Small Volumes of Enteral Feedings Normalise Immune Function in

Infants Receiving Parenteral Nutrition

By Yasuhiro Okada, Nigel Klein, H.K.F. van Saene, and Agostino Pierro

Background/Purpose: Parenteral nutrition (PN) is associated cantly after the addition of small enteral feeds (52.0 f 4.6%.
with a risk of septicaemia. This may be caused by impairment PC .005) approaching the levels measured in controls
of immune function related to PN. The authors investigated (65.1 + 3.4%). TNF-ol production was low during total PN
the effects of the addition of enteral feedings to PN on the (1467 + 297 pg/mL) and rose significantly after the addition
immune status of human newborn infants. of minimal enteral feeds (4,661 + 1,311 pg/mL, P < .05). The
increase in CNS killing after the addition of small enteral
Methods: Ten surgical infants (age less than 6 months)
feeds in patients on PN was significantly correlated with the
requiring PN were studied in two consecutive phases: (A)
duration of enteral feeding (r = 0.8, P = .006).
after 31.1 + 6.0 days (mean + SEMI of PN with no enteral
feeding (total PN); and (B) after 4.7 + 1.1 days from the
addition of small volumes of enteral feeding to PN. Full blood Conclusions: These results indicate that the introduction of
count and liver function tests were not significantly different small volumes of enteral feed improve the impaired killing of
between phases A and B. A control group (n = 9) of infants CNS and the abnormal cytokine response observed during
receiving a normal enteral diet was also studied. Host bacteri- total PN. This implies that stimulation of the gastrointestinal
cidal activity against coagulase-negative staphylococci (CNS) tract may modulate immune function in neonates and pre-
was measured by an in vitro whole blood model. Bacterial vent bacterial infection.
killing was measured after a 45-minute bacterial challenge J Pediatr Surg 33:16-19. Copyrighto 1998 by W.B. Saunders
using the Miles-Misra technique. Tumour necrosis factor-a Company.
(TNF-CK) was measured by enzyme-linked immunosorbent
assay (ELISA) after 2 hours of bacterial challenge.
INDEX WORDS: Parenteral nutrition, enteral feeding, bacteri-
Results: The lowest level of CNS killing (37.7 + 5.2%), was cidal activity, tumour necrosis factor, coagulase-negative
observed in patients receiving total PN. This increased signifi- staphylococci.

T OTAL PARENTERAL NUTRITION has contrib-

uted to the recent improvement in the survival rate
of newborn infants. Premature and low birth weight
neonates commonly receive parenteral nutrition (PN) as
their sole source of nutrition for the first few weeks of
life. Similarly, newborn infants with congenital or ac-
quired gastrointestinal abnormalities are often dependent
exclusively on intravenous feeding for long periods.
Prolonged duration of total PN increases the risk of
infection.‘J We have recently shown that newborn infants
receiving total PN for longer than 2 weeks have impaired
phagocytosis and killing of coagulase negative staphylo-
cocci (CNS), which may explain the high risk of infection
by CNS in these patients.3Enteral feeding has several
advantageswhen comparedwith total PN. Theseinclude
maintenance of intestinal barriers,4,5decreasedrisk of
infection,6,7and improvement of nitrogen balance.4,sTo
our knowledge there are no studies on the effects of
minimal enteral feeding during PN on the immune status
of humannewborn infants.
In this study we tested the hypothesis that the addition
of smallenteral feedingsduring PN improves the bacteri-
cidal activity of newborn infants.
MATERIALS AND METHODS

Patients
From the Paediatric Surgery Unit and Irnrma-tology Unit, Institute of
Child Health and Great Ormond Street Hospital for Children, Univer- Host bactericidal activity against CNS was investigated in 10 infants
sity College London Medical School, 30 Guilford St, London, and the (age less than 6 months) requiring PN for postoperative gut dysfunction
Department of Clinical Microbiology, Royal Children’s Hospital of and in nine control infants receiving normal enteral diet. The clinical
Alder Hey, Liverpool, England. indications for PN included necrotizing enterocolitis (n = 4), gastrosclr-
Presented at the 44th Annual International Congress of the British sis (n = 3) and intestinal atresia (n = 3). Patients on PN were studied in
Association of Paediatric Surgeons. Istanbul, Turkey, July 22-25, 1997. two consecutive phases: study A during total PN with no enteral feeding
Address reprint requests to Agostino Pierro, MD, Reader in Paediat- and study B after the introduction of small volumes of enteral feeding in
ric Surgery Unit, Institute of Child Health and Great Ormond Street combination with PN. Patients on PN were studied after the 7th
Hospital for Children, University College London Medical School, 30 postoperative day to avoid the influence of surgical stress. Infants
G&ford St, London WClN IEH, England. receiving antibiotics were studied at least 36 hours after the end of
Copyright o I998 by WB. Saunders Company antibiotic treatment. Patients with clinical symptoms of sepsis were
0022-3468/98/3301-0004$03.00/O excluded from the study. Control infants receiving a normal enteral diet

16 Journal ofPediatric Surgery, Vol33, No 1 (January), 1998: pp 16-19

SMALL ENTERAL FEEDING AND IMMUNITY 17

were not affected by gastrointestinal anomalies, were not receiving
antibiotics, and were studied before minor operations under general
anaesthesia such as inguinal hernia or lensectomy. All studies were
performed in Great Ormond Street Hospital for Children, London after
obtaining informed consent from the parents.
Whole Blood Model of Bacteraemia
An iir vitro whole blood model of bacteraemia,9,‘0 was used to
measure the host-pathogen interactions in CNS infection. Patients’
blood samples were taken from a central venous line and/or from a
peripheral vein without stasis, and were anticoagulated with heparin (1
pL/mL). Five microliters of a bacterial suspension was added to 500 pL
of whole blood and incubated at 37°C on a shaker.
Bacterial Suspension
A single strain of the coagulase-negative staphylococci (CNS)
isolated from the blood of a patient was used for the bacterial challenge.
The strain was inoculated into a broth for 18 hours and resuspended in
phosphate-buffered saline solution containing lo8 organisms per milli-
liter.
Bacterial Killing
One hundred microliters of the patient’s blood sample was used to
assess the in vitro killing of CNS after 45 minutes of incubation. The
number of viable organisms was determined by using a modification of
the Miles and Misra technique. 9~10A tenfold dilution of the samples
were repeated five times with sterile water, and 100 pL (20 pL X 5) of
each dilution were taken and spotted onto a blood agar plate. After 18
hours of incubation at 36°C in 6% carbon dioxide, the number of
colonies was counted and expressed as colony forming units (CFU) per
milliliter. The percentage of viable bacteria was determined to calculate
the rate of CNS killing.
Cytokine Analysis
After 2 hours of incubation, 200 pL of the patient’s blood sample
were taken for the analysis of tumour necrosis factor-o (TNFol).
Plasma was removed and stored at 70°C for subsequent TNF-a
analysis. TNF-ol was assayed by enzyme-linked immune sorbent assay
(ELISA) as described previously.”
Data Analysis
Data were analysed using the SPSS statistical program (Microsoft,
Redmond, WA). Results are expressed as mean 2 SEM. The two study
phases in patients on PN were compared using paired t test. Patients on
PN and controls were compared using the unpaired t test. Linear
regression analysis was used to correlate independent variables. Results
demonstrating probability levels of less than .05 were considered
significant.
reported in Table 1. There were no differences in full
blood count and liver function test results between the
two study phasesin patients receiving PN. Body weight
was smaller in patients on PN than in control patients
reflecting their impaired nutritional status. Serum total
bilirubin and serum alkaline phosphataselevels were
higher in patients on PN than in control patients as a
consequenceof PN-related liver dysfunction (Table 1).
The lowest level of CNS killing was observed in
patients receiving total PN with no enteral feeding (Fig
1). CNS killing increasedsignificantly after the addition
of small enteral feedings, almost approaching the values
measuredin normal controls. The increasein CNS killing
after the addition of small enteral feedings in patients on
PN (A-CNS killing) was positively correlated with the
duration of additional enteral feeding (r = 0.8, P = .006).
However, there was no correlation between A-CNS
killing and the amount of caloriesgiven enterally.
TNF-ol production after CNS challengeis shownin Fig
2. Patients receiving total PN had lower TNP-cx produc-
tion than control patients. The addition of small enteral
feeding to PN diet normalisedTNF-0~production.
DISCUSSION
PN has contributed to the recent improvement of
morbidity in newborn infants. However, infants receiving
this form of nutritional support are at high risk of
infection. Previous reports have shown that septicaemia
developed in up to 37% of patientsduring PN.Q We have
recently shown that, in newborn infants, there is a
significant correlation between duration of total PN and
impaired bactericidal activity, assessed
by the killing and
phagocytosis of CNS.3 These microorganisms, which
were formally regarded as harmlesscommensals,have
now emerged as a pathogenic organism of clinical
importance, associatedwith an increased incidence of
septicaemia,particularly in premature infants.1J2There
are recent data that the gastrointestinal tract is an
important reservoir of CNS .l3
Clinica16.7and laboratory14-l9studieshave shown that
Table 1. Clinical Data

PN Control
RESULTS Total PN PN + Enteral (Enteral)
Study A on patients receiving PN was performed after Postnatal age (d) 53.5 2 11.4 60.9 + 12.3 82.2 i 14.4
31.1 t: 6.0 days of total PN with no enteral feeding. Body weight (g) 2.97 2 0.34* 3.07 2 0.35* 5.05 i 0.50
Study B was performed after 4.7 ?I 1.1 days from the Red blood cell (xlO1*/L) 3.64 k 0.24 3.41 t 0.14 3.83 2 0.19

addition of small enteral feeding to the PN diet. The gap Nautrophil count (X109/L) 4.95 k 1.37 4.11 2 0.97 3.39 i 0.45
Lymphocyte count (XIOg/L) 6.90 i- 1.13 4.96 i 0.49* 6.84 IO.46
between phaseA and phaseB was 7.7 t 2.1 days. The Total bilirubin (pmol/L) 62.4 2 lO.O* 60.3 2 11.3* 9.2 i 1.2
calories given enterally at the time of study B were ALT (U/L) 52.6 -t 11.4 57.3 i 18.6 26.4 i 8.2
21.3 I 6.2 kcal/kg/d or 17.7 + 5.2% of total calorie Alkaline phosphatase (U/L) 507 2 81.6* 532 + 91.2* 211 ? 18.0
intake. NOTE. Data are expressed as mean c SEM.
The clinical characteristics of the patients studied are *PC .05 versus enteral group.
18 OKADA ET AL

80 neutrophils, macrophages, and lymphocytes and may

therefore be required for effective eradication of bacterial
E 70 infections.
z In very low birth weight and premature infants, enteral
3 60
feeding often results in vomiting, interruption of feeding,
FE 50 inadequate calorie intake, and occasionally necrotizing
enterocolitis. In infants with congenital gastrointestinal
6 40 anomalies, exclusive enteral feeding is commonly pre-
cluded for long periods because of gastric stasis and
intestinal dysmotility. Therefore, neonatologists and pae-
diatric surgeons commonly establish appropriate calorie
control intake by total PN. Supplementary enteral feeding is
total PN PN+
Enteral (enteral) introduced when intestinal motility and absorption im-
proves. The percentage of calories given enterally is
Fig 1. Bactericidal activity against CNS (CNS killing). Data are gradually increased at the expense of intravenous calorie
expressed as mean f SEM.
intake. This transition time from total PN to total enteral
feeding can be prolonged. The presence of significant
enteral feeding is associated with fewer infections and gastric aspirate often discourages clinicians and surgeons
immunologic complications than parenteral nutrition. from using the gut for nutrition. However, minimal
Moore et al7 demonstrated that patients receiving total enteral feeding can be implemented early in these pa-
enteral feeding experienced significantly fewer septic tients even if its nutritional value is questionable. Our
complications than patients on PN. Kudsk et a114,15 data indicate that minimal enteral feeding may be all that
showed that enteral feeding improved survival after is required to enhance some immunologic function. This
hemoglobin Esherichia coli peritonitis in both malnour- is supported by experiments in rats.*O Shou et alzO
ished and well-nourished rats compared with rats receiv- reported that supplementation of PN with just 10%
ing total PN. The reason for these findings remains poorly enteral calories as chew diet improved rat macrophage
understood. However, enteral feeding may act by stimu- and splenocyte function.
lating more effective immune response. Alverdy et all8 In our study, bactericidal activity against CNS and
documented in a rodent model that enteral feeding TNF-or production after a CNS challenge were measured
maintains normal biliary concentrations of secretory IgA as parameters of immune function. Patients receiving
(S-IgA), which is an important component of mucosal total PN had the lowest level of both CNS killing and
immunity. In contrast, total PN decreases the biliary TNF-o production. There was a significant increase in
levels of this immunogloblin. Furthermore, Lin et all9 these functions after the addition of small volume enteral
demonstrated that the level of TNF-ar in peritoneal lavage feedings to the PN diet. Small amounts of enteral
fluid was higher in enterally fed rats than in rats receiving feedings, in addition to parenteral nutrition, can norma-
total PN after 2 hours of peritoneal bacterial challenge. lise these parameters in infants and may reduce the risk of
TNF-o(, which is mainly produced by macrophages and septic complications. The increase in CNS killing after
lymphocytes, is an important factor in the activation of introducing small volume enteral feedings (A-CNS kill-
ing) was significantly correlated with duration of enteral
feeding, but not with the amount of calories given
enterally. This implies that the duration of gut stimulation
is more important than calorific contents for restoring
5ts 6000
immunologic competence.
It is possible that the variations observed in host
bactericidal activity in our study are not related to the
4000 addition of small enteral feedings but to an improvement
in clinical condition of the patients. We feel that this is
2000 unlikely because (1) patients analysed in study B, ie, after
* ~~0.05 vs others the introduction of small enteral feeds, showed no
0 significant improvement in liver function tests; (2) the
total PN PN+ control time span between study A and B was quite short
Enteral (enteral) (7.7 + 2.1 days); (3) patients were clinically stable in
Fig 2. TNF-cu production after 2 hours of CNS challenge. Data are both phases of the study.
expressed as mean f SEM. The mechanism causing impaired immune function
SMALL ENTERAL FEEDING AND IMMUNITY 19

during total PN is unknown, and it is probably multifacto- status of patients and may help to reduce the risk of
rial. Further investigations are required to clarify how infectious complications.
small enteral feeds affect host bactericidal activity. On the
basis of these preliminary observations, minimal enteral ACKNOWLEDGMENT
feeding, establishedas soonaspossiblein infants receiv- Yasuhiro Okada was supported by the Daiwa Anglo-Japanese Foun-
ing PN, may produce beneficial effects on the immune dation.

REFERENCES
1. Ogata ES, Schulman S, Raffensperger J, et al: Caval catheteriza- molecule expression and endothelial injury by Neisseria meningitidis. J
tion in the intensive care nursery: A useful means for providing Infect Dis 173:172-179, 1996
parenteral nutrition to the extremely low birth-weight infant. J Pediatr 11. Klein NJ, Kalabalikis P, Curtis N, et al: Ex-vivo assessment of
Surg 19:258-262, 1984 candidate anti-inflammatory agents in the treatment of Gram-negative
2. Wesley JR, Coran AG: Intravenous nutrition for the Pediatric sepsis. Immun Infect Dis 4:33-35, 1994
patient. Sem Pediatr Surg 1:212-230, 1992 12. Fleer A, Gerards LJ, Aerts P, et al: Opsonic defence to
3. Okada Y, Klein N, van Saene KHF, et al: The immune response to Staphylococcus epidermidis in the premature neonate. J Infect Dis
invading bacteria is impaired in neonates receiving parenteral nutrition. 152:930-937, 1985
Proc Nutr Sot 56:218A, 1997 13. Eastick K, Leeming JP, Bennett D, et al: Reservoirs of coagulase
4. Saito H, Trocki 0, Alexander JW, et al: The effect of route of negative staphylococci in preterm infants. Arch Dis Child 74:F99-F104,
nutrient administration on the nutritional state, catabolic hormone 1996
secretion and gut mucosal integrity after bum injury. JPEN ll:l-7, 1987 14. Kudsk KA, Carpenter G, Peterson S, et al: Effect of enteral and
5. Illig KA, Ryan CK, Hardy DJ, et al: Total parenteral nutrition parenteral feeding in malnourished rats with E co&hemoglobin adju-
induced changes in gut mucosal function: atrophy alone is not the issue. vantperitonitis. J Surg Res 31:105-110, 1981
Surgery 112:631-637, 1992 15. Kudsk KA, Stone JM, Carpenter G, et al: Enteral and parenteral
6. Moore FA, Moore EE, Jones TN, et al: TEN versus TPN following feeding influences mortality after hemoglobin-E coli peritonitis in
major abdominal trauma: reduced septic morbidity. J Trauma 29:916- normal rats. J Trauma 23:605-609, 1983
923, 1989 16. Deitch EA, Winterton J, Li M, et al: The gut as a portal of entry
7. Moore FA, Feliciano DV, Androssy RJ, et al: Early enteral for bacteremia. Ann Surg 205:681-692, 1987
feeding, compare with parenteral, reduces postoperative septic compli- 17. Alverdy JC, Chi HS, Sheldon GF: The effect of parenteral
cations: the results of meta-analysis. Ann Surg 216:172-183, 1992 nutrition on gastrointestinal immunity. Ann Surg 202:681-684, 1985
8. Fletcher JP, Little JM: A comparison of parenteral nutrition and 18. Alverdy JC, Aoys E, Moss GS: Total parenteral nutrition pro-
early postoperative enteral feeding on the nitrogen balance after major motes bacterial translocation from the gut. Surgery 104: 185-190, 1988
abdominal surgery. Surgery 100:21-24, 1986 19. Lin MT, Saito H, Fukushima R, et al: Route of nutritional supply
9. Ison CA, Heyderman R, Klein NJ, et al: Whole blood model of influences local systemic and remote organ response to intraperitoneal
meningococcal bacteraemia-A method for exploring host-bacterial bacterial challenge. Ann Surg 223:84-93,1996
interactions. Microbial Pathogenesis 18:97-107, 1995 20. Shou J, Lappin J, Minnard EA, et al: Total parenteral nutrition,
10. Klein NJ, Ison CA, Peakman M, et al: The influence of bacterial translocation and host immune function. Am J Surg 167:145-
capsulation and lipooligosaccarides structure on neutrophil adhesion 150,1994