Necrotizing
Enterocolitis
in the First 24 Hours
of Life
Elizabeth H. Thilo, MD, Raul A. Lazarte, MD, and
Jacinto
A. Hemandez,
MD
From the Department of Perinatology, The Children’s Hospital, Denver
ABSTRACT. Necrotizing enterocolitis (NEC) is
corn-monly thought of as occuring in the sick premature infant, usually in the first one to two weeks of life. A
review of NEC at the Children’s Hospital of Denver over a 5-year period, found that 13 of 79 infants (16.1%) had onset of NEC during the first day of life. These infants were larger (mean birth weight 2,624 ± 849 g), more mature (mean gestational age 37.9 ± 2.5 weeks), and less
asphyxiated asjudged by Apgar scores (mean five-minute score 8.15 ± 1.07) than infants with onset of NEC after
the first day of life (mean birth weight 1,519 ± 586 g,
mean gestational age 32.0 ± 3.5 weeks, P < .001, and
mean five-minute Apgar score 6.81 ± 1.84, P < .05).
Despite their large size and degree of maturity, eight of
these infants (62%) showed signs of respiratory distress;
four (31%) were polycythemic; four (31%) had either a
partial or double-volume exchange transfusion
per-formed; and 11 (85%) were fed prior to developing NEC.
Presenting signs of disease, occurrence of sepsis (31%),
requirement for surgical intervention (62%), and mortal-ity (30%) were similar for the two groups of infants. It is
suggested that term and near-term infants who have
significant illness after delivery be treated more like their premature counterparts with cautious introduction of feedings after an adequate period of stabilization. Pedi-atrics 1984;73:476-480; necrotizing enterocolitis,
polycy-themia, exchange transfusion.
Necrotizing enterocolitis (NEC) is commonly thought of as occurring in the sick premature
in-fant, usually in the first one to two weeks of life.
Indeed, 75% to 95% of cases of NEC occur in
infants of less than 38 weeks’ gestation,’ and with
onset typically at 4 to 6 days of age.5 Stoll et a14 in their series reporting on 35 infants with NEC, noted an inverse relationship between gestational age and age at onset of NEC. Kliegman and Fanaroff,3 in
Received for publication Dec 16, 1982; accepted May 16, 1983. Reprint requests to (E.H.T.) Department of Newborn Medicine, Rose Medical Center, 4567 East Ninth Aye, Denver, CO 80220.
PEDIATRICS (ISSN 0031 4005). Copyright © 1984 by the
American Academy of Pediatrics.
their large series of 123 cases, spanning 9 year’s
experience, did not note any such relationship. Al-though they found three patients (2.4%) with onset of disease in the first 24 hours of life, they did not comment any further on this phenomenon.
While reviewing our experience with NEC at the Newborn Center ofthe Denver Children’s Hospital, we found 13 infants (16% ofthe total) over a 6-year
period who had onset of their disease on the first
day of life. It is this group of patients that we would like to characterize further.
MATERIALS
AND
METHODS
The charts of all infants cared for in the Newborn
Center of the Denver Children’s Hospital (a re-gional referral center for newborn intensive care) from September 1975 through June 1981, whose condition had been diagnosed as NEC were
re-viewed. To be included in this analysis, the
diag-nosis was considered secure only if the infants had
clinical evidence of the disease (gastric retention, abdominal distension, and occult or gross blood in
the stool) together with radiographic evidence
(pneumatosis intestinahis, portal venous air, or
pneumoperitoneum) or histopathologic evidence of
the disease in specimens obtained at laparotomy or autopsy. Suspected cases without such confirma-tion were not included. The information gathered
from each record included: sex; birth weight;
ges-tational age; complications of pregnancy, labor, and delivery; Apgar scores; problems affecting the new-born (such as respiratory distress syndrome,
venti-latory support required, sepsis, shock,
hypoglyce-mia, catheter placement, polycythemia, exchange
transfusion, patent ductus arteriosus, etc); time of
onset of NEC; presenting signs of NEC; type of cultures obtained at the time of diagnosis and their results; laboratory parameters and radiographic findings at the time of diagnosis; therapy required;
TABLE 1. Clinical Characteristics of Patients with
Necrotizing Enterocolitis (NEC) as Related to Time of Onset of NEC*
P <.001 <.001 <.001 NS NS <.050 NS <.050 <.001 <.010 <.010 NS <.001 11.8±4.6 152.0 ± 171.3t
2(15) 4(6) NS
<.001 Polycythemia, defined by a venous hematocrit of 65% or more, was present much more frequently in
group I than group II (4/13 v 1/66, P <
.001).
Umbilical venous catheters were placed more
fre-quently in group I infants (3/13 v 4/66, P < .01),
and exchange transfusions, either partial or double-volume, were performed more frequently in group I than in group II infants (4/13 v 2/66, P < .01).
Medical management of the disease included im-mediate cessation of feedings, nasogastric suction for decompression, systemic antibiotics (usually ampicillin and gentamicin), intravenous fluid and colloid support as needed, frequent clinical
exami-nations, and frequent abdominal radiographs. In-dications for surgery were: clinical deterioration despite aggressive medical management, signs of peritonitis (abdominal tenderness with abdominal
wall erythema or induration), abdominal mass, pneumoperitoneum, or a persistently dilated loop of bowel unchanged in position on abdominal ra-diographs over a 12- to 24-hour period.6’7
Associations were tested for statistical signifi-cance by 2 and Fisher’s exact test, and means were
compared using Student’s t test.
RESULTS
Over the 6-year period reviewed, 79 cases of doc-umented NEC were found. Thirteen of these (16%)
were diagnosed in the first 24 hours of life. The
clinical characteristics of these 13 patients and those ofthe other 66 patients are presented in Table
1.
No. of infants
Birth weight (g)
Gestation (wk)
No. 38 wk
SGA infants Apgar score 1-mm 5-mm Respiratory distress UAC placed Polycythemia (venous Hct 65%) UVC placed Exchange transfusion
No. never fed
Age feedings began (h)
Interval between
feedings and onset of NEC (h) Associated anomaly
Groupl: Group II
Onset at Onset at
Age Age
24h(%) >24h(%)
13 (16) 66 (84) 2,624 ± 849 1,519 ± 586
37.9 ± 2.5 32.0 ± 3.5
7 (54) 4 (6) 5 (38) 10 (15)
6.08 ± 2.29 4.61 ± 2.53 8.15 ± 1.07 6.81 ± 1.84
8 (62) 39 (59)
2 (15) 40 (61) 4 (31) 1 (1.5)
3 (23) 4 (6)
4(31) 2(3)
2(15) 4(6)
7.0 ± 3.9 72.3 ± 42.7
* Values are means ± SD. Percent is shown in
parenthe-ses. Abbreviations used are: SGA, small for gestational
age; UAC, umbilical arterial catheter; UVC, umbilical venous catheter; Hct, hematocrit.
t Excluding one infant with onset at 5.5 months.
Birth
Weight
and Gestational
Age
Infants with onset of NEC in the first 24 hours of life (group I) were considerably larger and more
mature than their counterparts with later onset of disease (group II), with mean (±SD) birth weight
2,624 ± 849 g and mean gestational age 37.9 ± 2.5
week, respectively (P < .001). Seven (54%) of the
infants with onset of NEC in the first 24 hours were born at 38 weeks of gestation or more com-pared with only four infants who had (6%) later-onset NEC (P < .001). Infants in group I were more
often small for gestational age (SGA) as determined by the Colorado intrauterine growth chart8 than were those in group II (38% v 15%), but this
differ-ence did not reach statistical significance.
Apgar
Scores,
Respiratory
Distress,
and
Umbilical Arterial CathetersGroup I infants were less asphyxiated than their
counterparts with later onset of disease as judged by mean Apgar scores. Mean (±SD) score at one
minute was 6.08 ± 2.29 for group I u 4.61 ± 2.53 for group II (P = NS). The mean five-minute scores
were 8.15 ± 1.07 and 6.81 ±
1.84,
respectively (P <.05).
Although the infants in group I were larger and
gestationally more mature than those in group II,
the overall incidence of respiratory distress was similar in both groups (62% v 59%). However, the
infants in group I more frequently had mild
respi-ratory distress such as transient tachypnea (6/8) or due to hyperviscosity (1/8) requiring only hood oxygen for treatment, whereas those in group II
more frequently had hyaline membrane disease
re-quiring mechanical ventilation. Only two of the infants in group I had prolonged rupture of the fetal
membranes, and only one of these (a 35-week
ges-tation, 2,325-g male infant) had respiratory distress
which was diagnosed on clinical and radiologic
cri-teria as transient tachypnea rather than pneumo-nia. Because of their more severe respiratory dis-ease, infants in group II had umbilical arterial catheters placed more often (61% u 15%, P < .05).
Polycythemia, Umbilical Venous Catheters, and
Age at Feeding
and Onset of NEC
Two infants in group I and four in group II (15% and 6%, respectively, P = NS) were never fed prior
to developing signs of NEC. The feedings given
consisted of 5% dextrose in water, breast milk, or
standard 20-calorie infant formula in similar pro-portions in both groups. The mean age at which
feedings were begun was, of course, younger in
group I than group II infants (7.0 ± 3.9 hours v 72.3 ± 42.7 hours, P < .001) due to the criteria for
selection of the groups. More interesting, however, was the difference in lag-time between the intro-duction of feedings and the time of onset of NEC.
Infants in group I became symptomatic at a mean ofonly 11.8 ± 4.6 (SD) hours after the first feeding,
whereas symptoms of NEC did not occur for a mean of 152.0 ± 171.3 hour (six days) after feedings were begun in group II (P < .001).
Anomalies
and Other Complications
, Two infants in group I and four in group II had associated major congenital anomalies. These six infants included one with myelomeningocele, one
with Down’s syndrome who had onset of NEC within the first 24 hours, one hydropic infant, and three infants with congenital heart disease (valvu-lar pulmonic stenosis; hypoplastic left ventricle; atrial and ventricular septal defects with right-sided aortic arch) who had onset of NEC after 24 hours of age.
There were no significant differences between the two groups in the incidence of or types of complications of pregnancy, labor, or delivery other than premature labor.
Temporal
Occurrence
of Cases
of NEC
The base-line rate of occurrence of NEC during
the 6-year period reviewed was 0 to 3 cases per
month. There were two months, December 1976 and March 1981, with more than three cases.
In December 1976, there were five cases, four of which were acquired in our unit between Dec
16-28. All four infants were premature with birth
weight of 910 to 2,000 g and gestational age of 28
to 34 weeks. All had respiratory distress syndrome
and 3/4 required mechanical ventilation. One of the four grew Clostridia sp from a peritoneal aspirate; otherwise cultures were negative. All four infants became symptomatic after 24 hours of age.
In March 1981, there were again five cases, four of which were diagnosed at a single referring hos-pital in our region between March 8-29. Two of
these infants were dizygotic twins. None of them
had severe respiratory distress syndrome requiring mechanical ventilation, although all were
prema-ture with birth weight of 1,530 to 2,140 g and gestational age of 33 to 36 weeks. Despite blood,
CSF, and stool cultures, and surgery required in two of these infants, no common etiologic organism
was found. All infants became symptomatic after
24 hours of age.
Presenting Signs and Clinical Course (Table 2)
Presenting signs of NEC were similar in both
groups. Abdominal distension was the most
com-mon sign (85%), followed by heme-positive stools (55% to 75%) and gastric retention (44% to 54%).
Apnea and bradycardia were also common in group II (42%), the more premature infants.
Septicemia with positive blood, CSF, or
perti-oneal cultures occurred with equal frequency in
both groups (30%). The most common organisms isolated were Escherichia coli and Kiebsielki sp, which together accounted for more than half of the
isolates. Other organisms isolated were Clostridia sp, Staphylococcus aureus, Enterobacter sp, Bacte-roides fragilis, and Pseudomonas sp.
Pneumatosis, pneumoperitoneum, or portal ye-nous air was present in 13/13 patients in group I (100%) and in 59/66 (89%) patients in group II.
Surgical intervention was required with similar frequency in both groups (60% to 65%). Four of eight (50%) infants in group I had surgery because of perforation, and four (50%) because of signs of peritonitis with suspected necrotic bowel. In group II, 28/42 infants (67%) had surgery because of
perforations and 14 (33%) because of signs of pen-tonitis. There are no significant differences
be-tween group I and group II regarding indications
for surgery.
Mortality from the disease was similar in both
groups (31% for group I v 23% for group II, P =
NS) for an overall mortality of 24%. There was no statistically significant difference in mortality
be-TABLE 2. Clinical Course of Infants with Necrotizing
Enterocolitis (NEC) as Related to Time of Onset of NEC
Group I: Group II: P
Onset at Onset at Age Age >24h(%)
24 h (%)
Presenting sign
Abdominal distension 11 (85) 59 (89) NS
Heme-positive stools 7 (54) 50 (76) NS
Gastric residuals 7 (54) 29 (44) NS
Apnea/bradycardia 0 28 (42) <.003
Positive blood, CSF, or 4 (31) 20 (30) NS peritoneal cultures
Surgery required 8 (61) 42 (64) NS
Perforation 4 (50) 28 (67) NS
Signs of peritonitis 4 (50) 14 (33) NS
Mortality 4 (31) 15 (23) NS
tween those infants treated medically and those
treated surgically. Four of 29 patients treated
med-ically died (14%) v 15/50 who required surgery
(30%, P = NS). However, ofthose treated medically
who died, two in group I and two in group II, surgery
was not pursued because of either severe congenital anomalies (one infant with Down’s syndrome in
group I, one with hypoplastic left ventricular syn-drome in group II), or moribund condition at the
time of arrival in our unit (one in each group).
All
of those treated medically who did not have such a coexisting problem survived.
Post-NEC stricture occurred in none of the nine survivors in group I, and in 5/51 survivors in group
II (P = NS). The occurrence of stricture was not
different whether patients were treated medically
(2/25 survivors) or surgically (3/35 survivors) for
an overall incidence of 8% of survivors of NEC.
DISCUSSION
The fact that NEC can occur during the first day of life, in primarily term and near-term infants, has
not been appreciated in the past. Other authors4’9”#{176}
have noted that the age at onset of NEC is inversely proportional to gestational age, but none have fo-cused on the group of babies with onset in the first
24 hours ofhife. Better recognition ofthis possibility in these “lesser-risk” patients may lead to an
im-proved outcome for them just as a higher index of
suspicion and more aggressive early therapy have
improved the outcome for premature infants with NEC.”2
Although the precise etiology and
pathophysiol-ogy of NEC remains unclear, a variety of factors have been imphicated.’#{176} In addition, several re-cent papers4’2’ cast serious doubt on the concept of using “risk factors” to predict which infants are likely to develop NEC because the risk factors are
present just as frequently in premature infants who
don’t develop NEC as in those who do. Several of these risk factors were present in our infants with onset of NEC in the first 24 hours of life. These infants, however, differed from the majority of in-fants with NEC seen in our unit, and from the “typical” infant with NEC described in the literature5’7 in that they were larger, more mature, and less asphyxiated. Respiratory distress of van-ious degrees was recognized in 8/13 (62%), and, despite their respiratory difficulties, most of them were fed within the first 12 hours oflife. We suspect that these infants, because of their large size and degree of maturity were fed more aggressively than comparably ill but premature infants would have been. The records available to us were inadequate to further substantiate this suspicion.
Although conflicting opinions exist regarding the
relationship of oral feedings to the development of NEC,4”5”8 it appears from our cases that there was a close temporal association between the institution of feedings and the early development of NEC. A longer period of fasting and more cautious intro-duction of feedings could explain the much longer lag-time between first feeding and onset of NEC seen in the more premature group II infants com-pared with group I infants as has been suggested by Teasdale et al.9 Alternatively, as suggested by Law-rence et al,22 the lag-time in group II may have been prolonged due to the usual nursing practices af-forded these small, premature infants, practices that tend to delay their gastrointestinal bacterial colonization. The interval between time of first feeding and onset of NEC may have been shortened
in at least one infant in group I by in utero coloni-zation in the presence of chonioamnionitis. This infant had prolonged rupture of membranes and subsequently developed NEC with B fragili and Clostridium bacteremia. Chonioamnionitis did not seem to be a factor in any of the other infants in group I. Finally, as with the time of first feeding, the difference in lag-time between the groups could simply be inherent to the selection of the groups.
An additional risk factor was also present in our infants. Polycythemia was found in four ofthese 13 infants (31%). This incidence is much higher than the 3% to 5% reported for the general newborn population.’24 Three of these four infants were also SGA and were treated with a partial exchange transfusion in the first few hours of life, and all of them were fed early. Although the role of polycy-themia and hyperviscosity in the etiology of NEC is still a controversial issue,3’21’’ our findings are consistent with those reported by Kliegman and Fanaroff3 and Hakanson and Oh5, that polycythe-mia with hyperviscosity probably is a valid risk factor for the development of NEC, particularly in infants who are SGA. Hakanson and Oh5 reported that SGA infants who were hyperviscous had a 36% incidence of NEC, whereas in nonhyperviscous SGA infants the incidence of NEC was only 1.5%. In our polycythemic infants, viscosity studies were not available; however, the likelihood that they were hyperviscous with venous hematocnits of 65% or more approaches 100%.23
reported as a complication of exchange transfusion through an umbilical venous catheter that is posi-tioned in the inferior vena cava.’3 Whether the exchange procedure on the polycythemia itself was most contributory to the development of NEC can-not be delineated by this retrospective review. Hopefully, prospective studies will be able to answer this question.
Polin et al2 described two additional risk factors for the development of NEC in term infants: con-genital heart disease associated with poor cardiac output and protracted diarrhea. None of our infants with onset on the first day of life had either of these
conditions. Finally, Teasdale et al9 described a more severe course of early-onset NEC (<7 days of age)
compared with late-onset disease (>7 days of age). They noted a more frequent need for surgical inter-vention and a higher mortality in their early-onset group. This trend was not supported by our cases.
SUMMARY
Infants who develop NEC in the first 24 hours of
life tend to be term or near term, are frequently SGA on polycythemic, and frequently have mild respiratory distress. While we cannot determine the
mechanism of disease from our retrospective data,
we do wish to make two recommendations: (1) term
and near-term infants with respiratory distress
must be treated more like their premature
counter-parts with cautious introduction of feedings after an adequate period of stabilization, and (2) infants
with polycythemia, whether treated with partial exchange transfusion or not, should be carefully
observed and cautiously fed, with a high index of suspicion for early signs of NEC.
ACKNOWLEDGMENTS
The authors thank Dennis W. Luckey for performing the statistical analyses and Jean Hoffman for assistance
in preparation of this manuscript.
REFERENCES
1. Santulli TV, Schullinger JN, Heird WC, et al: Acute necro-tiring enterocolitis in infancy: A review of 64 cases.
Pediat-rics 1975;55:376-387
2. Polin RA, Pollack PF, Barlow B, et al: Necrotizing entero-colitis in term infants. J Pediatr 1976;89:460-462
3. Kliegman RM, Fanaroff AA: Neonatal necrotizing entero-colitis: A nine-year experience. I. Epidemiology and uncom-mon observations. Am J Dis Child 1981;135:603-607
4. Stoll BJ, Kanto WP, Glass RI, et al: Epidemiology of
nec-rotizing enterocolitis: A case control study. J Pediatr 1980;96:447-451
5. Hakanson DO, Oh W: Necrotizing enterocolitis and hyper-viscosity in the newborn infant. J Pediatr 1977;90:458-461 6. Wexler HA: The persistent loop sign in neonatal necrotizing
enterocolitis: A new indication for surgical intervention?
Radiology 1978;126:201-204
7. Burrington JD: Profile of the population at risk, in Moore TD (ed): Necrotizing Enterocolitis in the Newborn Infant,
Report of the Sixty-eighth Ross Conference on Pediatric Research. Columbus, OH, Ross Laboratories, 1975, pp 23-25
8. Lubchenco LO, Hansman C, Boyd E: Intrauterine growth in length and head circumference as estimated from live births at gestational ages from 26 to 42 weeks. Pediatrics
1966;37:403-408
9. Teasdale F, LeGuennec JC, Bard H, et al: Neonatal necro-tizing enterocolitis: The relation of age at the time of onset to prognosis. Can Med Assoc J 1980;123:387-390
10. Wilson R, Kanto WP, McCarthy BJ, et al: Age at onset of necrotizing enterocolitis. Am J Dis Child 1982;136:814-816 11. Schullinger JN, Mollitt DL, Vinocur CD, et al: Neonatal
necrotizing enterocolitis: Survival, management, and corn-plications: A 25-year study. Am J Dis Child 1981;135:612-614
12. Kliegrnan RM, Fanaroff AA: Neonatal necrotizing entero-colitis: A nine-year experience. II. Outcome assessment. Am
JDis Child 1981;135:608-611
13. Thangavel M, Pornerance JJ, Raj C: Catheter tip localiza-tion during umbilical venous exchange transfusion and nec-rotizing enterocolitis. Pediatrics 1982;69:799-800
14. Kosloske AM: Necrotizing enterocolitis in the neonate. Surg
Gyn Obstet 1979;148:259-269
15. Eidelrnan A!, Inwood RJ: Necrotizing enterocolitis and en-teral feeding Is too much just too much? Am J Dis Child 1980;134:545-456
16. Lehmiller DJ, Kanto WP: Relationships of mesenteric thromboembolism, oral feeding, and necrotizing enterocoli-tis. J Pediatr 1978;92:96-100
17. Chany C, Moscovici 0, Lebon P, et al: Association of coron-avirus infection with neonatal necrotizing enterocolitis.
Pe-diatrics 1982;69:209-214
18. Book LS, Herbst JJ, Jung AL: Comparison of fast and slow-feeding rate schedules to the development of necrotizing enterocolitis. J Pediatr 1976;89:463-466
19. Kitterinan JA: Effects of intestinal ischemia, in Moore TD (ed): Necrotizing Enterocolitis in the Newborn Infant. Report of the sixty-eighth Ross Conference on Pediatric Research. Columbus, OH, Ross Laboratories, 1975, pp 38-40 20. DeLemos BA: The role of hyperosmolar formulas in
necro-tizing enterocolitis-Animal Studies, in Moore TD (ed):
Necrotizing Enterocolitis in the Newborn Infant. Report of
the Sixty-eighth Ross Conference on Pediatric Research. Columbus, OH, Ross Laboratories, 1975, pp 75-77 21. Kliegman RM, Hack M, Jones P, et al: Epidemiologic study
of necrotizing enterocolitis among low-birth-weight in-fants-Absence of identifiable risk factors. J. Pediatr 1982;100:440-444
22. Lawrence G, Bates J, Gaul A: Pathogenesis of neonatal necrotizing enterocolitis. Lancet 1982;1:137-139
23. Wirth FH, Goldberg KE, Lubchenco LO: Neonatal hyper-viscosity. I. Incidence. Pediatrics 1979;63:833-836
24. Black VD, Lubchenco LO, Luckey DW, et al: Developmen-tel and neurologic sequelae of neonatal hyperviscosity syn-drome. Pediatrics 1982;69:426-431
25. Ramamurthy RS, Brans YW: Neonatal polycythemia: I.
