Received 24 July 2009; returned 28 August 2009; revised 9 October 2009; accepted 17 October 2009

Transcription

1 J Antimicrob Chemother 2010; 65: doi: /jac/dkp411 Advance publication 3 December 2009 Bacteraemia due to extended-spectrum b-lactamase-producing Escherichia coli (ESBL-EC) in cancer patients: clinical features, risk factors, molecular epidemiology and outcome Carlota Gudiol 1,2 *, Laura Calatayud 3, Carolina Garcia-Vidal 1,2, Jaime Lora-Tamayo 1,2, María Cisnal 3, Rafael Duarte 4, Montserrat Arnan 4, Mar Marin 5, Jordi Carratalà 1,2 and Francesc Gudiol 1,2 1 Infectious Disease Department, Hospital Universitari de Bellvitge, l Hospitalet de Llobregat, Barcelona, Spain; 2 REIPI (Spanish Network for Research in Infectious Disease), Instituto de Salud Carlos III, Madrid, Spain; 3 Microbiology Department, Hospital Universitari de Bellvitge, l Hospitalet de Llobregat, Barcelona, Spain; 4 Haematology Department, Hospital Duran i Reynals; Institut d Investigació Biomèdica de Bellvitge (IDIBELL), University of Barcelona, L Hospitalet de Llobregat, Barcelona, Spain; 5 Oncology Department, Hospital Duran i Reynals; Institut d Investigació Biomèdica de Bellvitge (IDIBELL), University of Barcelona, L Hospitalet de Llobregat, Barcelona, Spain *Corresponding author. Infectious Disease Department, Hospital Universitari de Bellvitge, Feixa Llarga s/n, 08907L Hospitalet de Llobregat, Barcelona, Spain. Tel: þ ; Fax: þ ; cgudiol@iconcologia.net Received 24 July 2009; returned 28 August 2009; revised 9 October 2009; accepted 17 October 2009 Objectives: To assess the clinical features, risk factors, molecular epidemiology and outcome of extended-spectrum b-lactamase-producing Escherichia coli (ESBL-EC) bacteraemia in hospitalized cancer patients. Methods: Episodes of ESBL-EC bacteraemia were compared with a susceptible control group in a 3 year prospective study. ESBL-EC strains were studied by PCR and isoelectric focusing, and molecular typing was performed by PFGE. Results: Out of 531 episodes of bacteraemia, 135 were caused by E. coli. Seventeen of these cases involved ESBL-EC-producing strains (12.6%). In the multivariate analysis, female gender [odds ratio (OR) 3.43; 95% confidence interval (CI) ] and previous antibiotic therapy (OR 3.22; 95% CI ) were found to be independent risk factors for ESBL acquisition. An analysis of ESBL-EC isolates revealed a polyclonal distribution with CTX-M predominance (59%). Patients with ESBL-EC bacteraemia were more likely to have received an inadequate empirical antibiotic therapy (65% versus 6%; P¼0.000), and the time to adequate therapy was longer in this group (0 versus 1.50 days; P¼0.000). The overall mortality rate was 22%, ranging from 20% to 35% (P¼0.20). Risk factors for mortality were solid tumour (OR 19.41; 95% CI ), corticosteroid therapy (OR % CI ) and intensive care unit admission (OR , 95% CI ). In neutropenic patients, ESBL-EC bacteraemia was associated with poorer outcome and a higher overall mortality rate (37.5% versus 6.5%; P¼0.01). Conclusions: In our centre, ESBL-EC bacteraemia is frequent among cancer patients, especially in those exposed to antibiotic pressure. All ESBL-EC strains were unrelated and most of them carried a CTX-M group enzyme. Patients with ESBL-EC bacteraemia received inadequate empirical antibiotic therapy more frequently than patients carrying a susceptible strain, but significant differences in mortality could not be demonstrated. Keywords: multidrug resistant, bloodstream infection, neutropenia Introduction Escherichia coli is one of the most common causes of Gramnegative bloodstream infection complicating immunosuppressed patients with cancer. Extended-spectrum b-lactamase-producing E. coli (ESBL-EC), particularly those producing CTX-M types of ESBL, are emerging pathogens worldwide. Recently, it has been shown that ESBL-EC faecal carriage is frequent in the community, and we have found that it is also common in neutropenic cancer patients in our geographic area. 1 3 Treatment options for ESBL-EC infections are limited, since these microorganisms are resistant to penicillins, cephalosporins (with the exception of cephamycins) and aztreonam, and frequently present associated resistance to other antimicrobials such as trimethoprim/ # The Author Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please journals.permissions@oxfordjournals.org 333

2 Gudiol et al. sulfamethoxazole, aminoglycosides and quinolones. 4 This situation can be aggravated in cancer patients, a high-risk population, who are often empirically treated for febrile neutropenia with a broad-spectrum penicillin or cephalosporin. The objective of our study was to assess the clinical features, risk factors and outcome of ESBL-EC bacteraemia in immunosuppressed patients with cancer, and to study the molecular epidemiology of ESBLcarrying isolates in the emerging CTX-M era. Methods Setting, patients and study design A prospective observational study was conducted in a 200 bed university cancer centre for adults in Barcelona, Spain, which has 6500 admissions per year. Participants included all hospitalized cancer patients and haematopoietic stem cell transplant recipients with at least one episode of bacteraemia from January 2006 to October Information on baseline characteristics, clinical features, empirical antibiotic therapy and outcome was carefully recorded in a specific database. All the episodes of bacteraemia that were caused by an ESBL-EC strain were compared with those caused by a non-esbl-ec strain. We also compared the characteristics of patients who died with those of patients who survived, to determine the factors influencing mortality. All episodes of bacteraemia in our hospital are reported and followed up by an infectious disease physician. Changes in antimicrobial treatment and general management were advised when necessary. During the study period, no antibacterial prophylaxis to prevent bacterial infections was administered to patients with neutropenia. This study was approved by the ethics committee of our institution. Definitions Neutropenia was defined as an absolute neutrophil count,500 cells/ mm 3. Prior antibiotic therapy was defined as the receipt of any systemic antibiotic for.48 h in the previous month. An endogenous source of bacteraemia was considered in patients who were neutropenic and who did not have any other evident source of infection. Empirical antibiotic therapy was considered inadequate if the treatment regimen did not include at least one antibiotic that is active in vitro against the infecting microorganism. In patients with ESBL-EC bacteraemia, treatment with an oxyimino-b-lactam (combined or not with an aminoglycoside) was considered inadequate, regardless of the MIC. Time to adequate antibiotic therapy was recorded in days, considering the closest day to the hour that the antibiotic was administered. Shock was defined as a systolic pressure,90 mmhg that was unresponsive to fluid treatment and/or required vasoactive drug therapy. 5 Overall mortality was defined as death by any cause within 30 days of the onset of bacteraemia. Early mortality was defined as death within 7 days of the onset of bacteraemia. Microbiological studies Two sets of two blood samples, separated by 30 min, and containing 15 ml of blood each, were drawn in patients who presented fever.388c or when bacteraemia was suspected for any clinical sign or symptom. Blood samples were processed by the BACTEC 9240 system (Becton-Dickinson Microbiology Systems) with an incubation period of 5 days. The identification of E. coli strains and susceptibility testing were performed using commercial panels from the MicroScan system (Dade Behring Inc., West Sacramento, CA, USA). CLSI criteria were used to define susceptibility or resistance to the antimicrobial agents studied. ESBL production was screened and confirmed in accordance with CLSI standards. 6 Molecular characterization of ESBL-EC was performed by macrorestriction analysis of genomic DNA restricted with XbaI (New England BioLabs). DNA fragments were separated by PFGE in a CHEF-DR III system (Bio-Rad). Restriction patterns were analysed by following previously established criteria. 7 ESBLs were characterized by a multiplex PCR and isoelectric focusing. The presence of bla TEM, bla SHV, bla CTX-M and bla OXA in each organism was studied by PCR, as described previously. 8 Isoelectric focusing analysis of crude extracts was done as previously described. Briefly, E. coli strains were grown for 4 h in Luria broth. Bacterial growth was centrifuged, resuspended in distilled water, and sonicated. Extract purification was performed by ultracentrifugation. Isoelectric focusing was performed with the PhastSystem apparatus (Pharmacia, Uppsala, Sweden) in polyacrylamide gels with a ph range of 3 9 (PhastGel 3-9; Pharmacia). Polyacrylamide gels were stained with nitrocefin (500 mg/ml; Oxoid, Hampsire, UK), and the pis were obtained by comparison with a set of different b-lactamases with known pis. 9 Statistical analysis Continuous variables were compared by the Mann Whitney U-test and the t-test. Qualitative variables were compared by the x 2 test; odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Multivariable conditional logistic regression analysis of factors potentially associated with ESBL-EC acquisition and mortality included all statistically significant variables in univariate analysis, gender and age, and all clinically important variables, whether they were statistically significant or not. 10 The analysis was performed with the stepwise logistic regression model of the SPSS software package (SPSS). Results Out of 531 episodes of bacteraemia documented during the study period, 135 (25.4%) were caused by E. coli. Seventeen (12.6%) of these cases were caused by an ESBL- producing strain. There was one episode of ESBL-EC bacteraemia per patient, whereas 10 patients in the control group experienced more than one episode, separated by at least a 4 week interval (118 episodes in 104 patients), occurring in a total of 121 patients. Table 1 shows the main baseline characteristics and clinical features of the cohort. Previous antibiotic therapy was common among patients (44%). There were no differences between groups in the type of antibiotic received (oxyimino-b-lactams 37.5%, quinolones 26%, glycopeptides 26%, b-lactamþb-lactamase inhibitors 22% and carbapenems 17.5%). When patients with ESBL-EC bacteraemia were compared with those with non-esbl-ec bacteraemia, we found that there was a trend towards haematological malignancy as the underlying condition in the ESBL-EC group (76.5% versus 53%; P¼0.073). More patients in this group were female (71% versus 37%; P¼0.016), had received antibiotic therapy in the previous month (71% versus 40%; P¼0.034) and needed a blood transfusion 5 days prior to the episode of bacteraemia (41% versus 18%; P¼0.026). In the multivariate analysis (Table 2), the only variables found to be independent risk factors for ESBL acquisition were female gender (OR 3.43; 95% CI ) and previous antibiotic therapy (OR 3.22; 95% CI ). Table 3 summarizes the results of b-lactamase characterization and resistance patterns of the 17 ESBL-EC isolates. All but one strain was available for genotype and ESBL characterization. The ESBLs were identified as eight from the CTX-M-9 334

3 ESBL E. coli bacteraemia in cancer patients JAC Table 1. Baseline characteristics and clinical features of all episodes of E. coli bacteraemia in cancer patients ESBL-EC n¼17 (%) Non-ESBL-EC n¼118 (%) Total n¼135 (%) P Gender (female) 12 (71) 44 (37) 56 (41) Age (years, mean+sd) Underlying disease solid tumour 4 (24) 55 (47) 59 (44) haematological malignancy 13 (76) 63 (53) 76 (56) Haematopoietic stem cell transplant 1 (6) 10 (8) 11 (8) 1.0 Graft versus host disease 0 3 (3) 3 (2) 1.0 Co-morbidities a 6 (35) 37 (31) 43 (32) 0.78 diabetes mellitus 4 (24) 15 (13) 19 (14) 0.26 chronic obstructive pulmonary disease 1 (6) 7 (6) 8 (6) 1.0 chronic heart disease 1 (6) 19 (16) 20 (15) 0.46 chronic renal failure 0 2 (2) 2 (1) 1.0 chronic liver disease 0 4 (3) 4 (3) 1.0 Pervious chemotherapy (1 month) 14 (82) 76 (64) 90 (67) 0.17 Neutropenia (,500 neutrophils/ml) 8 (47) 46 (39) 54 (40) 0.52 Corticosteroid therapy (1 month) 2 (12) 47 (40) 49 (36) Previous antibiotic therapy (1 month) 12 (71) 47 (40) 59 (44) Urinary catheter 2 (12) 19 (16) 21 (16) 1.0 Intravascular catheter 12 (71) 61 (52) 73 (54) 0.14 Parenteral nutrition (3 days) 0 2 (2) 2 (1) 1.0 Previous surgery (1 month) 0 7 (6) 7 (5) 0.59 Previous intensive care unit admission (3 months) 0 3 (3) 3 (2) 1.0 Previous hospital admission (3 months) 11 (65) 66 (56) 77 (57) 0.49 Healthcare acquisition 16 (94) 104 (88) 120 (89) Previous blood transfusion (5 days ) 7 (41) 20 (17) 27 (20) Previous platelet transfusion (5 days ) 3 (18) 17 (14) 20 (15) 0.72 Axillary temperature 388C 14 (82) 96 (81) 110 (81) 1.0 Polymicrobial bacteraemia 3 (18) 21 (18) 24 (18) 1.0 Shock at presentation 1 (6) 8 (7) 9 (7) 0.37 Source of bacteraemia endogenous bacteraemia 9 (53) 40 (34) 49 (36) urinary tract 1 (6) 32 (27) 33 (24) cholangitis 5 (29) 21 (18) 26 (19) other gastrointestinal tract b 1 (6) 17 (14) 18 (13) skin and soft tissue infection 0 1 (1) 1 (1) unknown origin 1 (6) 7 (6) 8 (6) Univariate analysis of risk factors for ESBL-EC acquisition was performed. ESBL-EC, extended-spectrum b-lactamase E. coli. a Some patients presented more than one co-morbidity. b Other gastrointestinal tract: colon, 11; intrabdominal abscess, 3; neutropenic enterocolitis, 2; spontaneous bacterial peritonitis, 1; secondary peritonitis 1. Table 2. Risk factors for extended-spectrum b-lactamase E. coli acquisition by multivariate analysis Characteristic Adjusted OR (95% CI) P Gender (female) 3.43 ( ) Age (years, mean+sd) 1.02 ( ) 0.28 Haematological underlying disease 1.46 ( ) 0.59 Previous antibiotic therapy 3.22 ( ) Previous blood transfusion 3.06 ( ) 0.97 ESBL-EC, extended-spectrum b-lactamase E. coli. group, four from the CTX-M-1 group, three from the SHV family and one strain from the TEM family. As determined using PFGE, all isolates were genetically unrelated (Figure 1). Resistance rates to non-b-lactam antibiotics were as follows: quinolones 65%, trimethoprim/sulfamethoxazole 59%, gentamicin 12%, tobramycin 12%. All strains remained fully susceptible to carbapenems and amikacin. Fifty-three percent of the strains were resistant to amoxicillin/clavulanic and 18% to piperacillin/tazobactam. All patients in the study received empirical antibiotic therapy (Table 4). Most of them received a combination of a b-lactam (mostly an oxyimino-b-lactam) þaminoglycoside (40%), 335

4 Gudiol et al. Table 3. Microbiological characteristics of extended-spectrum b-lactamase-producing E. coli strains PCR result bla CTX-M pi bla TEM Patient bla CTX-M9 bla CTX-M1 bla CTX-M2 bla SHV bla OXA PFGE pattern pi1 pi2 RP 1 þ E CIP 2 þ þ E CIP, SXT 3 þ E S 4 NA NA NA NA NA NA NA NA NA CIP, SXT 5 þ þ E GEN, TOB, CIP, SXT 6 þ þ E CIP 7 þ þ E TOB, CIP, SXT 8 þ þ E SXT 9 þ þ E CIP, SXT 10 þ þ E S 11 þ þ E GEN, TOB, CIP, SXT 12 þ E CIP 13 þ þ E CIP, SXT 14 þ þ E SXT 15 þ þ E CIP, SXT 16 þ þ E CIP 17 þ þ E CIP pi, isoelectric point(s) of b-lactamase(s); RP, resistance patterns; S, susceptible to all antimicrobial agents except b-lactams; CIP, ciprofloxacin resistant; GEN, gentamicin resistant; TOB, tobramycin resistant; SXT, cotrimoxazole resistant; NA, not available. M Figure 1. PFGE patterns of XbaI-digested genomic DNA of E. coli strains. Lane M, molecular weigh marker. Lanes 1 16 show the PFGE patterns of the 16 available ESBL-EC strains. followed by b-lactamþb-lactamase inhibitor (38.5%) and carbapenem (21%). When comparing patients with ESBL-EC bacteraemia with the susceptible control group, we observed that patients harbouring an ESBL strain more frequently received an inadequate initial empirical antibiotic therapy (65% versus 6%; P¼0.000), and time to adequate antibiotic therapy was longer in this group (1.50 versus 0 days; P¼0.000). There were no statistically differences regarding early (12% versus 8%; P¼0.63) and overall mortality (35% versus 19%; P¼0.20) between the two groups. Among the 17 patients with ESBL-EC bacteraemia, 11 received a b-lactam alone and 6 received a b-lactamþ aminoglycoside. In the first group, three patients receiving a b-lactamþb-lactamase inhibitor died (inadequate in two) and two more died while receiving a carbapenem. In the second group, only one patient receiving cefepimeþan aminoglycoside died. We also attempted to identify risk factors for mortality, and the results are shown in Table 5. The variables influencing mortality identified in the univariate analysis were solid tumour as the underlying disease (34% versus 69%; P, 0.001), simultaneous treatment with corticosteroids (28% versus 55%; P¼0.014) and intensive care unit (ICU) admission (1% versus 10%; P, 0.001). These variables were also found to be independent risk factors for mortality in the multivariate analysis (OR 19.41, 95% CI ; OR 3.04, 95% CI ; and OR , 95% CI , respectively). Patients with bacteraemia that had originated in the urinary tract were more likely to survive than patients with other sources of bacteraemia, although this finding was not statistically significant (27.5% versus 10%; P¼0.056). As patients with neutropenia are a high-risk population, who may present with severe sepsis and a poor outcome, we conducted a subanalysis of the group of 54 neutropenic patients with E. coli bacteraemia. We found that neutropenic patients 336

5 ESBL E. coli bacteraemia in cancer patients JAC Table 4. Antibiotic treatment and outcome of patients with extended-spectrum b-lactamase E. coli bacteraemia compared with the susceptible control group Characteristic ESBL-EC n¼17 (%) Non-ESBL-EC n¼118 (%) Total n¼135 (%) P Empirical antibiotic treatment b-lactamþb-lactamase inhibitor 6 (35) 46 (39) 52 (39) 1.00 carbapenem 5 (29) 23 (19) 28 (21) 0.34 oxyimino b-lactam 6 (35) 49 (42) 55 (41) 0.79 quinolone 0 4 (3) 4 (3) 1.00 aminoglycoside 6 (35) 49 (42) 55 (41) 0.79 b-lactamþaminoglycoside 6 (35) 48 (41) 54 (40) 0.79 other antibiotic combinations a 0 5 (4) 5 (4) 1.00 Inadequate initial empirical antibiotic therapy 11 (65) 7 (6) 18 (13) Time to adequate antibiotic therapy (days, median, range) 1.50 (0 4) 0 (0 10) 0 (0 10) Intensive care unit admission 2 (12) 5 (4) 7 (5) 0.21 Early mortality (7 days) 2 (12) 9 (8) 11 (8) 0.63 Overall mortality (30 days) 6 (35) 23 (19) 29 (21) 0.20 ESBL-EC, extended-spectrum b-lactamase E. coli. a Other antibiotic combinations: quinoloneþb-lactam, 4; quinoloneþaminoglycoside, 1. with ESBL-EC bacteraemia also received an inadequate initial empirical antibiotic therapy more frequently than the susceptible control group (50% versus 0%; P, 0.001), and presented a poorer outcome, with higher ICU admission (25% versus 4%, P¼0.04) and a higher overall mortality rate (37.5% versus 6.5%; P¼0.01). Discussion Over the last few years, a significant increase in the number of bloodstream infections due to ESBL-EC has being reported in several parts of the world Most published reports are retrospective, focused on non-immunosuppressed patients and frequently include E. coli and Klebsiella pneumoniae in the same analysis, even though several recent reports have suggested that these two microorganisms have different epidemiological 11 16,18 20 characteristics. In our study, the incidence of ESBL-EC bacteraemia was 12.6%. It is difficult to know whether our figure is in accordance with that found in other centres, since there is little information in the literature describing the incidence of ESBL-EC bacteraemia in this patient population. 14,17,21 However, it is in good agreement with the prevalence of ESBL-EC faecal colonization found among neutropenic cancer patients (31.8%) in our centre. 1 Patients in our study presented clinical features that are in line with those that one may expect in an immunosuppressed population. Most patients were receiving active chemotherapy at the time of the infection, which mostly occurred in the healthcare setting. The most common site of bacteraemia was an endogenous source, in neutropenic patients (40% of the population), followed by the urinary and biliary tract. The only difference between the group of patients with ESBL-EC bacteraemia and the group with non-esbl-producing strains was the previous use of antibiotics. These findings are in keeping with the fact that ESBL-EC colonization is mainly a problem in the community. Outpatients may be carriers of an ESBL-producing strain, which may be selected after exposure to several invasive procedures and especially after antibiotic pressure. Infection may develop afterwards. 14,15,19,20,22 The results obtained from the microbiological study also support this view, since we found extensive clonal diversity among all ESBL-EC isolates, with a clear predominance (75%) of ESBL type CTX-M. These findings have also been described previously in other studies performed in the same geographic area The high incidence of quinolone resistance among all ESBL-EC isolates (65%) is noteworthy, whereas all isolates remained fully susceptible to carbapenems and amikacin. The high rate of quinolone resistance reinforces the controversy of its role in decontaminating the gastrointestinal tract of neutropenic cancer patients. In our study, patients with ESBL-EC bacteraemia received an inadequate empirical antibiotic therapy more frequently than patients with non-esbl-ec bacteraemia. Despite the fact that inadequate empirical antibiotic therapy has previously been associated with a poor outcome in a different clinical settting, 11,17,23 25 we did not find significant differences in mortality between the two groups, although it was higher in the group of patients harbouring an ESBL-producing strain. Following the institutional guidelines in our centre, cancer patients who develop febrile neutropenia are empirically treated with a broadspectrum cephalosporin (cefepime) plus an aminoglycoside (amikacin). Even though we considered such combination therapy as inadequate in patients with ESBL bacteraemia, one may speculate that the addition of an aminoglycoside might confer a certain degree of protection against severe sepsis, shock or fatal outcome, until an appropriate antibiotic regimen is administered. Moreover, a short delay in appropriate definitive antimicrobial therapy may not be associated with higher mortality if antimicrobial therapy is promptly adjusted according to the susceptibility results. In this regard, early reporting of the 337

6 Gudiol et al. 338 Table 5. Characteristics, treatment and outcome of all 17 episodes of ESBL-EC bacteraemia Patient Age, gender Underlying disease Source of bacteraemia Polymicrobial bacteraemia NTP Shock/ICU Adequate empirical antibiotic therapy Definitive antibiotic Outcome 1 70, female non-hodgkin lymphoma endogenous no no no no, ATM IPM survived 2 68, female CLL cholangitis no no no no, AMC IPM survived 3 70, male disseminated pancreatic tumour cholangitis yes, E. aerogenes no no no, AMC CIP died (24th day) 4 59, male pancreatic tumour cholangitis yes, E. faecium no no no, TZP IPM survived 5 89, male non-hodgkin lymphoma cholangitis no no no no, TZP IPM survived 6 72, female advanced ovarian cancer GI tract no no no no, TZP AMC died (9th day) 7 59, male AML endogenous no yes yes/yes yes, IPM IPM died (16th day) 8 52, female AML endogenous no yes no no, FEPþAMK IPM died (29th day) 9 49, male AML endogenous no yes no no, FEPþAMK IPM survived 10 57, female AML endogenous no yes no no, FEPþAMK IPM survived 11 79, female non-hodgkin lymphoma endogenous no yes no no, FEPþAMK IPM survived 12 51, female AML urinary tract no no no no, FEPþAMK IPM survived 13 26, female AML endogenous no yes no yes, IPMþAMK IPM survived 14 66, female AML endogenous yes, P. mirabilis yes no/yes yes, IPM MEM died (4th day) 15 73, female multiple myeloma unknown no no no yes, IPM MEM survived 16 23, female AML, allogeneic TPH endogenous no yes no yes, IPM IPM survived 17 83, female metastatic breast cancer cholangitis no no no yes, TZP TZP died (5th day) CLL, chronic lymphocitic leukaemia; AML, acute myeloid leukaemia; GI, gastrointestinal; NTP, neutropenia; ICU, intensive care unit; ATM, aztreonam; IPM, imipenem; AMC, amoxicillin clavulanate; TZP, piperacillin tazobactam; FEP, cefepim; AMK, amikacin; CIP, ciprofloxacin; MEM, meropenem. Downloaded from by guest on December 14, 2014

7 ESBL E. coli bacteraemia in cancer patients JAC results from the microbiological laboratory is of paramount importance. The potential protective role of an aminoglycoside combined with a b-lactam in the empirical antibiotic therapy for ESBL-EC bacteraemia needs to be further investigated. Factors influencing mortality in our study were solid tumour as underlying disease, simultaneous treatment with corticosteroids, and ICU admission (Table 6). Mortality mainly occurred in debilitated patients with disseminated solid organ tumours who were receiving palliative chemotherapy, in order to stop disease progression rather than with a curative intention, and also receiving corticosteroid therapy in an attempt to mitigate the symptoms caused by the underlying disease. Patients who needed to be admitted to the ICU for severe sepsis or further complications were also more likely to have a poorer outcome. It has been suggested that urinary infections caused by ESBL-EC have a better outcome than non-urinary infections In our study, the urinary tract was the second most frequent source of bacteraemia, and it also tended to be associated with a better outcome than bacteraemia from any other source. The subset of neutropenic cancer patients with ESBL-EC bacteraemia in our study also received inadequate empirical antibiotic therapy more frequently and presented a higher mortality rate. This finding should be interpreted with caution, due to the small number of cases. A larger group of neutropenic patients need to be studied to identify risk factors for mortality in this particularly high-risk group of patients. Carbapenems are widely regarded as the drug of choice for treatment of serious ESBL-EC infections such as bacteraemia. 26,27 They are also recommended by the current clinical practice guidelines published by the Infectious Diseases Society of America as an option for the management of high-risk febrile neutropenic patients with cancer. 22,28 However, increased use of carbapenems may induce the appearance of resistance, especially in Enterobacteriaceae and Pseudomonas aeruginosa Our findings may help to better delineate the profile of patients who are suitable for receiving a carbapenem agent from the very beginning. However, controlled studies, particularly those involving neutropenic patients, are warranted. This study has some limitations. As it was carried out in a single centre the results may not be applicable to settings with a different epidemiological context. In addition, the small number of patients with ESBL-EC bacteraemia may not have allowed us to find significant differences between both groups, especially regarding mortality. Conclusions We found that in our centre ESBL-EC bacteraemia is frequent among cancer patients, especially in those exposed to antibiotic pressure. All ESBL-EC strains were unrelated and most of them carried a CTX-M group enzyme. Patients with ESBL-EC bacteraemia received an inadequate initial empirical antibiotic therapy more frequently than those with bacteraemia caused by a susceptible strain, but there were no significant differences in mortality between the groups. Nevertheless, in the subgroup of neutropenic patients, ESBL-EC bacteraemia was associated with a poorer outcome and a higher overall mortality rate. It remains unclear whether the use of amikacin has a potential Table 6. Risk factors for overall mortality by univariate and multivariate analysis Risk factor Died n¼29 (%) Survived n¼102 (%) P Adjusted OR (95% CI) P Gender (female) 13 (45) 43 (42) ( ) Age (years, mean+sd) ( ) 0.41 Solid tumour 20 (69) 35 (34) ( ) Polymicrobial bacteraemia 8 (28) 16 (16) 0.17 Simultaneous corticosteroid therapy 16 (55) 29 (28) ( ) 0.04 Inadequate initial empirical antibiotic therapy 4 (14) 14 (14) 1.00 Time to adequate antibiotic therapy (days, median, range) 0 (0 3) 0 (0 10) 0.53 Co-existing infection 2 (7) 3 (3) 0.30 Shock at presentation 3 (10) 6 (6) 0.41 Intensive care unit admission 6 (21) 1 (1) ( ) ESBL-EC 6 (21) 11 (11) 0.20 Bacteraemia of urinary source 3 (10) 28 (27) ESBL-EC, extended-spectrum b-lactamase E. coli. 339

8 Gudiol et al. role as a component of the initial empirical antibiotic therapy for febrile neutropenia in patients with ESBL-EC bacteraemia. Likewise, controlled studies are needed to define which patients should receive a carbapenem agent from the beginning of therapy. Funding This study was supported by research grant REIPI RD06/0008 from the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, Spanish Network for the Research in Infectious Diseases, and by research grant from Fundació la Marató de TV3, Generalitat de Catalunya, Barcelona. The funding sources had no role in the study design, the collection, analysis and interpretation of the data or the decision to submit the manuscript for publication. Only the authors had full access to the data files for the study. Transparency declarations None to declare. References 1 Calatayud L, Arnan M, Liñares J et al. Prospective study of fecal colonization by extended-spectrum b-lactamase-producing Escherichia coli in neutropenic patients with cancer. Antimicrob Agents Chemother 2008; 52: Valverde A, Coque TM, Sánchez-Moreno MP et al. Dramatic increase in prevalence of fecal carriage of extended-spectrum b-lactamaseproducing Enterobacteriaceae during nonoutbreak situations in Spain. J Clin Microbiol 2004; 42: Rodríguez-Baño J, López-Cerero L, Navarro MD et al. Faecal carriage of extended-spectrum b-lactamase-producing Escherichia coli: prevalence, risk factors and molecular epidemiology. J Antimicrob Chemother 2008; 62: Paterson DL, Bonomo RA. Extended-spectrum b-lactamases: a clinical update. Clin Microbiol Rev 2005; 18: Munford RS. Severe sepsis and septic shock. In: Kasper DL, Braunwald E, Hanser SL et al., eds. Harrison s Principles of Internal Medicine. 16th edn. New York: McGraw-Hill, 2005; Clinical and Laboratory Standards Institute. Performance Standards for Antimicrobial Susceptibility Testing: Sixteenth Informational Supplement M100-S16. CLSI, Wayne, PA, USA, Tenover FC, Arbeit RD, Goering RV et al. Interpreting chromosomal DNA restriction patterns produced by pulsed-field gel electrophoresis: criteria for bacterial strain typing. J Clin Microbiol 1995; 33: Fang H, Ataker F, Hedin G et al. Molecular epidemiology of extended-spectrum b-lactamases among Escherichia coli isolates collected in a Swedish hospital and its associated health care facilities from 2001 to J Clin Microbiol 2008; 46: Matthew M, Harris AM, Marshall MJ et al. The use of analytical isoelectric focusing for detection and identification of b-lactamases. J Gen Microbiol 1975; 88: Hosmer DW, Lemeshow S. Logistic regression: variable selection. In: Hosmer DW, Lemeshow S, eds. Applied Logistic Regression. 2nd edn. New York: John Wiley & Sons, 2000; Rodríguez-Baño J, Navarro MD, Romero L et al. Bacteremia due to extended-spectrum b-lactamase-producing Escherichia coli in the CTX-M era: a new clinical challenge. Clin Infect Dis 2006; 43: Ho PL, Chan WM, Tsang KWT et al. Bacteremia caused by Escherichia coli producing extended-spectrum b-lactamase: a case-control study of risk factors and outcomes. Scand J Infect Dis 2002; 34: Kang CI, Kim SH, Park WB et al. Bloodstream infections due to extended-spectrum b-lactamase-producing Escherichia coli and Klebsiella pneumonia: risk factors for mortality and treatment outcome, with special emphasis on antimicrobial therapy. Antimicrob Agent Chemother 2004; 48: Peña C, Gudiol C, Tubau F et al. Risk-factors for acquisition of extended-spectrum b-lactamase producing Escherichia coli among hospitalised patients. Clin Microbiol Infect 2006; 12: Ortega M, Marco F, Soriano A et al. Analysis of 4758 Escherichia coli bacteremia episodes: predictive factors for isolation of an antibiotic-resistant strain and the impact on outcome. J Antimicrob Chemother 2009; 63: Pitout JD, Laupland KB. Extended-spectrum b-lactamaseproducing Enterobacteriaceae: an emerging public-health concern. Lancet Infect Dis 2008; 8: Trecarichi EM, Tumbarello M, Spanu T et al. Incidence and clinical impact of extended-spectrum-b-lactamase (ESBL) production and fluoroquinolones resistance in bloodstream infections caused by Escherichia coli in patients with hematological malignancies. J Infect 2009; 58: Siu LK, Lu PL, Hsueh PR et al. Bacteremia due to extended-spectrum b-lactamase-producing Escherichia coli and Klebsiella pneumoniae in a pediatric oncology ward: clinical features and identification of different plasmids carrying both SHV-5 and TEM-1 genes. J Clin Microbiol 1999; 37: Kim YK, Pai H, Lee HJ et al. Bloodstream infections by extended-spectrum b-lactamase-producing Escherichia coli and Klebsiella pneumoniae in children: epidemiology and clinical outcome. Antimicrob Agent Chemother 2002; 46: Rodríguez-Baño J, Alcalá JC, Cisneros JM et al. Community infections caused by extended-spectrum b-lactamase-producing Escherichia coli. Arch Intern Med 2008; 168: Ariffin H, Navaratnam P, Mohamed M et al. Ceftazidimeresistant Klebsiella pneumoniae bloodstream infection in children with febrile neutropenia. Int J Infect Dis 1999; 4: Rodríguez-Baño J, Navarro MD, Romero L et al. Risk-factors for emerging bloodstream infections caused by extended-spectrum b-lactamase-producing Escherichia coli. Clin Microbiol Infect 2008; 14: Paterson DL, Ko WC, Von Gottberg A et al. Outcome of cephalosporin treatment for serious infections due to apparently susceptible organisms producing extended-spectrum b-lactamases: implications for the clinical microbiology laboratory. J Clin Microbiol 2001; 39: Peña C, Gudiol C, Calatayud L et al. Infections due to Escherichia coli producing extended-spectrum b-lactamase among hospitalised patients: factors influencing mortality. J Hosp Infect 2008; 68: Schiappa DA, Hayden MG, Matushek G et al. Ceftazidime-resistant Klebsiella pneumoniae and Escherichia coli bloodstream infection: a case-control and molecular epidemiologic investigation. J Infect Dis 1996; 174: Hyle EP, Lipworth AD, Zaoutis TE et al. Impact of inadequate initial antimicrobial therapy on mortality in infections due to extendedspectrum b-lactamase-producing enterobacteriaceae: variability by site of infection. Arch Intern Med 2005; 165: Paterson DL. Recommendations for treatment of severe infections caused by Enterobacteriaceae producing extended-spectrum b-lactamases (ESBLs). Clin Microbiol Infect 2000; 6:

9 ESBL E. coli bacteraemia in cancer patients JAC 28 Hughes WT, Armstrong D, Bodey GP et al Guidelines for the use of antimicrobial agents in neutropenic patients with cancer. Clin Infect Dis 2002; 34: Jeon MH, Choi SH, Kwak YG et al. Risk factors for the acquisition of carbapenem-resistant Escherichia coli among hospitalized patients. Diagn Microbiol Infect Dis 2008; 62: Falagas ME, Rafailidilis PI, Kofteridis D et al. Risc factors of carbapenem-resistant Klebsiella pneumoniae infections: a matched-control study. J Antimicrob Chemother 2007; 60: Troillet N, Samore MH, Carmeli Y. Imipenem-resistant Pseudomonas aeruginosa: risk factors and antibiotic susceptibility patterns. Clin Infect Dis 1997; 25: