Clinical characteristics of patients with community-acquired complicated intra-abdominal infections: A prospective, multicentre, observational study

International Journal of Antimicrobial Agents 44 (2014) 222–228

Contents lists available at ScienceDirect

International Journal of Antimicrobial Agents journal homepage: http://www.elsevier.com/locate/ijantimicag

Clinical characteristics of patients with community-acquired complicated intra-abdominal infections: A prospective, multicentre, observational study Shio-Shin Jean a , Wen-Chien Ko b , Yang Xie c , Vaishali Pawar d , Dongmu Zhang c , Girish Prajapati e , Myrna Mendoza f , Pattarachai Kiratisin g , Elmano Ramalheira h , Ana Paula Castro i , Fernando Rosso j , Po-Ren Hsueh k,∗ a

Department of Emergency Medicine, Wan Fang Hospital, Taipei Medical University, Taipei, Taiwan Department of Medicine at National Cheng Kung University Medical College and Hospital, Tainan, Taiwan c Merck & Co, Inc., Global Health Outcomes, Whitehouse Station, NJ, USA d Rutgers University, Ernest Mario School of Pharmacy, Piscataway, NJ, USA e All Source PPS, Huntington Beach, CA, USA f Section of Infectious Diseases, Department of Medicine, University of the Philippines, Philippine General Hospital, Taft, Manila, Philippines g Department of Microbiology, Faculty of Medicine Siriraj Hospital, Bangkok, Thailand h Clinical Pathology Department, Hospital Infante Dom Pedro – Aveiro, Aveiro, Portugal i Clinical Pathology Department/Microbiology Service, Hospital Geral de Santo António, Porto, Portugal j Infectious Diseases, Fundacion Clinica Valle del Lili, Cali, Colombia k Departments of Laboratory Medicine and Internal Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, No. 7 Chung-Shan South Road, Taipei 100, Taiwan b

a r t i c l e

i n f o

Article history: Received 2 March 2014 Accepted 20 May 2014 Keywords: Community-acquired complicated intra-abdominal infection Extended-spectrum ␤-lactamase ESBL Length of stay Outcomes

a b s t r a c t In this prospective, observational, multicentre study using data from five countries (Columbia, The Philippines, Portugal, Taiwan and Thailand), the clinical impact of extended-spectrum ␤-lactamase (ESBL)-producing organisms on hospitalised patients with community-acquired complicated intraabdominal infections (CA-cIAIs) was compared with that of non-ESBL-producing organisms during the period April 2010 to December 2011. Adult patients (aged ≥18 years) requiring surgery or percutaneous drainage were enrolled and were followed during the first hospitalisation course. An unadjusted statistical comparison of risk factors for ESBL-positive and ESBL-negative patients was performed. Multivariate regression analyses were performed to assess whether length of stay (LOS) in hospital, clinical cure rate and some important clinical characteristics were associated with ESBL positivity. During the study period, a total of 105 adult patients from five countries were enrolled, of whom 17 (16.2%) had CA-cIAI due to ESBL-positive organisms and 88 (83.8%) had CA-cIAI due to ESBL-negative organisms. Escherichia coli was isolated in 73.3% of all samples. Infections were cured in 8 (47.1%) of the patients with CA-cIAI due to ESBL-positive organisms and in 59 (67.0%) of the patients with CA-cIAI due to ESBL-negative organisms (P = 0.285). The median LOS was 11.6 days for patients with infections due to ESBL-negative organisms and 17.6 days for patients with infections due to ESBL-positive organisms (P = 0.011). Multivariate logistic regression analysis revealed that pre-existing co-morbidities, but not ESBL positivity, were adversely associated with clinical cure of CA-cIAIs. In contrast, duration of hospitalisation was longer for patients with CA-cIAI due to ESBL-positive organisms. © 2014 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

1. Introduction Intra-abdominal infections (IAIs) encompass a broad group of infections and are primarily classified as uncomplicated or

∗ Corresponding author. Tel.: +886 2 2312 3456x5355; fax: +886 2 2322 4263. E-mail address: [email protected] (P.-R. Hsueh).

complicated according to severity. In uncomplicated IAIs, only a single organ is involved and the infection does not spread to the peritoneum. These patients are treated either conservatively with antibiotics or with surgery. Complicated IAIs (cIAIs), however, extend beyond a single affected organ and result in either localised peritonitis (intra-abdominal abscess) or diffuse peritonitis. cIAIs often cause significant morbidity and are frequently associated with poor prognosis, especially in patients with certain risk factors.

http://dx.doi.org/10.1016/j.ijantimicag.2014.05.018 0924-8579/© 2014 Elsevier B.V. and the International Society of Chemotherapy. All rights reserved.

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

IAIs can be either community-acquired (CA) or hospital-acquired [1]. The major pathogens involved in CA-IAIs are Enterobacteriaceae spp. and anaerobic microbes (especially Bacteroides fragilis). However, the emergence of multidrug-resistant organisms such as vancomycin-resistant enterococci, meticillin-resistant Staphylococcus aureus (MRSA) and, more recently, extended-spectrum ␤-lactamase (ESBL)-producing Enterobacteriaceae has raised considerable concern regarding the appropriate treatment of IAIs. In a review of the literature, Paterson and Bonomo found that 5–8% of Enterobacteriaceae spp. isolated from patients in South Korea, Japan, Malaysia and Singapore were ESBL producers, whereas in Thailand, Taiwan and The Philippines the isolation rate of ESBL-producing Enterobacteriaceae spp. ranged from 12% to 24% [2]. The Study for Monitoring Antimicrobial Resistance Trends (SMART) revealed a gradually increasing trend in the worldwide prevalence of ESBL-producing organisms amongst Enterobacteriaceae spp. [3–5]. In addition, Chen et al. found that the rates of antimicrobial resistance as well as the prevalence of ESBL-producing intra-abdominal Gram-negative bacilli (GNB) were alarmingly high in the Asia-Pacific region, with the highest rates being found in India (57–67%), China (32–65%) and Thailand (48–56%) [6]. Although ESBL-producing members of the Enterobacteriaceae family are mainly isolated from the hospital setting, ESBL-producing organisms have been isolated with increasing frequency from community settings [7]. According to the preliminary report from the Complicated Intra-Abdominal Infections Observational European study (CIAO Study), ESBL-producing Escherichia coli accounted for 8.1% of all E. coli isolates, and 19.3% of all Klebsiella pneumoniae isolates were ESBL producers [8]. In the CIAO study, Enterobacteriaceae spp. still accounted for the majority of causative organisms (86%), of which most were E. coli (48%), followed by Klebsiella spp. (16%) [8]. Krobot et al. demonstrated that use of appropriate antibiotics for the treatment of CA-cIAIs was significantly associated with clinical success, which was defined as fewer days of using parenteral therapy and shorter length of hospital stay [9]. In contrast, inappropriate selection of antimicrobial agents frequently led to treatment failure, which was defined as re-operation, the necessity of switching to second-line antibiotics, longer hospitalisation and greater hospital-associated costs [10]. Although many previous studies have evaluated the acquisition risk factors, bacteriologic profiles and clinical outcomes of patients infected with hospitalacquired ESBL-producing organisms [11–13], very few studies have explored the risk factors, antimicrobial resistance profiles and outcomes of patients with ESBL-producing organisms isolated from IAIs of community origin [14]. In this prospective, observational, multicentre study using data from five countries, the clinical impact of ESBL-producing organisms on hospitalised patients with CA-cIAIs was compared with that of non-ESBL-producing organisms during the period April 2010 to December 2011.

2. Patients and methods 2.1. Study subjects and design This prospective, observational study of patients with CAcIAIs was conducted at multiple sites in five countries, including Thailand, Taiwan, The Philippines, Portugal and Columbia. Patients were recruited from sites that contributed isolates to the 2010 SMART surveillance programme. A total of 127 patients with cIAIs were enrolled from April 2010 through December 2011. Adult patients (aged ≥18 years) who required surgery or percutaneous drainage were included. All patients received initial empirical antibiotic therapy, and their clinical responses were recorded

223

in detail until discharge from hospital or until death. Patients with traumatic intestinal perforations (i.e. stomach or duodenum) requiring surgery within 12 h of hospital arrival, those with primary peritonitis or cholecystitis with inflammation confined to the gallbladder, non-perforated appendicitis without peri-appendicular abscess, and acute infected pancreatitis were excluded from the study. Patients with cIAIs underwent timely percutaneous drainage or surgery after evaluation by a general surgeon. Isolates collected during surgery or drainage were sent to SMART central laboratories to determine the antimicrobial susceptibility profiles and their ESBL status. Only patients with CA-cIAIs caused by E. coli or Klebsiella spp. were enrolled into the study. All patients were followed through their first hospitalisation course to assess their baseline characteristics, clinical responses and appropriateness of empirical antibiotic regimens. The Charlson Comorbidity Index (CCI) was used to determine the underlying co-morbid conditions in the patients. The CCI assigns weights for a number of major conditions present amongst secondary diagnoses. The index score is the sum of assigned weights and represents a measure of the burden of co-morbid disease. A baseline score of 0 indicates no co-morbidity, and scores of 1, 2 and ≥3 were assigned to stratify patients based upon their co-morbid conditions. 2.2. Data collection Data were collected prospectively by the investigators at each site and were reported using an electronic case report form. All investigators were blinded to the patients’ respective clinical and laboratory data. In this study, the following data were collected: demographic profile; co-morbid conditions; infectious origin; causative micro-organism; prior use of antibiotic therapy; urinary or venous catheterisation within 3 months prior to this hospitalisation course; and response to clinical management. Clinical response was recorded as (i) cure, which represented complete resolution of the symptoms/signs after surgery and antimicrobial therapy, (ii) improvement but not cure, which was defined as partial resolution of the infection with therapy, and (iii) failure, which was defined as no improvement in symptoms or signs, the need for surgical re-intervention, or death attributable to cIAI. The cause of death was recorded as death attributable to cIAI, or death not related to cIAI. The hospital length of stay (LOS) was calculated from the date of admission to the date of death or discharge. The appropriateness of empirical antimicrobial regimens to specific causative organisms was also evaluated. Adjustment of antibiotic therapy was done immediately after the antibiotic susceptibility testing results were available regardless of the degree of clinical severity. 2.3. Species identification Isolates collected within 48 h after admission were presumptively categorised as CA-IAIs, whereas those collected >48 h after admission were classified as hospital-acquired infections. Bacterial species were identified at each hospital of admission; however, resistance data and ESBL status were determined at the central SMART laboratory. Traditional ESBL confirmation methods employing a disc containing 30 ␮g of ceftazidime (or cefotaxime) alone or in combination with a 10 ␮g of clavulanic acid disc are known to have low sensitivity at detecting bacteria that potentially produce AmpC ␤-lactamase (E. coli and K. pneumoniae) [15]. Thus, in this survey, in addition to the standard method using cefotaxime/ceftazidime with or without clavulanic acid by potentiation of discs, the modified double-disc synergy test (DDST) was also applied, which involves a disc containing 30 ␮g of cefepime with or without a disc containing 10 ␮g of clavulanic acid (at a centreto-centre distance of 30 mm) to further confirm ESBL production.

224

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

Production of ESBL in the modified DDST was considered positive if the diameter of the cefepime disc increased by ≥5 mm or if the zone expansion (diameter) was ≥50% of the original size, as suggested by M’Zali et al. [16]. The criteria for pathological identification of cIAIs from surgical specimens unavoidably varied between different hospitals in different countries. In contrast, the criteria for microbiological identification in all hospitals were in accordance with the protocol of the SMART programme [4,5]. SMART is a global longitudinal surveillance programme designed to monitor the in vitro antimicrobial susceptibility of aerobic and facultative GNB isolated from hospitalised patients with CA-cIAIs [4,5]. Specifically, isolates obtained during open or laparoscopic surgery or percutaneous drainage were examined only for E. coli or Klebsiella spp. Antimicrobial susceptibility testing was performed using the broth microdilution method in accordance with the guidelines for minimum inhibitory concentration testing and breakpoints recommended by the Clinical and Laboratory Standards Institute (CLSI) in 2011 [17]. 2.4. Statistical analyses Descriptive statistics such as frequencies or percentages for categorical variables, and mean (±standard deviation), median, minimum and maximum for continuous variables were used to describe patients’ baseline demographic data and clinical characteristics. Baseline patient demographic and clinical characteristics (including the details of healthcare-acquisition factors), clinical outcomes and type of causative bacterial species were examined regarding the ESBL status (positive or negative) using the Fisher’s exact test for categorical variables and the t-test for continuous variables. Multivariate logistic regression was used to assess the association of clinical cure and important demographic items, ESBL status, important healthcare-associated factors (prior use of catheterisation or antibiotic agent, and hospitalisation within 3 months prior to inclusion in this study). In addition, as hospital LOS (days) had a skewed, long right tail distribution, the relationship between LOS and ESBL status was analysed by a generalised linear model with a negative binomial distribution and log link after controlling for the factors described above. Goodness-of-fit was assessed for both models. All statistical analyses were conducted at the 5% significance level and were performed using SAS v.9.1 (SAS Institute Inc., Cary, NC). 3. Results 3.1. Patient characteristics Data from 105 eligible patients from Columbia (n = 4), The Philippines (n = 41), Portugal (n = 16), Taiwan (n = 23) and Thailand (n = 21) were available for analysis, of whom 88 (83.8%) had infections due to ESBL-negative Enterobacteriaceae spp. and 17 (16.2%) had infections caused by ESBL-positive Enterobacteriaceae (Table 1). Microbiological assessment showed that E. coli was the predominant organism (n = 77; 73.3%). None of the patients had concomitant E. coli and Klebsiella spp. infections. Clinical characteristics of the patients are shown in Table 2. Patients were predominantly male (n = 64; 61.0%) and the mean age of all patients was 53 years. The majority of patients had chronic digestive diseases or biliary tract obstructions. Clinical assessment of the patients revealed that gallbladder infections, followed by generalised appendicular infections and bile duct infections, were the most frequent causes of CA-cIAIs resulting in hospitalisation. Non-ESBL-producing organisms were more commonly isolated from gallbladder infections (P = 0.0485) and ESBL-producing organisms were more frequently isolated from bile duct infections

Table 1 Number and percentage of patients listed by country and according to extendedspectrum ␤-lactamase (ESBL) status of the infecting micro-organism. Country

Columbia Portugal The Philippines Taiwan Thailand Overall

No. (%) of patients from participating country Total (n = 105)

ESBL-positive (n = 17)

ESBL-negative (n = 88)

4 (3.8) 16 (15.2) 41 (39.0) 23 (21.9) 21 (20.0) 105 (100)

1 (25.0) 3 (18.7) 4 (9.8) 2 (8.7) 7 (33.3) 17 (16.2)

3 (75.0) 13 (81.3) 37 (90.2) 21 (91.3) 14 (66.7) 88 (83.8)

(P = 0.0654). Moreover, 24 patients (22.9%) had been hospitalised within 3 months prior to this CA-cIAI episode. Other treatment procedures that had been performed within 3 months prior to this hospitalisation were venous or urinary tract catheterisation (n = 33; 31.4%), surgery (n = 12; 11.4%) and administration of antibiotics (n = 25; 23.8%). Venous catheterisation and prior hospitalisation were more frequently noted for patients infected with ESBLpositive species than patients infected by ESBL-negative organisms, whereas a history of prior antibiotic use was not. The CCI scores in the patients ranged from 0 to 9. Of the 105 patients, 45 (42.9%) had a CCI ranging from 1 to ≥3, whilst 60 (57.1%) had a score of 0, indicating no co-morbidity. The mean CCI score in patients with ESBL-positive infections was 2.1, whereas that of patients with non-ESBL infections was 1.1. However, there were no significant differences in the percentage of patients with a CCI of 0, 1, 2 or ≥3 between patients with ESBL-positive and those with ESBL-negative infections (Table 2). 3.2. Clinical outcomes Univariate analysis was applied to compare clinical outcomes (cure, improvement but no cure, failure but not death, and death) immediately before discharge between patients with CA-cIAIs due to ESBL-producing organisms and those with infections due to non-ESBL-producing pathogens (Table 3). The outcome could not be evaluated for one patient with CA-cIAI caused by non-ESBLproducing species. At discharge, 59 patients (67.0%) with infections due to non-ESBL-producing organisms were cured, 19 patients (21.8%) showed improvement in clinical symptoms/signs and 9 (10.2%) did not respond to treatment. Of the nine who did not respond, four (44.4%) died due to reasons attributable to the cIAIs. In comparison, 8 (47.1%) of the 17 patients infected with ESBLproducing species were cured, 7 (41.2%) showed improvement in clinical symptoms/signs and 2 (11.8%) did not respond to treatment and died due to infection-related complications. There were no significant differences in mortality rate or percentage of patients who showed improvement in symptoms/signs between patients with infections due to ESBL-producing organisms and those with infections due to non-ESBL-producing pathogens (P = 0.455 and 0.085, respectively). In addition, the median LOS was 11.6 days for patients with infections due to non-ESBL-producing organisms and 17.6 days for patients with infections due to ESBL-producing organisms (P = 0.011) (Table 3). The logistic regression model revealed that ESBL positivity was not significantly associated with failure to cure the CA-cIAIs when controlling for age, sex, site of infection (appendix, bile duct and gallbladder), involved organism and prior use of antibiotics or interventional procedures (urinary or venous catheter) within 3 months prior to current admission (Table 4). However, clinical cure was significantly associated with a CCI score of 1 or 2 [odds ratio (OR) = 0.177 and 0.175 (P = 0.032 and 0.025), respectively]. A borderline significance of association between clinical cure and a CCI ≥3 was also found (OR = 0.0213; P = 0.068).

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

225

Table 2 Clinical characteristics of 105 patients with community-acquired complicated intra-abdominal infections (cIAIs) due to either extended-spectrum ␤-lactamase (ESBL)producing Enterobacteriaceae or non-ESBL-producing organisms. No. (%) of patientsa

Characteristic

P-value

ESBL-positive (n = 17) Age (years) (mean ± S.D.) 53.2 ± 16.1 Sex Female 8 (47.1) 9 (52.9) Male b Site of infection Gallbladder 2 (11.8) 3 (17.6) Appendix 5 (29.4) Bile duct Other 7 (41.2) Causative organisms 5 (29.4) Klebsiella spp. 12 (70.6) Escherichia coli Charlson Comorbidity Indexc 8 (47.1) 0d 2 (11.8) 1 2 (11.8) 2 5 (29.4) ≥3 Three months prior to current hospitalisation for cIAI Previous antibiotic use 7 (41.2) 7 (41.2) Venous catheterisation Prior hospitalisation 8 (47.1)

ESBL-negative (n = 88)

Overall (n = 105)

52.4 ± 18.5

52.5 ± 18.0

33 (37.5) 55 (62.5)

41 (39.0) 64 (61.0)

34 (38.6) 24 (27.3) 10 (11.4) 26 (29.5)

36 (34.3) 27 (25.7) 15 (14.3) 33 (31.4)

23 (26.1) 65 (73.9)

28 (26.7) 77 (73.3)

52 (59.1) 10 (11.4) 12 (13.6) 14 (15.9)

60 (57.1) 12 (11.4) 14 (13.3) 19 (18.1)

0.8619 0.5883

0.0485 0.5500 0.0654 0.3963 0.7705

0.5862

18 (20.5) 15 (17.0) 16 (18.2)

0.1152 0.0455 0.0226

S.D., standard deviation. a Data are no. (%) unless otherwise stated. b Some patients had more than one intra-abdominal infectious source after detailed investigation. c Validated system that assigns a score of 1, 2 or ≥3 to a wide range of moderate-to-severe co-morbid conditions. d The no co-morbidity group was regarded as the reference group.

Table 3 Univariate analysis of clinical status at discharge amongst patients with community-acquired complicated intra-abdominal infections (CA-cIAIs) due to either extendedspectrum ␤-lactamase (ESBL)-producing Enterobacteriaceae or non-ESBL-producing organisms. No. (%) of patientsa

Clinical outcome

Cured at discharge Mortality related to cIAIb Treatment failure, but discharged home Improvement of symptoms or signs, but not cured Not assessable Median length of stay (days) a b

P-value

ESBL-positive (n = 17)

ESBL-negative (n = 88)

8 (47.1) 2 (11.8) 0 (0) 7 (41.2) 0 (0) 17.6

59 (67.0) 4 (4.5) 5 (5.7) 19 (21.6) 1 (1.1) 11.6

0.285 0.455 0.085 0.011

Data are no. (%) unless otherwise stated. A total of 12 deaths, 6 were related to cIAIs.

Table 4 Multivariate logistic regression analysis for predictors of clinical cure amongst 105 patients with community-acquired complicated intra-abdominal infections (cIAIs).a

Table 5 Multivariate linear regression model for predictors of length of stay amongst 105 patients with community-acquired complicated intra-abdominal infections (cIAIs).

Variable

Odds ratio (95% CI)

P-value

Variable

Age Charlson Comorbidity Index 1 (vs. 0) 2 (vs. 0) ≥3 (vs. 0) ESBL-positive (vs. negative) Prior hospitalisation (vs. no prior hospitalisation)b Male (vs. female) Prior antibiotic use (vs. no prior antibiotic use)b Venous catheterisation (vs. no catheterisation)b

1.005 (0.974–1.039)

0.740

Age Charlson Comorbidity Index 1 (vs. 0) 2 (vs. 0) ≥3 (vs. 0) ESBL-positive (vs. negative) Prior hospitalisation (vs. no hospitalisation)a Male (vs. female) Prior antibiotic use (vs. no prior antibiotic use)a Venous catheterisation (vs. no prior catheterisation)a

*

0.177 (0.036–0.863) 0.175 (0.038–0.801) 0.213 (0.041–1.121) 0.513 (0.126–2.091) 1.531 (0.185–12.670)

0.032 0.025* 0.068** 0.352 0.693

0.846 (0.310–2.309) 2.537 (0.547–11.761)

0.744 0.234

0.161 (0.021–1.211)

0.76

CI, confidence interval; ESBL, extended-spectrum ␤-lactamase. a Also adjusted for bacterial type (Escherichia coli or Klebsiella spp.), IAI source (focus of infection) and past exposure to urinary catheters. b Three months prior to current hospitalisation for cIAI. * Significantly different from zero at 0.95 confidence level. ** Significantly different from zero at 0.90 confidence level.

Estimate (Wald confidence interval) 0.002 (−0.007, 0.010)

P-value 0.718

0.266 (−0.129, 0.660) 0.349 (−0.025, 0.724) 0.382 (−0.057, 0.820) 0.424 (0.097, 0.752) −0.252 (−0.743, 0.240)

0.186 0.068** 0.088** 0.011* 0.316

0.159 (−0.083, 0.402) −0.078 (−0.407, 0.251)

0.198 0.640

−0.004 (−0.538, 0.530)

0.988

ESBL, extended-spectrum ␤-lactamase. a Three months prior to current hospitalisation for cIAI. * Significantly different from zero at 0.95 confidence level. ** Significantly different from zero at 0.90 confidence level.

226

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

Analysis of the generalised linear model (Table 5) revealed that ESBL positivity was associated with a 0.424-day increase in LOS after controlling for age, sex, site of infection, organism involved, CCI and prior use of antibiotics or interventional procedures (urinary or venous catheter) within 3 months prior to this hospitalisation (P = 0.011). In addition, patients with at least one underlying co-morbidity (CCI = 2 and ≥3) did not have a significantly longer LOS than patients without co-morbidities (P = 0.186, 0.068 and 0.088 for patients with a CCI of 1, 2 and ≥3, respectively). Table 6 illustrates the appropriateness of empirical treatment regimens for patients with infections due to non-ESBL-producing organisms and patients with infections caused by ESBL-producing pathogens. Third-generation cephalosporin agents were more frequently administered to patients with infections due to nonESBL-producing organisms (P = 0.0287). Table 7 shows that the probability of treating ESBL-positive patients with appropriate empirical antibiotic(s) was approximately one-half (range 45–53%) of that in ESBL-negative patients, if all of the regimens consisting of ␤-lactam/␤-lactamase inhibitor combination were regarded as inappropriate therapy against the intra-abdominal ESBL-producing E. coli or Klebsiella spp.

4. Discussion Most studies on the clinical impact of infections due to ESBLproducing species of Enterobacteriaceae have focused on mortality rates due to sepsis and on acquisition risk factors associated with these resistant pathogens [7,18–20]. In this study, we found that the overall mortality rate of patients with CA-IAIs was 11.4% (12/105) and that 50% of the deaths were directly attributable to CA-cIAIs. The case fatality rate in this study is indeed significantly lower than that reported in prior investigations of IAIs [21–23]. Factors related to mortality amongst patients with IAIs include high bacterial load (i.e. high inoculum) in the intra-abdominal cavity before surgery, the diversity of intra-abdominal organ structures and adequacy of surgical intervention [11,22]. Amongst these factors, timely surgical intervention is the most important factor affecting the outcome of patients with cIAIs. In the current study, timely surgical intervention was performed for every enrolled patient. Studies have shown that >80% of cases of community-onset bacteraemia caused by ESBL-producing E. coli can be treated medically [7,24]. The current study shows that patients with higher CCI scores do not necessarily have a predisposition of acquiring infections due to ESBL-producing organisms. However, in agreement with results of previous surveys [7,24], we found that the presence of some healthcare-associated factors (recent hospitalisation or venous catheterisation) was associated with a risk of acquiring IAIs caused by community-acquired ESBL-producing Enterobacteriaceae spp. Roehrborn et al. reported that the resistance profiles of pathogens causing IAIs were closely related to the spectra of latest preceding antimicrobial agents [11]. Nevertheless, their study enrolled patients with post-operative peritonitis, whereas the current study included patients with CA-cIAIs. The appropriateness of antibiotic agents might be dependent upon the clinical severity (e.g. haemodynamic status and Pitt bacteraemic score), underlying co-morbidities and the subjective judgement of the surgeon or attending physician. Therefore, although prior antibiotic usage predisposed some patients to colonisation by resistant organisms, some of the patients benefitted from therapy. These possibilities explain, at least in part, why prior antibiotic usage is statistically irrelevant to the clinical cure rate and hospital LOS in patients with CA-cIAIs. Results of the multivariate analysis show that the clinical cure rate is adversely affected by the existence of any underlying

co-morbidity after controlling for other covariates, although the statistical difference is borderline (P = 0.068) in the patient group with Charlson index ≥3. This association was also reported by Cattan et al. [10]. In addition, Cattan et al. also found that patients with co-morbid conditions were more likely to fail to respond to initial empirical antimicrobial drugs (OR = 1.023) [10]. Krobot et al. found that poor response to antibiotics prolonged the duration of hospital stay in patients with IAIs [9]. Consequently, patient co-morbidities possibly prolong LOS, which is partially consistent with our findings (Table 5). In contrast, ESBL-positive status does not significantly influence the cure rate in patients with CA-cIAIs. Lautenbach et al. had found that infections (mostly non-surgical aetiologies) caused by ESBL-producing E. coli or K. pneumoniae predispose patients to receive inappropriate antibiotic therapy initially [19]. Furthermore, patients with infections due to ESBL-producing organisms had longer LOS (average of LOS difference, 4 days) and had lower rates of clinical response (OR = 1.43, 95% confidence interval 0.95–2.14; P = 0.08) compared with patients with infections due to non-ESBLproducing pathogens [19]. Thus, the impact of ESBL positivity on clinical cure rate in their survey differs from that of ESBL positivity in patients with CA-cIAIs. ˜ et al. found that regimens comRecently, Rodríguez-Bano prising a ␤-lactam/␤-lactamase inhibitor (BLBLI) [such as piperacillin/tazobactam (TZP) and amoxicillin/clavulanic acid (AMC)] are appropriate alternatives to carbapenems in treating bacteraemia due to ESBL-producing E. coli if they are susceptible in vitro [25]. Nevertheless, the inoculum effect, possibly existing in some scenarios of cIAI, has been shown to exert a significant influence on the in vivo efficacies of these two agents (especially TZP) against ESBL-producing E. coli in a murine sepsis model [26,27]. Timely adequate surgery, which was performed for every patient in the current study, is beneficial in rapidly reducing the intraperitoneal bacterial burden. Thus, early surgical intervention might lessen the impact of high inoculum on the in vivo effectiveness of these antimicrobial agents. Although the susceptibility rate of ESBL-producing Klebsiella spp. to TZP was much lower (ranging from 26% to 47%) than that of ESBL-producing E. coli (ranging from 62% to 87%) [6,28], the majority (73.3%) of causative organisms in this series of CA-cIAIs was E. coli. If all of the ESBL-producing isolates from the CA-cIAIs of this survey are hypothesised to be susceptible to TZP, AMC and regimens of other BLBLIs in vitro, the probability of prescribing appropriate antibiotic(s) against ESBLpositive infections would increase to 0.80 [i.e. (12/18)/(80/96)] if BLBLI regimens are regarded as inappropriate therapy for ESBLnegative infections but appropriate for ESBL-positive producers, and to 0.68 [i.e. (12/18)/(94/96)] if BLBLIs are deemed appropriate therapy against both ESBL-positive and -negative infections. Thus, in real clinical settings of CA-cIAIs, the probability of empirically using appropriate antibiotics against ESBL-cIAI pathogens could be adjusted to the range of 0.53–0.80 and 0.45–0.68, respectively. The present investigation was not designed to survey the faecal carriage rate of ESBL-producing enteric GNB in healthy community subjects of participating countries. Amongst healthy paediatric or adult populations, the prevalence rate of faecal carriage of ESBLproducing Enterobacteriaceae was estimated to be 2.7% in Portugal [29], 3.7–10.7% in Spain [30] and 29.3–69.3% in Thailand [31,32]. Although faecal carriage of ESBL-producing Enterobacteriaceae was not investigated in Colombia, other countries in South America were reported as having 1.7% of ESBL-producing enteric GNB in faecal samples from healthy community children during the last decade [33]. There are four limitations to this investigation regarding CAcIAIs. First, the number of enrolled patients is small given the fact that the study continued for 1.5 years in multiple centres. The number of patients screened during the study period was ca. 10-fold the number of patients actually enrolled. The majority of cases

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

227

Table 6 Administration of empirical antibiotics for treatment of complicated intra-abdominal infections (cIAIs) and appropriateness based on known resistance patterns.a Empirical antimicrobial treatment for cIAI

Penicillin (ampicillin or amoxicillin) BLBLI regimen (TZP, AMC or other ␤-lactam combined with clavulanic acid, etc.) Carbapenems (imipenem, meropenem, ertapenem, doripenem) First-generation cephalosporin Second-generation cephalosporin Third-generation cephalosporin Anti-anaerobe (metronidazole or clindamycin) Fluoroquinolones (ciprofloxacin, levofloxacin or moxifloxacin) Other antibiotics

ESBL-positive

ESBL-negative

No. (%)

Appropriate

No. (%)

Appropriate

1 (5.9) 4 (23.5) 4 (23.5) 1 (5.9) 4 (23.5) 2 (11.8) 5 (29.4) 2 (11.8) 1 (5.9)

0 0 4 0 4 0 N/A 0 N/A

1 (1.1) 14 (15.9) 10 (11.4) 1 (1.1) 31 (35.2) 35 (39.8) 44 (50.0) 4 (4.5) 2 (2.3)

0 0 or 14b 10 0 31 35 N/A 4 N/A

Total no. (%)

P-value

2 (1.9) 18 (17.1) 14 (13.3) 2 (1.9) 35 (33.3) 37 (35.2) 49 (46.7) 6 (5.7) 3 (2.9)

0.2989 0.4850 0.2359 0.2989 0.4123 0.0287* 0.1836 0.2494 0.4146

ESBL, extended-spectrum ␤-lactamase; BLBLI, ␤-lactam/␤-lactamase inhibitor; TZP, piperacillin/tazobactam; AMC, piperacillin/tazobactam; N/A, not applicable. a Resistance patterns used to decide the appropriateness were obtained from the Study for Monitoring Antimicrobial Resistance Trends (SMART) registry for the countries contributing isolates in this study. b Considered the use of ␤-lactam/␤-lactamase inhibitor combination regimens for ESBL-negative infections as either appropriate or inappropriate. * , P-Value <0.05 Table 7 Summary of appropriate and inappropriate use of antibiotic(s) categorised by extended-spectrum ␤-lactamase (ESBL) status. ESBL-positive

ESBL-negative

Appropriate usea 8 Inappropriate use 10 18 Total instances of antibiotic(s) useb,c Probability of prescribing appropriate antibiotic(s) against ESBL-positive infections (vs. ESBL-negative ones)

BLBLIa inappropriate

BLBLIa appropriate

80 16 96 0.53d

94 2 96 0.45e

BLBLI, ␤-lactam(s)/␤-lactamase inhibitor. a Based on the Study for Monitoring Antimicrobial Resistance Trends (SMART) registry resistance patterns: if the susceptibilities of some class of antibiotic(s) against Escherichia coli and Klebsiella spp. were <80%, use of empirical use of an antibiotic agent from this class was considered as inappropriate. Since piperacillin/tazobactam (TZP) has a susceptibility level >80% against ESBL-negative infections for all participating countries, and we could not separate TZP from other BLBLI combination regimens that had decreased susceptible rates, both appropriateness and inappropriateness for ESBL-negative infections were analysed. b Excluded anti-anaerobes and ‘other’ antibiotics from the denominator. c Total instances of antibiotic(s) use may exceed the number of patients within each group as combination of at least two antibiotics might be used empirically. d Compared with ESBL-negative patients, the probability of appropriate antibiotic use in ESBL-positive infections is (8/18)/(80/96) = 0.53 when regimens of other BLBLI are considered inappropriate. e By contrast, if we consider the empirical use of other BLBLI regimens as appropriate, the probability of using appropriate antibiotic against ESBL-positive vs. ESBL-negative infections is (8/18)/(94/96) = 0.45.

were excluded due to healthcare-associated infections, isolates not specified being collected within 48 h of admission, localised cholecystitis requiring surgery or percutaneous drainage, and nonperforated appendicitis, or infections due to pathogens other than E. coli or K. pneumoniae. Second, we did not investigate risk factors related to mortality. Third, information regarding the clinical severity of every patient upon admission was not available so we could not explore its impact on clinical cure rate and LOS. Finally, this multicentre study was not intended to collect isolates from different countries for further evaluation, and the genotypes of the ESBL-producing isolates were not investigated. In conclusion, ESBL-producing E. coli and Klebsiella spp. should be suspected as the causative organisms in patients with CA-cIAIs who have healthcare-associated factors such as recent hospitalisation or venous catheterisation. Pre-existing co-morbidities, but not ESBL positivity, were found to be adversely associated with clinical cure of CA-cIAIs. In contrast, duration of hospitalisation is longer in CA-cIAI patients infected with ESBL-producing isolates or with co-morbidities. Further research is needed to determine the appropriateness of using BLBLI regimens against ESBL-producing organisms in patients with CA-cIAIs. Funding: This study was supported by Merck & Co. (Whitehouse Station, NJ, USA). Competing interests: None declared. Ethical approval: This study was approved by the Institutional Review Boards of the participating centres. The same document in Chinese was approved by the Institutional Review Board of the National Taiwan University Hospital [NTUH IRB 201005033R]

and National Cheng Kung University Hospital [NCKUH ER-99-123]. Signed consent forms for all enrolled subjects were obtained.

References [1] Sartelli M, Catena F, Ansaloni L, Lazzareschi DV, Taviloglu K, Van Goor H, et al. Complicated Intra-Abdominal Infections Observational European study (CIAO Study). World J Emerg Surg 2011;6:40. [2] Paterson DL, Bonomo RA. Extended-spectrum ␤-lactamases: a clinical update. Clin Microbiol Rev 2005;18:657–86. [3] Villegas MV, Blanco MG, Sifuentes-Osornio J, Rossi F. Increasing prevalence of extended-spectrum ␤-lactamase among Gram-negative bacilli in Latin America—2008 update from the Study for Monitoring Antimicrobial Resistance Trends (SMART). Braz J Infect Dis 2011;15:34–9. [4] Hawser SP, Bouchillon SK, Hoban DJ, Badal RE, Cantón R, Baquero F. Incidence and antimicrobial susceptibility of Escherichia coli and Klebsiella pneumoniae with extended-spectrum ␤-lactamases in community- and hospital-associated intra-abdominal infections in Europe: results of the 2008 Study for Monitoring Antimicrobial Resistance Trends (SMART). Antimicrob Agents Chemother 2010;54:3043–6. [5] Hawser SP, Bouchillon SK, Hoban DJ, Badal RE, Hsueh PR, Paterson DL. Emergence of high levels of extended-spectrum ␤-lactamase-producing Gramnegative bacilli in the Asia-Pacific region: data from the Study for Monitoring Antimicrobial Resistance Trends (SMART) program, 2007. Antimicrob Agents Chemother 2009;53:3280–4. [6] Chen YH, Hsueh PR, Badal RE, Hawser SP, Hoban DJ, Bouchillon SK, et al. Antimicrobial susceptibility profiles of aerobic and facultative Gram-negative bacilli isolated from patients with intra-abdominal infections in the Asia-Pacific region according to currently established susceptibility interpretive criteria. J Infect 2011;62:280–91. ˜ ˜ C, [7] Rodríguez-Bano J, Picón E, Gijón P, Hernández JR, Ruíz M, Pena et al. Community-onset bacteremia due to extended-spectrum ␤-lactamaseproducing Escherichia coli: risk factors and prognosis. Clin Infect Dis 2010;50:40–8.

228

S.-S. Jean et al. / International Journal of Antimicrobial Agents 44 (2014) 222–228

[8] Sartelli M, Catena F, Ansaloni L, Leppaniemi A, Taviloglu K, van Goor H, et al. Complicated intra-abdominal infections in Europe: preliminary data from the first three months of the CIAO Study. World J Emerg Surg 2012; 7:15. [9] Krobot K, Yin D, Zhang Q, Sen S, Altendorf-Hofmann A, Scheele J, et al. Effect of inappropriate initial empiric antibiotic therapy on outcome of patients with community-acquired intra-abdominal infections requiring surgery. Eur J Clin Microbiol Infect Dis 2004;23:682–7. [10] Cattan P, Yin DD, Sarfati E, Lyu R, De Zelicourt M, Fagnani F. Cost of care for inpatients with community-acquired intra-abdominal infections. Eur J Clin Microbiol Infect Dis 2002;21:787–93. [11] Roehrborn A, Thomas L, Potreck O, Ebener C, Ohmann C, Goretzki PE, et al. The microbiology of postoperative peritonitis. Clin Infect Dis 2001;33: 1513–9. [12] Marra AR, Wey SB, Castelo A, Gales AC, Cal RG, Filho JR, et al. Nosocomial bloodstream infections caused by Klebsiella pneumoniae: impact of extendedspectrum ␤-lactamase (ESBL) production on clinical outcome in a hospital with high ESBL prevalence. BMC Infect Dis 2006;6:24. ˜ C, Pujol M, Ardanuy C, Ricart A, Pallarés R, Linares ˜ [13] Pena J, et al. An outbreak of hospital-acquired Klebsiella pneumoniae bacteraemia, including strains producing extended-spectrum ␤-lactamase. J Hosp Infect 2001;47: 53–9. [14] Yang Q, Zhang H, Wang Y, Xu Y, Chen M, Badal RE, et al. A 10 year surveillance for antimicrobial susceptibility of Escherichia coli and Klebsiella pneumoniae in community- and hospital-associated intra-abdominal infections in China. J Med Microbiol 2013;62:1343–9. [15] Stürenburg E, Sobottka I, Laufs R, Mack D. Evaluation of a new screen agar plate for detection and presumptive identification of Enterobacteriaceae producing extended-spectrum ␤-lactamases. Diagn Microbiol Infect Dis 2005;51: 51–5. [16] M’Zali FH, Chanawong A, Kerr KG, Birkenhead D, Hawkey PM. Detection of extended-spectrum ␤-lactamases in members of the family Enterobacteriaceae: comparison of the MAST DD test, the double disc and the Etest ESBL. J Antimicrob Chemother 2000;45:881–5. [17] Clinical and Laboratory Standards Institute. Performance standards for antimicrobial susceptibility testing; twenty-first informational supplement. Document M100-S21. Wayne, PA: CLSI; 2011. [18] Hyle EP, Lipworth AD, Zaoutis TE, Nachamkin I, Bilker WB, Lautenbach E. Impact of inadequate initial antimicrobial therapy on mortality in infections due to extended-spectrum ␤-lactamase-producing Enterobacteriaceae: variability by site of infection. Arch Intern Med 2005;165:1375–80. [19] Lautenbach E, Patel JB, Bilker WB, Edelstein PH, Fishman NO. Extendedspectrum ␤-lactamase-producing Escherichia coli and Klebsiella pneumoniae: risk factors for infection and impact of resistance on outcomes. Clin Infect Dis 2001;32:1162–71. [20] Qureshi ZA, Paterson DL, Pakstis DL, Adams-Haduch JM, Sandkovsky G, Sordillo E, et al. Risk factors and outcome of extended-spectrum ␤-lactamaseproducing Enterobacter cloacae bloodstream infections. Int J Antimicrob Agents 2011;37:26–32. [21] Riché FC, Dray X, Laisné MJ, Matéo J, Raskine L, Sanson-Le Pors MJ, et al. Factors associated with septic shock and mortality in generalized peritonitis: comparison between community-acquired and postoperative peritonitis. Crit Care 2009;13:R99.

[22] Pacelli F, Doglietto GB, Alfieri S, Piccioni E, Sgadari A, Gui D, et al. Prognosis in intra-abdominal infections. Multivariate analysis on 604 patients. Arch Surg 1996;131:641–5. [23] Sturkenboom MC, Goettsch WG, Picelli G, in’t Veld B, Yin DD, de Jong RB, et al. Inappropriate initial treatment of secondary intra-abdominal infections leads to increased risk of clinical failure and costs. Br J Clin Pharmacol 2005;60:438–43. [24] Kang CI, Song JH, Chung DR, Peck KR, Ko KS, Yeom JS, et al. Risk factors and treatment outcomes of community-onset bacteraemia caused by extendedspectrum ␤-lactamase-producing Escherichia coli. Int J Antimicrob Agents 2010;36:284–7. ˜ J, Navarro MD, Retamar P, Picón E, Pascual Á. ␤-Lactam/␤[25] Rodríguez-Bano lactamase inhibitor combinations for the treatment of bacteremia due to extended-spectrum ␤-lactamase-producing Escherichia coli: a post hoc analysis of prospective cohorts. Clin Infect Dis 2012;54:167–74. [26] Docobo-Pérez F, López-Cerero L, López-Rojas R, Egea P, Domínguez˜ J, et al. Inoculum effect on the efficacies Herrera J, Rodríguez-Bano of amoxicillin–clavulanate, piperacillin–tazobactam, and imipenem against extended-spectrum ␤-lactamase (ESBL)-producing and non-ESBL-producing Escherichia coli in an experimental murine sepsis model. Antimicrob Agents Chemother 2013;57:2109–13. [27] López-Cerero L, Picón E, Morillo C, Hernández JR, Docobo F, Pachón J, et al. Comparative assessment of inoculum effects on the antimicrobial activity of amoxicillin–clavulanate and piperacillin–tazobactam with extended-spectrum ␤-lactamase-producing and extended-spectrum ␤-lactamase-non-producing Escherichia coli isolates. Clin Microbiol Infect 2010;16:132–6. [28] Hoban DJ, Bouchillon SK, Hawser SP, Badal RE, Labombardi VJ, DiPersio J. Susceptibility of Gram-negative pathogens isolated from patients with complicated intra-abdominal infections in the United States, 2007–2008: results of the Study for Monitoring Antimicrobial Resistance Trends (SMART). Antimicrob Agents Chemother 2010;54:3031–4. [29] Guimarães B, Barreto A, Radhouani H, Figueiredo N, Gaspar E, Rodrigues J, et al. Genetic detection of extended-spectrum ␤-lactamase-containing Escherichia coli isolates and vancomycin-resistant enterococci in fecal samples of healthy children. Microb Drug Resist 2009;15:211–6. ˜ [30] Fernández-Reyes M, Vicente D, Gomariz M, Esnal O, Landa J, Onate E, et al. High rate of fecal carriage of extended-spectrum-␤-lactamase-producing Escherichia coli in healthy children in Gipuzkoa, northern Spain. Antimicrob Agents Chemother 2014;58:1822–4. [31] Luvsansharav UO, Hirai I, Nakata A, Imura K, Yamauchi K, Niki M, et al. Prevalence of and risk factors associated with faecal carriage of CTXM ␤-lactamase-producing Enterobacteriaceae in rural Thai communities. J Antimicrob Chemother 2012;67:1769–74. [32] Luvsansharav UO, Hirai I, Niki M, Sasaki T, Makimoto K, Komalamisra C, et al. Analysis of risk factors for a high prevalence of extended-spectrum ␤lactamase-producing Enterobacteriaceae in asymptomatic individuals in rural Thailand. J Med Microbiol 2011;60:619–24. [33] Pallecchi L, Bartoloni A, Fiorelli C, Mantella A, Di Maggio T, Gamboa H, et al. Rapid dissemination and diversity of CTX-M extended-spectrum ␤lactamase genes in commensal Escherichia coli isolates from healthy children from low-resource settings in Latin America. Antimicrob Agents Chemother 2007;51:2720–5.