Fluoroquinolone-Resistant Pseudomonas aeruginosa: Assessment of Risk Factors and Clinical Impact

Transcription

1 The American Journal of Medicine (2006) 119, 526.e e25 CLINICAL RESEARCH STUDY Fluoroquinolone-Resistant Pseudomonas aeruginosa: Assessment of Risk Factors and Clinical Impact Leanne B. Gasink, MD, a,e,f Neil O. Fishman, MD, a,f Mark G. Weiner, MD, b Irving Nachamkin, DrPH, MPH, d Warren B. Bilker, PhD, c,e,f Ebbing Lautenbach, MD, MPH, MSCE a,c,e,f a Divisions of Infectious Diseases and b General Internal Medicine, Department of Medicine, c Department of Biostatistics and Epidemiology, d Department of Pathology and Laboratory Medicine, e Center for Clinical Epidemiology and Biostatistics, and f Center for Education and Research on Therapeutics, University of Pennsylvania School of Medicine, Philadelphia, Penn. ABSTRACT PURPOSE: Pseudomonas aeruginosa infections have been associated with considerable morbidity and mortality. Fluoroquinolones (FQ) are the only oral therapy available for P. aeruginosa infections, but resistance is increasingly prevalent. METHODS: We examined annual trends in FQ-resistant P. aeruginosa (FQRPA) from 1991 to Subsequently, inpatients with a clinical culture positive for P. aeruginosa between January 1, 1999 and December 31, 2000 were included in a case control study to identify risk factors for FQ resistance and a cohort study to examine the impact of FQ resistance on outcomes in P. aeruginosa. RESULTS: Annual prevalence of FQRPA increased from 15% in 1991 to 41% in 2000 (P trend). Between 1999 and 2000, 332 P. aeruginosa isolates were FQ resistant and 540 were FQ susceptible. Prior FQ use was the only independent risk factor for FQRPA (adjusted OR 3.43; 95% confidence interval [CI] 2.37, 4.96). Subjects with FQRPA had greater median hospital charges ($62,325 vs $48,734) (P.007) and higher mortality (47.5% vs 35.5%) (P.004). However, in a multivariate model, only imipenem resistance of the isolate was significantly associated with mortality. FQ resistance was not an independent risk factor. CONCLUSIONS: FQRPA has increased significantly and is associated with prior FQ use. Limiting FQ use may curb the emergence of resistance among P. aeruginosa. FQRPA is associated with increased hospital charges, but other resistance patterns may have a more significant impact on mortality Elsevier Inc. All rights reserved. KEYWORDS: Fluoroquinolones; Pseudomonas aeruginosa; Resistance; Pseudomonas aeruginosa is one of the most common hospital-acquired pathogens. 1,2 A delay in appropriate antimicrobial therapy has been shown to be associated with P. aeruginosa infection-related mortality, but selection of empiric and definitive therapy has become increasingly difficult due to high rates of antibiotic resistance. 3,4 Fluoroquinolones (FQs) are an important class of antipseudomonal antibiotics because they have favorable safety and pharmacokinetic profiles and convenient oral dosing. Requests for reprints should be addressed to Leanne B. Gasink, MD, 502 Johnson Pavillion, University of Pennsylvania, Philadelphia, PA address: leanne.gasink@uphs.upenn.edu. However, FQ resistance among P. aeruginosa isolates has increased at an alarming rate. In United States intensive care units, rates of resistance to ciprofloxacin increased from 15% in 1993 to 32% in Similar trends have been documented in other countries. 5 Moreover, resistance to FQs is associated with resistance to other antibiotics. 6 Few studies report patient-level data about specific risk factors for FQ-resistant Pseudmonas aeruginosa (FQRPA). Results have conflicted and may be limited by inadequate power. 7-9 Furthermore, the clinical and economic impact of FQRPA has not been well described. A better understanding of the epidemiology and impact of FQRPA could aid in the design of interventions to preserve the utility of this class of /$ -see front matter 2006 Elsevier Inc. All rights reserved. doi: /j.amjmed

2 526.e20 The American Journal of Medicine, Vol 119, No 6, June 2006 antibiotics in the treatment of P. aeruginosa. The purpose of this study, therefore, was to identify risk factors for FQRPA among a large cohort of subjects. In addition, we sought to investigate the clinical and economic impact of FQ-resistant P. aeruginosa. METHODS This study was conducted at the Hospital of the University of Pennsylvania (HUP), a 625-bed academic tertiary care medical center in Philadelphia, Pennsylvania. The study was approved by the Committee on Studies Involving Human Beings of the University of Pennsylvania. CLINICAL SIGNIFICANCE Analysis of Secular Trends All P. aeruginosa isolates identified by the Clinical Microbiology Laboratory at HUP between January 1, 1991 and December 31, 2000 were identified, and trends in the prevalence of FQRPA were examined. The following agents were used as markers of FQ susceptibility in different time periods throughout the study: ciprofloxacin ( ), ofloxacin ( ), and levofloxacin ( ). Case Control Study We conducted a case-control study to identify risk factors for FQRPA. Study subjects were identified through records of the clinical microbiology lab. All patients who had an inpatient clinical culture positive for P. aeruginosa between January 1, 1999 and December 31, 2000 were eligible for inclusion. Repeat isolates in individual patients were included only if they occurred more than 30 days after the initial P. aeruginosa isolate was identified. A new isolate after 30 days was felt to likely reflect a new clinical event. This time period also reflected a new window of potential exposures possibly influencing the subsequent isolation of a new P. aeruginosa isolate. Levofloxacin was used as a marker for FQ resistance. Patients with FQRPA were designated as cases. Patients with levofloxacin-susceptible P. aeruginosa (FQSPA) were designated as controls. Risk factors were assessed through the use of a comprehensive health system computer database, which has been used effectively for similar studies of antimicrobial use in the past. 10 Data obtained included age, sex, race, origin of patient at the time of hospital admission (home, transferred from another institution), hospital location at the time of infection, and number of hospital and intensive care unit (ICU) days before infection. The anatomic site of infection and the antibiotic susceptibility profile of the P. aeruginosa isolate were noted. Nosocomial acquisition was defined as P. aeruginosa infections cause considerable morbidity and mortality. Rates of drug resistance, particularly fluoroquinolone resistance, are increasing. Information about risk factors and impact of fluoroquinolone-resistant P. aeruginosa is needed to develop strategies for curbing the emergence of resistance and preserving the utility of fluoroquinolones. an isolate identified 48 hours after admission to the hospital, or an isolate identified 48 hours after admission in a patient transferred from another medical institution. Finally, all antimicrobial therapy administered in the 30 days before the positive culture was ascertained. The presence of comorbid conditions was documented: hepatic dysfunction/cirrhosis, anemia (indicated by hemoglobin of 10.0), malignancy, diabetes mellitus, renal insufficiency (indicated by creatinine level 2.0 mg/dl or the requirement of dialysis), and human immunodeficiency virus (HIV) infection. The Charlson comorbidity index also was calculated for each subject. 11 Cohort Study To investigate the association between FQRPA and clinical and economic outcomes, we conducted a retrospective cohort study. Those subjects previously identified as cases and controls comprised the exposed (patients with FQRPA) and unexposed (patients with FQSPA) subjects of the cohort study. Specific outcomes of interest were in-hospital mortality, length of hospital stay following the culture, and total hospital charges from the date of culture through the date of discharge or death. Microbiological Methods Susceptibilities to all antimicrobial agents were performed and interpreted according to the criteria of the National Committee for Clinical Laboratory Standards (NCCLS), by means of either a semiautomated system (MicroScan Walk- Away System, NC16 panel, Dade Behring; biomerieux, St. Louis, Missouri) or disk diffusion susceptibility testing. 12,13 Statistical Methods The annual prevalence of FQRPA isolates was calculated for the time period 1991 through 2000 and the Cochran- Armitage trend test (chi-square test for trend) was performed. 14 In the case control study, bivariable analyses were conducted to determine the association between potential risk factors and FQRPA, in particular, prior FQ use. Categorical variables were compared using the Fisher s exact test. An odds ratio (OR) and 95% confidence interval (CI) was calculated to evaluate the strength of any association. Continuous variables were compared using the Student s t test or the Wilcoxon rank-sum test, depending on the validity of the normality assumption. 15 Stratified analyses were then performed to identify where data were sparse and to elucidate where confounding and interactions were likely to exist. Interaction was assumed to

3 Gasink et al Fluoroquinolone-Resistant Pseudomonas aeruginosa 526.e21 be present when the test for heterogeneity between the ORs for different strata were significant (P.05); the Mantel- Haenszel test for summary statistics was used to evaluate the effects of each variable of interest as a possible confounder. 16 Multivariable analysis was performed using multiple logistic regression. 17 Building of the multivariable model began with inclusion of the key variable of interest (prior FQ use). Length of stay before the P. aeruginosa culture was apriori included in the final model given the importance of time at risk as a potential confounding variable. 18 Variables with a P value of 0.20 on bivariable analyses were considered for inclusion in a multivariable model as were variables noted to be confounders on stratified analysis. 19 Variables remained in the final model if their inclusion resulted in a 15% change in the effect size for the association of the primary association of interest (prior FQ use) and FQRPA. 20 Interaction between risk factor variables was also investigated. Bivariable and stratified analyses were similarly conducted in the evaluation of the association between FQ resistance and mortality. The impact of potential confounders was examined, specifically ICU location at the time of culture, nosocomial acquisition, anatomic site of culture, comorbidities, Charlson score, and resistance to other unique agents/classes of antibiotics (ie, aminoglycosides, carbapenems [imipenem], piperacillin, aztreonam and antipseudomonal cephalosporins [cefepime and ceftazidime]). As in the case-control analysis, the final multivariable model included length of stay before P. aeruginosa culture. 21 For all calculations, a two-tailed P value of.05 was considered significant. All statistical calculations were performed using standard programs in STATA v 8.0, (Stata- Corp, College Station, Tex). RESULTS Analysis of Secular Trends From 1991 through 2000, 4976 P. aeruginosa isolates were identified. Over this time period, the annual prevalence of FQ-resistant isolates increased significantly from 15% to 41% (P.001, trend) (Figure 1). Case Control Study During , 940 inpatient P. aeruginosa isolates fulfilling inclusion criteria were identified in 677 patients. Of these 940 isolates, 872 (92.8%) had full data available through the health system patient database. There were no substantial differences when comparing available data (ie, age, sex, anatomic site of culture, susceptibility profile of the organism) for those subjects who were and were not included. The anatomic site of culture for these 872 isolates was respiratory in 388 (44.5%); urinary in 198 (22.7%); skin/soft tissue in 184 (21.1%); blood in 75 (8.6%); abdominal in 10 (1.2%); catheter in 9 (1.0%); and other in 8 (0.9%). Although these 872 isolates demonstrated variable Percent Fluoroquinolone Resistant (%) Year P <.001, trend Figure 1 Trends in fluroquinolone resistance ( ). resistance to other antibiotics and antibiotic classes, FQRPA isolates were significantly more likely to be resistant to other agents (Figure 2). On bivariable analyses, several factors were associated with FQRPA (Table 1). Prior use of numerous antibiotics, including FQs, was significantly more common among cases. Cases were exposed to a greater number of antibiotics in the prior 30 days. The median (interquartile range [IQR]) number of antibiotic-days for cases and controls was 4 (0-47) and 3 (0-13), respectively (P.001). On multivariable analysis, the only independent risk factor for FQRPA was prior FQ use (adjusted OR [95% CI] 3.43 [2.37, 4.96], P.001). As noted previously, the variable for hospital days from admission to culture was included in the final model but was not significantly associated with FQRPA (adjusted OR [95% CI] 1.00 [1.00, 1.01]; P.34. We performed a secondary analysis in which each subject was included only once, even if subsequent P. aeruginosa isolates were identified after 30 days. The results of the final multivariable analysis were not substantively different from the primary analysis. Cohort Study Among the 872 subjects in the case control study, 847 (320 with FQRPA and 527 with FQSPA) had discharge disposition data available and were included in the cohort study. Patients not included in the cohort study were compared with the original 940 subjects with P. aeruginosa and no substantial differences were identified. The median (IQR) hospital charges for subjects with FQRPA and FQSPA were $62,325 ($22,129-$188,979) and $48,733 ($18,760-$124,829), respectively (P.008). Length of stay subsequent to culture did not differ significantly between the two groups. Median (IQR) days for subjects with FQRPA was 10 (4-14); compared with 9 (4-20) for subjects with FQSPA (P.13). In-hospital mortality among subjects with FQRPA was 77/320 (24.1%), compared with 85/527 (16.1%) for subjects with FQSPA (relative risk [95% CI] 1.49 [1.13, 1.96]; P.004).

4 526.e22 The American Journal of Medicine, Vol 119, No 6, June 2006 Ceftazidime Levofloxacin Susceptible (n=540) Levofloxacin Resistant (n=332) Amikacin Aztreonam Cefepime Agent Imipenim Gentamicin Tobramycin Piperacillin Percent Resistant For all comparisons, p<0.001 Figure 2 Resistance to other antimicrobials among FQ-resisistant and FQ-susceptible P. aeruginosa isolates ( ). However, in a multivariate analysis, only the imipenem resistance status of the P. aeruginosa isolate was significantly associated with mortality. FQRPA had no independent association (Table 2) Subsequent comparisons of resistance profiles among FQ-resistant and imipenemresistant P. aeruginosa isolates revealed that imipenemresistant isolates were resistant to a greater number of unique antipsuedomonal classes. Imipenem-resistant isolates were resistant to a median (IQR) of 4 (2-6) unique classes of drugs and FQRPA isolates were resistant to 2 (2-3) classes of drugs. Among imipenem-resistant isolates, 42.6% were resistant to 5 or more unique classes of drugs, compared with 7.4% of FQRPA isolates (Figure 3). DISCUSSION The high oral bioavailability, convenient dosing, and excellent safety profile of FQs makes them a particularly attractive treatment option for the treatment of P. aeruginosa. However, the emergence of FQ resistance among P. aeruginosa and several other bacteria has been noted frequently in recent years. 4,22,23 Given the already limited arsenal of antipseudomonal drugs, the loss of FQs as reliable agents is of major concern. Several ecologic studies have linked increases in national and hospital FQ use with decreased rates of P. aeruginosa susceptibility to FQs. 6,24,25 Although several authors have reported an association between previous exposure to a particular antipseudomonal agent and subsequent emergence of pseudomonal resistance to that particular agent, very few studies have focused specifically on risk factors for FQRPA. 26,27 Most recently, Hsu et al identified 91 cases of FQRPA and 86 controls, and found that FQ use in the prior 30 days, nosocomial acquisition, and diabetes mellitus were independent risk factors for FQRPA. 9 FQ use within 7 days has also been implicated as a risk factor for FQRPA pneumonia. 7 In contrast, a recent small study reported no association between FQ use and FQRPA. 28 Our study is the largest to examine FQ resistance specifically among P. aeruginosa and suggests that interventions designed to limit FQ use might be effective in curbing further emergence of FQRPA. Indeed, the high prevalence of inappropriate FQ use noted previously suggests there is substantial room for improvement in current FQ use practices. 29 High rates of resistance to other antibiotics among FQRPA also have been noted by other investigators and have serious implications on the selection of other antipseudomonal agents for empiric therapy. 6,9 Several investigators have implicated prior exposure to FQs as an independent risk factor for multi-drug-resistant P. aeruginosa. 30,31 Although the primary mechanism of FQ resistance is via alteration in 2 bacterial enzymes, DNA gyrase and topoisomerase IV, FQ resistance also may occur via alteration or reduction in outer membrane proteins or through overexpression of multi-drug resistance efflux pumps that enhance excretion of FQs and other agents from the cell. 32 Thus, cross-resistance of FQRPA is not surprising. We found that imipenem resistance, and not FQ resistance, is independently associated with increased mortality. Although the explanation for this is unclear, it is well recognized that a delay in effective therapy has been shown to be a major predictor of infection-related mortality among patients with P. aeruginosa. 3,33 Thus, although these data were not available in our dataset, it is possible that initiation of appropriate therapy was less common in those subjects with imipenem-

5 Gasink et al Fluoroquinolone-Resistant Pseudomonas aeruginosa 526.e23 Table 1 Bivariable Analysis (Risk Factors for FQ Resistance) Variable Cases (n 332) n (%) Controls (n 540) n (%) OR (95% CI) P value 1 * General Age 56 ( ) 62 (48-72).001 Hospital days 6 (1-30.5) 5 (1-14).02 ICU days 1 (0-20) 0.5 (0-6).001 ICU location 163 (49.1%) 210 (38.9%) 1.52 (1.14, 2.02).003 Comorbidities Renal insufficiency 130 (39.2%) 173 (32.0%) 1.37 (1.02, 1.83).03 Hepatic insufficiency 99 (29.8%) 110 (20.4%) 1.66 (1.20, 2.30).002 Anemia 281 (84.6%) 423 (78.3%) 1.52 (1.05, 2.23).02 Diabetes 93 (28.0%) 116 (21.5%) 1.42 (1.02, 1.97).03 Prior antimicrobial use Fluoroquinolone 130 (39.2%) 79 (14.6%) 3.76 (2.68, 5.27).001 Aminoglycoside 136 (41.0%) 143 (26.5%) 1.93 (1.42, 2.60).001 Imipenem 37 (11.1%) 17 (3.2%) 3.86 (2.07, 7.43).001 Pip-tazo 33 (9.9%) 25 (4.6%) 2.27 (1.28, 4.07).002 TMP-SMX 66 (19.9%) 66 (12.2%) 1.78 (1.21, 2.63).002 ESC 70 (21.1%) 63 (11.7) 2.02 (1.37, 2.99).001 Anti-anaerobic 178 (53.6%) 234 (43.3%) 1.51 (1.14, 2.01).003 agent Vancomycin 132 (39.8) 144 (26.7%) 1.82 (1.34, 2.45).001 Microbiology Skin/soft tissue 48 (14.5%) 136 (25.2%) 0.50 (0.34, 0.73).001 isolate Respiratory isolate 185 (55.7%) 203 (37.6) 2.09 (1.57, 2.79).001 Note: only significant associations shown (P.05). TMP-SMX trimethoprim sulfamethoxazole; ESC extended spectrum cephalosporin (ie, ceftriaxone, ceftazidime, cefepime). Amoxicillin/clavulanate, ampicillin/sulbactam, imipenem, metronidazole, clindamycin, chloramphenicol. *Fisher s exact test (categorical variables); Wilcoxon Rank Sum test (continuous variables). Median (interquartile range). Days from hospital admission until P. aeruginosa isolate. Number of days in an ICU in the past 30 days. resistant P. aeruginosa isolates, compared to those with FQRPA. In support of this hypothesis, P. aeruginosa isolates that were imipenem resistant displayed higher levels of cross-resistance compared with FQ-resistant isolates. Hsu et al found that FQRPA and empiric FQ were independently associated with mortality, but there was no attempt to adjust for resistance to other antibiotics. 9 Our study had several potential limitations. A full assessment of all P. aeruginosa isolates was limited by incomplete data. However, there were no substantive differences when Table 2 Multivariable Analysis (Association between FQRPA and Mortality) Variable Unadjusted OR Adjusted OR (95% CI) P value FQ resistance (0.85, 1.83).26 Imipenem (1.13, 2.84).01 resistance Duration of hospitalization* (1.00, 1.02).002 *Days from hospital admission until P. aeruginosa isolate. OR reflects the odds associated with each increase in 1 hospital day. comparing available data (ie, age, sex, anatomic site of culture, susceptibility profile of the organism) for those subjects who were and were not included. Selection bias may also have been introduced by allowing subjects to be included more than once (ie, if they had P. aeruginosa isolates 30 days apart). However, we found no substantive differences in the results of our case control or cohort studies when conducting a secondary analysis in which Percent of Isolates FQ resistant isolates Imipenem resistant isolates Total number of isolate-resistant classes Figure 3 Cross-resistance to other antibiotic classes: FQ-resistant and imipenim-resistant P. aeruginosa isolates ( ).

6 526.e24 The American Journal of Medicine, Vol 119, No 6, June 2006 subjects were included only once. The question of which group represents the optimal control group in case control studies of antimicrobial resistance has been recently debated. We believe the appropriate control group depends on the question being asked. 34 We selected patients with FQSPA because we sought to determine risk factors for FQ resistance among P. aeruginosa isolates. We did not determine if isolates were true pathogens or colonizers. We also did not include details about prior antibiotic exposure in an outpatient setting, because data was inconsistently reported and thus, unreliable for inclusion in meaningful analysis. Our study was conducted in a large tertiary care medical center and the results may not be generalizable to other institutions. In addition, our study focused only on levofloxacin because this was the only FQ used at our institution during the time of the study and susceptibility of P. aeruginosa to this agent was consistently reported. Whether these results are generalizable to other FQs and other institutions is unknown. The prevalence of FQRPA has increased substantially in recent years and is associated with increased hospital charges, but other resistance patterns may have a more significant impact on mortality. Prior exposure to FQs is the only independent risk factor for FQRPA and may result in high levels of cross-resistance to other antibiotics. FQs are no longer adequate for empiric therapy of infections caused by P. aeruginosa, particularly for patients who recently received these drugs. We urge clinicians to avoid the use of FQs when alternative agents are available in order to limit the emergence of resistance in P. aeruginosa and preserve the utility of these agents as effect therapy. ACKNOWLEDGMENTS This work was supported by the Public Health Service Grant DK of the National Institutes of Health (Dr. Lautenbach). This study was also supported in part by an Agency for Healthcare Research and Quality (AHRQ) Centers for Education and Research on Therapeutics cooperative agreement (U18-HS10399). The patient database on which this study was based was originally compiled as part of a study supported by a research grant from Merck Pharmaceuticals. For this current study, however, Merck Pharmaceuticals played no role in the data analysis, interpretation of results, or writing of the manuscript. References 1. Fluit AC, Jones ME, Schmitz FJ, Acar J, Gupta R, Verhoef J. Antimicrobial susceptibility and frequency of occurrence of clinical blood isolates in Europe from the SENTRY antimicrobial surveillance program, 1997 and Clin Infect Dis. 2000;30: Pfaller MA, Jones RN, Doern GV, Kugler K. Bacterial pathogens isolated from patients with bloodstream infection: frequencies of occurrence and antimicrobial susceptibility patterns from the SENTRY antimicrobial surveillance program (United States and Canada, 1997). Antimicrob Agents Chemother. 1998;42: Kang CI, Kim SH, Kim HB, et al. Pseudomonas aeruginosa bacteremia: risk factors for mortality and influence of delayed receipt of effective antimicrobial therapy on clinical outcome. Clin Infect Dis. 2003;37: Obritsch MD, Fish DN, MacLaren R, Jung R. National surveillance of antimicrobial resistance in Pseudomonas aeruginosa isolates obtained from intensive care unit patients from 1993 to Antimicrob Agents Chemother. 2004;48: Andrade SS, Jones RN, Gales AC, Sader HS. Increasing prevalence of antimicrobial resistance among Pseudomonas aeruginosa isolates in Latin American medical centres: 5 year report of the SENTRY Antimicrobial Surveillance Program ( ). J Antimicrob Chemother. 2003;52: Neuhauser MM, Weinstein RA, Rydman R, Danziger LH, Karam G, Quinn JP. Antibiotic resistance among gram-negative bacilli in US intensive care units: implications for fluoroquinolone use. JAMA. 2003;289: Paladino JA, Sunderlin JL, Forrest A, Schentag JJ. Characterization of the onset and consequences of pneumonia due to fluoroquinolonesusceptible or -resistant Pseudomonas aeruginosa. J Antimicrob Chemother. 2003;52: Krcmery V, Mateika F, Krupova I, Trupl J, Kunova A. Bacteremia due to ciprofloxacin resistant Pseudomonas aeruginosa in cancer patients: risk factors for resistance and outcome of 25 episodes. Infect Dis Clin Pract. 1999;8: Hsu DI, Okamoto MP, Murthy R, Wong-Beringer A. Fluoroquinoloneresistant Pseudomonas aeruginosa: risk factors for acquisition and impact on outcomes. J Antimicrob Chemother. 2005;55: Barton T, Fishman NO, Weiner MG, LaRosa L, Lautenbach E. High rate of coadministration of di- or tri-valent cation-containing compounds with oral fluoroquinolones: risk factors and potential implications. Infect Control Hosp Epidemiol. 2005;26: Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45: National Committee for Clinical Laboratory Standards (NCCLS). Performance Standards for Antimicrobial Susceptibility Testing: Eleventh Informational Supplement (NCCLS document M100-S11). Wayne, PA: National Committee for Clinical Laboratory Standards; National Committee for Clinical Laboratory Standards (NCCLS). Performance Standards for Antimicrobial Susceptibility Testing: Tenth Informational Supplement. NCCLS document M100-S10. Wayne, PA: National Committee for Clinical Laboratory Standards; Armitage P. Test for linear trend in proportions and frequencies. Biometrics. 1955;11: Kleinbaum DG, Kupper LL, Morgenstern H. Epidemiologic Research: Principles and Quantitative Methods. New York: Van Nostrand Reinhold; Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst. 1959;22: Hosmer DO, Lemeshow SL. Applied Logistic Regression. New York: Wiley and Sons; Harris AD, Karchmer TB, Carmeli Y, Samore MH. Methodological principles of case-control studies that analyzed risk factors for antibiotic resistance: a systematic review. Clin Infect Dis. 2001;32: Sun J. A non-parametric test for interval censored failure time data with applications to AIDS studies. Stat Med. 1996;15: Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol. 1989;129: Cosgrove SE, Carmeli Y. The impact of antimicrobial resistance on health and economic outcomes. Clin Infect Dis. 2003;36: Chen DK, McGeer A, de Azavedo JC, Low DE. Decreased susceptibility of Streptococcus pneumoniae to fluoroquinolones in Canada. Canadian Bacterial Surveillance Network. N Engl J Med. 1999;341: Lautenbach E, Fishman NO, Bilker WB, et al. Risk factors for fluoroquinolone resistance in nosocomial Escherichia coli and Klebsiella pneumoniae infections. Arch Intern Med. 2002;162:

7 Gasink et al Fluoroquinolone-Resistant Pseudomonas aeruginosa 526.e Polk RE, Johnson CK, McClish D, Wenzel RP, Edmond MB. Predicting hospital rates of fluoroquinolone-resistant Pseudomonas aeruginosa from fluoroquinolone use in US hospitals and their surrounding communities. Clin Infect Dis. 2004;39: Zervos MJ, Hershberger E, Nicolau DP, et al. Relationship between fluoroquinolone use and changes in susceptibility to fluoroquinolones of selected pathogens in 10 United States teaching hospitals, Clin Infect Dis. 2003;37: Carmeli Y, Troillet N, Eliopoulos M, Samore M. Emergence of antibiotic-resistant Pseudomonas aeruginosa: comparison of risks associated with different antipseudomonal agents. Antimicrob Agents Chemother. 1999;43: El Amari EB, Chamot E, Auckenthaler R, Pechere JC, Van Delden C. Influence of previous exposure to antibiotic therapy on the susceptibility pattern of Pseudomonas aeruginosa bacteremic isolates. Clin Infect Dis. 2001;33: Krcmery V, Mateicka F, Krupova I, Trupl J, Kunova A. Bacteremia due to ciprofloxacin-resistant Pseudomonas aeruginosa in cancer patients: risk factors for resistance and outcome of 25 episodes. A case-control study. Infect Dis Clin Pract. 1999;8: Lautenbach E, Strom BL, Bilker WB, Patel JB, Edelstein PH, Fishman NO. Epidemiological investigation of fluoroquinolone resistance in infections due to extended-spectrum -lactamase-producing Escherichia coli and Klebsiella pneumoniae. Clin Infect Dis. 2001;33: Paramythiotou E, Lucet JC, Timsit JF, et al. Acquisition of multidrugresistant Pseudomonas aeruginosa in patients in intensive care units: role of antibiotics with antipseudomonal activity. Clin Infect Dis. 2004;38: Defez C, Fabbro-Peray P, Bouziges N, et al. Risk factors for multidrug-resistant Pseudomonas aeruginosa nosocomial infection. J Hosp Infect. 2004;57: Piddock LJV. Mechanisms of resistance to fluoroquinolones: state-ofthe-art Drugs. 1995;49(Suppl 2): Vidal F, Mensa J, Almela M, et al. Epidemiology and outcome of Pseudomonas aeruginosa bacteremia, with special emphasis on the influence of antibiotic treatment. Analysis of 189 episodes. Arch Intern Med. 1996;156: Harris AD, Karchmer TB, Carmeli Y, Samore MH. Methodological principles of case-control studies that anlayzed risk factors for antibiotic resistance: a systematic review. Clin Infect Dis. 2001;32: