Calprotectin is a calcium- and zinc-binding protein that constitutes approximately 60% of soluble cytosol proteins in human neutrophil granulocytes. Its measurement in feces is now recognized as a reliable marker for the detection of organic intestinal diseases and the assessment of inflammatory activity in the gastrointestinal tract (1).
In preterm infants, intestinal complications such as enteropathy or necrotizing enterocolitis (NEC) are not uncommon during the first weeks of life. Clinical and radiological signs currently inform the diagnosis but can be inconclusive, especially for enteropathy. This uncertainty highlights the need for a noninvasive biological marker. The usefulness of calprotectin as such a marker remains controversial because of high interindividual variations in healthy full-term and preterm infants during the first weeks of life (2–4). In infants with mild digestive symptoms, such as gastrointestinal bleeding or abdominal distension, we have previously shown a transient increase of fecal calprotectin levels when symptoms occur (2). In a pilot study, Carroll et al (5) identified significantly higher levels of fecal calprotectin in 7 preterm neonates with proven NEC compared with gestational age–matched healthy infants. Josefsson et al (4) proposed a cutoff level above 2000 μg/g for identifying a severe intestinal inflammatory condition in very low birth weight preterm infants. To try to define cutoff levels related to the risk of mild or severe enteropathy, the aim of this study was to evaluate the level of fecal calprotectin in different populations of preterm infants according to digestive events in the first month of life.
PATIENTS AND METHODS
This multicenter study (2004–2007) enrolled preterm infants born at a gestational age of 35 weeks or less and originating from different cohorts monitored at the neonatal departments of Saint Vincent de Paul and Port Royal hospitals (Paris, France) and at the neonatal intensive care unit of the University of Bari (Italy). Infants staying less than 1 week in a neonatal unit were excluded. For each included neonate, fecal samples were collected weekly from the diaper, beginning at the end of the first week of life and continuing until the end of the first month; when any gastrointestinal events occurred, such as abdominal distension or gastrointestinal bleeding; with the presence of digestive symptoms leading to interruption in enteral feeding; or if NEC developed. NEC episodes were staged according to the Bell classification (6). Abdominal distension, gastrointestinal bleeding, or digestive symptoms leading to interruption of enteral feeding were considered symptoms of mild enteropathy; NEC Bell stages IIb and III were considered symptoms of severe enteropathy. Written informed consent was obtained from all of the parents before collection of fecal samples or recording of clinical parameters. Fecal samples were stored at −80°C, and calprotectin was measured by enzyme-linked immunosorbent assay (Calprest, Eurospital, Italy).
Quantitative variables are reported as median and range. Comparisons between different groups were performed by use of the Wilcoxon test or the Kruskal-Wallis test for continuous variables. The Pearson correlation coefficient was used to assess the relation between continuous variables. Receiver operating characteristic (ROC) curve–based analysis was used to assess optimal cutoff values for the calprotectin level. Because 2 levels of severity were defined, 2 cutoff values had to be determined. Two groups were thus successively compared: healthy neonates versus all those identified as having a gastrointestinal disorder and infants with NEC versus all other neonates (those with and without mild digestive symptoms). Multivariate linear and logistic models were used to evaluate the impact of birth weight and gestational age on the optimal cutoff values and on the sensitivity and specificity of the calculated cutoff, respectively. Independence in the data was questionable because some neonates contributed several datapoints; thus, a complementary analysis was performed in which only the first datapoint pertaining to each infant was used.
The study population included 126 infants (75 boys, 51 girls) born at a median gestational age of 33 weeks (range 25.7–35 weeks) with a median birth weight of 1760 g (range 730–2750 g). During the follow-up, 312 samples were taken. Of the group, 95 infants never experienced digestive symptoms (233 samples). A total of 24 neonates experienced mild enteropathy; 42 samples were taken during the symptomatic period in this group, and 19 samples were taken during the period without symptoms. Seven neonates experienced Bell stage IIb or III NEC episodes (18 samples; Fig. 1).
The median fecal calprotectin level was 226 μg/g (range 16–4775 μg/g). There was no significant correlation between calprotectin level and birth weight (r = 0.09, P = 0.30) or gestational age (r = 0.10, P = 0.25). By contrast, calprotectin levels were significantly lower (P < 0.0001) in neonates with a birth weight below 1800 g (median 200 μg/g, range 16–1240 μg/g) compared with neonates with a birth weight above 1800 g (median 279 μg/g, range 43–4775 μg/g). In healthy neonates, the median calprotectin level was 206 μg/g (range 16–1240 μg/g). In symptomatic infants, the medians were 393 μg/g (range 52–996 μg/g) and 832 μg/g (range 168–4775 μg/g) in infants with mild and severe enteropathy, respectively.
The ROC curves yielded a cutoff value of 363 μg/g (sensitivity 0.65, specificity 0.82) for the development of digestive symptoms and a cutoff of 636 μg/g (sensitivity 0.72, specificity 0.95) for the development of severe symptoms (Fig. 2). Neither gestational age nor birth weight significantly altered the cutoff values or the sensitivity and specificity of the calculated cutoff. Statistical analysis including only the first pertinent datapoint for each neonate confirmed these cutoffs as differences of less than 5% observed between the results corresponding to both analyses.
The results show that measurement of fecal calprotectin may provide evidence of variations related to the severity of intestinal distress in preterm neonates. They offer an attempt to determine potential cutoff levels for decision making. Intestinal complications in preterm infants, such as enteropathy or NEC, are not uncommon and represent a major concern for both the vital prognosis and the management of enteral feeding. Diagnosis of these complications is based on clinical and radiological signs, and current laboratory markers have poor specificity. This work forms the basis of using calprotectin levels as a potentially helpful marker.
This preliminary study confirms the previously identified high interindividual variability of fecal calprotectin in preterm infants. However, the cutoff values could be calculated according to the severity of intestinal disease, and gestational age and birth weight did not influence these values. Cutoffs (ie, >363 μg/g and >636 μg/g for mild and severe enteropathy, respectively) were much higher than those for adults and children older than 4 years of age (ie, 50 μg/g) (7,8) and gave the highest combined sensitivity and specificity for screening intestinal inflammation in neonates. More precisely, levels above 363 μg/g seem to have a sufficiently high specificity to justify suspicion of intestinal inflammation, thus rendering the prolongation of enteral feeding risky. Higher levels (>636 μg/g) correlated with a higher risk of severe distress, such as NEC, requiring specific monitoring. Lower levels led to higher sensitivity but much lower specificity: for example, with a cutoff of 200 μg/g, sensitivity was 0.83 and specificity was 0.57 for intestinal symptoms. This reduction could lead to the risk for an improper interruption in enteral feeding, although it is of crucial importance for intestinal maturation in preterm infants. Because of interindividual variability, a sequential measurement of calprotectin should also be considered as an alternative to using defined cutoff levels and deserves a prospective study.
Previous studies in healthy neonates have identified high calprotectin levels and have suggested increased intestinal permeability during the first weeks of life (3,9,10). These high basal calprotectin levels could also reflect a minimal secretion of antimicrobial peptides in response to alimentary allergens and to colonization of the gut by commensal microbiota to prevent enteric pathogens and pathogen–host cell interaction (11). Higher levels observed in pathological scenarios would thus reflect an increased intestinal inflammatory defense response with the release of molecules that act as antimicrobial agents.
Studies performed in preterm infants are scarce, especially under pathological conditions, and our results differ from those of other reports. In a first pilot study, Carroll et al (5) observed significantly higher calprotectin levels in patients with NEC than in cross-matched controls patients of similar age; however, calprotectin levels in the current investigation remained below our proposed cutoff level of 363 μg/g for mild intestinal events, even in the NEC population. These apparently conflicting results may be explained by a reduced sensitivity of the assay used in their study, in which calprotectin levels were notably lower in healthy infants than is usually observed in similar populations (2–4). For instance, the upper adult reference value in the study by Carroll et al (5) was 30 μg/g, rather than 50 μg/g. Josefsson et al (4), measuring calprotectin levels in very low birth weight infants, included 7 infants with severe abdominal diseases: 4 cases of NEC, 2 cases of focal intestinal perforation, and 1 case of intestinal obstruction. Because no reference infant had a level above 2000 μg/g, in contrast to 3 cases of NEC and 1 case of intestinal perforation, those authors proposed 2000 μg/g as a cutoff for severe abdominal diseases in this population. Of course, this level gives a specificity of 1, but the sensitivity is consequently low. Calculations based on our data indicate a sensitivity of 0.3, which would be of minor use for clinical purposes.
The interpretation of our preliminary study does carry some limitations, including the lower number of neonates with mild or major enteropathy and the use of several samples per patient. However, the current data suggest that measurement of calprotectin levels seems to be a promising noninvasive clinical screening test for intestinal distress in neonates.
1. Tibble J, Teahon K, Thjodleifsson B, et al
. A simple method for assessing intestinal inflammation in Crohn's disease. Gut 2000; 47:506–513.
2. Campeotto F, Kalach N, Lapillonne A, et al
. Time course of faecal calprotectin in preterm newborns during the first month of life. Acta Paediatr 2007; 96:1531–1533.
3. 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.
4. Josefsson S, Bunn SK, Domellof M. Fecal calprotectin in very low birth weight infants. J Pediatr Gastroenterol Nutr 2007; 44:407–413.
5. Carroll D, Corfield A, Spicer R, et al
. Faecal calprotectin concentrations and diagnosis of necrotising enterocolitis. Lancet 2003; 361:310–311.
6. Bell MJ, Ternberg JL, Feigin RD, et al
. Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging. Ann Surg 1978; 187:1–7.
7. Ton H, Brandsnes O, Dale S, et al
. Improved assay for fecal calprotectin. Clin Chim Acta 2000; 292:41–54.
8. Fagerberg UL, Loof L, Merzoug RD, et al
. Fecal calprotectin levels in healthy children studied with an improved assay. J Pediatr Gastroenterol Nutr 2003; 37:468–472.
9. 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.
10. van Elburg RM, Fetter WP, Bunkers CM, et al
. Intestinal permeability in relation to birth weight and gestational and postnatal age. Arch Dis Child Fetal Neonatal Ed 2003; 88:F52–F55.
11. Splichal I, Fagerhol MK, Trebichavsky I, et al
. The effect of intestinal colonization of germ-free pigs with Escherichia coli
on calprotectin levels in plasma, intestinal and bronchoalveolar lavages. Immunobiology 2005; 209:681–687.