Previous Article | Next Article ![]()
Antimicrobial Agents and Chemotherapy, June 2007, p. 1967-1971, Vol. 51, No. 6
0066-4804/07/$08.00+0 doi:10.1128/AAC.01483-06
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

Infectious Diseases Service,1 Microbiology Service, IDIBELL, Hospital de Bellvitge, University of Barcelona, Barcelona, Spain2
Received 24 November 2006/ Returned for modification 26 January 2007/ Accepted 2 April 2007
|
|
|---|
|
|
|---|
The isolation of carbapenem-resistant P. aeruginosa (CR-PA) strains because of the loss of a specific OprD porin following exposure of the microorganism to these antibiotics was already reported 2 decades ago (11). However, more recently multidrug resistance has progressively emerged in CR-PA isolates as a consequence of the development of other mechanisms of resistance such as efflux pumps or of the concomitant appearance of several mechanisms of resistance to antibiotics of different kinds (15).
While the general epidemiological characteristics of CR-PA infections appear to resemble those of nosocomial P. aeruginosa, there have been few epidemiological studies of ICUs with high rates of multidrug CR-PA (17). Determining the relevance of endogenous colonization versus cross-transmission and the influence of carbapenems and other families of antibiotics is essential to the design of targeted and efficacious strategies for infection control.
The high prevalence of CR-PA isolates in our ICUs, where they are apparently endemic, prompted us to conduct a prospective surveillance study in order to characterize the local epidemiology and to identify risk factors for CR-PA carriage.
|
|
|---|
Study design. We carried out an active surveillance program with ICU patients over two different 2-month periods (May and June of 2003 and 2004). Weekly rectal and oropharyngeal swab samples were obtained immediately on admission to detect digestive tract carriage of CR-PA between ICU admission and discharge. To study patients at risk for digestive tract carriage of CR-PA during their ICU stay, all patients admitted to the unit for more than 48 h were included.
Demographic characteristics and severity of acute illness on ICU admission were recorded by using the simplified acute physiologic score (SAPS) (14). We also assessed the devices in place, i.e., intravascular catheters, urinary catheters, and endotracheal tubes, and prior surgery. Antibiotic therapy was determined as the number of days of therapy with different groups of antibiotics for each patient since hospital admission until CR-PA colonization. The groups of antibiotics analyzed were penicillins (penicillin G, ampicillin, amoxicillin-clavulanic acid, cloxacillin, and piperacillin-tazobactam), cephalosporins (all cephalosporins and aztreonam), aminoglycosides, glycopeptides, fluoroquinolones, and carbapenems.
Antibiotic pressure was determined according to the consumption of different groups of antibiotics per month. The period studied included the 2-months previous to the surveillance program (March and April of 2003 and 2004) and antibiotic consumption during the study period (May and June of 2003 and 2004). As recommended by the Nordic Council on medicines (20), consumption was expressed as defined daily doses (DDD). During the study period, the use of antibiotics in the ICUs was not restricted, in accordance with hospital policy. However, scheduled antimicrobial cycling was used in these wards during the second period.
Microbiological surveillance and genotyping. Susceptibility studies. MICs were determined with the MicroScan automated microdilution system (Dade International, West Sacramento, CA). The antibiotics tested were piperacillin, ticarcillin, piperacillin-tazobactam, ceftazidime, cefepime, aztreonam, imipenem, meropenem, ciprofloxacin, gentamicin, tobramycin, and amikacin. The criteria of the National Committee for Clinical Laboratory Standards (19) were used to define susceptibility or resistance to these antimicrobial agents. Intermediate isolates were considered resistant.
CR-PA colonization detection. Rectal and oropharyngeal swabs were inoculated onto MacConkey agar plates. The plates were incubated at 37°C for 48 h. Gram-negative, oxidase-positive bacilli growing on MacConkey agar were identified as P. aeruginosa by conventional biochemical tests.
Pulsed-field gel electrophoresis (PFGE). Chromosomal DNAs from all CR-PA and earlier CS-PA digestive tract samples from ICU patients were prepared for PFGE analysis as previously described (23) and then digested with XbaI. DNA fragments were separated with a CHEF DR III apparatus (Bio-Rad Laboratories, Hercules, CA). Electrophoresis was run at 6 V/cm and 14°C for 23 h with pulses ranging from 5 to 25 s. DNA restriction patterns generated by PFGE were interpreted according to the guidelines proposed by Tenover et al. (24).
Definitions. Carriage was defined as at least one positive culture obtained from an oropharyngeal or rectal swab. CR-PA carriage on admission was defined as a positive swab culture within 48 h of ICU admission. Colonizations occurring more than 48 h after ICU admission were defined as ICU acquired.
We used the following criteria to provide an epidemiological classification of the colonization routes in the digestive tract. (i) Endogenous colonization was considered if the finding of CR-PA was preceded by isolation of carbapenem-susceptible P. aeruginosa (CS-PA) in the same patient and if the strains showed identical clonal patterns. (ii) Exogenous colonization or cross-transmission was considered if CR-PA isolates had a PFGE pattern similar to that of isolates from at least one other patient present in either ICU. (iii) Indeterminate colonization was considered if the patient did not have endogenous colonization according to the above criteria and if the colonization involved a CR-PA strain not identified in other patients by PFGE.
In addition, two patterns of carbapenem resistance were differentiated in CR-PA isolates. (i) An imipenem resistance (IR) pattern was considered when strains showed resistance to imipenem along with an increase in the MICs of meropenem but within the susceptibility range. (ii) A multiresistance (MR) pattern was considered if strains showed resistance to imipenem and meropenem and concomitantly to at least two other families of antibiotics.
Statistical analysis. Contingency tables were analyzed by two-tailed chi-square test. Continuous variables were compared by the t test. Univariate analysis was performed to determine the significance of risk factors for CR-PA digestive tract carriage acquisition. P < 0.05 was considered statistically significant. A multivariate analysis was performed to assess the independence of statistically significant variables in the univariate analysis by an unconditional stepwise logistic regression model. The probability of carriage of CR-PA in the digestive tract was calculated by using the Kaplan-Meier estimate. Statistical analysis was performed with the SPSS/PC version 11.0 statistical package.
|
|
|---|
The remaining 254 patients were included in the study cohort, 119 patients from May and June of 2003 and 135 patients from May and June of 2004. The epidemiological characteristics in the two periods were comparable, although patients in the first period had received more prior antibiotics than patients in the second period (87 [84%] versus 84 [62%]; P < 0.01), specifically, more carbapenems (43 [36%] versus 21 [15.5%]; P < 0.01).
Risk factors for CR-PA carriage. Thirty digestive tract CR-PA carriers were colonized during the two study periods, 17 (14%) out of 119 patients during the first period and 13 (10%) of 135 patients during the second period. The probabilities of CR-PA carriage in the digestive tract were 10 and 8%, respectively, 10 days (P = 0.8) after ICU admission (Fig. 1).
![]() View larger version (13K): [in a new window] |
FIG. 1. Probability of remaining free of digestive tract CR-PA colonization. Thin line, first period; thick line, second period.
|
The variables associated with CR-PA carriage are listed in Table 1. CR-PA carriers were more likely to have received antibiotics than non-CR-PA carriers (93 versus 64%). In addition, these CR-PA carriers were more likely to have been exposed to carbapenems (63 versus 20%) and fluoroquinolones (23 versus 7%). A logistic regression model with CR-PA digestive tract colonization as the dependent variable identified the severity of acute illness according to SAPS (odds ratio [OR], 1.0; 95% confidence interval [CI], 1.0 to 1.1), prior carbapenem use (OR, 7.8; 95% CI, 1.7 to 35.3), and prior use of fluoroquinolones (OR, 11.0; 95% CI, 1.7 to 67.9) as independent risk factors.
|
View this table: [in a new window] |
TABLE 1. Univariate analysis of risk factors for digestive tract CR-PA colonization of ICU patients
|
|
View this table: [in a new window] |
TABLE 2. Activities of antibiotics against CR-PA isolates
|
Acquisition of CR-PA carriage. Among the 30 colonized patients with ICU acquisition, 15 genotypes were found in a single patient only, whereas 1 genotype, B, was isolated from 5 patients, 2 other genotypes (J and L) were isolated from 4 and 2 patients, respectively, and isolates were nontypeable in four cases (Table 3).
|
View this table: [in a new window] |
TABLE 3. Routes of CR-PA colonization of ICU patients
|
Among the 12 CR-PA digestive tract colonizations before ICU admission, isolates were also typed by PFGE, which produced seven distinct genotypes. Six genotypes were isolated from individual patients only, and the remaining genotype (B) colonized six patients. Epidemiological analysis of these six patients with genetically related strains showed that none had previously been admitted to ICUs and only two patients presented some evidence of temporal clustering in the surgical ward.
|
|
|---|
In our study, acquisition of digestive tract CR-PA colonization was associated with similar risk factors already identified for carriage of other resistant bacteria in ICUs, but only severity of acute illness and, in particular, prior carbapenem and prior fluoroquinolone consumption presented independent associations.
Analysis of colonization routes has several potential limitations. First, patients may be colonized with multiple genotypes of P. aeruginosa; as only a single isolate was included in our genotyping analysis, it may not accurately represent the whole epidemiology. In addition, the stratification we used may have introduced a bias, and cultures of environmental sites and items from these ICU wards were not available. On balance, however, we think that it was a useful approach in this epidemiologic setting.
Colonization was endogenous in 33% of our patients, who showed pairs of clonally identical CS-PA and CR-PA isolates. All had been exposed to carbapenem therapy, and 80% showed an IR phenotype pattern. The risk of emergence of drug-resistant P. aeruginosa in the course of or after using antibiotics with high antipseudomonal activity, especially carbapenems, is well recognized (7, 12, 25).
On the other hand, 30% of our CR-PA carriers had an exogenous source and three clones were identified. Interestingly, we observed cross-transmission of clones J and L, whose index cases were among the endogenously acquired cases. Overall, the MR pattern prevailed in exogenously acquired colonizations; similar data were observed in another study (17), where 64% of the multidrug-resistant P. aeruginosa strains had been potentially transmitted via cross-colonization.
Although multiresistant P. aeruginosa strains have emerged worldwide in recent years, their prevalence and epidemiology have not yet been well established. An MR pattern was observed in 60% of our patients. Previous exposure to carbapenems was found in about 50% of these patients, and the influence of other antibiotics such as fluoroquinolones was also established. It has recently been postulated that fluoroquinolones may induce the expression of multidrug efflux pumps, producing an MR P. aeruginosa phenotype (1, 2). The mechanisms of resistance in our patients were not studied, but five of the seven patients with CR-PA colonization who had been exposed to fluoroquinolones showed an MR pattern.
The evaluation of patients showing colonization at ICU admission deserves some comment. While more than half of the patients colonized on admission had a diversity of multiresistant clones which were not found during the study in patients with colonization acquired in ICUs, six patients showed the same multiresistant clone, B, which was responsible for five additional cases of colonization acquisition in the ICU. This alerted us to the fact that, concomitantly with our endemic CR-PA, our ICUs and other wards of the hospital also face a limited outbreak caused by this multiresistant clone.
Leaving aside the clone B outbreak, the CR-PA digestive tract colonization endemic in our ICUs was characterized by a polyclonal map of strains. The causes of the emergence of endogenous resistance in patients previously colonized by CS-PA may be considered the large-scale use of carbapenems and also the selection of intestinal flora, making the host more susceptible to colonization by resistant strains. In parallel, widespread use of fluoroquinolones may induce efflux systems and therefore render P. aeruginosa less susceptible to antibiotics (8, 13, 21). These conditions also increase the number of colonized patients and consequently the pressure of colonization, a circumstance that has been demonstrated to predispose to higher cross-transmission (5).
While maintaining vigilance and avoiding cross-transmission in our ICUs is mandatory, antibiotic control policies must be revised in the setting of endemic colonization or infection with CR-PA. Our study provides strong arguments for limiting carbapenem use and maybe fluoroquinolone use in the antibiotic policies of ICUs in order to prevent the worrying problem of emerging CR-PA. Because a program of extensive antimicrobial restriction can yield the potential development of new and possibly unexpected resistance patterns, efforts to encourage heterogeneity through individualization of drug use should be recommend.
Published ahead of print on 9 April 2007. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»