- Email: [email protected]
Emergence of Enterobacter cloacae as a common pathogen in neonatal units: pulsed-field gel electrophoresis analysis
Journal of Hospital Infection (2004) 57, 119–125
www.elsevierhealth.com/journals/jhin
Emergence of Enterobacter cloacae as a common pathogen in neonatal units: pulsed-field gel electrophoresis analysis D. Talona,*, P. Mengetb, M. Thouvereza, G. Thiriezb, H. Gbaguidi Haorea, C. Fromentinb, A. Mullera, X. Bertranda a
`ne Hospitalie `re et d’Epide ´miologie mole ´culaire, 25030 Besanc¸on cedex, France Service d’Hygie ´onatologie, Centre Hospitalier Universitaire Jean Minjoz, 25030 Besanc¸on cedex, France Service de ne
b
Received 12 August 2003; accepted 21 January 2004 Available online 13 May 2004
KEYWORDS Enterobacter cloacae; Molecular epidemiology; Neonatal units; Antibiotic use
Summary In the first week of December 2002, three infants hospitalized in the neonatal department of our hospital had blood cultures positive with Enterobacter cloacae. Screening cultures and genotyping showed that 10 of 25 screened patients also carried E. cloacae and that nine isolates belonged to the same clone as that responsible for all three bacteraemias. This epidemic cluster was limited to one of the two units of the department. Surveillance of both units continued until the end of March 2003; 51 of 159 neonates screened were colonized with E. cloacae, 38 out of 80 (47.5%) in the premature unit (PU) and 13 out of 79 (16.4%) in the paediatric intensive care unit (PICU). Pulsed-field gel electrophoresis (PFGE) analysis of 130 available isolates revealed 30 different pulsotypes, including 24 unique pulsotypes from individual patients and six from multiple patients. Antibiotic (particularly b-lactam) use did not significantly vary from 1999 to 2003. The consumption of alcohol-based hand rub (four-fold higher in the PICU than in the PU) and nurse-to-patient ratio (1:2 in the PICU and 1:4 in the PU) might explain the higher cross-transmission rate in the PU. Finally, despite an epidemiological survey, we failed to identify the causes of the emergence of E. cloacae in our neonatology units. However, improved hygiene practices combined with restriction of admission led to the progressive disappearance of the epidemic strain. The increasing importance of this type of unit and the dramatic consequences of infections emphasize the need for additional research on the constitution of the flora of newborns and the mode of acquisition Gramnegative multi-resistant bacteria. Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
*Corresponding author. Tel.: þ33-381-668286; fax: þ33-381-668914. E-mail address: [email protected] 0195-6701/$ - see front matter Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2004.01.023
120
Introduction In recent years, Enterobacter cloacae has emerged as an important nosocomial pathogen in neonatal units,1,2 with numerous outbreaks of infection being reported.3 – 8 This bacterium may be transmitted to neonates through intravenous fluids,9 total parenteral nutrition solutions10 and medical equipment.11 Many single-clone outbreaks probably caused by cross-transmission via healthcare workers have been described, suggesting that inpatients are the major reservoir.3 – 5,7,8,11 and that contamination from an environmental reservoir is the exception rather than the rule. Common endogenous reservoirs for E. cloacae include the gastrointestinal tract of healthy adults and the urinary and respiratory tracts of sick patients. E. cloacae is isolated from the faeces of 10 – 70% of neonates.11 – 13 However, the mode of acquisition during the first days of life and the emergence of this pathogen in nosocomial infection remain partially unknown, even if broad-spectrum antibiotics such as third-generation cephalosporins are probably at the root of this. Indeed, there appears to be a positive correlation between previous use of third-generation cephalosporins and the rates of colonization and infection caused by E. cloacae.14 – 16 Due to their relative lack of toxicity and their ability to cross the blood– brain barrier, these antimicrobial agents have been increasingly used as first-line antibiotic therapy in neonates.17 The status of our neonatal unit changed in 1999 when it became the regional reference centre. Thus means patients were transferred in utero for highrisk pregnancies and all ‘immunocompromised’ neonates were admitted to our unit. Here, we report the microbiological investigation of an outbreak of E. cloacae in the unit and the emergence of E. cloacae as a major pathogen responsible for neonatal infections.
D. Talon et al.
gestational weeks). The department contains two units: a 15-bed paediatric intensive care unit (PICU) and a 24-bed unit called the premature unit (PU). The PU is further divided into a 12-bed critical-care facility and a 12-bed intermediate-care nursery. Premature neonates with severe organ dysfunction are transferred to the PICU for life-support treatment and return to the PU or other facilities when their condition stabilizes.
Bacteriological culture and antibiotic susceptibility testing
Methods
Enterobacteriaceae from diagnostic cultures were identified to the species level by biochemical characteristic analysis (API 20E and ID 32 E Strips, bioMe ´rieux, Marcy l’Etoile, France). At the start of the investigation (December 2002), rectal swabs were collected once a week from all the children in the neonatal unit. The swabs were inoculated into brain – heart broth (Bio-Rad, Ivry sur Seine, France) supplemented with 20 mg/L amoxicillin and 10 mg/ L clavulanic acid. The cultures were incubated at 37 8C for 24 h. All positive broth cultures were subcultured onto Drigalski agar plates (Oxoid, Basingstoke, UK) supplemented with 20 mg/L cephalothin and incubated at 37 8C for 24 h. Colonies morphologically resembling Enterobacteriaceae were primarily identified by Gram staining and a negative oxidase test. E. cloacae was identified by use of API 20E strips. Susceptibilities to common antibiotics were determined by disk diffusion. Isolates were classified as being susceptible, intermediate or resistant according to the criteria recommended by the Antibiogram Committee of the French Microbiology Society.18 All E. cloacae showing reduced susceptibility or resistance to ceftazidime (, 22 mm for 30 mg disk) and/or cefotaxime (, 27 mm for 30 mg disk) and/or aztreonam (, 27 mm for 30 mg disk) were also tested for extended-spectrum b-lactamase (ESBL) production by a classical double-disk diffusion test.18 According to the production of ESBL, isolates were classified as being high-level blactamase producers (HLBL) or ESBL producers.
Setting
Genotyping
Besanc ¸on is the largest city in the Franche-Comte ´ (130 000 inhabitants), a region of eastern France with approximately one million inhabitants. Besan¸ con Hospital is a 1220-bed university-affiliated public hospital. The neonatal department is a 39bed unit that admits approximately 1000 patients per year, accounting for 10 000 patient-days. Approximately 75% were preterm infants (, 32
All available E. cloacae isolated from November 2002 to March 2003 were typed. For this, the macrorestriction profile of total DNA (pulsotype) was determined by pulsed-field gel electrophoresis (PGFE; CHEF DRIII, Bio-Rad), according to a technique previously developed in our laboratory.19 Epidemiological relatedness was determined according to the criteria published by Tenover and
Epidemiology of E. cloacae in neonatal units colleagues.20 Isolates with indistinguishable PFGE patterns were assigned to the same pulsotype and subtype. Isolates that differed by less than three bands were considered to belong to different subtypes and isolates differing by more than three bands were considered to belong to different pulsotypes.
Environmental survey Multiple sites within the wards were sampled to seek an environmental source of E. cloacae. These sites included tap water, sinks, plug holes, tap handles, handles of the liquid soap dispensers, incubators, bottle warmers, disinfectants, soaps, feeding equipment (enteral feeding, nasogastric tubes, maternal milk), and all the medical equipment (electrodes, scopes, thermometers). Standard methods and media were used to isolate organisms.
Use of antibiotics and alcohol-based hand rub The quantities of each antimicrobial agent delivered to the neonatal unit from January 1999 to December 2003 were obtained from the pharmacy computer system. The amounts of antimicrobials used were converted into defined daily doses (DDD), following the recommendations of the WHO.21 The use of various classes of antibiotics was expressed in DDD per 1000 days of hospitalization. We also recorded the consumption of alcohol-based hand rub and expressed it in litres per 1000 days of hospitalization.
Definitions The following definitions were used: a case was defined as a patient from whom E. cloacae was recovered from any site irrespective of the sample type (screening or clinical diagnostic); carrier was defined as a patient from whom E. cloacae was recovered from screening samples in the absence of or before the isolation of E. cloacae from clinical diagnostic samples, and colonized/infected patient, a patient from whom E. cloacae was recovered from clinical samples. We did not collect any clinical information allowing us to differentiate infections from simple colonizations.
121
because blood cultures from three infants hospitalized in the PU were positive for E. cloacae in the same week. Several measures were taken to investigate the outbreak: (1) we retrospectively collected data concerning the frequency of E. cloacae isolation among clinical samples in the neonatal department and within the entire hospital from 1999, (2) we implemented weekly screening cultures to identify all E. cloacae carriers, (3) we performed an environmental survey, (4) we used PFGE to genotype all the isolates of E. cloacae collected, (5) we collected data on antibiotic use, and (6) we evaluated hand hygiene practices by recording the quantity of alcohol-based hand rub used.
Results of the investigations The retrospective study revealed that E. cloacae is a major pathogen in our neonatal units. The frequency of E. cloacae among all bacteria isolated from clinical samples (colonization/infection) increased considerably over time in the PU, but remained stable in the PICU and all other units (Table I). Figure 1 shows the number of patients colonized/infected according to the type of neonatal unit (PICU versus PU). From January 1999 to March 2003, 10 infants had E. cloacae bacteraemias: one in 2000, three in 2001, four in 2002 and two in 2003 (January to March; Table II). The combination of screening cultures and genotyping showed that, at the beginning of the survey, 10 of the 25 screened patients were positive with E. cloacae; nine isolates belonged to the same clone. The latter were all recovered in the PU, and the unrelated isolate was recovered in the PICU. Screening was continued until the end of March 2003. Fifty-one of the 159 newborns screened were
Results Outbreak alert and implementation of an investigation In December 2002, an intervention was initiated
Figure 1 Number of patients colonized/infected with E. cloacae from January 1999 to March 2003. Open bars, paediatric intensive care unit; filled bars, premature unit.
122
D. Talon et al.
Table I Frequency of Enterobacter cloacae among all bacterial isolations in clinical samples Year
Frequency of E. cloacae Neonatal unit
1999 2000 2001 2002 2003 (from Jan to March)
All other units in the hospital (%)
PICU (%)
Premature unit (%)
6.4 3.2 8.2 7.5 12.8
1.8 2.8 4.8 13.6 11.1
2 1.7 1.7 1.9 1.6
PICU, paediatric intensive care unit.
colonized with E. cloacae, 38 out of 80 in the PU (47.5%) and 13 out of 79 (16.4%) in the PICU. In total, 130 isolates were available for genotyping. Of them, 110 were isolated from screening cultures (from 51 patients), 14 from clinical cultures collected during the investigation and six from stored clinical cultures collected from patients discharged before the investigation began (four blood cultures). PFGE analysis of genomic DNA revealed 30 different pulsotypes, including 24 unique pulsotypes from individual patients and six multiple patterns (Table III). Unrelated strains of E. cloacae examined always had more than six differences in band profiles. Six patients were colonized and/or infected with two different pulsotypes, and one patient was colonized and/or infected with three different pulsotypes. Eleven patients had both screening and clinical cultures positive for E. cloacae and the isolates had similar pulsotypes in nine of them. During the environmental survey, 115 samples were analysed for E. cloacae contamination. This organism was isolated from nasogastric tubes of five
infants. No other environmental sites were contaminated with E. cloacae.
Antibiotic susceptibility results None of the isolates produced ESBL. Table III lists the main results of the antibiograms for ‘wild-type’ (WT) and HLBL. The isolates were susceptible to the other major antibiotic classes such as aminoglycosides and fluoroquinolones.
Antibiotic use Antibiotic use, particularly that of b-lactams, did not vary significantly during the survey period. The average use was 519.3 DDD/1000 hospital-days in the PICU and 70.7 in the PU. The only exception was that use of the third-generation cephalosporins increased in the PU from October 2001. This last use remains low, at , 20 DDD/1000 hospital-days (Figure 2).
Table II Characteristics of E. cloacae isolated from blood cultures in the neonatal department Patient 1 2 3 4 5 6 7 8 9 10 a
Date of positive blood culture
Other sites positive for E. cloacae (date)
Unit
Clone
19/02/2000 14/05/2001 08/12/2001 08/12/2001 28/05/2002 03/12/2002 03/12/2002 06/12/2002 11/02/2003 26/02/2003
Urine (16/02/00)a Nonea Wounda (12/12/01) Nonea Nonea Nonea Nonea Nonea Rectum (10/02/03) Rectum (24/02/03)
PICU PICU PU PICU PICU PU PU PU PU PU
ND Spb Spb Spb Spb A A A Spc Cc
As screening programme had not been started at the time of the isolation, we do not know whether these patients were carriers. As screening programme had not been started at the time of the isolation, we cannot draw conclusions concerning the sporadic nature of these strains. c Carriage and clinical isolates had similar pulsed-field gel electrophoresis patterns. PICU, paediatric intensive care unit; PU, premature unit. b
Epidemiology of E. cloacae in neonatal units
123
Table III Characteristics of the different clones of Enterobacter cloacae Clones of E. cloacae
Epidemic clones Clone A Clone B Clone C Clone D Clone E Clone F Sporadic clones
Number of positive patientsa
Screening/ clinical cultureb
23 11 4 4 3 2 24
18/6 11/1 4/1 4/1 3/0 2/0 17/8
Unit of isolation: PU/PICU
22/1 9/2 4/0 2/2 0/3 2/0 13/11
Period of spreadc
12/12/02–07/02/03g 13/01/03–31/03/03 17/02/03–10/03/03 30/12/02–20/01/03 25/12/03–13/01/03 21/12/02–03/01/03 –
Antibiotic resistance patternd
Wild-typee
High-level b-lactamase-producerf
17 4 0 0 1 2 15
6 7 4 4 2 0 9
Considering only the first E. cloacae isolate perform each patient, except if the patient was positive for several clones. a Six patients were colonized and/or infected with two different pulsotypes and one patient was colonized and/or infected with three different pulsotypes. b Some patients had both clinical and screening cultures positive for the same clone, some patients had only clinical culture because they were discharged before the epidemiological study began. c Established from the microbiological isolation dates, without considering the date of stay of the positive patients. d When several isolates belonging to the same clone were consecutively obtained from one patient, we report the first antibiotype. e Wild-type: resistant to amoxicillin, amoxicillin/clavulanic acid, ticarcillin and cefalothin. f Resistant or intermediate to ceftazidime or cefotaxime and extended-spectrum b-lactamase-negative. g This clone probably disseminated before the 12/12/02, date corresponding to the start of the survey. PICU, paediatric intensive care unit; PU, premature unit.
Use of alcohol-based hand rub The mean consumption of alcohol-based hand rub was 64.6, 16.0 and 9.3 L/1000 hospitalization-days in the PICU, the PU and the other units of the hospital, respectively (Figure 3).
Infection control measures
reinforced. Furthermore, carriers and colonized/ infected patients were kept separate from other patients and admissions were restricted.
Discussion
The infection control measures used by the units were reviewed, and medical and nursing staff were urged to comply with these measures. In particular, optimal hand disinfection practices were
This investigation was prompted by the concomitant identification of three E. cloacae-positive blood cultures in a neonatal unit. These preliminary results suggest that a ‘classical’ clonal outbreak occurred in one unit of our neonatal department. This epidemic strain spread to the PU as a
Figure 2 Antibiotic use in the premature unit from January 1999 to December 2002. (—) Penicillins; (---) third-generation cephalosporins; ( – – – ) others.
Figure 3 Alcohol-based hand rub solution use in Besan¸ con hospital in 2002. Each point represents one hospital unit; (…) mean; (—) median.
124
consequence of patient-to-patient transmission. A central concept in preventing cross-transmission is good compliance with hand hygiene procedures by healthcare workers after handling each patient.22 The consumption of alcohol-based hand rub was fourfold higher in the PICU than in the PU. This difference might explain the higher cross-transmission rate in the PU. It is also likely, that, as in other studies,3,4 understaffing and overcrowding played a role in disseminating the epidemic strain. Indeed, the nurse-to-patient ratio was 1:2 in the PICU and to 1:4 in the PU, although this difference does not reflect the difference in workload between the two units. We were unable to identify an environmental source of contamination. The only sites that tested positive for the epidemic strain were nasogastric tubes and these represent a simple consequence of intestinal carriage rather than a source of transmission of the strain. Moreover, as in other reports, the improvement of hygiene practices combined with the restriction of admissions led to the progressive disappearance of the epidemic strain.3 – 5 The surveillance programme (screening and genotyping), which was continued after the major epidemic episode, showed that the emergence of E. cloacae was not limited to the sequential spread of epidemic strains. Our study improved our knowledge of outbreaks of E. cloacae in NICUs by determining the genetic relatedness of clinical and screening isolates over a long period (four months). Clinical cultures do not present the full epidemiological picture and surveillance cultures are essential for an understanding of the mechanisms involved in the establishment and maintenance of E. cloacae colonization. Genotyping, particularly PFGE, is a very useful tool for identification, tracing and detailed analysis of the general epidemiology of a bacterial species within a unit. Our study shows that in our neonatal department, E. cloacae caused a succession of colonization outbreaks affecting from two to over 20 patients5 involving transmission by healthcare workers who failed to wash their hands adequately, in a background of sporadic cases. The epidemic potential of a given strain probably depends on the characteristics of the patient,3 the compliance with infection control measures by medical and nursing staff,3,4,6 the antibiotic use policy23 and the intrinsic properties of the strain such as antibiotic resistance,14 – 17 adhesion, fitness and survival in the environment.4 The increased use of third-generation cephalosporins in the PU (Figure 2) may have favoured the emergence of E. cloacae. However, its use remained low compared with that in the PICU (20 versus 70 DDD/1000 HD). Moreover, most of the
D. Talon et al.
isolates were susceptible to this class of antibiotics. So, antibiotic selective pressure can only partially explain the emergence of E. cloacae. There appears to be a correlation between previous use of thirdgeneration cephalosporins and the increase in the rates of colonization and infection caused by E. cloacae,14 – 17 but the strains used in these studies were HLBL. The use of narrow-spectrum antibiotics reduces bacterial resistance,23 but our clinicians found that the data concerning the role of thirdgeneration cephalosporins were not conclusive enough to limit the use of these important agents. More preterm infants in need of aggressive invasive therapy are being admitted to neonatal intensive care units than ever before. These patients are admitted to NICUs immediately after birth, often after delivery by Caesarean section. They are at great risk of becoming colonized by the strains present in their environment. The babies’ bacterial flora is acquired through contacts with the environment, which is constituted by the mother during the delivery and by healthcare workers and other relations during the hospitalization period. Antibiotics exert a selective pressure favouring blactamase-producing bacteria such as E. cloacae. This pressure is exerted within the NICUs and even during the pregnancy if antibiotics are prescribed (for example for urinary tract infection or preterm premature rupture of membranes). Contamination may occur during delivery. The frequency of detection of E. cloacae could therefore reflect the mothers’ intestinal flora. Women with high-risk pregnancies have a higher probability of requiring antibiotic treatment, with subsequent modification of the Enterobacteriaceae residing in the intestinal tract and on various mucous membranes. This would explain the polyclonal nature of E. cloacae isolates outside of clonal outbreaks. Our investigations concerning the frequency of E. cloacae intestinal carriage in patients in neonatal units do not allow us to conclude that the incidence of intestinal colonization increased concomitantly with that of infections. We cannot rule out the possibility that changes in the average status of the patients (prematurity, very low birthweight) increased the infectious risk among patients whose intestines were colonized, without there being a true increase in the frequency of such patients.
References 1. Sanders Jr WE, Sanders CC. Enterobacter spp: pathogens poised to flourish at the turn of the century. Clin Microbiol Rev 1997;10:220—241.
Epidemiology of E. cloacae in neonatal units
2. Gaston MA. Enterobacter: an emerging nosocomial pathogen. J Hosp Infect 1988;11:197—208. 3. Harbath S, Sudre P, Dharan S, Cadenas M, Pittet D. Outbreak of Enterobacter cloacae related to understaffing, overcrowding and poor hygienes practices. Infect Control Hosp Epidemiol 1999;20:598—603. 4. van Dijk Y, Bik EM, Hochstenbach-Vernooij S, et al. Management of an outbreak of Enterobacter cloacae in a neonatal unit using simple preventive measures. J Hosp Infect 2002; 51:21—26. 5. Fernandez-Baca V, Ballesteros F, Hervas JA, et al. Molecular epidemiological typing of Enterobacter cloacae from a neonatal intensive care unit: three-year prospective study. J Hosp Infect 2001;49:173—182. 6. Verweij PE, Van Belkum A, Melchers WJ, Voss A, HoogkampKorstanje JA, Meis JF. Interrepeat fingerprinting of thirdgeneration cephalosporin-resistant Enterobacter cloacae isolated during an outbreak in a neonatal intensive care unit. Infect Control Hosp Epidemiol 1995;16:25—29. 7. Liu SC, Leu HS, Yen MY, Lee PI, Chou MC. Study of an outbreak of Enterobacter cloacae sepsis in a neonatal intensive care unit: the application of epidemiologic chromosome profiling by pulsed-field gel electrophoresis. Am J Infect Control 2002;30:381—385. 8. van Nierop WH, Duse AG, Stewart RG, Bilgeri YR, Koornhof HJ. Molecular epidemiology of an outbreak of Enterobacter cloacae in the neonatal intensive care unit of a provincial hospital in Gauteng, South Africa. J Clin Microbiol 1998;36: 3085—3087. 9. Tresoldi AT, Padoveze MC, Trabasso P, et al. Enterobacter cloacae sepsis outbreak in a newborn unit caused by contaminated total parenteral nutrition solution. Am J Infect Control 2000;28:258—261. 10. van den Berg RWA, Claahsen HL, Niessen M, Muytjens HL, Liem K, Voss A. Enterobacter cloacae outbreak in the NICU related to disinfected thermometers. J Hosp Infect 1999;45: 29—34. 11. Acolet D, Ahmet Z, Houang E, Hurley R, Kaufmann ME. Enterobacter cloacae in a neonatal intensive care unit: account of an outbreak and its relationship to use of thirdgeneration cephalosporins. J Hosp Infect 1994;28:273—286. 12. Goldmann DA, Leclair J, Macone A. Bacterial colonization of
125
13.
14.
15.
16.
17.
18.
19.
20.
21.
22. 23.
neonates admitted to an intensive care environment. J Pediatr 1978;93:288—293. Burman LG, Berglund B, Huovinen P, Tullus K. Effect of ampicillin versus cefuroxime on the emergence of b-lactam resistance in faecal Enterobacter cloacae isolates from neonates. J Antimicrob Chemother 1993;31:111—116. Chow JW, Fine MJ, Shlaes DM, et al. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med 1991;115: 585—590. Finnstrom O, Isaksson B, Haeggman S, Burman LG. Control of an outbreak of a highly b-lactam-resistant Enterobacter cloacae strain in a neonatal special care unit. Acta Paediatr 1998;87:1070—1074. Calil R, Marba ST, von Nowakonski A, Tresoldi AT. Reduction in colonization and nosocomial infection by multiresistant bacteria in a neonatal unit after institution of educational measures and restriction in the use of cephalosporins. Am J Infect Control 2001;29:133—138. Kaplan SL, Patrick CC. Cefotaxime and aminoglycoside treatment of meningitidis caused by Gram-negative enteric organisms. Pediatr Infect Dis J 1990;9:810—814. Members of the SFM Antibiogram Committee. Comite ´ de l’antibiogramme de la Socie ¸aise de Microbiologie ´te ´ Franc Report 2003. Int J Antimicrob Agents 2003;21:364—391. Talon D, Dupont MJ, Lesne J, Thouverez M, Michel-Briand Y. Pulsed-field gel electrophoresis as an epidemiological tool for clonal identification of Aeromonas hydrophila. J Appl Bacteriol 1996;80:277—282. Tenover FC, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA restriction patterns produced by pulsedfield gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995;33:2233—2239. Natsch S, Hekster YA, de Jong R, Heerdink ER, Herings RM, van der Meer JW. Application of the ATC/DDD methodology to monitor antibiotic drug use. Eur J Clin Microbiol Infect Dis 1998;17:20—24. Jarvis WR. Handwashing—the Semmelweis lesson forgotten? Lancet 1994;344:1311—1312. de Man P, Verhoeven BAN, Verbrugh HA, Vos MC, van den Anker JN. An antibiotic policy to prevent emergence of resistant bacilli. Lancet 2000;355:973—978.
www.elsevierhealth.com/journals/jhin
Emergence of Enterobacter cloacae as a common pathogen in neonatal units: pulsed-field gel electrophoresis analysis D. Talona,*, P. Mengetb, M. Thouvereza, G. Thiriezb, H. Gbaguidi Haorea, C. Fromentinb, A. Mullera, X. Bertranda a
`ne Hospitalie `re et d’Epide ´miologie mole ´culaire, 25030 Besanc¸on cedex, France Service d’Hygie ´onatologie, Centre Hospitalier Universitaire Jean Minjoz, 25030 Besanc¸on cedex, France Service de ne
b
Received 12 August 2003; accepted 21 January 2004 Available online 13 May 2004
KEYWORDS Enterobacter cloacae; Molecular epidemiology; Neonatal units; Antibiotic use
Summary In the first week of December 2002, three infants hospitalized in the neonatal department of our hospital had blood cultures positive with Enterobacter cloacae. Screening cultures and genotyping showed that 10 of 25 screened patients also carried E. cloacae and that nine isolates belonged to the same clone as that responsible for all three bacteraemias. This epidemic cluster was limited to one of the two units of the department. Surveillance of both units continued until the end of March 2003; 51 of 159 neonates screened were colonized with E. cloacae, 38 out of 80 (47.5%) in the premature unit (PU) and 13 out of 79 (16.4%) in the paediatric intensive care unit (PICU). Pulsed-field gel electrophoresis (PFGE) analysis of 130 available isolates revealed 30 different pulsotypes, including 24 unique pulsotypes from individual patients and six from multiple patients. Antibiotic (particularly b-lactam) use did not significantly vary from 1999 to 2003. The consumption of alcohol-based hand rub (four-fold higher in the PICU than in the PU) and nurse-to-patient ratio (1:2 in the PICU and 1:4 in the PU) might explain the higher cross-transmission rate in the PU. Finally, despite an epidemiological survey, we failed to identify the causes of the emergence of E. cloacae in our neonatology units. However, improved hygiene practices combined with restriction of admission led to the progressive disappearance of the epidemic strain. The increasing importance of this type of unit and the dramatic consequences of infections emphasize the need for additional research on the constitution of the flora of newborns and the mode of acquisition Gramnegative multi-resistant bacteria. Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved.
*Corresponding author. Tel.: þ33-381-668286; fax: þ33-381-668914. E-mail address: [email protected] 0195-6701/$ - see front matter Q 2004 The Hospital Infection Society. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.jhin.2004.01.023
120
Introduction In recent years, Enterobacter cloacae has emerged as an important nosocomial pathogen in neonatal units,1,2 with numerous outbreaks of infection being reported.3 – 8 This bacterium may be transmitted to neonates through intravenous fluids,9 total parenteral nutrition solutions10 and medical equipment.11 Many single-clone outbreaks probably caused by cross-transmission via healthcare workers have been described, suggesting that inpatients are the major reservoir.3 – 5,7,8,11 and that contamination from an environmental reservoir is the exception rather than the rule. Common endogenous reservoirs for E. cloacae include the gastrointestinal tract of healthy adults and the urinary and respiratory tracts of sick patients. E. cloacae is isolated from the faeces of 10 – 70% of neonates.11 – 13 However, the mode of acquisition during the first days of life and the emergence of this pathogen in nosocomial infection remain partially unknown, even if broad-spectrum antibiotics such as third-generation cephalosporins are probably at the root of this. Indeed, there appears to be a positive correlation between previous use of third-generation cephalosporins and the rates of colonization and infection caused by E. cloacae.14 – 16 Due to their relative lack of toxicity and their ability to cross the blood– brain barrier, these antimicrobial agents have been increasingly used as first-line antibiotic therapy in neonates.17 The status of our neonatal unit changed in 1999 when it became the regional reference centre. Thus means patients were transferred in utero for highrisk pregnancies and all ‘immunocompromised’ neonates were admitted to our unit. Here, we report the microbiological investigation of an outbreak of E. cloacae in the unit and the emergence of E. cloacae as a major pathogen responsible for neonatal infections.
D. Talon et al.
gestational weeks). The department contains two units: a 15-bed paediatric intensive care unit (PICU) and a 24-bed unit called the premature unit (PU). The PU is further divided into a 12-bed critical-care facility and a 12-bed intermediate-care nursery. Premature neonates with severe organ dysfunction are transferred to the PICU for life-support treatment and return to the PU or other facilities when their condition stabilizes.
Bacteriological culture and antibiotic susceptibility testing
Methods
Enterobacteriaceae from diagnostic cultures were identified to the species level by biochemical characteristic analysis (API 20E and ID 32 E Strips, bioMe ´rieux, Marcy l’Etoile, France). At the start of the investigation (December 2002), rectal swabs were collected once a week from all the children in the neonatal unit. The swabs were inoculated into brain – heart broth (Bio-Rad, Ivry sur Seine, France) supplemented with 20 mg/L amoxicillin and 10 mg/ L clavulanic acid. The cultures were incubated at 37 8C for 24 h. All positive broth cultures were subcultured onto Drigalski agar plates (Oxoid, Basingstoke, UK) supplemented with 20 mg/L cephalothin and incubated at 37 8C for 24 h. Colonies morphologically resembling Enterobacteriaceae were primarily identified by Gram staining and a negative oxidase test. E. cloacae was identified by use of API 20E strips. Susceptibilities to common antibiotics were determined by disk diffusion. Isolates were classified as being susceptible, intermediate or resistant according to the criteria recommended by the Antibiogram Committee of the French Microbiology Society.18 All E. cloacae showing reduced susceptibility or resistance to ceftazidime (, 22 mm for 30 mg disk) and/or cefotaxime (, 27 mm for 30 mg disk) and/or aztreonam (, 27 mm for 30 mg disk) were also tested for extended-spectrum b-lactamase (ESBL) production by a classical double-disk diffusion test.18 According to the production of ESBL, isolates were classified as being high-level blactamase producers (HLBL) or ESBL producers.
Setting
Genotyping
Besanc ¸on is the largest city in the Franche-Comte ´ (130 000 inhabitants), a region of eastern France with approximately one million inhabitants. Besan¸ con Hospital is a 1220-bed university-affiliated public hospital. The neonatal department is a 39bed unit that admits approximately 1000 patients per year, accounting for 10 000 patient-days. Approximately 75% were preterm infants (, 32
All available E. cloacae isolated from November 2002 to March 2003 were typed. For this, the macrorestriction profile of total DNA (pulsotype) was determined by pulsed-field gel electrophoresis (PGFE; CHEF DRIII, Bio-Rad), according to a technique previously developed in our laboratory.19 Epidemiological relatedness was determined according to the criteria published by Tenover and
Epidemiology of E. cloacae in neonatal units colleagues.20 Isolates with indistinguishable PFGE patterns were assigned to the same pulsotype and subtype. Isolates that differed by less than three bands were considered to belong to different subtypes and isolates differing by more than three bands were considered to belong to different pulsotypes.
Environmental survey Multiple sites within the wards were sampled to seek an environmental source of E. cloacae. These sites included tap water, sinks, plug holes, tap handles, handles of the liquid soap dispensers, incubators, bottle warmers, disinfectants, soaps, feeding equipment (enteral feeding, nasogastric tubes, maternal milk), and all the medical equipment (electrodes, scopes, thermometers). Standard methods and media were used to isolate organisms.
Use of antibiotics and alcohol-based hand rub The quantities of each antimicrobial agent delivered to the neonatal unit from January 1999 to December 2003 were obtained from the pharmacy computer system. The amounts of antimicrobials used were converted into defined daily doses (DDD), following the recommendations of the WHO.21 The use of various classes of antibiotics was expressed in DDD per 1000 days of hospitalization. We also recorded the consumption of alcohol-based hand rub and expressed it in litres per 1000 days of hospitalization.
Definitions The following definitions were used: a case was defined as a patient from whom E. cloacae was recovered from any site irrespective of the sample type (screening or clinical diagnostic); carrier was defined as a patient from whom E. cloacae was recovered from screening samples in the absence of or before the isolation of E. cloacae from clinical diagnostic samples, and colonized/infected patient, a patient from whom E. cloacae was recovered from clinical samples. We did not collect any clinical information allowing us to differentiate infections from simple colonizations.
121
because blood cultures from three infants hospitalized in the PU were positive for E. cloacae in the same week. Several measures were taken to investigate the outbreak: (1) we retrospectively collected data concerning the frequency of E. cloacae isolation among clinical samples in the neonatal department and within the entire hospital from 1999, (2) we implemented weekly screening cultures to identify all E. cloacae carriers, (3) we performed an environmental survey, (4) we used PFGE to genotype all the isolates of E. cloacae collected, (5) we collected data on antibiotic use, and (6) we evaluated hand hygiene practices by recording the quantity of alcohol-based hand rub used.
Results of the investigations The retrospective study revealed that E. cloacae is a major pathogen in our neonatal units. The frequency of E. cloacae among all bacteria isolated from clinical samples (colonization/infection) increased considerably over time in the PU, but remained stable in the PICU and all other units (Table I). Figure 1 shows the number of patients colonized/infected according to the type of neonatal unit (PICU versus PU). From January 1999 to March 2003, 10 infants had E. cloacae bacteraemias: one in 2000, three in 2001, four in 2002 and two in 2003 (January to March; Table II). The combination of screening cultures and genotyping showed that, at the beginning of the survey, 10 of the 25 screened patients were positive with E. cloacae; nine isolates belonged to the same clone. The latter were all recovered in the PU, and the unrelated isolate was recovered in the PICU. Screening was continued until the end of March 2003. Fifty-one of the 159 newborns screened were
Results Outbreak alert and implementation of an investigation In December 2002, an intervention was initiated
Figure 1 Number of patients colonized/infected with E. cloacae from January 1999 to March 2003. Open bars, paediatric intensive care unit; filled bars, premature unit.
122
D. Talon et al.
Table I Frequency of Enterobacter cloacae among all bacterial isolations in clinical samples Year
Frequency of E. cloacae Neonatal unit
1999 2000 2001 2002 2003 (from Jan to March)
All other units in the hospital (%)
PICU (%)
Premature unit (%)
6.4 3.2 8.2 7.5 12.8
1.8 2.8 4.8 13.6 11.1
2 1.7 1.7 1.9 1.6
PICU, paediatric intensive care unit.
colonized with E. cloacae, 38 out of 80 in the PU (47.5%) and 13 out of 79 (16.4%) in the PICU. In total, 130 isolates were available for genotyping. Of them, 110 were isolated from screening cultures (from 51 patients), 14 from clinical cultures collected during the investigation and six from stored clinical cultures collected from patients discharged before the investigation began (four blood cultures). PFGE analysis of genomic DNA revealed 30 different pulsotypes, including 24 unique pulsotypes from individual patients and six multiple patterns (Table III). Unrelated strains of E. cloacae examined always had more than six differences in band profiles. Six patients were colonized and/or infected with two different pulsotypes, and one patient was colonized and/or infected with three different pulsotypes. Eleven patients had both screening and clinical cultures positive for E. cloacae and the isolates had similar pulsotypes in nine of them. During the environmental survey, 115 samples were analysed for E. cloacae contamination. This organism was isolated from nasogastric tubes of five
infants. No other environmental sites were contaminated with E. cloacae.
Antibiotic susceptibility results None of the isolates produced ESBL. Table III lists the main results of the antibiograms for ‘wild-type’ (WT) and HLBL. The isolates were susceptible to the other major antibiotic classes such as aminoglycosides and fluoroquinolones.
Antibiotic use Antibiotic use, particularly that of b-lactams, did not vary significantly during the survey period. The average use was 519.3 DDD/1000 hospital-days in the PICU and 70.7 in the PU. The only exception was that use of the third-generation cephalosporins increased in the PU from October 2001. This last use remains low, at , 20 DDD/1000 hospital-days (Figure 2).
Table II Characteristics of E. cloacae isolated from blood cultures in the neonatal department Patient 1 2 3 4 5 6 7 8 9 10 a
Date of positive blood culture
Other sites positive for E. cloacae (date)
Unit
Clone
19/02/2000 14/05/2001 08/12/2001 08/12/2001 28/05/2002 03/12/2002 03/12/2002 06/12/2002 11/02/2003 26/02/2003
Urine (16/02/00)a Nonea Wounda (12/12/01) Nonea Nonea Nonea Nonea Nonea Rectum (10/02/03) Rectum (24/02/03)
PICU PICU PU PICU PICU PU PU PU PU PU
ND Spb Spb Spb Spb A A A Spc Cc
As screening programme had not been started at the time of the isolation, we do not know whether these patients were carriers. As screening programme had not been started at the time of the isolation, we cannot draw conclusions concerning the sporadic nature of these strains. c Carriage and clinical isolates had similar pulsed-field gel electrophoresis patterns. PICU, paediatric intensive care unit; PU, premature unit. b
Epidemiology of E. cloacae in neonatal units
123
Table III Characteristics of the different clones of Enterobacter cloacae Clones of E. cloacae
Epidemic clones Clone A Clone B Clone C Clone D Clone E Clone F Sporadic clones
Number of positive patientsa
Screening/ clinical cultureb
23 11 4 4 3 2 24
18/6 11/1 4/1 4/1 3/0 2/0 17/8
Unit of isolation: PU/PICU
22/1 9/2 4/0 2/2 0/3 2/0 13/11
Period of spreadc
12/12/02–07/02/03g 13/01/03–31/03/03 17/02/03–10/03/03 30/12/02–20/01/03 25/12/03–13/01/03 21/12/02–03/01/03 –
Antibiotic resistance patternd
Wild-typee
High-level b-lactamase-producerf
17 4 0 0 1 2 15
6 7 4 4 2 0 9
Considering only the first E. cloacae isolate perform each patient, except if the patient was positive for several clones. a Six patients were colonized and/or infected with two different pulsotypes and one patient was colonized and/or infected with three different pulsotypes. b Some patients had both clinical and screening cultures positive for the same clone, some patients had only clinical culture because they were discharged before the epidemiological study began. c Established from the microbiological isolation dates, without considering the date of stay of the positive patients. d When several isolates belonging to the same clone were consecutively obtained from one patient, we report the first antibiotype. e Wild-type: resistant to amoxicillin, amoxicillin/clavulanic acid, ticarcillin and cefalothin. f Resistant or intermediate to ceftazidime or cefotaxime and extended-spectrum b-lactamase-negative. g This clone probably disseminated before the 12/12/02, date corresponding to the start of the survey. PICU, paediatric intensive care unit; PU, premature unit.
Use of alcohol-based hand rub The mean consumption of alcohol-based hand rub was 64.6, 16.0 and 9.3 L/1000 hospitalization-days in the PICU, the PU and the other units of the hospital, respectively (Figure 3).
Infection control measures
reinforced. Furthermore, carriers and colonized/ infected patients were kept separate from other patients and admissions were restricted.
Discussion
The infection control measures used by the units were reviewed, and medical and nursing staff were urged to comply with these measures. In particular, optimal hand disinfection practices were
This investigation was prompted by the concomitant identification of three E. cloacae-positive blood cultures in a neonatal unit. These preliminary results suggest that a ‘classical’ clonal outbreak occurred in one unit of our neonatal department. This epidemic strain spread to the PU as a
Figure 2 Antibiotic use in the premature unit from January 1999 to December 2002. (—) Penicillins; (---) third-generation cephalosporins; ( – – – ) others.
Figure 3 Alcohol-based hand rub solution use in Besan¸ con hospital in 2002. Each point represents one hospital unit; (…) mean; (—) median.
124
consequence of patient-to-patient transmission. A central concept in preventing cross-transmission is good compliance with hand hygiene procedures by healthcare workers after handling each patient.22 The consumption of alcohol-based hand rub was fourfold higher in the PICU than in the PU. This difference might explain the higher cross-transmission rate in the PU. It is also likely, that, as in other studies,3,4 understaffing and overcrowding played a role in disseminating the epidemic strain. Indeed, the nurse-to-patient ratio was 1:2 in the PICU and to 1:4 in the PU, although this difference does not reflect the difference in workload between the two units. We were unable to identify an environmental source of contamination. The only sites that tested positive for the epidemic strain were nasogastric tubes and these represent a simple consequence of intestinal carriage rather than a source of transmission of the strain. Moreover, as in other reports, the improvement of hygiene practices combined with the restriction of admissions led to the progressive disappearance of the epidemic strain.3 – 5 The surveillance programme (screening and genotyping), which was continued after the major epidemic episode, showed that the emergence of E. cloacae was not limited to the sequential spread of epidemic strains. Our study improved our knowledge of outbreaks of E. cloacae in NICUs by determining the genetic relatedness of clinical and screening isolates over a long period (four months). Clinical cultures do not present the full epidemiological picture and surveillance cultures are essential for an understanding of the mechanisms involved in the establishment and maintenance of E. cloacae colonization. Genotyping, particularly PFGE, is a very useful tool for identification, tracing and detailed analysis of the general epidemiology of a bacterial species within a unit. Our study shows that in our neonatal department, E. cloacae caused a succession of colonization outbreaks affecting from two to over 20 patients5 involving transmission by healthcare workers who failed to wash their hands adequately, in a background of sporadic cases. The epidemic potential of a given strain probably depends on the characteristics of the patient,3 the compliance with infection control measures by medical and nursing staff,3,4,6 the antibiotic use policy23 and the intrinsic properties of the strain such as antibiotic resistance,14 – 17 adhesion, fitness and survival in the environment.4 The increased use of third-generation cephalosporins in the PU (Figure 2) may have favoured the emergence of E. cloacae. However, its use remained low compared with that in the PICU (20 versus 70 DDD/1000 HD). Moreover, most of the
D. Talon et al.
isolates were susceptible to this class of antibiotics. So, antibiotic selective pressure can only partially explain the emergence of E. cloacae. There appears to be a correlation between previous use of thirdgeneration cephalosporins and the increase in the rates of colonization and infection caused by E. cloacae,14 – 17 but the strains used in these studies were HLBL. The use of narrow-spectrum antibiotics reduces bacterial resistance,23 but our clinicians found that the data concerning the role of thirdgeneration cephalosporins were not conclusive enough to limit the use of these important agents. More preterm infants in need of aggressive invasive therapy are being admitted to neonatal intensive care units than ever before. These patients are admitted to NICUs immediately after birth, often after delivery by Caesarean section. They are at great risk of becoming colonized by the strains present in their environment. The babies’ bacterial flora is acquired through contacts with the environment, which is constituted by the mother during the delivery and by healthcare workers and other relations during the hospitalization period. Antibiotics exert a selective pressure favouring blactamase-producing bacteria such as E. cloacae. This pressure is exerted within the NICUs and even during the pregnancy if antibiotics are prescribed (for example for urinary tract infection or preterm premature rupture of membranes). Contamination may occur during delivery. The frequency of detection of E. cloacae could therefore reflect the mothers’ intestinal flora. Women with high-risk pregnancies have a higher probability of requiring antibiotic treatment, with subsequent modification of the Enterobacteriaceae residing in the intestinal tract and on various mucous membranes. This would explain the polyclonal nature of E. cloacae isolates outside of clonal outbreaks. Our investigations concerning the frequency of E. cloacae intestinal carriage in patients in neonatal units do not allow us to conclude that the incidence of intestinal colonization increased concomitantly with that of infections. We cannot rule out the possibility that changes in the average status of the patients (prematurity, very low birthweight) increased the infectious risk among patients whose intestines were colonized, without there being a true increase in the frequency of such patients.
References 1. Sanders Jr WE, Sanders CC. Enterobacter spp: pathogens poised to flourish at the turn of the century. Clin Microbiol Rev 1997;10:220—241.
Epidemiology of E. cloacae in neonatal units
2. Gaston MA. Enterobacter: an emerging nosocomial pathogen. J Hosp Infect 1988;11:197—208. 3. Harbath S, Sudre P, Dharan S, Cadenas M, Pittet D. Outbreak of Enterobacter cloacae related to understaffing, overcrowding and poor hygienes practices. Infect Control Hosp Epidemiol 1999;20:598—603. 4. van Dijk Y, Bik EM, Hochstenbach-Vernooij S, et al. Management of an outbreak of Enterobacter cloacae in a neonatal unit using simple preventive measures. J Hosp Infect 2002; 51:21—26. 5. Fernandez-Baca V, Ballesteros F, Hervas JA, et al. Molecular epidemiological typing of Enterobacter cloacae from a neonatal intensive care unit: three-year prospective study. J Hosp Infect 2001;49:173—182. 6. Verweij PE, Van Belkum A, Melchers WJ, Voss A, HoogkampKorstanje JA, Meis JF. Interrepeat fingerprinting of thirdgeneration cephalosporin-resistant Enterobacter cloacae isolated during an outbreak in a neonatal intensive care unit. Infect Control Hosp Epidemiol 1995;16:25—29. 7. Liu SC, Leu HS, Yen MY, Lee PI, Chou MC. Study of an outbreak of Enterobacter cloacae sepsis in a neonatal intensive care unit: the application of epidemiologic chromosome profiling by pulsed-field gel electrophoresis. Am J Infect Control 2002;30:381—385. 8. van Nierop WH, Duse AG, Stewart RG, Bilgeri YR, Koornhof HJ. Molecular epidemiology of an outbreak of Enterobacter cloacae in the neonatal intensive care unit of a provincial hospital in Gauteng, South Africa. J Clin Microbiol 1998;36: 3085—3087. 9. Tresoldi AT, Padoveze MC, Trabasso P, et al. Enterobacter cloacae sepsis outbreak in a newborn unit caused by contaminated total parenteral nutrition solution. Am J Infect Control 2000;28:258—261. 10. van den Berg RWA, Claahsen HL, Niessen M, Muytjens HL, Liem K, Voss A. Enterobacter cloacae outbreak in the NICU related to disinfected thermometers. J Hosp Infect 1999;45: 29—34. 11. Acolet D, Ahmet Z, Houang E, Hurley R, Kaufmann ME. Enterobacter cloacae in a neonatal intensive care unit: account of an outbreak and its relationship to use of thirdgeneration cephalosporins. J Hosp Infect 1994;28:273—286. 12. Goldmann DA, Leclair J, Macone A. Bacterial colonization of
125
13.
14.
15.
16.
17.
18.
19.
20.
21.
22. 23.
neonates admitted to an intensive care environment. J Pediatr 1978;93:288—293. Burman LG, Berglund B, Huovinen P, Tullus K. Effect of ampicillin versus cefuroxime on the emergence of b-lactam resistance in faecal Enterobacter cloacae isolates from neonates. J Antimicrob Chemother 1993;31:111—116. Chow JW, Fine MJ, Shlaes DM, et al. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med 1991;115: 585—590. Finnstrom O, Isaksson B, Haeggman S, Burman LG. Control of an outbreak of a highly b-lactam-resistant Enterobacter cloacae strain in a neonatal special care unit. Acta Paediatr 1998;87:1070—1074. Calil R, Marba ST, von Nowakonski A, Tresoldi AT. Reduction in colonization and nosocomial infection by multiresistant bacteria in a neonatal unit after institution of educational measures and restriction in the use of cephalosporins. Am J Infect Control 2001;29:133—138. Kaplan SL, Patrick CC. Cefotaxime and aminoglycoside treatment of meningitidis caused by Gram-negative enteric organisms. Pediatr Infect Dis J 1990;9:810—814. Members of the SFM Antibiogram Committee. Comite ´ de l’antibiogramme de la Socie ¸aise de Microbiologie ´te ´ Franc Report 2003. Int J Antimicrob Agents 2003;21:364—391. Talon D, Dupont MJ, Lesne J, Thouverez M, Michel-Briand Y. Pulsed-field gel electrophoresis as an epidemiological tool for clonal identification of Aeromonas hydrophila. J Appl Bacteriol 1996;80:277—282. Tenover FC, Arbeit RD, Goering RV, et al. Interpreting chromosomal DNA restriction patterns produced by pulsedfield gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995;33:2233—2239. Natsch S, Hekster YA, de Jong R, Heerdink ER, Herings RM, van der Meer JW. Application of the ATC/DDD methodology to monitor antibiotic drug use. Eur J Clin Microbiol Infect Dis 1998;17:20—24. Jarvis WR. Handwashing—the Semmelweis lesson forgotten? Lancet 1994;344:1311—1312. de Man P, Verhoeven BAN, Verbrugh HA, Vos MC, van den Anker JN. An antibiotic policy to prevent emergence of resistant bacilli. Lancet 2000;355:973—978.