Secondary Logo

Journal Logo

Maternal-Neonatal Reports

Genetic Relatedness of Staphylococcus haemolyticus in Gut and Skin of Preterm Neonates and Breast Milk of Their Mothers

Soeorg, Hiie PhD*; Metsvaht, Hanna Kadri*; Keränen, Evamaria Elisabet*; Eelmäe, Imbi MD; Merila, Mirjam MD; Ilmoja, Mari-Liis MD§; Metsvaht, Tuuli PhD; Lutsar, Irja PhD*

Author Information
The Pediatric Infectious Disease Journal: March 2019 - Volume 38 - Issue 3 - p 308-313
doi: 10.1097/INF.0000000000002056

Abstract

Staphylococcus haemolyticus is the second most common coagulase-negative staphylococcus (CoNS) causing late-onset sepsis (LOS) in preterm neonates.1 LOS-causing strains are often multidrug resistant1,2 and belong to widespread predominant strains3 that may persist in neonatal intensive care units (NICU) for at least a decade.4 According to multilocus sequence typing (MLST), LOS-causing strains in Norway in 2004–2005 belonged to sequence type (ST) 1,4 that is one of the commonest STs of clonal complex 29 (CC29), a genetic lineage adapted to the hospital environment that causes the majority of infections in other patient groups as well.5–8

S. haemolyticus appears on the skin of preterm neonates hospitalized in the NICU within the first days of admission, whereas the prevalence of colonization and the antimicrobial resistance rate of colonizing strains increase during hospitalization.9,10 Over recent years, several studies have demonstrated that the gut could be a source of CoNS strains causing LOS.11–13 However, little is known about gut-colonizing compared with skin-colonizing isolates and their genetic relatedness with LOS-causing strains. As highlighted recently, clarification of infection route is warranted to further improve the prevention of LOS.14

We recently demonstrated that the gut of preterm neonates hospitalized in the NICU becomes initially colonized with S. epidermidis strains spreading in the unit but gradually is enriched with less pathogenic strains originating from mother’s breast milk (BM).15,16 On the other hand, BM of mothers of preterm neonates becomes colonized with mecA-positive CoNS17 and thus could serve as a source of invasive strains. However, the role of mother’s BM in S. haemolyticus colonization or infection in preterm neonates is unknown.

We aimed to describe gut and skin colonization of hospitalized preterm neonates with S. haemolyticus and to determine the role of mother’s own BM as a source of colonization and/or LOS by describing genetic relatedness of strains isolated from different sites. Healthy term neonates were included as a control group.

MATERIALS AND METHODS

Study Design

This is a substudy of a prospective longitudinal 2-group observational study describing colonization of gut and skin of neonates with CoNS from their mothers’ BM conducted from January 2014 to December 2015 in 2 third-level NICUs of Tallinn Children’s Hospital and Tartu University Hospital.16 Preterm neonates [gestational age (GA) <37 weeks] and their mothers were included if the neonate was hospitalized in the NICU and feeding with mother’s BM was initiated within the first week of life. Term neonates (GA ≥37 weeks, birth weight ≥2500 g) and their mothers were recruited if both were healthy and neonate was exclusively breastfed.

Sample Collection

Stool and skin of neonates and BM of mothers were sampled once a week in the first month after delivery, as described previously.16 All CoNS isolates from routine blood cultures from neonates with signs and symptoms of LOS, as described previously,16 were collected for typing.

Isolation of Staphylococci

Stool, skin and BM samples were cultured onto mannitol-salt agar, as described previously.16 After incubation at +37°C for 48 hours, 5 colonies, including each morphologically distinct, were randomly picked from each sample, identified to the species level by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (Bruker Daltonics, Bremen, Germany) and stored in skimmed milk at −80°C.

Multilocus variable number of tandem repeat analysis

Multilocus variable number tandem repeat analysis (MLVA) was used for typing as described elsewhere.5 For that purpose, DNA was extracted by boiling method18 from all S. haemolyticus isolates. Each distinct MLVA profile was defined as MLVA-type and designated with an arbitrary integer. The relatedness between MLVA types (excluding MLVA types yielding 2 bands in loci SH0999 or SH0324) was analyzed by goeBURST algorithm and Euclidean distance at single locus variant level in the software program PHYLOViZ 2.0.19 Simpson index of diversity was computed on the Comparing Partitions Website (http://www.comparingpartitions.info/; last accessed on 17 July 2017).

Multilocus sequence typing

Multilocus sequence typing (MLST) as described elsewhere5 was performed on the following 82 isolates: all LOS-causing isolates (n=8), randomly chosen isolates representing each distinct MLVA type in each term-neonate–mother pair (n=27) and randomly chosen isolates representing distinct MLVA types in unit A and unit B (n=47). STs were grouped into clonal complexes according to eBURST analysis (http://eburst.mlst.net/), using all STs in the S. haemolyticus MLST database as of August 29, 2017. Alleles and STs detected for the first time were submitted to the MLST database (https://pubmlst.org/shaemolyticus/).

Characterization of Virulence and Resistance Genes

From stool and skin of each neonate and BM of mother, the first isolate of each distinct MLVA type was characterized in terms of the presence of the mecA, icaA gene, and IS256 as described elsewhere.20,21

Data Analysis

Statistical analysis was performed with the software program R (version 3.4.1; 2017 The R Foundation for Statistical Computing). Fisher exact test was used for categorical and Mann–Whitney U test for continuous variables. The following characteristics were analyzed by logistic regression as potential risk factors associated with colonization by predominant NICU strains: GA; birth weight; delivery mode; unit; age at hospitalization; length of NICU stay; use and duration of invasive respiratory support, central venous catheter; use and duration of penicillin G, other beta-lactam antibiotics, gentamicin, vancomycin; age at initiation of BM feeding; and proportion of BM of total enteral feeds. In addition, logistic regression analysis of risk factors of LOS included proportion of S. haemolyticus among all staphylococci in gut and on skin. The software program EstimateS 9.1.0 was used for calculating Bray-Curtis similarity index.22

Ethics

The Research Ethics Committee of the University of Tartu approved the study protocol. The mother signed an informed consent on behalf of herself and her neonate before inclusion to the study.

RESULTS

Study Population

A total of 49 preterm (19 from unit A and 30 from unit B) and 20 term neonates were included in the study. Neonates in unit A had significantly lower GA, longer duration of NICU stay, invasive respiratory support and central venous catheter (Table, Supplemental Digital Content 1, https://links.lww.com/INF/D117). Term neonates [median (interquartile range (IQR)) GA 40 (39–40) weeks; birth weight 3651 (3324–3970) g; 35% (n=7) male] were delivered vaginally and discharged home at median (IQR) age of 2 (1.75–3) days. All preterm neonates hospitalized in NICU received enteral feeding by nasogastric tubes, and all term neonates were directly breastfed.

Colonization With S. haemolyticus

A total of 4403 staphylococci were isolated from stool, skin swabs and BM, of which 676 were S. haemolyticus. Overall, S. haemolyticus colonized the majority of preterm but less than half of term neonates and mothers (Table 1). Preterm neonates in unit B compared with unit A were colonized more commonly with S. haemolyticus (Table 1). S. haemolyticus constituted greater proportion of staphylococci in gut compared with skin of preterm neonates [36.4% (353/970) vs. 21.9% (237/1082); P < 0.001]. In term neonates and their mothers’ BM, less than 5% of staphylococci were S. haemolyticus. Colonization of gut and skin of preterm neonates occurred mostly within the first week of life, but their mothers’ BM within the second week postpartum (Fig., Supplemental Digital Content 2, https://links.lww.com/INF/D118). In preterm neonates, the proportion of S. haemolyticus among all staphylococci decreased in the first month of life in gut, but not on skin (Fig. 1).

TABLE 1.
TABLE 1.:
Prevalence of Colonization of Gut and Skin of Neonates and Their Mothers’ Breast Milk With Staphylococcus haemolyticus
FIGURE 1.
FIGURE 1.:
Proportion of S. haemolyticus of all staphylococci and the carriage rate of the mecA gene and IS256 and proportion of predominant NICU strains in S. haemolyticus from gut and skin of preterm neonates. *Statistically significant decrease in the proportion of S. haemolyticus among all staphylococci in gut of preterm neonates from 1–3 to 28–30 days of age.

MLVA types

Among 663 MLVA-typeable S. haemolyticus isolates, 44 distinct MLVA types were detected. Including only the first isolate of each distinct MLVA type from gut and skin of each neonate and BM of each mother, a total of 243 isolates were studied further. Overall, Simpson index of diversity was 0.938 [95% confidence interval (CI): 0.923–0.953] with no difference between preterm and term neonates and mothers. According to goeBURST algorithm, all except 4 MLVA types clustered into 1 group (Fig., Supplemental Digital Content 3, https://links.lww.com/INF/D119).

A total of 12 MLVA types colonized at least 5 neonates (Fig. 2). We defined 10 MLVA types that colonized at least 5 preterm neonates and none of term neonates or mothers of term neonates as predominant NICU strains (Fig. 2). Predominant NICU strains constituted the majority of all S. haemolyticus from preterm neonates and from BM of their mothers (Table 2).

TABLE 2.
TABLE 2.:
The Carriage Rate of the mecA Gene and IS256 and Proportion of Predominant NICU Strains in Staphylococcus haemolyticus Isolates and the Prevalence of Colonization With Such Strains in Gut and on Skin of Neonates and Breast Milk of Their Mothers
FIGURE 2.
FIGURE 2.:
Prevalence of MLVA types (MT) colonizing skin and/or gut of at least 5 neonates among preterm neonates in unit A (black bars) and unit B (gray bars), mothers of preterm neonates (white bars), term neonates (bars with gray horizontal lines) and their mothers (bars with black horizontal lines). Corresponding sequence types (ST) or its single-locus variants (SLVs) are shown in brackets. #Isolates of MT3 corresponded to several STs (ST30; single, double and triple locus variants of ST1; double locus variant of ST23). Black boxes around MLVA type numbers designate predominant NICU strains that were defined as MLVA types that colonized at least 5 preterm neonates, but none of term neonates or mothers of term neonates. Number of asterisks indicates number of bloodstream isolates from total of 7 episodes of late-onset sepsis (1 episode was caused by 2 distinct MLVA types).

The majority of preterm neonates harbored at least 1 predominant NICU strain in gut or on skin (Table 2). The odds of colonization of gut with predominant NICU strains were lower in unit A than that in unit B [odds ratio (95% CI): 0.19 (0.04–0.87)] and the odds of colonization of skin were lower in those receiving penicillin G [odds ratio (95% CI): 0.21 (0.05–0.81)].

Multilocus Sequence Typing

Of the 82 isolates typed by MLST, 2 did not yield a band in arc or SH1431 locus. Among the 80 typeable isolates, 24 distinct STs were detected, the most common of which were ST3 (n=24) and ST42 (n=19). According to eBURST analysis, 65 of 80 (81.3%) isolates belonged to CC29 (Fig., Supplemental Digital Content 4, https://links.lww.com/INF/D120), whereas CC29 was more common in preterm compared with term neonates and mothers [41/46 (89.1%) vs. 17/27 (63%); P = 0.014]. All predominant NICU strains were representatives of ST3, ST42 or single locus variant of ST3 (Fig. 2).

Virulence-Related Genes

S. haemolyticus from preterm neonates and their mothers carried the mecA gene and IS256 more commonly than isolates from term neonates and their mothers (Table 2). None of the isolates carried the icaA gene. Isolates belonging to predominant NICU strains, compared with other MLVA-types, carried more frequently the mecA gene [87% (134/154) vs. 44.9% (40/89); P < 0.001] and IS256 [56.5% (87/154) vs. 24.7% (22/89); P < 0.001]. The carriage rate of the mecA gene, IS256 and the proportion of predominant NICU strains did not change in isolates colonizing preterm neonates over time (Fig. 2).

Genetic Relatedness of Colonizing S. haemolyticus

Indistinguishable MLVA types in gut and on skin were common in preterm but rare in term neonates [39 (79.6%) vs. 5 (25%); P < 0.001]. Indistinguishable MLVA types in BM and gut were isolated from the gut of 14.3% (n=7) preterm and 15% (n=3) term neonates. Indistinguishable MLVA types in BM and on skin were isolated from the skin of 14.3% (n=7) preterm and 15% (n=3) term neonates. Low overall similarity of neonatal gut or skin and mother’s BM samples contrasted the high similarity of gut and skin (Fig., Supplemental Digital Content 5, https://links.lww.com/INF/D121).

LOS Caused by S. haemolyticus

S. haemolyticus caused 7 episodes of LOS in 6 preterm neonates [median (IQR) GA 26 (24–28) weeks] at median (IQR) age of 10 (6–15) days. All 8 isolates (1 episode was caused by 2 distinct MLVA types) were predominant NICU strains (Fig. 2). The LOS-causing MLVA type was isolated from gut and on skin in 4 and 5 episodes, respectively, similarly a median (IQR) of 4 (2.5–5) days before the onset of LOS. LOS-causing MLVA types were not present in mother’s BM before the onset of LOS. Only in 1 LOS episode LOS-causing MLVA type was isolated from BM, but the sample was collected 1 day after the onset of LOS.

In univariate logistic regression analysis, longer length of NICU stay and duration of treatment with β-lactam antibiotics other than penicillin G, larger proportion of S. haemolyticus of all staphylococci in gut and on skin increased the odds of LOS caused by S. haemolyticus (Table, Supplemental Digital Content 6, https://links.lww.com/INF/D122). All LOS episodes caused by S. haemolyticus occurred in unit B, where neonates had shorter duration of treatment with β-lactams other than penicillin G and length of NICU stay (Table, Supplemental Digital Content 1, https://links.lww.com/INF/D117). However, neonates with LOS caused by S. haemolyticus compared with neonates in unit B without LOS caused by S. haemolyticus and neonates in unit A had longer duration of treatment with β-lactam antibiotics other than penicillin G [median (IQR) 14 (13–15.5) vs. 7.5 (4–11) vs. 11 (3–14.5) days] and length of NICU stay [median (IQR) 28.5 (23.5–31.31) vs. 11.5 (7–17) vs. 26 (18–28) days]. In multivariate logistic regression analysis, only a larger proportion of S. haemolyticus of all staphylococci among skin isolates remained a significant risk factor increasing the odds of LOS caused by S. haemolyticus [OR (95% CI): 134 (4.8–3737)].

DISCUSSION

To the best of our knowledge, this is the first study analyzing the genetic relationship between S. haemolyticus colonizing skin and gut of preterm neonates and their mother’s BM and those causing LOS in preterm neonates. We demonstrated that mother’s BM is rarely a source of neonatal colonization with S. haemolyticus in contrast to S. epidermidis that often originates from mother’s BM.16 Instead, the gut and the skin of preterm neonates become colonized soon after admission to NICU with S. haemolyticus of which the majority are mecA-positive and/or predominant NICU strains. These strains belong to CC29 and may subsequently cause LOS.

The very high prevalence of colonization with S. haemolyticus in preterm neonates is in line with previous studies, where colonization of gut in preterm neonates was 89%,23 and S. haemolyticus constituted 58% of all staphylococci from skin.24 Our study suggests that this high carriage rate is most likely a result of exposure to NICU environment. First, the prevalence of colonization in preterm neonates was more frequent than in term neonates; it occurred soon after admission and mostly with predominant NICU strains that were not seen in term neonates. Second, the prevalence of colonization differed between the units. Such difference is in accordance with other studies that report varying prevalence of colonization with S. haemolyticus (from 0% to 87%).23,25,26 Unit was also the only risk factor for colonization with predominant NICU strains in the gut. Although we did not study possible environmental sources, the hands of NICU staff27 or inanimate objects, such as disinfectant bottles28 or feeding tubes,29 could be sources for colonization in NICU.

The facts that in the first 3 days S. haemolyticus form 60% of all staphylococci in gut but only 20% on skin, enteral feeding was initiated mostly within the first 2 days of life and all preterm neonates were fed by nasogastric tubes in NICU suggest that enteral nutrition and/or feeding tubes have an important role in colonization with S. haemolyticus in NICU. Indeed, feeding regimen influences development of microbiota in neonates. For example, S. haemolyticus is more typical of the gut of formula-fed compared with BM-fed neonates.30,31 This is in line with the scarcity of S. haemolyticus in BM32 that was also shown in our study. Although all neonates in our study received mother’s BM, the small amounts of BM could be insufficient to influence colonization in the first week of life when the colonization with predominant NICU strains mostly occurred. Still, we found a decrease in the abundance of S. haemolyticus among colonizing staphylococci in the gut within the first month of life. This could occur at the expense of more common colonization with S. epidermidis strains originating from mother’s own BM, as shown in our previous study.16 The impact of feeding regimen (mother’s own BM, pasteurized or frozen donor milk) on gut colonization of preterm neonates should be further studied.

In preterm neonates, most colonizing isolates, both on skin and in gut, were mecA-positive and/or predominant NICU strains that also caused LOS, similarly to previous studies.9,10,33 Large proportion of isolates from preterm neonates carried the insertion sequence IS256 that contributes to adaptation to hospital environment and/or were predominant NICU strains that had genetic background associated with infection development, that is, CC29.8 In line with domination of ica-independent biofilm formation34 and thus low carriage rate of ica-operon,35 none of the isolates carried the icaA gene. Such characteristics, that is, the mecA gene, IS256 and CC29, are also typical to LOS-causing strains,3,4,36 suggesting that colonization with strains with invasiveness potential is widespread in gut and on skin. Such predominant strains were found in BM of mothers of preterm neonates as well, indicating that BM could be potentially a source of strains with higher pathogenicity. However, none of the mothers of preterm neonates with LOS caused by S. haemolyticus had the strain that caused LOS in the BM before the onset of LOS, suggesting that BM is probably rare, if at all a source of LOS-causing S. haemolyticus. Instead, subsequently invasive strains were found on skin or in gut in more than 4 of 7 LOS cases. Although higher proportion of S. haemolyticus among staphylococci-colonizing skin was a significant risk factor for LOS caused by S. haemolyticus, the proportion of S. haemolyticus in gut was larger than on skin, underlining that skin and gut could be equally important reservoirs of S. haemolyticus, similarly to a previous study.9 Thus, for decreasing the incidence of LOS, one should reduce the prevalence of colonization with S. haemolyticus, for example, by detection and elimination of contaminated sources28 or changing feeding practices (eg, using fresh or frozen BM instead of formula or pasteurized BM).

Some limitations of the study should be noted. First, we did not use whole-genome sequencing that could more conclusively determine the genetic relatedness of isolates colonizing skin, gut and BM and causing LOS compared with typing methods with lower discriminatory power, such as MLVA.5 However, even by using MLVA, we could demonstrate that strains in BM only rarely colonize neonates. Second, we did not study NICU environment, and thus, we cannot conclusively determine what exactly in NICU environment is related to colonization in preterm neonates. Still, different prevalence of S. haemolyticus between the units in our study and large variation between previous studies,23,25 but also the high proportion of predominant NICU strains suggesting transmission, imply major role of NICU environment in colonization with S. haemolyticus. Third, we included only the first isolate from each neonate and mother in the analysis of MLVA typing results. However, we believe that such selection avoids overrepresentation of MLVA types from neonates with high proportion of S. haemolyticus among colonizing staphylococci. Finally, we did not include preterm neonates not receiving mother’s BM, and thus, we cannot conclusively determine the role of feeding regimen in colonization with S. haemolyticus. However, comparison of BM-fed and formula-fed neonates in terms of the prevalence or abundance of colonization with S. haemolyticus was beyond the scope of this study but warrants further studies. If BM reduces colonization with S. haemolyticus, early administration of colostrum that reduces the abundance of staphylococci in oral microbiome37 could be an approach to prevent the initial colonization with S. haemolyticus.

In conclusion, S. haemolyticus is a common colonizer of premature neonates in NICU, but BM is a rare source of gut and skin colonization or LOS in preterm neonates. Instead, gut and then skin of preterm neonates hospitalized in the NICU become colonized soon after birth with mecA-positive S. haemolyticus strains adapted to the hospital environment. As more abundant colonization is associated with development of LOS, the possibilities to limit spread of S. haemolyticus in NICU warrant further studies.

ACKNOWLEDGMENTS

The authors thank the study nurses Marika Zuihhina, Eve Kaur and Tuuli Tammekunn; clinical microbiologist Dr Marika Jürna-Ellam; laboratory assistants Dagmar Hoidmets, Tiiu Rööp and Sandra Sokmann; research fellow Kristi Huik; and all the study participants.

References

1. Dong H, Cao H, Zheng H. Pathogenic bacteria distributions and drug resistance analysis in 96 cases of neonatal sepsis. BMC Pediatr. 2017;17:44.
2. Hira V, Sluijter M, Estevão S, et al. Clinical and molecular epidemiologic characteristics of coagulase-negative staphylococcal bloodstream infections in intensive care neonates. Pediatr Infect Dis J. 2007;26:607–612.
3. Pereira PM, Binatti VB, Sued BP, et al. Staphylococcus haemolyticus disseminated among neonates with bacteremia in a neonatal intensive care unit in Rio de Janeiro, Brazil. Diagn Microbiol Infect Dis. 2014;78:85–92.
4. Cavanagh JP, Hjerde E, Holden MT, et al. Whole-genome sequencing reveals clonal expansion of multiresistant Staphylococcus haemolyticus in European hospitals. J Antimicrob Chemother. 2014;69:2920–2927.
5. Cavanagh JP, Klingenberg C, Hanssen AM, et al. Core genome conservation of Staphylococcus haemolyticus limits sequence based population structure analysis. J Microbiol Methods. 2012;89:159–166.
6. Panda S, Jena S, Sharma S, et al. Identification of novel sequence types among Staphylococcus haemolyticus isolated from variety of infections in India. PLoS One. 2016;11:e0166193.
7. McManus BA, Coleman DC, Deasy EC, et al. Comparative genotypes, staphylococcal cassette chromosome mec (SCCmec) genes and antimicrobial resistance amongst Staphylococcus epidermidis and Staphylococcus haemolyticus isolates from infections in humans and companion animals. PLoS One. 2015;10:e0138079.
8. Bouchami O, de Lencastre H, Miragaia M. Impact of insertion sequences and recombination on the population structure of Staphylococcus haemolyticus. PLoS One. 2016;11:e0156653.
9. Hira V, Kornelisse RF, Sluijter M, et al. Colonization dynamics of antibiotic-resistant coagulase-negative Staphylococci in neonates. J Clin Microbiol. 2013;51:595–597.
10. Ternes YM, Lamaro-Cardoso J, André MC, et al. Molecular epidemiology of coagulase-negative Staphylococcus carriage in neonates admitted to an intensive care unit in Brazil. BMC Infect Dis. 2013;13:572.
11. Soeorg H, Huik K, Parm U, et al. Genetic relatedness of coagulase-negative Staphylococci from gastrointestinal tract and blood of preterm neonates with late-onset sepsis. Pediatr Infect Dis J. 2013;32:389–393.
12. Cossey V, Vanhole C, Verhaegen J, et al. Intestinal colonization patterns of staphylococci in preterm infants in relation to type of enteral feeding and bacteremia. Breastfeed Med. 2014;9:79–85.
13. Shaw AG, Sim K, Randell P, et al. Late-onset bloodstream infection and perturbed maturation of the gastrointestinal microbiota in premature infants. PLoS One. 2015;10:e0132923.
14. Decousser JW. Prevention of paediatric nosocomial infections: adapting before acting. Lancet Infect Dis. 2017;17:350–351.
15. Soeorg H, Huik K, Parm Ü, et al. Molecular epidemiology of Staphylococcus epidermidis in neonatal intensive care units. APMIS. 2017;125:63–73.
16. Soeorg H, Metsvaht T, Eelmäe I, et al. The role of breast milk in the colonization of neonatal gut and skin with coagulase-negative staphylococci. Pediatr Res. 2017;82:759–767.
17. Soeorg H, Metsvaht T, Eelmäe I, et al. Coagulase-negative Staphylococci in human milk from mothers of preterm compared with term neonates. J Hum Lact. 2017;33:329–340.
18. Zhang K, Sparling J, Chow BL, et al. New quadriplex PCR assay for detection of methicillin and mupirocin resistance and simultaneous discrimination of Staphylococcus aureus from coagulase-negative staphylococci. J Clin Microbiol. 2004;42:4947–4955.
19. Nascimento M, Sousa A, Ramirez M, et al. PHYLOViZ 2.0: providing scalable data integration and visualization for multiple phylogenetic inference methods. Bioinformatics. 2017;33:128–129.
20. Kondo Y, Ito T, Ma XX, et al. Combination of multiplex PCRs for staphylococcal cassette chromosome mec type assignment: rapid identification system for mec, ccr, and major differences in junkyard regions. Antimicrob Agents Chemother. 2007;51:264–274.
21. Ziebuhr W, Krimmer V, Rachid S, et al. A novel mechanism of phase variation of virulence in Staphylococcus epidermidis: evidence for control of the polysaccharide intercellular adhesin synthesis by alternating insertion and excision of the insertion sequence element IS256. Mol Microbiol. 1999;32:345–356.
22. Colwell R. EstimateS: statistical estimation of species richness and shared species from samples. Version 9.1.0. 2013. Available at: http://purl.oclc.org/estimates. Accessed July 26, 2016.
23. Aujoulat F, Roudière L, Picaud JC, et al. Temporal dynamics of the very premature infant gut dominant microbiota. BMC Microbiol. 2014;14:325.
24. Jain A, Agarwal J, Bansal S. Prevalence of methicillin-resistant, coagulase-negative staphylococci in neonatal intensive care units: findings from a tertiary care hospital in India. J Med Microbiol. 2004;53(pt 9):941–944.
25. Moles L, Gómez M, Heilig H, et al. Bacterial diversity in meconium of preterm neonates and evolution of their fecal microbiota during the first month of life. PLoS One. 2013;8:e66986.
26. Keyworth N, Millar MR, Holland KT. Development of cutaneous microflora in premature neonates. Arch Dis Child. 1992;67(7 Spec No):797–801.
27. Hira V, Sluijter M, Goessens WH, et al. Coagulase-negative staphylococcal skin carriage among neonatal intensive care unit personnel: from population to infection. J Clin Microbiol. 2010;48:3876–3881.
28. Ben Saida N, Marzouk M, Ferjeni A, et al. A three-year surveillance of nosocomial infections by methicillin-resistant Staphylococcus haemolyticus in newborns reveals the disinfectant as a possible reservoir. Pathol Biol (Paris). 2009;57:e29–e35.
29. Mehall JR, Kite CA, Saltzman DA, et al. Prospective study of the incidence and complications of bacterial contamination of enteral feeding in neonates. J Pediatr Surg. 2002;37:1177–1182.
30. Parm Ü, Metsvaht T, Ilmoja ML, et al. Gut colonization by aerobic microorganisms is associated with route and type of nutrition in premature neonates. Nutr Res. 2015;35:496–503.
31. 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:F167–173.
32. Jost T, Lacroix C, Braegger C, et al. Assessment of bacterial diversity in breast milk using culture-dependent and culture-independent approaches. Br J Nutr. 2013;110:1253–1262.
33. Kornienko M, Ilina E, Lubasovskaya L, et al. Analysis of nosocomial Staphylococcus haemolyticus by MLST and MALDI-TOF mass spectrometry. Infect Genet Evol. 2016;39:99–105.
34. Fredheim EG, Klingenberg C, Rohde H, et al. Biofilm formation by Staphylococcus haemolyticus. J Clin Microbiol. 2009;47:1172–1180.
35. de Silva GD, Kantzanou M, Justice A, et al. The ica operon and biofilm production in coagulase-negative Staphylococci associated with carriage and disease in a neonatal intensive care unit. J Clin Microbiol. 2002;40:382–388.
36. Foka A, Chini V, Petinaki E, et al. Clonality of slime-producing methicillin-resistant coagulase-negative staphylococci disseminated in the neonatal intensive care unit of a university hospital. Clin Microbiol Infect. 2006;12:1230–1233.
37. Sohn K, Kalanetra KM, Mills DA, et al. Buccal administration of human colostrum: impact on the oral microbiota of premature infants. J Perinatol. 2016;36:106–111.
Keywords:

late-onset sepsis; colonization; molecular epidemiology; multilocus sequence typing; term neonate

Supplemental Digital Content

Copyright © 2018 Wolters Kluwer Health, Inc. All rights reserved.