Coagulase-negative staphylococci (CoNS) are the most prevalent pathogens causing late onset sepsis in neonates. These infections are often associated with use of indwelling vascular catheters.1 In contrast to Staphylococcus aureus, CoNS produce a very limited number of virulence factors.2 Neonatal CoNS infections are rarely fatal, but they cause significant morbidity, especially among very low birth weight infants.1,3,4 In addition, CoNS frequently display multiresistance to antibiotics.5 However, the relationship between antibiotic resistance and clinical outcome of CoNS infections is unknown.
The ability of CoNS to form an adherent multilayered biofilm on polymer surfaces is considered their main virulence determinant.2 The early stage of biofilm formation may arbitrarily be divided into 2 phases: a rapid bacterial adherence to inorganic and/or organic surfaces followed by a more prolonged cell accumulation phase. The autolysin AtlE, belonging to a family of cell wall-lytic enzymes with adhesive properties, is central in the rapid adherence phase. Formation of cell clusters in the accumulation phase depends on production of polysaccharide intercellular adhesin, synthesized by icaADBC-encoded proteins.2,6,7 The icaADBC operon is highly prevalent among invasive CoNS isolates and has been proposed as a marker to discriminate between invasive and contaminating isolates.8,9 In animal models, atlE- and icaADBC-negative mutants have shown reduced capacity to colonize indwelling devices compared with wild-type strains.10–12 However, increased levels of biofilm production are conversely associated with a decreased incidence of tissue invasion.12,13 In general, there is a paucity of data on the relationship between virulence factors and clinical outcome of CoNS infections.
The aim of this study was to describe the association between antibiotic resistance, biofilm production and genetic determinants for biofilm formation in blood isolates of different CoNS species and their correlation with neonatal inflammatory response.
MATERIALS AND METHODS
Patients eligible for inclusion in this analysis were neonates admitted to the tertiary care neonatal intensive care unit (NICU) at Rikshospitalet University Hospital, Oslo during the period of January 1989 through April 2000 and diagnosed with growth of CoNS in blood culture. This NICU has an annual admission rate of ∼800 patients, among these on average 100–110 very low birth weight infants. During the study period, ampicillin and netilmicin were administered empirically to all neonates with suspected sepsis and to all preterm infants with respiratory distress until a group B streptococcal pneumonia could be ruled out. Blood cultures were collected from a peripheral vein after skin cleansing with a chlorhexidin-ethanol solution or from an intravascular catheter at the time of insertion. Duplicate cultures were not routinely obtained. We identified 150 neonates, from whom a total of 180 CoNS blood isolates were retrieved. The diagnosis of sepsis caused by CoNS required clinical signs of sepsis in a neonate older than 72 hours of age and a positive monomicrobial blood culture and elevated CRP >10 mg/L within 2 days of blood culture. All other CoNS blood isolates, including cultures growing more than one organism, were considered to be contaminants. Data collected on each patient included birth weight (BW), gestational age (GA), maximum C-reactive protein (CRP) values within 2 days of blood culture, current and previous use of umbilical and central venous catheters, ventilator treatment and days in hospital. The regional committee for medical research ethics approved the collection and analysis of patient data.
All isolates were identified to the species level by ID 32 Staph (bioMérieux, Marcy l'Etoile, France). We determined minimum inhibitory concentrations (MIC) of oxacillin, gentamicin, erythromycin, clindamycin, fusidic acid, rifampin and vancomycin with E-test (AB Biodisk, Solna, Sweden). Susceptibility was interpreted according to Clinical and LaboratoryStandards Institute guidelines.14 Semiquantitative determination of biofilm production was performed as previously described.15 Briefly each strain was inoculated in 8 parallel wells in polystyrene microtiter plates (Nunclon, Roskilde, Denmark). All strains were tested independently on 3 separate occasions. We determined the optical density (OD) of the crystal violet-stained adherent biofilm with a spectrophotometer at 570 nm. The highest and lowest OD570 value of each run was excluded from the analyses, and the remaining 18 values were averaged. Staphylococcus epidermidis ATCC 35984 (RP62A) and S. epidermidis ATCC 12228 served as positive and negative controls, respectively. Isolates with OD570 < 0.20, 0.20 ≤ OD570 ≤ 1.0 and OD570 > 1.0 were defined as biofilm-negative, weak biofilm producers and strong biofilm producers, respectively.16
DNA extractions and PCRs for the methicillin resistance gene (mecA), the most frequent aminoglycoside resistance gene (aac(6′)-Ie-aph(2′′)-Ia) and icaD, as a marker for the ica operon, were conducted as previously reported.17–19 The primer sequences for atlE were forward 5′-TCC GAC AGA TTA CTT ATC TTG GG-3′ and reverse 5′-ATT TGA GCA ACA CCA CGA TTA G-3′, corresponding to bases 3822-4256 of the atlE gene (S. epidermidis gene sequence, GenBank accession number U71377). S. epidermidis ATCC 35984 served as positive control for both icaD and atlE. Primers targeting the 16S ribosomal RNA gene were used to amplify an internal control sequence to identify potentially false negative results.20 Polymerase chain reaction (PCR) products were analyzed with gel electrophoresis. Genomic DNA from all biofilm forming strains with negative icaD PCR by 2 independent analyses were further analyzed with dot-blot hybridization, and some were in addition confirmed with Southern blot hybridization. The icaD PCR amplicon was used as probe. Probe labeling, hybridization and detection were performed as previously described.20 A positive hybridization result was interpreted as verified identification of the icaD gene. Pulsed field gel electrophoresis (PFGE) was performed on all isolates as previously described.20 PFGE patterns were analyzed by GelCompar II version 2.5 (Applied Maths, Sint-Martens-Latem, Belgium). The Dice coefficient was calculated with the use of a position tolerance of 1.6%. PFGE dendrograms were constructed by the unweighted pair group method with arithmetic mean. Isolates with 95% similarity were considered to be indistinguishable strains.
To compare differences between groups, the Mann-Whitney U test was used for continuous variables without normal distribution and the χ2 test or Fisher's exact test were used for dichotomous variables. Among patients with sepsis, we performed a linear multivariable regression analysis to identify bacterial and clinical variables potentially influencing CRP, as a marker of inflammatory response. To reduce skewness of residuals CRP was log-transformed. All analyses were 2-tailed, and P < 0.05 was considered to be significant. Statistical analysis was performed with SPSS for Windows software version 11.0 (2001).
Background demographic data are listed in Table 1. Growth of CoNS in blood culture was identified in 164 suspected septic episodes. These episodes were divided into a“sepsis” and a “contaminant” group. In 16 patients, blood cultures revealed growth of 2 different CoNS isolates and were thus by definition classified as contaminants. Ten patients experienced episodes classified as sepsis as well as episodes where an isolate in blood culture was considered as contamination. Two of 81 patients [body weights, 430 and 550 g, respectively; gestational age, 24 and 26 weeks, respectively] died of the infection within 3 days of the positive blood culture, giving a CoNS sepsis-attributable mortality rate of 2.5%.1,21 None of the neonates in the contamination group died within 3 days of blood culture.
Table 2 shows the species distribution. Eighty-five isolates from 81 patients (4 patients had 2 separate episodes of CoNS sepsis) were considered as invasive and 95 isolates from 79 patients as contaminants. PFGE dendrograms revealed a diverse population. The largest cluster of indistinguishable strains contained 5 isolates. Additionally 4 clusters contained 3 isolates each. We have therefore chosen to include all isolates in the analyses. The antibiotic resistance rates and the prevalence of antibiotic resistance genes are listed in Table 3. Significantly higher prevalences of both phenotypic and genotypic β-lactam and aminoglycoside resistance were seen in invasive isolates. All 19 Staphylococcus haemolyticus isolates carried mecA and aac(6′)-Ie-aph(2′′)-Ia and were phenotypically resistant to oxacillin and gentamicin.
Biofilm production was more frequent in S. epidermidis than in other CoNS species (Table 2) (P < 0.001). The icaD prevalences in S. epidermidis isolates classified as negative, weak and strong biofilm producers were 30, 87 and 96%, respectively. However, 15 icaD-positive S. epidermidis isolates did not produce biofilm. We found no correlation between phenotypic biofilm formation and icaD carriage in CoNS non-epidermidis isolates. The atlE prevalences in S. epidermidis isolates classified as negative, weak and strong biofilm producers were 72, 79 and 82%, respectively.
Table 3 shows that there were no differences in biofilm production and icaD or atlE carriage between invasive isolates and contaminants. There was no correlation between the presence of atlE and icaD in S. epidermidis (P = 0.176).
We found no correlation between CRP values and gestational age, body weight, sex or age at onset of infection. Figure 1 shows the association between CRP, biofilm production and methicillin resistance. CRP values were on average 74% higher in CoNS sepsis caused by isolates containing mecA compared with sepsis with mecA-negative isolates. In contrast, we observed on average 34% lower CRP values in sepsis caused by biofilm-positive versus biofilm-negative isolates. Because of high colinearity between mecA and aac(6′)-Ie-aph(2′′)-Ia (Pearson correlation coefficient 0.708), the latter resistance gene was not included in the multivariable model. In a separate linear multivariable regression analysis including aac(6′)-Ie-aph(2′′)-Ia, but excluding mecA, we also found a significant association between an increased inflammatory response and isolates containing aac(6′)-Ie-aph(2′′)-Ia (P = 0.023). No other antibiotic resistance patterns were associated with inflammatory response. There were no significant differences in CRP values in sepsis caused by different CoNS species.
Antibiotic resistance was more frequent in biofilm-positive than biofilm-negative S. epidermidis isolates (Table 4). Among 98 mecA-positive S. epidermidis isolates the median oxacillin MIC was 4 mg/L in 70 biofilm-positive isolates compared with median MIC of 48 mg/L in 28 biofilm-negative isolates (P < 0.001). In contrast to S. epidermidis, there was no correlation between phenotypic biofilm production and antibiotic resistance rates in CoNS non-epidermidis isolates (Table 4). There was also a lack of correlation between atlE and antibiotic resistance (data not shown).
Consistent with previous studies from NICUs, we found a high overall level of antibiotic resistance among the isolates in this study.5 In addition, there were higher rates of resistance to β-lactams and aminoglycosides in invasive isolates versus contaminants. There may be different explanations for this observation. Sepsis patients were more premature, received more mechanical ventilation and were more exposed to antibiotic treatment, which in turn may lead to selection of resistant microbes.22 The blood cultures of sepsis patients were also obtained significantly later during the hospital course, and CoNS isolates from the skin surface of preterm neonates may become increasingly resistant to antibiotics during the first weeks after birth.23 However, the prevalence of genes encoding resistance to β-lactams and aminoglycosides was high even among contaminant isolates. Previous studies have shown that the antimicrobial resistance patterns reflect the antibiotic use in that unit,24 and it is probable that the predominant use of β-lactams and aminoglycosides in our NICU have exerted a selective pressure on the commensal CoNS population.
As expected, there was a strong correlation between the presence of icaD and biofilm production in S. epidermidis isolates. The expression of the icaADBC genes is controlled by a complex variety of conditions and factors, as illustrated by the fact that 17% of icaD-positive isolates did not produce biofilm.2,25 Although many non-epidermidis strains can cause similarly serious infections as S. epidermidis, less is known about their putative virulence determinants.2 Like others, we found that S. epidermidis produced more biofilm than other CoNS species.19 However, our results furthermore indicate presence of homologous ica genes in other CoNS species.16,26 We found an association between biofilm production and both phenotypic and genotypic antibiotic resistance.8,19,27 The close contact of bacteria within a biofilm may facilitate horizontal exchange of genetic information, including antimicrobial resistance genes and virulence determinants.
Neonates with CoNS sepsis may present with a highly variable clinical picture, but in general the CRP response is lower than in other bacterial infections.28 Recent investigations have in detail delineated how polysaccharide intercellular adhesin of S. epidermidis has a crucial role in evasion of the human innate immune system.29,30 Despite a limited number of observations, we identified a significantly lower CRP response in neonates with CoNS sepsis caused by biofilm-positive versus -negative isolates. This may be in line with the current hypothesis that CoNS evade from the host's immune system simply by hiding behind a biofilm layer.2,7
We found higher CRP values in sepsis episodes with methicillin/aminoglycoside-resistant versus methicillin/aminoglycoside-susceptible CoNS. A delay in adequate antibiotic treatment could explain why CRP seemed to be higher in sepsis with resistant strains. We are not aware of any genetic linkage between antibiotic resistance and proinflammatory virulence determinants in CoNS, but such a linkage might be an explanation for our results. It is striking that biofilm production and antibiotic resistance were so closely correlated, whereas these factors seem to create an opposite inflammatory host response. However, it is possible that biofilm production and antibiotic resistance are independently selected as determinants that confer a selective advantage for colonization and survival in a hospital environment. Recently the proinflammatory properties of staphylococcal phenol-soluble modulins (PSM) were described.31 PSM and biofilm formation are both under control by the quorum-sensing system agr. Suppression of agr promotes biofilm formation but limits PSM production.7,31 Furthermore naturally occurring agr mutants in S. epidermidis may enhance biofilm formation.12 Mutations in the agr locus might be an alternative explanation for differences in proinflammatory capacity of CoNS sepsis isolates in our study.
This study has several potential limitations. Distinguishing between true CoNS bacteremia and blood culture contamination in neonates is notoriously difficult.19,32 To determine neonatal CoNS sepsis in the scenario of only 1 single positive blood culture requires the addition of clinical signs of sepsis and either at least 5 days of antibacterial therapy33,34 or elevated CRP1,35 values. These additional criteria all have limitations. CRP values may sometimes be elevated by noninfectious conditions.36 However, a definition based on CRP values may in retrospect be more objective and robust than criteria based on duration of antibacterial treatment. Still we acknowledge that a proportion of the CoNS sepsis episodes may have represented contamination and vice versa, potentially skewing our comparisons. Finally, when analyzing the inflammatory response, we performed a multivariable analysis including the variables biofilm production, methicillin/aminoglycoside resistance, CoNS species, gestational age, sex and age at onset of infection to adjust for possible confounders. However, we cannot exclude other unknown and potential confounding patient conditions or bacterial factors having influenced the CRP values.
A valid and reliable method to distinguish clinically relevant from contaminant CoNS isolates would be of great value to clinicians, but no such methods exist. There are multiple studies with contradictory results on the correlation between the presence of ica genes and invasiveness.8,9,19,25,37,38 In accordance with the latest studies, we conclude that detection of theica operon is not suitable for discriminating invasive from contaminating CoNS isolates.37,38 Furthermore targeting the combination of mecA and the ica operon did not yield any improved discriminatory value in our study. In the future PCR-based detection of CoNS in the blood of suspected septic neonates may improve diagnostic accuracy.39
To our knowledge, this is the first clinical study to assess the correlation between host inflammatory response, as a marker of CoNS virulence, and the ability of CoNS to produce biofilm as well as to express multiple antimicrobial resistance. We found an enhanced neonatal inflammatory response in CoNS sepsis caused by antibiotic-resistant versus -susceptible isolates. In contrast, we observed a significantly decreased inflammatory response in sepsis caused by biofilm-positive versus -negative isolates. The latter finding suggests that evasion of the host immune system is a clinical relevant mechanism in neonatal CoNS sepsis.
We thank the Department of Microbiology, University Hospital of North Norway for kindly providing control strains; Bettina Aasnæs, Jarle Mikalsen and Lise Nordgård for excellent technical assistance; Arnfinn Sundsfjord and Gunnar S. Simonsen for critical reading of the manuscript; and Tom Wilsgaard for statistical advice.
1. Stoll BJ, Hansen N, Fanaroff AA, Wright LL, et al. Late-onset sepsis in very low birth weight neonates: the experience of the NICHD Neonatal Research Network. Pediatrics
2. Otto M. Virulence factors of the coagulase-negative staphylococci. Front Biosci
3. Maayan-Metzger A, Linder N, Marom D, Vishne T, Ashkenazi S, Sirota L. Clinical and laboratory impact of coagulase-negative staphylococci bacteremia in preterm infants. Acta Paediatr
4. Isaacs D. A ten year, multicentre study of coagulase negative staphylococcal infections in Australasian neonatal units. Arch Dis Child Fetal Neonatal Ed
5. de Silva GD, Justice A, Wilkinson AR, et al. Genetic population structure of coagulase-negative staphylococci associated with carriage and disease in preterm infants. Clin Infect Dis
6. von Eiff C, Proctor RA, Peters G. Coagulase-negative staphylococci: pathogens have major role in nosocomial infections. Postgrad Med
. 2001;110:63–70, 73.
7. Yao Y, Sturdevant DE, Otto M. Genomewide analysis of gene expression in Staphylococcus epidermidis
biofilms: insights into the pathophysiology of S. epidermidis
biofilms and the role of phenol-soluble modulins in formation of biofilms. J Infect Dis
8. Frebourg NB, Lefebvre S, Baert S, Lemeland JF. PCR-based assay for discrimination between invasive and contaminating Staphylococcus epidermidis
strains. J Clin Microbiol
9. Galdbart JO, Allignet J, Tung HS, Ryden C, El Solh N. Screening for Staphylococcus epidermidis
markers discriminating between skin-flora strains and those responsible for infections of joint prostheses. J Infect Dis
10. Rupp ME, Fey PD, Heilmann C, Gotz F. Characterization of the importance of Staphylococcus epidermidis
autolysin and polysaccharide intercellular adhesin in the pathogenesis of intravascular catheter-associated infection in a rat model. J Infect Dis
11. Rupp ME, Ulphani JS, Fey PD, Mack D. Characterization of Staphylococcus epidermidis
polysaccharide intercellular adhesin/hemagglutinin in the pathogenesis of intravascular catheter-associated infection in a rat model. Infect Immun
12. Vuong C, Kocianova S, Yao Y, Carmody AB, Otto M. Increased colonization of indwelling medical devices by quorum-sensing mutants of Staphylococcus epidermidis
in vivo. J Infect Dis
13. Patrick CC, Plaunt MR, Hetherington SV, May SM. Role of the Staphylococcus epidermidis
slime layer in experimental tunnel tract infections. Infect Immun
14. Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing. 12th Informational Supplement. M100-S12. Wayne, PA: Clinical and Laboratory Standards Institute, 2002.
15. Christensen GD, Simpson WA, Younger JJ, et al. Adherence of coagulase-negative staphylococci to plastic tissue culture plates: a quantitative model for the adherence of staphylococci to medical devices. J Clin Microbiol
16. Moretro T, Hermansen L, Holck AL, Sidhu MS, Rudi K, Langsrud S. Biofilm formation and the presence of the intercellular adhesion locus ica
among staphylococci from food and food processing environments. Appl Environ Microbiol
17. Predari SC, Ligozzi M, Fontana R. Genotypic identification of methicillin-resistant coagulase-negative staphylococci by polymerase chain reaction. Antimicrob Agents Chemother
18. Klingenberg C, Sundsfjord A, Ronnestad A, Mikalsen J, Gaustad P, Flaegstad T. Phenotypic and genotypic aminoglycoside resistance in blood culture isolates of coagulase-negative staphylococci from a single neonatal intensive care unit, 1989–2000. J Antimicrob Chemother
19. 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
20. Hanssen AM, Kjeldsen G, Sollid JU. Local variants of staphylococcal cassette chromosome mec
in sporadic methicillin-resistant Staphylococcus aureus
and methicillin-resistant coagulase-negative staphylococci: evidence of horizontal gene transfer? Antimicrob Agents Chemother
21. Makhoul IR, Sujov P, Smolkin T, Lusky A, Reichman B. Pathogen-specific early mortality in very low birth weight infants with late-onset sepsis: a national survey. Clin Infect Dis
22. Archer GL. Alteration of cutaneous staphylococcal flora as a consequence of antimicrobial prophylaxis. Rev Infect Dis
. 1991;13(suppl 10):S805–S809.
23. D'Angio CT, McGowan KL, Baumgart S, St Geme J, Harris MC. Surface colonization with coagulase-negative staphylococci in premature neonates. J Pediatr
24. Burnie JP, Naderi-Nasab M, Loudon KW, Matthews RC. An epidemiological study of blood culture isolates of coagulase-negative staphylococci demonstrating hospital-acquired infection. J Clin Microbiol
25. Ziebuhr W, Heilmann C, Gotz F, et al. Detection of the intercellular adhesion gene cluster (ica
) and phase variation in Staphylococcus epidermidis
blood culture strains and mucosal isolates. Infect Immun
26. Allignet J, Aubert S, Dyke KG, El SolhN. Staphylococcus caprae
strains carry determinants known to be involved in pathogenicity: a gene encoding an autolysin-binding fibronectin and the ica
operon involved in biofilm formation. Infect Immun
27. Deighton MA, Franklin JC, Spicer WJ, Balkau B. Species identification, antibiotic sensitivity and slime production of coagulase-negative staphylococci isolated from clinical specimens. Epidemiol Infect
28. Ronnestad A, Abrahamsen TG, Gaustad P, Finne PH. C-reactive protein (CRP) response patterns in neonatal septicaemia. APMIS
29. Vuong C, Voyich JM, Fischer ER, et al. Polysaccharide intercellular adhesin (PIA) protects Staphylococcus epidermidis
against major components of the human innate immune system. Cell Microbiol
30. Vuong C, Kocianova S, Voyich JM, et al. A crucial role for exopolysaccharide modification in bacterial biofilm formation, immune evasion, and virulence. J Biol Chem
31. Vuong C, Durr M, Carmody AB, Peschel A, Klebanoff SJ, Otto M. Regulated expression of pathogen-associated molecular pattern molecules in Staphylococcus epidermidis
: quorum-sensing determines pro-inflammatory capacity and production of phenol-soluble modulins. Cell Microbiol
32. Craft A, Finer N. Nosocomial coagulase negative staphylococcal (CoNS) catheter-related sepsis in preterm infants: definition, diagnosis, prophylaxis, and prevention. J Perinatol
33. Makhoul IR, Sujov P, Smolkin T, Lusky A, Reichman B. Epidemiological, clinical, and microbiological characteristics of late-onset sepsis among very low birth weight infants in Israel: a national survey. Pediatrics
34. Vermont Oxford Network Database Manual of Operations, Release 2.0.
Burlington, VT: Vermont Oxford Network; 1993.
35. Bjorkqvist M, Soderquist B, Tornqvist E, et al. Phenotypic and genotypic characterisation of blood isolates of coagulase-negative staphylococci in the newborn. APMIS
36. Benitz WE, Han MY, Madan A, Ramachandra P. Serial serum C-reactive protein levels in the diagnosis of neonatal infection. Pediatrics
37. Vandecasteele SJ, Peetermans WE, Merckx R, Rijnders BJ, Van Eldere J. Reliability of the ica, aap
genes in the discrimination between invasive, colonizing and contaminant Staphylococcus epidermidis
isolates in the diagnosis of catheter-related infections. Clin Microbiol Infect
38. Rohde H, Kalitzky M, Kroger N, et al. Detection of virulence-associated genes not useful for discriminating between invasive and commensal Staphylococcus epidermidis
strains from a bone marrow transplant unit. J Clin Microbiol
39. Makhoul I, Smolkin T, Sujov P, Kassis I, Tamir A, Sprecher H. PCR-based detection of neonatal staphylococcal infections. Pediatr Res