- Email: [email protected]
Risk factors for the acquisition of carbapenem-resistant Escherichia coli among hospitalized patients
Available online at www.sciencedirect.com
Diagnostic Microbiology and Infectious Disease 62 (2008) 402 – 406 www.elsevier.com/locate/diagmicrobio
Clinical Studies
Risk factors for the acquisition of carbapenem-resistant Escherichia coli among hospitalized patients Min-Hyok Jeona , Sang-Ho Choib,c , Yee Gyung Kwakb,c , Jin-Won Chungb,c , Sang-Oh Leeb,c , Jin-Yong Jeongb,c,d , Jun Hee Woob,c , Yang Soo Kimb,c,⁎ a
Department of Infectious Diseases, University of Soonchunhyang College of Medicine, Choongchungnam-do 336-745, Republic of Korea b Division of Infectious Diseases, Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea c Center for Antimicrobial Resistance and Microbial Genetics, University of Ulsan, Seoul 138-736, Republic of Korea d Asan Institute of Life Sciences, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea Received 5 December 2007; accepted 14 August 2008
Abstract Carbapenem resistance among Gram-negative bacilli has become an increasingly serious problem worldwide, and the emergence and spread of carbapenem-resistant Escherichia coli (CREC) is also becoming a serious problem. To date, however, risk factors for CREC acquisition have not been determined, so we decided to evaluate this in hospitalized patients through matched case-control study. Nosocomially acquired CREC was isolated from 46 patients between January 1997 and December 2007. For each patient, 3 matched-control subjects were selected. Previous use of carbapenem (adjusted odds ratio [AOR], 6.50) and metronidazole (AOR, 4.25), the presence of biliary drainage catheter (AOR, 4.59), and prior hospital stay (AOR 1.02) were found as independent risk factors for CREC. Our results suggest that the nosocomial acquisition of CREC may be favored by the selection pressure of carbapenems and metronidazole and also related to prior hospital stay and the presence of biliary drainage catheter. © 2008 Elsevier Inc. All rights reserved. Keywords: Carbapenem; Resistance; Escherichia coli
1. Introduction Carbapenems, a class of β-lactam antibiotics with a broad spectrum of antibacterial activity, have been one of the last remaining therapeutic choices for infections caused by cephalosporin-resistant microorganisms. However, their effectiveness is being challenged by the emergence and spread of carbapenem-resistant Gramnegative rods, such as Pseudomonas aeruginosa and Acinetobacter baumannii (Gaynes and Culver, 1992; Lee et al., 2004; NNIS, 2004). Recently, carbapenem resistance in Escherichia coli is also emerging worldwide (Gulmez et al., 2008; Hong et al., 2005; Miriagou et al.,
⁎ Corresponding author. Division of Infectious Diseases, Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea. Tel.: +82-2-3010-3303; fax: +82-2-3010-6970. E-mail address: [email protected] (Y.S. Kim). 0732-8893/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2008.08.014
2003; Moloughney et al., 2005; Poirel et al., 2004; Stapleton et al., 1999; Urban et al., 2008). Mechanisms causing resistance to the carbapenems in E. coli have been suggested to include the presence of an AmpC β-lactamase in association with the loss of porin (Poirel et al., 2004; Stapleton et al., 1999), metallo-βlactamases (Miriagou et al., 2003), and class A carbapenemases such as Klebsiella pneumoniae carbapenemase type 2 (KPC-2) (Navon-Venezia et al., 2006) and KPC-3 (Hong et al., 2005). Recently, OXA-48–like carbapenemases and outer membrane protein loss were also reported (Gulmez et al., 2008). Carbapenem-resistant E. coli (CREC) strains were 1st isolated in our hospital in 1997, and since then, they are being recovered steadily. After our limited observation and clinical experience with CREC, which poses a potential threat to the hospital infection control, we assessed the risk factors for CREC acquisition in hospitalized patients.
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
403
2. Materials and methods
2.3. Statistical analysis
2.1. Case definition, control definition, and study design
Continuous variables are presented as mean ± SD, and categorical variables are presented as numbers and percentages. Univariate analyses were performed separately for each of the variables using conditional logistic regression to account for the matching between cases and controls. Variables for which the P value was b0.2 in univariate analysis or clinically plausible for the acquisition of carbapenem resistance (i.e., previous exposure to antibiotics, prior ICU stay, and time at risk) were included in a conditional logistic regression model for multivariate analysis. A backward elimination process was used, and adjusted odds ratios (AOR) and 95% confidence intervals (CIs) were calculated. P value of b0.05 was considered significant. All statistical analyses were performed using SAS release 9.1 (SAS, Cary, NC).
A case-control study was conducted at the Asan Medical Center in Seoul, Korea, a 2200-bed tertiary-care teaching hospital with 159 intensive care unit (ICU) beds and approximately 10 000 admissions per year. The use of imipenem or meropenem at this institution is not restricted to specific wards or to treatment of infections due to microorganisms resistant to other agents, but, in principle, it is monitored and supervised by infectious disease specialists. The microbiology laboratory database was searched to identify all clinical cultures positive for CREC from patients admitted between January 1997 and December 2007. Identification of E. coli and susceptibility testing were performed using the MicroScan (Dade Behring, West Sacramento, CA) system and the NEG Combo Type 21 panel (Dade Behring), using breakpoints recommended by the Clinical and Laboratory Standards Institute (formerly National Committee for Clinical Laboratory Standards) (NCCLS, 2004). Resistance to carbapenem was defined as an MIC to imipenem or meropenem exceeding 8 μg/mL, thus, including isolates of intermediate resistance (MIC, 8 μg/mL). Patients were assigned as cases if their isolates had an MIC to imipenem or meropenem ≥8 μg/mL. Nosocomial acquisition of CREC was defined as isolation N48 h after admission to the hospital. For the purpose of comparison, control patients were selected at a ratio of 3:1 from patients present in the same ward on the date that CREC was isolated, matching as closely as possible for sex, age, and admission date. When there were more than 3 subjects satisfying these conditions, an independent person chose 3 subjects randomly. 2.2. Risk factors investigated Potential risk factors for CREC were ascertained by review of medical records. Variables explored as possible risk factors included age, sex, underlying diseases and comorbid conditions, Charlson score (Charlson et al., 1994, 1987), ICU stay before nosocomial isolation of E. coli (within 1 month), prior hospital stay including transfer from another hospital in the previous year, recent surgery (within 1 month), and length of hospital stay before the outcome of interest, which we called the time at risk (Harris et al., 2001, 2002b). For case patients, time at risk was defined as length of hospital stay before CREC isolation, whereas for control patients, time at risk was defined as complete length of hospital stay. The presence of a central venous catheter, biliary drainage catheter, urinary catheter, or mechanical ventilation was also assessed. Prior antibiotic exposure was defined as at least 24 h of therapy during the 14 days before CREC isolation for case patients and before discharge for control patients (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004).
3. Results During the 11-year study period, CREC was isolated from 46 patients who met the criteria for nosocomial acquisition: 1 patient in 1998, 8 in 1999, 3 in 2000, 2 in 2001, 10 each in 2002 and 2003, 1 in 2004, 3 in 2005, 5 in 2006, and 3 in 2007. A total of 138 patients were included in the control group. Carbapenem-resistant E. coli was most frequently recovered from surgical drainage specimens (28.3%), followed by bile (21.7%), blood (13.0%), respiratory secretions (10.9%), ascitic fluid (8.7%), and urine (6.5%). Twenty-nine (63%) of 46 CREC isolates were recovered with other concomitant bacteria: 17 (37%) P. aeruginosa including 9 (19.7%) carbapenem-resistant organisms and 12 (26%) Enterococcus spp. including 3 vancomycin-resistant bacteria. One CREC isolate was recovered with both Carbepenem-resistant Pseudomonas aeruginosa (CRPA) and Enterococcus spp. On the date that a positive culture result was obtained, patients with CREC were most frequently in surgical wards (32.6%), followed by medical wards (28.3%), solid organ transplantation services (15.2%), surgical ICU (15.2%), medical ICU (4.3%), and hemato-oncology services (4.3%). Almost all (45/46) CREC isolates were resistant to 3rdgeneration cephalosporins. Of the 46 isolates, 36 (78.3%) were susceptible in vitro to amikacin, 18 (39.5%) to trimethoprim/sulfamethoxazole, 13 (28.3%) to gentamicin, and 2 (4.3%) to ciprofloxacin. Comparison of the 2 groups and results of univariate analysis regarding risk factors for nosocomial acquisition of CREC are listed in Table 1. The results of multivariable analysis are presented in Table 2. Multivariable conditional logistic regression analysis demonstrated that patients with nosocomial isolation of CREC were more likely to have been exposed to carbapenems (AOR, 6.50; 95% CI, 2.33–18.16) and metronidazole (AOR, .4.25; 95% CI, 1.56–11.59) in the 14 days before the date of a positive culture. Also, the presence of biliary drainage catheter and prior hospital stay within 1
404
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
Table 1 Univariate analysis of risk factors for the emergence of CREC Variable
Demographic characteristics Gender, male Age (mean ± SD) Related to hospitalization Prior hospital stay (b1 year) ICU stay Abdominal surgery (b30 days) Time at risk (mean ± SD)b Steroid administration Total parenteral nutrition Central catheter Mechanical ventilation Biliary drainage Urinary catheter insertion Underlying disorder Diabetes mellitus Malignancy Hepatic disease Biliary disease Renal disease Cardiac disease Pulmonary disease Renal transplantation Charlson score (median ± SD) Prior antibiotic exposure(s) Carbapenems Ampicillin/sulbactam Fluoroquinolones Piperacillin/tazobactam Cephalosporins 1st generation 2nd generation 3rd generation Aminoglycosides Metronidazole Vancomycin Both carbapenems and MTZ
Study group Casea (n = 46)
Controla (n = 138)
25 (54.3%) 56.98 ± 13.23
75 (54.3%) 58.15 ± 12.11
23 (50.0%) 24 (52.2%) 20 (43.5%) 74.34 ± 109.60 10 (21.7%) 9 (19.6%) 21 (45.7%) 12 (26.1%) 19 (41.3%) 20 (43.5%)
49 (35.5%) 61 (44.2%) 56 (40.6%) 37.41 ± 32.24 30 (21.7%) 33 (23.9%) 59 (42.8%) 38 (27.5%) 26 (18.8%) 64 (46.4%)
.082 .261 .686 .005 1.000 .523 .670 .823 .001 .674
1.85 1.60 1.17 1.02 1.00 0.76 1.20 0.91 4.85 0.84
0.93–3.69 0.70–3.66 0.54–2.59 1.01–1.03 0.41–2.43 0.32–1.79 0.52–2.76 0.38–2.18 1.85–12.75 0.37–1.91
11 (23.9%) 21 (45.7%) 19 (41.3%) 22 (47.8%) 10 (21.7%) 2 (4.3%) 4 (8.7%) 1 (2.2%) 3.72 ± 2.08
28 (20.3%) 63 (45.7%) 45 (32.6%) 38 (27.5%) 22 (15.9%) 14 (10.1%) 16 (11.6%) 3 (2.2%) 3.20 ± 2.21
.593 1.000 .209 .007 .348 .208 .565 1.000 .129
1.25 1.00 1.68 3.00 1.52 0.36 0.70 1.00 1.14
0.55–2.85 0.49–2.06 0.75–3.79 1.34–6.70 0.63–3.68 0.07–1.78 0.21–2.35 0.10–9.61 0.96–1.34
29 (63.0%) 0 (0.0%) 16 (34.8%) 3 (6.5%)
28 (20.3%) 10 (7.2%) 50 (36.2%) 8 (5.8%)
b.001 .990 .858 .827
8.14 b0.001 0.94 1.21
3.47–19.05 −INF, INF 0.46–1.90 0.22–6.65
1 (2.2%) 1 (2.2%) 28 (60.9%) 6 (13.0%) 25 (54.3%) 21 (45.7%) 16 (34.8%)
3 (2.2%) 18 (13.0%) 77 (55.8%) 20 (14.5%) 45 (32.6%) 34 (24.6%) 6 (4.3%)
1.000 .051 .495 .803 .005 .007 b.001
1.00 0.12 1.31 0.88 3.02 3.53 9.24
0.10–9.61 0.02–1.01 0.60–2.87 0.32–2.40 1.39–6.56 1.52–8.20 3.38–25.29
P
OR
95% CI
– –
– –
MTZ = metronidazole; INF = infinity. a Data are number (%) of patients unless otherwise indicated. b For cases, time at risk before isolation of CREC from clinical culture; for controls, complete length of hospital stay.
year were statistically significant risk factors associated with the acquisition of CREC (AOR, 4.59; 95% CI, 1.18–17.78; and odds ratio [OR], 1.02; 95% CI, 1.00–1.03, respectively). 4. Discussion To date, there have been a few descriptions of CREC strains, with most reports describing the mechanisms responsible for carbapenem resistance (Bratu et al., 2007; Gulmez et al., 2008; Hong et al., 2005; Moloughney et al., 2005; Navon-Venezia et al., 2006; Poirel et al., 2004; Stapleton et al., 1999). Moreover, to our knowledge, the potential risk factors for the acquisition of CREC have not been assessed, although risk factors for carbapenem resistance to P. aeruginosa, A. baumannii, and K. pneumoniae have been determined (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004; Troillet et al., 1997). Thus, this
matched case-control study is the 1st to assess potential risk factors for the acquisition of CREC in clinical specimens from hospitalized patients. Our results show that imipenem and metronidazole were the antibiotics associated with acquisition of CREC. The finding that prior use of carbapenems was associated with isolation of CREC was comparable with results showing that prior exposure to imipenem was a major risk factor for Table 2 Multivariate analysis of risk factors for the emergence of CREC Variables
Odds ratio
95% C.I.
P
Carbapenem use Metronidazole use Presence of biliary drainage catheter Prior hospital stay (b1 year)
6.50 4.25 4.59
2.33–18.16 1.56–11.59 1.18–17.78
b.001 .005 .028
1.02
1.00–1.03
.012
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
imipenem-resistant P. aeruginosa, A. baumannii, and K. pneumoniae in hospitalized patients (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004; Troillet et al., 1997). The association between metronidazole exposure and acquisition of CREC is intriguing. Although conditional logistic regression analyses adjust for all other explanatory variables in the model, they do not necessarily identify causal associations. Thus, the association between exposure to metronidazole and acquisition of CREC may actually be due to the exposure of these patients to a number of different antibiotics before CREC is cultured. Indeed, CREC isolates were recovered most frequently from surgical drain specimens and bile, and majority (15 of 23) of patients with these isolates were exposed to metronidazole. Metronidazole may also have an indirect effect on the acquisition of CREC in the intestinal tract. Of all enteric bacterial groups, anaerobic bacteria seem to be the key in preventing the enteric colonization of exogenous bacteria, thus, conferring “colonization resistance” (van der Waaij et al., 1971). In contrast, animal models have suggested that disruption of the anaerobic flora by metronidazole, an antibiotic agent with potent activity against anaerobic organisms but minimal activity against other intestinal flora, may promote the overgrowth of E. coli in the gastrointestinal tract and increase the frequency of translocation (Edmond et al., 1995; Lucas et al., 1998; Wells et al., 1987; Wells et al., 1988a, 1988b). Thus, patients prone to gastrointestinal tract colonization with CREC may be placed at further risk for CREC infection by prior treatment with metronidazole. This may also explain the association of metronidazole exposure and Vancomycin-resistant Entorococcus (VRE) bacteremia (Lucas et al., 1998) and deserves further investigation. It was surprising that carbapenem resistance emerged in patients with biliary drainage catheter. After a careful review of the individual cases, we found that the majority of patients had biliary tract infection associated with unresectable malignant bile duct invasion. Although percutaneous bile duct drainage was performed in all these patients, adequate drainage was not possible in many of them. Therefore, larger inoculum of infection may contribute in acquisition of carbapenem resistance in E. coli. The identification of prior hospital stay as a risk factor is not unexpected. Patients with this risk factor may have had more chances to be exposed to additional antibiotics and to other patients with antibiotic-resistant organisms. This has been introduced as a risk factor in previous study of antibiotic-resistant organisms (Harris et al., 2002a). One limitation of our study was our inability to assess the role of patient-to-patient transmission. Because our study was retrospective in design, isolates were not available for molecular epidemiologic analysis, preventing us from determining their clonality. In addition, because the control patients were not screened for E. coli by active surveillance culture, some may actually have been case patients. However, this type of misclassification would make the case and control patients more similar. This would mean that
405
the associations observed in this study are, if anything, underestimates of the true associations (Harris et al., 2002c). Other limitations of this study are its retrospective nature and the small sample size. In summary, our results suggest that the nosocomial acquisition of CREC may be favored by the selection pressure of carbapenems and metronidazole. Moreover, CREC occurrence may also be related to prior hospital stay and the presence of biliary drainage. The effect of metronidazole and biliary drainage on the acquisition of CREC warrants further investigation. Acknowledgments The authors thank the Medical Information Team, Asan Medical Center, Seoul, Korea, for database maintenance and data extraction and Sung-Cheol Yun of Division of Biostatistics, Center for Medical Research and Information, University of Ulsan College of Medicine, Seoul, Korea, for assistance with statistical analyses. This study was supported by a grant (2002-131) from the Asan Institute of Life Science, Seoul, Korea. References CDC NNIS System (2004) National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 32:470–485. Bratu S, Brooks S, Burney S, Kochar S, Gupta J, Landman D, Quale J (2007) Detection and spread of Escherichia coli possessing the plasmidborne carbapenemase KPC-2 in Brooklyn, New York. Clin Infect Dis 44:972–975. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383. Charlson M, Szatrowski TP, Peterson J, Gold J (1994) Validation of a combined comorbidity index. J Clin Epidemiol 47:1245–1251. Edmond MB, Ober JF, Weinbaum DL, Pfaller MA, Hwang T, Sanford MD, Wenzel RP (1995) Vancomycin-resistant Enterococcus faecium bacteremia: risk factors for infection. Clin Infect Dis 20:1126–1133. Gaynes RP, Culver DH (1992) Resistance to imipenem among selected gram-negative bacilli in the United States. Infect Control Hosp Epidemiol 13:10–14. Gulmez D, Woodford N, Palepou MF, Mushtaq S, Metan G, Yakupogullari Y, Kocagoz S, Uzun O, Hascelik G, Livermore DM (2008) Carbapenemresistant Escherichia coli and Klebsiella pneumoniae isolates from Turkey with OXA-48–like carbapenemases and outer membrane protein loss. Int J Antimicrob Agents 31:523–526. Harris AD, Karchmer TB, Carmeli Y, Samore MH (2001) Methodological principles of case-control studies that analyzed risk factors for antibiotic resistance: a systematic review. Clin Infect Dis 32:1055–1061. Harris AD, Perencevich E, Roghmann MC, Morris G, Kaye KS, Johnson JA (2002a) Risk factors for piperacillin–tazobactam-resistant Pseudomonas aeruginosa among hospitalized patients. Antimicrob Agents Chemother 46:854–858. Harris AD, Samore MH, Lipsitch M, Kaye KS, Perencevich E, Carmeli Y (2002b) Control-group selection importance in studies of antimicrobial resistance: examples applied to Pseudomonas aeruginosa, enterococci, and Escherichia coli. Clin Infect Dis 34:1558–1563. Harris AD, Smith D, Johnson JA, Bradham DD, Roghmann MC (2002c) Risk factors for imipenem-resistant Pseudomonas aeruginosa among hospitalized patients. Clin Infect Dis 34:340–345.
406
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
Hong T, Moland ES, Abdalhamid B, Hanson ND, Wang J, Sloan C, Fabian D, Farajallah A, Levine J, Thomson KS (2005) Escherichia coli: development of carbapenem resistance during therapy. Clin Infect Dis 40:e84–e86. Kwak YG, Choi SH, Choo EJ, Chung JW, Jeong JY, Kim NJ, Woo JH, Ryu J, Kim YS (2005) Risk factors for the acquisition of carbapenemresistant Klebsiella pneumoniae among hospitalized patients. Microb Drug Resist 11:165–169. Lee S-O, Kim NJ, Choi S-H, Hyong Kim T, Chung J-W, Woo J-H, Ryu J, Kim YS (2004) Risk factors for acquisition of imipenem-resistant Acinetobacter baumannii: a case-control study. Antimicrob Agents Chemother 48:224–228. Lucas GM, Lechtzin N, Puryear DW, Yau LL, Flexner CW, Moore RD (1998) Vancomycin-resistant and vancomycin-susceptible enterococcal bacteremia: comparison of clinical features and outcomes. Clin Infect Dis 26:1127–1133. Miriagou V, Tzelepi E, Gianneli D, Tzouvelekis LS (2003) Escherichia coli with a self-transferable, multiresistant plasmid coding for metallo-betalactamase VIM-1. Antimicrob Agents Chemother 47:395–397. Moloughney JG, Thomas JD, Toney JH (2005) Novel IMP-1 metallo-[beta]lactamase inhibitors can reverse meropenem resistance in Escherichia coli expressing IMP-1. FEMS Microbiol Lett 243:65–71. Navon-Venezia S, Chmelnitsky I, Leavitt A, Schwaber MJ, Schwartz D, Carmeli Y (2006) Plasmid-mediated imipenem-hydrolyzing enzyme KPC-2 among multiple carbapenem-resistant Escherichia coli clones in Israel. Antimicrob Agents Chemother 50:3098–4101. National Committee for Clinical Laboratory Standards (2004) Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement. M100-S14, NCCLS; Wayne, PA.
Poirel L, Heritier C, Spicq C, Nordmann P (2004) In vivo acquisition of high-level resistance to imipenem in Escherichia coli. J Clin Microbiol 42:3831–3833. Stapleton PD, Shannon KP, French GL (1999) Carbapenem resistance in Escherichia coli associated with plasmid-determined CMY-4 betalactamase production and loss of an outer membrane protein. Antimicrob Agents Chemother 43:1206–1210. Troillet N, Samore MH, Carmeli Y (1997) Imipenem-resistant Pseudomonas aeruginosa: risk factors and antibiotic susceptibility patterns. Clin Infect Dis 25:1094–1098. Urban C, Bradford PA, Tuckman M, Segal-Maurer S, Wehbeh W, Grenner L, Colon-Urban R, Mariano N, Rahal JJ (2008) Carbapenem-resistant Escherichia coli harboring Klebsiella pneumoniae carbapenemase beta-lactamases associated with long-term care facilities. Clin Infect Dis 46:e127–e130. van der Waaij D, Berghuis-de Vries JM, Lekkerkerk L-v (1971) Colonization resistance of the digestive tract in conventional and antibiotic-treated mice. J Hyg (Lond) 69:405–411. Wells CL, Maddaus MA, Reynolds CM, Jechorek RP, Simmons RL (1987) Role of anaerobic flora in the translocation of aerobic and facultatively anaerobic intestinal bacteria. Infect Immun 55: 2689–2694. Wells CL, Jechorek RP, Maddaus MA, Simmons RL (1988a) Effects of clindamycin and metronidazole on the intestinal colonization and translocation of enterococci in mice. Antimicrob Agents Chemother 32: 1769–1775. Wells CL, Maddaus MA, Jechorek RP, Simmons RL (1988b) Role of intestinal anaerobic bacteria in colonization resistance. Eur J Clin Microbiol Infect Dis 7:107–113.
Diagnostic Microbiology and Infectious Disease 62 (2008) 402 – 406 www.elsevier.com/locate/diagmicrobio
Clinical Studies
Risk factors for the acquisition of carbapenem-resistant Escherichia coli among hospitalized patients Min-Hyok Jeona , Sang-Ho Choib,c , Yee Gyung Kwakb,c , Jin-Won Chungb,c , Sang-Oh Leeb,c , Jin-Yong Jeongb,c,d , Jun Hee Woob,c , Yang Soo Kimb,c,⁎ a
Department of Infectious Diseases, University of Soonchunhyang College of Medicine, Choongchungnam-do 336-745, Republic of Korea b Division of Infectious Diseases, Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea c Center for Antimicrobial Resistance and Microbial Genetics, University of Ulsan, Seoul 138-736, Republic of Korea d Asan Institute of Life Sciences, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea Received 5 December 2007; accepted 14 August 2008
Abstract Carbapenem resistance among Gram-negative bacilli has become an increasingly serious problem worldwide, and the emergence and spread of carbapenem-resistant Escherichia coli (CREC) is also becoming a serious problem. To date, however, risk factors for CREC acquisition have not been determined, so we decided to evaluate this in hospitalized patients through matched case-control study. Nosocomially acquired CREC was isolated from 46 patients between January 1997 and December 2007. For each patient, 3 matched-control subjects were selected. Previous use of carbapenem (adjusted odds ratio [AOR], 6.50) and metronidazole (AOR, 4.25), the presence of biliary drainage catheter (AOR, 4.59), and prior hospital stay (AOR 1.02) were found as independent risk factors for CREC. Our results suggest that the nosocomial acquisition of CREC may be favored by the selection pressure of carbapenems and metronidazole and also related to prior hospital stay and the presence of biliary drainage catheter. © 2008 Elsevier Inc. All rights reserved. Keywords: Carbapenem; Resistance; Escherichia coli
1. Introduction Carbapenems, a class of β-lactam antibiotics with a broad spectrum of antibacterial activity, have been one of the last remaining therapeutic choices for infections caused by cephalosporin-resistant microorganisms. However, their effectiveness is being challenged by the emergence and spread of carbapenem-resistant Gramnegative rods, such as Pseudomonas aeruginosa and Acinetobacter baumannii (Gaynes and Culver, 1992; Lee et al., 2004; NNIS, 2004). Recently, carbapenem resistance in Escherichia coli is also emerging worldwide (Gulmez et al., 2008; Hong et al., 2005; Miriagou et al.,
⁎ Corresponding author. Division of Infectious Diseases, Asan Medical Center, University of Ulsan College of Medicine, Seoul 138-736, Republic of Korea. Tel.: +82-2-3010-3303; fax: +82-2-3010-6970. E-mail address: [email protected] (Y.S. Kim). 0732-8893/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.diagmicrobio.2008.08.014
2003; Moloughney et al., 2005; Poirel et al., 2004; Stapleton et al., 1999; Urban et al., 2008). Mechanisms causing resistance to the carbapenems in E. coli have been suggested to include the presence of an AmpC β-lactamase in association with the loss of porin (Poirel et al., 2004; Stapleton et al., 1999), metallo-βlactamases (Miriagou et al., 2003), and class A carbapenemases such as Klebsiella pneumoniae carbapenemase type 2 (KPC-2) (Navon-Venezia et al., 2006) and KPC-3 (Hong et al., 2005). Recently, OXA-48–like carbapenemases and outer membrane protein loss were also reported (Gulmez et al., 2008). Carbapenem-resistant E. coli (CREC) strains were 1st isolated in our hospital in 1997, and since then, they are being recovered steadily. After our limited observation and clinical experience with CREC, which poses a potential threat to the hospital infection control, we assessed the risk factors for CREC acquisition in hospitalized patients.
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
403
2. Materials and methods
2.3. Statistical analysis
2.1. Case definition, control definition, and study design
Continuous variables are presented as mean ± SD, and categorical variables are presented as numbers and percentages. Univariate analyses were performed separately for each of the variables using conditional logistic regression to account for the matching between cases and controls. Variables for which the P value was b0.2 in univariate analysis or clinically plausible for the acquisition of carbapenem resistance (i.e., previous exposure to antibiotics, prior ICU stay, and time at risk) were included in a conditional logistic regression model for multivariate analysis. A backward elimination process was used, and adjusted odds ratios (AOR) and 95% confidence intervals (CIs) were calculated. P value of b0.05 was considered significant. All statistical analyses were performed using SAS release 9.1 (SAS, Cary, NC).
A case-control study was conducted at the Asan Medical Center in Seoul, Korea, a 2200-bed tertiary-care teaching hospital with 159 intensive care unit (ICU) beds and approximately 10 000 admissions per year. The use of imipenem or meropenem at this institution is not restricted to specific wards or to treatment of infections due to microorganisms resistant to other agents, but, in principle, it is monitored and supervised by infectious disease specialists. The microbiology laboratory database was searched to identify all clinical cultures positive for CREC from patients admitted between January 1997 and December 2007. Identification of E. coli and susceptibility testing were performed using the MicroScan (Dade Behring, West Sacramento, CA) system and the NEG Combo Type 21 panel (Dade Behring), using breakpoints recommended by the Clinical and Laboratory Standards Institute (formerly National Committee for Clinical Laboratory Standards) (NCCLS, 2004). Resistance to carbapenem was defined as an MIC to imipenem or meropenem exceeding 8 μg/mL, thus, including isolates of intermediate resistance (MIC, 8 μg/mL). Patients were assigned as cases if their isolates had an MIC to imipenem or meropenem ≥8 μg/mL. Nosocomial acquisition of CREC was defined as isolation N48 h after admission to the hospital. For the purpose of comparison, control patients were selected at a ratio of 3:1 from patients present in the same ward on the date that CREC was isolated, matching as closely as possible for sex, age, and admission date. When there were more than 3 subjects satisfying these conditions, an independent person chose 3 subjects randomly. 2.2. Risk factors investigated Potential risk factors for CREC were ascertained by review of medical records. Variables explored as possible risk factors included age, sex, underlying diseases and comorbid conditions, Charlson score (Charlson et al., 1994, 1987), ICU stay before nosocomial isolation of E. coli (within 1 month), prior hospital stay including transfer from another hospital in the previous year, recent surgery (within 1 month), and length of hospital stay before the outcome of interest, which we called the time at risk (Harris et al., 2001, 2002b). For case patients, time at risk was defined as length of hospital stay before CREC isolation, whereas for control patients, time at risk was defined as complete length of hospital stay. The presence of a central venous catheter, biliary drainage catheter, urinary catheter, or mechanical ventilation was also assessed. Prior antibiotic exposure was defined as at least 24 h of therapy during the 14 days before CREC isolation for case patients and before discharge for control patients (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004).
3. Results During the 11-year study period, CREC was isolated from 46 patients who met the criteria for nosocomial acquisition: 1 patient in 1998, 8 in 1999, 3 in 2000, 2 in 2001, 10 each in 2002 and 2003, 1 in 2004, 3 in 2005, 5 in 2006, and 3 in 2007. A total of 138 patients were included in the control group. Carbapenem-resistant E. coli was most frequently recovered from surgical drainage specimens (28.3%), followed by bile (21.7%), blood (13.0%), respiratory secretions (10.9%), ascitic fluid (8.7%), and urine (6.5%). Twenty-nine (63%) of 46 CREC isolates were recovered with other concomitant bacteria: 17 (37%) P. aeruginosa including 9 (19.7%) carbapenem-resistant organisms and 12 (26%) Enterococcus spp. including 3 vancomycin-resistant bacteria. One CREC isolate was recovered with both Carbepenem-resistant Pseudomonas aeruginosa (CRPA) and Enterococcus spp. On the date that a positive culture result was obtained, patients with CREC were most frequently in surgical wards (32.6%), followed by medical wards (28.3%), solid organ transplantation services (15.2%), surgical ICU (15.2%), medical ICU (4.3%), and hemato-oncology services (4.3%). Almost all (45/46) CREC isolates were resistant to 3rdgeneration cephalosporins. Of the 46 isolates, 36 (78.3%) were susceptible in vitro to amikacin, 18 (39.5%) to trimethoprim/sulfamethoxazole, 13 (28.3%) to gentamicin, and 2 (4.3%) to ciprofloxacin. Comparison of the 2 groups and results of univariate analysis regarding risk factors for nosocomial acquisition of CREC are listed in Table 1. The results of multivariable analysis are presented in Table 2. Multivariable conditional logistic regression analysis demonstrated that patients with nosocomial isolation of CREC were more likely to have been exposed to carbapenems (AOR, 6.50; 95% CI, 2.33–18.16) and metronidazole (AOR, .4.25; 95% CI, 1.56–11.59) in the 14 days before the date of a positive culture. Also, the presence of biliary drainage catheter and prior hospital stay within 1
404
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
Table 1 Univariate analysis of risk factors for the emergence of CREC Variable
Demographic characteristics Gender, male Age (mean ± SD) Related to hospitalization Prior hospital stay (b1 year) ICU stay Abdominal surgery (b30 days) Time at risk (mean ± SD)b Steroid administration Total parenteral nutrition Central catheter Mechanical ventilation Biliary drainage Urinary catheter insertion Underlying disorder Diabetes mellitus Malignancy Hepatic disease Biliary disease Renal disease Cardiac disease Pulmonary disease Renal transplantation Charlson score (median ± SD) Prior antibiotic exposure(s) Carbapenems Ampicillin/sulbactam Fluoroquinolones Piperacillin/tazobactam Cephalosporins 1st generation 2nd generation 3rd generation Aminoglycosides Metronidazole Vancomycin Both carbapenems and MTZ
Study group Casea (n = 46)
Controla (n = 138)
25 (54.3%) 56.98 ± 13.23
75 (54.3%) 58.15 ± 12.11
23 (50.0%) 24 (52.2%) 20 (43.5%) 74.34 ± 109.60 10 (21.7%) 9 (19.6%) 21 (45.7%) 12 (26.1%) 19 (41.3%) 20 (43.5%)
49 (35.5%) 61 (44.2%) 56 (40.6%) 37.41 ± 32.24 30 (21.7%) 33 (23.9%) 59 (42.8%) 38 (27.5%) 26 (18.8%) 64 (46.4%)
.082 .261 .686 .005 1.000 .523 .670 .823 .001 .674
1.85 1.60 1.17 1.02 1.00 0.76 1.20 0.91 4.85 0.84
0.93–3.69 0.70–3.66 0.54–2.59 1.01–1.03 0.41–2.43 0.32–1.79 0.52–2.76 0.38–2.18 1.85–12.75 0.37–1.91
11 (23.9%) 21 (45.7%) 19 (41.3%) 22 (47.8%) 10 (21.7%) 2 (4.3%) 4 (8.7%) 1 (2.2%) 3.72 ± 2.08
28 (20.3%) 63 (45.7%) 45 (32.6%) 38 (27.5%) 22 (15.9%) 14 (10.1%) 16 (11.6%) 3 (2.2%) 3.20 ± 2.21
.593 1.000 .209 .007 .348 .208 .565 1.000 .129
1.25 1.00 1.68 3.00 1.52 0.36 0.70 1.00 1.14
0.55–2.85 0.49–2.06 0.75–3.79 1.34–6.70 0.63–3.68 0.07–1.78 0.21–2.35 0.10–9.61 0.96–1.34
29 (63.0%) 0 (0.0%) 16 (34.8%) 3 (6.5%)
28 (20.3%) 10 (7.2%) 50 (36.2%) 8 (5.8%)
b.001 .990 .858 .827
8.14 b0.001 0.94 1.21
3.47–19.05 −INF, INF 0.46–1.90 0.22–6.65
1 (2.2%) 1 (2.2%) 28 (60.9%) 6 (13.0%) 25 (54.3%) 21 (45.7%) 16 (34.8%)
3 (2.2%) 18 (13.0%) 77 (55.8%) 20 (14.5%) 45 (32.6%) 34 (24.6%) 6 (4.3%)
1.000 .051 .495 .803 .005 .007 b.001
1.00 0.12 1.31 0.88 3.02 3.53 9.24
0.10–9.61 0.02–1.01 0.60–2.87 0.32–2.40 1.39–6.56 1.52–8.20 3.38–25.29
P
OR
95% CI
– –
– –
MTZ = metronidazole; INF = infinity. a Data are number (%) of patients unless otherwise indicated. b For cases, time at risk before isolation of CREC from clinical culture; for controls, complete length of hospital stay.
year were statistically significant risk factors associated with the acquisition of CREC (AOR, 4.59; 95% CI, 1.18–17.78; and odds ratio [OR], 1.02; 95% CI, 1.00–1.03, respectively). 4. Discussion To date, there have been a few descriptions of CREC strains, with most reports describing the mechanisms responsible for carbapenem resistance (Bratu et al., 2007; Gulmez et al., 2008; Hong et al., 2005; Moloughney et al., 2005; Navon-Venezia et al., 2006; Poirel et al., 2004; Stapleton et al., 1999). Moreover, to our knowledge, the potential risk factors for the acquisition of CREC have not been assessed, although risk factors for carbapenem resistance to P. aeruginosa, A. baumannii, and K. pneumoniae have been determined (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004; Troillet et al., 1997). Thus, this
matched case-control study is the 1st to assess potential risk factors for the acquisition of CREC in clinical specimens from hospitalized patients. Our results show that imipenem and metronidazole were the antibiotics associated with acquisition of CREC. The finding that prior use of carbapenems was associated with isolation of CREC was comparable with results showing that prior exposure to imipenem was a major risk factor for Table 2 Multivariate analysis of risk factors for the emergence of CREC Variables
Odds ratio
95% C.I.
P
Carbapenem use Metronidazole use Presence of biliary drainage catheter Prior hospital stay (b1 year)
6.50 4.25 4.59
2.33–18.16 1.56–11.59 1.18–17.78
b.001 .005 .028
1.02
1.00–1.03
.012
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
imipenem-resistant P. aeruginosa, A. baumannii, and K. pneumoniae in hospitalized patients (Harris et al., 2002c; Kwak et al., 2005; Lee et al., 2004; Troillet et al., 1997). The association between metronidazole exposure and acquisition of CREC is intriguing. Although conditional logistic regression analyses adjust for all other explanatory variables in the model, they do not necessarily identify causal associations. Thus, the association between exposure to metronidazole and acquisition of CREC may actually be due to the exposure of these patients to a number of different antibiotics before CREC is cultured. Indeed, CREC isolates were recovered most frequently from surgical drain specimens and bile, and majority (15 of 23) of patients with these isolates were exposed to metronidazole. Metronidazole may also have an indirect effect on the acquisition of CREC in the intestinal tract. Of all enteric bacterial groups, anaerobic bacteria seem to be the key in preventing the enteric colonization of exogenous bacteria, thus, conferring “colonization resistance” (van der Waaij et al., 1971). In contrast, animal models have suggested that disruption of the anaerobic flora by metronidazole, an antibiotic agent with potent activity against anaerobic organisms but minimal activity against other intestinal flora, may promote the overgrowth of E. coli in the gastrointestinal tract and increase the frequency of translocation (Edmond et al., 1995; Lucas et al., 1998; Wells et al., 1987; Wells et al., 1988a, 1988b). Thus, patients prone to gastrointestinal tract colonization with CREC may be placed at further risk for CREC infection by prior treatment with metronidazole. This may also explain the association of metronidazole exposure and Vancomycin-resistant Entorococcus (VRE) bacteremia (Lucas et al., 1998) and deserves further investigation. It was surprising that carbapenem resistance emerged in patients with biliary drainage catheter. After a careful review of the individual cases, we found that the majority of patients had biliary tract infection associated with unresectable malignant bile duct invasion. Although percutaneous bile duct drainage was performed in all these patients, adequate drainage was not possible in many of them. Therefore, larger inoculum of infection may contribute in acquisition of carbapenem resistance in E. coli. The identification of prior hospital stay as a risk factor is not unexpected. Patients with this risk factor may have had more chances to be exposed to additional antibiotics and to other patients with antibiotic-resistant organisms. This has been introduced as a risk factor in previous study of antibiotic-resistant organisms (Harris et al., 2002a). One limitation of our study was our inability to assess the role of patient-to-patient transmission. Because our study was retrospective in design, isolates were not available for molecular epidemiologic analysis, preventing us from determining their clonality. In addition, because the control patients were not screened for E. coli by active surveillance culture, some may actually have been case patients. However, this type of misclassification would make the case and control patients more similar. This would mean that
405
the associations observed in this study are, if anything, underestimates of the true associations (Harris et al., 2002c). Other limitations of this study are its retrospective nature and the small sample size. In summary, our results suggest that the nosocomial acquisition of CREC may be favored by the selection pressure of carbapenems and metronidazole. Moreover, CREC occurrence may also be related to prior hospital stay and the presence of biliary drainage. The effect of metronidazole and biliary drainage on the acquisition of CREC warrants further investigation. Acknowledgments The authors thank the Medical Information Team, Asan Medical Center, Seoul, Korea, for database maintenance and data extraction and Sung-Cheol Yun of Division of Biostatistics, Center for Medical Research and Information, University of Ulsan College of Medicine, Seoul, Korea, for assistance with statistical analyses. This study was supported by a grant (2002-131) from the Asan Institute of Life Science, Seoul, Korea. References CDC NNIS System (2004) National Nosocomial Infections Surveillance (NNIS) System Report, data summary from January 1992 through June 2004, issued October 2004. Am J Infect Control 32:470–485. Bratu S, Brooks S, Burney S, Kochar S, Gupta J, Landman D, Quale J (2007) Detection and spread of Escherichia coli possessing the plasmidborne carbapenemase KPC-2 in Brooklyn, New York. Clin Infect Dis 44:972–975. Charlson ME, Pompei P, Ales KL, MacKenzie CR (1987) A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis 40:373–383. Charlson M, Szatrowski TP, Peterson J, Gold J (1994) Validation of a combined comorbidity index. J Clin Epidemiol 47:1245–1251. Edmond MB, Ober JF, Weinbaum DL, Pfaller MA, Hwang T, Sanford MD, Wenzel RP (1995) Vancomycin-resistant Enterococcus faecium bacteremia: risk factors for infection. Clin Infect Dis 20:1126–1133. Gaynes RP, Culver DH (1992) Resistance to imipenem among selected gram-negative bacilli in the United States. Infect Control Hosp Epidemiol 13:10–14. Gulmez D, Woodford N, Palepou MF, Mushtaq S, Metan G, Yakupogullari Y, Kocagoz S, Uzun O, Hascelik G, Livermore DM (2008) Carbapenemresistant Escherichia coli and Klebsiella pneumoniae isolates from Turkey with OXA-48–like carbapenemases and outer membrane protein loss. Int J Antimicrob Agents 31:523–526. Harris AD, Karchmer TB, Carmeli Y, Samore MH (2001) Methodological principles of case-control studies that analyzed risk factors for antibiotic resistance: a systematic review. Clin Infect Dis 32:1055–1061. Harris AD, Perencevich E, Roghmann MC, Morris G, Kaye KS, Johnson JA (2002a) Risk factors for piperacillin–tazobactam-resistant Pseudomonas aeruginosa among hospitalized patients. Antimicrob Agents Chemother 46:854–858. Harris AD, Samore MH, Lipsitch M, Kaye KS, Perencevich E, Carmeli Y (2002b) Control-group selection importance in studies of antimicrobial resistance: examples applied to Pseudomonas aeruginosa, enterococci, and Escherichia coli. Clin Infect Dis 34:1558–1563. Harris AD, Smith D, Johnson JA, Bradham DD, Roghmann MC (2002c) Risk factors for imipenem-resistant Pseudomonas aeruginosa among hospitalized patients. Clin Infect Dis 34:340–345.
406
M.-H. Jeon et al. / Diagnostic Microbiology and Infectious Disease 62 (2008) 402–406
Hong T, Moland ES, Abdalhamid B, Hanson ND, Wang J, Sloan C, Fabian D, Farajallah A, Levine J, Thomson KS (2005) Escherichia coli: development of carbapenem resistance during therapy. Clin Infect Dis 40:e84–e86. Kwak YG, Choi SH, Choo EJ, Chung JW, Jeong JY, Kim NJ, Woo JH, Ryu J, Kim YS (2005) Risk factors for the acquisition of carbapenemresistant Klebsiella pneumoniae among hospitalized patients. Microb Drug Resist 11:165–169. Lee S-O, Kim NJ, Choi S-H, Hyong Kim T, Chung J-W, Woo J-H, Ryu J, Kim YS (2004) Risk factors for acquisition of imipenem-resistant Acinetobacter baumannii: a case-control study. Antimicrob Agents Chemother 48:224–228. Lucas GM, Lechtzin N, Puryear DW, Yau LL, Flexner CW, Moore RD (1998) Vancomycin-resistant and vancomycin-susceptible enterococcal bacteremia: comparison of clinical features and outcomes. Clin Infect Dis 26:1127–1133. Miriagou V, Tzelepi E, Gianneli D, Tzouvelekis LS (2003) Escherichia coli with a self-transferable, multiresistant plasmid coding for metallo-betalactamase VIM-1. Antimicrob Agents Chemother 47:395–397. Moloughney JG, Thomas JD, Toney JH (2005) Novel IMP-1 metallo-[beta]lactamase inhibitors can reverse meropenem resistance in Escherichia coli expressing IMP-1. FEMS Microbiol Lett 243:65–71. Navon-Venezia S, Chmelnitsky I, Leavitt A, Schwaber MJ, Schwartz D, Carmeli Y (2006) Plasmid-mediated imipenem-hydrolyzing enzyme KPC-2 among multiple carbapenem-resistant Escherichia coli clones in Israel. Antimicrob Agents Chemother 50:3098–4101. National Committee for Clinical Laboratory Standards (2004) Performance standards for antimicrobial susceptibility testing; fourteenth informational supplement. M100-S14, NCCLS; Wayne, PA.
Poirel L, Heritier C, Spicq C, Nordmann P (2004) In vivo acquisition of high-level resistance to imipenem in Escherichia coli. J Clin Microbiol 42:3831–3833. Stapleton PD, Shannon KP, French GL (1999) Carbapenem resistance in Escherichia coli associated with plasmid-determined CMY-4 betalactamase production and loss of an outer membrane protein. Antimicrob Agents Chemother 43:1206–1210. Troillet N, Samore MH, Carmeli Y (1997) Imipenem-resistant Pseudomonas aeruginosa: risk factors and antibiotic susceptibility patterns. Clin Infect Dis 25:1094–1098. Urban C, Bradford PA, Tuckman M, Segal-Maurer S, Wehbeh W, Grenner L, Colon-Urban R, Mariano N, Rahal JJ (2008) Carbapenem-resistant Escherichia coli harboring Klebsiella pneumoniae carbapenemase beta-lactamases associated with long-term care facilities. Clin Infect Dis 46:e127–e130. van der Waaij D, Berghuis-de Vries JM, Lekkerkerk L-v (1971) Colonization resistance of the digestive tract in conventional and antibiotic-treated mice. J Hyg (Lond) 69:405–411. Wells CL, Maddaus MA, Reynolds CM, Jechorek RP, Simmons RL (1987) Role of anaerobic flora in the translocation of aerobic and facultatively anaerobic intestinal bacteria. Infect Immun 55: 2689–2694. Wells CL, Jechorek RP, Maddaus MA, Simmons RL (1988a) Effects of clindamycin and metronidazole on the intestinal colonization and translocation of enterococci in mice. Antimicrob Agents Chemother 32: 1769–1775. Wells CL, Maddaus MA, Jechorek RP, Simmons RL (1988b) Role of intestinal anaerobic bacteria in colonization resistance. Eur J Clin Microbiol Infect Dis 7:107–113.