ARTICLE |
a Service de Bactériologie, Hôpital Cochin-Saint Vincent de Paul, Paris, France
b Service de Microbiologie, Hôpital de Bicêtre, Kremlin-Bicêtre, France
c Service de Microbiologie
g Service de Néonatologie, Hôpital Robert Debré, Paris, France
d Service de Microbiologie, Hôpital Trousseau, Paris, France
e Service de Microbiologie, Hôpital Antoine Béclère, Clamart, France
f Service de Microbiologie, Hôpital Necker Enfants-Malades, Paris, France
| ABSTRACT |
|---|
|
|
|---|
METHODS. This multicenter study was conducted in 5 hospitals in the Paris area from 1999 to 2003. We recorded all isolations of multidrug-resistant bacteria from clinical samples that were obtained from hospitalized children. Strains that were isolated during systematic screening for carriers were excluded.
RESULTS. The mean incidences were 0.9 per 1000 hospitalization-days for methicillin-resistant Staphylococcus aureus, 0.45 for extended-spectrum ß-lactamaseproducing Klebsiella pneumoniae, 0.32 for extended-spectrum ß-lactamaseproducing Enterobacteriaceae other than Klebsiella pneumoniae, 0.40 for Enterobacter species with derepressed cephalosporinase, and 0.01 for vancomycin-resistant Enterococcus. The incidences per 1000 hospitalization-days of methicillin-resistant Staphylococcus aureus, extended-spectrum ß-lactamaseproducing Klebsiella pneumoniae, extended-spectrum ß-lactamaseproducing Enterobacteriaceae other than Klebsiella pneumoniae, and Enterobacter species with derepressed cephalosporinase decreased significantly from 1999 to 2003, whereas the incidence of vancomycin-resistant Enterococcus remained very low. The proportion of resistant strains within the species did not vary significantly for methicillin-resistant Staphylococcus aureus (11% to 9.6%), extended-spectrum ß-lactamaseproducing Enterobacteriaceae other than Klebsiella pneumoniae (1.1%), and vancomycin-resistant Enterococcus (0.03% to 0.023%). In contrast, the frequency of extended-spectrum ß-lactamaseproducing Klebsiella pneumoniae decreased from 31.6% to 7.4%, and that of Enterobacter species with derepressed cephalosporinase decreased from 38.8% to 18.5%.
CONCLUSIONS. We report significant decreases in the incidence of methicillin-resistant Staphylococcus aureus, extended-spectrum ß-lactamaseproducing Klebsiella pneumoniae, extended-spectrum ß-lactamaseproducing Enterobacteriaceae other than Klebsiella pneumoniae, and Enterobacter species with derepressed cephalosporinase in hospitalized children during a 5-year period.
Key Words: multidrug-resistant bacteria children nosocomial infections
Abbreviations: MDRBmultidrug-resistant bacteria MRSAmethicillin-resistant Staphylococcus aureus VREvancomycin-resistant Enterococcus ESBLextended-spectrum ß-lactamase ESBLKpextended-spectrum ß-lactamaseproducing Klebsiella pneumoniae UTIurinary tract infection ESBLEextended-spectrum ß-lactamaseproducing Enterobacteriaceae other than K pneumoniae HDhospitalization-days
Multidrug-resistant bacteria (MDRB) present an emerging threat worldwide in hospitalized children as in adult patients.1,2 They are responsible of an increase in mortality rate and additional financial costs.3 The most clinically relevant MDRB include methicillin-resistant Staphylococcus aureus (MRSA), extended-spectrum cephalosporin-resistant Gram-negative bacilli, and vancomycin-resistant enterococci (VRE).1 MDRB colonization occurs as a result of patient-to-patient transmission and/or the selective effects of antimicrobial therapy.4 MDRB colonization/infection can be used as an indicator of compliance with hygiene measures.
Incidence of MDRB is widely known in hospitalized adults. MRSA strains accounted for 5% to 20% of the S aureus that was isolated from adult patients who were hospitalized in ICUs in Europe5 and 64.4% of those who were isolated in similar situations in the United States in 2006.6 In Europe, >20% of the Klebsiella strains that were isolated from adult patients in ICUs were found to produce an extended-spectrum ß-lactamase (ESBL; ESBL-producing Klebsiella pneumoniae [ESBLKp]) in 19987 and often are responsible of outbreaks.8 The frequency of VRE ranges from <1% in Europe to >20% in ICUs in the United States.9 However, few data are available concerning the incidence of MDRB in hospitalized children. In a study by Cosseron et al,10 11% of the S aureus strains that were isolated from NICUs and PICUs were MRSA. Kim et al11 reported that ESBLKp strains accounted for 52.9% of the Klebsiella strains that were isolated from bacteremic patients in Korea. In a previous multicenter study that was conducted in pediatric units in 8 European countries, we found that 18% of the S aureus strains were MRSA and 14% of the Klebsiella strains were ESBLKp.2 We report the results of a multicenter study that was conducted to determine the frequency of MDRB and the incidence of infections with these bacteria in hospitalized children during a 5-year period in 5 university hospitals in the Paris area.
| METHODS |
|---|
|
|
|---|
Nosocomial bloodstream infections, catheter-related infections, and urinary tract infections (UTIs) were defined as previously reported.12,13 Clinical signs of bacteremia, such as fever and chills, were always present. Urine was collected using sterile bags in children who were younger than 2 years.
The antimicrobial treatments that were administered before infection were not analyzed. Antibiotic treatment protocols for suspected nosocomial sepsis was a combination of vancomycin, a third-generation cephalosporin, and an aminoglycoside was given when the patient had an indwelling catheter; vancomycin was omitted from this combination for the treatment of suspected sepsis in patients without an indwelling catheter. The use of carbapenem was restricted to children who previously were identified as carriers of MDRB. These empiric treatment recommendations were maintained throughout the study and were applied in all hospitals.
Bacteria
All MRSA, ESBLKp, ESBL-producing Enterobacteriaceae other than K pneumoniae (ESBLE), Enterobacter species with derepressed cephalosporinase, and VRE that were isolated from clinical specimens in cases of suspected or confirmed infection were recorded. Only strains that were cultured from clinical samples that were taken 48 hours after admission were counted. Strains that were isolated during systematic screening for carriage were not considered. When multiple strains of the same species with the same antimicrobial susceptibility were recovered from a single patient, only the first isolate was included. We also recorded the sampling site of each strain.
We did not systematically record cases of bacteremia that was caused by non-MDRB. The overall frequency of MDRB among all bacteria that were isolated from patients with infection therefore was not known.
Infection Control Measures
The French National Guidelines for Control Measures that were published in 1999 were applied at each hospital.14 They included recommendations (1) to identify reservoirs of MDRB by systematic screening for MDRB in newly hospitalized patients; briefly, samples from nose and anus were taken at admission from patients in ICUs; (2) for the early notification of the clinician by the laboratory and notification of the new hospital if the patient is transferred (3) to prevent transmission between patients by means of barrier precautions for colonized and infected patients, such as antiseptic hand-washing, the wearing of gloves, and keeping infected patients isolated in single rooms or keeping patients who are infected with the same bacterium together in the same room; and (4) to organize systematic detection and monitoring of carriers. In case of an extended hospitalization, samples from nose and anus were taken each week to search MDRB.
Microbiologic Methods
The Kirby-Bauer disk diffusion method, microbroth dilution, and E-tests (AB Biodisk, Solna, Sweden) were performed according to National Committee for Clinical Laboratory Standards guidelines, using the recommended thresholds to define resistance.15 S aureus strains were considered to be resistant to methicillin when a diameter of inhibition of <20 mm was observed around a 5-µg oxacillin disk after 24 hours to 48 hours of incubation at 35°C. We used the double-disk synergy method to assess ESBL production in Enterobacteriaceae isolates as previously described.16 The derepression of cephalosporinase in Enterobacter (Enterobacter expressing chromosomal ampC ß-lactamases) was detected on the basis of cefotaxime resistance (minimum inhibitory concentration
16 µg/mL). For detection of a possible ß-lactamase, each cefotaxime-resistant Enterobacter was tested on an MH medium (Mérieux, Marcy l'Etoile, France) with 100 mg/L of cloxacillin as cephalosporinase inhibitor.15 VRE strains were detected by vancomycin and teicoplanin E-tests (AB Biodisk), conducted according to the manufacturer's instructions.
Data Analysis
We used 3 different surveillance indicators: (1) percentage of MRSA strains among isolated strains of S aureus, of ESBL Klebsiella strains among isolated Klebsiella species strains, of cefotaxime-resistant Enterobacter strains among Enterobacter species, and of VRE strains among Enterococcus species; (2) the incidence of MDRB infection per 1000 hospitalization-days (HD); and (3) the incidence of MDRB infection per 100 admissions. Changes over time in incidence per 1000 HD and per 100 admissions were analyzed from 1999 to 2003 for bacteremia, catheter-related infections, and UTI.
Statistical Analysis
Time trends were analyzed by
2 tests for linear trend on proportions and goodness of fit, comparing the observed and expected incidence rate distributions. For all tests, P < .05 was considered significant.
| RESULTS |
|---|
|
|
|---|
|
The percentages of MRSA, ESBLE, and VRE strains in the corresponding species did not vary significantly during the course of the study. Conversely, the proportion of ESBLKp decreased considerably, from 31.6% to 7.4% (P < .0001), and that of derepressed Enterobacter species decreased from 38.8% to 18.5% (P < .0001; Table 1).
The global incidences of MRSA, ESBLKp, ESBLE, and derepressed Enterobacter infections per 1000 HD decreased significantly from 1999 to 2003 (P < .001). Similarly, incidences per 100 admissions of ESBLKp, ESBLE, and derepressed Enterobacter infections decreased significantly during the same period except for MRSA.
Between 1999 and 2003, incidence per 100 admissions of ESBLKp decreased significantly in 3 centers but not significantly in 2 centers, probably in relation with the low number of isolated bacteria. The incidence of derepressed Enterobacter decreased significantly in 3 centers and not significantly in 2 centers; incidence of ESBLE decreased in 3 centers, increased in 1 (with a low number of isolated strains), and was stable for the last 1. Finally, global trend of MRSA incidence did not decrease because bacterial evolution was different according to the centers (Table 2).
|
Of the 3856 MDRB isolated, 176 (4.6%) were responsible for primary septicemia: 88 (51.2%) MRSA, 34 (19.3%) ESBLKp, 21 (11.9%) ESBLE, and 33 (18.7%) derepressed Enterobacter species. For catheter-related infections, 119 (3%) MDRB were isolated: 52 (43.7%) MRSA, 29 (24.4%) ESBLKp, 17 (14.3%) ESBLE, and 21 (17.6%) derepressed Enterobacter. For UTI, 492 (12.8%) MDRB were isolated: 54 (11%) MRSA, 137 (27.8%) ESBLKp, 131 (26.6%) ESBLE, 170 (34.5%) derepressed Enterobacter, and 1 VRE. The other 3069 MDRB were isolated from various sites, including respiratory samples and wounds, but it was not possible to differentiate between colonization and infection. No case of meningitis as a result of MDRB was recorded.
| DISCUSSION |
|---|
|
|
|---|
Numerous outbreaks of ESBLKp have been reported in several countries in adults as well as in children.2024 The prevalence of ESBLKp in adult patients is higher in Latin America (45%) and Europe (16.7% to 23%) than in the United States (8%).2527 We found a higher proportion of ESBLKp in French children (23.2%) than in French adults (11.4%).14 However, incidence of ESBLKp decreased in all centers during the study period. Conversely, our results confirm those of Hernandez et al,27 who reported a low proportion and incidence of ESBLE infections in children, in contrast to recent data from the Microbiology Surveillance Network of Northern France for adult patients.14,28 This incidence decreased in all centers except 1, in which was reported a low number of strains. The low prevalence of VRE (1%) also was reported previously in adults in other studies in Europe (0%0.3%) and France.18,29
In our study, trends in incidence per 100 admissions may have varied from center to center. When a decrease was observed in all centers for ESBLKp and derepressed E cloacae, incidence was more variable for MRSA. Therefore, it can be speculated that these findings may not be applicable to other metropolitan areas or other countries. Because this prospective study was conducted in the microbiology laboratories, data from PICUs and NICUs could not be differentiated from the pediatric units.
Antibiotic treatment may affect the incidence of infections that are caused by MDRB, MRSA, and Gram-negative bacilli.30 Colonization with resistant Gram-negative bacilli has been shown to be associated with the empiric use of amoxicillin and cefotaxime, leading to selection pressure that promotes ESBL-producing bacteria emergence.31 Cefotaxime use also may select Enterobacteriaceae with naturally derepressed cephalosporinase. Restriction of the antibiotic use can lead to a decrease in MDRB colonization.32,33 During the study, no changes in antimicrobial drug use, which could explain the MDRB decreased incidence, were observed in the various pediatric departments.
The major reservoir of MDRB is colonized/infected adult or pediatric patients. The hand carriage accounts for the major mechanism for patient-to-patient transmission.34 Wendt et al35 stated that bacterial overgrowth in the gut of patients in the ICU is a relevant mechanism to acquired carriage of MDRB during a nonoutbreak situation.35 Therefore, barrier precautions are important to avoid patient-to-patient transmission. The national guidelines, including recommendations similar to those of the US Hospital Infection Control Practices Advisory Committee, were distributed to all hospitals.36 However, no external audit was conducted to assess the validity of hygiene control procedures.
Methicillin-resistant coagulase-negative Staphylococcus, the pathogen that most commonly is responsible for systemic infection that is associated with a central venous catheter in children, was excluded from this analysis. The inclusion of methicillin-resistant coagulase-negative Staphylococcus in future studies of MDRB infections in children would be of clinical relevance.
| CONCLUSIONS |
|---|
|
|
|---|
| FOOTNOTES |
|---|
Address correspondence to Josette Raymond, MD, PhD, Service de Bactériologie, Hôpital Cochin-Saint Vincent de Paul, 27 rue du faubourg Saint-Jacques, 75679 Paris Cedex 14, France. E-mail: j.raymond{at}svp.ap-hop-paris.fr
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||