Reply

Abstract

Listen

Translate article

To the Editor: We thank Savino and coworkers their interest in our article. They note that many studies have shown that fecal calprotectin (FC) levels have a nonparametric distribution and should be evaluated with appropriate statistical tools. We agree; indeed, in our report we did not explain that a preliminary evaluation of the distribution of the variables had been performed. We observed that even if FC values were not normally distributed in the total sample studied, we opted to perform classic parametric procedures for the following reasons:1.When the distribution of FC values was examined separately at enrollment and 4 weeks after enrollment, we found that values did have a normal distribution at enrollment and quite close to a normal distribution 4 weeks later.2.We tested a variety of traditional transformations, none of which improved normality.3.To make the decision between parametric and nonparametric procedures, we referred to Conover: “The recommended procedure in experimental designs for which no nonparametric test exist is to use the usual analysis of variance on the data and then to use the same procedure on the rank transformed data. If the two procedures give nearly identical results, the assumptions underlying the usual analysis of variance are likely to be reasonable and the regular parametric analysis valid”.1Conover WJ. In: Practical Nonparametric Statistics. 3rd ed. New York: Wiley, 199X. p. 419.Google Scholar4.The parametric procedures used are well known to be quite robust to nonnormality, even for quite small samples.2Box G.E.P. Hunter J.S. Hunter W.G. Statistics for Experimenters: Design, Innovation, and Discovery.2nd ed. Wiley, New York2005Google Scholar Savino et al imply that the elevated FC levels in the samples from our patients could be the result of blood in the stools, not from inflammation, and reference a study by Campeotto et al3Campeotto F. Kalach N. Lapillonne A. Butel M.J. Dupont C. Kapel N. Time course of faecal calprotectin in preterm newborns during the first month of life.Acta Paediatr. 2007; 96: 1531-1535Crossref PubMed Scopus (34) Google Scholar to support this. We find it difficult to compare FC results from preterm infants at 27-34 weeks gestation with our results from term infants age 3-5 months.4Baldassarre ME, Laforgia N, Fanelli M, Laneve A, Grosso R, Lifschitz C. Lactobacillus GG improves recovery in infants with blood in the stools and presumptive allergic colitis compared with extensively hydrolyzed formula alone. J Pediatr, in pressGoogle Scholar Nonetheless, if anything, the mean values and range of FC in the study of Campeotto et al were greater in patients with diarrhea than in those with blood in the stool (645 g/g [334-928 g/g] vs 471 g/g [177-604 g/g]). In our study, FC level was used as a marker to evaluate subclinical intestinal inflammation once macroscopic blood was no longer present, which allowed us to identify any differences in recovery between the 2 dietary regimens evaluated, extensively hydrolyzed protein formula with and without the addition of lactobacillus GG. If baseline values were artificially higher because of the presence of blood, it should have affected stools from both dietary groups equally and thus had no impact on results. Savino et al state that we did not clarify whether patients had anal fissures or gastroesophageal reflux. None of the patients studied had anal fissures, and we are not sure why gastroesophageal reflux had to be ruled out in infants whose only symptoms were blood and mucus in the stool. Finally, Savino et al state that they found higher FC levels in breast-fed infants compared with formula-fed infants. That seems to be in contrast with the findings of Campeotto et al,3Campeotto F. Kalach N. Lapillonne A. Butel M.J. Dupont C. Kapel N. Time course of faecal calprotectin in preterm newborns during the first month of life.Acta Paediatr. 2007; 96: 1531-1535Crossref PubMed Scopus (34) Google Scholar who reported no differences related to type of feeding. Fecal calprotectin in infants with presumptive allergic colitisThe Journal of PediatricsVol. 157Issue 1PreviewTo the Editor: Full-Text PDF