Early Human Development 90 (2014) 579–585
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
Gut microbiota in preterm infants with gross blood in stools: A prospective, controlled study Mohamed Ben Said a,b, Stephane Hays a,c, Deplhine Maucort-Boulch d,e,f, Abdallah Oulmaati a, Stefanyia Hantova g, Claire-Marie Loys a, Estelle Jumas-Bilak g,h, Jean-Charles Picaud a,c,f,⁎ a
Department of Neonatology, University Hospital Croix Rousse, Hospices Civils de Lyon, F-69004 Lyon, France Faculty of Medicine of Tunis, University of Tunis El-Manar, 1007 Bab Saâdoun, Tunis, Tunisia Rhone-Alpes Human Nutrition Research Center, F-69310 Pierre Bénite, France d Department of Biostatistics, Hospices Civils de Lyon, F-69003 Lyon, France e CNRS, Laboratoire Biostatistique Santé, UMR 5558, F-69310 Pierre Bénite, France f Lyon-Sud Charles Merieux Medical School, Claude Bernard University Lyon 1, F-69310 Pierre Bénite, France g University Montpellier 1, Laboratoire de Bacteriologie-Virologie, EA 3755 UM1, Faculte de Pharmacie, F-34093 Montpellier, France h Department of Hospital Hygiene, CHU de Montpellier, F-34000 Montpellier, France b c
a r t i c l e
i n f o
Article history: Received 21 March 2014 Received in revised form 6 July 2014 Accepted 8 July 2014 Available online xxxx Keywords: Prematurity Gut microbiota Staphylococcus Infection Rectal bleeding Human milk
a b s t r a c t Objective: Gross blood in stools is a peculiar entity in preterm infants, but little is known about its etiology. As gut microbiota can be distorted in preterm infants, we aimed to evaluate the gut microbiota in infants with gross blood in stools. Study design: In a prospective, controlled, single-center study, we enrolled all infants born before 34 weeks of gestational age presenting gross blood in stools that was either completely isolated or associated with mild clinical symptoms or radiological signs. Each case was paired with two controls who were hospitalized in the same unit and were matched for gestational age and birth weight. The diversity of the gut microbiota was analyzed using 16S rRNA gene PCR and temporal temperature gel electrophoresis. We calculated a diversity score corresponding to the number of operational taxonomic units present in the microbiota. Results: Thirty-three preterm infants with gross blood in stools were matched with 57 controls. Clinical characteristics were similar in cases and controls. There was no statistically significant difference in the diversity score between the two groups, but microbiota composition differed. The proportion of infants with Escherichia coli was significantly higher in cases than in controls (p = 0.045) and the opposite pattern occurred for Staphylococcus sp. (p = 0.047). Conclusion: Dysbiosis could be a risk factor for gross blood in stools in preterm infants. Additional, larger studies are needed to confirm the implications of the presence of different genotypes of E. coli and to evaluate preventive actions such as the prophylactic use of probiotics and/or prebiotics. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction In premature infants gross blood in stools may be related to severe diseases (necrotizing enterocolitis (NEC), colitis in Hirschsprung's disease, infectious colitis, or hemorrhagic disease of the newborn) or to milder diseases (allergic colitis, anal fissure, or swallowing blood syndrome) [1,2]. Apart from these conditions, gross blood in stools is a peculiar entity either completely isolated or is associated with mild clinical symptoms or radiological signs (gastric residuals, vomiting,
Abbreviations: NEC, necrotizing enterocolitis; OTUs, operational taxonomic units; TTGE, temporal temperature gel electrophoresis. ⁎ Corresponding author at: Service de Néonatologie, Hôpital de la Croix Rousse, 103 Grande rue de la Croix Rousse, 69004 Lyon, France. Tel.: +33 472 001 550. E-mail address: [email protected] (J.-C. Picaud).
http://dx.doi.org/10.1016/j.earlhumdev.2014.07.004 0378-3782/© 2014 Elsevier Ltd. All rights reserved.
mild abdominal distension, and mild radiological intestinal dilation). Isolated rectal bleeding has been related to ecchymotic colitis [3]. Occurrence of gross blood in stools has a significant impact on neonatal care, as the management of this disorder often requires a fasting period to start or to extend parenteral nutrition, which increases the risk of catheter-related sepsis [4]. However, published data on gross blood in stools in preterm infants are scarce. While the predisposing factors of severe NEC (≥stage II) are wellknown, including colonization by potentially pathogenic bacteria or gut dysbiosis [5–8], it is not the case for gross blood in stools. MaayanMetzger et al. reported that a feeding regimen that did not include breast milk was the only variable that predicted isolated rectal bleeding and emphasized the benign nature of this bleeding [9]. Luoto et al. did not identify any risk factors in a small number of very-low-birth weight infants fed human milk supplemented with probiotics [10]. Luoto et al.
580
M.B. Said et al. / Early Human Development 90 (2014) 579–585
did not detect significant changes in the composition of the gut microflora in preterm infants with gross blood in stools, but they focused their investigation on a few main bacterial genera and species [10]. In the present investigation, we evaluated the relationship between the composition of the intestinal microbiota and gross blood in stools in preterm infants. 2. Methods 2.1. Study design and population We performed a prospective, controlled, single-center study. All preterm infants hospitalized between January and October 2011 in our tertiary care unit at the University Hospital Croix Rousse, Lyon, France were eligible for study enrolment if they met the inclusion criteria of birth gestational age ≤34 weeks, absence of congenital malformation, and gross blood in stools. The nurses in charge of daily care for cases and controls collected systematically few times per day information about digestive tolerance as routinely recommended in our unit: gastric residuals, vomiting, abdominal distension, stool consistency and gross blood in stools. Diagnosis was confirmed by physicians by clinical observation. Gross blood in stools was either completely isolated or associated with mild clinical symptoms or radiological signs (gastric residuals, vomiting, mild abdominal distension, and mild radiological intestinal dilation). We did not include infants who developed NEC ≥stage II according to the modified Bell classification [11] or exhibited spontaneous intestinal perforation. The study protocol was approved by the ethics committee of Lyon (Comité de protection des personnes Sud Est IV Lyon). Each case was matched with two controls. Controls were the first two preterm infants hospitalized in the same unit during the same period, and whose birth weight and gestational age were similar (±100 g and ±1 week, respectively) to those of the cases. These infants were selected by research nurses among hospitalized infants fulfilling the above criteria, independently from clinicians who took care of these infants. Infants included as controls were monitored as routinely recommended in our unit: daily monitoring of gastric residuals, vomiting, abdominal distension, stool consistency and blood in stools. When infants included as controls later presented rectal bleeding they were excluded from the control group and not replaced by another one, then included in the group of cases and paired with two controls. 2.2. Routine care protocol The feeding regimen of the very-low-birth weight infants was not changed during the study period. Cases and controls were fed according to the same NICU protocol. Enteral feeding was started at day 1 or 2 and complementary parenteral feeding was administered until the enteral intake reached 100 mL/kg/day. All infants were fed pasteurized human milk (their own mother's milk or donor milk) according to French regulations [12] until their body weight was ~1500 g. Then, if the mother had no milk, feeding with the same preterm formula (PreMilumel, Milumel, Torce, France) was commenced. Human milk was supplemented with a cow's milk protein based multicomponent fortifier (Eoprotine, Milumel, Torce France). None of the infants included in our study received probiotics. Full enteral feeding was 160 mL/kg/day and was achieved by increasing the feeding by 10–20 mL/kg/day depending on the feeding tolerance of the infant. According to the NICU protocol enteral feeding was started as continuous feeding followed by bolus feeding when full enteral feeding was reached and digestive tolerance was satisfactory. After full enteral feeding was achieved, a thickener was used when infants presented signs of gastro-esophageal reflux (carob-based when the infant was fed human milk and starch-based when the infant was fed preterm formula). If the signs of gastro-esophageal reflux persisted, we administered domperidone with or without a proton pump inhibitor (omeprazole or esomeprazole) depending on the
severity of the signs. Infants who presented clinical and ultrasound signs of persistent ductus arteriosus were treated with ibuprofen (10, 5, and 5 mg/kg/day at days 1, 2, and 3, respectively). 2.3. Management of gross blood in stools When gross blood in stools occurred, we systematically assessed blood count cells, serum C-reactive protein levels, serum procalcitonin levels, and blood culture (aerobic and anaerobic). Stool samples were collected for bacteriological culture and for the identification of rotavirus and adenovirus. Abdominal x-rays were also performed. Then, infants received no enteral feeding for a few days, depending on the clinical and radiographic data collected at the time of rectal bleeding and in the following days. For each case with gross blood in stools, stool samples were collected from the two matched control infants. 2.4. Data collection We collected data on the pregnancies, the deliveries (antenatal steroids, mode of delivery, maternal diseases), the infants' characteristics at birth (gestational age, birth weight, gender, Apgar score), growth restriction (body weight less than −2SD for gestational age) [13] treatments during hospitalization before gross blood in stools (ventilation, antibiotics, ibuprofen treatment, postnatal steroids, anti-reflux treatment, breast milk or formula), treatments at the time of gross blood in stools (body weight, clinical examination, x-ray results, C-reactive protein levels, procalcitonin levels, complete blood count, blood culture), and information about the infant after gross blood in stools (fasting period, recurrence of rectal bleeding). 2.5. Analysis of gut microbiota The qualitative composition of the gut microbiota was assessed via extraction of bacterial DNA from stool samples, PCR, temporal temperature gradient electrophoresis (TTGE), and identification of amplified sequences. Stool samples collected from cases and controls were stored at −80 °C until analysis. Approximately 50 mg of each homogenized stool specimen was placed in 1 mL of sterile DNA-free water in a 1.5-mL tube. The suspension was centrifuged for 10 min at 10,000 ×g. DNA was extracted from the pellet using the MasterPure Gram Positive DNA Purification Kit (Epicentre, Madison, WI, USA) according to the supplier's instructions and optimized by Roudière et al. [14]. The V2–V3 region (233 bp) of the 16S rRNA gene was amplified with primers HDA1 (with a GC-clamp) and HDA2 [14]. The reaction mixture (50 μL) contained 200 μM of deoxynucleoside triphosphate mix, 10 pmol of each primer, 2.5 U of Taq DNA polymerase in the appropriate buffer (FastStart High Fidelity PCR system, Roche, Basel, Switzerland), and 1 μL of template DNA. The amplification program, which was carried out with a Mastercycler apparatus (Eppendorf, Le Pecq, France), was 95 °C for 2 min, 30 cycles of 95 °C for 1 min, 62 °C for 30 s, and 72 °C for 1 min, with a final extension of 72 °C for 7 min. PCRs were checked by electrophoresis migration on a 1.5% agarose gel stained with ethidium bromide (500 μg/mL) and visualized on an ultraviolet transillumination system. The DCode universal mutation detection system (Bio-Rad Laboratories, Marne La Coquette, France) was used for TTGE. Five microliters of the PCR product was added to 5 μL of loading buffer. The gels were prepared with 8% (wt/vol) bisacrylamide (37.5:1) and 7 M urea and were run with 1X Tris–acetate–EDTA buffer. Denaturating electrophoresis was performed at 46 V for 16 h with a temperature gradient of 63–70 °C (0.4 °C/h). Gels were stained with 10 mg/mL ethidium bromide and photographed with an ultraviolet transillumination system. Identification of the TTGE bands was performed as described previously, via comparison of migration distances to the “gut microbiota diversity ladder” and sequencing on an ABI 3730XL sequencer (Takeley, United Kingdom) [14,15]. The resulting DNA sequences were compared
M.B. Said et al. / Early Human Development 90 (2014) 579–585
with Genbank (http://www.ncbi.nlm.nih.gov/) and with the Ribosomal Database Project II (http://rdp.cme.msu.edu/) using the Basic Local Alignment Search Tool (BLAST) and Seqmatch, respectively. Each sequence was affiliated with an operational taxonomic unit (OTU) that may be defined first at the species level, and otherwise at the genus level on the basis of the percent of sequence identity with the closest relative in the sequence databases. Each OTU was further classified into groups of taxonomic or clinical relevance (genus or family). The diversity score corresponded to the number of different OTUs in a sample as previously validated [14,15]. Complementary TTGE experiments were performed to obtain species level affiliation, for enterobacteria and staphylococci. Tuf-PCRTTGE [16] and rpoB-PCR-TTGE [17] were performed as described in order to determine the species in the genus Staphylococcus and in the family Enterobacteriaceae, respectively. 2.6. Statistical analysis Main variables were described with numbers (percentage) for categorical variables and medians [minimum, maximum] for continuous variables. We used the chi-squared test or Fisher's exact test to compare qualitative variables and the Wilcoxon test to compare quantitative variables. p-Values less than 0.05 were considered significant. We used SPSS version 16.0 (SPSS Inc., Chicago, IL, USA) to perform the statistical analyses.
581
Table 1 Clinical characteristics of 33 cases before gross blood in stools and 57 controls before inclusion. Cases n = 33 Pregnancy Hypertension Gestational diabetes Transfusion syndrome Cesarean section Intrapartum antibiotics
07 (21) 0 (0) 2 (6) 28 (85) 11 (33)
Controls n = 57 9 (16) 2 (3) 3 (5) 46 (81) 27 (47)
p-Value
0.517 0.530 1.000 0.620 0.194
At birth Gestational age in weeks Birth weight in g Birth weight ≤ 2 SDs Male gender
30 [25, 34] 1200 [635, 2160] 1 (3) 11 (33)
30 [23, 34] 1240 [570, 2280] 2 (3) 28 (49)
0.694 0.714 1.000 0.145
Drugs before rectal bleeding Domperidone Proton pump inhibitor Ibuprofen Postnatal steroids Postnatal antibiotics Duration of postnatal antibiotics
3 (9) 1 (3) 6 (18) 2 (6) 18 (54) 2 [2, 10]
0 (0) 1 (2) 9 (16) 1 (2) 30 (53) 2 [2, 3]
0.046 1.000 0.769 0.552 1.000 0.489
0 [0, 1512] 190 [0, 1655] 20 [0, 1500]
0.530 0.933 0.443
Respiratory support before rectal bleeding Assisted ventilation, in hours 0 [0, 1479] CPAP, in hours 218 [0, 1600] Oxygen therapy, in hours 3 [0, 1291]
Values are given as number (percentage) or median [minimum, maximum]. SD, standard deviation; CPAP, continuous positive airway pressure.
3. Results Of the 270 preterm infants with gestational age less than 34 weeks that were hospitalized in our neonatal intensive care unit between January and October 2011, 33 (12.2%) presented gross blood in stools. We matched 57 controls to these 33 cases (Fig. 1). Approximately two out three cases had isolated rectal bleeding (n = 22). Nine infants initially selected as controls presented a gross blood in stools.
270 preterm infants ≤34 wks
Rectal bleeding (n=26)
No rectal bleeding (n=244)
Not included for : . Spontaneous intestinal perforation (n=1)
The clinical characteristics of the two groups before and at inclusion were comparable, with the exceptions of treatment with domperidone and use of thickener. There was no significant difference in antibiotic exposure between the two groups (Tables 1 and 2). The proportion of singletons was similar in both groups (57% of cases vs. 47% of controls, p = 0.351). Most infants received antenatal steroids (91% of cases vs. 96% of controls, p = 0.352). The bacterial diversity scores observed in the overall study population were low, ranging from 1 to 10, but the scores varied between patients. There was no significant difference in the diversity score between cases and controls (3 [1, 9] vs. 3 [1, 10] OTUs, respectively; p = 0.654). A total of 39 OTUs were detected in the populations studied. These OTUs were distributed through four phyla, with a predominance of OTUs in Firmicutes and Proteobacteria (Fig. 2), and 10 major genera or families, with a predominance of OTUs in Staphylococcus, Enterobacteriaceae and Clostridium (Fig. 3). The distributions of phyla (Fig. 2) and most-represented genera or families (Fig. 3) did not significantly differ between cases and controls. However, enterobacteria occurred more frequently in cases but the difference was not statistically significant (Fig. 3). The proportion of infants harboring Escherichia coli was
. NEC grade 2 (n=1) Gross blood in stools (n=24)
Table 2 Clinical characteristics of preterm infants at the time of gross blood in stools (cases) and at inclusion (controls).
Controls (n=48)
Cases n = 33
Gross blood in stools (n=9)
Gross blood in stools (n=33)
Controls n=18
Controls (n=57)
Fig. 1. Flow chart of study design.
Postmenstrual age, in weeks Body weight, in g
33 [27, 46] 1752 [880, 3454] Postnatal age, in days 21.5 [3, 102] Postnatal growth restriction 3 (9.4) Enteral feeding at time of rectal bleeding 18 (54) Human milka Preterm formula 11 (33) 4 (12) Human milka + preterm formula Thickener 10 (30)
Controls n = 57
p-Value
33 [29, 38] 1670 [940, 2900] 17 [4, 75] 1 (1.8)
0.427 0.110 0.147 0.131
31 (54) 16 (28) 10 (17) 6 (10)
0.988 0.600 0.494 0.018
Values are presented as number (percentage) or median [minimum, maximum]. a Human milk fortified with a multicomponent fortifier (Eoprotine, Milumel, Torce France: 4 g/dL).
582
M.B. Said et al. / Early Human Development 90 (2014) 579–585
Fig. 2. Distribution of phyla in the gut microbiota of 90 preterm infants (33 cases and 57 controls). No significant differences were detected between the two groups.
significantly higher in cases than in controls (27.3% of cases vs. 8.8% of controls, p = 0.045) (Fig. 4a); the opposite pattern occurred for the Staphylococcus sp. (12% of cases vs. 32% of controls, p = 0.047) (Fig. 4b). Staphylococcus sp. corresponded to a species level OTU that could not be affiliated with certainty to a taxonomic species. No other significant differences were identified at the genus or species levels, particularly concerning colonization with bacteria of the genus Clostridium. A systematic search for evidence of rotavirus and adenovirus in infants with gross blood in stools revealed the presence of both viruses in one child and the presence of rotavirus in another infant. We analyzed isolated rectal bleeding separately and we found results close to what is observed in the whole population (significantly higher proportion of infants with E. coli in the cases).
4. Discussion In our population of preterm infants born before a gestational age of 34 weeks, we observed modifications in gut microbiota in infants with gross blood in stools which was recently reported with NEC [18,19].
We observed a high incidence of rectal bleeding which is in-line with the study from Luoto et al. in a similar population of preterm infants [10]. The number of subjects in each group was not equal because we were not able to pair each case with two controls infants. Nine infants initially included as controls later presented gross blood in stools. However, the matching of cases and controls was successful, as there were no significant differences in the characteristics of the included subjects. Differences in the use of anti-reflux medication and thickener between the two groups indicated that feeding tolerance was lower in infants presenting gross blood in stools. The potential for thickened foods to increase the risk of NEC has been suggested but not confirmed [20,21]. We were not able to carry out multivariate analyses due to the small number of events (only three infants received domperidone and thickening) and the small number of subjects in each group. We cannot exclude that thickening could modify intestinal microbiota but there is no data from the literature that supports that hypothesis. One strength of our study is that our controls were selected from the set of infants hospitalized in the same unit at the same time as the cases, which is crucial when considering the influence of environment on gut
Fig. 3. Distribution of the most-represented bacterial groups defined to the genus or to the family level in the gut microbiota of 90 preterm infants (33 cases and 57 controls). There were no significant between-group differences.
M.B. Said et al. / Early Human Development 90 (2014) 579–585
583
a p = 0.045
b p = 0.047
Fig. 4. Enterobacterial (a) and staphylococcal (b) operational taxonomic units (OTUs) in stool samples from 90 preterm infants (33 cases and 57 controls). A significant difference in distribution was detected for E. coli (p = 0.045) and Staphylococcus sp. (p = 0.047).
microbiota in these infants [22]. The diversity of the gut microbiota was low, as suggested previously [15], and the composition of the microbiota was roughly similar between cases and controls; OTUs were distributed through four bacterial phyla and 10 major genera or families, with the absence of an OTU specific to one patient group. However, the relative frequencies of E. coli and Staphylococcus sp. varied significantly between cases and controls, suggesting that gut dysbiosis (more E. coli and less Staphylococcus) could be another cause for gross blood in stools. Dysbiosis has been suggested to explain the occurrence of NEC [8,18] and late-onset sepsis [23]. Mai et al. suggested that a distortion in the normal microbiota composition, and not an enrichment of pathogens, could be associated with late-onset sepsis in preterm infants [23]. Jenke et al. suggested that expansion of E. coli in ELBW infants is associated with NEC [24]. In preterm infants, staphylococci are predominant during the first month of life [15]. In our study, stool sample was collected from cases and controls at approximately three weeks of age. Therefore, the reduction of the presence of Staphylococcus together with an increase in the presence of E. coli in cases represents a substantial distortion of the gut microbiota in these subjects. The significant increase in the presence of E. coli in infants with gross blood in stools is of particular interest. This species consists of several genotypes (pathovars) that display particular pathogen behaviors. Classical enteropathogenic pathovars, genotypes that are considered to be strict pathogens, are rarely involved in the infections of infants
hospitalized in neonatal intensive care units. However, some genotypes seem to display virulence that differs from the classical acute gastroenteritis associated with most enteropathogenic pathovars. For instance, the enterohemorrhagic strain O157:H7 has been associated with a case of NEC [25], and strains with particular virulence determinants for adhesion and invasion seem to be involved in bacteremia [26] and in inflammatory bowel diseases in infants [27]. A recently described pathovar, diffusely adherent E. coli, not only harbors genes associated with virulence, but it also contributes to the normal microbiota present in asymptomatic patients [28]. Consequently, this pathovar may exhibit an opportunistic behavior linked to dysbiosis, with the augmentation of its representation in the microbiota leading to symptoms. The microbiological approach used in this study allowed the identification of bacteria to the species level, but not to the genotype level. Consequently, we cannot associate the clones of E. coli detected in patients with a particular pathovar. Additional investigations will be required in order to describe the E. coli genotypic diversity in the gut microbiota of premature neonates. In preterm infants human milk prophylactically supplemented with Lactobacillus rhamnosus GG, Luoto et al. found no alterations in gut microecology associated with gross blood in stool samples from a very small number of infants (25 cases in Luoto et al. vs. 89 in the present investigation) [10]. In contrast to our results, Luoto et al. were unable to identify differences between cases and controls, probably because
584
M.B. Said et al. / Early Human Development 90 (2014) 579–585
they focused a priori on a limited number of species. It is possible that prophylactic supplementation with a probiotic contributed in their study to successful gut colonization with bifidobacteria and lactobacilli, but further investigation will be necessary before concluding that a particular gut microbiota composition is typical of healthy breastfed infants, as claimed by Luoto et al. [10]. A limitation of our study was the limited number of subjects. Therefore we were unable to correlate, in the cases, the clinical and/or radiographic characteristics with micro-organisms. We were neither able to investigate whether clinical pattern of infants with E. coli was worse than the others. Further prospective studies in larger population should be performed to answer these questions. Another limitation of our study was that the diagnosis was performed by sole clinical observation, which could be associated to a degree of subjectivity. However, the diagnosis was assessed by experienced physicians. We did not confirm the diagnosis as biological testing (e.g. lab analysis or bedside test such as Hemoccult®) is not considered reliable in pediatric patients and there is no published validation in neonates [29]. Another limitation could have been related to the bacteriological method we used (TTGE). Microbiota descriptions based on fingerprint methods allowed the detection of only majority sub-populations forming the community. Recent studies comparing fingerprinting and high-throughput sequencing showed that fingerprint methods gave a valuable survey for the detection of major differences between groups of patients [30]. High-throughput sequencing should be a second step to study deeply the communities in order to detect dysbiosis affecting minority bacterial populations. As other sequencing methods focused on a small variable part of the 16S rRNA gene, communities TTGE fingerprints lack of resolution to identify to the species level with accuracy, particularly in some particular taxa. Here in, we confirmed and complete the identification by the use of alternative markers tuf and rpoB recognized to have higher resolution for species level identification in staphylococci and enterobacteria. The very high incidence of cesarean section in our study population could represent a limitation to the generalizability of our results. There are still debates about the influence of the mode of delivery on gut microbiota. In term neonates, cesarean deliver may hamper the development of digestive microbiota [31,32]. In preterm infants, there are still discrepancies about the influence of the delivery mode on gut microbiota. Jacquot et al. reported a faster development of gut microbiota after a cesarean delivery [15] while other authors did not find any association between the mode of delivery and the number of bacterial species, in small number of subjects [33,34]. Finally, Penders et al. observed a significant impact on colonization rates, which were lower for bifidobacteria and higher for Clostridium difficile and counts of E. coli in infants born after C-section [35]. In conclusion, our study suggests that dysbiosis could be a risk factor for gross blood in stools. Additional, larger studies are needed to confirm the implications of the presence of various genotypes of E. coli and to evaluate preventive actions such as the use of probiotics and/or prebiotics. Conflicts of interest statement The authors have no conflicts of interest to declare. Authors have no actual or potential conflict of interest including any financial, personal or other relationships with other people or organizations within that could inappropriately influence their work. All authors made substantial contribution to the design of the study, acquisition of data, analysis and interpretation of data, drafting or revising the article and approved the final version of the manuscript. Acknowledgments The authors thank Brigitte Guy, Blandine Pastor-Diez, Marion Masclef, and Bernadette Reygrobellet for their assistance in collecting stools, and Paul Kretchmer for editing of the manuscript.
References [1] Thompson EC, Brown MF, Bowen EC, Smith LM, vander Griten D. Causes of gastrointestinal hemorrhage in neonates and children. South Med J 1996;89:370–4. [2] Levene MI. Rectal bleeding in the first month of life. Postgrad Med J 1979;55:22–3. [3] Canioni D, Pauliat S, Gaillard JL, Mougenot JF, Bompard Y, Berche P, et al. Histopathology and microbiology of isolated rectal bleeding in neonates: the so-called ‘ecchymotic colitis’. Histopathology 1997;30:472–7. [4] Brotschi B, Baenziger O, Frey B, Bucher HU, Ersch J. Early enteral feeding in conservatively managed stage II necrotizing enterocolitis is associated with a reduced risk of catheter-related sepsis. J Perinat Med 2009;37:701–5. [5] Stewart CJ, Marrs EC, Magorrian S, Nelson A, Lanyon C, Perry JD, et al. The preterm gut microbiota: changes associated with necrotizing enterocolitis and infection. Acta Paediatr 2012;101:1121–7. [6] Morrow AL, Lagomarcino AJ, Schibler KR, Taft DH, Yu Z, Wang B, et al. Early microbial and metabolomic signatures predict later onset of necrotizing enterocolitis in preterm infants. Microbiome 2013;1(13). [7] Mai V, Young CM, Ukhanova M, Wang X, Sun Y, Casella G, et al. Fecal microbiota in premature infants prior to necrotizing enterocolitis. PLoS One 2011;6:e20647. [8] Normann E, Fahlén A, Engstrand L, Lilja HE. Intestinal microbial profiles in extremely preterm infants with and without necrotizing enterocolitis. Acta Paediatr 2013; 102:129–36. [9] Maayan-Metzger A, Ghanem N, Mazkereth R, Kuint J. Characteristics of neonates with isolated rectal bleeding. Arch Dis Child Fetal Neonatal Ed 2004;89:F68–70. [10] Luoto R, Rodriguez-Diaz J, Collado MC, Salminen S, Isolauri E, Lehtonen L. Gross blood in stools of premature neonates, a clinical and microbiological follow-up study. Acta Paediatr 2013;102:486–91. [11] Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 1986;33:179–201. [12] Règles de bonnes pratiques de collecte, de préparation, de qualification, de traitement, de conservation, de distribution et de délivrance sur prescription médicale du lait humain par les lactariums; Afssaps, Décision du 3 décembre 2007– JO du 5 janvier 2008. english version available on ADLF Website http://sdp. perinat-france.org/ADLF/document/bonnes_pratiques/lactarium_guide_bonnes_ pratiques_5_janvier_2008_traduction_anglais.pdf. [Accessed september 9th, 2013]. [13] Usher R, McLean F. Intrauterine growth of live-born Caucasian infants at sea level: standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr 1969;74:901–10. [14] Roudière L, Jacquot A, Marchandin H, Aujoulat F, Devine R, Zorgniotti I, et al. Optimized PCR-Temporal Temperature Gel Electrophoresis compared to cultivation to assess diversity of gut microbiota in neonates. J Microbiol Methods 2009; 79:156–65. [15] Jacquot A, Neveu D, Aujoulat F, Mercier G, Marchandin H, Jumas-Bilak E, et al. Dynamics and clinical evolution of bacterial gut microflora in extremely premature patients. J Pediatr 2011;158:390–6. [16] Filleron A, Simon M, Hantova S, Jacquot A, Cambonie G, Marchandin H, et al. tuf-PCRtemporal temperature gradient gel electrophoresis for molecular detection and identification of staphylococci: application to breast milk and neonate gut microbiota. J Microbiol Methods 2014;98:67–75. [17] Navarrete P, Magne F, Mardones P, Riveros M, Opazo R, Suau A, et al. Molecular analysis of intestinal microbiota of rainbow trout (Oncorhynchus mykiss). FEMS Microbiol Ecol 2010;71:148–56. [18] Torrazza RM, Neu J. The altered gut microbiome and necrotizing enterocolitis. Clin Perinatol 2013;40:93–108. [19] Stewart CJ, Marrs EC, Nelson A, Lanyon C, Perry JD, Embleton ND, et al. Development of the preterm gut microbiome in twins at risk of necrotising enterocolitis and sepsis. PLoS One 2013;8:e73465. [20] Clarke P, Robinson MJ. Thickening milk feeds may cause necrotizing enterocolitis. Arch Dis Child Fetal Neonatal Ed 2004;89:F280. [21] Drenckpohl D, Knaub L, Schneider C, McConell C, Wang H, Macwan K. Risk factors that may predispose premature infants to increased incidence of necrotizing enterocolitis. Infant Child Adolesc Nutr 2010;2:37–44. [22] Murono K, Fujita K, Yoshikawa M, Saijo M, Inyaku F, Kakehashi H, et al. Acquisition of nonmaternal Enterobacteriaceae by infants delivered in hospitals. J Pediatr 1993; 122:120–4. [23] Mai V, Torrazza RM, Ukhanova M, Wang X, Sun Y, Li N, et al. Distortions in development of intestinal microbiota associated with late onset sepsis in preterm infants. PLoS One 2013;8:e52876. [24] Jenke AC, Postberg J, Mariel B, Hensel K, Foell D, Däbritz J, et al. S100A12 and hBD2 correlate with the composition of the fecal microflora in ELBW infants and expansion of E. coli is associated with NEC. Biomed Res Int 2013;2013:150372. [25] Guner YS, Malhotra A, Ford HR, Stein JE, Kelly LK. Association of Escherichia coli O157:H7 with necrotizing enterocolitis in a full-term infant. Pediatr Surg Int 2009; 25:459–63. [26] Mahjoub-Messai F, Bidet P, Caro V, Diancourt L, Biran V, Aujard Y, et al. Escherichia coli isolates causing bacteremia via gut translocation and urinary tract infection in young infants exhibit different virulence genotypes. J Infect Dis 2011;203:1844–9. [27] Sobieszczańska BA, Duda-Madej AB, Turniak MB, Franiczek R, Kasprzykowska U, Duda AK, et al. Invasive properties, adhesion patterns and phylogroup profiles among Escherichia coli strains isolated from children with inflammatory bowel disease. Adv Clin Exp Med 2012;21:591–9. [28] Mansan-Almeida R, Pereira AL, Giugliano LG. Diffusely adherent Escherichia coli strains isolated from children and adults constitute two different populations. BMC Microbiol 2013;13(22). [29] Rosenthal P, Jennings MT. Comparison of fecal occult blood tests for detection of gastrointestinal bleeding in pediatric patients. Am J Gastroenterol 1992;87:1575–9.
M.B. Said et al. / Early Human Development 90 (2014) 579–585 [30] Li W, Han L, Yu P, Ma C, Wu X, Moore JE, et al. Molecular characterization of skin microbiota between cancer cachexia patients and healthy volunteers. Microb Ecol Apr. 2014;67(3):679–89. [31] Grönlund MM, Lehtonen OP, Eerola E, Kero P. 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(1):19–25. [32] Biasucci G, Rubini M, Riboni S, Morelli L, Bessi E, Retetangos C. Mode of delivery affects the bacterial community in the newborn gut. Early Hum Dev 2010; 86:13–5.
585
[33] Sakata H, Yoshioka H, Fujita K. Development of the intestinal flora in very low birth weight infants compared to normal full-term newborns. Eur J Pediatr 1985; 144(2):186–90. [34] Gewolb IH, Schwalbe RS, Taciak VL, Harrison TS, Panigrahi P. Stool microflora in extremely low birthweight infants. Arch Dis Child Fetal Neonatal Ed 1999;80(3): F167–73. [35] Penders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006; 118:511–21.
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
Gut microbiota in preterm infants with gross blood in stools: A prospective, controlled study Mohamed Ben Said a,b, Stephane Hays a,c, Deplhine Maucort-Boulch d,e,f, Abdallah Oulmaati a, Stefanyia Hantova g, Claire-Marie Loys a, Estelle Jumas-Bilak g,h, Jean-Charles Picaud a,c,f,⁎ a
Department of Neonatology, University Hospital Croix Rousse, Hospices Civils de Lyon, F-69004 Lyon, France Faculty of Medicine of Tunis, University of Tunis El-Manar, 1007 Bab Saâdoun, Tunis, Tunisia Rhone-Alpes Human Nutrition Research Center, F-69310 Pierre Bénite, France d Department of Biostatistics, Hospices Civils de Lyon, F-69003 Lyon, France e CNRS, Laboratoire Biostatistique Santé, UMR 5558, F-69310 Pierre Bénite, France f Lyon-Sud Charles Merieux Medical School, Claude Bernard University Lyon 1, F-69310 Pierre Bénite, France g University Montpellier 1, Laboratoire de Bacteriologie-Virologie, EA 3755 UM1, Faculte de Pharmacie, F-34093 Montpellier, France h Department of Hospital Hygiene, CHU de Montpellier, F-34000 Montpellier, France b c
a r t i c l e
i n f o
Article history: Received 21 March 2014 Received in revised form 6 July 2014 Accepted 8 July 2014 Available online xxxx Keywords: Prematurity Gut microbiota Staphylococcus Infection Rectal bleeding Human milk
a b s t r a c t Objective: Gross blood in stools is a peculiar entity in preterm infants, but little is known about its etiology. As gut microbiota can be distorted in preterm infants, we aimed to evaluate the gut microbiota in infants with gross blood in stools. Study design: In a prospective, controlled, single-center study, we enrolled all infants born before 34 weeks of gestational age presenting gross blood in stools that was either completely isolated or associated with mild clinical symptoms or radiological signs. Each case was paired with two controls who were hospitalized in the same unit and were matched for gestational age and birth weight. The diversity of the gut microbiota was analyzed using 16S rRNA gene PCR and temporal temperature gel electrophoresis. We calculated a diversity score corresponding to the number of operational taxonomic units present in the microbiota. Results: Thirty-three preterm infants with gross blood in stools were matched with 57 controls. Clinical characteristics were similar in cases and controls. There was no statistically significant difference in the diversity score between the two groups, but microbiota composition differed. The proportion of infants with Escherichia coli was significantly higher in cases than in controls (p = 0.045) and the opposite pattern occurred for Staphylococcus sp. (p = 0.047). Conclusion: Dysbiosis could be a risk factor for gross blood in stools in preterm infants. Additional, larger studies are needed to confirm the implications of the presence of different genotypes of E. coli and to evaluate preventive actions such as the prophylactic use of probiotics and/or prebiotics. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction In premature infants gross blood in stools may be related to severe diseases (necrotizing enterocolitis (NEC), colitis in Hirschsprung's disease, infectious colitis, or hemorrhagic disease of the newborn) or to milder diseases (allergic colitis, anal fissure, or swallowing blood syndrome) [1,2]. Apart from these conditions, gross blood in stools is a peculiar entity either completely isolated or is associated with mild clinical symptoms or radiological signs (gastric residuals, vomiting,
Abbreviations: NEC, necrotizing enterocolitis; OTUs, operational taxonomic units; TTGE, temporal temperature gel electrophoresis. ⁎ Corresponding author at: Service de Néonatologie, Hôpital de la Croix Rousse, 103 Grande rue de la Croix Rousse, 69004 Lyon, France. Tel.: +33 472 001 550. E-mail address: [email protected] (J.-C. Picaud).
http://dx.doi.org/10.1016/j.earlhumdev.2014.07.004 0378-3782/© 2014 Elsevier Ltd. All rights reserved.
mild abdominal distension, and mild radiological intestinal dilation). Isolated rectal bleeding has been related to ecchymotic colitis [3]. Occurrence of gross blood in stools has a significant impact on neonatal care, as the management of this disorder often requires a fasting period to start or to extend parenteral nutrition, which increases the risk of catheter-related sepsis [4]. However, published data on gross blood in stools in preterm infants are scarce. While the predisposing factors of severe NEC (≥stage II) are wellknown, including colonization by potentially pathogenic bacteria or gut dysbiosis [5–8], it is not the case for gross blood in stools. MaayanMetzger et al. reported that a feeding regimen that did not include breast milk was the only variable that predicted isolated rectal bleeding and emphasized the benign nature of this bleeding [9]. Luoto et al. did not identify any risk factors in a small number of very-low-birth weight infants fed human milk supplemented with probiotics [10]. Luoto et al.
580
M.B. Said et al. / Early Human Development 90 (2014) 579–585
did not detect significant changes in the composition of the gut microflora in preterm infants with gross blood in stools, but they focused their investigation on a few main bacterial genera and species [10]. In the present investigation, we evaluated the relationship between the composition of the intestinal microbiota and gross blood in stools in preterm infants. 2. Methods 2.1. Study design and population We performed a prospective, controlled, single-center study. All preterm infants hospitalized between January and October 2011 in our tertiary care unit at the University Hospital Croix Rousse, Lyon, France were eligible for study enrolment if they met the inclusion criteria of birth gestational age ≤34 weeks, absence of congenital malformation, and gross blood in stools. The nurses in charge of daily care for cases and controls collected systematically few times per day information about digestive tolerance as routinely recommended in our unit: gastric residuals, vomiting, abdominal distension, stool consistency and gross blood in stools. Diagnosis was confirmed by physicians by clinical observation. Gross blood in stools was either completely isolated or associated with mild clinical symptoms or radiological signs (gastric residuals, vomiting, mild abdominal distension, and mild radiological intestinal dilation). We did not include infants who developed NEC ≥stage II according to the modified Bell classification [11] or exhibited spontaneous intestinal perforation. The study protocol was approved by the ethics committee of Lyon (Comité de protection des personnes Sud Est IV Lyon). Each case was matched with two controls. Controls were the first two preterm infants hospitalized in the same unit during the same period, and whose birth weight and gestational age were similar (±100 g and ±1 week, respectively) to those of the cases. These infants were selected by research nurses among hospitalized infants fulfilling the above criteria, independently from clinicians who took care of these infants. Infants included as controls were monitored as routinely recommended in our unit: daily monitoring of gastric residuals, vomiting, abdominal distension, stool consistency and blood in stools. When infants included as controls later presented rectal bleeding they were excluded from the control group and not replaced by another one, then included in the group of cases and paired with two controls. 2.2. Routine care protocol The feeding regimen of the very-low-birth weight infants was not changed during the study period. Cases and controls were fed according to the same NICU protocol. Enteral feeding was started at day 1 or 2 and complementary parenteral feeding was administered until the enteral intake reached 100 mL/kg/day. All infants were fed pasteurized human milk (their own mother's milk or donor milk) according to French regulations [12] until their body weight was ~1500 g. Then, if the mother had no milk, feeding with the same preterm formula (PreMilumel, Milumel, Torce, France) was commenced. Human milk was supplemented with a cow's milk protein based multicomponent fortifier (Eoprotine, Milumel, Torce France). None of the infants included in our study received probiotics. Full enteral feeding was 160 mL/kg/day and was achieved by increasing the feeding by 10–20 mL/kg/day depending on the feeding tolerance of the infant. According to the NICU protocol enteral feeding was started as continuous feeding followed by bolus feeding when full enteral feeding was reached and digestive tolerance was satisfactory. After full enteral feeding was achieved, a thickener was used when infants presented signs of gastro-esophageal reflux (carob-based when the infant was fed human milk and starch-based when the infant was fed preterm formula). If the signs of gastro-esophageal reflux persisted, we administered domperidone with or without a proton pump inhibitor (omeprazole or esomeprazole) depending on the
severity of the signs. Infants who presented clinical and ultrasound signs of persistent ductus arteriosus were treated with ibuprofen (10, 5, and 5 mg/kg/day at days 1, 2, and 3, respectively). 2.3. Management of gross blood in stools When gross blood in stools occurred, we systematically assessed blood count cells, serum C-reactive protein levels, serum procalcitonin levels, and blood culture (aerobic and anaerobic). Stool samples were collected for bacteriological culture and for the identification of rotavirus and adenovirus. Abdominal x-rays were also performed. Then, infants received no enteral feeding for a few days, depending on the clinical and radiographic data collected at the time of rectal bleeding and in the following days. For each case with gross blood in stools, stool samples were collected from the two matched control infants. 2.4. Data collection We collected data on the pregnancies, the deliveries (antenatal steroids, mode of delivery, maternal diseases), the infants' characteristics at birth (gestational age, birth weight, gender, Apgar score), growth restriction (body weight less than −2SD for gestational age) [13] treatments during hospitalization before gross blood in stools (ventilation, antibiotics, ibuprofen treatment, postnatal steroids, anti-reflux treatment, breast milk or formula), treatments at the time of gross blood in stools (body weight, clinical examination, x-ray results, C-reactive protein levels, procalcitonin levels, complete blood count, blood culture), and information about the infant after gross blood in stools (fasting period, recurrence of rectal bleeding). 2.5. Analysis of gut microbiota The qualitative composition of the gut microbiota was assessed via extraction of bacterial DNA from stool samples, PCR, temporal temperature gradient electrophoresis (TTGE), and identification of amplified sequences. Stool samples collected from cases and controls were stored at −80 °C until analysis. Approximately 50 mg of each homogenized stool specimen was placed in 1 mL of sterile DNA-free water in a 1.5-mL tube. The suspension was centrifuged for 10 min at 10,000 ×g. DNA was extracted from the pellet using the MasterPure Gram Positive DNA Purification Kit (Epicentre, Madison, WI, USA) according to the supplier's instructions and optimized by Roudière et al. [14]. The V2–V3 region (233 bp) of the 16S rRNA gene was amplified with primers HDA1 (with a GC-clamp) and HDA2 [14]. The reaction mixture (50 μL) contained 200 μM of deoxynucleoside triphosphate mix, 10 pmol of each primer, 2.5 U of Taq DNA polymerase in the appropriate buffer (FastStart High Fidelity PCR system, Roche, Basel, Switzerland), and 1 μL of template DNA. The amplification program, which was carried out with a Mastercycler apparatus (Eppendorf, Le Pecq, France), was 95 °C for 2 min, 30 cycles of 95 °C for 1 min, 62 °C for 30 s, and 72 °C for 1 min, with a final extension of 72 °C for 7 min. PCRs were checked by electrophoresis migration on a 1.5% agarose gel stained with ethidium bromide (500 μg/mL) and visualized on an ultraviolet transillumination system. The DCode universal mutation detection system (Bio-Rad Laboratories, Marne La Coquette, France) was used for TTGE. Five microliters of the PCR product was added to 5 μL of loading buffer. The gels were prepared with 8% (wt/vol) bisacrylamide (37.5:1) and 7 M urea and were run with 1X Tris–acetate–EDTA buffer. Denaturating electrophoresis was performed at 46 V for 16 h with a temperature gradient of 63–70 °C (0.4 °C/h). Gels were stained with 10 mg/mL ethidium bromide and photographed with an ultraviolet transillumination system. Identification of the TTGE bands was performed as described previously, via comparison of migration distances to the “gut microbiota diversity ladder” and sequencing on an ABI 3730XL sequencer (Takeley, United Kingdom) [14,15]. The resulting DNA sequences were compared
M.B. Said et al. / Early Human Development 90 (2014) 579–585
with Genbank (http://www.ncbi.nlm.nih.gov/) and with the Ribosomal Database Project II (http://rdp.cme.msu.edu/) using the Basic Local Alignment Search Tool (BLAST) and Seqmatch, respectively. Each sequence was affiliated with an operational taxonomic unit (OTU) that may be defined first at the species level, and otherwise at the genus level on the basis of the percent of sequence identity with the closest relative in the sequence databases. Each OTU was further classified into groups of taxonomic or clinical relevance (genus or family). The diversity score corresponded to the number of different OTUs in a sample as previously validated [14,15]. Complementary TTGE experiments were performed to obtain species level affiliation, for enterobacteria and staphylococci. Tuf-PCRTTGE [16] and rpoB-PCR-TTGE [17] were performed as described in order to determine the species in the genus Staphylococcus and in the family Enterobacteriaceae, respectively. 2.6. Statistical analysis Main variables were described with numbers (percentage) for categorical variables and medians [minimum, maximum] for continuous variables. We used the chi-squared test or Fisher's exact test to compare qualitative variables and the Wilcoxon test to compare quantitative variables. p-Values less than 0.05 were considered significant. We used SPSS version 16.0 (SPSS Inc., Chicago, IL, USA) to perform the statistical analyses.
581
Table 1 Clinical characteristics of 33 cases before gross blood in stools and 57 controls before inclusion. Cases n = 33 Pregnancy Hypertension Gestational diabetes Transfusion syndrome Cesarean section Intrapartum antibiotics
07 (21) 0 (0) 2 (6) 28 (85) 11 (33)
Controls n = 57 9 (16) 2 (3) 3 (5) 46 (81) 27 (47)
p-Value
0.517 0.530 1.000 0.620 0.194
At birth Gestational age in weeks Birth weight in g Birth weight ≤ 2 SDs Male gender
30 [25, 34] 1200 [635, 2160] 1 (3) 11 (33)
30 [23, 34] 1240 [570, 2280] 2 (3) 28 (49)
0.694 0.714 1.000 0.145
Drugs before rectal bleeding Domperidone Proton pump inhibitor Ibuprofen Postnatal steroids Postnatal antibiotics Duration of postnatal antibiotics
3 (9) 1 (3) 6 (18) 2 (6) 18 (54) 2 [2, 10]
0 (0) 1 (2) 9 (16) 1 (2) 30 (53) 2 [2, 3]
0.046 1.000 0.769 0.552 1.000 0.489
0 [0, 1512] 190 [0, 1655] 20 [0, 1500]
0.530 0.933 0.443
Respiratory support before rectal bleeding Assisted ventilation, in hours 0 [0, 1479] CPAP, in hours 218 [0, 1600] Oxygen therapy, in hours 3 [0, 1291]
Values are given as number (percentage) or median [minimum, maximum]. SD, standard deviation; CPAP, continuous positive airway pressure.
3. Results Of the 270 preterm infants with gestational age less than 34 weeks that were hospitalized in our neonatal intensive care unit between January and October 2011, 33 (12.2%) presented gross blood in stools. We matched 57 controls to these 33 cases (Fig. 1). Approximately two out three cases had isolated rectal bleeding (n = 22). Nine infants initially selected as controls presented a gross blood in stools.
270 preterm infants ≤34 wks
Rectal bleeding (n=26)
No rectal bleeding (n=244)
Not included for : . Spontaneous intestinal perforation (n=1)
The clinical characteristics of the two groups before and at inclusion were comparable, with the exceptions of treatment with domperidone and use of thickener. There was no significant difference in antibiotic exposure between the two groups (Tables 1 and 2). The proportion of singletons was similar in both groups (57% of cases vs. 47% of controls, p = 0.351). Most infants received antenatal steroids (91% of cases vs. 96% of controls, p = 0.352). The bacterial diversity scores observed in the overall study population were low, ranging from 1 to 10, but the scores varied between patients. There was no significant difference in the diversity score between cases and controls (3 [1, 9] vs. 3 [1, 10] OTUs, respectively; p = 0.654). A total of 39 OTUs were detected in the populations studied. These OTUs were distributed through four phyla, with a predominance of OTUs in Firmicutes and Proteobacteria (Fig. 2), and 10 major genera or families, with a predominance of OTUs in Staphylococcus, Enterobacteriaceae and Clostridium (Fig. 3). The distributions of phyla (Fig. 2) and most-represented genera or families (Fig. 3) did not significantly differ between cases and controls. However, enterobacteria occurred more frequently in cases but the difference was not statistically significant (Fig. 3). The proportion of infants harboring Escherichia coli was
. NEC grade 2 (n=1) Gross blood in stools (n=24)
Table 2 Clinical characteristics of preterm infants at the time of gross blood in stools (cases) and at inclusion (controls).
Controls (n=48)
Cases n = 33
Gross blood in stools (n=9)
Gross blood in stools (n=33)
Controls n=18
Controls (n=57)
Fig. 1. Flow chart of study design.
Postmenstrual age, in weeks Body weight, in g
33 [27, 46] 1752 [880, 3454] Postnatal age, in days 21.5 [3, 102] Postnatal growth restriction 3 (9.4) Enteral feeding at time of rectal bleeding 18 (54) Human milka Preterm formula 11 (33) 4 (12) Human milka + preterm formula Thickener 10 (30)
Controls n = 57
p-Value
33 [29, 38] 1670 [940, 2900] 17 [4, 75] 1 (1.8)
0.427 0.110 0.147 0.131
31 (54) 16 (28) 10 (17) 6 (10)
0.988 0.600 0.494 0.018
Values are presented as number (percentage) or median [minimum, maximum]. a Human milk fortified with a multicomponent fortifier (Eoprotine, Milumel, Torce France: 4 g/dL).
582
M.B. Said et al. / Early Human Development 90 (2014) 579–585
Fig. 2. Distribution of phyla in the gut microbiota of 90 preterm infants (33 cases and 57 controls). No significant differences were detected between the two groups.
significantly higher in cases than in controls (27.3% of cases vs. 8.8% of controls, p = 0.045) (Fig. 4a); the opposite pattern occurred for the Staphylococcus sp. (12% of cases vs. 32% of controls, p = 0.047) (Fig. 4b). Staphylococcus sp. corresponded to a species level OTU that could not be affiliated with certainty to a taxonomic species. No other significant differences were identified at the genus or species levels, particularly concerning colonization with bacteria of the genus Clostridium. A systematic search for evidence of rotavirus and adenovirus in infants with gross blood in stools revealed the presence of both viruses in one child and the presence of rotavirus in another infant. We analyzed isolated rectal bleeding separately and we found results close to what is observed in the whole population (significantly higher proportion of infants with E. coli in the cases).
4. Discussion In our population of preterm infants born before a gestational age of 34 weeks, we observed modifications in gut microbiota in infants with gross blood in stools which was recently reported with NEC [18,19].
We observed a high incidence of rectal bleeding which is in-line with the study from Luoto et al. in a similar population of preterm infants [10]. The number of subjects in each group was not equal because we were not able to pair each case with two controls infants. Nine infants initially included as controls later presented gross blood in stools. However, the matching of cases and controls was successful, as there were no significant differences in the characteristics of the included subjects. Differences in the use of anti-reflux medication and thickener between the two groups indicated that feeding tolerance was lower in infants presenting gross blood in stools. The potential for thickened foods to increase the risk of NEC has been suggested but not confirmed [20,21]. We were not able to carry out multivariate analyses due to the small number of events (only three infants received domperidone and thickening) and the small number of subjects in each group. We cannot exclude that thickening could modify intestinal microbiota but there is no data from the literature that supports that hypothesis. One strength of our study is that our controls were selected from the set of infants hospitalized in the same unit at the same time as the cases, which is crucial when considering the influence of environment on gut
Fig. 3. Distribution of the most-represented bacterial groups defined to the genus or to the family level in the gut microbiota of 90 preterm infants (33 cases and 57 controls). There were no significant between-group differences.
M.B. Said et al. / Early Human Development 90 (2014) 579–585
583
a p = 0.045
b p = 0.047
Fig. 4. Enterobacterial (a) and staphylococcal (b) operational taxonomic units (OTUs) in stool samples from 90 preterm infants (33 cases and 57 controls). A significant difference in distribution was detected for E. coli (p = 0.045) and Staphylococcus sp. (p = 0.047).
microbiota in these infants [22]. The diversity of the gut microbiota was low, as suggested previously [15], and the composition of the microbiota was roughly similar between cases and controls; OTUs were distributed through four bacterial phyla and 10 major genera or families, with the absence of an OTU specific to one patient group. However, the relative frequencies of E. coli and Staphylococcus sp. varied significantly between cases and controls, suggesting that gut dysbiosis (more E. coli and less Staphylococcus) could be another cause for gross blood in stools. Dysbiosis has been suggested to explain the occurrence of NEC [8,18] and late-onset sepsis [23]. Mai et al. suggested that a distortion in the normal microbiota composition, and not an enrichment of pathogens, could be associated with late-onset sepsis in preterm infants [23]. Jenke et al. suggested that expansion of E. coli in ELBW infants is associated with NEC [24]. In preterm infants, staphylococci are predominant during the first month of life [15]. In our study, stool sample was collected from cases and controls at approximately three weeks of age. Therefore, the reduction of the presence of Staphylococcus together with an increase in the presence of E. coli in cases represents a substantial distortion of the gut microbiota in these subjects. The significant increase in the presence of E. coli in infants with gross blood in stools is of particular interest. This species consists of several genotypes (pathovars) that display particular pathogen behaviors. Classical enteropathogenic pathovars, genotypes that are considered to be strict pathogens, are rarely involved in the infections of infants
hospitalized in neonatal intensive care units. However, some genotypes seem to display virulence that differs from the classical acute gastroenteritis associated with most enteropathogenic pathovars. For instance, the enterohemorrhagic strain O157:H7 has been associated with a case of NEC [25], and strains with particular virulence determinants for adhesion and invasion seem to be involved in bacteremia [26] and in inflammatory bowel diseases in infants [27]. A recently described pathovar, diffusely adherent E. coli, not only harbors genes associated with virulence, but it also contributes to the normal microbiota present in asymptomatic patients [28]. Consequently, this pathovar may exhibit an opportunistic behavior linked to dysbiosis, with the augmentation of its representation in the microbiota leading to symptoms. The microbiological approach used in this study allowed the identification of bacteria to the species level, but not to the genotype level. Consequently, we cannot associate the clones of E. coli detected in patients with a particular pathovar. Additional investigations will be required in order to describe the E. coli genotypic diversity in the gut microbiota of premature neonates. In preterm infants human milk prophylactically supplemented with Lactobacillus rhamnosus GG, Luoto et al. found no alterations in gut microecology associated with gross blood in stool samples from a very small number of infants (25 cases in Luoto et al. vs. 89 in the present investigation) [10]. In contrast to our results, Luoto et al. were unable to identify differences between cases and controls, probably because
584
M.B. Said et al. / Early Human Development 90 (2014) 579–585
they focused a priori on a limited number of species. It is possible that prophylactic supplementation with a probiotic contributed in their study to successful gut colonization with bifidobacteria and lactobacilli, but further investigation will be necessary before concluding that a particular gut microbiota composition is typical of healthy breastfed infants, as claimed by Luoto et al. [10]. A limitation of our study was the limited number of subjects. Therefore we were unable to correlate, in the cases, the clinical and/or radiographic characteristics with micro-organisms. We were neither able to investigate whether clinical pattern of infants with E. coli was worse than the others. Further prospective studies in larger population should be performed to answer these questions. Another limitation of our study was that the diagnosis was performed by sole clinical observation, which could be associated to a degree of subjectivity. However, the diagnosis was assessed by experienced physicians. We did not confirm the diagnosis as biological testing (e.g. lab analysis or bedside test such as Hemoccult®) is not considered reliable in pediatric patients and there is no published validation in neonates [29]. Another limitation could have been related to the bacteriological method we used (TTGE). Microbiota descriptions based on fingerprint methods allowed the detection of only majority sub-populations forming the community. Recent studies comparing fingerprinting and high-throughput sequencing showed that fingerprint methods gave a valuable survey for the detection of major differences between groups of patients [30]. High-throughput sequencing should be a second step to study deeply the communities in order to detect dysbiosis affecting minority bacterial populations. As other sequencing methods focused on a small variable part of the 16S rRNA gene, communities TTGE fingerprints lack of resolution to identify to the species level with accuracy, particularly in some particular taxa. Here in, we confirmed and complete the identification by the use of alternative markers tuf and rpoB recognized to have higher resolution for species level identification in staphylococci and enterobacteria. The very high incidence of cesarean section in our study population could represent a limitation to the generalizability of our results. There are still debates about the influence of the mode of delivery on gut microbiota. In term neonates, cesarean deliver may hamper the development of digestive microbiota [31,32]. In preterm infants, there are still discrepancies about the influence of the delivery mode on gut microbiota. Jacquot et al. reported a faster development of gut microbiota after a cesarean delivery [15] while other authors did not find any association between the mode of delivery and the number of bacterial species, in small number of subjects [33,34]. Finally, Penders et al. observed a significant impact on colonization rates, which were lower for bifidobacteria and higher for Clostridium difficile and counts of E. coli in infants born after C-section [35]. In conclusion, our study suggests that dysbiosis could be a risk factor for gross blood in stools. Additional, larger studies are needed to confirm the implications of the presence of various genotypes of E. coli and to evaluate preventive actions such as the use of probiotics and/or prebiotics. Conflicts of interest statement The authors have no conflicts of interest to declare. Authors have no actual or potential conflict of interest including any financial, personal or other relationships with other people or organizations within that could inappropriately influence their work. All authors made substantial contribution to the design of the study, acquisition of data, analysis and interpretation of data, drafting or revising the article and approved the final version of the manuscript. Acknowledgments The authors thank Brigitte Guy, Blandine Pastor-Diez, Marion Masclef, and Bernadette Reygrobellet for their assistance in collecting stools, and Paul Kretchmer for editing of the manuscript.
References [1] Thompson EC, Brown MF, Bowen EC, Smith LM, vander Griten D. Causes of gastrointestinal hemorrhage in neonates and children. South Med J 1996;89:370–4. [2] Levene MI. Rectal bleeding in the first month of life. Postgrad Med J 1979;55:22–3. [3] Canioni D, Pauliat S, Gaillard JL, Mougenot JF, Bompard Y, Berche P, et al. Histopathology and microbiology of isolated rectal bleeding in neonates: the so-called ‘ecchymotic colitis’. Histopathology 1997;30:472–7. [4] Brotschi B, Baenziger O, Frey B, Bucher HU, Ersch J. Early enteral feeding in conservatively managed stage II necrotizing enterocolitis is associated with a reduced risk of catheter-related sepsis. J Perinat Med 2009;37:701–5. [5] Stewart CJ, Marrs EC, Magorrian S, Nelson A, Lanyon C, Perry JD, et al. The preterm gut microbiota: changes associated with necrotizing enterocolitis and infection. Acta Paediatr 2012;101:1121–7. [6] Morrow AL, Lagomarcino AJ, Schibler KR, Taft DH, Yu Z, Wang B, et al. Early microbial and metabolomic signatures predict later onset of necrotizing enterocolitis in preterm infants. Microbiome 2013;1(13). [7] Mai V, Young CM, Ukhanova M, Wang X, Sun Y, Casella G, et al. Fecal microbiota in premature infants prior to necrotizing enterocolitis. PLoS One 2011;6:e20647. [8] Normann E, Fahlén A, Engstrand L, Lilja HE. Intestinal microbial profiles in extremely preterm infants with and without necrotizing enterocolitis. Acta Paediatr 2013; 102:129–36. [9] Maayan-Metzger A, Ghanem N, Mazkereth R, Kuint J. Characteristics of neonates with isolated rectal bleeding. Arch Dis Child Fetal Neonatal Ed 2004;89:F68–70. [10] Luoto R, Rodriguez-Diaz J, Collado MC, Salminen S, Isolauri E, Lehtonen L. Gross blood in stools of premature neonates, a clinical and microbiological follow-up study. Acta Paediatr 2013;102:486–91. [11] Walsh MC, Kliegman RM. Necrotizing enterocolitis: treatment based on staging criteria. Pediatr Clin North Am 1986;33:179–201. [12] Règles de bonnes pratiques de collecte, de préparation, de qualification, de traitement, de conservation, de distribution et de délivrance sur prescription médicale du lait humain par les lactariums; Afssaps, Décision du 3 décembre 2007– JO du 5 janvier 2008. english version available on ADLF Website http://sdp. perinat-france.org/ADLF/document/bonnes_pratiques/lactarium_guide_bonnes_ pratiques_5_janvier_2008_traduction_anglais.pdf. [Accessed september 9th, 2013]. [13] Usher R, McLean F. Intrauterine growth of live-born Caucasian infants at sea level: standards obtained from measurements in 7 dimensions of infants born between 25 and 44 weeks of gestation. J Pediatr 1969;74:901–10. [14] Roudière L, Jacquot A, Marchandin H, Aujoulat F, Devine R, Zorgniotti I, et al. Optimized PCR-Temporal Temperature Gel Electrophoresis compared to cultivation to assess diversity of gut microbiota in neonates. J Microbiol Methods 2009; 79:156–65. [15] Jacquot A, Neveu D, Aujoulat F, Mercier G, Marchandin H, Jumas-Bilak E, et al. Dynamics and clinical evolution of bacterial gut microflora in extremely premature patients. J Pediatr 2011;158:390–6. [16] Filleron A, Simon M, Hantova S, Jacquot A, Cambonie G, Marchandin H, et al. tuf-PCRtemporal temperature gradient gel electrophoresis for molecular detection and identification of staphylococci: application to breast milk and neonate gut microbiota. J Microbiol Methods 2014;98:67–75. [17] Navarrete P, Magne F, Mardones P, Riveros M, Opazo R, Suau A, et al. Molecular analysis of intestinal microbiota of rainbow trout (Oncorhynchus mykiss). FEMS Microbiol Ecol 2010;71:148–56. [18] Torrazza RM, Neu J. The altered gut microbiome and necrotizing enterocolitis. Clin Perinatol 2013;40:93–108. [19] Stewart CJ, Marrs EC, Nelson A, Lanyon C, Perry JD, Embleton ND, et al. Development of the preterm gut microbiome in twins at risk of necrotising enterocolitis and sepsis. PLoS One 2013;8:e73465. [20] Clarke P, Robinson MJ. Thickening milk feeds may cause necrotizing enterocolitis. Arch Dis Child Fetal Neonatal Ed 2004;89:F280. [21] Drenckpohl D, Knaub L, Schneider C, McConell C, Wang H, Macwan K. Risk factors that may predispose premature infants to increased incidence of necrotizing enterocolitis. Infant Child Adolesc Nutr 2010;2:37–44. [22] Murono K, Fujita K, Yoshikawa M, Saijo M, Inyaku F, Kakehashi H, et al. Acquisition of nonmaternal Enterobacteriaceae by infants delivered in hospitals. J Pediatr 1993; 122:120–4. [23] Mai V, Torrazza RM, Ukhanova M, Wang X, Sun Y, Li N, et al. Distortions in development of intestinal microbiota associated with late onset sepsis in preterm infants. PLoS One 2013;8:e52876. [24] Jenke AC, Postberg J, Mariel B, Hensel K, Foell D, Däbritz J, et al. S100A12 and hBD2 correlate with the composition of the fecal microflora in ELBW infants and expansion of E. coli is associated with NEC. Biomed Res Int 2013;2013:150372. [25] Guner YS, Malhotra A, Ford HR, Stein JE, Kelly LK. Association of Escherichia coli O157:H7 with necrotizing enterocolitis in a full-term infant. Pediatr Surg Int 2009; 25:459–63. [26] Mahjoub-Messai F, Bidet P, Caro V, Diancourt L, Biran V, Aujard Y, et al. Escherichia coli isolates causing bacteremia via gut translocation and urinary tract infection in young infants exhibit different virulence genotypes. J Infect Dis 2011;203:1844–9. [27] Sobieszczańska BA, Duda-Madej AB, Turniak MB, Franiczek R, Kasprzykowska U, Duda AK, et al. Invasive properties, adhesion patterns and phylogroup profiles among Escherichia coli strains isolated from children with inflammatory bowel disease. Adv Clin Exp Med 2012;21:591–9. [28] Mansan-Almeida R, Pereira AL, Giugliano LG. Diffusely adherent Escherichia coli strains isolated from children and adults constitute two different populations. BMC Microbiol 2013;13(22). [29] Rosenthal P, Jennings MT. Comparison of fecal occult blood tests for detection of gastrointestinal bleeding in pediatric patients. Am J Gastroenterol 1992;87:1575–9.
M.B. Said et al. / Early Human Development 90 (2014) 579–585 [30] Li W, Han L, Yu P, Ma C, Wu X, Moore JE, et al. Molecular characterization of skin microbiota between cancer cachexia patients and healthy volunteers. Microb Ecol Apr. 2014;67(3):679–89. [31] Grönlund MM, Lehtonen OP, Eerola E, Kero P. 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(1):19–25. [32] Biasucci G, Rubini M, Riboni S, Morelli L, Bessi E, Retetangos C. Mode of delivery affects the bacterial community in the newborn gut. Early Hum Dev 2010; 86:13–5.
585
[33] Sakata H, Yoshioka H, Fujita K. Development of the intestinal flora in very low birth weight infants compared to normal full-term newborns. Eur J Pediatr 1985; 144(2):186–90. [34] Gewolb IH, Schwalbe RS, Taciak VL, Harrison TS, Panigrahi P. Stool microflora in extremely low birthweight infants. Arch Dis Child Fetal Neonatal Ed 1999;80(3): F167–73. [35] Penders J, Thijs C, Vink C, Stelma FF, Snijders B, Kummeling I, et al. Factors influencing the composition of the intestinal microbiota in early infancy. Pediatrics 2006; 118:511–21.
