Skip to main content
Antimicrobial Agents and Chemotherapy logoLink to Antimicrobial Agents and Chemotherapy
. 2015 Jul 16;59(8):5049–5051. doi: 10.1128/AAC.00046-15

Risk Factors for Resistance to β-Lactam/β-Lactamase Inhibitors and Ertapenem in Bacteroides Bacteremia

Janessa M Smith a,, Edina Avdic a, Pranita D Tamma b, Long Zhang c, Karen C Carroll d, Sara E Cosgrove e
PMCID: PMC4505236  PMID: 26033720

Abstract

The objective of this study was to determine risk factors for the development of resistance to β-lactams/β-lactamase inhibitors (βL/βLIs) and ertapenem among Bacteroides species bacteremia. We conducted a retrospective case-control study of 101 adult patients with Bacteroides species bacteremia at a 1,051-bed tertiary care medical center. The duration of exposure to βL/βLIs (odds ratio [OR], 1.25; 95% confidence interval [CI], 1.08 to 2.31) was the only independent risk factor for resistance.

TEXT

The incidence of Bacteroides species bacteremia has increased over the past few decades (15). This increase is likely related to enhanced detection resulting from improved anaerobic culture methods and a more complex at-risk patient population (1). Anaerobic bacteremia has been associated with recent surgery, malignancy, and immunosuppression (1, 58).

There are 24 species of Bacteroides in the Bacteroides fragilis group, and they vary considerably in their resistance patterns although rates of resistance to β-lactams have increased among most species (24, 9, 10). B. fragilis remains the most susceptible while Bacteroides distasonis and Bacteroides thetaiotaomicron have demonstrated higher rates of resistance to β-lactams (9).

The primary objective of this study was to determine the prevalence of and risk factors for Bacteroides species with reduced susceptibility to β-lactams/β-lactamase inhibitors (βL/βLIs) and ertapenem in patients with bacteremia.

We performed a retrospective case-control study at a 1,051-bed tertiary care medical center in Baltimore, Maryland. All adult patients hospitalized between 1 January 2007 and 31 August 2013 with blood cultures growing Bacteroides species with β-lactam susceptibility testing available were eligible for inclusion. Gram-negative organisms believed to be anaerobes in positive blood culture bottles were subcultured to CDC anaerobic blood agar, laked kanamycin-vancomycin (LKV) agar, and Bacteroides bile esculin (BBE) agar and incubated under anaerobic conditions in an anaerobe chamber. Prior to 2012, organisms that grew on these media were further identified as Bacteroides species by the RapID ANA II System (Thermo Fisher Scientific, Waltham, MA, USA). Starting in 2012, organisms were identified via matrix-assisted laser desorption ionization–time of flight mass spectrometry (MALDI-TOF MS) using a Bruker microflex instrument, Biotyper software v3.0, and database v3.1.66 (Bruker Daltonics, Billerica, MA).

Antibiotic susceptibility testing was performed via Etest strips (bioMérieux). For the purpose of this study, all isolates with MICs interpreted as intermediate or resistant to amoxicillin-clavulanate, piperacillin-tazobactam, and/or ertapenem according to Clinical and Laboratory Standards Institute (CLSI) recommendations, were classified as resistant (11). For a majority of isolates, the MIC was not available, so isolates were evaluated based on their susceptibility to the respective antibiotic. Ertapenem is the only carbapenem routinely tested against anaerobes at our institution.

Patients were identified using the TheraDoc clinical surveillance software system. Cases were defined as patients with blood isolates of Bacteroides species resistant to βL/βLIs and/or ertapenem. Patients who had positive blood cultures for Bacteroides species but did not have susceptibility testing performed were excluded. Three controls were matched to each case patient by year of positive culture. A random number generator was used to select controls.

Baseline characteristics of cases and controls were compared using chi-square testing and Fisher's exact test for categorical variables, as appropriate, and the Wilcoxon rank sum test for continuous variables. Variables with a P value of ≤0.20 were entered into a multivariable logistic regression model and automatically selected using a backwards stepwise approach. Data were analyzed using Stata statistical software v12.0 (Stata Corp LP, TX). This study was approved by the Johns Hopkins University School of Medicine Institutional Review Board with a waiver of informed consent.

There were 159 patients with Bacteroides bacteremia identified during the study period. The identified isolates and their corresponding resistance rates are listed in Table 1. Of these, 26 (16.0%) patients had resistant Bacteroides. Amoxicillin-clavulanate was the most common agent to which isolates were resistant (11.5%), followed by ertapenem (7.0%) and piperacillin-tazobactam (6.8%). There were five isolates (19%) that were resistant to ertapenem but not βL/βLIs.

TABLE 1.

Resistance by species

Bacteroides speciesa No. of isolates No. (%) resistant to amoxicillin-clavulanate No. (%) resistant to piperacillin-tazobactam No. (%) resistant to ertapenem
B. caccae 12 3 (25.0) 0 0
B. distasonis 6 1 (16.0) 0 1 (16.0)
B. eggerthii 4 0 0 0
B. fragilis 71 8 (11.3) 2 (2.8) 5 (7.0)
B. loescheii 1 0 0 0
B. melaninogenicus 1 0 0 0
B. stercoris 1 0 0 0
B. thetaiotaomicron 43 3 (7.0) 6 (14.0) 2 (4.7)
B. uniformis 15 3 (20.0) 0 0
B. ureolyticus 2 0 0 0
B. vulgatus 3 0 0 0
a

A total of 159 Bacteroides bacteremia isolates were identified during the study period.

There were 101 patients included in the case-control analysis (26 case patients and 75 control patients). Only one suitable control patient was found for one of the case patients. The characteristics of these patients are summarized in Table 2. Of note, 23 (88.5%) cases and 48 (64%) controls were considered health care-associated infections (defined as positive blood culture >48 h after hospitalization or surgical procedure or hospitalization within 30 days of a positive culture). Factors associated with resistance on univariable analysis included time at risk (odds ratio [OR], 1.07; 95% confidence interval [CI], 1.02 to 1.12), health care-associated infection (OR, 4.3; 95% CI, 1.18 to 15.7), and duration of exposure to βL/βLIs (OR, 1.5; 95% CI, 0.93 to 2.45). On multivariable analysis, the only independent predictor of resistance was duration of therapy with βL/βLIs prior to infection (OR, 1.25; 95% CI, 1.08 to 2.31) and after being adjusted for time at risk.

TABLE 2.

Risk factors for resistance to Bacteroides species

Characteristicb Cases (n = 26)a Controls (n = 75)a OR (95% CI) P value
Age (yr) 60 (50–67) 61 (47–71) 1.00 (0.98–1.03) 0.79
Pitt score 2 (0–4) 1 (0–2) 1.13 (0.95–1.34) 0.16
Health care-associated infections 23 (88.5) 46 (60.5) 4.21 (1.18–15.7) 0.03
ICU admissions at time of positive culture 13 (50) 22 (29.33) 2.41 (0.96–6.02) 0.06
Source of bacteremia
    Abdominal surgery 17 (65.38) 35 (47.30) 2.05 (0.81–5.19)
    Perforation 3 (11.54) 5 (6.67) 1.83 (0.40–8.24)
    Biliary procedure 0 (0) 2 (2.67) 1.17 (0–15.29)
    SSTI 1 (3.85) 6 (8.00) 0.46 (0.05–4.01)
Concurrent Gram-negative bacteremia 4 (15.38) 4 (5.33) 2.87 (0.89–8.42) 0.12
Source controlsc 2 (7.69) 6 (8.00) 0.95 (0.88–1.08) 1.0
Comorbidity
    Diabetes 5 (19.23) 12 (16.22) 1.23 (0.39–3.90)
    GI diseased 2 (7.69) 7 (9.33) 0.81 (0.16–4.17)
    Malignancy 9 (34.62) 35 (47.30) 0.57 (0.23–1.46)
    Solid organ transplant 2 (7.69) 1 (1.33) 6.17 (0.54–71.05)
    Liver disease 4 (15.38) 4 (5.56) 3.09 (0.71–13.40)
Time at riske (days) 5.83 (3.04–16.92) 3.07 (0.80–6.61) 1.07 (1.02–1.12) 0.005
Duration of prior exposure to β-lactam (days)f
    βL/βLI exposure 13 (8–15) 5 (2–9) 1.5 (0.93–2.45) 0.008
    Ertapenem exposure 14 (2–19) 10 (7–27) 0.95 (0.81–1.12) 0.044
a

Values shown are median (IQR) or number (%).

b

βL/βLI, β-lactam/β-lactamase inhibitor; GI, gastrointestinal; ICU, intensive care unit; SSTI, skin and soft tissue infection.

c

Defined as no evidence of ongoing contamination and/or an undrained collection of infection.

d

Includes ulcerative colitis, Crohn's disease, diverticulosis.

e

The time (days) from admission to positive culture.

f

Within 90 days before positive culture.

A higher proportion of case patients died in the hospital (n = 10 [38.5%]) than control patients (n = 4 [5.3%]; P < 0.001). There was a higher percentage of concurrent Gram-negative bacteremia in case patients (n = 4 [15.3%]) than in control patients (n = 4 [5.3%]), but this difference was not statistically significant (P = 0.12).

Since we started routine antibiotic susceptibility testing in 2010, we have not documented metronidazole resistance among Bacteroides species isolates and therefore limited our evaluation to β-lactams. Resistance rates in this study were similar to what has previously been reported for amoxicillin-clavulanate but were higher for ertapenem and piperacillin-tazobactam (4, 9). Karlowsky et al. reported that 2.3%, 0.5%, and 12.7% of B. fragilis species were intermediate or resistant to ertapenem, piperacillin-tazobactam, and amoxicillin-clavulanate, respectively (4). In a report by Snydman et al., the proportions of resistant B. fragilis and B. thetaiotaomicron isolates from 2008 to 2009 to βL/βLIs and ertapenem were <5% (9). Resistance rates for Bacteroides spp. may vary among other regions, and higher resistance rates have been reported outside North America (12).

The duration of exposure to βL/βLIs was an important risk factor for resistance in our study. Each additional day of βL/βLI therapy was associated with a 25% increased risk of developing a resistant Bacteroides isolate. Similarly, Nguyen et al. found a significant association between in vitro resistance to β-lactams and previous exposure to β-lactam antianaerobic agents (e.g., ticarcillin-clavulanate, piperacillin, cefotetan, ampicillin-sulbactam) within the previous 14 days (6). We did not find exposure to metronidazole to be protective against β-lactam resistance, as more case patients were exposed to metronidazole than were control patients (19.2% versus 6.7%, respectively).

We found a higher mortality rate in patients with resistant Bacteroides species. This result was likely confounded by more severe underlying disease with a higher percentage of patients with concurrent Gram-negative bacteremia, admissions to the ICU, and longer hospital stays prior to the first positive culture in case patients. In our study, 90% of patients who died received initial appropriate therapy. Only one patient received combination antianaerobic therapy with a βL/βLI and metronidazole. Of the 10 patients with resistant Bacteroides spp. that died, 4 (40%) were treated with βL/βLIs and 4 (40%) were treated with metronidazole. The remaining 2 patients received ertapenem. Because of these small numbers, we are unable to draw any meaningful conclusions about the role of empirical combination antianaerobic therapy.

Resistance patterns vary among regions and institutions; therefore, the single center design limits the generalizability of our findings. We were limited by a small sample size, so we grouped together the βL/βLI agents for analysis. Grouping these agents together makes it difficult to determine if different risk factors for resistance to amoxicillin-clavulanate and piperacillin-tazobactam exist.

Duration of exposure to βL/βLIs is a significant risk factor for the development of resistance to these classes of agents. It is important to use these agents judiciously in an effort to decrease the development of resistance to anaerobic organisms. Routine antimicrobial susceptibility testing and the development of an antibiogram for Bacteroides species can ensure that appropriate empirical therapy is selected.

REFERENCES

  • 1.Lassmann B, Gustafson DR, Wood CM, Rosenblatt JE. 2007. Reemergence of anaerobic bacteremia. Clin Infect Dis 44:895–900. doi: 10.1086/512197. [DOI] [PubMed] [Google Scholar]
  • 2.Snydman DR, McDermott L, Cuchural GJ, Hecht DW, Iannini PB, Harrell LJ, Jenkins SG, O'Keefe JP, Pierson CL, Rihs JD, Yu VL, Finegold SM, Gorbach SL. 1996. Analysis of trends in antimicrobial resistance patterns among clinical isolates of Bacteroides fragilis group species from 1990 to 1994. Clin Infect Dis 23(Suppl):S54–S65. doi: 10.1093/clinids/23.Supplement_1.S54. [DOI] [PubMed] [Google Scholar]
  • 3.Snydman DR, Jacobus NV, McDermott LA, Ruthazer R, Goldstein EJC, Finegold SM, Harrell LJ, Hecht DW, Jenkins SG, Pierson C, Venezia R, Rihs J, Gorbach SL. 2002. National survey on the susceptibility of Bacteroides fragilis group: report and analysis of trends for 1997-2000. Clin Infect Dis 35(Suppl):S126–S134. doi: 10.1086/341934. [DOI] [PubMed] [Google Scholar]
  • 4.Karlowsky JA, Walkty AJ, Adam HJ, Baxter MR, Hoban DJ, Zhanel GG. 2012. Prevalence of antimicrobial resistance among clinical isolates of Bacteroides fragilis group in Canada in 2010-2011: CANWARD Surveillance Study. Antimicrob Agents Chemother 56:1247–1252. doi: 10.1128/AAC.05823-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Salonen JH, Eerola E, Meurman O. 1998. Clinical significance and outcome of anaerobic bacteremia. Clin Infect Dis 26:1413–1417. doi: 10.1086/516355. [DOI] [PubMed] [Google Scholar]
  • 6.Nguyen MH, Yu VL, Morris AJ, McDermott L, Wagener MW, Harrell L, Snydman DR. 2000. Antimicrobial resistance and clinical outcome of Bacteroides bacteremia: findings of a multicenter prospective observational trial. Clin Infect Dis 30:870–876. doi: 10.1086/313805. [DOI] [PubMed] [Google Scholar]
  • 7.Wilson JR, Limaye AP. 2004. Risk factors for mortality in patients with anaerobic bacteremia. Eur J Clin Microbiol Infect Dis 23:310–316. doi: 10.1007/s10096-004-1111-y. [DOI] [PubMed] [Google Scholar]
  • 8.Chen YM, Lee HC, Chang CM, Chuang YC, Ko WC. 2001. Clostridium bacteremia: emphasis on the poor prognosis in cirrhotic patients. J Microbiol Immunol Infect 34:113–118. [PubMed] [Google Scholar]
  • 9.Snydman DR, Jacobus NV, McDermott LA, Golan Y, Goldstein EJ, Harrell L, Jenkins S, Newton D, Pierson C, Rosenblatt J, Venezia R, Gorbach SL, Queenan AM, Hecht DW. 2011. Update on resistance of Bacteroides fragilis group and related species with special attention to ertapenem 2006-2009. Anaerobe 17:147–151. doi: 10.1016/j.anaerobe.2011.05.014. [DOI] [PubMed] [Google Scholar]
  • 10.Hecht DW. 2004. Prevalence of antibiotic resistance in anaerobic bacteremia: worrisome developments. Clin Infect Dis 39:92–97. doi: 10.1086/421558. [DOI] [PubMed] [Google Scholar]
  • 11.Clinical and Laboratory Standards Institute. 2012. Performance standards for antimicrobial susceptibility testing; 22nd informational supplement CLSI M100-S22. Clinical and Laboratory Standards Institute, Wayne, PA. [Google Scholar]
  • 12.Lie C-Y, Huang Y-T, Liao C-H, Yen L-C, Lin H-Y, Hsueh P-R. 2008. Increasing trends in antimicrobial resistance among clinically important anaerobes and Bacteroides fragilis isolates causing nosocomial infections: emerging resistance to carbapenems. Antimicrob Agents Chemother 52:3161–3168. doi: 10.1128/AAC.00355-08. [DOI] [PMC free article] [PubMed] [Google Scholar]

Articles from Antimicrobial Agents and Chemotherapy are provided here courtesy of American Society for Microbiology (ASM)

RESOURCES