Abstract
Extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBLPE) were isolated from clinical specimens from 130 to 140 patients/year in 1989–1991 in our hospital. In February 1992, a control program was initiated: screening tests in 3 intensive care units (ICUs) and contact-isolation precautions in all units. The septic surgical unit served as an isolation ward for surgical patients from whom ESBLPE was isolated. In 1992, the incidence of ESBLPE acquisition failed to decrease, and most acquisitions occurred in 3 ICUs. Critical evaluation of implementation of isolation procedures in these ICUs prompted corrective measures for barrier precautions. The incidence of acquired cases subsequently decreased, and a second evaluation determined that these measures had been correctly applied. The incidence of acquired cases in the septic surgical unit was lower than those in the other units. Decreases were also found in the incidence of acquisition of other hand-transmitted multidrug-resistant organisms. Barrier precautions, screening tests for ICU patients, and grouping of cohorts after ICU discharge are effective in controlling the spread of multidrug-resistant microorganisms by cross-contamination. The outbreak was effectively controlled without restricting antimicrobial use.
Since extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBLPE) were first recognized in 1983 [1], many outbreaks have been reported in France [2–9], other European countries [10–16], and North America [17–23]. These multidrug-resistant organisms have now spread worldwide [24–27]. ESBLPE were first isolated in intensive care units (ICUs) and subsequently spread to other units, most notably extended-care facilities. In 1993, 8.7% of Klebsiella pneumoniae isolates reported to the Centers for Disease Control and Prevention (CDC) were suspected to harbor extended-spectrum β-lactamases [23]. In France, multicenter surveys done in 1988–1990 and 1993 found that 13.1% and 24.8% of K. pneumoniae strains isolated from hospitalized patients, respectively, produced extended-spectrum β-lactamases [28, 29].
Despite growing concern about ESBLPE, few studies have reported successful interventions to control its dissemination. Two types of interventions for infection control have been proposed. First, because there were reports of an association between increased antimicrobial use and spread of ESBLPE, several studies were initiated to examine restricted use of antimicrobials (particularly third-generation cephalosporins), the goal of which was to decrease their selection effect [17–21]. Second, barrier precautions designed to prevent transmission via the hands of health care workers were recommended [13, 30, 31]. Most studies, however, used both types of interventions for infection control [19–21], making it difficult to differentiate the respective roles of the selection effect of antimicrobials and cross-contamination. We report successful control of a long-lasting ESBLPE outbreak in a university hospital by using isolation measures alone, without restricting antimicrobial use.
Patients and Methods
Background. The Bichat-Claude Bernard Hospital in Paris is a 1200-bed public hospital serving both as a primary and tertiary referral center; there are >30,000 admissions/year. The hospital includes 5 ICUs: a cardiac ICU, a surgical cardiac ICU, an 18-bed infectious diseases ICU, a 17-bed medical ICU, and a 10-bed surgical ICU. The latter 3 (hereafter referred to as “the 3 ICUs”) admit patients referred from other ICUs or institutions; their total number of admissions/year is ∼1200. There are 19 other acute care units (including a 28-bed septic surgical ward) and 205 beds for intermediate and extended care.
ESBLPE were first isolated in our hospital in 1985 [4]. Barrier precautions were not or were only partially implemented during the next 5 years. Clinical isolates of ESBLPE were recovered from 130 to 140 patients each year from 1989 to 1991 (incidence, 0.43–0.47 cases/100 admissions).
Infection control program. A program to minimize ESBLPE dissemination was started in February 1992. All inpatients from whom ESBLPE were isolated were prospectively investigated and followed up until discharge. The following data were collected: basic demographic characteristics, location before admission, date and unit of admission, date(s) and unit(s) of transfer(s) within the hospital, date of first isolation of ESBLPE, site(s) of colonization and infection, date of discharge from the hospital, and destination after discharge. ESBLPE infections were defined according to standard criteria [32]. In addition, quantitative criteria were required to define pulmonary infection (at least 103 cfu/mL in cultures of protected specimen brushes [33]) and catheter-related infection (at least 103 cfu/mL in semiquantitative cultures of catheter tips [34]). Colonization was defined as the presence of ESBLPE in patients who failed to meet criteria for ESBLPE infection. Throughout the 4-year study period, routine rectal swabs and urine specimens for culture were obtained at admission and weekly from all patients in the 3 ICUs and in the rehabilitation unit, as well as patients sharing rooms with, or in rooms adjacent to, patients in other units from whom ESBLPE were isolated.
Barrier precautions similar to recent CDC guidelines for contact isolation [35] were recommended for patients colonized or infected with ESBLPE. Briefly, these measures consisted of marking microbiological reports, charts, and doors of rooms of ESBLPE-positive patients with a symbol; wearing gowns and gloves when caring for these patients; hand washing with antiseptic soap; and reporting ESBLPE positivity when patients were transferred to another unit. Preventive isolation precautions were also recommended for patients transferred to the 3 ICUs from another unit where ESBLPE were known to be present, particularly if this unit was another ICU or a rehabilitation unit. Isolation measures were maintained pending results of screening tests performed at admission. In addition, selective digestive decontamination was used during the first year of the intervention for infection control in the 3 ICUs and in the rehabilitation unit [36, 37]. Data for 2 of the 3 ICUs have been reported elsewhere and will not be discussed here [38, 39].
On the basis of results of routine surveillance cultures and clinical surveillance data, cases of colonization or infection with ESBLPE were classified as imported to or acquired in our institution. Patients referred to the hospital were those who met any of the following criteria: ESBLPE isolated at admission, ESBLPE isolated from an infected site within 72 h after admission, or history of colonization and/or infection with ESBLPE. Other cases were considered acquired.
Before starting the infection control program, meetings were arranged with medical and nursing staff of all clinical and technical units of the hospital. These meetings emphasized the importance of hand washing. Subsequently, ICU staff met at 3- to 4-month intervals. Updates on ESBLPE spread in the hospital were sent to the other units every 6 months. When a new case of infection or colonization with ESBLPE was diagnosed outside the ICUs, an infection control practitioner visited the unit where the patient was hospitalized and helped institute isolation precautions. Patients with ESBLPE who were discharged from the surgical ICU or identified in surgical wards were transferred to the septic surgical unit. Nurses in this unit, however, also cared for ESBLPE-negative patients.
Microbiological studies. Routine specimens were cultured on a selective medium (Drigalski lactose agar; Sanofi-Pasteur, Marnes-la-Coquette, France) supplemented with cefotaxime (0.5 mg/L). ESBLPE isolated from clinical and routine specimens were studied prospectively. Susceptibility to antimicrobials was tested by using a standard agar diffusion method with use of Mueller-Hinton agar (Sanofi-Pasteur) [40]. Double-disk diffusion testing of routine or clinical specimens detected production of extended-spectrum β-lactamase by a synergistic effect between clavulanic acid/amoxicillin and cefotaxime, ceftazidime, or aztreonam [3, 5].
Data on other multidrug-resistant organisms were obtained by yearly review of computerized databases of the microbiology laboratory. We recorded the number of patients with ≥1 clinical isolates (excluding screening specimens) of methicillin-resistant Staphylococcus aureus (MRSA), ticarcillin-resistant Pseudomonas aer-u-ginosa, Acinetobacter baumannii, and Enterobacteriaceae that were resistant to third-generation cephalosporins as a result of production of class C derepressed mutants (Enterobacter species, Citrobacter species, and Serratia marcescens) [41].
Evaluation of compliance with infection control procedures. During 5 consecutive days in January 1993 and June 1994, compliance with isolation procedures was evaluated in the 3 ICUs by a group of nurses and head nurses from these units. They first received 4 h of training on observation techniques and developed a standardized evaluation sheet, with the help of infection control practitioners. Evaluations in each of the 3 ICUs were performed by unobtrusive observation by nurses from 1 of the 2 other ICUs. In each ICU, 1 or several rooms occupied by patients positive or negative for ESBLPE were observed for 2 h on 8 different occasions during each 5-day period. Contacts with patients, body fluids, and the environment of each person entering the room were recorded. The following data were recorded: reason for each person entering the room, whether the symbol indicating isolation was on the room door, whether each person washed his or her hands before and after contacts in the room, how hand washing was performed, and whether recommendations for glove and gown use were followed.
Antimicrobial use surveillance. Antimicrobial use was not restricted during the study period. Computerized databases of the pharmacy were used to determine the annual number of defined daily doses of selected antimicrobials per 100 admissions, including imipenem and third-generation cephalosporins (cefotaxime, ceftriaxone, and ceftazidime).
Statistical analysis. Categorical variables were compared by use of the χ2 test or Fisher's exact test as appropriate. Rates were compared by means of the large-sample test for comparison of incidences. Confidence intervals (CIs) for simple proportions were determined by using exact confidence limits for binomial distribution. Data collection and analysis were performed with use of Epi-Info Version 6.0 (CDC, Atlanta) and SPSS (SPSS, Chicago). All P values were two-sided; P values <.05 were considered significant.
Results
Between 18 February 1992 and 18 February 1996, ESBLPE were isolated from 454 patients. One hundred and ten cases were classified as imported from outside the hospital, and 328 cases were classified as acquired in the hospital. Sixteen cases could not be classified. Four hundred seventy-nine ESBLPE strains were isolated from 454 patients: K. pneumoniae, 416 (86.8%); Escherichia coli, 33 (7.0%); Enterobacter species, 15 (3.1%); and other Enterobacteriaceae, 15 (3.1%). The percentage of K. pneumoniae isolates among ESBLPE decreased over time, from 94.1% in 1992 to 76.6% in 1995. In 1997, 50 ESBLPE isolates (including those from routine specimens) were recovered from 38 patients: K. pneumoniae, 27 (54%); Enterobacter species, 8 (16%); E. coli, 6 (12%); and other Enterobacteriaceae, 9 (18%).
Of the 454 ESBLPE-positive patients, 249 had 321 ESBLPE infections, 167 (52%) of which were urinary tract infections. In the ICUs, where routine screening was performed, 105 (43%) of the 245 patients who acquired ESBLPE had only positive rectal screening tests. Of the remaining 140 patients, 110 had 174 ESBLPE infections. The remaining 30 patients were colonized at sites other than the rectum.
Rates of colonization and/or infection are presented in figure 1 and table 1. The prevalence of imported cases (total, 110) remained stable over the study period (1.1, 0.7, 1.1, and 0.6 cases/100 admissions in 1992, 1993, 1994, and 1995, respectively). These 110 patients were from ICUs (30 [27.2%]), medical or surgical wards (20 [18.2%]), nursing homes or rehabilitation units (19 [17.3%]), their homes (24 [21.8%]), or foreign countries (10 [9.1%]); origin could not be determined for 7 patients (6.4%). Twenty of the 24 patients admitted from home had a history of hospitalization; whether the remaining 4 patients had been hospitalized previously could not be determined. Of the 79 patients admitted to the 3 ICUs with ESBLPE, 50 were from outside our institution, and 29 were from other units in our institution; 40 (74%) of the 54 patients not known to be ESBLPE carriers had positive rectal screening tests, whereas only 30% would have been identified by testing of clinical specimens. The 69 patients with ESBLPE transferred to our hospital were from >40 different health care institutions.
No. of patients from whom extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBLPE) were isolated at Bichat-Claude Bernard Hospital in Paris between February 1992 and February 1996. Results are presented by quarter, starting on 18 February 1992. Shaded bars, imported cases of colonization and/or infection with ESBLPE; open bars, acquired cases of colonization and/or infection with ESBLPE. Yearly rates of acquired cases/100 admissions for each year from 1992 to 1995 were 0.56, 0.27, 0.16, and 0.06, respectively.
Number of patients, by unit, from whom extended spectrum β-lactamase-producing Enterobacteriaceae (ESBLPE) were isolated at Bichat-Claude Bernard Hospital in Paris during 1992–1995.
Of the 328 acquired cases in our hospital, 173 occurred in 1992, 84 in 1993, 52 in 1994, and 19 in 1995; incidences during these years were 0.56, 0.27, 0.16, and 0.06 cases/100 admissions, respectively (figure 1). All differences from year to year were highly significant (P <.0001, =.0035, and <.0001, respectively). Distribution among units is presented in table 1.
Of the 454 ESBLPE-positive patients, 156 patients (with 206 transfers) were transferred within the hospital: 126 transfers (61.1%) were from the 3 ICUs, and 48 transfers (23.3%) were to the septic surgical unit. The septic surgical unit served as a cohort unit for surgical patients with ESBLPE. Fifty-three patients were admitted to this unit, and 3 acquired ESBLPE. One-hundred fifty-three ESBLPE-positive patients were referred to the 922 medical or surgical beds (excluding the 3 ICUs and the rehabilitation unit), and 46 patients acquired ESBLPE in these units. The 56 ESBLPE-positive patients in the 28-bed septic surgical unit were hospitalized a total of 1340 days over the study period; the mean duration of isolation was 23.9 days/patient. The 199 ESBLPE-positive patients in the medical or surgical units were hospitalized a total of 3099 days. Incidences of ESBLPE acquisition in the septic surgical unit and the other medical or surgical units were 2.24 and 14.84 cases/1000 patient-days, respectively (incidence ratio, 6.63; 95% CI, 2.06–21.3; P =.0002).
Of the 454 ESBLPE-positive patients, 130 (28.6%) died in the hospital and 324 were discharged to a nursing home or rehabilitation unit (123 [38%]), their home (96 [29.6%]), a medical or surgical unit (91 [28.1%]), an ICU (2 [0.6%]), or an unknown destination (12 [3.7%]).We identified >80 different health care institutions to which patients were sent at discharge from our institution. Of the 221 ESBLPE-positive patients who were discharged from the 4 ICUs, 3 were discharged to home, 126 were transferred to other units in our hospital, and 90 were transferred to other hospitals. The destinations of 2 patients were unknown.
Compliance evaluations. During the first year of the intervention for infection control, we identified 121 patients with ≥1 clinical isolates of ESBLPE. This number was comparable with those during previous years. Most acquired cases occurred in the 3 ICUs. These findings prompted us to evaluate implementation of isolation procedures in the 3 ICUs in January 1993 and again in June 1994. The numbers of contacts were 281 in 1993 and 275 in 1994; 210 and 190 of these contacts were patient-health care worker interactions. The main findings of this evaluation are shown in table 2.
Compliance with isolation procedures for control of extended-spectrum β-lactamase-producing Enterobacteriaceae (ESBLPE) infection during 2 evaluations at Bichat-Claude Bernard Hospital in Paris.
Of observed contacts, 50% were with ESBLPE-positive patients in 1993 and 16.3% in 1994. The rate of compliance with isolation procedures for ESBLPE-positive patients increased from 57.5%–79.2% in 1993 to 71.0%–93.5% in 1994. For ESBLPE-negative patients, the rate of compliance with isolation procedures remained unchanged or increased from the first to the second evaluation (28.8%–72.1% to 28.3%–78.6%). The difference was significant for hand washing after contact with ESBLPE-negative patients between 1993 and 1994 (P <.0001). The isolation symbol was present in 96% and 90% of observations in 1993 and 1994, respectively. Nurses and assistant nurses were more likely than physicians and other staff to wash their hands after contact with a patient (in 1993, nurses, 88.1%; physicians, 73.3%; other staff, 40% in 1993, in 1994, 89.1%, 66.7%, and 70.4%, respectively). Breaks in the continuity of care were recorded for 79 (74.5%) of 106 observations of contacts between ESBLPE-positive patients and health care workers in 1993, and for 4 (12.9%) of 31 in 1994 (P <.0001). Compliance with isolation procedures was observed in 11.4% of these breaks in 1993 and 50% in 1994.
Results from the first compliance evaluation were used as the basis for a drive for improvement in isolation procedures that involved meetings of staff from the 3 ICUs. The importance and technique of hand washing was emphasized, and a booklet on hand washing was designed by the health care staff. Specific measures were decided on. For example, phone calls from relatives were previously received by nurses at the bedside; relatives were asked not to call these phone numbers during morning hours, at the time of maximum workload. In the surgical ICU, a list of the material required for complex dressings was drawn up and updated on a daily basis to avoid interruptions during dressing changes because of the nurse's having to leave the room to get missing material. A special emphasis was placed on compliance with isolation procedures by health care staff from outside the ICU, such as surgeons, chest physiotherapists, and radiology technicians.
Antimicrobial use. Use of selected antimicrobials is presented in figure 2. Use of third-generation cephalosporins increased steadily, by 60% from 1991 to 1995 (23% in ICUs and 69% in the other units). During the same period, imipenem use decreased by 33%, as a result of a 44% reduction in imipenem use in ICUs, contrasting with no change in the other units.
Use of selected antimicrobials at Bichat-Claude Bernard Hospital in Paris during 1991–1995. Use of third-generation cephalosporins (A) and imipenem (B). Results are presented as defined daily doses/100 admissions over 1 calendar year. Open bars, antimicrobial use in the 3 intensive care units; solid bars, antimicrobial use in the other units. Third-generation cephalosporins included cefotaxime, ceftriaxone, and ceftazidime.
Other multidrug-resistant organisms. The incidence of new cases (i.e., the number of clinical isolates of selected multidrug-resistant organisms/100 admissions) is displayed in figure 3. Between 1991 and 1995, the incidence of MRSA decreased by 38% (1.90–1.16 clinical isolates/100 admissions; P<.0001), and the incidence of A. baumannii decreased by 62% (1.21–0.46 clinical isolates/100 admissions; P<.0001). By contrast, the incidence of ticarcillin-resistant P. aeruginosa remained unchanged (0.87 clinical isolates/100 admissions), and the incidence of class C derepressed mutants of Enterobacteriaceae increased from 0.16 to 0.27 clinical isolate/100 admissions (P =.004).
No. of clinical isolates (excluding those from screening specimen and replicates) of selected multidrug-resistant organisms/100 admissions at Bichat-Claude Bernard Hospital in Paris during 1991–1995. ●, methicillin-resistant Staphylococcus aureus; ■, Acinetobacter baumannii; △, ticarcillin-resistant Pseudomonas aeruginosa; ●, class C derepressed mutants of Enterobacteriaceae.
Discussion
During the past decade, ESBLPE have emerged as one of the major multidrug-resistant organisms. In 1993, 9.4% of K. pneumoniae isolates in the National Nosocomial Infection System were ESBLPE [23]. The same year, the Paris Area Public Hospital Surveillance Network found that as many as 24.8% of K. pneumoniae isolates were ESBLPE, indicating an outbreak more extensive than any other reported worldwide [29]. As part of a multicenter infection control program initiated in the Paris area, we conducted a surveillance study of this outbreak.
The distribution of infections was as expected for nosocomial K. pneumoniae, with >50% of urinary tract infections [30, 42]. Fifty-four percent of the 321 infections occurred in the 3 ICUs, where ESBLPE were endemic. As in other studies, many of the cases acquired in the 3 ICUs were identified only by screening tests for gastrointestinal organisms [20, 42, 43]. Because these patients constitute a reservoir for subsequent cross-contamination in ICUs and in downstream units, screening probably played a pivotal role in identifying ESBLPE carriers and limiting spread of ESBLPE.
A major obstacle to effective control of the outbreak was the large number of patient transfers between units and hospitals. One-third of patients had at least 1 transfer within the hospital. Good communication among units and use of a specific symbol to flag charts of ESBLPE-positive patients are essential to ensure maintenance of isolation precautions until discharge. The outbreak was also fueled by hospital-to-hospital transfers, as shown by the very large number of health care institutions where ESBLPE-positive patients were transferred to or from. These data are in keeping with several reports [13, 23, 44]. A recent study examined the first SHV-4 extended-spectrum β-lactamase-producing K. pneumoniae strain isolated in 14 French hospitals, mainly in the Paris area [44]. Molecular methods showed that 12 of 14 isolates were probably the same strain. These data suggest that measures designed to control the spread of ESBLPE must be implemented simultaneously in all health care institutions linked by patient transfers [13].
The percentage of K. pneumoniae strains among ESBLPE decreased in our hospital from 94% in 1992 to 54% in 1997. This finding may be ascribable to successful control of a single epidemic strain. Nonetheless, our results demonstrate production of extended-spectrum β-lactamases by other Enterobacteriaceae, thus involving a risk of production of extended-spectrum β-lactamases in the community [24, 25, 45]. In our study, however, there were no documented cases of ESBLPE infection or colonization in patients admitted from their homes in the absence of a history of recent hospitalization.
Our ESBLPE control program was ineffective during the first year, primarily because of persistent acquisition of ESBLPE in the 3 ICUs. There is ample evidence that ICUs are the epicenters of outbreaks of multidrug-resistant organisms [2, 3, 6, 13, 20, 31]. This role of ICUs is ascribable to several factors. First, many patients are referred to ICUs from other health care institutions; thus, in our study nearly one-half (50/110) of the ESBLPE-positive patients referred to our institution were admitted to the 3 ICUs. Second, health care contacts, use of invasive procedures, and antimicrobial therapy are frequent in ICUs. Third, ICU patients with multidrug-resistant organisms are rarely discharged to home; in our study, only 3 of 221 ESBLPE-positive patients went home from the ICU. The other patients were transferred to other units, providing ESBLPE an opportunity to disseminate.
The failure of our program after the first year prompted us to evaluate isolation procedures in the 3 ICUs. We found that despite a high rate of hand washing compared with data from the literature [46], isolation procedures were not fully applied, especially by physicians and non-ICU health care workers. The other major feature was poor compliance in cases of breaks in the continuity of a patient's care. Our evaluation of isolation procedures was prepared and performed by nurses from the 3 ICUs. We believe that this promoted full understanding, development, and implementation of corrective measures. ESBLPE acquisitions decreased dramatically during the year after the evaluation. During the second evaluation, compliance with isolation procedures was similar to or slightly better than during the first evaluation. Breaks in care were rarely observed. In addition, compliance with hand washing was improved for ESBLPE-negative patients.
Surgical patients positive for ESBLPE were transferred to the septic surgical unit. Nurses in this unit, however, also cased for ESBLPE-negative patients. Nevertheless, only 3 acquisitions were documented in this unit. We cannot exclude the occurrence of other acquisitions in this unit, because routine screening was not performed. Sampling practices for microbiological cultures in this unit, however, were not different from those in other units. The risk of ESBLPE acquisition was significantly lower in the septic surgical unit than in the other medical or surgical units. Nurses and head nurses in this unit had extensive experience with the care of infected patients and were motivated to comply with isolation procedures. As in other outbreaks of hand-transmitted multidrug-resistant microorganisms and involving many patients, a cohort unit appeared to be useful in controlling dissemination of ESBLPE [47, 48].
Both restriction of antimicrobial use and isolation procedures have been suggested to curb ESBLPE spread. Several studies have found an association between increased use of antimicrobials, most notably third-generation cephalosporins, and ESBLPE outbreaks and reported a decrease in the incidence of ESBLPE infection and colonization after restriction of the use of some antimicrobials, usually ceftazidime [17–21]. Other measures, including isolation procedures, were also used in these studies; however, the effect of each intervention on infection control remained unclear.
Two studies have evaluated risk factors for ESBLPE acquisition in ICUs [20, 43]. Invasive procedures, but not antimicrobial use, were independently associated with ESBLPE acquisition. The relationship between invasive procedures and ESBLPE acquisition, which usually occurs in the digestive tract, is not clear. Transmission via the hands can occur during nonsterile manipulation of devices used for invasive procedures. In addition, use of invasive procedures may be a marker for intensity of care. In outbreaks involving other multidrug-resistant organisms transmitted by cross-contamination, such as MRSA, intensity of care, workload, and overcrowding with understaffing were risk factors for their acquisition [49, 50]. Because the hands are the vector for transmission of multidrug-resistant Klebsiella [30], these data support a role for intensity of care and workload rather than antimicrobial use in ESBLPE acquisition.
Our finding that use of third-generation cephalosporins increased at the same time that the outbreak was successfully controlled lends support to this possibility. Use of third-generation cephalosporins for an ESBLPE-positive patient, however, may amplify the burden of ESBLPE in this patient and therefore the risk of cross-transmission. Therefore, adding restriction of antimicrobial use to our infection control program would perhaps have expedited its effect.
It is interesting that the incidence of other multidrug-resistant organisms acquired by cross-transmission, primarily MRSA and A. baumannii, declined significantly during the study period. Routine screening and isolation procedures for MRSA were started in the 3 ICUs in early 1995. The ESBLPE control program therefore seems to have reduced the transmission of other multidrug-resistant organisms carried on the hands. Therefore, targeting control of hand-transmitted multidrug-resistant organisms may be effective in improving overall hygiene precautions. Conversely, antimicrobial drugs are well-known risk factors for the emergence of other microorganisms, such as ticarcillin-resistant P. aeruginosa and class C derepressed mutants of Enterobacteriaceae [51, 52]. Surveillance showed that the incidence of these microorganisms remained unchanged or increased; during the same time, we recorded an increase in use of third-generation cephalosporins. The decreasing use of imipenem in the ICUs was probably due to a decrease in the number of infections due to ESBLPE and A. baumannii.
In conclusion, our program aimed at controlling an ESBLPE outbreak became successful after we specifically targeted interventions for infection control in the ICUs, where most acquisitions occurred. Evaluation of infection control procedures proved useful for identifying deficiencies and introducing corrective measures. Involvement of the nursing staff may have enhanced the efficacy of the program. The incidence of multidrug-resistant organisms acquired by cross-transmission decreased in lockstep with the incidence of ESBLPE, whereas the incidence of other multidrug-resistant microorganisms emerging under the selection effect of antimicrobials remained stable or increased with use of third-generation cephalosporins.
Acknowledgments
We thank the following personnel at Bichat-Claude Bernard Hospital in Paris: Bruno Baune (Pharmacy Unit) for providing pharmacy data, Dr. Sylvie Calvat (Medical ICU) for her technical assistance in the conduct of the study, the staff of the 3 ICUs for their active participation, and Nadine Sauteret and Antoinette Wolfe for their preparation of the manuscript.
References
See editorial response by Paterson and Yu on pages 1419–22.
This work was presented in part at the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, 15–17 September 1996, New Orleans.
Author notes
Current affiliations: Department of Microbiology, Hôpital Saint Antoine (D.D.), and Emergency Unit (A.F.) and Infectious Diseases Intensive Care Unit (B.R.), Hôpital Bichat-Claude Bernard, Paris, Department of Microbiology, Hôpital Louis Mourier, Colombes (M.-L.J.-G.); Hôpital Jean Verdier, Bondy (M.P.), France.
Comments