Advertisement

Cluster of necrotizing enterocolitis in a neonatal intensive care unit: New Mexico, 2007

  • Aaron M. Wendelboe
    Correspondence
    Address correspondence to Aaron Wendelboe, PhD, assistant professor, Biostatistics and Epidemiology, University of Oklahoma Health Sciences Center, 801 NE 13th St, CHB 323, Oklahoma City, OK 73104.
    Affiliations
    Centers for Disease Control and Prevention, Office of Workforce and Career Development, EIS Field Assignments Branch, Atlanta, GA

    New Mexico Department of Health, Epidemiology and Response Division, Santa Fe, NM
    Search for articles by this author
  • Chad Smelser
    Affiliations
    New Mexico Department of Health, Epidemiology and Response Division, Santa Fe, NM
    Search for articles by this author
  • Cynthia A. Lucero
    Affiliations
    Centers for Disease Control and Prevention, National Center for Preparedness, Detection, and Control of Infectious Diseases, Division of Healthcare Quality Promotion, Atlanta, GA
    Search for articles by this author
  • L. Clifford McDonald
    Affiliations
    Centers for Disease Control and Prevention, National Center for Preparedness, Detection, and Control of Infectious Diseases, Division of Healthcare Quality Promotion, Atlanta, GA
    Search for articles by this author
Published:October 13, 2009DOI:https://doi.org/10.1016/j.ajic.2009.06.009

      Background

      Although the cause of necrotizing enterocolitis (NEC) is unknown, infection control practices have been shown to play an important role in containing many outbreaks. We investigated the etiology of a cluster of NEC in a level 3 neonatal intensive care unit and monitored for new cases following the implementation of enhanced infection control measures.

      Methods

      Investigators performed a chart and laboratory review for neonates with a diagnosis of NEC during January 1, 2007, to February 13, 2007, to identify risk factors. Enhanced environmental cleaning, cohorting of infants and nurses, and increased attention to hand hygiene were instituted. Commercial feeding products in the unit were tested for bacterial contamination. Close monitoring for new cases continued for 2 months following the identification of the cluster.

      Results

      Eleven cases of NEC were identified during the study period. Patients had a median of 5 disease risk factors (range, 3-8). Four distinct pathogens were detected in blood or stool specimens from 4 different patients. One sample of human milk fortifier (HMF) tested contained a colony count of Bacillus cereus at the US Food and Drug Administration's upper microbiologic limit for contamination. Seven (65%) patients received HMF before symptom onset, and 9 (82%) patients received 1 or more types of liquid formula. Only 1 new case was identified during the period of close monitoring.

      Conclusion

      A microbiologic cause was not identified, and, although the cluster might have resolved spontaneously, enhanced infection control and changing batches of HMF might have played a role in controlling this outbreak.

      Key Words

      Necrotizing enterocolitis (NEC) is one of the most common serious gastrointestinal diseases among newborns
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      and is a major cause of morbidity and mortality among preterm infants.
      • Neu J.
      Neonatal necrotizing enterocolitis: an update.
      Incidence of NEC nationally in neonatal intensive care units (NICUs) has ranged from 1 per 100 infants to 5 per 100 infants,
      • Stoll B.J.
      Epidemiology of necrotizing enterocolitis.
      • Guthrie S.O.
      • Gordon P.V.
      • Thomas V.
      • Thorp J.A.
      • Peabody J.
      • Clark R.H.
      Necrotizing enterocolitis among neonates in the United States.
      and the case fatality rate has been estimated to range from 10% to 30%.
      • Caplan M.S.
      • Jilling T.
      New concepts in necrotizing enterocolitis.
      Although the pathogenesis is not understood, it is likely multifactorial (eg, prematurity, aggressive feeding regimens, infection, and inflammation).
      • Neu J.
      Neonatal necrotizing enterocolitis: an update.
      • Stoll B.J.
      Epidemiology of necrotizing enterocolitis.
      Neonates experiencing NEC typically present with intestinal symptoms (including feeding intolerance, bloody stool, and abdominal distention), systemic symptoms (including apnea, bradycardia, temperature instability, and septic shock), and characteristic radiographic findings of pneumatosis intestinalis. Limited progress has been achieved in delineating the etiology of NEC because both older
      • Stoll B.J.
      Epidemiology of necrotizing enterocolitis.
      and recent
      • Neu J.
      Neonatal necrotizing enterocolitis: an update.
      articles have described its enigmatic etiology. The majority of cases is sporadic; however, the occurrence of clusters indicates an infectious component to the disease.
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      Multiple causative organisms have been proposed, including coagulase-negative Staphylococcus species, various gram-negative bacilli, and Clostridium species. However, in approximately 35% of published outbreaks, no etiologic organism was identified.
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      This report summarizes an investigation of a NEC cluster among 11 premature infants in a NICU in New Mexico from January 22 to February 13, 2007. Personnel from the New Mexico Department of Health (NMDOH), the hospital in which the cluster occurred, and the Centers for Disease Control and Prevention conducted an investigation to identify the cause of this cluster, evaluate risk factors associated with development of NEC, and implement appropriate control measures.

      Materials and methods

      This cluster of NEC was identified by the physicians on February 5, 2007, after the first 5 cases were diagnosed. Assistance from NMDOH was requested on February 12, 2007. Cases were defined according to modified Bell's staging.
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      • Walsh M.C.
      • Kliegman R.M.
      Necrotizing enterocolitis: treatment based on staging criteria.
      (In 1978, Bell et al
      • Bell M.J.
      • Ternberg J.L.
      • Feigin R.D.
      • Keating J.P.
      • Marshall R.
      • Barton L.
      • et al.
      Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging.
      developed a uniform staging system for NEC, which was modified by Walsh et al
      • Walsh M.C.
      • Kliegman R.M.
      Necrotizing enterocolitis: treatment based on staging criteria.
      in 1986 to more explicitly describe the systemic, intestinal, and radiographic signs.) Annual rates of NEC at this hospital were calculated for 2005, 2006, and 2007. Medical charts for each case-patient were reviewed, and the following information was collected based on factors identified from previous reports of NEC clusters and the judgment of the clinical staff and investigative team: date of birth; estimated gestational age; birth weight; sex; admission date; APGAR score; date of NEC onset; NEC staging and accompanying signs and symptoms; previous diagnosis of NEC; detailed feeding history for ≤7 days before symptom onset; antimicrobial therapy and other pharmaceutical treatments; umbilical catheter; surgical treatment; hyaline membrane disease; and maternal risk factors, including prenatal care, gestational diabetes, drug abuse, and preeclampsia. Information regarding bed location and number of nurses and other health care providers caring for case-patients during the 2 days before symptom onset was also collected. Enhanced surveillance for NEC was conducted for 2 months following the last case of the cluster.
      All clinical samples were processed through the hospital laboratory. Blood cultures were ordered for each case-patient, and, after the outbreak was detected, stool was tested for the following pathogens: Clostridium difficile via toxin test (n=7), norovirus (n=7), rotavirus (n=6), and other enteric pathogens (n=6). Information regarding test results and their accompanying dates was collected from hospital medical records. Anaerobic testing of stool or environmental samples was not available through the New Mexico Scientific Laboratory Division or local private laboratories limiting the ability to identify anaerobic pathogens.
      When the cluster was identified, the nursing and dietary staff immediately isolated all remaining commercial feeding products for possible testing. Human milk fortifier (HMF) (Enfamil; Mead Johnson, Glenview, IL) and single-use liquid formula with and without iron in 20- and 24-calories/oz formulations were among the isolated products and were sent to New Mexico Scientific Laboratory Division for aerobic bacterial contamination testing. HMF, used to increase levels of nutrients and calories in expressed breast milk, was available in single-use, 0.71-g packets of dry powder. Because of the limited volume in each HMF packet, contents from multiple packets sharing the same prefix lot code were combined to produce sufficient volume for testing. Microsoft Excel, 2003, (Microsoft Corp, Redmond, WA) was used for all data management, descriptive analyses, and cluster analysis.

      Results

      The rate of NEC in this NICU during the outbreak period of January 22 to February 13, 2007, was 16.9 of 100 infants, compared with 3.3 of 100 infants and 2.4 of 100 infants in 2006 and 2005, respectively. (The annual rate of NEC decreased to 3.4 of 100 infants by the end of 2007 in this NICU.) The median estimated gestational age of case-patients was 33 weeks (range, 27-37 weeks), and the median time from birth to NEC onset was 8 days (range, 5-81 days). Seven case-patients were male, and 4 were female. Distribution of staging was as follows: 3 stage I (suspect), 5 stage II (definite), and 3 stage III (advanced). All patients experiencing advanced NEC required surgical resection, and 1 died. No other infants died.
      Health care providers maintained a heightened index of suspicion (increased awareness of Bell's stage I criteria) for NEC for 2 months from the last identified patient. One case of NEC was identified 22 days (March 7) after the last case associated with the cluster. Although data were collected regarding this infant (by using 2 months of surveillance data), we decided to exclude it from the cluster under investigation because of the relatively substantial time between the preceding cluster and this case. The epidemic curve summarizing the time line for this outbreak is depicted in Fig 1, stratified by NEC staging. Each patient had multiple, previously recognized risk factors for NEC (Table 1). Patients had a median of 5 (range, 3-8) risk factors for NEC. The median 1-minute (and 5 minute) APGAR scores for stages I, II, and III were 7 (8), 7 (9), and 6 (8), respectively. No appreciable patterns were detected regarding antibiotic use or other medications prescribed. Three of the infants' mothers (1 from each stage) had <5 prenatal care visits. A total of 48 nurses cared for the patients in the 2 days before diagnosis. No single nurse or physician was identified as a common caregiver to all case-patients. No other health care providers were identified as potential common sources. Case-patients were distributed throughout the NICU.
      Figure thumbnail gr1
      Fig 1Number of necrotizing enterocolitis (NEC) cases (n=11) by stage in a neonatal intensive care unit in New Mexico, January-March 2007.
      Table 1Risk factors for necrotizing enterocolitis patients by stage in a neonatal intensive care unit in New Mexico, January 22 to February 13, 2007
      Stage I (n=3)Stage II (n=5)Stage III (n=3)
      Potential risk factors for NECn%n%n%
      Infant risk factors
       Human milk fortifier004803100
       Liquid formula26751002100
       Very low birth weight (<1500 g)2672403100
       Very low gestational age (≤32 wk)1331203100
       Small for gestational age13312000
       Umbilical catheter2672403100
       Hyaline membrane disease1331203100
       Positive blood culture00120266
       Positive C difficile toxin test0012000
       Previous NEC episode1330000
      Maternal risk factors
       Gestational diabetes13324000
       Preeclampsia13324000
       Drug abuse00120133
      NEC, necrotizing enterocolitis.
      NOTE. Number of identified NEC patients: n=11.
      Three different organisms (Klebsiella pneumonia, Staphylococcus aureus, and Enterobacter hormaechei) were isolated from blood cultures from 3 different patients, and C difficile toxin was detected in a fourth patient. No stool cultures were positive for enteric pathogens.
      Infant formula and HMF lot codes were not routinely recorded or tracked by the NICU staff. Seven (64%) of the case-patients received HMF before symptom onset, and 9 (82%) affected infants received 1 or more types of liquid formula. One affected infant was fed an amino acid-based formula and was not exposed to HMF or standard formula.
      Of the 4 lot codes of HMF tested, 1 contained 100 colony-forming units (cfu)/g of Bacillus cereus, 1 contained 10 cfu/g of B cereus, and 2 contained <10 cfu/g of B cereus. No other bacteria were recovered from HMF, and all single-use liquid formula sampled tested negative for bacterial contamination. None of the infant blood specimens tested positive for B cereus, and stool samples were not specifically cultured for B cereus because it is not part of the standard microbiologic panel and there was no stool available for additional testing once the environmental tests were performed. The finding of 100 cfu/g of B cereus in 1 sample is at the US Food and Drug Administration's (FDA) upper microbiologic limit for healthy term infants. No limit is available for preterm infants.
      We recommended that patients be placed under contact isolation and reemphasized the need for strict adherence to hand hygiene and appropriate use of personal protective equipment for both health care personnel and visitors. A 1-time thorough environmental cleaning of the entire NICU was performed with a 1:10 dilution of household bleach (ie, 5000 parts/million of hypochlorite), and infants and nurses cohorted when possible. The hospital also elected to change to new lot codes of HMF and formula. Only 1 isolated case of NEC (previously mentioned) was detected 3 weeks following the implementation of these infection control measures; however, the NEC cluster possibly would have spontaneously resolved regardless.

      Discussion

      A cluster of NEC in a New Mexico NICU was detected and investigated. No single pathogen was identified, a fact that might not be surprising given the likely multifactorial etiology of NEC. NICU staff initiated the following additional infection control measures: enhanced cleaning of all surfaces and instruments by using a sporicidal disinfectant, increased attention to hand hygiene, isolating patients until outbreak termination, staff cohorting, and increased index of suspicion for new cases.
      This report highlights the need for systematic diagnostic testing to cover the spectrum of potential causal organisms when investigating nosocomial outbreaks of NEC, especially for those organisms not routinely isolated by standard blood or stool culture (eg, anaerobic bacteria and B cereus). Three blood cultures were positive with 3 different organisms (K pneumonia, S aureus, and E hormaechei). This finding would not be unusual for sporadic cases of NEC, but, given the clustering of these cases, a common exposure increasing the risk of sepsis with the identified organisms is possible. Although comprehensive testing might not always be necessary in sporadic cases of NEC, it is critical in establishing the etiology when a substantial cluster or outbreak is identified. Specifically, testing for anaerobic pathogens (eg, Clostridium species), coronavirus, rotavirus, and echovirus might be added to the initial set of diagnostic tests because they have been implicated in past outbreaks.
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      Unfortunately, systematic testing was not performed on all infants in this cluster. Echovirus and coronavirus testing were not performed, stool cultures were not ordered for the first 5 patients, and tissue samples were not available. The only testing for Clostridium species consisted of C difficile toxin testing. The clinical significance of a positive C difficile toxin test in this age group is uncertain.
      The potential role of HMF in this outbreak cannot be excluded. Specifically, prematurity is a widely reported risk factor for NEC,
      • Stoll B.J.
      Epidemiology of necrotizing enterocolitis.
      which might be associated with prematurity of the gastrointestinal tract.
      • Neu J.
      Neonatal necrotizing enterocolitis: an update.
      • Stoll B.J.
      Epidemiology of necrotizing enterocolitis.
      • Bjornvad C.R.
      • Schmidt M.
      • Petersen Y.M.
      • Jensen S.K.
      • Offenberg H.
      • Elnif J.
      • et al.
      Preterm birth makes the immature intestine sensitive to feeding-induced intestinal atrophy.
      In addition, formula-fed infants are reported to have 6 to 20 times the risk of experiencing NEC compared with breast milk-fed infants.
      • McGuire W.
      • Anthony M.Y.
      Donor human milk versus formula for preventing necrotising enterocolitis in preterm infants: systematic review.
      • Quigley M.A.
      • Henderson G.
      • Anthony M.Y.
      • McGuire W.
      Formula milk versus donor breast milk for feeding preterm or low birth weight infants.
      • Lambert D.K.
      • Christensen R.D.
      • Henry E.
      • Besner G.E.
      • Bael V.L.
      • Viedmeier S.E.
      • et al.
      Necrotizing enterocolitis in term neonates: data from a multihospital health-care system.
      An increased risk of NEC among preterm infants who were fed HMF has also been reported,
      • Hallstrom M.
      • Koivisto A.M.
      • Janas M.
      • Tammela O.
      Frequency of and risk factors for necrotizing enterocolitis in infants born before 33 weeks of gestation.
      • Lucas A.
      • Fewtrell M.S.
      • Morley R.
      • Lucas P.J.
      • Baker B.A.
      • Lister G.
      • et al.
      Randomized outcome trial of human milk fortification and developmental outcome in preterm infants.
      and contamination
      • CDC
      Enterobacter sakazakii infections associated with the use of powdered infant formula—Tennessee, 2001.
      (including with B cereus
      • Adler A.
      • Gottesman G.
      • Dolfin T.
      • Arnon S.
      • Regev R.
      • Bauer S.
      • et al.
      Bacillus species sepsis in the neonatal intensive care unit.
      • Girisch M.
      • Ries M.
      • Zenker M.
      • Carbon R.
      • Rauch R.
      • Hofbeck M.
      Intestinal perforations in a premature infant caused by Bacillus cereus.
      ) has been postulated as a potential mechanism. In this investigation, 1 lot code of sequestered HMF tested positive for B cereus at the FDA's upper microbiologic limit (100 cfu/g).

      US Food and Drug Administration. Chapter 21. Food composition, standards, labeling and economics: compliance program guidance manual. Available at: http://www.cfsan.fda.gov/∼acrobat/cp21006.pdf. Accessed May 20, 2009.

      Risk assessment models have concluded that powdered infant formula containing ≤100 cfu/g of B cereus and reconstituted with cooled boiled water (25 °C) and stored for ≤24 hours at ≤10 °C would not expose term infants to an infectious dose of B cereus; however, the risk to premature infants is unknown. Thus, because of similarities in ingredients and manufacturing processes between powdered formula and HMF, potential to improperly reconstitute and store fortified breast milk, and the unknown risk that B cereus poses to premature infants, a common mechanism might be shared by formula and HMF and warrants additional examination. Because anaerobic cultures of HMF were not performed, isolation of B cereus might be a marker of similar or even greater levels of contamination with anaerobic spore-forming organisms (eg, Clostridium species), which have been previously indicated in the etiology of NEC.
      • de la Cochetiere M.F.
      • Piloquet H.
      • des Robert C.
      • Darmaun D.
      • Galmiche J.P.
      • Roze J.C.
      Early intestinal bacterial colonization and necrotizing enterocolitis in premature infants: the putative role of Clostridium.
      In this instance, the volume of HMF was insufficient to conduct both aerobic and anaerobic cultures.
      Findings in this report are subject to some limitations. First, pathogens previously implicated in certain NEC outbreaks were not uniformly tested for in this investigation, in part because an outbreak was not apparent until 5 cases had been identified. This includes the lack of cultures for anaerobic bacteria in both clinical and food product specimens. Second, no tissue samples from patients' gastrointestinal tracts were available for confirmatory testing. Third, a case-control study would have allowed us to better characterize risk factors for NEC (including HMF) in this outbreak. (A case-control study was not conducted because of the insufficient power this small number of cases would have yielded.) Despite these limitations, this investigation demonstrates that considering other risk factors is critical when investigating NEC clusters and might serve as the basis for conducting more rigorous studies to evaluate the role of HMF consumption in the pathogenesis of NEC.
      No new cases in this NEC cluster occurred after institution of enhanced infection control interventions and switching lot codes of HMF and formula fed to the neonates; however, no pathogens discovered in HMF were identified in any of the infants, and no causal associations were made to explain this temporal association. Because powdered formula products, including HMF, are not sterile, careful attention should be given to contraindications for their use. NICU staff should consider documenting lot codes of nutritional products administered to their patients. Strict adherence to proper infection control practices is a key recommendation in controlling NEC clusters.
      • Boccia D.
      • Stolfi I.
      • Lana S.
      • Moro M.L.
      Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
      Systematic testing of blood, stool, and tissue specimens (when available) is necessary for better understanding of the etiology of NEC clusters.
      The authors thank Erika Fernandez, MD, in the newborn intensive care unit at the University of New Mexico hospitals; Paul Torres, MS, Scientific Laboratory Division, NMDOH; Mark Mueller, MPH, Epidemiology and Response Division, NMDOH; Susan Kellie, MD, Ann-Marie Yaroslaski, RD, Pam Iwamoto, CIC, University of New Mexico hospitals; Matthew Arduino, DrPH, Division of Healthcare Quality Promotion, National Center for Preparedness, Detection, and Control of Infectious Diseases; Anna Bowen, MD, Division of Bacterial and Mycotic Diseases, National Center for Zoonotic, Vector-Borne, and Enteric Diseases for contributing the data on which this report, in part, is based.

      References

        • Boccia D.
        • Stolfi I.
        • Lana S.
        • Moro M.L.
        Nosocomial necrotising enterocolitis outbreaks: epidemiology and control measures.
        Eur J Pediatr. 2001; 160: 385-391
        • Neu J.
        Neonatal necrotizing enterocolitis: an update.
        Acta Paediatr Suppl. 2005; 94: 100-105
        • Stoll B.J.
        Epidemiology of necrotizing enterocolitis.
        Clin Perinatol. 1994; 21: 205-218
        • Guthrie S.O.
        • Gordon P.V.
        • Thomas V.
        • Thorp J.A.
        • Peabody J.
        • Clark R.H.
        Necrotizing enterocolitis among neonates in the United States.
        J Perinatol. 2003; 23: 278-285
        • Caplan M.S.
        • Jilling T.
        New concepts in necrotizing enterocolitis.
        Curr Opin Pediatr. 2001; 13: 111-115
        • Walsh M.C.
        • Kliegman R.M.
        Necrotizing enterocolitis: treatment based on staging criteria.
        Pediatr Clin North Am. 1986; 33: 179-201
        • Bell M.J.
        • Ternberg J.L.
        • Feigin R.D.
        • Keating J.P.
        • Marshall R.
        • Barton L.
        • et al.
        Neonatal necrotizing enterocolitis: therapeutic decisions based upon clinical staging.
        Ann Surg. 1978; 187: 1-7
        • Bjornvad C.R.
        • Schmidt M.
        • Petersen Y.M.
        • Jensen S.K.
        • Offenberg H.
        • Elnif J.
        • et al.
        Preterm birth makes the immature intestine sensitive to feeding-induced intestinal atrophy.
        Am J Physiol Regul Integr Comp Physiol. 2005; 289: R1212-R1222
        • McGuire W.
        • Anthony M.Y.
        Donor human milk versus formula for preventing necrotising enterocolitis in preterm infants: systematic review.
        Arch Dis Child Fetal Neonatal Ed. 2003; 88: F11-F14
        • Quigley M.A.
        • Henderson G.
        • Anthony M.Y.
        • McGuire W.
        Formula milk versus donor breast milk for feeding preterm or low birth weight infants.
        Cochrane Database Syst Rev. 2007; (CD002971)
        • Lambert D.K.
        • Christensen R.D.
        • Henry E.
        • Besner G.E.
        • Bael V.L.
        • Viedmeier S.E.
        • et al.
        Necrotizing enterocolitis in term neonates: data from a multihospital health-care system.
        J Perinatol. 2007; 27: 437-443
        • Hallstrom M.
        • Koivisto A.M.
        • Janas M.
        • Tammela O.
        Frequency of and risk factors for necrotizing enterocolitis in infants born before 33 weeks of gestation.
        Acta Paediatr. 2003; 92: 111-113
        • Lucas A.
        • Fewtrell M.S.
        • Morley R.
        • Lucas P.J.
        • Baker B.A.
        • Lister G.
        • et al.
        Randomized outcome trial of human milk fortification and developmental outcome in preterm infants.
        Am J Clin Nutr. 1996; 64: 142-151
        • CDC
        Enterobacter sakazakii infections associated with the use of powdered infant formula—Tennessee, 2001.
        MMWR Morb Mortal Wkly Rep. 2002; 51: 297-300
        • Adler A.
        • Gottesman G.
        • Dolfin T.
        • Arnon S.
        • Regev R.
        • Bauer S.
        • et al.
        Bacillus species sepsis in the neonatal intensive care unit.
        J Infect. 2005; 51: 390-395
        • Girisch M.
        • Ries M.
        • Zenker M.
        • Carbon R.
        • Rauch R.
        • Hofbeck M.
        Intestinal perforations in a premature infant caused by Bacillus cereus.
        Infection. 2003; 31: 192-193
      1. US Food and Drug Administration. Chapter 21. Food composition, standards, labeling and economics: compliance program guidance manual. Available at: http://www.cfsan.fda.gov/∼acrobat/cp21006.pdf. Accessed May 20, 2009.

        • de la Cochetiere M.F.
        • Piloquet H.
        • des Robert C.
        • Darmaun D.
        • Galmiche J.P.
        • Roze J.C.
        Early intestinal bacterial colonization and necrotizing enterocolitis in premature infants: the putative role of Clostridium.
        Pediatr Res. 2004; 56: 366-370