- Email: [email protected]
Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae
ARTICLE IN PRESS International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
Contents lists available at ScienceDirect
International Journal of Antimicrobial Agents j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j a n t i m i c a g
Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae Jean-Ralph Zahar a,b, Philippe Lesprit c, Stephane Ruckly b,d, Aurelia Eden e, Hitoto Hikombo f, Louis Bernard g, Stephan Harbarth h, Jean-François Timsit b,d,i, Christian Brun-Buisson j,* for the BacterCom Study Group 1 a
Infection Control Unit, Avicenne University Hospital, APHP, Paris-Nord University (UFR SMBH), Bobigny, France IAME U1137 Decision Science in Infectious Diseases, Inserm, Paris Diderot University, Paris, France c Antibiotic Team, Department of Biology, Hôpital Foch, Suresnes, France d Department of Biostatistics, Outcomerea, Bobigny, France e Infectious and Tropical Diseases Department, Perpignan General Hospital, Perpignan, France f Infectious and Tropical Diseases Department, Le Mans General Hospital, Le Mans, France g Internal Medicine and Infectious Diseases, Bretonneau University Hospital, Tours, France h Infection Control Program, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland i Medical and Infectious Diseases ICU, APHP, Bichat-Claude Bernard University Hospital, Paris, France j Medical Intensive Care Unit and Infection Control Unit, Henri Mondor University Hospital, APHP, Paris-Est Créteil University (UPEC), 51 Avenue du Maréchal de Lattre de Tassigny, 94000 Créteil, France b
A R T I C L E
I N F O
Article history: Received 10 June 2016 Accepted 24 September 2016 Keywords: Extended-spectrum β-lactamase Bacteraemia Antibiotic appropriateness Risk factors
A B S T R A C T
Extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) are endemic pathogens worldwide. Infection with ESBL-PE may be associated with inadequate antibiotic therapy and a poor outcome. However, risk factors for ESBL-PE community-acquired infections are ill-defined. An observational multicentre study was performed in 50 hospitals to identify the prevalence of and risk factors for communityacquired ESBL-PE bacteraemia. All patients presenting with community-onset Enterobacteriaceae bacteraemia were recorded over a 2-month period (between June and November 2013). Risk factors and 14-day outcomes of patients were investigated. Among 682 Enterobacteriaceae bacteraemia episodes recorded, 58 (8.5%) were caused by ESBL-PE. The most frequent species isolated were Escherichia coli (537; 76.7%) and Klebsiella spp. (68; 9.7%), of which 49 (9.1%) and 8 (11.8%), respectively, were ESBL-producers. Most ESBLPE episodes were healthcare-associated, and only 22 (38%) were apparently community-acquired. The main risk factor for community-acquired ESBL-PE bacteraemia was a prior hospital stay of ≥5 days within the past year. The overall 14-day survival was 90%; only 4 (6.9%) of 58 patients with ESBL-PE bacteraemia died. Inadequate initial antibiotic therapy was administered to 55% of patients with ESBL-PE bacteraemia but was not associated with increased 14-day mortality. Although many patients had community-onset ESBL-PE bacteraemia, almost two-thirds of the episodes were actually healthcare-associated, and true community-acquired ESBL-PE bacteraemia remains rare. In our essentially non-severely ill population, inappropriate initial therapy was not associated with a higher risk of mortality. © 2016 Published by Elsevier B.V.
1. Introduction Infections due to Enterobacteriaceae are the main cause of community and hospital-acquired bacteraemia, often leading to sepsis
* Corresponding author. Medical Intensive Care Unit, Henri Mondor University Hospital, APHP, Paris-Est Créteil University (UPEC), 51 Avenue du Maréchal de Lattre de Tassigny, 94000 Créteil, France. Fax: +33 1 4981 2542. E-mail address: [email protected] (C. Brun-Buisson). 1 Members of the BacterCom Study Group are listed in the Acknowledgements.
and septic shock [1]. In addition, the rising incidence of extendedspectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) both in hospitals and the community has led to an increased risk of inappropriate antimicrobial therapy and to increasing use of broadspectrum antibiotics [2]. The increasing incidence of ESBL-PE in the community questions the conventional therapeutic approach in patients with severe community-onset infection. A better knowledge of risk factors associated with infection caused by ESBL-PE could help physicians to identify the at-risk population in whom broader-spectrum empirical antimicrobial therapy might be warranted to lessen the risk of inappropriate treatment.
http://dx.doi.org/10.1016/j.ijantimicag.2016.09.032 0924-8579/© 2016 Published by Elsevier B.V.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
2
Indeed, some authors have attempted to develop and validate a tool for identifying patients harbouring ESBL-PE [3–10]. Therapeutic decision-making in these patients may be guided in part by the stratification between healthcare-associated infections (HCAIs) and true community-acquired cases. However, more information of the prevalence of and risk factors for ESBL-PE within the two categories of patients presenting with community-onset bacteraemia would be useful for this purpose. The first aim of this study was to evaluate the current prevalence of apparently community-acquired ESBL-PE infection among patients presenting with community-onset Enterobacteriaceae bacteraemia as well as the impact of ESBL-PE infection on the appropriateness of antimicrobial therapy and 14-day outcome. The second objective was to identify potential risk factors associated with truly community-acquired ESBL-PE bacteraemia. 2. Patients and methods This prospective observational study was conducted between June and November 2013 in 50 voluntary hospitals (49 in France and 1 in Geneva, Switzerland) among hospitals with an infectious diseases team who were contacted by the investigators. In each hospital, all or up to 20 (or 30 in large centres) consecutive adult patients admitted during a 2-month period with community-onset bacteraemia caused by Enterobacteriaceae were included. To exclude patients with hospital-acquired infections, those admitted from longterm care facilities or rehabilitation centres or with a recent (<7 days) history of acute-care hospital admission were not included. Patients with polymicrobial bacteraemia with micro-organisms other than Enterobacteriaceae (n = 48) were secondarily excluded from the analysis of the adequacy of therapy. Community-onset bacteraemia was defined as any positive blood culture drawn within the first 48 h of admission in patients coming directly from home or from a non-medical nursing home. Community-onset bacteraemia episodes were subsequently stratified into healthcare-associated bloodstream infection (HCA-BSI) or community-acquired bloodstream infection (CA-BSI) cases. The former group included patients fulfilling any of the criteria published by Friedman et al [11], i.e. receiving home intravenous (i.v.) therapy or home care, having a recent hospital stay (≥24 h) in the past 90 days, receiving haemodialysis, or having received i.v. chemotherapy within 30 days before occurrence of the bacteraemia episode; patients admitted from a nursing home were also included in this group. True CA-BSIs were defined as those recorded in patients without the above risk factors for HCAI. Data were collected prospectively by the corresponding clinical microbiologist and/or infectious diseases consultant of the antimicrobial therapy team in each hospital. Data were checked retrospectively for consistency and accuracy by two of the co-authors (CB-B and J-RZ). Demographics and clinical data were recorded, including age, sex, birth country and history of travel abroad within the past 3 years, underlying diseases, and co-morbidities [human immunodeficiency virus (HIV) infection, haematological malignancies, solid tumour, diabetes mellitus, solid organ transplant and chronic diseases]. Previous hospitalisation (≥24 h) and antibiotic therapy received within the past 12 months before admission were also recorded, as well as the time elapsed from therapy to the onset of bacteraemia. For statistical analysis, antibiotics were grouped into 11 different classes (see Supplementary material). Known previous ESBL colonisation or infection within the last 6 months was also recorded. Indwelling invasive devices were recorded at admission (urinary tract or intravascular catheters, tracheostomy). The source of bacteraemia was determined according to the US Centers for Disease Control and Prevention (CDC) criteria [12]; when no definite source was identified (including intravascular lines), bacteraemia was qualified as ‘primary’. The severity of underlying
diseases and co-morbidities were assessed using the revised Charlson’s weighted index of morbidity [13], and the severity of acute infection and organ dysfunction was defined as sepsis, severe sepsis or septic shock according to Bone’s criteria [14] as well as by the SOFA (Sepsis-related Organ Failure Assessment) score [15] recorded within 24 h of admission. As there are notable differences in ESBL-PE incidence rates in different regions of France [16], participating hospitals were stratified by their regional incidence according to data from the French antimicrobial surveillance network and this variable was included among potential risk factors for ESBL-PE bacteraemia. The Geneva area was considered to be at similar risk as the South-East France region with regard to ESBL-PE epidemiology [17]. As this study was observational, no specific guidelines were used for empirical treatment. Antimicrobial therapy was considered appropriate if at least one antimicrobial agent with in vitro activity against the isolate had been administered. Nonetheless, therapy combining an active aminoglycoside with another agent to which the strain was resistant was considered inappropriate, except for urinary tract infections (UTIs). The time elapsed (in hours) between the first positive blood sample and the first appropriate antibiotic was assessed, and the appropriateness of antimicrobial treatment administered at 24, 48 and 72 h after hospital admission was analysed according to definitive susceptibility test results. The endpoint was assessed as discharged alive within 14 days of admission; a poor outcome was defined as death within the first 2 weeks of hospitalisation or prolonged hospitalisation beyond Day 14. The study protocol was approved by the Ethics Committee (CPP Ile-de-France IX) in October 2012. Written informed consent was waived according to French law given the non-interventional design of the study, and patients were provided oral information. 2.1. Statistical analysis Results are reported as the median and interquartile range (IQR) or as counts (%). Categorical variables were compared using the χ2 test, and continuous variables were compared using Wilcoxon’s ranksum test for unpaired data. Multivariate analysis of variables associated (P ≤ 0.10) in univariate analysis with ESBL-PE bacteraemia was performed using stepwise backward logistic regression. Odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated. The effect of ESBL-PE infection on outcome at Day 14 was also examined, including all non-redundant variables yielding a P-value of <0.10 in the univariate analysis. Stepwise variable elimination was then performed and ESBL-PE and time to appropriate antimicrobial therapy were forced in the final model. Statistical tests were performed using SAS 9.4 (SAS Institute, Cary, NC) and Stata 13.1 (Stata Corp., College Station, TX). Differences were considered significant when the probability (p) of a type I error was ≤5%. 3. Results Of the 50 hospitals participating in this prospective study, 27 (54%) were teaching hospitals (26 in France and 1 in Geneva) and 23 (46%) were non-teaching hospitals; all French regions were represented. A mean of 14 patients (range, 3–30 patients) was included per participating hospital over the 2-month period of the survey. A total of 693 patients with a single episode of Enterobacteriaceae bacteraemia were recorded; 11 patients were secondarily excluded because of missing information (1 patient), missing susceptibility (1 patient) or hospital-acquired infection (9 patients) (Fig. 1). The remaining 682 patients with community-onset Enterobacteriaceae bacteraemia remained in the analysis, of whom 325 (47.7%) had risk factors for HCAI and 357 (52.3%) had CA-BSI.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
3
Fig. 1. Flow chart of the 693 patients with Enterobacteriaceae bacteraemia included in the study cohort. ESBL-PE, extended-spectrum β-lactamase producing Enterobacteriaceae.
The 682 patients had a median (IQR) age of 74.9 (61–84) years; 352 (51.6%) were female. A total of 226 patients (33.1%) had at least one co-morbid condition; 170 (24.9%) patients had diabetes mellitus, 39 (5.7%) received long-term steroid therapy and 73 (11%) received immunosuppressive therapy, and 82 (12%) had active cancer (Table 1). The vast majority of patients (89%) were admitted directly from home, and 70 (10.3%) were nursing home residents. A total of 133 patients (19.5%) were native from ‘high-risk countries’ as defined above, and most of the patients were French or Swiss residents (96%). A total of 108 patients (15.8%) had travelled abroad within the last year, among which 92% had travelled to a high-risk country. Apart from characteristics defining HCAI, the two subgroups of CA-BSI and HCA-BSI bacteraemia differed in several respects (Table 1): patients having CA-BSI were more often women (55.5% vs. 47.4%; P = 0.03), were less often admitted to a university hospital (54.3% vs. 65.8%; P = 0.002) and were less often in a region with a higher level of ESBL (58.5% vs. 68.3%; P = 0.008) than those with HCA-BSI; more CA-BSI patients had travelled abroad within the past 3 years (18.8% vs 12.6%; P = 0.03), including in high-risk countries (16.8% vs. 10.5%; P = 0.02). Previous hospital admission (for ≥24 h) or administration of an antibiotic course within the past year was recorded in 71 (19.9%) and 85 (23.8%) patients with CA-BSI, respectively; as expected, patients with HCA-BSI had higher corresponding rates at 75.4% and 46.8%, respectively; 10.2% of patients with CABSI and 52.1% of those with HCA-BSI had been hospitalised for >5 days. Of 26 patients (3.8%) previously known as ESBL-PE carriers, 3 (0.8%) and 23 (7.1%) belonged to the CA- and HCA-BSI groups, respectively (Table 1). Eighty-four patients (12.3%) required admission to the intensive care unit within the first 24 h of hospital admission, with a similar rate between the two groups (13.7% vs. 10.8%; P = 0.24).
A source of infection was confirmed in 519 cases (76.1%) and was considered probable on clinical grounds and/or imaging studies in 153 patients (22.4%) (Table 2). Their distribution differed between CA-BSI and HCA-BSI because of a higher proportion of intraabdominal infections and a lower proportion of intravascular infections among CA-BSI patients. Among patients having ESBLPE bacteraemia, however, this distribution did not differ between CA-BSI and HCA-BSI episodes. Bacterial species identified are shown in Supplementary Table S1; 621 episodes (91.1%) were monomicrobial, whereas 61 patients (8.9%) had polymicrobial bacteraemia. Among the latter, 21 (34.4%) blood cultures grew two or three different Enterobacteriaceae. ESBL-producers accounted for 8.7% (61/700) of all Enterobacteriaceae recovered. Of the 58 patients having infection with an ESBL-PE, 36 (62%) had risk factors for HCAI and 22 (38%) were considered truly community-acquired, representing 3.2% of all community-onset episodes and 6.2% of community-acquired Enterobacteriaceae bacteraemia episodes. Compared with other patients having CABSI, patients having infection with ESBL-PE tended to be younger (P = 0.09), were more often hospitalised in a region with a higher mean level of ESBL (P = 0.02) and had been more often previously hospitalised (P = 0.01) within the past year and for longer periods. After multivariate analysis, the only variable significantly associated with ESBL-PE infection among CA-BSI episodes was a previous hospital stay of ≥5 days (P = 0.01), although the regional incidence fell short of statistical significance (P = 0.058). Variables associated with healthcare-associated ESBL-PE bacteraemia are shown in Supplementary Table S2. A total of 614 patients (90.0%) received empirical therapy within the first 24 h of hospital admission and the others received delayed therapy; 373 (60.7%) initially received combination therapy, including 99 (16.1%) receiving three or more antibiotics within the
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
4
Table 1 Clinical characteristics of patients with community-onset Enterobacteriaceae bacteraemia, contrasting healthcare-associated bloodstream infection (HCA-BSI) with truly community-acquired bloodstream infection (CA-BSI). Variable a
Total (N = 682)
HCA-BSI ESBL-PE (n = 36)
Non-ESBL-PE (n = 289)
ESBL-PE (n = 22)
Non-ESBL-PE (n = 335)
Female sex Age (years) [median (IQR)] Native high-risk country Living in low-risk country Travel to high-risk country within 1 year Known ESBL carriage Prior antibiotic therapy within 1 year Prior hospitalisation within 1 year Co-morbid conditions Chronic respiratory disease Chronic cardiac disease Chronic liver disease Haematological disease Chemotherapy treatment Chronic haemodialysis Home care Admission hospital category University Non-university Regional ESBL incidence Low risk High risk Type of admission Medical Surgical Emergency surgery Origin of patient Home Nursing home SOFA score within 24 h of admission [median (IQR)] Charlson score [median (IQR)] Clinical severity of sepsis b Sepsis Severe sepsis Septic shock
352 (51.6) 74.9 (61–84) 133 (19.5) 669 (98.1) 94 (13.8) 26 (3.8) 237 (34.8) 316 (46.3) 226 (33.1) 23 (3.4) 25 (3.7) 62 (9.1) 36 (5.3) 79 (11.6) 11 (1.6) 70 (10.3)
13 (36.1) 79.4 (62–88) 12 (33.3) 35 (97.2) 9 (25.0) 10 (27.8) 28 (77.8) 32 (88.9) 24 (66.7) 0 1 (2.8) 9 (25.0) 3 (8.3) 7 (19.4) 1 (2.8) 10 (27.8)
141 (48.8) 74.8 (64–85) 50 (17.3) 285 (98.6) 25 (8.7) 13 (4.5) 124 (42.9) 213 (73.7) 157 (54.3) 13 (4.5) 13 (4.5) 42 (14.5) 29 (10.0) 72 (24.9) 10 (3.5) 60 (20.8)
13 (59.1) 68.5 (48–82) 5 (22.7) 22 (100) 4 (18.2) 1 (4.5) 8 (36.4) 9 (40.9) 1 (4.5) 0 0 0 0 0 0 0
185 (55.2) 75 (61–83) 66 (19.7) 327 (97.6) 56 (16.7) 2 (0.6) 77 (23.0) 62 (18.5) 44 (13.1) 10 (3.0) 11 (3.3) 11 (3.3) 4 (1.2) 0 0 0
408 (59.8) 274 (40.2)
25 (69.4) 11 (30.6)
189 (65.4) 100 (34.6)
14 (63.6) 8 (36.4)
180 (53.7) 155 (46.3)
251 (36.8) 431 (63.2)
4 (11.1) 32 (88.9)
99 (34.3) 190 (65.7)
4 (18.2) 18 (81.8)
144 (43.0) 191 (57.0)
608 (89.1) 57 (8.4) 17 (2.5)
35 (97.2) 1 (2.8) 0
261 (90.3) 21 (7.3) 7 (2.4)
20 (90.9) 1 (4.5) 1 (4.5)
292 (87.2) 34 (10.1) 9 (2.7)
607 (89.0) 75 (11.0) 1 (0–3) 2 (0–3)
27 (75.0) 9 (25.0) 2 (0.5–4) 3.5 (2–5.5)
223 (77.2) 66 (22.8) 2 (0–4) 2 (1–4)
431 (63.2) 91 (13.3) 49 (7.2)
22 (61.1) 6 (16.7) 0
175 (60.6) 42 (14.5) 20 (6.9)
CA-BSI
22 (100) 0 1 (0–3) 0.5 (0–2)
335 (100) 0 1 (0–3) 0 (0–2)
12 (54.5) 4 (18.2) 3 (13.6)
222 (66.3) 39 (11.6) 26 (7.8)
ESBL-PE, extended-spectrum β-lactamase-producing Enterobacteriaceae; IQR, interquartile range; SOFA, Sepsis-related Organ Failure Assessment. a Values are presented as n (%) unless otherwise stated. b 111 patients (16.3%) did not fulfil criteria for sepsis.
first 24 h. Initial therapy included a third-generation cephalosporin in 497 cases (72.9%), amoxicillin or amoxicillin/clavulanic acid in 69 cases (10.1%) and a fluoroquinolone in 39 cases (5.7%). Overall, an appropriate empirical regimen was administered within the first 24 h of bacteraemia in 534 (85.6%) of 624 patients with non-ESBLPE bacteraemia and 26 (44.8%) of the 58 patients with ESBL-PE bacteraemia (Table 3). The median (IQR) time elapsed between drawing the first blood culture and the first administration of adequate antibiotic therapy was 0.1 (0–0.8) days overall and, respectively, 0.1 (0–0.5) days and 1.7 (0.7–3.2) days for non-ESBLPE and ESBL-PE bacteraemia (P < 0.001). Among the 22 patients having community-acquired ESBL-PE bacteraemia, 50%, 77% and 82%
received an adequate antibiotic regimen within the first 24, 48 and 72 h, respectively. The overall 14-day mortality in patients with communityonset bacteraemia was 9.8%. As expected, patients having HCABSI had a poorer outcome than those with CA-BSI, with a higher 14-day mortality (P = 0.036), poorer outcome at Day 14 (P = 0.004) and longer hospital stay (P = 0.001). Among those with CA-BSI, patients having infection with ESBL-PE had a comparable outcome to those without ESBL-PE infection. By multivariable analysis, only age and clinical severity (i.e. SOFA score), but not ESBL-PE infection (P = 0.54) or time to adequate antibiotic therapy (P = 0.46), were associated with death at 14 days. However, a time to adequate therapy
Table 2 Source of infection identified in patients with community-onset Enterobacteriaceae bacteraemia. Source of infection a
Urinary tract Respiratory tract Intra-abdominal infection Intravascular infection Others b None identified (primary bacteraemia)
Total (N = 682)
CA-BSI ESBL-PE (n = 22)
Non-ESBL-PE (n = 335)
HCA-BSI ESBL-PE (n = 36)
Non-ESBL-PE (n = 289)
383 (56.2%) 27 (4.0%) 200 (29.3%) 21 (3.1%) 19 (2.8%) 37 (5.4%)
13 0 4 0 1 3
196 11 113 1 7 12
21 0 9 0 0 3
153 12 74 20 8 19
CA-BSI, community-acquired bloodstream infection; HCA-BSI, healthcare-associated bloodstream infection; ESBL-PE, extended-spectrum β-lactamase-producing Enterobacteriaceae. a The source was considered confirmed in 519 patients (when identical micro-organisms were recovered from the infectious focus and blood) and presumed in the remaining 153; 15 patients had multiple infection sites and 10 patients had missing information on the source of infection (3 CA-BSI and 7 HCA-BSI). b Including soft-tissue infection, meningitis, osteomyelitis and arthritis.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
Table 3 Timing of adequacy of antibiotic therapy in 682 patients with community-onset bacteraemia, according to extended-spectrum β-lactamase (ESBL) status. Time to adequate antibiotic therapy Prior to hospital admission In hospital ≤24 h >24–48 h >48–72 h >72 h Total
Non-ESBL-PE bacteraemia [n (%)] 3 (0.5) 534 (85.6) 57 (9.1) 16 (2.6) 14 (2.2) 624 (100)
ESBL-PE bacteraemia [n (%)] 0 26 (44.8) 16 (27.6) 6 (10.3) 10 (17.2) 58 (100)
Total evaluable bacteraemia 3 (0.6) 560 (82.1) 73 (10.7) 22 (3.2) 24 (3.5) 682 (100)
of >24 h was marginally associated with a prolonged hospital stay (OR = 1.83, 95% CI 0.97–3.48; P = 0.06). 4. Discussion In this large prospective observational cohort study, the prevalence of ESBL-PE among patients admitted with communityonset bacteraemia reached 8% of all episodes. However, twothirds of episodes actually qualified as HCAI, and true communityacquired ESBL-PE bacteraemia was identified in only 22 patients (3.2%). Community-acquired ESBL-PE bacteraemia was associated with a history of prior hospital stay of ≥5 days and with regional ESBL incidence. Despite a much higher rate of inadequate antimicrobial therapy than in other patients, patients with ESBL-PE bacteraemia had a comparable outcome to other patients. Recent epidemiological data suggest increasing rates of ESBLPE invasive infections throughout European countries [18]. In a previous study conducted 3 years before ours in 10 French hospitals [19], ESBL-PE accounted for 3.8% of all Enterobacteriaceae bacteraemia diagnosed within the first 48 h of hospital admission, and most ESBL-PE bacteraemias were healthcare-associated. The current study suggests an increased rate of community-onset ESBL-PE bacteraemia. Despite ESBL-PE spreading in the community in France [20], the prevalence of truly community-acquired ESBLPE invasive infection however appears to remain low. Indeed, twothirds of the patients in the current study had criteria for healthcareassociated ESBL-PE bacteraemia according to Friedman et al [11]. Moreover, the major risk factor for community-acquired ESBL-PE infection appeared to be a prior hospital stay of ≥5 days in the previous year, suggesting that many of these community-onset cases without the accepted risk factors for HCAI could be in fact acquired in the hospital. Indeed, carriage of ESBL-PE is known to persist for months in some patients [21]. Patients having had a prior hospitalisation dating back more than 3 months can be a persistent carrier if colonisation occurred during that hospitalisation. The current results suggest that, beyond the conventional criteria for HCAI, patients having had a prior hospitalisation within the preceding year should be considered at risk for ESBL-PE infection. Applying this rule to our patients would have resulted in considering at risk 44 (75.9%) of all patients with community-onset ESBLPE infection. Together with the potential influence of regional ESBL incidence, these results suggest a major driving role for healthcareassociated factors in the occurrence of community-onset ESBL-PE infections. The low rate of truly community-acquired infection in this study can be explained in light of recent data from the literature. Indeed, only 10–20% of carriers appear to be exposed to a risk of secondary infection [22,23]. In a multicentre prospective study conducted in ESBL-producing Escherichia coli carriers, Ruppé et al suggested that the occurrence of UTI in previously known carriers was related to the relative faecal abundance of ESBL-PE induced by antibiotic therapy [24]. Likewise, in a retrospective case–control study where
5
controls were selected on the basis of age, hospital location and exposure time, the only two factors associated with infection in previously colonised patients were urinary catheterisation and therapy with a β-lactam/β-lactamase inhibitor combination prior to infection. Thus, aside from healthcare-associated risk factors, antibiotic exposure also appears to play a major role in the occurrence of ESBL-PE infection [22], with cephalosporin and fluoroquinolone classes most commonly implicated [25]. In the current study, previous antibiotic therapy was not significantly associated with community-acquired ESBL-PE infection, although investigation into the various antibiotic classes and time elapsed since receipt of antibiotics suggested that previous therapy with antibiotics having antianaerobic activity was consistently associated with a higher risk of community onset ESBL-PE bacteraemia (Supplementary Table S3). These findings are consistent with animal models [26] and clinical studies suggesting that preservation of the anaerobic flora may be critical for preventing the acquisition of multidrug-resistant microorganisms and subsequent infection [27]. The rate of inadequate first-line antibiotic therapy in this study underlines the clinical importance of prediction rules to help clinicians identify those patients who should be treated with a broadspectrum antibiotic. A number of studies have demonstrated that patients with healthcare-associated Gram-negative BSI experience delay in receipt of appropriate antibiotic therapy [28]. Several studies attempted to identify risk factors associated with ESBL-PE community-onset infection [3–10]. However, many methodological issues limit the interpretation of these studies. Indeed, many have focused on specific species, mostly on E. coli [4,5,7–9]. Tumbarello et al [10] attempted to establish a predictive clinical score to identify patients having ESBL-PE recovered in clinical samples from a retrospective chart review of 849 adult patients admitted to a large hospital in Rome (Italy). The score showed good discrimination; however, this study combined nosocomial and community-acquired infections. Other studies sought to identify risk factors associated with carriage [29] or infection with a specific species such as ESBLproducing Enterobacter spp. [30] or ESBL-producing E. coli [31] or at a specific site of infection, mostly the urinary tract [32]. In the current study, no association between the appropriateness of antimicrobial therapy and mortality was found, irrespective of the time elapsed between admission and the first appropriate dose of antibiotic. However, inappropriate empirical antibiotic therapy has been shown to increase mortality essentially in critical ill patients with severe infection [33]. The results of the current study could thus be explained, at least in part, by the expected low proportion of patients having severe sepsis or septic shock in this hospital-based study; they are also consistent with some conflicting results with regard to the association between inappropriate initial therapy of ESBL-PE infection and mortality, especially for UTIs [34,35]. Nevertheless, delayed adequate therapy tended to be associated with prolonged hospital stay. This study has several limitations. First, it was conducted in France and Geneva and the results may not be applicable to other countries with a higher or lower prevalence of ESBL. However, the prevalence of ESBL-PE carriage in non-hospitalised subjects is estimated at 6% in France [20], which is similar to reported prevalences from other Western European countries such as Germany [36] and The Netherlands [37]. Second, the relatively low number of patients with community-acquired bacteraemia reduced the ability to identify potential risk factors for ESBL-PE infection. In conclusion, despite the increased carriage rate of ESBL-PE in the community, community-acquired ESBL-PE bacteraemia remains rare in France and Geneva, although patients may remain at risk for ESBL-PE infection after a prolonged hospitalisation, beyond the conventional 90 days included in the definition of HCAI. In our general hospital population, inadequate first-line antibiotic therapy was common but was not associated with a higher risk of death.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
6
These results thus do not suggest altering the conventional approach to therapy in most instances and support a highly selective use of broad-spectrum antibiotic therapy in non-severely ill patients with community-acquired Gram-negative bacteraemia.
Acknowledgements The BacterCom Study Group: Steering committee: Christian Brun-Buisson, Jean Carlet, Pierre Dellamonica, Rémy Gauzit, Christian Rabaud. Scientific committee: Jean-Ralph Zahar, Jean-François Timsit, Serge Alfandari, Robert Cohen, Bruno Coignard, Céline Pulcini, Jérôme Robert. Investigators: Nejla Aissa, Alain Losniewski, Sandrine Henard, Christian Rabaud (CHU Nancy); Cécile Janssen, Léonardo Pagani, Stéphane Bland, Jacques Gaillat, Jean Pierre Bru, Pierre Bonnin, Gaelle Clavère, Virginie Vitrat (CHG Annecy); Laurence Maulin (CHG Aix); Luc Quarsaet, Helene Roger, Jean-Baptiste Euzen (CHU Brest); Eric Denes, Elodie Couve-Daecon (CHU Limoges); Catherine Lechiche, Jean-Philippe Lavigne (CHU Nimes); Véronique Mondain, Celine Pulcini (CHU Nice); Veronique Garnier, Dorothée Lambert, Maxime Hentzien, Yohan Nguyen (CHU Reims); Jean Paul Talarmin, Pascaline Rameau, Françoise Geffroy (CH Intercommunal de Cornouaille); Arnaud Salmon-Rousseau, Vincent Cattoir (CHU Côte de Nacre); Antoine Froissart, Said Aberrane (CH Intercommunal de Créteil); Laurence Legout (CH Alpes-Leman); France Roblot, Guillaume Beraud (CHU Poitiers); Patricia Pavese (CHU Grenoble): Olivier Belmonte, Nicolas Traversier, Laurent Bellec, Marie Lagrange-Xelot, MariePierre Moiton, Barbara Kuli (CHU la Réunion); Pascal Longuet, Agnes Scanvic (CH Argenteuil); Lea Colombain, Milagros Ferreyra, Maxime Jean, Aurelia Eden, Etienne Laurens, Matthieu Saada, Hugues Aumaitre, Martine Malet, Frederique Roustant (CH Perpignan); Vincent LeMoing, Jean-Pierre Helene (CHU Montpellier); Jonathan Theodore, Caroline Garandeau (CH Angoulême); Jocelyne Caillon, Benjamin Olivero, David Boutoille, Didier Lepelletier (CHU Nantes); Lucia Grandière-Perez, Aurélie Beaudron, Céline Ramanantsoa, Catherine Varache, Nicolas Varache, Nicolas Crochette, Hikombo Hitoto, Sophie Blanchi, Pascale Penn (CH Le Mans); Bernard Louis (CHU Tours); Dominique Vanjak, Marie-Christine Escande (Institut Curie); René-Gilles Patrigeon, Stéphanie Honoré (CH Auxerre); Audrey Barrelet, Marlène Amara (CH Versailles); Jean-Marc Galempoix, JeanClaude Réveil, Patricia Bucchiotty (CH Charleville-Mézières); Carine Couzigou, Alban LeMonnier (CH St Joseph); Emilie Boidin, Stephane Sire (CH Cahors); Nelly Luizy, Caroline Bronstain (CHG Montfermeil); Stephan Harbarth, Carolina Fankhauser, Caroline Brossier, Elodie von Dach (HU Genève); Colette Sauve (CH Longjumeau); Colette Sauve, Marie-Noelle Adam (CH Orsay); Laurent Hocqueloux, Laurent Bret (CH Orléans); Marie-Hélène André, Régine Barruet, Marion Challier (CH Montreuil); Andre Cabié, Benoit Roze (CHU Martinique); Patrick Plesiat, Katy Jeanot, Isabelle Patry, Joséphine Moreau, Catherine Chirouze (CHU Besançon); Aaron Laurent, Marie-Nadège Bachelier (CH Bourges); Guillaume Kac, Nicolas Farthouat, Nicolas Feuillebois, Thomas Guimard, Sandra Bourdon (CHD Vendée); Sophie Nguyen, Serge Alfandari (CHU Tourcoing); Houcine Bentayeb, Mari-Thérèse Albertini (CH St Quentin); Bruno Marchou, Noémie Gaudré, Damien Dubois (CHU Toulouse, Assistance Publique-Hôpitaux de Paris); Christelle Guillet-Caruba, Florence Doucet Populaire (Antoine Béclère); Nicolas Fortineau, Leila Escaut (Kremlin-Bicêtre); Philippe Lesprit, Vincent Fihman (Henri Mondor); Rémy Gauzit, Josette Raymond, Solene Kerneis (Cochin); Veronique Leflon-Guibout, Agnes Lefort (Beaujon); Aurelien Dinh, Anne-Laure Roux (Ambroise Paré); Sophie Abgrall, Françoise Jauréguy (Avicenne); Jean-Christophe Lucet (Bichat); Aurelien Dinh, Nathalie Soismier (Raymond Poincaré); Amanda Lopes, Emmanuelle Cambau (Lariboisière); Veronique Manceron, Guilène Barnaud (Louis Mourier).
Supporting institutions: The BacterCom study was supported by the Institut Maurice Rapin, the Société française de microbiologie (SFM), the Société de pathologie infectieuse de langue française (SPILF) and the association AC-de-BMR. Funding: This study was funded in part by the Institut Maurice Rapin and the Société de pathologie infectieuse de langue française (SPILF). Competing interests: J-RZ participates on an advisory board for Merck Sharp & Dohme; SH has participated on advisory boards and has received honoraria from bioMérieux, Novartis, Johnson & Johnson and Da Volterra. All other authors declare no competing interests. Ethical approval: The study protocol was approved by the Ethics Committee (CPP Ile-de-France IX) in October 2012. Written informed consent was waived according to French law. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijantimicag.2016.09.032. References [1] Laupland KB, Church DL. Population-based epidemiology and microbiology of community-onset bloodstream infections. Clin Microbiol Rev 2014;27:647–64. [2] Zilberberg MD, Shorr AF, Micek ST, Vazquez-Guillamet C, Kollef MH. Multi-drug resistance, inappropriate initial antibiotic therapy and mortality in Gramnegative severe sepsis and septic shock: a retrospective cohort study. Crit Care 2014;18:596. [3] Ben-Ami R, Rodriguez-Bano J, Arslan H, Pitout JD, Quentin C, Calbo ES, et al. A multinational survey of risk factors for infection with extended-spectrum β-lactamase-producing Enterobacteriaceae in nonhospitalized patients. Clin Infect Dis 2009;49:682–90. [4] Han JH, Kasahara K, Edelstein PH, Bilker WB, Lautenbach E. Risk factors for infection or colonization with CTX-M extended-spectrum-β-lactamase-positive Escherichia coli. Antimicrob Agents Chemother 2012;56:5575–80. [5] Kang CI, Wi YM, Lee MY, Ko KS, Chung DR, Peck KR, et al. Epidemiology and risk factors of community onset infections caused by extended-spectrum β-lactamase-producing Escherichia coli strains. J Clin Microbiol 2012;50:312–17. [6] Kang CI, Wi YM, Ko KS, Chung DR, Peck KR, Lee NY, et al. Outcomes and risk factors for mortality in community-onset bacteremia caused by extendedspectrum β-lactamase-producing Escherichia coli, with a special emphasis on antimicrobial therapy. Scand J Infect Dis 2013;45:519–25. [7] Kuster SP, Hasse B, Huebner V, Bansal V, Zbinden R, Ruef C, et al. Risks factors for infections with extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella pneumoniae at a tertiary care university hospital in Switzerland. Infection 2010;38:33–40. [8] Rodríguez-Baño J, Navarro MD, Romero L, Muniain MA, de Cueto M, Ríos MJ, et al. Bacteremia due to extended-spectrum β-lactamase-producing Escherichia coli in the CTX-M era: a new clinical challenge. Clin Infect Dis 2006;43:1407–14. [9] Rodriguez-Bano J, Picon E, Gijon P, Hernandez JR, Ruiz M, Pena C, et al. Community-onset bacteremia due to extended-spectrum β-lactamaseproducing Escherichia coli: risk factors and prognosis. Clin Infect Dis 2010;50:40–8. [10] Tumbarello M, Trecarichi EM, Bassetti M, De Rosa FG, Spanu T, Di Meco E, et al. Identifying patients harboring extended-spectrum-β-lactamase-producing Enterobacteriaceae on hospital admission: derivation and validation of a scoring system. Antimicrob Agents Chemother 2011;55:3485–90. [11] Friedman ND, Kaye KS, Stout JE, McGarry SA, Trivette SL, Briggs JP, et al. Health care-associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med 2002;137:791–7. [12] Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309–32. [13] Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol 2011;173:676–82. [14] Bone RC, Balk RA, Cerra FB, Dellinger EP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest 1992;101:1656–62. [15] Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22:707–10. [16] Carbonne A, Arnaud I, Maugat S, Marty N, Dumartin C, Bertrand X, et al. National multidrug-resistant bacteria (MDRB) surveillance in France through the RAISIN network: a 9 year experience. J Antimicrob Chemother 2013;68:954–9.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
[17] Pasricha J, Koessler T, Harbarth S, Schrenzel J, Camus V, Cohen G, et al. Carriage of extended-spectrum β-lactamase-producing Enterobacteriaceae among internal medicine patients in Switzerland. Antimicrob Resist Infect Control 2013;2:20. [18] European Centre for Disease Prevention and Control (ECDC). Antimicrobial resistance surveillance in Europe. Annual report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) 2013. Stockholm, Sweden: ECDC; 2014. Available from: http://ecdc.europa.eu/en/publications/Publications/ antimicrobial-resistance-surveillance-europe-2013.pdf. [Accessed 7 November 2016]. [19] Shoai Tehrani M, Hajage D, Fihman V, Tankovic J, Cau S, Day N, et al. Gramnegative bacteremia: which empirical antibiotic therapy? Med Mal Infect 2014;44:159–66. [20] Nicolas-Chanoine MH, Gruson C, Bialek-Davenet S, Bertrand X, Thomas-Jean F, Bert F, et al. 10-Fold increase (2006–11) in the rate of healthy subjects with extended-spectrum β-lactamase-producing Escherichia coli faecal carriage in a Parisian check-up centre. J Antimicrob Chemother 2012;68:562–8. [21] Titelman E, Hasan CM, Iversen A, Naucler P, Kais M, Kalin M, et al. Faecal carriage of extended-spectrum β-lactamase-producing Enterobacteriaceae is common 12 months after infection and is related to strain factors. Clin Microbiol Infect 2014;20:O508–15. [22] Goulenok T, Ferroni A, Bille E, Lecuyer H, Join-Lambert O, Descamps P, et al. Risk factors for developing ESBL E. coli: can clinicians predict infection in patients with prior colonization? J Hosp Infect 2013;84:294–9. [23] Reddy P, Malczynski M, Obias A, Reiner S, Jin N, Huang J, et al. Screening for extended-spectrum β-lactamase-producing Enterobacteriaceae among high-risk patients and rates of subsequent bacteremia. Clin Infect Dis 2007;45:846–52. [24] Ruppé E, Lixandru B, Cojocaru R, Buke C, Paramythiotou E, Angebault C, et al. Relative fecal abundance of extended-spectrum-β-lactamase-producing Escherichia coli strains and their occurrence in urinary tract infections in women. Antimicrob Agents Chemother 2013;57:4512–17. [25] Biehl LM, Schmidt-Hieber M, Liss B, Cornely OA, Vehreschild MJ. Colonization and infection with extended spectrum β-lactamase producing Enterobacteriaceae in high-risk patients—review of the literature from a clinical perspective. Crit Rev Microbiol 2016;42:1–16. [26] Bhalla A, Pultz NJ, Ray AJ, Hoyen CK, Eckstein EC, Donskey CJ. Antianaerobic antibiotic therapy promotes overgrowth of antibiotic-resistant, Gram-negative bacilli and vancomycin-resistant enterococci in the stool of colonized patients. Infect Control Hosp Epidemiol 2003;24:644–9.
7
[27] Donskey CJ. The role of the intestinal tract as a reservoir and source for transmission of nosocomial pathogens. Clin Infect Dis 2004;39:219–26. [28] Moehring RW, Sloane R, Chen LF, Smathers EC, Schmader KE, Fowler VG Jr, et al. Delays in appropriate antibiotic therapy for Gram-negative bloodstream infections: a multicenter, community hospital study. PLoS ONE 2013;8:e76225. [29] Leistner R, Gurntke S, Sakellariou C, Denkel LA, Bloch A, Gastmeier P, et al. Bloodstream infection due to extended-spectrum β-lactamase (ESBL)-positive K. pneumoniae and E. coli: an analysis of the disease burden in a large cohort. Infection 2014;42:991–7. [30] Qureshi ZA, Paterson DL, Pakstis DL, Adams-Haduch JM, Sandkovsky G, Sordillo E, et al. Risk factors and outcome of extended-spectrum β-lactamase-producing Enterobacter cloacae bloodstream infections. Int J Antimicrob Agents 2011;37:26–32. [31] Rodriguez-Bano J, Alcala JC, Cisneros JM, Grill F, Oliver A, Horcajada JP, et al. Community infections caused by extended-spectrum β-lactamase-producing Escherichia coli. Arch Intern Med 2008;168:1897–902. [32] Calbo E, Romani V, Xercavins M, Gomez L, Vidal CG, Quintana S, et al. Risk factors for community-onset urinary tract infections due to Escherichia coli harbouring extended-spectrum β-lactamases. J Antimicrob Chemother 2006;57:780–3. [33] Zahar JR, Timsit JF, Garrouste-Orgeas M, Francais A, Vesin A, Descorps-Declere A, et al. Outcomes in severe sepsis and patients with septic shock: pathogen species and infection sites are not associated with mortality. Crit Care Med 2011;39:1886–95. [34] Peralta G, Lamelo M, Alvarez-Garcia P, Velasco M, Delgado A, Horcajada JP, et al. Impact of empirical treatment in extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella spp. bacteremia. A multicentric cohort study. BMC Infect Dis 2012;12:245. [35] Rottier WC, Ammerlaan HS, Bonten MJ. Effects of confounders and intermediates on the association of bacteraemia caused by extended-spectrum β-lactamaseproducing Enterobacteriaceae and patient outcome: a meta-analysis. J Antimicrob Chemother 2012;67:1311–20. [36] Valenza G, Nickel S, Pfeifer Y, Eller C, Krupa E, Lehner-Reindl V, et al. Extendedspectrum-β-lactamase-producing Escherichia coli as intestinal colonizers in the German community. Antimicrob Agents Chemother 2014;58:1228– 30. [37] Huijbers PM, de Kraker M, Graat EA, van Hoek AH, van Santen MG, de Jong MC, et al. Prevalence of extended-spectrum β-lactamase-producing Enterobacteriaceae in humans living in municipalities with high and low broiler density. Clin Microbiol Infect 2013;19:E256–9.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
Contents lists available at ScienceDirect
International Journal of Antimicrobial Agents j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / l o c a t e / i j a n t i m i c a g
Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae Jean-Ralph Zahar a,b, Philippe Lesprit c, Stephane Ruckly b,d, Aurelia Eden e, Hitoto Hikombo f, Louis Bernard g, Stephan Harbarth h, Jean-François Timsit b,d,i, Christian Brun-Buisson j,* for the BacterCom Study Group 1 a
Infection Control Unit, Avicenne University Hospital, APHP, Paris-Nord University (UFR SMBH), Bobigny, France IAME U1137 Decision Science in Infectious Diseases, Inserm, Paris Diderot University, Paris, France c Antibiotic Team, Department of Biology, Hôpital Foch, Suresnes, France d Department of Biostatistics, Outcomerea, Bobigny, France e Infectious and Tropical Diseases Department, Perpignan General Hospital, Perpignan, France f Infectious and Tropical Diseases Department, Le Mans General Hospital, Le Mans, France g Internal Medicine and Infectious Diseases, Bretonneau University Hospital, Tours, France h Infection Control Program, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland i Medical and Infectious Diseases ICU, APHP, Bichat-Claude Bernard University Hospital, Paris, France j Medical Intensive Care Unit and Infection Control Unit, Henri Mondor University Hospital, APHP, Paris-Est Créteil University (UPEC), 51 Avenue du Maréchal de Lattre de Tassigny, 94000 Créteil, France b
A R T I C L E
I N F O
Article history: Received 10 June 2016 Accepted 24 September 2016 Keywords: Extended-spectrum β-lactamase Bacteraemia Antibiotic appropriateness Risk factors
A B S T R A C T
Extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) are endemic pathogens worldwide. Infection with ESBL-PE may be associated with inadequate antibiotic therapy and a poor outcome. However, risk factors for ESBL-PE community-acquired infections are ill-defined. An observational multicentre study was performed in 50 hospitals to identify the prevalence of and risk factors for communityacquired ESBL-PE bacteraemia. All patients presenting with community-onset Enterobacteriaceae bacteraemia were recorded over a 2-month period (between June and November 2013). Risk factors and 14-day outcomes of patients were investigated. Among 682 Enterobacteriaceae bacteraemia episodes recorded, 58 (8.5%) were caused by ESBL-PE. The most frequent species isolated were Escherichia coli (537; 76.7%) and Klebsiella spp. (68; 9.7%), of which 49 (9.1%) and 8 (11.8%), respectively, were ESBL-producers. Most ESBLPE episodes were healthcare-associated, and only 22 (38%) were apparently community-acquired. The main risk factor for community-acquired ESBL-PE bacteraemia was a prior hospital stay of ≥5 days within the past year. The overall 14-day survival was 90%; only 4 (6.9%) of 58 patients with ESBL-PE bacteraemia died. Inadequate initial antibiotic therapy was administered to 55% of patients with ESBL-PE bacteraemia but was not associated with increased 14-day mortality. Although many patients had community-onset ESBL-PE bacteraemia, almost two-thirds of the episodes were actually healthcare-associated, and true community-acquired ESBL-PE bacteraemia remains rare. In our essentially non-severely ill population, inappropriate initial therapy was not associated with a higher risk of mortality. © 2016 Published by Elsevier B.V.
1. Introduction Infections due to Enterobacteriaceae are the main cause of community and hospital-acquired bacteraemia, often leading to sepsis
* Corresponding author. Medical Intensive Care Unit, Henri Mondor University Hospital, APHP, Paris-Est Créteil University (UPEC), 51 Avenue du Maréchal de Lattre de Tassigny, 94000 Créteil, France. Fax: +33 1 4981 2542. E-mail address: [email protected] (C. Brun-Buisson). 1 Members of the BacterCom Study Group are listed in the Acknowledgements.
and septic shock [1]. In addition, the rising incidence of extendedspectrum β-lactamase-producing Enterobacteriaceae (ESBL-PE) both in hospitals and the community has led to an increased risk of inappropriate antimicrobial therapy and to increasing use of broadspectrum antibiotics [2]. The increasing incidence of ESBL-PE in the community questions the conventional therapeutic approach in patients with severe community-onset infection. A better knowledge of risk factors associated with infection caused by ESBL-PE could help physicians to identify the at-risk population in whom broader-spectrum empirical antimicrobial therapy might be warranted to lessen the risk of inappropriate treatment.
http://dx.doi.org/10.1016/j.ijantimicag.2016.09.032 0924-8579/© 2016 Published by Elsevier B.V.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
2
Indeed, some authors have attempted to develop and validate a tool for identifying patients harbouring ESBL-PE [3–10]. Therapeutic decision-making in these patients may be guided in part by the stratification between healthcare-associated infections (HCAIs) and true community-acquired cases. However, more information of the prevalence of and risk factors for ESBL-PE within the two categories of patients presenting with community-onset bacteraemia would be useful for this purpose. The first aim of this study was to evaluate the current prevalence of apparently community-acquired ESBL-PE infection among patients presenting with community-onset Enterobacteriaceae bacteraemia as well as the impact of ESBL-PE infection on the appropriateness of antimicrobial therapy and 14-day outcome. The second objective was to identify potential risk factors associated with truly community-acquired ESBL-PE bacteraemia. 2. Patients and methods This prospective observational study was conducted between June and November 2013 in 50 voluntary hospitals (49 in France and 1 in Geneva, Switzerland) among hospitals with an infectious diseases team who were contacted by the investigators. In each hospital, all or up to 20 (or 30 in large centres) consecutive adult patients admitted during a 2-month period with community-onset bacteraemia caused by Enterobacteriaceae were included. To exclude patients with hospital-acquired infections, those admitted from longterm care facilities or rehabilitation centres or with a recent (<7 days) history of acute-care hospital admission were not included. Patients with polymicrobial bacteraemia with micro-organisms other than Enterobacteriaceae (n = 48) were secondarily excluded from the analysis of the adequacy of therapy. Community-onset bacteraemia was defined as any positive blood culture drawn within the first 48 h of admission in patients coming directly from home or from a non-medical nursing home. Community-onset bacteraemia episodes were subsequently stratified into healthcare-associated bloodstream infection (HCA-BSI) or community-acquired bloodstream infection (CA-BSI) cases. The former group included patients fulfilling any of the criteria published by Friedman et al [11], i.e. receiving home intravenous (i.v.) therapy or home care, having a recent hospital stay (≥24 h) in the past 90 days, receiving haemodialysis, or having received i.v. chemotherapy within 30 days before occurrence of the bacteraemia episode; patients admitted from a nursing home were also included in this group. True CA-BSIs were defined as those recorded in patients without the above risk factors for HCAI. Data were collected prospectively by the corresponding clinical microbiologist and/or infectious diseases consultant of the antimicrobial therapy team in each hospital. Data were checked retrospectively for consistency and accuracy by two of the co-authors (CB-B and J-RZ). Demographics and clinical data were recorded, including age, sex, birth country and history of travel abroad within the past 3 years, underlying diseases, and co-morbidities [human immunodeficiency virus (HIV) infection, haematological malignancies, solid tumour, diabetes mellitus, solid organ transplant and chronic diseases]. Previous hospitalisation (≥24 h) and antibiotic therapy received within the past 12 months before admission were also recorded, as well as the time elapsed from therapy to the onset of bacteraemia. For statistical analysis, antibiotics were grouped into 11 different classes (see Supplementary material). Known previous ESBL colonisation or infection within the last 6 months was also recorded. Indwelling invasive devices were recorded at admission (urinary tract or intravascular catheters, tracheostomy). The source of bacteraemia was determined according to the US Centers for Disease Control and Prevention (CDC) criteria [12]; when no definite source was identified (including intravascular lines), bacteraemia was qualified as ‘primary’. The severity of underlying
diseases and co-morbidities were assessed using the revised Charlson’s weighted index of morbidity [13], and the severity of acute infection and organ dysfunction was defined as sepsis, severe sepsis or septic shock according to Bone’s criteria [14] as well as by the SOFA (Sepsis-related Organ Failure Assessment) score [15] recorded within 24 h of admission. As there are notable differences in ESBL-PE incidence rates in different regions of France [16], participating hospitals were stratified by their regional incidence according to data from the French antimicrobial surveillance network and this variable was included among potential risk factors for ESBL-PE bacteraemia. The Geneva area was considered to be at similar risk as the South-East France region with regard to ESBL-PE epidemiology [17]. As this study was observational, no specific guidelines were used for empirical treatment. Antimicrobial therapy was considered appropriate if at least one antimicrobial agent with in vitro activity against the isolate had been administered. Nonetheless, therapy combining an active aminoglycoside with another agent to which the strain was resistant was considered inappropriate, except for urinary tract infections (UTIs). The time elapsed (in hours) between the first positive blood sample and the first appropriate antibiotic was assessed, and the appropriateness of antimicrobial treatment administered at 24, 48 and 72 h after hospital admission was analysed according to definitive susceptibility test results. The endpoint was assessed as discharged alive within 14 days of admission; a poor outcome was defined as death within the first 2 weeks of hospitalisation or prolonged hospitalisation beyond Day 14. The study protocol was approved by the Ethics Committee (CPP Ile-de-France IX) in October 2012. Written informed consent was waived according to French law given the non-interventional design of the study, and patients were provided oral information. 2.1. Statistical analysis Results are reported as the median and interquartile range (IQR) or as counts (%). Categorical variables were compared using the χ2 test, and continuous variables were compared using Wilcoxon’s ranksum test for unpaired data. Multivariate analysis of variables associated (P ≤ 0.10) in univariate analysis with ESBL-PE bacteraemia was performed using stepwise backward logistic regression. Odds ratios (ORs) and their 95% confidence intervals (CIs) were calculated. The effect of ESBL-PE infection on outcome at Day 14 was also examined, including all non-redundant variables yielding a P-value of <0.10 in the univariate analysis. Stepwise variable elimination was then performed and ESBL-PE and time to appropriate antimicrobial therapy were forced in the final model. Statistical tests were performed using SAS 9.4 (SAS Institute, Cary, NC) and Stata 13.1 (Stata Corp., College Station, TX). Differences were considered significant when the probability (p) of a type I error was ≤5%. 3. Results Of the 50 hospitals participating in this prospective study, 27 (54%) were teaching hospitals (26 in France and 1 in Geneva) and 23 (46%) were non-teaching hospitals; all French regions were represented. A mean of 14 patients (range, 3–30 patients) was included per participating hospital over the 2-month period of the survey. A total of 693 patients with a single episode of Enterobacteriaceae bacteraemia were recorded; 11 patients were secondarily excluded because of missing information (1 patient), missing susceptibility (1 patient) or hospital-acquired infection (9 patients) (Fig. 1). The remaining 682 patients with community-onset Enterobacteriaceae bacteraemia remained in the analysis, of whom 325 (47.7%) had risk factors for HCAI and 357 (52.3%) had CA-BSI.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
3
Fig. 1. Flow chart of the 693 patients with Enterobacteriaceae bacteraemia included in the study cohort. ESBL-PE, extended-spectrum β-lactamase producing Enterobacteriaceae.
The 682 patients had a median (IQR) age of 74.9 (61–84) years; 352 (51.6%) were female. A total of 226 patients (33.1%) had at least one co-morbid condition; 170 (24.9%) patients had diabetes mellitus, 39 (5.7%) received long-term steroid therapy and 73 (11%) received immunosuppressive therapy, and 82 (12%) had active cancer (Table 1). The vast majority of patients (89%) were admitted directly from home, and 70 (10.3%) were nursing home residents. A total of 133 patients (19.5%) were native from ‘high-risk countries’ as defined above, and most of the patients were French or Swiss residents (96%). A total of 108 patients (15.8%) had travelled abroad within the last year, among which 92% had travelled to a high-risk country. Apart from characteristics defining HCAI, the two subgroups of CA-BSI and HCA-BSI bacteraemia differed in several respects (Table 1): patients having CA-BSI were more often women (55.5% vs. 47.4%; P = 0.03), were less often admitted to a university hospital (54.3% vs. 65.8%; P = 0.002) and were less often in a region with a higher level of ESBL (58.5% vs. 68.3%; P = 0.008) than those with HCA-BSI; more CA-BSI patients had travelled abroad within the past 3 years (18.8% vs 12.6%; P = 0.03), including in high-risk countries (16.8% vs. 10.5%; P = 0.02). Previous hospital admission (for ≥24 h) or administration of an antibiotic course within the past year was recorded in 71 (19.9%) and 85 (23.8%) patients with CA-BSI, respectively; as expected, patients with HCA-BSI had higher corresponding rates at 75.4% and 46.8%, respectively; 10.2% of patients with CABSI and 52.1% of those with HCA-BSI had been hospitalised for >5 days. Of 26 patients (3.8%) previously known as ESBL-PE carriers, 3 (0.8%) and 23 (7.1%) belonged to the CA- and HCA-BSI groups, respectively (Table 1). Eighty-four patients (12.3%) required admission to the intensive care unit within the first 24 h of hospital admission, with a similar rate between the two groups (13.7% vs. 10.8%; P = 0.24).
A source of infection was confirmed in 519 cases (76.1%) and was considered probable on clinical grounds and/or imaging studies in 153 patients (22.4%) (Table 2). Their distribution differed between CA-BSI and HCA-BSI because of a higher proportion of intraabdominal infections and a lower proportion of intravascular infections among CA-BSI patients. Among patients having ESBLPE bacteraemia, however, this distribution did not differ between CA-BSI and HCA-BSI episodes. Bacterial species identified are shown in Supplementary Table S1; 621 episodes (91.1%) were monomicrobial, whereas 61 patients (8.9%) had polymicrobial bacteraemia. Among the latter, 21 (34.4%) blood cultures grew two or three different Enterobacteriaceae. ESBL-producers accounted for 8.7% (61/700) of all Enterobacteriaceae recovered. Of the 58 patients having infection with an ESBL-PE, 36 (62%) had risk factors for HCAI and 22 (38%) were considered truly community-acquired, representing 3.2% of all community-onset episodes and 6.2% of community-acquired Enterobacteriaceae bacteraemia episodes. Compared with other patients having CABSI, patients having infection with ESBL-PE tended to be younger (P = 0.09), were more often hospitalised in a region with a higher mean level of ESBL (P = 0.02) and had been more often previously hospitalised (P = 0.01) within the past year and for longer periods. After multivariate analysis, the only variable significantly associated with ESBL-PE infection among CA-BSI episodes was a previous hospital stay of ≥5 days (P = 0.01), although the regional incidence fell short of statistical significance (P = 0.058). Variables associated with healthcare-associated ESBL-PE bacteraemia are shown in Supplementary Table S2. A total of 614 patients (90.0%) received empirical therapy within the first 24 h of hospital admission and the others received delayed therapy; 373 (60.7%) initially received combination therapy, including 99 (16.1%) receiving three or more antibiotics within the
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
4
Table 1 Clinical characteristics of patients with community-onset Enterobacteriaceae bacteraemia, contrasting healthcare-associated bloodstream infection (HCA-BSI) with truly community-acquired bloodstream infection (CA-BSI). Variable a
Total (N = 682)
HCA-BSI ESBL-PE (n = 36)
Non-ESBL-PE (n = 289)
ESBL-PE (n = 22)
Non-ESBL-PE (n = 335)
Female sex Age (years) [median (IQR)] Native high-risk country Living in low-risk country Travel to high-risk country within 1 year Known ESBL carriage Prior antibiotic therapy within 1 year Prior hospitalisation within 1 year Co-morbid conditions Chronic respiratory disease Chronic cardiac disease Chronic liver disease Haematological disease Chemotherapy treatment Chronic haemodialysis Home care Admission hospital category University Non-university Regional ESBL incidence Low risk High risk Type of admission Medical Surgical Emergency surgery Origin of patient Home Nursing home SOFA score within 24 h of admission [median (IQR)] Charlson score [median (IQR)] Clinical severity of sepsis b Sepsis Severe sepsis Septic shock
352 (51.6) 74.9 (61–84) 133 (19.5) 669 (98.1) 94 (13.8) 26 (3.8) 237 (34.8) 316 (46.3) 226 (33.1) 23 (3.4) 25 (3.7) 62 (9.1) 36 (5.3) 79 (11.6) 11 (1.6) 70 (10.3)
13 (36.1) 79.4 (62–88) 12 (33.3) 35 (97.2) 9 (25.0) 10 (27.8) 28 (77.8) 32 (88.9) 24 (66.7) 0 1 (2.8) 9 (25.0) 3 (8.3) 7 (19.4) 1 (2.8) 10 (27.8)
141 (48.8) 74.8 (64–85) 50 (17.3) 285 (98.6) 25 (8.7) 13 (4.5) 124 (42.9) 213 (73.7) 157 (54.3) 13 (4.5) 13 (4.5) 42 (14.5) 29 (10.0) 72 (24.9) 10 (3.5) 60 (20.8)
13 (59.1) 68.5 (48–82) 5 (22.7) 22 (100) 4 (18.2) 1 (4.5) 8 (36.4) 9 (40.9) 1 (4.5) 0 0 0 0 0 0 0
185 (55.2) 75 (61–83) 66 (19.7) 327 (97.6) 56 (16.7) 2 (0.6) 77 (23.0) 62 (18.5) 44 (13.1) 10 (3.0) 11 (3.3) 11 (3.3) 4 (1.2) 0 0 0
408 (59.8) 274 (40.2)
25 (69.4) 11 (30.6)
189 (65.4) 100 (34.6)
14 (63.6) 8 (36.4)
180 (53.7) 155 (46.3)
251 (36.8) 431 (63.2)
4 (11.1) 32 (88.9)
99 (34.3) 190 (65.7)
4 (18.2) 18 (81.8)
144 (43.0) 191 (57.0)
608 (89.1) 57 (8.4) 17 (2.5)
35 (97.2) 1 (2.8) 0
261 (90.3) 21 (7.3) 7 (2.4)
20 (90.9) 1 (4.5) 1 (4.5)
292 (87.2) 34 (10.1) 9 (2.7)
607 (89.0) 75 (11.0) 1 (0–3) 2 (0–3)
27 (75.0) 9 (25.0) 2 (0.5–4) 3.5 (2–5.5)
223 (77.2) 66 (22.8) 2 (0–4) 2 (1–4)
431 (63.2) 91 (13.3) 49 (7.2)
22 (61.1) 6 (16.7) 0
175 (60.6) 42 (14.5) 20 (6.9)
CA-BSI
22 (100) 0 1 (0–3) 0.5 (0–2)
335 (100) 0 1 (0–3) 0 (0–2)
12 (54.5) 4 (18.2) 3 (13.6)
222 (66.3) 39 (11.6) 26 (7.8)
ESBL-PE, extended-spectrum β-lactamase-producing Enterobacteriaceae; IQR, interquartile range; SOFA, Sepsis-related Organ Failure Assessment. a Values are presented as n (%) unless otherwise stated. b 111 patients (16.3%) did not fulfil criteria for sepsis.
first 24 h. Initial therapy included a third-generation cephalosporin in 497 cases (72.9%), amoxicillin or amoxicillin/clavulanic acid in 69 cases (10.1%) and a fluoroquinolone in 39 cases (5.7%). Overall, an appropriate empirical regimen was administered within the first 24 h of bacteraemia in 534 (85.6%) of 624 patients with non-ESBLPE bacteraemia and 26 (44.8%) of the 58 patients with ESBL-PE bacteraemia (Table 3). The median (IQR) time elapsed between drawing the first blood culture and the first administration of adequate antibiotic therapy was 0.1 (0–0.8) days overall and, respectively, 0.1 (0–0.5) days and 1.7 (0.7–3.2) days for non-ESBLPE and ESBL-PE bacteraemia (P < 0.001). Among the 22 patients having community-acquired ESBL-PE bacteraemia, 50%, 77% and 82%
received an adequate antibiotic regimen within the first 24, 48 and 72 h, respectively. The overall 14-day mortality in patients with communityonset bacteraemia was 9.8%. As expected, patients having HCABSI had a poorer outcome than those with CA-BSI, with a higher 14-day mortality (P = 0.036), poorer outcome at Day 14 (P = 0.004) and longer hospital stay (P = 0.001). Among those with CA-BSI, patients having infection with ESBL-PE had a comparable outcome to those without ESBL-PE infection. By multivariable analysis, only age and clinical severity (i.e. SOFA score), but not ESBL-PE infection (P = 0.54) or time to adequate antibiotic therapy (P = 0.46), were associated with death at 14 days. However, a time to adequate therapy
Table 2 Source of infection identified in patients with community-onset Enterobacteriaceae bacteraemia. Source of infection a
Urinary tract Respiratory tract Intra-abdominal infection Intravascular infection Others b None identified (primary bacteraemia)
Total (N = 682)
CA-BSI ESBL-PE (n = 22)
Non-ESBL-PE (n = 335)
HCA-BSI ESBL-PE (n = 36)
Non-ESBL-PE (n = 289)
383 (56.2%) 27 (4.0%) 200 (29.3%) 21 (3.1%) 19 (2.8%) 37 (5.4%)
13 0 4 0 1 3
196 11 113 1 7 12
21 0 9 0 0 3
153 12 74 20 8 19
CA-BSI, community-acquired bloodstream infection; HCA-BSI, healthcare-associated bloodstream infection; ESBL-PE, extended-spectrum β-lactamase-producing Enterobacteriaceae. a The source was considered confirmed in 519 patients (when identical micro-organisms were recovered from the infectious focus and blood) and presumed in the remaining 153; 15 patients had multiple infection sites and 10 patients had missing information on the source of infection (3 CA-BSI and 7 HCA-BSI). b Including soft-tissue infection, meningitis, osteomyelitis and arthritis.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
Table 3 Timing of adequacy of antibiotic therapy in 682 patients with community-onset bacteraemia, according to extended-spectrum β-lactamase (ESBL) status. Time to adequate antibiotic therapy Prior to hospital admission In hospital ≤24 h >24–48 h >48–72 h >72 h Total
Non-ESBL-PE bacteraemia [n (%)] 3 (0.5) 534 (85.6) 57 (9.1) 16 (2.6) 14 (2.2) 624 (100)
ESBL-PE bacteraemia [n (%)] 0 26 (44.8) 16 (27.6) 6 (10.3) 10 (17.2) 58 (100)
Total evaluable bacteraemia 3 (0.6) 560 (82.1) 73 (10.7) 22 (3.2) 24 (3.5) 682 (100)
of >24 h was marginally associated with a prolonged hospital stay (OR = 1.83, 95% CI 0.97–3.48; P = 0.06). 4. Discussion In this large prospective observational cohort study, the prevalence of ESBL-PE among patients admitted with communityonset bacteraemia reached 8% of all episodes. However, twothirds of episodes actually qualified as HCAI, and true communityacquired ESBL-PE bacteraemia was identified in only 22 patients (3.2%). Community-acquired ESBL-PE bacteraemia was associated with a history of prior hospital stay of ≥5 days and with regional ESBL incidence. Despite a much higher rate of inadequate antimicrobial therapy than in other patients, patients with ESBL-PE bacteraemia had a comparable outcome to other patients. Recent epidemiological data suggest increasing rates of ESBLPE invasive infections throughout European countries [18]. In a previous study conducted 3 years before ours in 10 French hospitals [19], ESBL-PE accounted for 3.8% of all Enterobacteriaceae bacteraemia diagnosed within the first 48 h of hospital admission, and most ESBL-PE bacteraemias were healthcare-associated. The current study suggests an increased rate of community-onset ESBL-PE bacteraemia. Despite ESBL-PE spreading in the community in France [20], the prevalence of truly community-acquired ESBLPE invasive infection however appears to remain low. Indeed, twothirds of the patients in the current study had criteria for healthcareassociated ESBL-PE bacteraemia according to Friedman et al [11]. Moreover, the major risk factor for community-acquired ESBL-PE infection appeared to be a prior hospital stay of ≥5 days in the previous year, suggesting that many of these community-onset cases without the accepted risk factors for HCAI could be in fact acquired in the hospital. Indeed, carriage of ESBL-PE is known to persist for months in some patients [21]. Patients having had a prior hospitalisation dating back more than 3 months can be a persistent carrier if colonisation occurred during that hospitalisation. The current results suggest that, beyond the conventional criteria for HCAI, patients having had a prior hospitalisation within the preceding year should be considered at risk for ESBL-PE infection. Applying this rule to our patients would have resulted in considering at risk 44 (75.9%) of all patients with community-onset ESBLPE infection. Together with the potential influence of regional ESBL incidence, these results suggest a major driving role for healthcareassociated factors in the occurrence of community-onset ESBL-PE infections. The low rate of truly community-acquired infection in this study can be explained in light of recent data from the literature. Indeed, only 10–20% of carriers appear to be exposed to a risk of secondary infection [22,23]. In a multicentre prospective study conducted in ESBL-producing Escherichia coli carriers, Ruppé et al suggested that the occurrence of UTI in previously known carriers was related to the relative faecal abundance of ESBL-PE induced by antibiotic therapy [24]. Likewise, in a retrospective case–control study where
5
controls were selected on the basis of age, hospital location and exposure time, the only two factors associated with infection in previously colonised patients were urinary catheterisation and therapy with a β-lactam/β-lactamase inhibitor combination prior to infection. Thus, aside from healthcare-associated risk factors, antibiotic exposure also appears to play a major role in the occurrence of ESBL-PE infection [22], with cephalosporin and fluoroquinolone classes most commonly implicated [25]. In the current study, previous antibiotic therapy was not significantly associated with community-acquired ESBL-PE infection, although investigation into the various antibiotic classes and time elapsed since receipt of antibiotics suggested that previous therapy with antibiotics having antianaerobic activity was consistently associated with a higher risk of community onset ESBL-PE bacteraemia (Supplementary Table S3). These findings are consistent with animal models [26] and clinical studies suggesting that preservation of the anaerobic flora may be critical for preventing the acquisition of multidrug-resistant microorganisms and subsequent infection [27]. The rate of inadequate first-line antibiotic therapy in this study underlines the clinical importance of prediction rules to help clinicians identify those patients who should be treated with a broadspectrum antibiotic. A number of studies have demonstrated that patients with healthcare-associated Gram-negative BSI experience delay in receipt of appropriate antibiotic therapy [28]. Several studies attempted to identify risk factors associated with ESBL-PE community-onset infection [3–10]. However, many methodological issues limit the interpretation of these studies. Indeed, many have focused on specific species, mostly on E. coli [4,5,7–9]. Tumbarello et al [10] attempted to establish a predictive clinical score to identify patients having ESBL-PE recovered in clinical samples from a retrospective chart review of 849 adult patients admitted to a large hospital in Rome (Italy). The score showed good discrimination; however, this study combined nosocomial and community-acquired infections. Other studies sought to identify risk factors associated with carriage [29] or infection with a specific species such as ESBLproducing Enterobacter spp. [30] or ESBL-producing E. coli [31] or at a specific site of infection, mostly the urinary tract [32]. In the current study, no association between the appropriateness of antimicrobial therapy and mortality was found, irrespective of the time elapsed between admission and the first appropriate dose of antibiotic. However, inappropriate empirical antibiotic therapy has been shown to increase mortality essentially in critical ill patients with severe infection [33]. The results of the current study could thus be explained, at least in part, by the expected low proportion of patients having severe sepsis or septic shock in this hospital-based study; they are also consistent with some conflicting results with regard to the association between inappropriate initial therapy of ESBL-PE infection and mortality, especially for UTIs [34,35]. Nevertheless, delayed adequate therapy tended to be associated with prolonged hospital stay. This study has several limitations. First, it was conducted in France and Geneva and the results may not be applicable to other countries with a higher or lower prevalence of ESBL. However, the prevalence of ESBL-PE carriage in non-hospitalised subjects is estimated at 6% in France [20], which is similar to reported prevalences from other Western European countries such as Germany [36] and The Netherlands [37]. Second, the relatively low number of patients with community-acquired bacteraemia reduced the ability to identify potential risk factors for ESBL-PE infection. In conclusion, despite the increased carriage rate of ESBL-PE in the community, community-acquired ESBL-PE bacteraemia remains rare in France and Geneva, although patients may remain at risk for ESBL-PE infection after a prolonged hospitalisation, beyond the conventional 90 days included in the definition of HCAI. In our general hospital population, inadequate first-line antibiotic therapy was common but was not associated with a higher risk of death.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
6
These results thus do not suggest altering the conventional approach to therapy in most instances and support a highly selective use of broad-spectrum antibiotic therapy in non-severely ill patients with community-acquired Gram-negative bacteraemia.
Acknowledgements The BacterCom Study Group: Steering committee: Christian Brun-Buisson, Jean Carlet, Pierre Dellamonica, Rémy Gauzit, Christian Rabaud. Scientific committee: Jean-Ralph Zahar, Jean-François Timsit, Serge Alfandari, Robert Cohen, Bruno Coignard, Céline Pulcini, Jérôme Robert. Investigators: Nejla Aissa, Alain Losniewski, Sandrine Henard, Christian Rabaud (CHU Nancy); Cécile Janssen, Léonardo Pagani, Stéphane Bland, Jacques Gaillat, Jean Pierre Bru, Pierre Bonnin, Gaelle Clavère, Virginie Vitrat (CHG Annecy); Laurence Maulin (CHG Aix); Luc Quarsaet, Helene Roger, Jean-Baptiste Euzen (CHU Brest); Eric Denes, Elodie Couve-Daecon (CHU Limoges); Catherine Lechiche, Jean-Philippe Lavigne (CHU Nimes); Véronique Mondain, Celine Pulcini (CHU Nice); Veronique Garnier, Dorothée Lambert, Maxime Hentzien, Yohan Nguyen (CHU Reims); Jean Paul Talarmin, Pascaline Rameau, Françoise Geffroy (CH Intercommunal de Cornouaille); Arnaud Salmon-Rousseau, Vincent Cattoir (CHU Côte de Nacre); Antoine Froissart, Said Aberrane (CH Intercommunal de Créteil); Laurence Legout (CH Alpes-Leman); France Roblot, Guillaume Beraud (CHU Poitiers); Patricia Pavese (CHU Grenoble): Olivier Belmonte, Nicolas Traversier, Laurent Bellec, Marie Lagrange-Xelot, MariePierre Moiton, Barbara Kuli (CHU la Réunion); Pascal Longuet, Agnes Scanvic (CH Argenteuil); Lea Colombain, Milagros Ferreyra, Maxime Jean, Aurelia Eden, Etienne Laurens, Matthieu Saada, Hugues Aumaitre, Martine Malet, Frederique Roustant (CH Perpignan); Vincent LeMoing, Jean-Pierre Helene (CHU Montpellier); Jonathan Theodore, Caroline Garandeau (CH Angoulême); Jocelyne Caillon, Benjamin Olivero, David Boutoille, Didier Lepelletier (CHU Nantes); Lucia Grandière-Perez, Aurélie Beaudron, Céline Ramanantsoa, Catherine Varache, Nicolas Varache, Nicolas Crochette, Hikombo Hitoto, Sophie Blanchi, Pascale Penn (CH Le Mans); Bernard Louis (CHU Tours); Dominique Vanjak, Marie-Christine Escande (Institut Curie); René-Gilles Patrigeon, Stéphanie Honoré (CH Auxerre); Audrey Barrelet, Marlène Amara (CH Versailles); Jean-Marc Galempoix, JeanClaude Réveil, Patricia Bucchiotty (CH Charleville-Mézières); Carine Couzigou, Alban LeMonnier (CH St Joseph); Emilie Boidin, Stephane Sire (CH Cahors); Nelly Luizy, Caroline Bronstain (CHG Montfermeil); Stephan Harbarth, Carolina Fankhauser, Caroline Brossier, Elodie von Dach (HU Genève); Colette Sauve (CH Longjumeau); Colette Sauve, Marie-Noelle Adam (CH Orsay); Laurent Hocqueloux, Laurent Bret (CH Orléans); Marie-Hélène André, Régine Barruet, Marion Challier (CH Montreuil); Andre Cabié, Benoit Roze (CHU Martinique); Patrick Plesiat, Katy Jeanot, Isabelle Patry, Joséphine Moreau, Catherine Chirouze (CHU Besançon); Aaron Laurent, Marie-Nadège Bachelier (CH Bourges); Guillaume Kac, Nicolas Farthouat, Nicolas Feuillebois, Thomas Guimard, Sandra Bourdon (CHD Vendée); Sophie Nguyen, Serge Alfandari (CHU Tourcoing); Houcine Bentayeb, Mari-Thérèse Albertini (CH St Quentin); Bruno Marchou, Noémie Gaudré, Damien Dubois (CHU Toulouse, Assistance Publique-Hôpitaux de Paris); Christelle Guillet-Caruba, Florence Doucet Populaire (Antoine Béclère); Nicolas Fortineau, Leila Escaut (Kremlin-Bicêtre); Philippe Lesprit, Vincent Fihman (Henri Mondor); Rémy Gauzit, Josette Raymond, Solene Kerneis (Cochin); Veronique Leflon-Guibout, Agnes Lefort (Beaujon); Aurelien Dinh, Anne-Laure Roux (Ambroise Paré); Sophie Abgrall, Françoise Jauréguy (Avicenne); Jean-Christophe Lucet (Bichat); Aurelien Dinh, Nathalie Soismier (Raymond Poincaré); Amanda Lopes, Emmanuelle Cambau (Lariboisière); Veronique Manceron, Guilène Barnaud (Louis Mourier).
Supporting institutions: The BacterCom study was supported by the Institut Maurice Rapin, the Société française de microbiologie (SFM), the Société de pathologie infectieuse de langue française (SPILF) and the association AC-de-BMR. Funding: This study was funded in part by the Institut Maurice Rapin and the Société de pathologie infectieuse de langue française (SPILF). Competing interests: J-RZ participates on an advisory board for Merck Sharp & Dohme; SH has participated on advisory boards and has received honoraria from bioMérieux, Novartis, Johnson & Johnson and Da Volterra. All other authors declare no competing interests. Ethical approval: The study protocol was approved by the Ethics Committee (CPP Ile-de-France IX) in October 2012. Written informed consent was waived according to French law. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ijantimicag.2016.09.032. References [1] Laupland KB, Church DL. Population-based epidemiology and microbiology of community-onset bloodstream infections. Clin Microbiol Rev 2014;27:647–64. [2] Zilberberg MD, Shorr AF, Micek ST, Vazquez-Guillamet C, Kollef MH. Multi-drug resistance, inappropriate initial antibiotic therapy and mortality in Gramnegative severe sepsis and septic shock: a retrospective cohort study. Crit Care 2014;18:596. [3] Ben-Ami R, Rodriguez-Bano J, Arslan H, Pitout JD, Quentin C, Calbo ES, et al. A multinational survey of risk factors for infection with extended-spectrum β-lactamase-producing Enterobacteriaceae in nonhospitalized patients. Clin Infect Dis 2009;49:682–90. [4] Han JH, Kasahara K, Edelstein PH, Bilker WB, Lautenbach E. Risk factors for infection or colonization with CTX-M extended-spectrum-β-lactamase-positive Escherichia coli. Antimicrob Agents Chemother 2012;56:5575–80. [5] Kang CI, Wi YM, Lee MY, Ko KS, Chung DR, Peck KR, et al. Epidemiology and risk factors of community onset infections caused by extended-spectrum β-lactamase-producing Escherichia coli strains. J Clin Microbiol 2012;50:312–17. [6] Kang CI, Wi YM, Ko KS, Chung DR, Peck KR, Lee NY, et al. Outcomes and risk factors for mortality in community-onset bacteremia caused by extendedspectrum β-lactamase-producing Escherichia coli, with a special emphasis on antimicrobial therapy. Scand J Infect Dis 2013;45:519–25. [7] Kuster SP, Hasse B, Huebner V, Bansal V, Zbinden R, Ruef C, et al. Risks factors for infections with extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella pneumoniae at a tertiary care university hospital in Switzerland. Infection 2010;38:33–40. [8] Rodríguez-Baño J, Navarro MD, Romero L, Muniain MA, de Cueto M, Ríos MJ, et al. Bacteremia due to extended-spectrum β-lactamase-producing Escherichia coli in the CTX-M era: a new clinical challenge. Clin Infect Dis 2006;43:1407–14. [9] Rodriguez-Bano J, Picon E, Gijon P, Hernandez JR, Ruiz M, Pena C, et al. Community-onset bacteremia due to extended-spectrum β-lactamaseproducing Escherichia coli: risk factors and prognosis. Clin Infect Dis 2010;50:40–8. [10] Tumbarello M, Trecarichi EM, Bassetti M, De Rosa FG, Spanu T, Di Meco E, et al. Identifying patients harboring extended-spectrum-β-lactamase-producing Enterobacteriaceae on hospital admission: derivation and validation of a scoring system. Antimicrob Agents Chemother 2011;55:3485–90. [11] Friedman ND, Kaye KS, Stout JE, McGarry SA, Trivette SL, Briggs JP, et al. Health care-associated bloodstream infections in adults: a reason to change the accepted definition of community-acquired infections. Ann Intern Med 2002;137:791–7. [12] Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309–32. [13] Quan H, Li B, Couris CM, Fushimi K, Graham P, Hider P, et al. Updating and validating the Charlson comorbidity index and score for risk adjustment in hospital discharge abstracts using data from 6 countries. Am J Epidemiol 2011;173:676–82. [14] Bone RC, Balk RA, Cerra FB, Dellinger EP, Fein AM, Knaus WA, et al. Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis. Chest 1992;101:1656–62. [15] Vincent JL, Moreno R, Takala J, Willatts S, De Mendonca A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure. On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine. Intensive Care Med 1996;22:707–10. [16] Carbonne A, Arnaud I, Maugat S, Marty N, Dumartin C, Bertrand X, et al. National multidrug-resistant bacteria (MDRB) surveillance in France through the RAISIN network: a 9 year experience. J Antimicrob Chemother 2013;68:954–9.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032
ARTICLE IN PRESS J.-R. Zahar et al. / International Journal of Antimicrobial Agents ■■ (2016) ■■–■■
[17] Pasricha J, Koessler T, Harbarth S, Schrenzel J, Camus V, Cohen G, et al. Carriage of extended-spectrum β-lactamase-producing Enterobacteriaceae among internal medicine patients in Switzerland. Antimicrob Resist Infect Control 2013;2:20. [18] European Centre for Disease Prevention and Control (ECDC). Antimicrobial resistance surveillance in Europe. Annual report of the European Antimicrobial Resistance Surveillance Network (EARS-Net) 2013. Stockholm, Sweden: ECDC; 2014. Available from: http://ecdc.europa.eu/en/publications/Publications/ antimicrobial-resistance-surveillance-europe-2013.pdf. [Accessed 7 November 2016]. [19] Shoai Tehrani M, Hajage D, Fihman V, Tankovic J, Cau S, Day N, et al. Gramnegative bacteremia: which empirical antibiotic therapy? Med Mal Infect 2014;44:159–66. [20] Nicolas-Chanoine MH, Gruson C, Bialek-Davenet S, Bertrand X, Thomas-Jean F, Bert F, et al. 10-Fold increase (2006–11) in the rate of healthy subjects with extended-spectrum β-lactamase-producing Escherichia coli faecal carriage in a Parisian check-up centre. J Antimicrob Chemother 2012;68:562–8. [21] Titelman E, Hasan CM, Iversen A, Naucler P, Kais M, Kalin M, et al. Faecal carriage of extended-spectrum β-lactamase-producing Enterobacteriaceae is common 12 months after infection and is related to strain factors. Clin Microbiol Infect 2014;20:O508–15. [22] Goulenok T, Ferroni A, Bille E, Lecuyer H, Join-Lambert O, Descamps P, et al. Risk factors for developing ESBL E. coli: can clinicians predict infection in patients with prior colonization? J Hosp Infect 2013;84:294–9. [23] Reddy P, Malczynski M, Obias A, Reiner S, Jin N, Huang J, et al. Screening for extended-spectrum β-lactamase-producing Enterobacteriaceae among high-risk patients and rates of subsequent bacteremia. Clin Infect Dis 2007;45:846–52. [24] Ruppé E, Lixandru B, Cojocaru R, Buke C, Paramythiotou E, Angebault C, et al. Relative fecal abundance of extended-spectrum-β-lactamase-producing Escherichia coli strains and their occurrence in urinary tract infections in women. Antimicrob Agents Chemother 2013;57:4512–17. [25] Biehl LM, Schmidt-Hieber M, Liss B, Cornely OA, Vehreschild MJ. Colonization and infection with extended spectrum β-lactamase producing Enterobacteriaceae in high-risk patients—review of the literature from a clinical perspective. Crit Rev Microbiol 2016;42:1–16. [26] Bhalla A, Pultz NJ, Ray AJ, Hoyen CK, Eckstein EC, Donskey CJ. Antianaerobic antibiotic therapy promotes overgrowth of antibiotic-resistant, Gram-negative bacilli and vancomycin-resistant enterococci in the stool of colonized patients. Infect Control Hosp Epidemiol 2003;24:644–9.
7
[27] Donskey CJ. The role of the intestinal tract as a reservoir and source for transmission of nosocomial pathogens. Clin Infect Dis 2004;39:219–26. [28] Moehring RW, Sloane R, Chen LF, Smathers EC, Schmader KE, Fowler VG Jr, et al. Delays in appropriate antibiotic therapy for Gram-negative bloodstream infections: a multicenter, community hospital study. PLoS ONE 2013;8:e76225. [29] Leistner R, Gurntke S, Sakellariou C, Denkel LA, Bloch A, Gastmeier P, et al. Bloodstream infection due to extended-spectrum β-lactamase (ESBL)-positive K. pneumoniae and E. coli: an analysis of the disease burden in a large cohort. Infection 2014;42:991–7. [30] Qureshi ZA, Paterson DL, Pakstis DL, Adams-Haduch JM, Sandkovsky G, Sordillo E, et al. Risk factors and outcome of extended-spectrum β-lactamase-producing Enterobacter cloacae bloodstream infections. Int J Antimicrob Agents 2011;37:26–32. [31] Rodriguez-Bano J, Alcala JC, Cisneros JM, Grill F, Oliver A, Horcajada JP, et al. Community infections caused by extended-spectrum β-lactamase-producing Escherichia coli. Arch Intern Med 2008;168:1897–902. [32] Calbo E, Romani V, Xercavins M, Gomez L, Vidal CG, Quintana S, et al. Risk factors for community-onset urinary tract infections due to Escherichia coli harbouring extended-spectrum β-lactamases. J Antimicrob Chemother 2006;57:780–3. [33] Zahar JR, Timsit JF, Garrouste-Orgeas M, Francais A, Vesin A, Descorps-Declere A, et al. Outcomes in severe sepsis and patients with septic shock: pathogen species and infection sites are not associated with mortality. Crit Care Med 2011;39:1886–95. [34] Peralta G, Lamelo M, Alvarez-Garcia P, Velasco M, Delgado A, Horcajada JP, et al. Impact of empirical treatment in extended-spectrum β-lactamase-producing Escherichia coli and Klebsiella spp. bacteremia. A multicentric cohort study. BMC Infect Dis 2012;12:245. [35] Rottier WC, Ammerlaan HS, Bonten MJ. Effects of confounders and intermediates on the association of bacteraemia caused by extended-spectrum β-lactamaseproducing Enterobacteriaceae and patient outcome: a meta-analysis. J Antimicrob Chemother 2012;67:1311–20. [36] Valenza G, Nickel S, Pfeifer Y, Eller C, Krupa E, Lehner-Reindl V, et al. Extendedspectrum-β-lactamase-producing Escherichia coli as intestinal colonizers in the German community. Antimicrob Agents Chemother 2014;58:1228– 30. [37] Huijbers PM, de Kraker M, Graat EA, van Hoek AH, van Santen MG, de Jong MC, et al. Prevalence of extended-spectrum β-lactamase-producing Enterobacteriaceae in humans living in municipalities with high and low broiler density. Clin Microbiol Infect 2013;19:E256–9.
Please cite this article in press as: Jean-Ralph Zahar, et al. for the BacterCom Study Group, Predominance of healthcare-associated cases among episodes of community-onset bacteraemia due to extended-spectrum β-lactamase-producing Enterobacteriaceae, International Journal of Antimicrobial Agents (2016), doi: 10.1016/j.ijantimicag.2016.09.032