| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1Centro de Pesquisa Experimental, Instituto de Ensino e Pesquisa Albert Einstein - Av. Albert Einstein 627, 05651-901, São Paulo, SP, Brazil 2Infection Control Service, Hospital da Beneficência Portuguesa de São Paulo, São Paulo, Brazil 3Infection Control Service, Hospital Israelita Albert Einstein, São Paulo, Brazil 4Laboratorium voor Microbiologie, Faculteit Wetenschappen, Universiteit Gent, Gent, Belgium
Correspondence Vanda D. Magalhães vanda{at}einstein.br
Received April 15, 2004
Accepted July 2, 2004
An outbreak of bacteraemia in a haemodialysis unit where 65 episodes of infection involved 35 outpatients is reported. Burkholderia cepacia complex was the agent most frequently recovered from blood. Thirty-three environmental and clinical isolates of B. cepacia complex were characterized by whole-cell protein electrophoresis and recA-RFLP profile. Fourteen isolates were genomovar I and 16 isolates were not classifiable by their recA-RFLP pattern. Ribotyping, random amplification of polymorphic DNA (RAPD) and integron profile were used to explore the clonality of the isolates, and revealed multiple strain genotypes. Four ribotypes and RAPD types and three integron patterns were identified. The water supply was identified as the source of the outbreak, and inappropriate cleaning and a leak in the reverse osmosis tubing connection were the probable causes of contamination. B. cepacia complex was still recovered from blood of patients even after apparently adequate measures were taken and water quality standards were met, suggesting that higher standards for water quality should be adopted in haemodialysis units. The genomovars recovered here were distinct from those commonly reported for cystic fibrosis isolates.
 |
INTRODUCTION |
|---|
 TOP INTRODUCTION METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Inadequate catheter care, contamination of water supply, defects in membrane integrity and reprocessed dialysers have all been implicated in outbreaks in haemodialysis units. Haemodialysis fluids and equipment support the growth of hydrophilic bacteria and contamination of them is directly related to the quality of the water supply. Although the water supply for haemodialysis is not required to be sterile, the number of bacteria present must fall within specific standards to reduce the risk of bloodstream infection (Roth & Jarvis, 2000). Therefore, monitoring of water quality in a haemodialysis setting is mandatory.
Members of the Burkholderia cepacia complex are taxonomically complex, but only B. cepacia complex from the respiratory tract of cystic fibrosis patients has been extensively studied (Coenye et al., 2001; Vermis et al., 2002). Species in this group cannot be reliably identified by phenotypic characteristics and have been classified into at least nine genomic species or genomovars (Vandamme et al., 2002). Here we describe extensive molecular characterization of uncommon B. cepacia isolates recovered from blood and environmental samples in a haemodialysis setting. The data unequivocally implicated the water supply as a source of infection.
|
METHODS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
Epidemiological investigation was carried out as a cohort study of chronic renal failure patients being dialysed three times a week. A case was defined as any long term haemodialysis patient with an episode of bloodstream infection documented with a positive blood culture.
Bacterial isolation, culture and identification.
Isolates were recovered from blood of patients undergoing haemodialysis treatment who presented with symptoms of bacteraemia (chills and fever). Ten millilitres of blood from the venous return, collected after wiping the venous port with povidine/alcohol, were inoculated into an enriched medium (Bactec NR 660). Water culture was done by standard methods using a filtration technique (Millipore) and spread on nutrient agar (3 % beef extract, 5 % peptone) for quantitative culture at room temperature for 7 days. Phenotypic identification and antibiotic susceptibility tests were carried out in the Microscan system (Baxter). Identification of B. cepacia complex bacteria was confirmed by whole-cell protein electrophoresis (Pot et al., 1994) and recA restriction fragment length polymorphism (RFLP) analysis (Mahenthiralingam et al., 2000; Vandamme et al., 2002).
Genomic DNA extraction.
Genomic DNA was extracted using a modified method described by Mekalanos (1983). Briefly, an overnight culture was washed with 50 mM Tris/HCl (pH 8.0), 50 mM EDTA, and resuspended in 500 µl of the same buffer. The cell suspension was treated with 20 µl lysozyme (10 mg ml–1; Sigma) for 30 min at 37 °C, followed by incubation with 25 µl proteinase K (20 mg ml–1; Invitrogen) and 20 µl SDS (10 %) for another 30 min at 55 °C. DNA was purified by phenol/chloroform extraction, precipitated with 3 M NaCl (0.1 volume) and absolute ethanol (2.5 volume) and was resuspended in 100 µl deionized water.
Ribotyping.
Ribotyping was carried out as previously described by Gruner et al. (1993), using the endonuclease PvuII for genomic DNA digestion. Restriction fragments were separated by 0.8 % agarose gel electrophoresis in 0.5x TBE buffer (89 mM Tris, 89 mM boric acid, 2 mM EDTA, pH 8.3) for 18 h at 1 V/cm. Southern blots were performed as described by Sambrook & Russell (2001) and the plasmid pKK3535 (Brosius et al., 1981), carrying the Escherichia coli rrn operon, was used as a probe. The plasmid was linearized with BglII and labelled by the digoxigenin system (DIG DNA labelling and detection kit; Roche Molecular Biochemicals). DNA fingerprinting patterns were visually analysed and each ribotype pattern was designated R followed by a number.
RAPD.
Fifty nanograms of total B. cepacia DNA was used in a 50 µl PCR with primer 272 (5'-AGCGGGCCAA) as previously described by Mahenthiralingam et al. (1996). Amplification products were photographed under UV transillumination after electrophoresis on agarose gel (2 % in 1x TAE, 40 mM Tris-acetate, 1 mM EDTA buffer) and staining with ethidium bromide.
Integron detection.
The presence of integrons was investigated by using primers directed to the conservative regions flanking those elements (3'CS and 5'CS) as described by Lévesque & Roy (1993). Three millilitres of overnight culture were washed and resuspended in 1 ml distilled water from which 5 µl was immediately used in PCR. Fifty microlitres of the reaction mixture contained 200 µM dNTP, 50 pmol each primer, 1x PCR buffer with additional MgCl2 to the final concentration of 3 mM, 2.5 U Taq DNA polymerase. Reaction mixture constituents were supplied by Amersham Biosciences. Gene Amp 9600 PCR System thermal cycler (Perkin Elmer Applied Biosystems) was programmed, after a modification of the protocol proposed by Lévesque & Roy (1993), for 10 min at 95 °C, followed by 12 cycles of 1 min at 94 °C, 1 min at 55 °C and 5 min at 72 °C, 12 cycles of 1 min at 94 °C, 1 min at 55 °C and 6 min at 72 °C and 12 cycles of 1 min at 94 °C, 1 min at 55 °C and 7 min at 72 °C. After electrophoresis at 5 V/cm for 2 h on 1.2 % agarose gels, PCR products were stained with ethidium bromide and photographed under UV transillumination.
|
RESULTS |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
|
A retrospective review of patients undergoing dialysis in the unit showed that only three episodes of bacteraemia had occurred during the preceding year (Fig. 1).
|
The haemodialysis procedure, including equipment disinfection, and the whole water system were reviewed. No backflow was detected in the machines and nurses were following good practices for dialysis set-up. Water quality monitoring exceeded limits of the American Association of Medical Instrumentation (AAMI) standards for heterotrophic bacteria on four occasions. In July 2001, one point of reservoir 2 yielded 700 c.f.u. ml–1 and a sample point before dialysis machine 1 produced 330 c.f.u. ml–1. In August 2001, reservoir 2 yielded 720 c.f.u. ml–1 and in February 2002, a sample before another dialysis machine showed 1650 c.f.u. ml–1. All other water samples showed less than 200 c.f.u. ml–1 (upper limit by AAMI) during the outbreak. Faecal coliforms were not detected.
B. cepacia was isolated from the water sample in July 2001. In August 2001, reservoir 2 and all pipes and tubing were replaced. Nevertheless, B. cepacia continued to be isolated from blood and from water on six further occasions, although water quality monitoring showed less than 200 c.f.u. ml–1 of heterotrophic bacteria (Table 2). The probable source of contamination was finally identified in December 2002 when a loose connection in the reverse osmosis tubing was detected, leading to a leak that could allow the entrance of bacteria into the water system. Repair of this particular part of the system stopped the outbreak.
Identification of B. cepacia complex isolates
B. cepacia complex formed a homogeneous group. All B. cepacia isolates gave the same antibiogram pattern, being multi-resistant but fully susceptible to sulfa/trimethoprin. Thirty-three B. cepacia isolates, from 17 clinical and 16 environmental samples, were available for further characterization. A B. cepacia complex isolate (IEP198) recovered from the bloodstream of a patient admitted to another ward during the outbreak period was included as a control in the study.
Isolates were confirmed as B. cepacia complex by whole-cell protein electrophoresis (data not shown) and genomovar status was identified by HaeIII-recA RFLP analysis (Mahenthiralingam et al., 2000). A recA amplicon of the expected size was generated for each isolate and two distinct RFLP patterns were obtained (Table 2). Fourteen isolates, three from patients and 11 from water, were identified as type K which corresponds to genomovar I (Vermis et al., 2002) and 16 isolates were designated AT (unknown genomovar), 13 from patients and three from water. The control strain (IEP 198) was designated Se27, a unique profile not seen before.
Molecular typing of B. cepacia complex isolates by ribotyping and RAPD
The B. cepacia complex isolates were characterized by ribotyping and RAPD. Three ribotypes were identified from epidemiologically linked patients and from water. Thirteen (76 %) isolates from blood were ribotype 1 and this strain was also found once in water. Ribotype 2, found in blood, dialysate and arterial wound, was the most frequent genotype in water samples from reservoirs (13/16, 81 %). Ribotype 3 was recovered from the bloodstream of a patient and, on the same day, was also found in water (Table 2 and Fig. 2).
|
RAPD fingerprinting resulted in five to 10 bands over a size range of 0.3–5 kb. Each assay was repeated twice to assure its reproducibility. Although some minor variations were noted between experiments, by computing only the main amplified bands, four different patterns were consistently obtained (Table 2 and Fig. 3). All isolates, apart from two, identified as R1 presented the A pattern for RAPD. IEP188 and IEP190 showed a one band difference when compared to strains classified as A, which suggests a close relationship between them, and they were classified as pattern A1. All R2 isolates were identified as pattern B and all R3 showed RAPD pattern C. The control strain IEP198 yielded unique patterns both for RAPD (D) and ribotyping (R4).
|
Integron profiles of B. cepacia complex isolates
Among the epidemiologically linked isolates, three different profiles of integrons were obtained. Three bands at positions 0.8, 0.85 and 1.4 kb were detected in R1 isolates (data not shown). Eleven R2 isolates from clinical and water samples showed a single band of 0.8 kb (I2). Five B. cepacia complex from water, all isolated on the same day and ribotyped as R2, had no detectable integron structure (Table 1). Isolates identified as R3 contained a 0.7 kb integron. The control strain was distinct.
|
DISCUSSION |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
The present report describes the investigation of bloodstream infection of patients undergoing long-term dialysis. B. cepacia complex was the most frequently isolated micro-organism and recA-RFLP pointed to a polyclonal nature of the outbreak, which was confirmed by other molecular typing data. The results are consistent with bacterial diversity found in environmental samples (Salles et al., 2002).
Data on the distribution of genomovars of B. cepacia complex have been mostly restricted to isolates from cystic fibrosis patients (Crowley et al., 2002; Mahenthiralingam et al., 2002), where genomovar III is consistently the most common, followed by genomovars II (37.8 %) and I (26 %) (Vermis et al., 2002). The identification of genomovar I isolates in this study represents, to our knowledge, the first time it has been confirmed in a haemodialysis setting. Another unassigned recA-RFLP profile (AT) was identified in the majority of isolates and further taxonomic analysis, including DNA–DNA hybridization, is required. Our data suggest that types of B. cepacia complex involved in bacteraemia are different from those isolated from the respiratory tract of cystic fibrosis patients. Nevertheless, widespread genomovar III and genomovar V (B. vietnamiensis) has been previously identified in another haemodialysis setting (Magalhães et al., 2003).
Ribotyping and RAPD have been successfully applied to molecular epidemiology studies of B. cepacia (Bingen et al., 1993; Brisse et al., 2000; Okazaki et al., 1999). RAPD correctly grouped epidemiologically related B. cepacia strains from a large collection according to previous data obtained by PFGE and ribotyping (Mahenthiralingam et al., 1996; Bingen et al., 1993; Brisse et al., 2000). Investigating a nosocomial outbreak, Okazaki et al. (1999) concluded that RAPD offered an alternative discriminatory method to PFGE, providing results within a shorter time and with less complexity than the latter.
B. cepacia isolates that showed identical ribotypes in this study were considered a single clone. RAPD patterns agreed with ribotyping results, except for two isolates IEP188 and IEP190, which showed a single band difference in the RAPD profile. In the light of ribotyping and integron data, we considered both isolates as part of the R1 clone.
A previous work suggested that ribotyping could provide reliable identification of B. cepacia complex genomovars (Brisse et al., 2000). Here, all R2 strains were identified as genomovar I. R1 and R3 strains showed novel recA-RFLP patterns and further study may reveal heterogeneity among them, as observed for other B. cepacia isolates of novel RFLP type (unpublished observations).
Integrons are plastic genetic elements that acquire and lose gene cassettes related mostly to antibiotic resistance. Horizontal transfer of integrons is mediated by plasmids and transposons (Stokes & Hall, 1989). Different bacterial genera can exhibit the same integron content and, conversely, a genetically unique strain can show a different integron profile (Severino & Magalhães, 2002). Recently, they have been used as strain markers for investigation of outbreaks (Severino & Magalhães, 2003; Martin et al., 2001). The rationale behind the use of this new molecular epidemiological tool is that isolates of a clonal origin, in a limited period of time, would not have enough time to modify their integron content. The integron profile of the B. cepacia isolates showed good correlation with the other typing methods and only isolates recovered from a reservoir on 14 May 2002, which were classified as R2, did not show inserted cassettes in their integron structure. They could represent part of a B. cepacia R2 population that lost the resistance cassettes in the absence of selective pressure.
The epidemiological investigation and typing data linked the bacteraemia episodes to water contamination. The probable source of contamination was related to inadequate cleaning procedures that left leaking connections of the reverse osmosis tubing. Environmental biofilm-forming bacteria or micro-organisms present in cleaning solutions could have entered the water system through this opening. Tubing connections are known to be critical segments of the system and biofilm formation is recognized as a risk for haemodialysis patients (Man et al., 1998; Dasgupta, 2002).
The counts of bacteria in water are influenced by the microbiological methods used. In this study the measurement of water contamination followed optimized AAMI recommendations for bacterial counts. During the outbreak period, in four instances more than 200 c.f.u. ml–1 were detected, but only once was B. cepacia identified. Unfortunately, this isolate was not available for molecular typing. Paradoxically, other B. cepacia isolated from the reservoir occurred in water samples that met the AAMI standards. In all instances when B. cepacia isolates recovered from water were available for typing, they were linked to bloodstream infection episodes. In another outbreak of bacteraemia of a polyclonal nature most water samples were also within the limits of AAMI guidelines (Magalhães et al., 2003).
The microbiological standard for dialysate and water purity in haemodialysis has been a matter of discussion (Lonnemann, 2000; Tielemans et al., 2001). While AAMI recommends an upper limit of 200 c.f.u. ml–1 in haemodialysis water, the Deutsche Arbeitgemeinschaft für Klinische Nephrologie (DafKN) from the German Renal Society recommends a maximum of 100 c.f.u. ml–1 (Lonnemann, 2000). In the light of the outbreak described here and considering the increased risk of chronic renal patients, we suggest that higher water quality standards should be adopted in haemodialysis units, as proposed by DafKN.
|
REFERENCES |
|---|
|
TOPINTRODUCTIONMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  W. Assaad, M. Magalhaes, M. Plesa, C. A. Hart, P. Cornelis, and C. Winstanley Identical Burkholderia cepacia complex strain types isolated from multiple patients attending a hospital in Brazil J. Med. Microbiol., February 1, 2006; 55(2): 247 - 249. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| INT J SYST EVOL MICROBIOL | J MED MICROBIOL | MICROBIOLOGY | J GEN VIROL | ALL SGM JOURNALS |
HindIII molecular size standards. Lanes 8, 9 and 10, ribotype 1; lanes 5, 6 and 7, ribotype 2; lanes 1 and 2, ribotype 3; lane 3, control strain, ribotype 4.