ORIGINAL ARTICLE: NUTRITION
Early Gut Colonization of Preterm Infants: Effect of Enteral Feeding Tubes
Marta Go´mez, Laura Moles, yAna Melgar, yNoelia Ureta, yGerardo Bustos, z Leo´nides Ferna´ndez, zJuan M. Rodrı´guez, and zEsther Jime´nez
ABSTRACT Objective: The aim of the study was to evaluate the potential colonization of nosocomial bacteria in enteral feeding systems and its effect on early gut colonization of preterm neonates. Methods: Mother’s own milk, donor milk, and preterm formula samples obtained after passing through the external part of the enteral feeding tubes were cultured. In addition, meconium and fecal samples from 26 preterm infants collected at different time points until discharge were cultured. Random amplification polymorphism DNA and pulse field gel electrophoresis were performed to confirm the presence of specific bacterial strains in milk and infant fecal samples. Results: Approximately 4000 bacterial isolates were identified at the species level. The dominant species in both feces from preterm infants and milk samples were Staphylococcus epidermidis, S aureus, Enterococcus faecalis, E faecium, Serratia marcescens, Klebsiella pneumoniae, and Escherichia coli. All of them were present at high concentrations independently of the feeding mode. Random amplification polymorphism DNA and pulse field gel electrophoresis techniques showed that several bacteria strains were found in both type of samples. Furthermore, scanning electron microscopy revealed the presence of a dense bacterial biofilm in several parts of the feeding tubes and the tube connectors. Conclusions: There is a sharing of bacterial strains between the neonates’ gastrointestinal microbiota and the feeding tubes used to feed them. Key Words: biofilms, hospital-associated bacteria, intestinal microbiota
Received August 11, 2015; accepted December 29, 2015. From the Departamento Nutricio´n, Bromatologı´a y Tecnologı´a de los Alimentos, Universidad Complutense de Madrid, the yServicio de Neonatologı´a y Red Samid, Hospital Universitario 12 de Octubre, and the zProbiSearch, Tres Cantos, Madrid, Spain. Address correspondence and reprint requests to Esther Jime´nez, Departamento de Nutricio´n, Bromatologı´a y Tecnologı´a de los Alimentos. Unviersidad Complutense de Madrid. Av. Puerta de Hierro, s/n. 28040, Madrid, Spain (e-mail: [email protected]). Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.jpgn.org). www.clinicaltrials.gov. registration number: NCT02502916. This work was supported by the projects CSD2007-00063 (FUN-C-FOOD, Consolider-Ingenio 2010) and AGL2013-41980-P from the Ministerio de Economı´a y Competitividad (Spain), and by the project FIS PS09/00040 (Ministerio de Sanidad y Consumo, Spain). Drs Go´mez and Moles contributed equally to the article. M.G. was the recipient of predoctoral fellowship from the Ministerio de Educacio´n, Cultura y Deporte. L.M. was the recipient of predoctoral fellowship from the Ministerio de Economı´a y Competitividad. The other authors report no conflicts of interest. Copyright # 2016 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000001104
JPGN
What Is Known
Preterm birth is associated with an aberrant intestinal colonization pattern. Preterm infants are routinely tube fed for several days until they are ready for sucking. The inner portion of nasogastric enteral feeding tubes has been shown to be colonized by neonatal intensive care unit–associated microorganisms.
What Is New
A thick bacterial biofilm is formed inside the external feeding tube and connectors within 24 hours contributing to the bacterial composition of the milk that passes through it. Sharing of enterobacterial, staphylococcal, and enterococcal strains between milk samples collected after they pass through the feeding system and infant feces was observed.
(JPGN 2016;62: 893–900)
T
he colonization of the infant gastrointestinal tract is an essential process that has important short- and long-term consequences for human health (1). Many factors affect the acquisition, composition, and evolution of the infant gut microbiota, including gestational age, mode of delivery, diet, or medical treatments (2). Preterm infants are known to have an abnormal intestinal colonization pattern during the first weeks of life (3,4), which may lead to increased susceptibility to disease (5–7). Compared with infants born at term, the intestinal microbiota of preterm infants exhibits a significantly reduced bacterial diversity and an abundance of microorganisms usually related to hospital environments (8–10). Breast-feeding is the natural and most recommended way of supporting the growth and development of healthy term infants (11,12). When, for any reason, breast-feeding is not possible and own mother’s milk (OMM) is not available, donor human milk (DM) becomes the next best alternative (11,13). Preterm neonates frequently receive a mixed diet regimen, including alternating OMM, DM, and/or preterm infant formula, depending on their health status, internal hospital management, and availability of human milk during their stay at the neonatal intensive care unit (NICU). Preterm infants are routinely tube fed until they are physiologically ready for the coordination of sucking, swallowing, and breathing, which often occurs at 33 to 36 weeks of postmenstrual
Volume 62, Number 6, June 2016
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
893
Go´mez et al
JPGN
age (14). Therefore, any type of feed must be administered through a feeding device. The inner portion of nasogastric enteral feeding tubes has been shown to be colonized by NICU-associated microorganisms (15–17). As a consequence, any nutritional source passing through the tubes may take along bacteria and have a strong effect on the infant intestinal colonization. In this context, the objective of this work was to evaluate whether the external feeding tube, which is connected with the nasogastric enteral feeding tube through a connecting device, may also act as a site that sustains growth of nosocomial bacteria and thereby affecting the early gut colonization of preterm neonates.
METHODS Subjects and Sampling Thirty-one preterm infants were recruited among those born at the Hospital Universitario 12 de Octubre of Madrid (Spain) from October 2009 to June 2010. The protocol of this prospective study was approved by the local ethics committee (09/157) and written informed parental consent was obtained for each preterm before inclusion. To be eligible for enrolment, preterm infants had to be born at a gestational age of 32 weeks or less or with a birth weight of 1200 g or less. Neonates with any malformation or those experiencing any genetic metabolic disorder were excluded from the study. All of the infants were fed with human milk (OMM and/or DM) and, occasionally, with preterm formula; however, there was a high individual variability in the feeding pattern. That issue had made impossible to distinguish whether the microbiota composition differed depending on the type of milk fed. Mother’s milk was extracted using an electric pump and stored either refrigerated (58C) for maximum of 24 hours or frozen (18 8C) up to 6 months. DM is normally pasteurized (62.58C, 30 minutes) and stored frozen (188C) after collection up to 3 months. The commercial sterilized formula milk used in the hospital is already prepared in individual doses. All milks were warmed during 10–15 minutes at 378C–408C before administration. Syringe barrels used as reservoirs were connected through an external feeding tube (EFT) to the infant’s nasogastric EFT (NEFT) (Fig. 1). Feeding tubes were routinely replaced and discarded every 24 hours, which means that different feed types could pass through the same tube during that period.
Volume 62, Number 6, June 2016
First spontaneously released meconium and weekly fecal samples were collected by the medical staff of the Department of Neonatology of the Hospital from the diapers of the infants during their stay at the NICU. All the samples were stored at 208C until analysis. Routinely nonused diapers were placed inside the incubators to be used as controls. Details regarding culture analysis, bacterial identification, and genetic relatedness among selected bacteria from mother and infants with random amplification polymorphism DNA and pulse field gel electrophoresis, and statistical analysis of microbial composition and clinical data are presented in the supplementary methods document (http://links.lww.com/MPG/A597). Furthermore, 6 different nasogastric enteral feeding tubes, connectors, and external feeding tubes were selected from 6 neonates taking into account numerous conditions that could be involved in microbial growth, such as temperature (cot or incubator), last type of feeding that passed through the tube, time that the tube was placed into the neonate, or infusion rate (gavage or pump). Bacteria present on the internal surface of these devices were analyzed by scanning electron microscopy (SEM) as explained in the supplementary methods document (http://links.lww.com/MPG/A597).
RESULTS Characteristics of the Infants Of the 31 infants recruited, 5 dropped out from the trial: 2 infants died before the meconium release and the parents of 3 infants revoked the consent. Twenty-six infants were included in the study. Their main clinical characteristics are described in Table 1. All of them, except 2, received antibacterial prophylaxis at least for the first 3 days of life. Infants were fed with their OMM, DM, and/or formula by nasogastric enteral feeding tube for, at least, 17 days after delivery.
Microbiological Characterization of OMM, DM, and Preterm Formula After Their Pass Through the External Feeding Tubes The 135 feeding (OMM: 85; DM: 35; and infant formula: 15) samples analyzed in the present study using culture-based methods
Milk sample collection
Own mother’s milk extraction
Refrigeration Frost/Defrost
Pasteurization
*1
Incubation
*2
*3
Formula Reconstitution
Donor milk extraction
Outer part
2m External feeding tube
Inner part
50 cm Nasogastric tube
FIGURE 1. Diagram of the food preparation and the enteral feeding system. A syringe is connected to an external feeding tube of approximately 2 m of length that is connected to a nasogastric tube of 50 cm. Scanning electron microscopy (SEM) images were taken from 1 cm sections marked as 1, external feeding tube, 2 connector, and 3 nasogastric feeding tube.
894
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
JPGN
Volume 62, Number 6, June 2016
Early Gut Colonization of Preterm Infants
TABLE 1. Demographic data and clinical characteristics of preterm infants (n ¼ 26) participating in the present study Characteristics
Mean (95% CI) or no. (%)
Gestational age, wk Sex Male Female Birth weight, g z Score Delivery mode Vaginal section Cesarean section Antibiotherapy No Yes 3 days Sepsis Parenteral nutrition, n ¼ 22, days Enteral feeding tube, days Mechanical ventilation, n ¼ 16, days CPAP, n ¼ 21, days Oxigenotherapy, n ¼ 20 (days) NICU stay, days Hospital stay, days Type of feeding Human milky Mixed feed from 6th week of lifez Mixed feed§
27.7 (26.6;28.7) 13 13 1167.3 0.23
(50%) (50%) (987.3;1347.3) (0.50;0.04)
12 (46%) 14 (54%) 2 24 11 13 7 12 59 23 23 43.5 52 69
(8%) (92%) (42%) (50%) (27%) (5–13) (46;72) (1–36) (12–34) (1–83) (18–81) (41–92)
8 (30.77%) 10 (38.46%) 8 (30.77%)
CI ¼ confidence interval; CPAP ¼ continuous positive airway pressure; NICU ¼ neonatal intensive care unit. Median (interquartile range). y Infants received their own mother’s milk and/or donor milk during all the hospital stay. z Infants fed with their own mother’s milk and/or donor milk exclusively during the first 6 weeks of life and formula milk was introduced after then. § Infants received their own mother’s milk and/or donor milk and/or formula milk from birth.
were the last fraction obtained after their passage through the external feeding tube, immediately before entering the nasogastric tube at the connector level (Fig. 1). The same bacterial profile could be observed in the 3 different feeding types (supplementary Table 1, http://links.lww.com/MPG/A598). Staphylococcus was the genus most frequently isolated from OMM samples (93%) in contrast to DM and formula feeding, in which it was present at 37% and 11%, respectively. Enterococcus was the Gram-positive genus most frequently found in DM and formula samples (49% and 27%, respectively) but it could also be isolated from a high percentage (61%) of OMM samples. In addition, some Gram-negative bacteria, such as Escherichia coli, Klebsiella spp, or Serratia spp, were also isolated from the 3 feeding types (supplementary Table 1, http:// links.lww.com/MPG/A598). Mean bacterial counts of Enterococcus (P ¼ –0.004), Staphylococcus (P < 0.001), and Serratia (P ¼ 0.05) were significantly higher in OMM samples (5.03, 4.82, and 5.58 log10 CFU/mL, respectively) than in the other 2 feeding types (supplementary Table 1, http://links.lww.com/MPG/A598). At the species level, Staphylococcus epidermidis was the dominant species in OMM samples, whereas Enterococcus faecalis was the most abundant in those of DM and infant formula (supplementary Fig. 1, http://links.lww.com/MPG/A599). Serratia marcescens accounted for >23% of the total bacteria isolated from infant formula samples. Three Lactobacillus species (L fermentum,
L gasseri, and L salivarius) and 2 Bifidobacterium species (B longum and B breve) were detected in the present study. Among them, L salivarius, B longum, and B breve were only detected in OMM samples. The diversity and evenness of the microbial communities of the different feed samples after their passage through the external portion of the tubes were determined using the Shannon diversity index. The results obtained showed that the diversity present in OMM (1.15 0.09) samples was higher than that observed in those of DM and infant formula (0.59 0.11 and 0.41 0.06, respectively).
Culture Analysis of the Meconium and Fecal Samples The time required for spontaneous release of the first meconium varied between the first minutes after birth to 6 days of age. A total of 17 meconium and 128 weekly fecal samples were collected during hospitalization of the infants; on average, 6.4 samples per infant were analyzed. Globally, inoculation of suitable dilutions of all the samples led to bacterial growth on the culture media tested, with the exception of 5 meconium samples. The dominant classes in all the samples analyzed were Bacilli and Gammaproteobacteria. Meconium samples were characterized by an absence of Gram-negative bacteria, such as E coli, Klebsiella spp, or Serratia spp (P < 0.001), a higher frequency of Streptococcus spp (P ¼ 0.089) and a lower frequency of Staphylococcus spp (P ¼ 0.047) when compared to the fecal samples (supplementary Table 2, http:// links.lww.com/MPG/A600). Mean counts of all bacterial groups were higher in feces than in meconium samples but, at the genus level the differences were statistically significant for Klebsiella (P ¼ 0.012), Serratia (P ¼ 0.034), and other Gram-negative bacteria (P ¼ 0.049) (supplementary Table 2, http://links.lww.com/MPG/A600). In relation to the succession of the different microbial groups in the fecal samples throughout the infants’ hospitalization, the prevalence of the genus Enterococcus remained high in all the samples, whereas that of Staphylococcus decreased after the first month of life (supplementary Table 2, http://links.lww.com/MPG/ A600). Although the number of tested samples decreased as the number of months increased, the prevalence of Klebsiella was significantly lower (P ¼ 0.003) after the third month of life (data not shown). The number of species detected in the meconium and fecal samples ranged from 1 to 4 and 2 to 9, respectively, increasing with participants’ growth and including up to 5 classes and 12 genera (supplementary Fig. 2, http://links.lww.com/MPG/A601). Despite the interindividual variability observed in the composition of the cultivable fecal microbiota of the infants, E faecalis and S epidermidis were isolated most frequently, being present in 100% and 96% of samples, respectively. Other abundant species belonging to the Class Bacilli were E faecium (77%), L fermentum (42%), and S aureus (38%). In addition, L gasseri and L salivarius were isolated from 8% of the samples. With respect to the Class Gammaproteobacteria, the predominant species were Klebsiella pneumoniae (88%), E coli (81%), S marcescens (81%), Enterobacter cancerogenus (46%), K oxytoca (35%), and E fergusonii (31%). Finally, 6 different genera of the phylum Actinobacteria (Rothia, Micrococcus, Microbacterium, Dermabacter, Corynebacterium, and Bifidobacterium) were detected. Two species of Bifidobacterium, B breve and B longum, could be isolated from a total of 4 infants. The diversity of the microbial communities, assessed using the Shannon-Weaver index, was lower in meconium samples (0.42)
895
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
Go´mez et al
JPGN Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 21 24 25 26
Sample Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces
0 days
7 days
14 days
21 days
28 days
Volume 62, Number 6, June 2016
35 days 42 days 49 days
FIGURE 2. Pulse field gel electrophoresis genotypes isolated from feces and milk samples are shown in the figure. Different figures represent different species (Enterococcus faecalis , E faecium , Staphylococcus aureus , S epidermidis and Klebiella pneumoniae ). Different colors refer to a different pulse field gel electrophoresis profile.
than in the fecal ones (1.19–1.35), whereas the highest microbial diversity was reached after the third month of life (1.35). Potential associations between demographic and clinical parameters and isolation of the different genera were assessed using the Fisher exact test (supplementary Fig. 3, http://links.lww.com/MPG/A602). The isolation of the genus Serratia seemed to be strongly influenced by demographic or clinical variables related to prematurity, whereas the presence of E coli was higher in fecal samples from infants with a lower degree of prematurity (supplementary Fig. 3, http://links.lww.com/MPG/A602).
Genotyping of Milk and Fecal Isolates The isolates belonging to the most frequently detected species in all the samples and sampling points (E faecalis, E faecium, S epidermidis, S aureus, K pneumoniae) were genotyped by random amplification polymorphism DNA and pulse field gel electrophoresis.
Sharing of 34 bacterial strains between milk samples collected after their passage through the feeding system and infant feces was observed at different time points in 23 of the 26 cases (Fig. 2). In most of the cases (26 times), the genotypes were detected simultaneously in milk samples collected after their passage through the feeding system and in infant feces. In addition to this, in 15 occasions the strain was observed first in the milk and at a later time point in the feces, although the opposite occurred in 16 occasions and only 1 time was found 1 strain only in fecal samples. The number of bacterial genotypes shared in each case varied between 1 and 6. The highest strain diversity was detected in the S epidermidis and E faecalis species in which 17 and 9 different genotypes were distinguished, respectively, followed by K pneumoniae with 8 different genotypes. Examples of the genotypes diversity observed in S epidermidis and E faecalis species are shown in supplementary Figure 4, (http://links.lww.com/MPG/ A603). Among the S aureus isolates 4 different genotypes were found and only 1 among the E faecium isolates (Fig. 2).
896
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
JPGN
Volume 62, Number 6, June 2016
Some genotypes of E faecalis, K pneumonia, or S epidermidis were found in different cases suggesting that these strains may have an environmental origin (Fig. 2). In fact, one of the genotypes of E faecalis (orange circle in Fig. 2) was found in 9 different cases and a different one (blue circle in Fig. 2) in 4 cases. Additionally, one genotype of K. pneumonia (pink cross in Fig. 2) was detected in 3 cases as well as one genotype of S. epidermidis (maroon diamond in Fig. 2).
SEM Analysis of the Nasogastric Enteral Feeding Tubes The selected parts of 6 external feeding tubes with their respective connectors and nasogastric enteral feeding tubes were analyzed by SEM. The time that the tubes were placed into a preterm infant was the factor that exerted the highest influence on bacterial growth. Thick bacterial biofilms were observed inside the external feeding tube (Fig. 3A–C) and connectors (Fig. 3D–F) that were used for 24 hours and they seemed to be particularly complex in nasogastric enteral feeding tubes that were used for >48 hours (Fig. 3G–J). In contrast, only milk residues (but no bacteria) could be observed in the inner surfaces of those nasogastric enteral feeding tubes that were placed for
Early Gut Colonization of Preterm Infants: Effect of Enteral Feeding Tubes
Marta Go´mez, Laura Moles, yAna Melgar, yNoelia Ureta, yGerardo Bustos, z Leo´nides Ferna´ndez, zJuan M. Rodrı´guez, and zEsther Jime´nez
ABSTRACT Objective: The aim of the study was to evaluate the potential colonization of nosocomial bacteria in enteral feeding systems and its effect on early gut colonization of preterm neonates. Methods: Mother’s own milk, donor milk, and preterm formula samples obtained after passing through the external part of the enteral feeding tubes were cultured. In addition, meconium and fecal samples from 26 preterm infants collected at different time points until discharge were cultured. Random amplification polymorphism DNA and pulse field gel electrophoresis were performed to confirm the presence of specific bacterial strains in milk and infant fecal samples. Results: Approximately 4000 bacterial isolates were identified at the species level. The dominant species in both feces from preterm infants and milk samples were Staphylococcus epidermidis, S aureus, Enterococcus faecalis, E faecium, Serratia marcescens, Klebsiella pneumoniae, and Escherichia coli. All of them were present at high concentrations independently of the feeding mode. Random amplification polymorphism DNA and pulse field gel electrophoresis techniques showed that several bacteria strains were found in both type of samples. Furthermore, scanning electron microscopy revealed the presence of a dense bacterial biofilm in several parts of the feeding tubes and the tube connectors. Conclusions: There is a sharing of bacterial strains between the neonates’ gastrointestinal microbiota and the feeding tubes used to feed them. Key Words: biofilms, hospital-associated bacteria, intestinal microbiota
Received August 11, 2015; accepted December 29, 2015. From the Departamento Nutricio´n, Bromatologı´a y Tecnologı´a de los Alimentos, Universidad Complutense de Madrid, the yServicio de Neonatologı´a y Red Samid, Hospital Universitario 12 de Octubre, and the zProbiSearch, Tres Cantos, Madrid, Spain. Address correspondence and reprint requests to Esther Jime´nez, Departamento de Nutricio´n, Bromatologı´a y Tecnologı´a de los Alimentos. Unviersidad Complutense de Madrid. Av. Puerta de Hierro, s/n. 28040, Madrid, Spain (e-mail: [email protected]). Supplemental digital content is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.jpgn.org). www.clinicaltrials.gov. registration number: NCT02502916. This work was supported by the projects CSD2007-00063 (FUN-C-FOOD, Consolider-Ingenio 2010) and AGL2013-41980-P from the Ministerio de Economı´a y Competitividad (Spain), and by the project FIS PS09/00040 (Ministerio de Sanidad y Consumo, Spain). Drs Go´mez and Moles contributed equally to the article. M.G. was the recipient of predoctoral fellowship from the Ministerio de Educacio´n, Cultura y Deporte. L.M. was the recipient of predoctoral fellowship from the Ministerio de Economı´a y Competitividad. The other authors report no conflicts of interest. Copyright # 2016 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology, Hepatology, and Nutrition DOI: 10.1097/MPG.0000000000001104
JPGN
What Is Known
Preterm birth is associated with an aberrant intestinal colonization pattern. Preterm infants are routinely tube fed for several days until they are ready for sucking. The inner portion of nasogastric enteral feeding tubes has been shown to be colonized by neonatal intensive care unit–associated microorganisms.
What Is New
A thick bacterial biofilm is formed inside the external feeding tube and connectors within 24 hours contributing to the bacterial composition of the milk that passes through it. Sharing of enterobacterial, staphylococcal, and enterococcal strains between milk samples collected after they pass through the feeding system and infant feces was observed.
(JPGN 2016;62: 893–900)
T
he colonization of the infant gastrointestinal tract is an essential process that has important short- and long-term consequences for human health (1). Many factors affect the acquisition, composition, and evolution of the infant gut microbiota, including gestational age, mode of delivery, diet, or medical treatments (2). Preterm infants are known to have an abnormal intestinal colonization pattern during the first weeks of life (3,4), which may lead to increased susceptibility to disease (5–7). Compared with infants born at term, the intestinal microbiota of preterm infants exhibits a significantly reduced bacterial diversity and an abundance of microorganisms usually related to hospital environments (8–10). Breast-feeding is the natural and most recommended way of supporting the growth and development of healthy term infants (11,12). When, for any reason, breast-feeding is not possible and own mother’s milk (OMM) is not available, donor human milk (DM) becomes the next best alternative (11,13). Preterm neonates frequently receive a mixed diet regimen, including alternating OMM, DM, and/or preterm infant formula, depending on their health status, internal hospital management, and availability of human milk during their stay at the neonatal intensive care unit (NICU). Preterm infants are routinely tube fed until they are physiologically ready for the coordination of sucking, swallowing, and breathing, which often occurs at 33 to 36 weeks of postmenstrual
Volume 62, Number 6, June 2016
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
893
Go´mez et al
JPGN
age (14). Therefore, any type of feed must be administered through a feeding device. The inner portion of nasogastric enteral feeding tubes has been shown to be colonized by NICU-associated microorganisms (15–17). As a consequence, any nutritional source passing through the tubes may take along bacteria and have a strong effect on the infant intestinal colonization. In this context, the objective of this work was to evaluate whether the external feeding tube, which is connected with the nasogastric enteral feeding tube through a connecting device, may also act as a site that sustains growth of nosocomial bacteria and thereby affecting the early gut colonization of preterm neonates.
METHODS Subjects and Sampling Thirty-one preterm infants were recruited among those born at the Hospital Universitario 12 de Octubre of Madrid (Spain) from October 2009 to June 2010. The protocol of this prospective study was approved by the local ethics committee (09/157) and written informed parental consent was obtained for each preterm before inclusion. To be eligible for enrolment, preterm infants had to be born at a gestational age of 32 weeks or less or with a birth weight of 1200 g or less. Neonates with any malformation or those experiencing any genetic metabolic disorder were excluded from the study. All of the infants were fed with human milk (OMM and/or DM) and, occasionally, with preterm formula; however, there was a high individual variability in the feeding pattern. That issue had made impossible to distinguish whether the microbiota composition differed depending on the type of milk fed. Mother’s milk was extracted using an electric pump and stored either refrigerated (58C) for maximum of 24 hours or frozen (18 8C) up to 6 months. DM is normally pasteurized (62.58C, 30 minutes) and stored frozen (188C) after collection up to 3 months. The commercial sterilized formula milk used in the hospital is already prepared in individual doses. All milks were warmed during 10–15 minutes at 378C–408C before administration. Syringe barrels used as reservoirs were connected through an external feeding tube (EFT) to the infant’s nasogastric EFT (NEFT) (Fig. 1). Feeding tubes were routinely replaced and discarded every 24 hours, which means that different feed types could pass through the same tube during that period.
Volume 62, Number 6, June 2016
First spontaneously released meconium and weekly fecal samples were collected by the medical staff of the Department of Neonatology of the Hospital from the diapers of the infants during their stay at the NICU. All the samples were stored at 208C until analysis. Routinely nonused diapers were placed inside the incubators to be used as controls. Details regarding culture analysis, bacterial identification, and genetic relatedness among selected bacteria from mother and infants with random amplification polymorphism DNA and pulse field gel electrophoresis, and statistical analysis of microbial composition and clinical data are presented in the supplementary methods document (http://links.lww.com/MPG/A597). Furthermore, 6 different nasogastric enteral feeding tubes, connectors, and external feeding tubes were selected from 6 neonates taking into account numerous conditions that could be involved in microbial growth, such as temperature (cot or incubator), last type of feeding that passed through the tube, time that the tube was placed into the neonate, or infusion rate (gavage or pump). Bacteria present on the internal surface of these devices were analyzed by scanning electron microscopy (SEM) as explained in the supplementary methods document (http://links.lww.com/MPG/A597).
RESULTS Characteristics of the Infants Of the 31 infants recruited, 5 dropped out from the trial: 2 infants died before the meconium release and the parents of 3 infants revoked the consent. Twenty-six infants were included in the study. Their main clinical characteristics are described in Table 1. All of them, except 2, received antibacterial prophylaxis at least for the first 3 days of life. Infants were fed with their OMM, DM, and/or formula by nasogastric enteral feeding tube for, at least, 17 days after delivery.
Microbiological Characterization of OMM, DM, and Preterm Formula After Their Pass Through the External Feeding Tubes The 135 feeding (OMM: 85; DM: 35; and infant formula: 15) samples analyzed in the present study using culture-based methods
Milk sample collection
Own mother’s milk extraction
Refrigeration Frost/Defrost
Pasteurization
*1
Incubation
*2
*3
Formula Reconstitution
Donor milk extraction
Outer part
2m External feeding tube
Inner part
50 cm Nasogastric tube
FIGURE 1. Diagram of the food preparation and the enteral feeding system. A syringe is connected to an external feeding tube of approximately 2 m of length that is connected to a nasogastric tube of 50 cm. Scanning electron microscopy (SEM) images were taken from 1 cm sections marked as 1, external feeding tube, 2 connector, and 3 nasogastric feeding tube.
894
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
JPGN
Volume 62, Number 6, June 2016
Early Gut Colonization of Preterm Infants
TABLE 1. Demographic data and clinical characteristics of preterm infants (n ¼ 26) participating in the present study Characteristics
Mean (95% CI) or no. (%)
Gestational age, wk Sex Male Female Birth weight, g z Score Delivery mode Vaginal section Cesarean section Antibiotherapy No Yes 3 days Sepsis Parenteral nutrition, n ¼ 22, days Enteral feeding tube, days Mechanical ventilation, n ¼ 16, days CPAP, n ¼ 21, days Oxigenotherapy, n ¼ 20 (days) NICU stay, days Hospital stay, days Type of feeding Human milky Mixed feed from 6th week of lifez Mixed feed§
27.7 (26.6;28.7) 13 13 1167.3 0.23
(50%) (50%) (987.3;1347.3) (0.50;0.04)
12 (46%) 14 (54%) 2 24 11 13 7 12 59 23 23 43.5 52 69
(8%) (92%) (42%) (50%) (27%) (5–13) (46;72) (1–36) (12–34) (1–83) (18–81) (41–92)
8 (30.77%) 10 (38.46%) 8 (30.77%)
CI ¼ confidence interval; CPAP ¼ continuous positive airway pressure; NICU ¼ neonatal intensive care unit. Median (interquartile range). y Infants received their own mother’s milk and/or donor milk during all the hospital stay. z Infants fed with their own mother’s milk and/or donor milk exclusively during the first 6 weeks of life and formula milk was introduced after then. § Infants received their own mother’s milk and/or donor milk and/or formula milk from birth.
were the last fraction obtained after their passage through the external feeding tube, immediately before entering the nasogastric tube at the connector level (Fig. 1). The same bacterial profile could be observed in the 3 different feeding types (supplementary Table 1, http://links.lww.com/MPG/A598). Staphylococcus was the genus most frequently isolated from OMM samples (93%) in contrast to DM and formula feeding, in which it was present at 37% and 11%, respectively. Enterococcus was the Gram-positive genus most frequently found in DM and formula samples (49% and 27%, respectively) but it could also be isolated from a high percentage (61%) of OMM samples. In addition, some Gram-negative bacteria, such as Escherichia coli, Klebsiella spp, or Serratia spp, were also isolated from the 3 feeding types (supplementary Table 1, http:// links.lww.com/MPG/A598). Mean bacterial counts of Enterococcus (P ¼ –0.004), Staphylococcus (P < 0.001), and Serratia (P ¼ 0.05) were significantly higher in OMM samples (5.03, 4.82, and 5.58 log10 CFU/mL, respectively) than in the other 2 feeding types (supplementary Table 1, http://links.lww.com/MPG/A598). At the species level, Staphylococcus epidermidis was the dominant species in OMM samples, whereas Enterococcus faecalis was the most abundant in those of DM and infant formula (supplementary Fig. 1, http://links.lww.com/MPG/A599). Serratia marcescens accounted for >23% of the total bacteria isolated from infant formula samples. Three Lactobacillus species (L fermentum,
L gasseri, and L salivarius) and 2 Bifidobacterium species (B longum and B breve) were detected in the present study. Among them, L salivarius, B longum, and B breve were only detected in OMM samples. The diversity and evenness of the microbial communities of the different feed samples after their passage through the external portion of the tubes were determined using the Shannon diversity index. The results obtained showed that the diversity present in OMM (1.15 0.09) samples was higher than that observed in those of DM and infant formula (0.59 0.11 and 0.41 0.06, respectively).
Culture Analysis of the Meconium and Fecal Samples The time required for spontaneous release of the first meconium varied between the first minutes after birth to 6 days of age. A total of 17 meconium and 128 weekly fecal samples were collected during hospitalization of the infants; on average, 6.4 samples per infant were analyzed. Globally, inoculation of suitable dilutions of all the samples led to bacterial growth on the culture media tested, with the exception of 5 meconium samples. The dominant classes in all the samples analyzed were Bacilli and Gammaproteobacteria. Meconium samples were characterized by an absence of Gram-negative bacteria, such as E coli, Klebsiella spp, or Serratia spp (P < 0.001), a higher frequency of Streptococcus spp (P ¼ 0.089) and a lower frequency of Staphylococcus spp (P ¼ 0.047) when compared to the fecal samples (supplementary Table 2, http:// links.lww.com/MPG/A600). Mean counts of all bacterial groups were higher in feces than in meconium samples but, at the genus level the differences were statistically significant for Klebsiella (P ¼ 0.012), Serratia (P ¼ 0.034), and other Gram-negative bacteria (P ¼ 0.049) (supplementary Table 2, http://links.lww.com/MPG/A600). In relation to the succession of the different microbial groups in the fecal samples throughout the infants’ hospitalization, the prevalence of the genus Enterococcus remained high in all the samples, whereas that of Staphylococcus decreased after the first month of life (supplementary Table 2, http://links.lww.com/MPG/ A600). Although the number of tested samples decreased as the number of months increased, the prevalence of Klebsiella was significantly lower (P ¼ 0.003) after the third month of life (data not shown). The number of species detected in the meconium and fecal samples ranged from 1 to 4 and 2 to 9, respectively, increasing with participants’ growth and including up to 5 classes and 12 genera (supplementary Fig. 2, http://links.lww.com/MPG/A601). Despite the interindividual variability observed in the composition of the cultivable fecal microbiota of the infants, E faecalis and S epidermidis were isolated most frequently, being present in 100% and 96% of samples, respectively. Other abundant species belonging to the Class Bacilli were E faecium (77%), L fermentum (42%), and S aureus (38%). In addition, L gasseri and L salivarius were isolated from 8% of the samples. With respect to the Class Gammaproteobacteria, the predominant species were Klebsiella pneumoniae (88%), E coli (81%), S marcescens (81%), Enterobacter cancerogenus (46%), K oxytoca (35%), and E fergusonii (31%). Finally, 6 different genera of the phylum Actinobacteria (Rothia, Micrococcus, Microbacterium, Dermabacter, Corynebacterium, and Bifidobacterium) were detected. Two species of Bifidobacterium, B breve and B longum, could be isolated from a total of 4 infants. The diversity of the microbial communities, assessed using the Shannon-Weaver index, was lower in meconium samples (0.42)
895
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
Go´mez et al
JPGN Case 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 21 24 25 26
Sample Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces Milk Feces
0 days
7 days
14 days
21 days
28 days
Volume 62, Number 6, June 2016
35 days 42 days 49 days
FIGURE 2. Pulse field gel electrophoresis genotypes isolated from feces and milk samples are shown in the figure. Different figures represent different species (Enterococcus faecalis , E faecium , Staphylococcus aureus , S epidermidis and Klebiella pneumoniae ). Different colors refer to a different pulse field gel electrophoresis profile.
than in the fecal ones (1.19–1.35), whereas the highest microbial diversity was reached after the third month of life (1.35). Potential associations between demographic and clinical parameters and isolation of the different genera were assessed using the Fisher exact test (supplementary Fig. 3, http://links.lww.com/MPG/A602). The isolation of the genus Serratia seemed to be strongly influenced by demographic or clinical variables related to prematurity, whereas the presence of E coli was higher in fecal samples from infants with a lower degree of prematurity (supplementary Fig. 3, http://links.lww.com/MPG/A602).
Genotyping of Milk and Fecal Isolates The isolates belonging to the most frequently detected species in all the samples and sampling points (E faecalis, E faecium, S epidermidis, S aureus, K pneumoniae) were genotyped by random amplification polymorphism DNA and pulse field gel electrophoresis.
Sharing of 34 bacterial strains between milk samples collected after their passage through the feeding system and infant feces was observed at different time points in 23 of the 26 cases (Fig. 2). In most of the cases (26 times), the genotypes were detected simultaneously in milk samples collected after their passage through the feeding system and in infant feces. In addition to this, in 15 occasions the strain was observed first in the milk and at a later time point in the feces, although the opposite occurred in 16 occasions and only 1 time was found 1 strain only in fecal samples. The number of bacterial genotypes shared in each case varied between 1 and 6. The highest strain diversity was detected in the S epidermidis and E faecalis species in which 17 and 9 different genotypes were distinguished, respectively, followed by K pneumoniae with 8 different genotypes. Examples of the genotypes diversity observed in S epidermidis and E faecalis species are shown in supplementary Figure 4, (http://links.lww.com/MPG/ A603). Among the S aureus isolates 4 different genotypes were found and only 1 among the E faecium isolates (Fig. 2).
896
www.jpgn.org
Copyright © ESPGHAL and NASPGHAN. All rights reserved.
JPGN
Volume 62, Number 6, June 2016
Some genotypes of E faecalis, K pneumonia, or S epidermidis were found in different cases suggesting that these strains may have an environmental origin (Fig. 2). In fact, one of the genotypes of E faecalis (orange circle in Fig. 2) was found in 9 different cases and a different one (blue circle in Fig. 2) in 4 cases. Additionally, one genotype of K. pneumonia (pink cross in Fig. 2) was detected in 3 cases as well as one genotype of S. epidermidis (maroon diamond in Fig. 2).
SEM Analysis of the Nasogastric Enteral Feeding Tubes The selected parts of 6 external feeding tubes with their respective connectors and nasogastric enteral feeding tubes were analyzed by SEM. The time that the tubes were placed into a preterm infant was the factor that exerted the highest influence on bacterial growth. Thick bacterial biofilms were observed inside the external feeding tube (Fig. 3A–C) and connectors (Fig. 3D–F) that were used for 24 hours and they seemed to be particularly complex in nasogastric enteral feeding tubes that were used for >48 hours (Fig. 3G–J). In contrast, only milk residues (but no bacteria) could be observed in the inner surfaces of those nasogastric enteral feeding tubes that were placed for
