| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Guest Access | Sign In via User Name/Password
|
|
|
|||||||
Guest Access | Sign In via User Name/Password |
|||||||||
* From the Divisions of Infectious Diseases (Ms. Merz, and Drs. Warren and Fraser), and Pulmonary and Critical Care Medicine (Dr. Kollef), Washington University School of Medicine, Saint Louis, MO; and the Division of Bacterial and Mycotic Diseases (Dr. Fridkin), Centers for Disease Control and Prevention, Atlanta, GA.
Correspondence to: Liana R. Merz, MPH, Washington University School of Medicine, Box 8051, 660 S Euclid Ave, Saint Louis, MO 63110; e-mail: lmerz{at}im.wustl.edu
Abstract
Background: Antimicrobial-resistant organisms are an emerging problem in the ICU. Therapy cycling empiric antibiotics between various classes may influence bacterial resistance patterns. Understanding the impact of cycling on the appropriate treatment of suspected Gram-negative infections is important.
Methods: Data were prospectively collected on patients who were admitted to a 19-bed medical ICU (MICU). A total of 1,172 patients were admitted to the MICU for > 48 h and were evaluated during a 28.5-month period. After 4.5 months of baseline data collection, an antibiotic-cycling protocol was implemented, using four different antibiotic classes with Gram-negative activity that were cycled every 3 to 4 months. Therapy was considered to be inappropriate if the subsequent bacterial isolate was resistant to the empiric drug used.
Results: There were 59 bloodstream infections (BSIs), 17 ventilator-associated pneumonias (VAPs), and 101 urinary tract infections (UTIs) involving Gram-negative bacteria among 139 patients. Fifty-five infections (31%) were due to Gram-negative bacteria resistant to one or more antibiotic agents (BSIs, 18 [30%]; VAPs, 4 [23%]; and UTIs, 33 [33%]). Fifteen patients received inappropriate empiral therapy for 18 resistant Gram-negative infections (BSIs, 7 [39%]; VAPs, 3 [75%]; UTIs, 8 [24%]). Patients receiving inappropriate therapy were more likely to die (10 patients [67%] vs 40 patients [32%], respectively; p < 0.01). There was no difference in the receipt of appropriate empirical antibiotic therapy during the baseline compared to cycling (infectious episodes, 15% vs 10%, respectively; p = 0.4).
Conclusions: Antimicrobial resistance occurred in almost 30% of ICU infections involving Gram-negative bacteria. Antibiotic cycling was not associated with significant changes in the receipt of appropriate empirical antimicrobial therapy for the treatment of ICU infections.
Key Words: antibiotics • antibiotic rotation • antimicrobial resistance • ICU • outcomes
Antimicrobial-resistant bacteria are a significant problem in ICUs. Infections due to antimicrobial-resistant bacteria, particularly Gram-negative bacteria, are associated with higher mortality, prolonged hospitalization, and increased health-care costs.123 In addition, the presence of antimicrobial resistance decreases the ability of clinicians to optimally treat suspected infections. Inappropriate initial antimicrobial therapy of nosocomial infections can also lead to increased hospital mortality, increased hospital LOS, and increased morbidity.456
Steps need to be taken to slow the emergence and spread of antimicrobial-resistant organisms. Antibiotic cycling is one mechanism that has been proposed to control the emergence of antimicrobial-resistant bacteria in ICUs.78910111213 In theory, by not using either a class of antibiotics or a specific drug for a specified period of time and then reintroducing it, bacterial resistance to the cycled agent may be minimized. Moss et al8 examined the impact of antibiotic cycling in a pediatric ICU and reported a 5.4% decline in the absolute prevalence of colonization with antimicrobial-resistant bacteria. Gruson et al13 conducted a cycling project in a medical ICU (MICU) and reported a decrease in the incidence of ventilator-associated pneumonia (VAP) caused by antimicrobial-resistant bacteria. There are few data, however, on the impact of antibiotic cycling protocols on the adequacy of initial empirical antimicrobial therapy for ICU infections. Kollef et al14 examined the impact of scheduled changes in antimicrobial classes on the incidence of inadequate antimicrobial therapy for nosocomial infections and concluded that scheduled changes could decrease the rate of inadequate antimicrobial therapy in the ICU. However, this study examined the impact of switching antibiotics not cycling. Thus, further study is needed to determine the impact of antibiotic cycling on the adequacy of initial empirical antimicrobial therapy for ICU infections. Additional research is also needed to assess the safety of antibiotic cycling, and its impact on both patient outcomes and antimicrobial resistance.
In this study, we prospectively examined the effect of an antibiotic cycling intervention on the frequency of inappropriate empirical antimicrobial use for patients with Gram-negative infections in an ICU, and the relationship between appropriate empirical antimicrobial therapy for Gram-negative bacterial infections and patient outcomes.
Methods and Materials
Study Location and Patients
The study was conducted at Barnes-Jewish Hospital in St. Louis, MO, a 1,400-bed urban, tertiary care teaching hospital. The 19-bed MICU served as the study unit. The MICU is staffed by a multidisciplinary care team whose members rotate on a monthly basis. All MICU attending physicians are board-certified in critical care medicine. During a 28.5-month period (February 14, 2000, to June 30, 2002), all patients admitted to the MICU for > 48 h were enrolled into the study.
Study Design and Data Collection
This was a prospective before-and-after study. For all study patients, the following characteristics were collected: age; gender; race; and the presence of preexisting medical conditions including congestive heart failure (CHF), COPD, diabetes, and malignancy. Additional data pertaining to patient demographics, medical history, hospital and ICU admission dates, and acute physiology and chronic health evaluation (APACHE) II score15 on admission to the hospital and ICU were also collected. In addition, data on processes of care, including the use of mechanical ventilation, a central venous catheter, and antimicrobial therapy were prospectively collected. Data pertaining to ICU treatment and events, including organ failure and the acquisition of Clostridium difficile-associated diarrhea, were also recorded.
All definitions were selected prospectively as part of the original study design, including acquired organ insufficiency as described by Rubin and colleagues.16 Urinary tract infections (UTIs), bloodstream infections (BSIs), and VAPs were defined using the Centers for Disease Control and Prevention/National Nosocomial Infections Surveillance System (NNIS) criteria17; the microbiological causes of all infections were determined using available clinical cultures. Antimicrobial susceptibility results were noted for all isolates. Empirical therapy for these infections was defined as being inappropriate if the subsequent Gram-negative bacteria isolated were not sensitive to the initial drug used.
Baseline data were collected for 4.5 months (February 14 to June 3, 2000). The details of this cycling program have been previously described.718 During the baseline period, the prescription of antibiotics for the empirical coverage of presumed infections by Gram-negative bacteria was at the discretion of the ordering physician, but all antibiotic classes with broad-spectrum activity against Gram-negative bacteria (ie, third-generation and fourth-generation cephalosporins, fluoroquinolones, carbapenems, and extended-spectrum penicillin/ß-lactamase inhibitors) required approval by the Barnes-Jewish Hospital Antibiotic Management Program prior to being dispensed. The only exception to this was cefepime, which could be prescribed without prior approval for 72 h. After 72 h, the ordering physician had to obtain pharmacy approval for continued use.
After the baseline observation period, an antibiotic-cycling protocol was implemented, using four antibiotic classes with Gram-negative activity for empirical use cycled every 3 to 4 months over a 2-year period. The four antibiotic classes cycled included cephalosporins, fluoroquinolones, carbapenems, and extended-spectrum penicillin/ß-lactamase inhibitors. This four-drug rotation was cycled twice, with the cycle drug changing every 4 months during the year 1 (rotation 1), and every 3 months during year 2 (rotation 2). Forty-eight percent of the total cycle antibiotic days were compliant with the cycling protocol, and a mean of 8.8 days per patient were spent receiving the specified cycle drug.7
Statistical analysis was performed using a statistical software package (SPSS, version 11.0 for Windows; SPSS Inc; Chicago, IL). Categoric variables were compared using the 
2 test or Fisher exact test, as appropriate. Continuous variables were compared using the Wilcoxon ranked sum test. A p value of < 0.05 on two-tailed testing was considered to be significant, and the Bonferroni correction was used in univariate analysis to adjust for multiple comparisons. Because of the small sample size of patients with inappropriate empirical therapy, multivariate analysis could not be performed.
The institutional review boards at Washington University and the Centers for Disease Control and Prevention approved this study.
Results
Characteristics of Gram-Negative Infections
During the study period, 3,239 patients were admitted to the study ICU; 1,172 patients (36%) were eligible and were enrolled into the study. A total of 5,016 positive cultures were obtained from 932 patients (mean, 5.38 cultures per patient; range, 1 to 52 cultures per patient). Among these positive cultures, 219 BSIs (8.3 per 1,000 patient days), 32 VAPs (3.5 per 1,000 ventilator days), and 292 UTIs (11.1 per 1,000 patient days) were identified.
A total of 177 infections involving Gram-negative bacteria occurred in 139 patients (BSIs, 59 [27%]; VAPs, 17 [53%]; UTIs, 101 [35%]) [Table 1 ]. The microbiology of these Gram-negative bacterial infections varied, with 55 infections (31%) due to Gram-negative bacteria that were resistant to one or more antimicrobial agents. Nine BSIs (15%), 11 VAPs (65%), and 17 UTIs (17%) were caused by Pseudomonas aeruginosa, while Escherichia coli was isolated in 10 BSIs (17%) and 37 UTIs (37%) [Table 1]. P aeruginosa was isolated in 5 of the 15 inadequately treated infections (33%) caused by Gram-negative bacteria, while Acinetobacter spp was isolated in 3 inadequately treated infections (20%) and E coli in 2 inadequately treated infections (13.3%).
|
Thirty-four of 242 patients (14%) enrolled during the baseline period and 105 of 930 patients (11%) enrolled during the cycling part of the protocol had one or more ICU-related Gram-negative infections (Table 2 ). There were no differences between the patients enrolled during the two periods with respect to age, race, or gender. Compared to patients who had been treated during the baseline period, patients enrolled during the cycling period had greater antacid use (cycling period, 17%; baseline period, 0%; p < 0.01) and hospital mortality (baseline period, 7 patients [21%]; cycling period, 43 patients [41%]; p = 0.03) [Table 2]. In addition, patients enrolled during the cycling period had greater mean ICU LOS (13.8 vs 9.1 days, respectively; p = 0.045). There was no difference in the receipt of inappropriate antimicrobial therapy between the baseline and cycling periods (baseline period, 5 patients [15%]; cycling period, 10 patients [10%]; p = 0.40). When comparing the appropriateness of therapy at baseline compared to rotation 1 and rotation 2 by infection type, there was also no significant difference (Fig 1 ).
|
|
|
|
Discussion
In this prospective cohort study of MICU patients, we found that the implementation of an antibiotic cycling protocol did not adversely impact the receipt of appropriate empirical antimicrobial therapy. There was not a significant difference in the delivery of appropriate therapy when comparing the baseline period to the cycling period. Corticosteroid use, longer hospital LOS, and greater risk of in-hospital mortality were significantly associated with inappropriate initial antimicrobial therapy for Gram-negative bacterial infections on univariate analysis.
Previous studies419202122 have identified an association between hospital mortality and the receipt of inadequate or inappropriate antimicrobial therapy for BSIs and VAPs. Weinstein and colleagues21 found that patients who received appropriate antimicrobial therapy for BSIs had the lowest mortality rate (13.3%). Similarly, Ibrahim et al4 demonstrated that increased hospital mortality was associated with inappropriate antimicrobial therapy for patients with BSIs compared to patients receiving appropriate initial antimicrobial treatment.2324 Alvarez-Lerma23 found that attributable mortality was significantly higher in patients who had received inappropriate initial antimicrobial therapy. Similar results were found in our study. Patients receiving inappropriate antimicrobial therapy for the treatment of infections in the ICU were at increased risk of in-hospital mortality. In agreement with this prior research, our findings suggest that increasing the amount of appropriate antimicrobial therapy will improve patient outcomes. This can be achieved by using knowledge of the hospital antibiogram and patient risk factors when prescribing antibiotics in order to improve the likelihood of adequate initial empirical therapy.
Factors that contribute to inappropriate antimicrobial therapy include prior exposure to antibiotics, lengthy hospital stay, and use of mechanical ventilation.5 While we were unable to examine prior exposure to antimicrobial agents as an independent risk factor in this study, we did find longer hospital LOS to be a risk factor for inappropriate antimicrobial therapy.
Antimicrobial resistance is emerging as a major cause of morbidity and mortality in both hospital and community settings. It influences both patient outcomes and the costs associated with hospitalization.5 It is important to establish the risks and benefits of implementing an antimicrobial cycling program. In this study, the routine cycling of antimicrobial classes with Gram-negative activity for empirical use did not increase the risk of receiving inappropriate antimicrobial therapy.
There were several limitations to this analysis. The small number of Gram-negative infections and patients receiving inappropriate antimicrobial therapy did not allow all outcomes and factors of interest to be fully explored. The number of patients receiving inappropriate therapy is too small to perform a multivariate analysis to determine whether a relationship between inappropriate antimicrobial therapy and death is independent of the underlying severity of the illness. Therefore, this study did not establish causality between inadequate initial treatment and worse outcomes. Also, the sensitivities of all bacteria to extended-spectrum penicillin/ß-lactamase inhibitors were not tested for all isolates due to a national shortage of piperacillin/tazobactam testing discs at the time of the study. If sensitivities were not available, appropriate treatment was assumed. In addition, this study was conducted at a single ICU, and therefore may not be generalizable to other ICUs with dissimilar infection rates and patient populations. However, the frequency of resistant Gram-negative infections seen in the study ICU is very similar to those seen in NNIS data (eg, 29.7% of P aeruginosa isolates were resistant to ciprofloxacin in the study ICU vs 29.3% of isolates were resistant in the NNIS ICUs).25
Antimicrobial resistance to one or more study drug classes occurred in almost one third of ICU-related infections involving Gram-negative bacteria. The use of antibiotic cycling was not associated with significant changes in the receipt of appropriate empirical antimicrobial therapy. The receipt of inappropriate empirical therapy for Gram-negative bacterial infections in the ICU was also associated with adverse patient outcomes, although a larger sample size is needed to investigate this further.
Acknowledgements
The authors would like to acknowledge the following people whose support and participation facilitated the completion of this study: the nursing staff of the Barnes-Jewish Hospital Medical Intensive Care Unit; Ms. Cherie Hill; Ms. Sondra Seiler; and Dr. David Sterling.
Footnotes
Abbreviations: APACHE = acute physiology and chronic health evaluation; BSI = bloodstream infection; CHF = congestive heart failure; LOS = length of stay; MICU = medical ICU; NNIS = National Nosocomial Infections Surveillance System; UTI = urinary tract infection; VAP = ventilator-associated pneumonia
This work was supported in part by Centers for Disease Control grants U50/CCU717925-03-01 and UR8/CCU715087 (CDC Prevention EpiCenter Program), and National Institute of Allergy and Infectious Diseases Career Development Award 1K2-AI50585-01A1 (to Dr. Warren).
The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.
Received for publication February 8, 2006. Accepted for publication May 26, 2006.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
