaloric enteral ali- patients in whom serum alkaline phosphatase activity was
mentation in very low birth weight infants. fractionated, it did prove to be of bone origin. In many
The present investigation of sick premature infants has patients, alkaline phosphatase activity did not revert to
also demonstrated specific beneficial effects of hypocaloric normal until vitamin D, with or without calcium or
628 Dunn et al. The Journal of Pediatrics
April 1988

phosphorus, was provided. We thus interpret the elevated REFERENCES

serum alkaline phosphatase activity to represent osteope- 1. Churella HR, Bachhuber WL, MacLean WC: Survey: meth-
nia of prematurity.26 ods of feeding low-birth-weight infants. Pediatrics 1985;
The precise cause of osteopenia of prematurity~6and the 76:243-9.
2. Moyer-Mileur L, Chan GM. Nutritional support of very-
metabolic bone disease seen during parenteral alimenta-
low:birth-weightinfants requiring prolonged assisted ventila-
tion among older subjects has not been defined. 16Potential tion. Am J Dis Chi!d 1986;140:929-32.
causes of osteopenia of prematurity include calcium or 3. Unger A, Goctzman BW, Chan C, Lyons AB, Miller MF.
phosphorus deficiency.26,27 It is possible that the higher Nutritional practices and outcome of extremely premature
calcium intake noted in the PO infants prevented the infants. Am J Dis Child 1986;140:1027-33.
4. Davies DP. The first feed of low-birthweight infants: chang-
development of the metabolic bone disease of prematuri-
ing attitudes in the twentieth century. Arch Dis Child
ty?6.27 Increased calcium intake has been reported either to 1978;53:187-92.
Prevent bone demineralization or to permit normal rates of 5. Lehmiller DJ, Kanto WP. Relationships of mesenteric throm-
bone mineralization in preterm infants.~s,29 In addition, boembolism, oral feeding, and necrotizing enterocolitis., J
reduced vitamin D intake during the first 2 weeks of life PEDIATR1978;92:96-100.
6. Gaisford W, Schofield D. Prolongation of the initial starva-
among the NPO infant may contribute to osteopenia.
tion period in premature infants. Br Med J 1950;1:1404-5.
Furthermore, infants with liver disease, as demonstrated 7. BransYW, Sumners JE, Dweck HS, Cassady G. Feeding the
by cholestasis, may have reduced absorption of vitamin D loiw-birgh-weightinfant: orally or parenterally? Preliminary
or attenuated metabolism to a more active metabo- results of a comparative study. Pediatrics 1974;54:15-22.
lite.l~, 14 8. Yu ~r James B, Hendry P, Macmahon RA. Total
parentera! nutrition in very low-birthweight infants: a con-
Compared with intravenous alimentation, oral alimenta-
trolled trial. Arch Dis Child 1979;54:653-61.
tion has been associated with the secretion of various 9. Pallares R, Sitges-Serra A, Fuentes J, Sitges-Creus A,
intestinal hormones that may augment the action of insulin Guardia J. Cholestasis associated with total parenteral nutri-
and improve the glycemic response to glucose alimenta- tion. Lancet 1983;1:758-9.
tion.20,z~ Studies in adult humans have demonstrated simi- 10. Klein GL, Ament ME, Bluestone R, et al. Bone disease
associated with total parenteral nutrition. Lancet 1980;
lar rates of energy production but lower levels of insulin
2:1041-4.
and glucose among enterally fed subjects compared with 11. Mactier H, Alroomi LG, Young DG, Raine PAM. Central
patients given parenterat a!imentation.2~.22,28 In our study, venous catheterisation in very low-birthweight infants. Arch
glucose tolerance was demonstrated in three PO infants, Dis Child 1986;61:449-53.
whereas five NPO infants demonstrated significantly 12. Cikrit D, West K, Schreiner R, Grosfeld J. Long-term
follow-up after surgical management of necrotizing enteroco-
greater excursions of plasma glucose levels after the oral
litis: sixty-three cases. J Pediatr Surg 1986;21:553-5.
tolerance test. Although these are pilot data, they suggest 13. Beale EF, Nelson RM, Bucciarelli RL, Donnelly DH, Eitz-
that oral glucose tolerance may be improved in infants who man DV. Intrahepatic cholestasis associated with parenteral
received enteral stimulation. nutrition in premature infants. Pediatrics 1979;64:342-7.
Recently, Slagle and Gross29 demonstrated beneficial 14. Patterson K, Kapur SP, Chandra RS. Hepatocellulm"carcino-
ma in a noncirrhotic infant after prolonged parenterat nutri-
effects of constant enteral feedings on neonatal gastroin- tion. J PEmATR1985;106:797-800.
testinal function. ComPared with infants who had received 15. Hodes JE, Grosfeld JL, Weber TR, Schreiner RL, Fitzgerald
only parenteral alimentati0n, those who initially received JF, Mirkin LD. Hepatic failure in infants on total parenteral
enteral feedings demonstrated an earlier time to full nutrition (TPN): clinical and histopathologic observations. J
enteral alimentation, less feeding intolerance, and lower Pediatr Surg 1982;17:463-8.
16. Brooke OG, Lucas A. Metabolic bone disease in preterm
direct bilirubin levels. There was no difference in the infants. Arch Dis Child 1985;60:682-5.
incidence of NEC among early or late enterally fed 17. Feldman E, Dowling R, McNaughton J, Peters T. Effects of
infants.29 oral versus intravenous nutrition on intestinal adaptation after
We Conclude that hypocaloric feeding has beneficial small bowel resection in the dog. Gastroenterology 1974;
effects on many of the common complications of the very 70:712-9.
18. Dworkin LD, Levin GM, Farber NJ, Spector MH. Small
immature infant. These problems include physiologic jaun- intestinal mass of the rat is partially determined by indirect
dice, cholestatic jaundice, and osteopenia of prematurity. effects of intraluminal nutrition. Gastroenterology 1976;
These benefits appear to occur without an increased 71:626-30.
morbidity related to early feedings. 19. Rager R, Finegold MJ. Cholestasis in immature newborn
infants: is parenteral alimentation responsible? J PEDIATR
We thank the neonatal nurses and the pediatric residents at 1975;86:264-69.
Rainbow Babies and Childrens Hospital. Without their coopera- 20. AI-Jurf AS, Younoszai MK, Chapman-Furr F. Effect of
tion this study could have never been performed. nutritional method on adaptation of the intestinal remnant
Volume 112 Early hypocaloric enteral feeding 629
Number 4

after massive bowel resection. J PEDIATRGastroenterol Nutr 26. Black DD, Suttle EA. Whitington PF, Whitington GL,
1985;4:245-52. Korones SD. The effect of short-term total parenteral nutri-
21. Aynsley-Green A. Hormones and postnatal adaptation to tion on hepatic function in the human neonate: a prospective
enteral nutrition. J Pediatr Gastroenterol Nutr 1983;2:418- randomized study demonstrating alteration of hepatic cana-
27. licular function. J PEDIATR 1981;99:445-9.
22. McArdle AH, Palmason C, Moreney I, Brown RA. A 27. Chan GM, Mileur L, Hansen JW. Effects of increased
rationale for enteral feeding as the preferable route for calcium and phosphorus formulas and human milk on bone
hyperalimentation. Surgery 1981;90:616-23. mineralization in preterm infants. J Pediatr Gastroenterol
23. Heymsfield SB, Bethel RA, Ansley J, Nixon DW, Rudman Nutr 1986;5:444-9.
D. Enternal hyperalimentation: an alternative to central 28. Vernet O, Christin L, Schutz Y, Danforth E, Jequier E.
venous hyperalimentation. Ann Intern Med 1979;90:63-71. Enteral versus parenteral nutrition: comparison of energy
24. De Carvalho M, Robertson S, Klaus M. Fecal bilirubin metabolism in healthy subjects. Am J Physiol 1986;250:E47-
excretion and serum bilirubin concentrations in breast-fed 54.
and bottle-fed infants. J PEDIATR 1985;107:786-90. 29. Slagle T, Gross S. Feeding tolerance and necrotizing entero-
25. Steichen J J, Gratton TL, Tsang RC. Osteopenia of prema- colitis (NEC): effect of early feeding to stimulate gastrointes-
turity: the cause and possible treatment. J PEDIATR 1980; tinal (GI) function [abstract]. Pediatr Res 1987;21:438a.
96:528-34.

FELLOWSHIPS
Available fellowships in pediatric subspecialties and those for general academic pediatric training
are listed once a year, in May, in The Journal of Pediatrics. Each October, forms for listing such
fellowships are sent to the Chairman of the Department of Pediatrics at most major hospitals in
the United States and Canada. Should you desire to list fellowships, a separate application must
be made each year for each position. All applications must be returned to The C. V. Mosby
Company by February 15 of the listing year to ensure publication. Additional forms will be
supplied on request from the Journal Editing Department, The C. V. Mosby Company, 11830
Westline Industrial Drive, St. Louis, M O 63146/314-872-8370.