- Email: [email protected]
Plasma l -arginine concentrations in premature infants with necrotizing enterocolitis
~ ~
~
~ #
~ © ~ i~ ~ ~ ~
~
~ ~
~ ~: ~! ~! ~
~ ~
~ ~!~ ~ '~ ~ ~1 ~i~ ~
~ ~
~i ~
I~ ~ ~
~ ~ ~
~i~ ~ !~ ~! ~
~il~ ~
~ ~ ~
~i ~l~ ~
~ ~
~ii ~i~ ~ ~ ~iil i~ ~i~ ~
~ ~
~ii ~
~: ~i~ ,ii~ ~ i~ ~ ~i~¸ ~i! ~il ~i~ ~i~ ~ii i~¸ !~ <~ ~i~ ~i
L-arginine concentrations in premature vith necrotizing enterocolitis ora, ~rD, Harisb J. Amin, MBBS~Douglas 1). Mc2/lillan, 211),Paul Kubes, PhD, Gordon H. Fick, PbD, J. Decker Butzner, MD, Howard G. Parsons, ~D, and 1g Brent Scott, MDC~I
Objective: To determine whether L-arginine concentrations (the substrate for nitric oxide synthesis) are lower in premature infants in whom necrotizing enterocolitis (NEC) develops than in unaffected infants.
Methods: We measured arglnine and nutritional intake, plasma arginine, glutamine, total amino acids, and ammonia concentrations in 53 premature infants (mean gestational age + SD: 27 e 1.7 weeks) at risk of NEC. Measurements were clone on days 3, 7, 14 and 21 and just before treatment in infants with NEC.
Results: Necrotizing enterocolitis developed in 11 infants between postnatal days 1 and 26. On day 3, plasma arglnine concentrations were decreased compared with norreal published values (mean + SE, 41 btmol/L -+4). Arginine concentrations increased with day of life of measurement (p < 0.001) and arginine intake (p < 0.001). Plasma arginine concentrations were significantly lower at the time of diagnosis in infants with N E e compared with control subjects, even after adjusting for arginine intake and clay of life (p = 0.032). Plasma glutamine and total amino acid concentrations were not significantly different in infants with N E C compared with control subjects. Plasma ammonia concentrations were elevated on clay 3 (mean _+SE, 72 + 3.3 gmol/L) and decreased with postnatal age (p < 0.001) and increasing plasma arginine concentrations (p < 0.001).
L-arginine, plays an important role in maintaining baseline vasodilator tone. 4 In the gastrointestinal system NO, the principal inhibitory neurotransmitter indueing gut smooth muscle relaxation, plays an important role in the regulation of mucosal blood flow. In the face of injury or inflammation N O is a mediator critical to maintenance of mucosal integrity and intestinal barrier function, s'6 Inhibition of N O synthesis in a variety of animal models in which bowel injury has been induced by ischemia, chemical toxins, hypoxia, or platelet-activating factor is associated with increased intestinal damage. 7-11 Exogenous sources of N O attenuate these effects. 7'9 In a neonatal piglet model of NEC, continuous infusion with
Conclusion: Plasma arginine concentrations are decreased at the time of diagnosis in premature infants with NEC. The potential benefit of arginine supplementation in the prevention of the disease deserves evaluation. (J Pediatr 1997; 131:226-32)
The majority of neonates with NEC are premature infants. Mucosal injury resulting from ischemia, bacterial colonization, and formula feeding are recognized as potentially important contributors to pathogenesis. 13 A combination of these predis-
posing factors that exceeds a certain threshold level initiates mucosal injury that has been postulated to progress to intestinal necrosis. 2 Nitric oxide, a biologic mediator derived from the terminal nitrogen atom of
From the Divisions of Gastroenterologyand Neonatology, and the Departments of Pediatrics,Physiology,and Community Health Sciences, Universityof Calgary, Canary, Alberta, Canada.
Supported in part by the Swiss FoundationEugenioLitta, the Swiss Societyfor Gastroenterologyand Hepatology,and the Alberta Children'sHospital Research Foundation. Submitted for publication June 26, 1996;acceptedDec. 5, 1996. Reprint requests: R. Brent Scott, MDCM, Departmentof Pediatrics, Health ScienceCentre, 3330 Hospital Drive N.W., Calgm3r,AB T2N 4N1, Canada. Copyright © 1997by Mosby-YearBook,Inc. 0022-3476/97/$5.00 + 0 9/21179701
226
L-arginine, the substrate for N O synthase, markedly reduced intestinal injury, suggesting a potential therapeutic use of this amino acid. 12 L-Arginine is essential for ammonia detoxification by means of the urea cycle and for ereatine and polyamine synthesis. The presence of arginine-responsive hyperammonemia in premature infants suggests that arginine may be limiting for the urea cycle. 13 We hypothesize that the availability of L-arginine may be a factor limiting N O production, predisposing the immature gut to NEC. To determine whether Larginine concentrations are lower in premature infants with N E C than in unaffected infants, we prospectively measured arginine and nutritional intake, plasma arginine, and ammonia concentrations in
ZAMORA ET AL.
THE JOURNALOF PEDIATRICS
Volume 13 I, Number 2 Table L Diagnostic criteria, clinical presentation, and management of patients with NEC
premature infants and compared these parameters in infants with N E C and those who were unaffected.
METHODS This protocol was approved by the ethics committee of our institution and written informed parental consent was obtained before infant participation. During a 14-month period, consecutive premature infants with a gestational age <32 weeks (assessed by maternal menstrual history, antenatal fetal ultrasound scans when available, and Ballard scoring system) and weighing <1250 gm who were admitted to the Foothills Hospital Neonatal Intensive Care Unit were, subject to informed consent, enrolled in the study before the third day of life. Dining the first month each case was followed prospectively for the development of NEC. Exclusion criteria included congenital abnormalities, evidence of liver @function
(alanine aminotransferase level greater than three times the upper limit of normal, or direct hyperbilirnbinemia) or an inborn error of metabolism, and exchange transfusion during the study period. Necrotizing enterocolitis was diagnosed as suspected (NEC stage I) or definite (NEC stage II) according to the criteria proposed by Bell et al. 14 Stage I N E C was diagnosed if at least two of the criteria for suspected N E C (Table I) were present and feedings were withheld for at least S days. Stage II N E C was diagnosed if the subjects met criteria for suspected N E C and had pneumatosis intestinalis demonstrated on abdominal radiographs. The diagnosis of pneumatosis intestinalis was made by a radiologist independent of the study and unaware of the infant's plasma arginine levels. No attempts were made to control inrant nutritional management, and all the decisions regarding infant care were made independently by the attending neonatologist. Weight, fluid and nutri-
tional intake, type of feeding, route of nutrition, and transfused blood products were recorded daily for each infant during the first 3 weeks of life. Parenteral nutrition was generally started in the first 3 clays of life and contained TrophAmine (Kendall McGaw, Irvine, Calif.) administered as a 3% protein mixture with an arginine content of 365 mg/100 ml (2.1 retool/100 ml). No infant with an umbilical artery catheter in place was fed; Formula-fed infants received Similac Special Care 20 or 24 (Ross Laboratories, Montreal, Quebec, Canada), which contains 63 mg/100 ml (0.36 mmol/100 ml) of arginine (Special Care 24). Breast-fed infants received milk from their own mothers; in these infants arginine intake was calculated according to the reported arginine content of pooled breast milk from mothers of premature infants averaging 54 mg/100 ml (0.31 mmol/100 ml). 15 Feedings were commenced at 15 to 25 ml/kg per day, augmented by 15 to 25 ml/kg per day as tol-
227
ZAMORA ETAL.
THE JOURNAL OF PEDIATRICS AUGUST 1997
Table II. Perinatalcharacteristics
amino acid concentrations were also compared to determine whether any difference in plasma arginine concentrations reflected a generalized amino acicl pattern differing between patients with N E C and control subjects. The t test, the Mann-Whitney U test, and Fisher Exact test were used to determine whether a difference in clinical characteristics existed between patients and control subjects. All tests are two-tailed with a p value of 0.05 considered statistically significant.
RESULTS
Table III. Frequency of common neonatal complications
eratecl, up to a maximum of 150 to 180 ml/kg per clay. Plasma amino acids and blood ammonia were measured on postnatal clays 5, 7, 1"4, and 21 and when blood could be obtained at the time of diagnosis of N E C before the institution of parenteral nutrition. Blood samples were obtained in the hour before a feeding for enterally fed babies to avoid postprandial fluctuations of amino acids 16 and at least 12 hours after any blood product transfusion. Each blood sample was immediately placed on ice and the ammonia concentration was measured within 30 minutes after collection. Plasma was stored at -70 ° C until amino acid determination with an LKB d151 Alpha Plus amino acid analyzer equipped with a 51/2-hour physiologic Ultra PAC 8 col-
228
umn (Pharmacia) and dual 570 nm and
440 nm flow cell photometers. 17 We estimated that to attain 80% power of detecting a 30% difference in plasma arginine concentrations when comparing at least two control subjects per case of NEC, the required sample size would be 18 control subjects and 9 patients with N E C ((Z = 5%, two-tailed test). A randora effects model for longitudinal data is was used to compare arginine concentrations between patients and control subjects (babies in the study without NEC) adjusting for the effects of postnatal age and arginine intake on plasma arginine concentrations. In infants in whom N E C developed, only the sample drawn before and closest to the diagnosis was used in this comparison. Plasma glutamine and
Fifty-three infants met the study criteria; 51 infants were appropriate for gestational age and 2 infants (in whom N E C did not develop) were small for gestational age. Eleven infants had suspected (n = 4) or definite (n = 7) N E C between postnatal clays 1 and 26 (median day 12) (Table I). The incidence of suspected and definite N E C in the study population (11/55, 20.8%) was not significantly different from the incidence of N E C in the last 4 years inclusive of 1992 to 1995 (22.1%). All infants with NEC/suspected N E C were treated with antibiotics ancl received parenteral nutrition during the period without oral feeding (median 10 days, range 5 to 14 days). No infant required surgical management. There were no differences between the control ancl NEC/suspected N E C groups in their baseline perinatal characteristics (Table lI) or in the frequency of common neonatal complications (Table Ill). Two infants in the control group born at 27 and 25 weeks of gestational age clied on clays 7 and 25, respectively, of severe respiratory distress syndrome with intractable hypoxia. Plasma arginine, glutamine, total amino acids, and ammonia concentrations, and arginine and caloric intake on postnatal days 3, 7, 14, and 21 for affected infants and control subjects are shown in Table IV. There was no difference between the two groups for each of these parameters at the four respective times. Plasma arginine correlated with day of life and argi-
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 2
ZAMORA ET AL.
Table IV. Plasma arginine, glutamine, total amino acid, ammonia concentrations, and arginine and nutritional intake obtained on days 3,
7, 4, and 21 for infants who subsequently had NEC but had not yet developed symptoms, and control subjects*
nme intake. Specifically, with the same daily arginine intake, plasma arginine concentration increased on average by 2.9 gmol/L per day of life (p < 0.001). Furthermore, at any clay of life, plasma arginine concentrations increased by 25 ~tmol/L per 100 mg/kg per day arginine intake (p < 0.001). When only the one sample obtained before and closest to the diagnosis of N E C was compared with the data from control subjects, plasma arginine concentrations were significantly lower in infants with N E C (p = 0.032) (Fig. 1). After adjusting for arginine intake and day of life (using the random effects model for longitudinal data), plasma arginine concentrations were on average 55 gmol/L lower in patients with N E C (95% confidence interval 2.8 to 65 ~tmol/L). At the time of diagnosis arginine intake was generally lower in infants with N E C compared with the control subjects, but this difference did not reach statistical significance (p = 0.067, Fig. 2). At the time of diagnosis glutamine concentratlons in infants with N E C were not statistically different from control subjects (p = 0.5"4). Similarly total amino acids were not significantly different when the
sample obtained before and closest to the diagnosis in patients with N E C was compared with control values (p = 0.29). Plasma ammonia concentrations also correlated with day of life and plasma arginine levels. Ammonia concentrations decreased an average of 1 gmol/L by day of life (p < 0.001), and by 0.9 gmol/L for each 10 gmol/L increase in plasma arginine level (p < 0.001).
DISCUSSION In animal studies intestinal ischemia increases intestinal permeability, 19 which may favor penetration of bacteria or endotoxins through the intestinal mucosa leading to inflammation. The presence of nutrients, particularly lipids, in the lumen aggravates the increase in intestinal permeability as a result of isehemia-reperfusion and may increase permeability even in the absence of ischemia. 20'21 Nitric oxide likely plays a protective role by its effects on the regulation of intestinal mucosal blood flow, mucosal protection, and modulation of the inflammatory response.S-7,10
We investigated plasma arginine concentrations, the sole precursor for N O synthesis, in infants with NEC. Plasma arginlne concentrations were significantly lower at the time of diagnosis in babies with N E C compared with control subjects, even after adjusting for arginine intake and day of life of measurement. This conclusion is dependent on a prospective study design and analysis inclusive of infants with both suspected (stage I) and definite (stage II and III) NEC. Although stages I, II, and IlI N E C are a continuum of the same disease, stage I disease lacks definitive objective criteria for its identification. It is thus a legitimate concern that some of the infants with stage I N E C may not, in fact, have had the disease and that misclassifieation might bias our results. Yet if some of the infants with stage I N E C did not have the disease, including them as cases should actually have decreased the difference in arginine levels between the two groups and reduced the probability of finding a significant difference between the two groups. Nonetheless, if the data are reanalyzed for only the seven infants with stage II N E C compared with the 42 control subjects, the co-
229
~MORA
ET AL.
THE JOURNAL OF PEDIATRICS AUGUST 1997
A
NEC Controls
o
300 ~L 250
8 0 o
200 ¢
0 O0 0
150
o
0
0
g
0 0
0
0
0 0
0
100
o o
8
0 0 0
0
~2
0
o °° O
r,
0
0 0
elm
c~
ooo
0
o
0 0
0
oo
ooo °
o
°
50 ° °o
•t
80 0 I
!
3
7
--
"
o,
, .
.
o .
.
oA l
'"
|
~
14
21
Day of life F~7. 1. Plasmaarginine concentrations at time of diagnosis of NEC are on average 33 gmol/L lower compared with control values obtained on days 3, 7, 14, and 21 (p = 0.032). Note that data points representing control subjects and patients have been spread out slightly around day 3, 7, 14, or 21 to prevent overlap and the obscuring of data on patients who had NEC.
400
• o
NEC Controls O O
goO
8
8°
300
O O O
o
=~ 2OO
0
0
8
0
g
O
O O
O
O 0
O0
o
3
O
oo
o~ 0 0
O 00 0
o
• &
"'
O
O
o
O~o~ ~t
0
•
o
O
°l~k°o A
.<
00
0 o oo
8o
o8 oo
"4 100
0
o !
I
7
1~
21
D a y o f life Ft~. 2. Arginine intake at time of diagnosis in patients with NEC compared with control subjects (p = 0.067). Note that data points representing control subjects and patients have been spread out slightly around day 3, 7, 14, or 2t to prevent overlap and the obscuring of data on patients who had NEC.
efficients derived are similar to those reported for the entire group. After adjusting for arginine intake and day of life, plasma arginine levels were, on average,
230
27 gmol/L lower in patients with NEC, but the difference does not reach statistical significance (p = 0.16). The failure to reach significance when only patients
with stage II N E C are evaluated is, in our opinion, a reflection of the lower power of the smaller sample size. Plasma arginine homeostasis is a complex process affected by arginine intake, arginine consumption, and endogenous arginine production, which occurs principally in the kidneys (conversion of citrulline to arginine) and in the gut (production of citnflline and direct conversion of glutamine to arginine). 22'23 The decreased plasma arginine concentrations in infants with N E C may be partially explained by limited arginine intake at the time of diagnosis, but the lack of correlation between intake and plasma arginine in some of the infants suggests an increased metabolic demand for arginine or limited endogenous synthesis. For the entire group studied on day 3, plasma arginine concentrations were at the lower reference range reported for normal term breast-fed infants 24 and were similar to previously published concentrations in sick premature babies. 25'26 Arginine intake on day 5 averaged 110 mg/kg (0.6 mmol/kg) and was almost exclusively provicled by the parenteral nutrition. With predictions from a random effects model, it can be estimated that on day 5 an arginine supplement of approximately 200 mg/kg would be necessary to normalize plasma arginine concentrations. On day 5 of life the majority of infants had hyperammonemia, which was corrected progressively as plasma arginine concentrations increased. As in previous studies of premature infants, 15 the low availability of arginine appears to be a physiologically important factor limiting the effectiveness of the urea cycle. The availability of arginine might also be limiting for N O synthesis. Over a short period of time, healthy adults receiving an arginine-free diet can avoid metabolic disturbances by de novo arginine synthesis and arginine restriction does not alter the rate of whole-body N O synthesisY "29 Nevertheless, in adults with hypercholesterolemia, the local cellular L-arginine concentrations may be insufficient for optimal N O synthesis and arginine supplementation improves endothelium-dependent vasodilatation in the coronary
ZAMORA ET AL.
THE JOURNALOF PEDIATRICS Volume 13 I, Number 2 microcirculation. 30,31 Animal studies suggest that N O synthase activity undergoes an ontogenic increase in enterocytes 32 and p u l m o n a r y endothelium. 33 Nitric oxide synthase may not have reached flail maturity in premature infants leading to insufficient N O production in the presence of decreased plasma arginine concentra~ons. Glutamine is the principal fuel for the intestinal epithelium. 34 Clinical studies in adults suggest its importance in maintaining intestinal structure a n d function in stress-related situations. 35 We found no statistical difference in glutamine or total plasma amino acid concentrations at the time of diagnosis in infants with N E C compared with control subjects. This suggests that the low plasma arginine concentrations in infants with N E C are not the result of generalized amino acid deficiency or increased catabolism at the time of diagnosis. In conclusion, plasma arginine concentrations are decreased at the time of diagnosis of N E C in premature infants compared with control subjects of the same gestational age. Low arglnlne concentration may be one of the predisposing factors for NEC. This suggests a rationale for future studies investigating the potential benefit of arginine supplementation in the prevention of disease; however, initial trials should carefully monitor the safety of arginine supplementation in this population.
We thank E E Snyder and S. Hodgesfor their assistance in amino acid determination and the nursing staff from the neonatal intensive care unit, Foothills Hospital for their collaboration.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
REFERENCES 1. Kllengrnan RM. Models of the pathogenesis of necrotizing enterocolitis. J Pediatr 1990;117:$2-5. 2. Koloske AM. Epidemiology of necrotizing enterocolltis. Acta Paedlatr 1994; suppl 396:2-7. 3. Stoll BJ. Epidemiology ofnecrotizing enterocolitis. Clin Perinatol 1994;21:205-17. 4. Moncada S, Palmer RMJ, Higgs EA. Nitric oxide: physiology, pathophysinlogy and pharmacology. Pharmacol Rev 1991; 43:109-42. 5. Stark ME, Szurszeski JH. Role of nitric oxide in gastrointestinal and hepatic func-
18.
19.
tion and diseases. Gastroenterology 1992; 103:1928-49. Allcan I, Kubes P. A critical role for nitric oxide in intestinal barrier function and dysfunction. Am J Physinl 1996;33:G225-37. Payne D, Kubes P. Nitric oxide donors reduce the rise in reperfusion-induced intestinal mucosal permeability. Am J Physinl 1993;265:G189-95. Miller MJS, Zhang X-J, SadowskaKrowicka H, Chotinaruemol S, McIntyre JA, Clark DA, et al. Nitric oxide release in response to gut injury. Scand J Gastroenterol 1993;28:149-54. Caplan MS, Hedlund E, Hill N, MacKendrlck W. The role of endogenous nitric oxide and platelet-activating factor in hypoxia-lnduced intestinal injury in rats. Gastroenterology 1994;106:346-52. MacKenclrlck W, Caplan M, Hsueh W. Endogenous nitric oxide protects against platelet-activatlng factor-induced bowel injury in the rat. Pediatr Res 1993;34:222-8. Kubes P. Ischemia-reperfusion in feline small intestine: a role for nitric oxide. Am J Physiol 1993;264:G143-9. Di Lorenzo M, Bass J, Krantis A. Use of Larginine in the treatment of experimental necrotizing enterocolitis. J Pediatr Surg 1995;30:235-41. Batshaw ML, Wachtel RC, Thomas GH, Starrett A, Brusilow SW. Arginine-responsire asymptomatic hyperammonemia in the premature infant. J Pediatr 1984;105:86-91. Bell M J, Ternberg JL, Feigin RD, Keating JE Marshall E Barton L, et aL Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging. Ann Surg 1978;187:1-7. Lemons JA, Reyman D, Moye L. Amino acid composition of preterm and term breast milk during early lactation. Early Hum Dev 1983;8:323-9. Tikanoja T, Simell O, Jarvenpaa A-L, Raiha NCR. Plasma amino acids in preterm infants after a feed of human milk or formula. J Pediatr 1982;101:248-52. Carter RJ, Snycler FF. Measurement ofaspartylglucosamine in physiological fluids with an amino acid analyzer: fused peak analysis with dual photometers. Anal Biochem 1981;116:273-9. Dwyer J, Feinlah M. Introduction to statistical models for longitudinal observation. In: Dwyer J, Feinlab M, Lippert P, Hoffmeister H, editors. Statistical models for longitudinal studies in health. New York: Oxford University Press; 1992. p. 348. Langer JC, Sarvjit SS, Riddell RH. Mucosal permeability to 51Cr EDTA following subclinical intestinal ischemia-reperfusion injury in the weanling rat. J Pediatr Surg 1993;28:601-5.
20. Crlssinger KD, %0 P. The role of hpids in ischemia/reperfusion-induced changes in mucosal permeability in developing piglets. Gastroenterology 1992;102:1693-9. 21. CrissingerKD, BurneyDL, Velasquez OR, Gonzalez E. An animal model of necrotizing enterocolitis induced by infant formula and ischemia in developing piglets. Gastroenterology 1994;106:1215-22. 22. Featherston WR, Rogers QR, Freedland RA. Relative importance of kidney and liver in synthesis of arginine by the rat. Am J Physiol 1973;224:127-9. 23. Wu G, Knabe DA. Arginine synthesis in enterocytes of neonatal pigs. Am J Physiol 1995;269:R621-9. 24. Wu PYK, Edwards N, Storm MC. Plasma amino acid pattern in normal term breastfed infants. J Pediatr 1986;109:347-9. 25. Van Gouvender JB, Colen T, Wattimena JLD, Hnijmans JGM, Carnielli VP, Saner PJJ. Immediate commencement of amino acid supplementation in preterm infants: effect on serum amino acid concentrations and protein kinetics on the first day of life. J Pediatr 1995;127:458-65. 26. Rivera A, Bell EE Stegink LD, Ziegler EE. Plasma amino acid profiles during the first three days of llfe in infants with respiratory distress syndrome: effect of parenteral amino acid supplementation. J Pediatr 1989;115:465-8. 27. Carey GP, Kime Z, Rogers QR, Morris JG, Hargrove D, But~gton CA, et al. An arginine-deficient diet in human does not evoke hyperammonemia or orotic aciduria. J Nutr 1987;117:1734-9. 28. CastilloL, Ajami A, Branch S, Chapman TE, Yu Y-M, Burke JE et al. Plasma arginine kinetics in adult man: response to an argininefree diet. Metabolism 1994;43:114-22. 29. Castillo L, Sanchez M, Vogt J, Chapman TE, DeRojas-Walker TC, Tannenhaum SR, et al. Plasma arginine, citrulline, and ornithine kinetics in adults, with observations on nitric oxide synthesis. Am J Physiol 1995;268:E360-7. 30. Drexler H, Zeiher AM, Meinzer K, Just H. Correction of endothelial dysfunction in coronary micro circulation of hypercholesterolemic patients by L-arginine. Lancet
1991;358:1546-50. 31. Creager M_A_,Gallagher S J, Girerd X J, Coleman SM, Dzan VJ, Cooke dE L-arginine improves endothelium-dependent vasodilatation in hypercholesterolemic humans. J Clin Invest 1992;90:1248-55. 32. M'Rahet-Touil H, Blachier E Morel M-T, Darcy-Vrillon B, Duee P-H. Characterization and ontogenesis of nitric oxide synthase activity in pig enterocytes. FEBS Lett 1993;331:243-7. 33. Shaul PW, Farrar MA, Magness RR. Pulmonary endothelial nitric oxide produc-
231
ZAMORA ET AL.
tiou is developmentally regulated in the fetus and newborn. Am J Physio11993;265: H1056-63. 34. Windmueller HG, Spaeth AE. Respiratory fuels and nitrogen metabolism in
THE JOURNAL OF PEDIATRICS AUGUST 1997 vivo in small intestine of fed rats: quantitative importance of glutamine, glutamate and aspartate. J Biol Chem 1980;255: 107-12. 35. Tremel H, Kienle B, Weilemann LS,
Stehle P, Furst P. Glutamine dipeptidesupplemented parenteral nutrition maintains intestinal function in the critically ill. Gastroenterology 1994;107: 1595-601.
CORRECTION The following table is the corrected version of the table appearing on page 815 of the article "Recurrent
Streptococcuspneumoniae Sepsis in Children With Sickle Cell Disease," by Hongeng et al., published in the May 1997 issue of The Journal (volume 130, pages 814 to 816):
Table. Characteristics of patients, penicillin prophylaxis use, and outcome of recurrent S, pneumoniae sepsis
232
~
~ #
~ © ~ i~ ~ ~ ~
~
~ ~
~ ~: ~! ~! ~
~ ~
~ ~!~ ~ '~ ~ ~1 ~i~ ~
~ ~
~i ~
I~ ~ ~
~ ~ ~
~i~ ~ !~ ~! ~
~il~ ~
~ ~ ~
~i ~l~ ~
~ ~
~ii ~i~ ~ ~ ~iil i~ ~i~ ~
~ ~
~ii ~
~: ~i~ ,ii~ ~ i~ ~ ~i~¸ ~i! ~il ~i~ ~i~ ~ii i~¸ !~ <~ ~i~ ~i
L-arginine concentrations in premature vith necrotizing enterocolitis ora, ~rD, Harisb J. Amin, MBBS~Douglas 1). Mc2/lillan, 211),Paul Kubes, PhD, Gordon H. Fick, PbD, J. Decker Butzner, MD, Howard G. Parsons, ~D, and 1g Brent Scott, MDC~I
Objective: To determine whether L-arginine concentrations (the substrate for nitric oxide synthesis) are lower in premature infants in whom necrotizing enterocolitis (NEC) develops than in unaffected infants.
Methods: We measured arglnine and nutritional intake, plasma arginine, glutamine, total amino acids, and ammonia concentrations in 53 premature infants (mean gestational age + SD: 27 e 1.7 weeks) at risk of NEC. Measurements were clone on days 3, 7, 14 and 21 and just before treatment in infants with NEC.
Results: Necrotizing enterocolitis developed in 11 infants between postnatal days 1 and 26. On day 3, plasma arglnine concentrations were decreased compared with norreal published values (mean + SE, 41 btmol/L -+4). Arginine concentrations increased with day of life of measurement (p < 0.001) and arginine intake (p < 0.001). Plasma arginine concentrations were significantly lower at the time of diagnosis in infants with N E e compared with control subjects, even after adjusting for arginine intake and clay of life (p = 0.032). Plasma glutamine and total amino acid concentrations were not significantly different in infants with N E C compared with control subjects. Plasma ammonia concentrations were elevated on clay 3 (mean _+SE, 72 + 3.3 gmol/L) and decreased with postnatal age (p < 0.001) and increasing plasma arginine concentrations (p < 0.001).
L-arginine, plays an important role in maintaining baseline vasodilator tone. 4 In the gastrointestinal system NO, the principal inhibitory neurotransmitter indueing gut smooth muscle relaxation, plays an important role in the regulation of mucosal blood flow. In the face of injury or inflammation N O is a mediator critical to maintenance of mucosal integrity and intestinal barrier function, s'6 Inhibition of N O synthesis in a variety of animal models in which bowel injury has been induced by ischemia, chemical toxins, hypoxia, or platelet-activating factor is associated with increased intestinal damage. 7-11 Exogenous sources of N O attenuate these effects. 7'9 In a neonatal piglet model of NEC, continuous infusion with
Conclusion: Plasma arginine concentrations are decreased at the time of diagnosis in premature infants with NEC. The potential benefit of arginine supplementation in the prevention of the disease deserves evaluation. (J Pediatr 1997; 131:226-32)
The majority of neonates with NEC are premature infants. Mucosal injury resulting from ischemia, bacterial colonization, and formula feeding are recognized as potentially important contributors to pathogenesis. 13 A combination of these predis-
posing factors that exceeds a certain threshold level initiates mucosal injury that has been postulated to progress to intestinal necrosis. 2 Nitric oxide, a biologic mediator derived from the terminal nitrogen atom of
From the Divisions of Gastroenterologyand Neonatology, and the Departments of Pediatrics,Physiology,and Community Health Sciences, Universityof Calgary, Canary, Alberta, Canada.
Supported in part by the Swiss FoundationEugenioLitta, the Swiss Societyfor Gastroenterologyand Hepatology,and the Alberta Children'sHospital Research Foundation. Submitted for publication June 26, 1996;acceptedDec. 5, 1996. Reprint requests: R. Brent Scott, MDCM, Departmentof Pediatrics, Health ScienceCentre, 3330 Hospital Drive N.W., Calgm3r,AB T2N 4N1, Canada. Copyright © 1997by Mosby-YearBook,Inc. 0022-3476/97/$5.00 + 0 9/21179701
226
L-arginine, the substrate for N O synthase, markedly reduced intestinal injury, suggesting a potential therapeutic use of this amino acid. 12 L-Arginine is essential for ammonia detoxification by means of the urea cycle and for ereatine and polyamine synthesis. The presence of arginine-responsive hyperammonemia in premature infants suggests that arginine may be limiting for the urea cycle. 13 We hypothesize that the availability of L-arginine may be a factor limiting N O production, predisposing the immature gut to NEC. To determine whether Larginine concentrations are lower in premature infants with N E C than in unaffected infants, we prospectively measured arginine and nutritional intake, plasma arginine, and ammonia concentrations in
ZAMORA ET AL.
THE JOURNALOF PEDIATRICS
Volume 13 I, Number 2 Table L Diagnostic criteria, clinical presentation, and management of patients with NEC
premature infants and compared these parameters in infants with N E C and those who were unaffected.
METHODS This protocol was approved by the ethics committee of our institution and written informed parental consent was obtained before infant participation. During a 14-month period, consecutive premature infants with a gestational age <32 weeks (assessed by maternal menstrual history, antenatal fetal ultrasound scans when available, and Ballard scoring system) and weighing <1250 gm who were admitted to the Foothills Hospital Neonatal Intensive Care Unit were, subject to informed consent, enrolled in the study before the third day of life. Dining the first month each case was followed prospectively for the development of NEC. Exclusion criteria included congenital abnormalities, evidence of liver @function
(alanine aminotransferase level greater than three times the upper limit of normal, or direct hyperbilirnbinemia) or an inborn error of metabolism, and exchange transfusion during the study period. Necrotizing enterocolitis was diagnosed as suspected (NEC stage I) or definite (NEC stage II) according to the criteria proposed by Bell et al. 14 Stage I N E C was diagnosed if at least two of the criteria for suspected N E C (Table I) were present and feedings were withheld for at least S days. Stage II N E C was diagnosed if the subjects met criteria for suspected N E C and had pneumatosis intestinalis demonstrated on abdominal radiographs. The diagnosis of pneumatosis intestinalis was made by a radiologist independent of the study and unaware of the infant's plasma arginine levels. No attempts were made to control inrant nutritional management, and all the decisions regarding infant care were made independently by the attending neonatologist. Weight, fluid and nutri-
tional intake, type of feeding, route of nutrition, and transfused blood products were recorded daily for each infant during the first 3 weeks of life. Parenteral nutrition was generally started in the first 3 clays of life and contained TrophAmine (Kendall McGaw, Irvine, Calif.) administered as a 3% protein mixture with an arginine content of 365 mg/100 ml (2.1 retool/100 ml). No infant with an umbilical artery catheter in place was fed; Formula-fed infants received Similac Special Care 20 or 24 (Ross Laboratories, Montreal, Quebec, Canada), which contains 63 mg/100 ml (0.36 mmol/100 ml) of arginine (Special Care 24). Breast-fed infants received milk from their own mothers; in these infants arginine intake was calculated according to the reported arginine content of pooled breast milk from mothers of premature infants averaging 54 mg/100 ml (0.31 mmol/100 ml). 15 Feedings were commenced at 15 to 25 ml/kg per day, augmented by 15 to 25 ml/kg per day as tol-
227
ZAMORA ETAL.
THE JOURNAL OF PEDIATRICS AUGUST 1997
Table II. Perinatalcharacteristics
amino acid concentrations were also compared to determine whether any difference in plasma arginine concentrations reflected a generalized amino acicl pattern differing between patients with N E C and control subjects. The t test, the Mann-Whitney U test, and Fisher Exact test were used to determine whether a difference in clinical characteristics existed between patients and control subjects. All tests are two-tailed with a p value of 0.05 considered statistically significant.
RESULTS
Table III. Frequency of common neonatal complications
eratecl, up to a maximum of 150 to 180 ml/kg per clay. Plasma amino acids and blood ammonia were measured on postnatal clays 5, 7, 1"4, and 21 and when blood could be obtained at the time of diagnosis of N E C before the institution of parenteral nutrition. Blood samples were obtained in the hour before a feeding for enterally fed babies to avoid postprandial fluctuations of amino acids 16 and at least 12 hours after any blood product transfusion. Each blood sample was immediately placed on ice and the ammonia concentration was measured within 30 minutes after collection. Plasma was stored at -70 ° C until amino acid determination with an LKB d151 Alpha Plus amino acid analyzer equipped with a 51/2-hour physiologic Ultra PAC 8 col-
228
umn (Pharmacia) and dual 570 nm and
440 nm flow cell photometers. 17 We estimated that to attain 80% power of detecting a 30% difference in plasma arginine concentrations when comparing at least two control subjects per case of NEC, the required sample size would be 18 control subjects and 9 patients with N E C ((Z = 5%, two-tailed test). A randora effects model for longitudinal data is was used to compare arginine concentrations between patients and control subjects (babies in the study without NEC) adjusting for the effects of postnatal age and arginine intake on plasma arginine concentrations. In infants in whom N E C developed, only the sample drawn before and closest to the diagnosis was used in this comparison. Plasma glutamine and
Fifty-three infants met the study criteria; 51 infants were appropriate for gestational age and 2 infants (in whom N E C did not develop) were small for gestational age. Eleven infants had suspected (n = 4) or definite (n = 7) N E C between postnatal clays 1 and 26 (median day 12) (Table I). The incidence of suspected and definite N E C in the study population (11/55, 20.8%) was not significantly different from the incidence of N E C in the last 4 years inclusive of 1992 to 1995 (22.1%). All infants with NEC/suspected N E C were treated with antibiotics ancl received parenteral nutrition during the period without oral feeding (median 10 days, range 5 to 14 days). No infant required surgical management. There were no differences between the control ancl NEC/suspected N E C groups in their baseline perinatal characteristics (Table lI) or in the frequency of common neonatal complications (Table Ill). Two infants in the control group born at 27 and 25 weeks of gestational age clied on clays 7 and 25, respectively, of severe respiratory distress syndrome with intractable hypoxia. Plasma arginine, glutamine, total amino acids, and ammonia concentrations, and arginine and caloric intake on postnatal days 3, 7, 14, and 21 for affected infants and control subjects are shown in Table IV. There was no difference between the two groups for each of these parameters at the four respective times. Plasma arginine correlated with day of life and argi-
THE JOURNAL OF PEDIATRICS Volume 13 I, Number 2
ZAMORA ET AL.
Table IV. Plasma arginine, glutamine, total amino acid, ammonia concentrations, and arginine and nutritional intake obtained on days 3,
7, 4, and 21 for infants who subsequently had NEC but had not yet developed symptoms, and control subjects*
nme intake. Specifically, with the same daily arginine intake, plasma arginine concentration increased on average by 2.9 gmol/L per day of life (p < 0.001). Furthermore, at any clay of life, plasma arginine concentrations increased by 25 ~tmol/L per 100 mg/kg per day arginine intake (p < 0.001). When only the one sample obtained before and closest to the diagnosis of N E C was compared with the data from control subjects, plasma arginine concentrations were significantly lower in infants with N E C (p = 0.032) (Fig. 1). After adjusting for arginine intake and day of life (using the random effects model for longitudinal data), plasma arginine concentrations were on average 55 gmol/L lower in patients with N E C (95% confidence interval 2.8 to 65 ~tmol/L). At the time of diagnosis arginine intake was generally lower in infants with N E C compared with the control subjects, but this difference did not reach statistical significance (p = 0.067, Fig. 2). At the time of diagnosis glutamine concentratlons in infants with N E C were not statistically different from control subjects (p = 0.5"4). Similarly total amino acids were not significantly different when the
sample obtained before and closest to the diagnosis in patients with N E C was compared with control values (p = 0.29). Plasma ammonia concentrations also correlated with day of life and plasma arginine levels. Ammonia concentrations decreased an average of 1 gmol/L by day of life (p < 0.001), and by 0.9 gmol/L for each 10 gmol/L increase in plasma arginine level (p < 0.001).
DISCUSSION In animal studies intestinal ischemia increases intestinal permeability, 19 which may favor penetration of bacteria or endotoxins through the intestinal mucosa leading to inflammation. The presence of nutrients, particularly lipids, in the lumen aggravates the increase in intestinal permeability as a result of isehemia-reperfusion and may increase permeability even in the absence of ischemia. 20'21 Nitric oxide likely plays a protective role by its effects on the regulation of intestinal mucosal blood flow, mucosal protection, and modulation of the inflammatory response.S-7,10
We investigated plasma arginine concentrations, the sole precursor for N O synthesis, in infants with NEC. Plasma arginlne concentrations were significantly lower at the time of diagnosis in babies with N E C compared with control subjects, even after adjusting for arginine intake and day of life of measurement. This conclusion is dependent on a prospective study design and analysis inclusive of infants with both suspected (stage I) and definite (stage II and III) NEC. Although stages I, II, and IlI N E C are a continuum of the same disease, stage I disease lacks definitive objective criteria for its identification. It is thus a legitimate concern that some of the infants with stage I N E C may not, in fact, have had the disease and that misclassifieation might bias our results. Yet if some of the infants with stage I N E C did not have the disease, including them as cases should actually have decreased the difference in arginine levels between the two groups and reduced the probability of finding a significant difference between the two groups. Nonetheless, if the data are reanalyzed for only the seven infants with stage II N E C compared with the 42 control subjects, the co-
229
~MORA
ET AL.
THE JOURNAL OF PEDIATRICS AUGUST 1997
A
NEC Controls
o
300 ~L 250
8 0 o
200 ¢
0 O0 0
150
o
0
0
g
0 0
0
0
0 0
0
100
o o
8
0 0 0
0
~2
0
o °° O
r,
0
0 0
elm
c~
ooo
0
o
0 0
0
oo
ooo °
o
°
50 ° °o
•t
80 0 I
!
3
7
--
"
o,
, .
.
o .
.
oA l
'"
|
~
14
21
Day of life F~7. 1. Plasmaarginine concentrations at time of diagnosis of NEC are on average 33 gmol/L lower compared with control values obtained on days 3, 7, 14, and 21 (p = 0.032). Note that data points representing control subjects and patients have been spread out slightly around day 3, 7, 14, or 21 to prevent overlap and the obscuring of data on patients who had NEC.
400
• o
NEC Controls O O
goO
8
8°
300
O O O
o
=~ 2OO
0
0
8
0
g
O
O O
O
O 0
O0
o
3
O
oo
o~ 0 0
O 00 0
o
• &
"'
O
O
o
O~o~ ~t
0
•
o
O
°l~k°o A
.<
00
0 o oo
8o
o8 oo
"4 100
0
o !
I
7
1~
21
D a y o f life Ft~. 2. Arginine intake at time of diagnosis in patients with NEC compared with control subjects (p = 0.067). Note that data points representing control subjects and patients have been spread out slightly around day 3, 7, 14, or 2t to prevent overlap and the obscuring of data on patients who had NEC.
efficients derived are similar to those reported for the entire group. After adjusting for arginine intake and day of life, plasma arginine levels were, on average,
230
27 gmol/L lower in patients with NEC, but the difference does not reach statistical significance (p = 0.16). The failure to reach significance when only patients
with stage II N E C are evaluated is, in our opinion, a reflection of the lower power of the smaller sample size. Plasma arginine homeostasis is a complex process affected by arginine intake, arginine consumption, and endogenous arginine production, which occurs principally in the kidneys (conversion of citrulline to arginine) and in the gut (production of citnflline and direct conversion of glutamine to arginine). 22'23 The decreased plasma arginine concentrations in infants with N E C may be partially explained by limited arginine intake at the time of diagnosis, but the lack of correlation between intake and plasma arginine in some of the infants suggests an increased metabolic demand for arginine or limited endogenous synthesis. For the entire group studied on day 3, plasma arginine concentrations were at the lower reference range reported for normal term breast-fed infants 24 and were similar to previously published concentrations in sick premature babies. 25'26 Arginine intake on day 5 averaged 110 mg/kg (0.6 mmol/kg) and was almost exclusively provicled by the parenteral nutrition. With predictions from a random effects model, it can be estimated that on day 5 an arginine supplement of approximately 200 mg/kg would be necessary to normalize plasma arginine concentrations. On day 5 of life the majority of infants had hyperammonemia, which was corrected progressively as plasma arginine concentrations increased. As in previous studies of premature infants, 15 the low availability of arginine appears to be a physiologically important factor limiting the effectiveness of the urea cycle. The availability of arginine might also be limiting for N O synthesis. Over a short period of time, healthy adults receiving an arginine-free diet can avoid metabolic disturbances by de novo arginine synthesis and arginine restriction does not alter the rate of whole-body N O synthesisY "29 Nevertheless, in adults with hypercholesterolemia, the local cellular L-arginine concentrations may be insufficient for optimal N O synthesis and arginine supplementation improves endothelium-dependent vasodilatation in the coronary
ZAMORA ET AL.
THE JOURNALOF PEDIATRICS Volume 13 I, Number 2 microcirculation. 30,31 Animal studies suggest that N O synthase activity undergoes an ontogenic increase in enterocytes 32 and p u l m o n a r y endothelium. 33 Nitric oxide synthase may not have reached flail maturity in premature infants leading to insufficient N O production in the presence of decreased plasma arginine concentra~ons. Glutamine is the principal fuel for the intestinal epithelium. 34 Clinical studies in adults suggest its importance in maintaining intestinal structure a n d function in stress-related situations. 35 We found no statistical difference in glutamine or total plasma amino acid concentrations at the time of diagnosis in infants with N E C compared with control subjects. This suggests that the low plasma arginine concentrations in infants with N E C are not the result of generalized amino acid deficiency or increased catabolism at the time of diagnosis. In conclusion, plasma arginine concentrations are decreased at the time of diagnosis of N E C in premature infants compared with control subjects of the same gestational age. Low arglnlne concentration may be one of the predisposing factors for NEC. This suggests a rationale for future studies investigating the potential benefit of arginine supplementation in the prevention of disease; however, initial trials should carefully monitor the safety of arginine supplementation in this population.
We thank E E Snyder and S. Hodgesfor their assistance in amino acid determination and the nursing staff from the neonatal intensive care unit, Foothills Hospital for their collaboration.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
REFERENCES 1. Kllengrnan RM. Models of the pathogenesis of necrotizing enterocolitis. J Pediatr 1990;117:$2-5. 2. Koloske AM. Epidemiology of necrotizing enterocolltis. Acta Paedlatr 1994; suppl 396:2-7. 3. Stoll BJ. Epidemiology ofnecrotizing enterocolitis. Clin Perinatol 1994;21:205-17. 4. Moncada S, Palmer RMJ, Higgs EA. Nitric oxide: physiology, pathophysinlogy and pharmacology. Pharmacol Rev 1991; 43:109-42. 5. Stark ME, Szurszeski JH. Role of nitric oxide in gastrointestinal and hepatic func-
18.
19.
tion and diseases. Gastroenterology 1992; 103:1928-49. Allcan I, Kubes P. A critical role for nitric oxide in intestinal barrier function and dysfunction. Am J Physinl 1996;33:G225-37. Payne D, Kubes P. Nitric oxide donors reduce the rise in reperfusion-induced intestinal mucosal permeability. Am J Physinl 1993;265:G189-95. Miller MJS, Zhang X-J, SadowskaKrowicka H, Chotinaruemol S, McIntyre JA, Clark DA, et al. Nitric oxide release in response to gut injury. Scand J Gastroenterol 1993;28:149-54. Caplan MS, Hedlund E, Hill N, MacKendrlck W. The role of endogenous nitric oxide and platelet-activating factor in hypoxia-lnduced intestinal injury in rats. Gastroenterology 1994;106:346-52. MacKenclrlck W, Caplan M, Hsueh W. Endogenous nitric oxide protects against platelet-activatlng factor-induced bowel injury in the rat. Pediatr Res 1993;34:222-8. Kubes P. Ischemia-reperfusion in feline small intestine: a role for nitric oxide. Am J Physiol 1993;264:G143-9. Di Lorenzo M, Bass J, Krantis A. Use of Larginine in the treatment of experimental necrotizing enterocolitis. J Pediatr Surg 1995;30:235-41. Batshaw ML, Wachtel RC, Thomas GH, Starrett A, Brusilow SW. Arginine-responsire asymptomatic hyperammonemia in the premature infant. J Pediatr 1984;105:86-91. Bell M J, Ternberg JL, Feigin RD, Keating JE Marshall E Barton L, et aL Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging. Ann Surg 1978;187:1-7. Lemons JA, Reyman D, Moye L. Amino acid composition of preterm and term breast milk during early lactation. Early Hum Dev 1983;8:323-9. Tikanoja T, Simell O, Jarvenpaa A-L, Raiha NCR. Plasma amino acids in preterm infants after a feed of human milk or formula. J Pediatr 1982;101:248-52. Carter RJ, Snycler FF. Measurement ofaspartylglucosamine in physiological fluids with an amino acid analyzer: fused peak analysis with dual photometers. Anal Biochem 1981;116:273-9. Dwyer J, Feinlah M. Introduction to statistical models for longitudinal observation. In: Dwyer J, Feinlab M, Lippert P, Hoffmeister H, editors. Statistical models for longitudinal studies in health. New York: Oxford University Press; 1992. p. 348. Langer JC, Sarvjit SS, Riddell RH. Mucosal permeability to 51Cr EDTA following subclinical intestinal ischemia-reperfusion injury in the weanling rat. J Pediatr Surg 1993;28:601-5.
20. Crlssinger KD, %0 P. The role of hpids in ischemia/reperfusion-induced changes in mucosal permeability in developing piglets. Gastroenterology 1992;102:1693-9. 21. CrissingerKD, BurneyDL, Velasquez OR, Gonzalez E. An animal model of necrotizing enterocolitis induced by infant formula and ischemia in developing piglets. Gastroenterology 1994;106:1215-22. 22. Featherston WR, Rogers QR, Freedland RA. Relative importance of kidney and liver in synthesis of arginine by the rat. Am J Physiol 1973;224:127-9. 23. Wu G, Knabe DA. Arginine synthesis in enterocytes of neonatal pigs. Am J Physiol 1995;269:R621-9. 24. Wu PYK, Edwards N, Storm MC. Plasma amino acid pattern in normal term breastfed infants. J Pediatr 1986;109:347-9. 25. Van Gouvender JB, Colen T, Wattimena JLD, Hnijmans JGM, Carnielli VP, Saner PJJ. Immediate commencement of amino acid supplementation in preterm infants: effect on serum amino acid concentrations and protein kinetics on the first day of life. J Pediatr 1995;127:458-65. 26. Rivera A, Bell EE Stegink LD, Ziegler EE. Plasma amino acid profiles during the first three days of llfe in infants with respiratory distress syndrome: effect of parenteral amino acid supplementation. J Pediatr 1989;115:465-8. 27. Carey GP, Kime Z, Rogers QR, Morris JG, Hargrove D, But~gton CA, et al. An arginine-deficient diet in human does not evoke hyperammonemia or orotic aciduria. J Nutr 1987;117:1734-9. 28. CastilloL, Ajami A, Branch S, Chapman TE, Yu Y-M, Burke JE et al. Plasma arginine kinetics in adult man: response to an argininefree diet. Metabolism 1994;43:114-22. 29. Castillo L, Sanchez M, Vogt J, Chapman TE, DeRojas-Walker TC, Tannenhaum SR, et al. Plasma arginine, citrulline, and ornithine kinetics in adults, with observations on nitric oxide synthesis. Am J Physiol 1995;268:E360-7. 30. Drexler H, Zeiher AM, Meinzer K, Just H. Correction of endothelial dysfunction in coronary micro circulation of hypercholesterolemic patients by L-arginine. Lancet
1991;358:1546-50. 31. Creager M_A_,Gallagher S J, Girerd X J, Coleman SM, Dzan VJ, Cooke dE L-arginine improves endothelium-dependent vasodilatation in hypercholesterolemic humans. J Clin Invest 1992;90:1248-55. 32. M'Rahet-Touil H, Blachier E Morel M-T, Darcy-Vrillon B, Duee P-H. Characterization and ontogenesis of nitric oxide synthase activity in pig enterocytes. FEBS Lett 1993;331:243-7. 33. Shaul PW, Farrar MA, Magness RR. Pulmonary endothelial nitric oxide produc-
231
ZAMORA ET AL.
tiou is developmentally regulated in the fetus and newborn. Am J Physio11993;265: H1056-63. 34. Windmueller HG, Spaeth AE. Respiratory fuels and nitrogen metabolism in
THE JOURNAL OF PEDIATRICS AUGUST 1997 vivo in small intestine of fed rats: quantitative importance of glutamine, glutamate and aspartate. J Biol Chem 1980;255: 107-12. 35. Tremel H, Kienle B, Weilemann LS,
Stehle P, Furst P. Glutamine dipeptidesupplemented parenteral nutrition maintains intestinal function in the critically ill. Gastroenterology 1994;107: 1595-601.
CORRECTION The following table is the corrected version of the table appearing on page 815 of the article "Recurrent
Streptococcuspneumoniae Sepsis in Children With Sickle Cell Disease," by Hongeng et al., published in the May 1997 issue of The Journal (volume 130, pages 814 to 816):
Table. Characteristics of patients, penicillin prophylaxis use, and outcome of recurrent S, pneumoniae sepsis
232