Appendicitis is a common gastrointestinal inflammatory disease in the Western world1 with peak incidences among young children and adolescents. Until recently, it was considered to be an irreversible progressive disease provoked by intraluminal obstruction, eventually leading to a perforation. Novel insights question this progressive nature and rather support that 2 distinct types of appendicitis exist: simple (uncomplicated) and complex (complicated) appendicitis.2–4 These types differ in terms of epidemiologic, clinical, biochemical and radiologic variables.5–10 Moreover, safe and potentially effective results were demonstrated for a nonsurgical (initial antibiotics alone) treatment for simple appendicitis.11,12 Accurate preoperative discrimination is thus important for both implementations of different treatment strategies as their comparison in research. Unfortunately, none of the previous developed clinical prediction rules have shown 100% accuracy and new diagnostic strategies need to be explored 13,14. Recently, it has been reported that gut microbiota might play a role in the pathogenesis of a variety of gastrointestinal diseases [eg, inflammatory bowel disease (IBD)] and extra-intestinal (autoimmune) diseases. In this light, the microbiota of the appendix is worth mentioning, since its role in the pathogenesis of appendicitis has been previously considered.15 In addition, novel, culture-independent techniques for microbiota analysis have shown compositional differences between noninflamed and inflamed appendices.16–21 Nevertheless, the role of the microbiota in specifically each type is largely unknown. To contribute to the available literature, we aimed to elucidate the microbial composition of the appendix in children with simple and complex appendicitis, using a culture-independent DNA-based microbiota detection technique to detect differences and to identify specific disease-related bacteria.
MATERIALS AND METHODS
Design and Participants
This pilot cohort study was performed between November 2015 and November 2016 in 2 hospitals in the Netherlands (VU Medical Centre, Amsterdam and Red Cross Hospital, Beverwijk). The medical ethics committee of the VU Medical Centre confirmed that the medical research involving human subjects act (WMO) did not apply to this study. All children (0–17 years old), presenting at the Emergency Department for the suspicion of appendicitis underwent routine diagnostic work-up as described in our national guideline.22 In summary, in case of clinical suspicion of appendicitis, based upon medical history, physical examination and biochemical signs of inflammation (elevated white blood cell count and/or C-reactive protein), imaging studies (ultrasound and if required magnetic resonance imaging) were undertaken. In case appendicitis was highly suspected upon this (since confirmation is only possible on histopathologic examination), children were scheduled for appendectomy and eligible for inclusion. Prior to surgery, informed consent was obtained from parents and the child (if ≥12 years) for data collection from their clinical files and to acquire a part of the removed appendix for further analysis. We excluded children with the suspicion of an alternative diagnosis or other pathology (intraoperative or at histopathology) than appendicitis, such as a malignancy. All patients were included in a consecutive manner.
Surgical Procedure and Sample Collection
Appendectomy was performed according to local/national protocol. All children received prophylactic antibiotics ~30 minutes before incision: (1) metronidazole (<12 years: 7.5 mg/kg/dose iv with a maximum of 500 mg/dose, ≥12 years: 500mg/dose iv) with a cephalosporin (cefazolin 100–150 mg/kg/day in 3 doses or cefuroxime 50 mg/kg/dose with maximum of 1.5 g/dose) or (2) Amoxicillin-clavulanic acid (<40 kg: 100/10 mg/kg/day in 3–4 doses, >40 kg: 1000/100 to 2000/200 mg/dose iv) with gentamycin (7 mg/kg/day). Approach was either by the gridiron incision (open procedure) or by laparoscopy based upon the surgeon’s preference. Postoperative care (antibiotics) was according to the respective diagnosis and in line with national guidelines.22 After resection of the appendix, 0.5 cm of the tip of the appendix was immediately frozen at −20°C until further microbiota analysis. The remainder of the appendix was sent for routine histopathologic examination. Additional data collection of patient files was performed according to a predefined data form.
Clinical and Pathologic Classification
Only children with histopathologically confirmed appendicitis were included in final microbiota analysis. Samples of these children were classified as either simple or complex based upon predefined criteria (for both hospitals in the same manner) by 2 of the authors (R.B. and R.R.G.) independently. Complex appendicitis was defined as a gangrenous or perforated appendix found intraoperatively with or without purulent intra-abdominal fluid or abscess formation or; histopathologic signs of extensive necrosis, ulceration or perforation of the appendix. Simple appendicitis was defined as inflammation of the appendix, with transmural invasion of neutrophils, in the absence of all of the abovementioned findings of complexity. In case of disagreement, discussion was held until consensus was reached. The researchers performing the microbial cluster analysis were blinded for the classified type of appendicitis (S.M.L.T., A.E.B., and L.P.).
Sample Work-up and DNA Isolation
All frozen samples were thawed on ice and 0.3 × 0.3cm of the center of the sample, including all available layers, was taken for analysis. Remainders were frozen at −80 for future research. A routine molecular diagnostic protocol was used to isolate DNA.23 At first, 360 μl ATL buffer and 40 μl proteinase K were added to each sample. After vortexing and shaking, 400 μL of AL buffer was added. We used an easyMAG automated DNA isolation machine of Biomerieux for further DNA extraction. Samples were put in easyMAG containers and suspended in nucliSENS lysis buffers. After incubation silica beads were added. The specific “A” protocol of the machine was used and DNA was eluted in nuclisens easyMAG extraction buffer 3. After the full DNA isolation, it was stored at 2–8°C for PCR amplification. During every DNA extraction and PCR reaction, controls (empty samples) were used to evaluate reagent contamination for internal control.
IS-pro Profiling and Analysis
A PCR-based IS-pro assay was performed by IS-diagnostics, Amsterdam, The Netherlands.23 DNA was isolated and amplified with a GeneAmp PCR system 9700 and used for 2 multiplex PCR reactions. The first consisted of FAFV and Bacteroidetes. The second consisted of Proteobacteria. Subsequently, the product was mixed with IS-pro eMix (IS-Diagnostics) and fragment analysis was performed on an ABI Prism 3500 Genetic Analyzer. If the PCR analysis was inhibited, the DNA was diluted 1:10 or 1:100 and the analysis was rerun. All IS profiles contained 3 levels of information. First, the fluorescent labels or peak-colors sorted species into the 3 main gastrointestinal phyla (FAFV, Bacteroidetes and Proteobacteria). Second, the length of 16S-23S rDNA IS regions, expressed by the number of nucleotides, were used to identify individual IS fragments. These individual IS fragments reflect bacterial operational taxonomic units and were linked to a database to identify bacteria at species level. This database was based upon all publicly available bacterial whole genome sequences (approximately ±1.2 million) in 2018. Third, the intensity of the PCR product is displayed by peak heights, visualized as (relative) abundance and expressed in relative fluorescence units (RFU).
Microbiota and Statistical Analysis
The elementary analysis of this research was to determine if two or more types of appendicitis could be distinguished upon microbiota profiles. For this, a hierarchical cluster analysis was performed using the unweighted pair group method with arithmetic mean.24 In short, the aim of a cluster analysis is to statistically, and objectively, classify samples into several categories or clusters. Subsequently, the Fisher exact can be used to analyze the association of the identified clusters with type of appendicitis (SPSS software version 22). Level of significance was determined as P < 0.05. The visualization of acquired IS-profiles was performed with TIBCO Spotfire software. Complementary to the cluster analysis, a comparison between simple and complex appendicitis was made at phylum and species level, the highest and lowest taxonomic rank within the domain of bacteria, respectively; presence of each phylum, based upon the previously mentioned allocation of peak-colors, was reported descriptively (numbers and percentages); the Shannon Diversity index was used to calculate diversity (number of different species) for each phylum; (relative) abundance in RFU’s, intensity, of bacteria for each phylum was determined; presence of individual species was determined and reported descriptively. Mann-Whitney U test (diversity and abundance) and Fisher’s exact test (presence) were used when appropriate. Association of presence was only calculated in case species were present in at least 5 samples of the same group. Visualization was performed with Prism Graphpad software, version 7.
In total, 40 children who underwent an appendectomy for suspected appendicitis were included in this study. The diagnosis of appendicitis (nor any other diagnosis) could not histopathologically be confirmed in 4 children. Of the 36 children with confirmed appendicitis, immediate agreement of classification was reached in 29 (81%). After discussion, the remaining samples were, classified as simple in 2 cases and complex in 5. Main topic of debate was the assessment and extent of necrosis and ulceration of these samples. Baseline characteristics of both groups, simple (n = 16) and complex (n = 20), are presented in Table 1.
Cluster Analysis and Characteristics
Figure 1 displays the cluster analysis of the 36 included samples. It shows 2 distinct high-order clusters encompassing 34 of 36 (94%) appendicitis samples. Two samples were not allocated to one of the two clusters based upon their microbiota profiles (1 simple and 1 complex, respectively). Cluster 1 encompassed 15 and Cluster 2 19 patients. Both clusters were closely related to the blinded classification of appendicitis phenotype: Distribution of simple and complex appendicitis was 12/15 (80%) and 3/15 (20%) versus 3/19 (16%) and 16/19 (84%) patients for Clusters 1 and 2, respectively (P < 0.001).
Species within the phyla Bacteroidetes, FAFV and Proteobacteria were detected in all samples with complex appendicitis compared with 100%, 94% and 56% of the samples with simple appendicitis, respectively. In addition, complex appendicitis was associated with increased diversity of the phyla Bacteroidetes (P = 0.001) and Proteobacteria (P = 0.016), but not FAFV (P = 0.062) (Figure 2A). Moreover, a relative increased abundance, intensity (RFU), of all phyla was found in complex appendicitis: Bacteroidetes (P = 0.001), FAFV (P = 0.005) and Proteobacteria (P < 0.001) (Figure 2B).
In all samples combined, a total of 57 species could be identified within 31 genera (Supplemental Table 1, http://links.lww.com/INF/D571). Fifteen species were detected in 5 or more samples of the same group, with their relative abundance presented in Figure 3. Results of these 15 species in terms of presence and relative abundance are listed in Table 2. Of these species, 5 were significantly more often present in complex appendicitis: Alistipes finegoldii (6% vs. 50%, P = 0.009), Bacteroides fragilis (25% vs. 80%, P = 0.002), Escherichia coli (44% vs. 85%, P = 0.014), Parvimonas micra (6% vs. 45%, P = 0.022) and Sutterella spp (6% vs. 40%, P = 0.026).
To our knowledge, this study is the first to show differences in microbial composition in the appendix of children with simple and complex appendicitis (not only perforated), using a culture-independent DNA-based microbiota detection technique, IS-pro. We observed 2 microbial clusters, closely related to the intraoperative and histopathologic classification, strengthening the theory of 2 different phenotypes of appendicitis. In this study, complex appendicitis showed an increased diversity of species within the phyla Bacteroidetes and Proteobacteria and an increased intensity of the phyla Bacteroidetes, FAFV and Proteobacteria compared with simple appendicitis. More specifically, 5 different species (A. finegoldii, B. fragilis, E. coli, P. micra and Sutterella spp) were found more often present in the appendix of children with complex appendicitis. Results from this study are of importance as they form the basis of future studies, investigating whether or not these found microbial differences reflect underlying different pathogeneses and if they can be useful in accurate preoperative (noninvasive) discrimination.
Microbiota in Children With Appendicitis
In 2000, Carr15 stated that multiple factors could cause appendicitis, eventually resulting in the invasion of the appendix by intraluminal bacteria. Preceding studies investigating intraluminal bacteria during appendicitis used conventional in vitro culture techniques. Most common findings were the presence of E. coli, Bacteroides spp or B. fragilis, Pseudomonas aeruginosa and P. micra in the inflamed appendix.25–28 Relevance of this conventional, culture-dependent technique lies in the determination of antibiotic resistance and thus proper antibiotic choice. Unfortunately, the majority of gastrointestinal microbial species is not cultivable. In our opinion, this technique is therefore not suitable to describe the complex microbial composition in the appendix. Luckily, possibilities expanded with the emergence of DNA-based detection techniques, like PCR, providing a culture-independent analysis and several studies have been published using these techniques in the field of appendicitis.16–21
Of interest in these studies is the reported increased abundance of Fusobacteria, especially of Fusobacterium nucleatum (an oral pathogen) in appendicitis, since its increase was also linked to severity.16–19 We found Fusobacteria in about a third of all patients in our study, and an increased trend was found in complex appendicitis (25% simple vs. 50% complex), but the difference was not statistically significant and we could therefore not confirm its association with complex appendicitis. Although Fusobacteria was most often mentioned, it was not the only found oral pathogen. Consistent with previous literature, we reported an increased presence of P. micra in complex appendicitis.18,19 Taking abovementioned findings together, it may indicate a potential role for oral pathogens in the pathogenesis of particularly complex appendicitis.
Both B. fragilis and E. coli were commonly found with conventional culture techniques in patients with appendicitis. While most studies reported an increase of E.coli in appendicitis, and especially complex appendicitis, the findings of Bacteroides and B. fragilis in appendicitis are inconsistent. Our study supports the association of an increased presence of B. fragilis with complex appendicitis, but the interpretation of this finding is challenging. A review into the beneficial effects of B. fragilis noted that the species is considered commensal in the healthy gut with immune-modulatory functions.29 While on the other hand, the specific strain of Enterotoxigenic B. fragilis (ETBF) is associated with IBD, especially in active disease.30 Could these specific Enterotoxigenic B. fragilis be present in complex appendicitis?
The presence of Sutterella spp in predominantly complex appendicitis is to our knowledge not previously described. The finding is notable since it has been suggested that Sutterella spp are able to interact via intra epithelial cells that are capable of directing antigen-presenting cells in the lamina propria into a pro-inflammatory, Th17, response.31–34 Although, the pro-inflammatory functions of Sutterella spp were found to be mild, it is specifically this pro-inflammatory Th17 response that is also mentioned in association with complex appendicitis.9,10
Bacterial Diversity in Appendicitis
Bacterial diversity remains a common topic in microbiota research and in inflammatory diseases, such as IBD, often a decreased diversity is found.35 On the other hand, an increased diversity, as found in complex appendicitis, was demonstrated in patients with Clostridium difficile and patients with diverticulitis.36,37 The consequence or role of this increased bacterial diversity is however still unknown. It must be noted that an increased bacterial diversity in complex appendicitis, or decreased diversity in simple appendicitis may also reflected in the calculated mean relative abundance of all species in Table 2.
Strengths and Limitations
As opposed to earlier research, we used a culture-independent DNA-based microbiota detection technique, the IS-pro technique, which is a nonselective detection method, able to unravel the highly complex intestinal microbiota composition.23 In a recent validation study, comparable results of intestinal microbiota characterization were shown for IS-pro and 454-pyrosequencing.38 An advantage of the IS-pro technique is the relative fast determination of bacteria (several hours compared with days of culture techniques) with potential benefits for clinical implementation. In addition of this novel technique, a nonsupervised cluster analysis was performed, to objectively determine, instead of subjectively classify, if 2 phenotypes of appendicitis exist. This method was previously and successfully demonstrated in diverticulitis and necrotizing enterocolitis (A.E.B., T.G.J.d.M.]. Moreover, since bacterial infiltration, especially of Fusobacterium, was demonstrated in the submucosa of samples of appendicitis, we chose to use appendiceal tissue, instead of swabs of the lumen for analysis (Swidsinksi, 2011).
Firstly, a limitation of this study could be the use of the tip of the appendix since it was therefore not available for histopathologic confirmation of appendicitis. Since inflammation of the appendix tends to start in the tip, it might have led to sampling bias. Secondly, we could not investigate previously described difference between the proximal and distal (tip) parts, since usage of the base of the appendix was reviewed as nonethical (histopathologic assessment is essential for detection and indicated treatment of carcinoid of the appendix).21 Thirdly, a limitation of this study is the relatively small sample size, which is also reflected in the comparison between species after Bonferroni correction for multiple comparisons. Results of species analysis should therefore be interpreted with caution. Fourthly, in this study, we could not answer a potential correlation between the presence of a fecalith and the microbiome in appendicitis. It is however of interest for future research. At last a limitation of this study is the use of prophylactic antibiotics. However, the beneficial effects of antibiotics have been widely accepted and since they were only administered a short time before operation we believe that this does not explain the found differences.
Results of this study will lead to future studies from our group. Firstly, we hypothesize that the microbial differences are not only present in the appendix itself, but will also be reflected in stool, since previous studies already describe differences in rectal swabs of children with and without appendicitis.18 Potential future findings may provide us with a novel noninvasive diagnostic modality, such as oral and rectal swabs and analysis of volatile organic compound (VOC). The latter has shown promising results discriminating between children with and without necrotizing enterocolitis, based upon fecal VOC profiling by an electronic nose device.39 Secondly, we will further investigate the role of specific bacteria in the pathogenesis of simple and complex appendicitis. It must be noted that to date we cannot certainly assess whether the reported findings of this study, or previous studies, are cause or consequence. It would therefore be of great interest to investigate the duration of abdominal pain, as well as other factor such as dietary influences or days of fasting, on the microbiome. We would suggest either a larger study into the topic with stratification of days of abdominal pain or an animal model on appendicitis. Thirdly, it would be of interest to investigate whether or not the reported bacteria in children with complex appendicitis are able to initiate the earlier mentioned Th17-response.
In conclusion, we demonstrate the presence of 2 clusters, based upon the microbial composition of the inflamed appendix, that are closely related to simple and complex appendicitis in the pediatric population, respectively. Complex appendicitis was associated with an increased diversity and intensity, different relative abundance of bacteria and with the increased presence of 5 species compared with simple appendicitis in the appendix itself. Whether these findings and specific bacteria play an etiologic role in either simple or complex appendicitis or are rather a consequence of an inflamed state remains still unclear. Further studies should focus on the exact role of these bacteria and in their potential for use in preoperative discrimination between simple and complex appendicitis.
1. Addiss DG, Shaffer N, Fowler BS, et al. The epidemiology of appendicitis and appendectomy in the United States. Am J Epidemiol. 1990;132:910–925.
2. Andersson RE. The natural history and traditional management of appendicitis revisited: spontaneous resolution and predominance of prehospital perforations imply that a correct diagnosis is more important than an early diagnosis. World J Surg. 2007;31:86–92.
3. Livingston EH, Woodward WA, Sarosi GA, et al. Disconnect between incidence of nonperforated and perforated appendicitis: implications for pathophysiology and management. Ann Surg. 2007;245:886–892.
4. Cobben LP, de Van Otterloo AM, Puylaert JB. Spontaneously resolving appendicitis: frequency and natural history in 60 patients. Radiology. 2000;215:349–352.
5. Bröker ME, van Lieshout EM, van der Elst M, et al. Discriminating between simple and perforated appendicitis. J Surg Res. 2012;176:79–83.
6. Williams RF, Blakely ML, Fischer PE, et al. Diagnosing ruptured appendicitis preoperatively in pediatric patients. J Am Coll Surg. 2009;208:819–25; discussion 826.
7. Peng YS, Lee HC, Yeung CY, et al. Clinical criteria for diagnosing perforated appendix in pediatric patients. Pediatr Emerg Care. 2006;22:475–479.
8. Oliak D, Yamini D, Udani VM, et al. Can perforated appendicitis Be diagnosed preoperatively based on admission factors? J Gastrointest Surg. 2000;4:470–474.
9. Rubér M, Berg A, Ekerfelt C, et al. Different cytokine profiles in patients with a history of gangrenous or phlegmonous appendicitis. Clin Exp Immunol. 2006;143:117–124.
10. Rubér M, Andersson M, Petersson BF, et al. Systemic Th17-like cytokine pattern in gangrenous appendicitis but not in phlegmonous appendicitis. Surgery. 2010;147:366–372.
11. Gorter RR, The SML, Gorter-Stam MAW, et al. Systematic review of nonoperative versus operative treatment of uncomplicated appendicitis. J Pediatr Surg. 2017;52:1219–1227.
12. Salminen P, Tuominen R, Paajanen H, et al. Five-Year follow-up of antibiotic therapy for uncomplicated acute appendicitis in the APPAC randomized clinical trial. JAMA. 2018;320:1259–1265.
13. Gorter RR, van den Boom AL, Heij HA, et al. A scoring system to predict the severity of appendicitis in children. J Surg Res. 2016;200:452–459.
14. Atema JJ, van Rossem CC, Leeuwenburgh MM, et al. Scoring system to distinguish uncomplicated from complicated acute appendicitis. Br J Surg. 2015;102:979–990.
15. Carr NJ. The pathology of acute appendicitis. Ann Diagn Pathol. 2000;4:46–58.
16. Swidsinski A, Dörffel Y, Loening-Baucke V, et al. Acute appendicitis is characterised by local invasion with Fusobacterium nucleatum
/necrophorum. Gut. 2011;60:34–40.
17. Rogers MB, Brower-Sinning R, Firek B, et al. Acute appendicitis in children is associated with a local expansion of fusobacteria
. Clin Infect Dis. 2016;63:71–78.
18. Jackson HT, Mongodin EF, Davenport KP, et al. Culture-independent evaluation of the appendix and rectum microbiomes in children with and without appendicitis. PLoS One. 2014;9:e95414.
19. Zhong D, Brower-Sinning R, Firek B, et al. Acute appendicitis in children is associated with an abundance of bacteria from the phylum fusobacteria
. J Pediatr Surg. 2014;49:441–446.
20. Guinane CM, Tadrous A, Fouhy F, et al. Microbial composition of human appendices from patients following appendectomy. MBio. 2013;4:e00366–12.
21. Salö M, Marungruang N, Roth B, et al. Evaluation of the microbiome in children’s appendicitis. Int J Colorectal Dis. 2017;32:19–28.
22. Bakker OJ, Go PM, Puylaert JB, et al. Richtlijn voor diagnostiek en behandeling van acute appendicitis [Guideline on diagnosis and treatment of acute appendicitis]. Beeldvorming voor appendectomie aanbevolen. Ned Tijdschr Geneeskd. 2010;154:A303
23. Budding AE, Grasman ME, Lin F, et al. IS-pro: high-throughput molecular fingerprinting of the intestinal microbiota
. FASEB J. 2010;24:4556–4564.
24. Michener CD, Sokal RR. A quantitative approach to a problem in classification. Evolution (N Y). 1957;11:130–162.
25. Obinwa O, Casidy M, Flynn J. The microbiology of bacterial peritonitis due to appendicitis in children. Ir J Med Sci. 2014;183:585–591.
26. Brook I. Bacterial studies of peritoneal cavity and postoperative surgical wound drainage following perforated appendix in children. Ann Surg. 1980;192:208–212.
27. Chan KW, Lee KH, Mou JW, et al. Evidence-based adjustment of antibiotic in pediatric complicated appendicitis in the era of antibiotic resistance. Pediatr Surg Int. 2010;26:157–160.
28. Guillet-Caruba C, Cheikhelard A, Guillet M, et al. Bacteriologic epidemiology and empirical treatment of pediatric complicated appendicitis. Diagn Microbiol Infect Dis. 2011;69:376–381.
29. Troy EB, Kasper DL. Beneficial effects of Bacteroides fragilis polysaccharides on the immune system. Front Biosci (Landmark Ed. 2010;15:25–34.
30. Prindiville TP, Sheikh RA, Cohen SH, et al. Bacteroides fragilis enterotoxin gene sequences in patients with inflammatory bowel disease. Emerg Infect Dis. 2000;6:171–174.
31. Hiippala K, Kainulainen V, Kalliomäki M, et al. Mucosal Prevalence and Interactions with the Epithelium Indicate Commensalism of Sutterella spp. Front Microbiol. 2016;7:1706.
32. Gaboriau-Routhiau V, Rakotobe S, Lécuyer E, et al. The key role of segmented filamentous bacteria in the coordinated maturation of gut helper T cell responses. Immunity. 2009;31:677–689.
33. Ivanov II, Honda K. Intestinal commensal microbes as immune modulators. Cell Host Microbe. 2012;12:496–508.
34. Atarashi K, Tanoue T, Ando M, et al. Th17 cell induction by adhesion of microbes to intestinal epithelial cells. Cell. 2015;163:367–380.
35. Ott SJ, Musfeldt M, Wenderoth DF, et al. Reduction in diversity of the colonic mucosa associated bacterial microflora in patients with active inflammatory bowel disease. Gut. 2004;53:685–693.
36. Chang JY, Antonopoulos DA, Kalra A, et al. Decreased diversity of the fecal microbiome in recurrent Clostridium difficile
-associated diarrhea. J Infect Dis. 2008;197:435–438.
37. Daniels L, Budding AE, de Korte N, et al. Fecal microbiome analysis as a diagnostic test for diverticulitis. Eur J Clin Microbiol Infect Dis. 2014;33:1927–1936.
38. de Meij TG, Budding AE, de Groot EF, et al. Composition and stability of intestinal microbiota
of healthy children within a Dutch population. FASEB J. 2016;30:1512–1522.
39. De Meij TGJ, Van Der Schee MPC, Berkhout DJC, et al. Early detection of necrotizing enterocolitis by fecal volatile organic compounds analysis. J Pediatr. 2015;167:562–567.e1.