Abstract
Among 477 patients with susceptible Enterobacter spp., 49 subsequently harbored third-generation cephalosporin-resistant Enterobacter spp. Broad-spectrum cephalosporins were independent risk factors for resistance (relative risk [OR] = 2.3, P = 0.01); quinolone therapy was protective (OR = 0.4, P = 0.03). There were trends toward decreased risk for resistance among patients receiving broad-spectrum cephalosporins and either aminoglycosides or imipenem. Of the patients receiving broad-spectrum cephalosporins, 19% developed resistance.
Enterobacter spp. are among the most common gram-negative pathogens associated with hospital infections, representing 6% of all nosocomial isolates recovered and 11% of pneumonia isolates (8).
Resistance to β-lactam antibiotics often complicates the treatment of Enterobacter infections (2, 6). In a recent report, 36% of Enterobacter spp. infections in intensive care units were resistant to broad-spectrum cephalosporins (9). Most commonly, resistance to third-generation cephalosporins in this organism is mediated by chromosomal AmpC cephalosporinase. Although isolates often appear susceptible in vitro, antibiotic pressure can facilitate the emergence of derepressed mutant Enterobacter cells which produce AmpC β-lactamases at high levels constitutively (12; A. A. Medeiros, Editorial, Clin. Infect. Dis. 25:341–342, 1996). In addition, exposure to certain β-lactam antibiotics results in increased synthesis of AmpC and induction of resistance to broad-spectrum cephalosporins.
A landmark study by Chow et al. showed a strong correlation between previous broad-spectrum cephalosporin exposure and the isolation of Enterobacter spp. resistant to these agents (2). Although the study demonstrated emergence of resistance during therapy of Enterobacter bacteremia with broad-spectrum cephalosporins in 19% of patients, the prospective nature of this study precluded a large enough sample size on which conclusive analysis of risk factors for emergence of resistance could be performed (only six patients showed the emergence of a resistant strain). No study to date has had enough statistical power to study this question comprehensively. We applied here effective analytical methods to a large study cohort to measure the effects of antimicrobial agent exposures on the emergence of broad-spectrum cephalosporin resistance among Enterobacter spp. We analyzed antimicrobial risk factors as time-dependent variables (7) so that risk estimates would account for the duration of time an individual was exposed to an antimicrobial agent only after therapy with the agent had commenced, thus decreasing the potential for bias.
The study design was a retrospective cohort. All patients admitted to Beth Israel Deaconess Medical Center, West Campus, Boston, Mass., between October 1993 and September 1997 with cultures positive for Enterobacter spp. susceptible to broad-spectrum cephalosporins were included in the study. Patients remained in the cohort until Enterobacter spp. resistant to broad-spectrum cephalosporins were isolated (this was the outcome of interest) or until hospital discharge or death. Data were collected from administrative, laboratory, and pharmacy databases. Antibiotics analyzed included narrow-, expanded-, and broad-spectrum cephalosporins (ceftriaxone and ceftazidime were the only broad-spectrum cephalosporins used during the study period); ampicillin; penicillin; aminoglycosides; quinolones; imipenem; piperacillin; ampicillin-sulbactam; and piperacillin-tazobactam. During the course of this study, other agents such as cefepime and meropenem were rarely used at our institution.
Statistical analyses were performed using the SAS software (SAS Institute, Cary, N.C.) system for Windows. Cox proportional-hazard models were used to analyze time-dependent variables and to account for variable durations of time spent in the cohort by study patients.
A total of 477 patients with broad-spectrum cephalosporin-susceptible Enterobacter spp. satisfied the criteria for entry into the cohort. Forty-nine patients (10% of the cohort) had broad-spectrum cephalosporin-resistant Enterobacter spp. isolated subsequently. Among the initial strains susceptible to broad-spectrum cephalosporins, 343 E. cloacae isolates, 108 E. aerogenes isolates, and 26 other Enterobacter spp. isolates were identified. Resistance emerged subsequently in only two species: in 31 of 343 of the patients with initial E. cloace isolates (9%) and in 18 of 108 of the patients with initial E. aerogenes isolates (17%) (P = 0.03). Among patients in whom resistance emerged, species of resistant and susceptible isolates were identical in all but three patients.
The sites of initial Enterobacter isolation are listed in Table 1.
TABLE 1.
Sites of initial isolation of Enterobacter spp.a
| Site | No. (%) |
|---|---|
| Wounds | 174 (36.5) |
| Respiratory secretions | 119 (24.9) |
| Urine | 102 (21.4) |
| Effusions | 41 (8.6) |
| Blood | 26 (5.5) |
| Vascular line tips | 8 (1.7) |
| Tissue | 7 (1.5) |
Among patients with the emergence of resistant Enterobacter spp., susceptible and resistant isolates were recovered from the same anatomic site in all but nine patients. In six of these patients, blood was one site of isolation and the other anatomic sites likely represented the source of bacteremia (four pulmonary, one effusion, and one urine). In the other three patients, the two Enterobacter isolates were recovered from anatomically distinct sites and might have been spread via self-inoculation.
Descriptive patient characteristics and crude results are shown in Table 2. Exposure to broad-spectrum cephalosporins was a risk factor for the emergence of resistant Enterobacter spp. (relative risk [RR] = 3.3, P < 0.001). Nineteen percent of the patients initially treated with these agents subsequently showed the emergence of a resistant Enterobacter isolate. Among patients treated with broad-spectrum cephalosporins, resistance emerged significantly more frequently when the initial site of isolation was the blood (4 of 14, 29%) than if the initial site was urine, tissue or wounds (5 of 67, 7%) (P = 0.04). Exposure to quinolones was associated with a decreased risk for emergence of broad-spectrum cephalosporin-resistant Enterobacter spp. (RR = 0.4, P = 0.03).
TABLE 2.
Descriptive characteristics of cohort and univariate analysis
| Variable (U)a | Value for patients with (n = 49) or without (n = 428) emergence of Enterobacter spp. resistant to broad-spectrum cephalosporins
|
RR (95% CI)b | P | |
|---|---|---|---|---|
| With | Without | |||
| Demographics | ||||
| Mean age in yr (SD) | 63.8 (11.5) | 63.5 (14.9) | NA | 0.85 |
| Male gender (%) | 31 (63.3) | 240 (56.1) | 1.3 (0.7–2.3) | 0.37 |
| Comorbidities | ||||
| Liver disease (%) | 13 (26.5) | 56 (15.1) | 2.7 (1.4–5.1) | 0.003 |
| Lung disease (%) | 9 (18.4) | 47 (11.0) | 1.3 (0.6–2.7) | 0.48 |
| Renal disease (%) | 10 (20.4) | 82 (19.2) | 0.8 (0.4–1.5) | 0.44 |
| Cardiovascular disease (%) | 38 (77.6) | 334 (78.0) | 0.8 (0.4–1.5) | 0.44 |
| Cancer (%) | 13 (26.5) | 73 (17.1) | 1.7 (0.9–3.2) | 0.10 |
| Diabetes (%) | 17 (34.7) | 231 (54.0) | 0.5 (0.3–0.9) | 0.02 |
| AIDS (%) | 0 (0.0) | 7 (1.6) | ND | 1.00 |
| Transplant (%) | 4 (8.2) | 21 (4.9) | 1.1 (0.4–1.4) | 0.95 |
| Median no. of patient comorbidities (IQR) | 2 (1–3) | 2 (1–2) | ||
| Hospital exposures | ||||
| Median length of hospital stay in days before entry into cohort (IQR) | 7.0 (5–12) | 4.0 (2–11) | NA | 0.005 |
| Median no. days in cohort (IQR) | 8 (4–14) | 7 (4–13) | NA | 0.79 |
| Transfer from other institution (%) | 15 (30.6) | 113 (26.4) | 1.1 (0.6–2.0) | 0.87 |
| ICU stay (%) | 40 (58.0) | 185 (43.2) | 2.8 (1.4–5.5) | 0.003 |
| Surgical procedure (%) | 38 (77.6) | 208 (48.6) | 2.6 (1.3–5.0) | 0.006 |
| Median culture score (1) (IQR)c | 1.4 (0.9–1.9) | 0.6 (0.3–1.0) | 3.6 (1.8–7.1)c | <0.001d |
| Exposures to antimicrobial agents | ||||
| Ampicillin and penicillin (%) | 22 (44.9) | 184 (43.0) | 1.0 (0.5–1.7) | 0.90 |
| Narrow-spectrum cephalosporins (%) | 12 (24.5) | 115 (26.9) | 0.9 (0.5–1.8) | 0.79 |
| Expanded-spectrum cephalosporins (%) | 3 (6.1) | 31 (7.2) | 1.0 (0.3–3.2) | 0.93 |
| Broad-spectrum cephalosporins (%) | 31 (63.3) | 130 (30.4) | 3.3 (1.8–6.0) | <0.001 |
| Aminoglycosides (%) | 10 (20.4) | 68 (15.9) | 0.9 (0.4–1.9) | 0.73 |
| Quinolones (%) | 6 (12.2) | 116 (27.1) | 0.4 (0.1–0.9) | 0.03 |
| Imipenem (%) | 8 (16.3) | 35 (8.2) | 1.0 (0.5–2.3) | 0.97 |
| Piperacillin (%) | 8 (16.3) | 55 (12.9) | 1.0 (0.5–2.2) | 0.98 |
| Ampicillin-sulbactam (%) | 13 (26.5) | 122 (28.5) | 1.0 (0.5–1.8) | 0.87 |
| Piperacillin-tazobactam (%) | 5 (10.2) | 35 (8.2) | 1.1 (0.4–2.7) | 0.90 |
Units apply to data in columns two and three only. IQR, interquartile range; ICU, intensive care unit.
CI, confidence interval; ND, not determinable; NA, not applicable.
The culture score is the average number of cultures obtained per day.
Analyzed as a dichotomous variable.
In multivariate analysis (Table 3), exposure to broad-spectrum cephalosporins remained a strong, independent predictor for the emergence of resistant Enterobacter spp. (RR = 2.3, P = 0.01), and the association between quinolone exposure and a decreased risk for emergence of resistance was unchanged (RR = 0.4, P = 0.03).
TABLE 3.
Multivariate analysis
| Variable | RR (95% CI)a | P |
|---|---|---|
| Broad-spectrum cephalosporins | 2.3 (1.2–4.3) | 0.01 |
| Quinolones | 0.4 (0.1–0.9) | 0.03 |
| Culture scoreb | 3.0 (1.5–6.1) | 0.002 |
| Liver disease | 1.7 (0.9–3.3) | 0.10 |
CI, confidence interval.
See Table 2, footnote c. The score was analyzed as a dichotomous variable.
When patients treated with broad-spectrum cephalosporins received either imipenem or aminoglycosides, the risk for emergence of resistance tended to decrease. These effects were not statistically significant (RR = 0.5, P = 0.38 and RR = 0.5, P = 0.32, respectively).
This was the first analytical study to comprehensively investigate risk factors for the emergence of resistance to broad-spectrum cephalosporins among Enterobacter spp. Isolates were not available for molecular typing, and it is possible that some resistant strains might have been different clones than the initial susceptible isolates. However, species of susceptible and resistant isolates were identical in >90% of cases, and the site of isolation was the same in >80% of cases. Thus, extrapolating from the molecular analyses of Chow et al., it is likely that susceptible and resistant isolates were of the same clone in most instances (2). Therapy with broad-spectrum cephalosporins was a strong risk factor for the emergence of Enterobacter spp. resistant to these agents. Similar relationships have been demonstrated in vitro and in the clinical setting (2, 10, 11). Interestingly, exposures to narrow- and expanded-spectrum cephalosporins were not associated with the emergence of resistance, nor was exposure to ureidopenicillins or to β-lactam–β-lactamase inhibitor combination agents. This finding is in accord with some studies (2, 6), but not with others (4, 5). Quinolone therapy was associated with decreased risk for emergence of broad-spectrum cephalosporin-resistant Enterobacter spp. This important association has not been previously demonstrated. We did not study the frequency of quinolone resistance. We detected a trend toward a protective effect when patients received both broad-spectrum cephalosporins and either aminoglycosides or imipenem, but this was not statistically significant. These findings deserve further investigation.
Our study suggests that if broad-spectrum cephalosporins are used to treat patients with Enterobacter-positive isolates, approximately 19% will develop resistance. This number was identical to the frequency of emergence of resistance among patients with Enterobacter spp. bacteremia reported by Chow et al. (2). However, our study found rates of emergence of resistance to be significantly higher in patients with Enterobacter bacteremia than in patients whose isolates were initially recovered from tissue, wounds, or urine. Also, resistance occurred more frequently among E. aerogenes than E. cloacae. Because the emergence of resistance to broad-spectrum cephalosporins in Enterobacter spp. is associated with adverse clinical outcomes (3), knowledge of the specific risk factors for resistance should aid in the selection of appropriate antibiotic therapy.
REFERENCES
- 1.Carmeli Y, Troillet N, Eliopoulos G M, Samore M H. Emergence of antibiotic-resistant Pseudomonas aeruginosa: comparison of risks associated with different antipseudomonal agents. Antimicrob Agents Chemother. 1999;43:1379–1382. doi: 10.1128/aac.43.6.1379. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chow J W, Fine M J, Shlaes D M, Quinn J P, Hooper D C, Johnson M P, Ramphal R, Wagener M M, Miyashiro D K, Yu V L. Enterobacter bacteremia: clinical features and emergence of antibiotic resistance during therapy. Ann Intern Med. 1991;115:585–590. doi: 10.7326/0003-4819-115-8-585. [DOI] [PubMed] [Google Scholar]
- 3.Cosgrove, S. E., K. S. Kaye, G. M. Eliopoulos, and Y. Carmeli. Impact of emergence of third-generation cephalosporin resistance in Enterobacter species on patient outcomes: mortality, length of stay, and hospital cost. Arch. Intern. Med., in press. [DOI] [PubMed]
- 4.D'Agata E M, Venkataraman L, DeGirolami P, Burke P, Eliopoulos G M, Karchmer A W, Samore M H. Colonization with broad-spectrum cephalosporin-resistant gram-negative bacilli in intensive care units during a nonoutbreak period: prevalence, risk factors, and rate of infection. Crit Care Med. 1999;27:1090–1095. doi: 10.1097/00003246-199906000-00026. [DOI] [PubMed] [Google Scholar]
- 5.Flynn D M, Weinstein R A, Kabins S A. Infections with gram-negative bacilli in a cardiac surgery intensive care unit: the relative role of enterobacter. J Hosp Infect. 1988;11(Suppl A):367–373. doi: 10.1016/0195-6701(88)90212-5. [DOI] [PubMed] [Google Scholar]
- 6.Jacobson K L, Cohen S H, Inciardi J F, King J H, Lippert W E, Iglesias T, VanCouwenberghe C J. The relationship between antecedent antibiotic use and resistance to extended-spectrum cephalosporins in group I beta-lactamase-producing organisms. Clin Infect Dis. 1995;21:1107–1113. doi: 10.1093/clinids/21.5.1107. [DOI] [PubMed] [Google Scholar]
- 7.Kleinbaum D G. Survival analysis: a self-learning text. New York, N.Y: Springer-Verlag, Inc.; 1996. [Google Scholar]
- 8.National Nosocomial Infections Surveillance System. A report from the National Nosocomial Infections Surveillance (NNIS) System. Am J Infect Control. 1996;24:380–388. [PubMed] [Google Scholar]
- 9.National Nosocomial Infections Surveillance System. National Nosocomial Infections Surveillance (NNIS) System report, data summary from January 1990–May 1999, issued June 1999. Am J Infect Control. 1999;27:520–532. doi: 10.1016/s0196-6553(99)70031-3. [DOI] [PubMed] [Google Scholar]
- 10.Sanders C C, Sanders W E., Jr Microbial resistance to newer generation beta-lactam antibiotics: clinical and laboratory implications. J Infect Dis. 1985;151:399–406. doi: 10.1093/infdis/151.3.399. [DOI] [PubMed] [Google Scholar]
- 11.Sanders C C, Sanders W E., Jr Type I beta-lactamases of gram-negative bacteria: interactions with beta-lactam antibiotics. J Infect Dis. 1986;154:792–800. doi: 10.1093/infdis/154.5.792. [DOI] [PubMed] [Google Scholar]
- 12.Sanders W E, Jr, Sanders C C. Enterobacter spp.: pathogens poised to flourish at the turn of the century. Clin Microbiol Rev. 1997;10:220–241. doi: 10.1128/cmr.10.2.220. [DOI] [PMC free article] [PubMed] [Google Scholar]