Neonatal necrotizing enterocolitis (NEC) is the most common serious gastrointestinal disorder affecting preterm infants (1). Among infants with very low birth weight (VLBW; <1500 g), the incidence is 10% to 15% with a mortality of 10% to 30% (2). A multifactorial pathogenesis has been suggested, involving risk factors such as gastrointestinal ischemia, enteral alimentation, and microorganisms in combination with immaturity of gastrointestinal function and host defense mechanisms. However, prematurity is the primary risk factor (3). The diagnosis of NEC is based on clinical, radiographic, and laboratory signs, but there is a lack of a simple diagnostic test that, at an early stage of the disease, would differentiate NEC from conditions such as neonatal sepsis or focal intestinal perforation (FIP), which can present themselves with similar clinical pictures to that of NEC (4). Inflammation and necrosis of the bowel are typical features of NEC, and neutrophils are important in this process (5).
Calprotectin is a calcium-binding protein found predominantly in neutrophils and macrophages; in the former it may account for as much as 60% of the protein in the cytosol (6). In vivo it seems to have a regulatory function in the inflammatory process as well as antimicrobial and antiproliferative activity (7). A method for determination of fecal calprotectin (f-calprotectin) concentration has recently been developed (8,9). F-calprotectin is regarded as a marker of inflammation in the gastrointestinal tract, with its presence in stool most likely caused by transepithelial migration of myeloid cells, and has been used clinically to follow the activity in inflammatory bowel disease (10–12). It may therefore be a marker of NEC and would be of particular use if its increase predated the clinical features. Concentrations of f-calprotectin in healthy term infants are more than 5 times higher than in healthy children and adults (13). We are aware of no studies of f-calprotectin in VLBW infants.
The aims of this study were to investigate the distribution of f-calprotectin in VLBW infants during the first 8 weeks of life, assess what factors influence f-calprotectin levels in healthy VLBW infants, and describe changes in f-calprotectin in infants who develop severe abdominal disease.
PATIENTS AND METHODS
Between December 2002 and May 2004, 59 VLBW infants were recruited from the neonatal unit at Umeå University Hospital in Sweden. Written informed consent was obtained from the parents and the study was approved by the university's ethical committee. The infants were followed until term or discharge. Inclusion criteria were birth weight <1500 g and postnatal age no greater than 2 weeks. Of the 59 infants all except 2 were included at the date of birth. Infants staying <1 week at the neonatal unit were subsequently excluded. Seven infants fulfilled criteria for disease as defined by NEC according to Bell's criteria (14) or any other condition that led to abdominal surgery. These disease cases are presented as case studies, and the remaining 52 infants are considered reference infants. F-calprotectin results did not affect diagnosis or treatment because they were not available to the clinicians.
Perinatal data were retrieved from hospital charts. Maternal data included preeclampsia, diabetes, amnionitis, premature rupture of the membranes, and mode of delivery (vaginal or cesarean section). Neonatal data for each infant included birth weight, sex, gestational age, first pH value measured within 2 hours of birth, Apgar score, number of days on a respirator, and number of days on nasal continuous positive airway pressure (CPAP). For the calendar day of each sample, the following data were collected: postnatal age, type of enteral feeding (own mother's milk, donated breast milk, fortified breast milk, preterm formula, term formula), volume of enteral feeding, total volume of fluids (enteral and parenteral), ratio between enteral and parenteral volume, withheld feeds (yes/no), antibiotics/antimycotics (ampicillin, gentamicin, vancomycin, cefotaxime, meropenem, fluconazole, amphotericin B), systemic corticosteroids (betamethasone), and c-reactive protein (CRP). Because CRP was not analyzed daily, we used the value closest in time to the stool sample, within the range of 3 days before to 1 day after. At the time of collection of each sample, the following binary data were recorded: blood in stool, mucus in stool, and presence of abdominal symptoms.
Disease cases were reviewed separately and the diagnoses were reevaluated by a radiologist, a pediatric surgeon, and a pathologist with access to hospital charts but not to f-calprotectin results.
Entire stool samples were taken from the infants' diapers with sterile plastic spoons and put in sterile plastic screw-cap tubes (76 × 20 mm; Sarstedt, Nümbrecht, Germany). At every diaper change (normally every 2–3 hours) during the initial hospital stay, all available stool was collected, making it possible to follow the changes in f-calprotectin level in detail before abdominal disease diagnosis in disease cases. When infants reached a postnatal age of 4 weeks, with a postconceptional age of >33 weeks, and showed no signs of sepsis or abdominal disease, sample collection was limited to 1 stool per 24 hours (the first available stool every calendar day). Date and time of each stool were recorded and the presence of blood in stools and gastrointestinal symptoms were noted. Within 30 minutes from the time of collection, samples were stored at −20°C until analysis.
Sample Preparation and Analysis
For all infants, starting at week 0, the first stool after each 7 days of postnatal age until 8 weeks was selected for analysis. Because the sample at week 0 was the first stool available after birth, we will refer to this as meconium. In addition, in disease cases daily samples were selected from 1 week before until 1 week after abdominal disease diagnosis and, within 24 hours of diagnosis, all of the available samples were selected. F-calprotectin was analyzed using a commercially available enzyme-linked immunosorbent assay kit (Calprest; Eurospital, Trieste, Italy) as previously described (9). In brief, ≈100 mg of feces was taken from each stool sample and fecal extraction buffer was added. Samples were then homogenized and centrifuged. Supernatants were applied in duplicate to the assay plate. Blanks, standards, and calprotectin controls were included in each run. Optical densities were read at 405 nm, and f-calprotectin was calculated from the standard curve and expressed as micrograms per gram of feces. Coefficients of variation were <4% within and <10% between assays. Duplicate extractions were analyzed in 22 stool samples and the mean coefficient of variation between extractions was 12%.
Statistical analyses were performed using SPSS version 12.0 (SPSS, Chicago, IL). Because f-calprotectin values were skewed to the right, they were log-transformed in all analyses and then transformed back to the original units for presentation.
Using repeated measures analysis to study crude variations in f-calprotectin levels over time, a statistically significant time effect (P = 0.033) was found. This effect was caused by an initial decrease in median f-calprotectin level from week 0 to week 1 followed by a slow increase in f-calprotectin level from week 1 to week 8 (Fig. 1). Meconium (week 0) and postmeconium (weeks 1–8) samples were analyzed separately because f-calprotectin levels were significantly higher in meconium.
Linear regression was used to evaluate how each variable influenced f-calprotectin. For meconium, all of the factors that showed a statistically significant correlation to f-calprotectin with simple linear regression were analyzed together in a stepwise, multivariate regression analysis. For postmeconium samples, data for each postnatal week were first analyzed separately. Variables that showed statistical significance for ≥2 weeks were analyzed in a stepwise multivariate regression analysis, adjusting for postnatal age and infant identification. P < 0.05 was considered statistically significant.
Reference infants had a mean (±SD) gestational age of 27.2 ± 2.6 weeks (range, 23.0–33.7 weeks) and a birth weight of 939 ± 273 g (range, 519–1442 g). There were 28 boys and 24 girls. Ten infants were delivered vaginally and 42 by cesarean section. Median (range) Apgar scores at 1, 5, and 10 minutes were 7 (2–9), 8 (3–10), and 8 (4–10), respectively. Median number of days on a respirator was 2 (range, 0–55) and median number of days on nasal CPAP (at our hospital) was 29 (range, 0–86). Seven of the 52 infants were transferred to other hospitals while still receiving nasal CPAP. Most infants (90%) received minimal enteral nutrition on the first day of life. The average time to full (>90%) enteral nutrition was 2.8 ± 1.6 weeks. During the first 8 weeks of life, 96% of the infants received their own mother's breast milk, 88% received donated breast milk, 60% received human breast milk fortifier, 12% received preterm formula, and 2% received standard infant formula. On the day of each sample, antibiotic use was as follows: ampicillin, 31%; gentamicin, 24%; vancomycin, 13%; cefotaxime, 16%; meropenem, 1%; fluconazole, 17%; and amphotericin B, 9%. At these time points the incidence of increased CRP (>10 mg/L) was 19% and the incidence of betamethasone treatment was 6%.
Concentrations of f-calprotectin are presented in Fig. 1. There was a significant decrease in f-calprotectin level from week 0 (ie, meconium) to week 1 (median, 332 μg/g; range, 12–9386 μg/g vs median, 190 μg/g; range, 9–942 μg/g; P = 0.024; using Mann-Whitney test). This was followed by a significant increase in f-calprotectin from week 1 to week 8 (r = 0.14; P = 0.020; linear regression). Median f-calprotectin in all postmeconium samples (weeks 1– 8) was 253 μg/g (range, 9–1867 μg/g). The interindividual variation in log f-calprotectin was approximately twice the intraindividual variation between time points (0.057 vs 0.024).
In meconium there were significant negative correlations between f-calprotectin and Apgar scores at 1, 5, and 10 min. In the stepwise multivariate regression analysis, only Apgar score at 10 min remained statistically significant (r = −0.40; P = 0.004). Otherwise, no statistically significant correlations were found between f-calprotectin in meconium and maternal factors (preeclampsia, diabetes, amnionitis, premature rupture of membranes, mode of delivery) or neonatal factors (birth weight, sex, gestational age, pH at birth, enteral feeds, antibiotics, corticosteroids, CRP). At week 0 no infant had abdominal symptoms recorded.
For postmeconium samples the explanatory factors for f-calprotectin that remained significant in the multivariate regression analysis are shown in Table 1, with postnatal age included as a possible confounder. Including infant identification in the model did not change the results. The other variables (maternal preeclampsia, diabetes, amnionitis, premature rupture of the membranes, birth weight, sex, gestational age, first pH value, Apgar score, number of days on respirator, number of days on nasal CPAP, postnatal age, type of enteral feed, volume of enteral feed, total volume of fluids [enteral and parenteral], ratio between enteral and parenteral volume, withheld feeding, ampicillin, gentamicin, vancomycin, fluconazole, amphotericin B, CRP, blood in stool, mucus in stool, and presence of abdominal symptoms) were not significant explanatory factors for f-calprotectin in the multivariate analysis.
To assess whether there was any association between high f-calprotectin and NEC-related symptoms among infants in whom a diagnosis of severe abdominal disease was ultimately not made, we identified reference infants with any suspicion of NEC, which we defined as any occasion of blood in stools, mucus in stools, or abdominal symptoms during the first 8 weeks of life. Of the 52 reference infants, 20 were classified as having “any suspicion of NEC.” Average postmeconium f-calprotectin at weeks 1 to 8 in these infants was not significantly different from median f-calprotectin in the remaining reference infants (238 vs 186 μg/g feces; P = 0.24). The variation was also similar (range, 88–924 vs 30–913 μg/g feces). Furthermore, the patient journals of all infants (n = 6) with postmeconium f-calprotectin levels >97.5th percentile (+2 SD; 1191 μg/g) were separately reviewed, and none of them had suspected NEC or other severe abdominal disorder during the week before or after the maximum f-calprotectin measurement, even though 2 of them had blood in stool. Of all of the reference infants, 6 had recurring gastrointestinal symptoms (blood in stool, mucus in stool, or abdominal symptoms). In all of the fecal samples from these infants, f-calprotectin was >97.5th percentile in 1.9% of the samples compared with 2.4% of the samples from infants without recurring gastrointestinal symptoms.
Disease Case Studies
Data on disease cases are presented in Table 2. There were 4 cases of NEC. In 1 of these (case 1) f-calprotectin was elevated above the range of reference infants (>2000 μg/g) in a stool sample obtained 4.5 hours before a radiological diagnosis was made. In 3 of the NEC cases, f-calprotectin was increased in the first stool sample after diagnosis, and in the remaining case (case 4) no postdiagnosis sample was available due to death within 24 hours. In case 3 no increase in f-calprotectin was seen 1 hour before radiological diagnosis, but f-calprotectin was increased in the subsequent stool sample 34 hours later. In case 4 no sample was available before diagnosis because the patient was transferred from another hospital, but f-calprotectin was increased in the first sample obtained 25 hours after radiological diagnosis.
There were 2 cases of FIP, and in both, f-calprotectin was not increased before or after diagnosis. There was 1 case of intestinal obstruction caused by subacute volvulus of the small intestine, leading to a covered perforation. In this case (case 7) mild abdominal symptoms started 20 days before surgery and f-calprotectin was increased in all of the fecal samples during the last 11 days preceding surgery.
We have measured f-calprotectin in 318 stool samples from 52 VLBW infants without severe abdominal disease. Our observed median f-calprotectin level in postmeconium samples of 253 μg/g in these reference infants is consistent with previous findings in term and moderately preterm infants. Olafsdottir et al found a mean (±SD) f-calprotectin level of 277 ± 109 μg/g in healthy term infants (13). Rugtveit et al found a median f-calprotectin level of 269 μg/g in 6-week-old healthy infants (15). Nissen et al studied f-calprotectin in 11 spot stool samples of healthy preterm infants (mean age, 31 weeks) and found a median (range) f-calprotectin level of 150 (81–221) μg/g (16). Our results support previous observations that healthy term and preterm infants have higher f-calprotectin than healthy adults, who have a mean f-calprotectin level of ≈50 μg/g (9). Possible explanations for the high f-calprotectin levels seen in infants include immaturity of the intestinal mucosa with increased intestinal permeability and physiological inflammation of the gastrointestinal tract, both leading to migration of neutrophil granulocytes into the gut lumen (17,18).
A weakness of the study is that samples were collected from diapers because absorption of water into the diaper may increase the concentration of calprotectin by as much as 30% (13). However, collection from diapers is the only method that is practically useful in the clinical setting, and this method has been used in all of the previous studies of f-calprotectin in infants (13,19,20).
Because antibiotic treatment included prophylactic and therapeutic indications, CRP was used as a proxy for current infection. However, there was no correlation between CRP and f-calprotectin in reference infants, suggesting that systemic infection does not affect f-calprotectin concentrations in the absence of severe abdominal disease. Our findings of significant negative correlations between f-calprotectin and antibiotic treatment may rather suggest that f-calprotectin concentrations in VLBW infants are influenced by bacterial colonization of the gut. The positive correlation between f-calprotectin and delivery by cesarean section also supports this theory because this procedure has been shown to delay the normal colonization of the gut (21,22). The opposite effects of antibiotic treatment and cesarean section on f-calprotectin may be explained by different effects of these factors on individual bacterial species. The positive correlation between f-calprotectin and increasing amounts of enteral feeds could be the result of the immunological stimulation of the mucosal immune system by novel intraluminal antigens and colonization of the gut by Gram-negative bacteria, as has been described during infancy (23,24).
In meconium we found significant negative correlations between f-calprotectin and Apgar scores at 1, 5, and 10 min, with the strongest correlation for 10-minute scores. In a study by Laforgia et al (19), f-calprotectin in meconium was investigated in 84 term and 47 preterm infants. Mean f-calprotectin was 145 ± 79 μg/g and negatively correlated to gestational age, birth weight, and 5-minute Apgar score (1- and 10-minute scores not investigated). Our observed median f-calprotectin level in meconium (332 μg/g) was higher than that in the study of Laforgia et al (19), possibly because of the lower gestational age (27.2 ± 2.6 weeks vs 34.3 ± 2.5 weeks) or the lower 5-minute Apgar scores of the infants in our study. We identified no correlation between f-calprotectin in meconium and gestational age or birth weight within this cohort of VLBW infants. Our observation of a negative correlation between f-calprotectin in meconium and Apgar score supports those authors' theory that the migration of neutrophils into the gut lumen (and consequently f-calprotectin in meconium) is increased in perinatal asphyxia because of ischemic changes in the intestinal mucosa (19). Through this mechanism the hypoxia associated with a normal delivery may also explain the higher f-calprotectin concentrations observed in meconium compared with postmeconium samples.
In the multivariate analysis of postmeconium samples from reference infants, there was no significant correlation between f-calprotectin and any variable associated with NEC (abdominal symptoms, blood in stool, mucus in stool, CRP). Furthermore, we found that f-calprotectin was not increased in infants with NEC-like symptoms who do not develop NEC and that moderately increased f-calprotectin in reference infants (1200–2000 μg/g) was not associated with NEC. In our study no reference infant had a postmeconium f-calprotectin level >2000 μg/g, whereas f-calprotectin was increased to >2000 μg/g in 3 cases of NEC and 1 case of covered perforation with microscopic bowel inflammation. A lower cutoff point (1200 μg/g) resulted in the same sensitivity (identical results for the disease cases) but a lower specificity (false-positive reference infants). We therefore propose a cutoff for f-calprotectin level in VLBW infants at 2000 μg/g. The highest fecal calprotectin concentrations in our study (>20 000 μg/g) were seen in 1 of the NEC cases and were similar to concentrations found by Tøn et al in hospitalized adult patients with pancolitis caused by ulcerative colitis (9).
One previous pilot study by Carroll et al (25) has shown that f-calprotectin was increased in patients after the diagnosis of NEC, and it has therefore been speculated that f-calprotectin may be an early marker of NEC. We have obtained a unique collection of consecutive stool samples from cases with NEC and other severe abdominal diseases, covering the period from before the development of clinical symptoms and throughout the course of the disease. Because only 1 of the 3 NEC cases for which prediagnosis samples were available had increased f-calprotectin before radiological diagnosis, it seems that f-calprotectin is not generally an early marker of NEC. However, in all 3 NEC cases with postdiagnosis samples available, f-calprotectin was increased in the first stool sample after the radiological diagnosis, suggesting that f-calprotectin is a sensitive marker of NEC.
In a brief report by Carroll et al (25), spot fecal samples were analyzed from 7 preterm infants (mean gestational age, 30 weeks) after the diagnosis of NEC and from 7 matching controls. They also found significantly higher postdiagnosis f-calprotectin levels in patients with NEC compared with controls (288 ± 49 mg/L vs 98 ± 61 mg/L; P < 0.001), even though controls and patients with NEC had lower f-calprotectin levels than in our study. A possible explanation for the lower f-calprotectin levels in Carroll's disease cases compared with ours may be that they analyzed only 1 spot stool sample from each infant, whereas we studied our infants prospectively. It is more difficult to explain why f-calprotectin levels in Carroll's control infants were notably lower than levels observed in normal infants in other studies, as discussed earlier (13,15,16).
In the 2 cases of FIP f-calprotectin was not elevated above the range of the reference infants before or after diagnosis, even though 1 of these patients underwent abdominal surgery. In case 5 abdominal radiography showed pneumatosis intestinalis suggesting NEC, but f-calprotectin was negative. At laparotomy there was no macroscopic bowel necrosis and no microscopic bowel inflammation. A clinical diagnosis of FIP was made. Therefore, in this case f-calprotectin was more useful than radiology in differentiating between NEC and FIP.
In case 7 f-calprotectin was increased but there was no pneumatosis intestinalis and this patient did not fulfill the criteria for NEC. However, there was microscopic bowel inflammation, which suggests that f-calprotectin may be a useful marker not only in NEC but also in a case like this with a covered perforation and pronounced intestinal inflammation, which requires similar management.
A diagnosis of NEC is based on clinical, radiographical, and laboratory signs. Based on the clinical presentation, NEC can be difficult to differentiate from neonatal sepsis or FIP. The radiographic diagnosis of NEC can also be inconclusive. Tam et al (26) found high specificity but low sensitivity for the radiological signs traditionally associated with NEC (ie, pneumatosis intestinalis and portal venous gas), and Hwang et al (4) found that pneumatosis intestinalis, although more common in NEC, could also be seen in FIP.
It is possible that the disease entity of NEC, as defined by Bell's criteria, includes cases with different initial pathogenesis, ranging from an acute ischemic event to a more slowly developing inflammatory process. The inclusion of f-calprotectin as a diagnostic aid may help in differentiating these subtypes of NEC.
Our measurements of f-calprotectin in VLBW infants agrees with the findings of other groups in showing that f-calprotectin in meconium and postmeconium samples is considerably higher than in older children and adults. Our findings support f-calprotectin measurement as a noninvasive test of intestinal inflammation in this group—both physiological inflammation as seen with increasing enteral feeds and pathological inflammation after intrauterine asphyxia or postnatal disease processes such as NEC. Our results suggest that the increase in f-calprotectin does not usually predate the symptoms of NEC and that the concentration in many cases increases only after the occurrence of radiological changes. However, f-calprotectin measurement in repeated samples may be a helpful diagnostic adjunct in patients with suspected NEC who do not require immediate surgery. F-calprotectin levels >2000 μg/g in a diaper sample strongly suggest NEC or another severe intestinal inflammatory condition. It may also be useful to follow f-calprotectin after NEC surgery to monitor the improvement of the intestinal inflammation. The possible relations between f-calprotectin and gut flora in VLBW infants need to be studied further.
The authors thank Hans Stenlund for statistical advice; Rickard Palmqvist (Department of Pathology, Umeå), Birger Sandzén (Department of Surgery, Umeå), Staffan Meurling (Department of Paediatric Surgery, Uppsala), and Anna-Maja Zdunek (Department of Radiology, Umeå) for reviewing disease cases. The authors also thank Carina Lagerqvist and Yvonne Andersson for their assistance with laboratory analyses.
1. Kliegman RM. Neonatal necrotizing enterocolitis: bridging the basic science with the clinical disease. J Pediatr 1990; 117:833–835.
2. Kliegman RM. Neonatal necrotizing enterocolitis. In: Wyllie R, Hyams JS, editors. Pediatric Gastrointestinal Disease. Philadelphia: WB Saunders; 1999. pp. 455–465.
3. Kliegman RM. Models of the pathogenesis of necrotizing enterocolitis. J Pediatr 1990; 117(Suppl):S2–S5.
4. Hwang H, Murphy JJ, Gow KW, et al
. Are localized intestinal perforations distinct from necrotizing enterocolitis: J Pediatr Surg 2003; 38:763–767.
5. Musemeche C, Caplan M, Hsueh W, et al
. Experimental necrotizing enterocolitis: the role of polymorphonuclear neutrophils. J Pediatr Surg 1991; 26:1047–1049.
6. Dale I, Brandtzaeg P, Fagerhol MK, et al
. Distribution of a new myelomonocytic antigen (L1) in human peripheral blood leukocytes: immunofluorescence and immunoperoxidase staining features in comparison with lysozyme and lactoferrin. Am J Clin Pathol 1985; 84:24–34.
7. Johne B, Fagerhol MK, Lyberg T, et al
. Functional and clinical aspects of the myelomonocyte protein calprotectin. Mol Pathol 1997; 50:113–123.
8. Roseth AG, Fagerhol MK, Aadland E, et al
. Assessment of the neutrophil dominating protein calprotectin in feces: a methodologic study. Scand J Gastroenterol 1992; 27:793–798.
9. Tøn H, Brandsnes O, Dale S, et al
. Improved assay for fecal calprotectin. Clin Chim Acta 2000; 292:41–54.
10. Berstad A, Arslan G, Folvik G. Relationship between intestinal permeability and calprotectin concentration in gut lavage fluid. Scand J Gastroenterol 2000; 35:64–69.
11. Roseth AG, Aadland E, Jahnsen J, et al
. Assessment of disease activity in ulcerative colitis by faecal calprotectin, a novel granulocyte marker protein. Digestion 1997; 58:176–180.
12. Bunn SK, Bisset WM, Main MJ, et al
. Fecal calprotectin as a measure of disease activity in childhood inflammatory bowel disease. J Pediatr Gastroenterol Nutr 2001; 32:171–177.
13. Olafsdottir E, Aksnes L, Fluge G, et al
. Faecal calprotectin levels in infants with infantile colic, healthy infants, children with inflammatory bowel disease, children with recurrent abdominal pain and healthy children. Acta Paediatr 2002; 91:45–50.
14. Bell MJ, Ternberg JL, Feigin RD, et al
. Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging. Ann Surg 1978; 187:1–7.
15. Rugtveit J, Fagerhol MK. Age-dependent variations in fecal calprotectin concentrations in children. J Pediatr Gastroenterol Nutr 2002; 34:323–324.
16. Nissen AC, van Gils CE, Menheere PP, et al
. Fecal calprotectin in healthy term and preterm infants. J Pediatr Gastroenterol Nutr 2004; 38:107–108.
17. Weaver LT, Laker MF, Nelson R. Intestinal permeability in the newborn. Arch Dis Child 1984; 59:236–241.
18. Cummins AG, Thompson FM. Postnatal changes in mucosal immune response: a physiological perspective of breast feeding and weaning. Immunol Cell Biol 1997; 75:419–429.
19. Laforgia N, Baldassarre ME, Pontrelli G, et al
. Calprotectin levels in meconium. Acta Paediatr 2003; 92:463–466.
20. Campeotto F, Butel MJ, Kalach N, et al
. High faecal calprotectin concentrations in newborn infants. Arch Dis Child Fetal Neonatal Ed 2004; 89:F353–F355.
21. Long SS, Swenson RM. Development of anaerobic fecal flora in healthy newborn infants. J Pediatr 1977; 91:298–301.
22. Gronlund MM, Lehtonen OP, Eerola E, et al
. Fecal microflora in healthy infants born by different methods of delivery: permanent changes in intestinal flora after cesarean delivery. J Pediatr Gastroenterol Nutr 1999; 28:19–25.
23. Amer MD, Hedlund E, Rochester J, et al
. Platelet-activating factor concentration in the stool of human newborns: effects of enteral feeding and neonatal necrotizing enterocolitis. Biol Neonate 2004; 85:159–166.
24. Cummins AG, Eglinton BA, Gonzalez A, et al
. Immune activation during infancy in healthy humans. J Clin Immunol 1994; 14:107–115.
25. Carroll D, Corfield A, Spicer R, et al
. Faecal calprotectin concentrations and diagnosis of necrotising enterocolitis. Lancet 2003; 361:310–311.
26. Tam AL, Camberos A, Applebaum H. Surgical decision making in necrotizing enterocolitis and focal intestinal perforation: predictive value of radiologic findings. J Pediatr Surg 2002; 37:1688–1691.