Relationship between increased levofloxacin use and decreased susceptibility of Streptococcus pneumoniae in the United States

Diagnostic Microbiology and Infectious Disease 51 (2005) 31 – 37 www.elsevier.com/locate/diagmicrobio

Pharmacology

Relationship between increased levof loxacin use and decreased susceptibility of Streptococcus pneumoniae in the United StatesB Sujata M. Bhavnania,b,e,*, Jeffrey P. Hammela,f, Ronald N. Jonesc,d, Paul G. Ambrosea,b,e a Cognigen Corporation, Buffalo, NY 14221-5831, USA School of Pharmacy and Pharmaceutical Sciences, University at Buffalo, Buffalo, NY 14221-5831, USA c The JONES Group/JMI Laboratories, North Liberty, IA 52317, USA d Tuft’s University School of Medicine, Boston, MA 02111, USA e Current address: Institute for Clinical Pharmacodynamics, Ordway Research Institute, Albany, NY 12208, USA f Current address: Collaborative Biostatistics Center, Department of Biostatistics and Epidemiology, Cleveland Clinic Foundation, Cleveland, OH 44195-5196, USA Received 30 June 2004; accepted 31 August 2004 b

Abstract Increasing reports of fluoroquinolone-non-susceptible Streptococcus pneumoniae are of clinical concern. We examined the relationship between outpatient fluoroquinolone use and susceptibility of community-acquired S. pneumoniae isolates. Using multivariable general linear modeling, US SENTRY Antimicrobial Surveillance Program and Intercontinental Medical Statistics data (1997–2002) were analyzed to determine the influence of selected patient-, institution-, and geographic region-specific factors, including local fluoroquinolone usage, on the minimum inhibitory concentration (MIC) of levofloxacin against S. pneumoniae. Levofloxacin MIC50, MIC90, and MIC range (n = 384 from 26 hospitals) were 1, 1, and V 0.5 to > 4 Ag/mL, respectively. Variables associated with changes in geometric mean MIC included geographical region ( P b 0.0001), medical service ( P = 0.0002), study year ( P = 0.0006), primary diagnosis group ( P = 0.02), and 2 interactions (duration of hospital stay before isolate collection by bed capacity, P = 0.06, and levofloxacin use by geographical region, P = 0.08; P b 0.001 when study year was removed from the model). MIC increased with levofloxacin use across all geographical regions, with increases of 54% and 126% in the southwest and west, respectively. In contrast to other fluoroquinolones, increased levofloxacin use, along with other variables, was associated with decreased pneumococcal susceptibility. Given the US environment of increasing pneumococcal resistance, these data may be useful in better understanding factors related to emergence of fluoroquinolone resistance. D 2005 Published by Elsevier Inc. Keywords: Levofloxacin; Streptococcus pneumoniae; Fluoroquinolone resistance; United States

1. Introduction Streptococcus pneumoniae remains a leading cause of morbidity and mortality in young children, patients with debilitating diseases, and the elderly in the United States,

B

This Antimicrobial Resistance Rate Epidemiology Study Team analysis was not supported by any pharmaceutical company. This work was presented at the 43rd Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, September 16, 2003 (abstract no. K-1397). * Corresponding author. Tel.: +1-518- 641- 6473; fax: +1-518- 6416303. E-mail address: [email protected] (S.M. Bhavnani). 0732-8893/$ – see front matter D 2005 Published by Elsevier Inc. doi:10.1016/j.diagmicrobio.2004.08.017

with over 7 million cases of pneumonia, acute bacterial exacerbation of chronic bronchitis, sinusitis, and otitis media per year (Centers for Disease Control and Prevention, 1996). Over the last decade, multidrug-resistant S. pneumoniae has emerged as a clinical concern, resulting in revised treatment guidelines (Heffelfinger et al, 2000; Mandell et al, 2000, 2003; Niederman et al, 2001) and subsequent changes in prescribing patterns. During the same period, fluoroquinolone use has steadily increased, especially in the treatment of community-acquired respiratory tract infections. Identification of S. pneumoniae with reduced susceptibility to respiratory fluoroquinolones has been increasing in

32

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37

North America and several other parts of the world (Brueggemann et al, 2002; Chen et al, 1999; Ho et al, 1999; Jones & Pfaller, 2000). Although the incidence of fluoroquinolone-resistant S. pneumoniae remains low (b 1% in the United States; Brueggemann et al, 2002; Jones & Pfaller, 2000), recently reported cases of fluoroquinolone treatment failures, particularly with levofloxacin, among patients with pneumococcal respiratory infections warrant concern (Anderson et al, 2003; Davidson et al, 2002; Kays et al, 2002; Ross et al, 2002; Urban et al, 2001; Wortmann & Bennett, 1999). There is some evidence suggesting that the increased prevalence of pneumococci with reduced fluoroquinolone susceptibility is related to the increased use of quinolones (Beekmann et al, 2003; Centers for Disease Control and Prevention, 2001; Chen et al, 1999; Ho et al, 2001). However, these reports have been limited in number and have failed to differentiate among fluoroquinolones in terms of selection of pneumococcal resistance. Given that S. pneumoniae is primarily a pathogen involved in community-acquired respiratory infections, evaluation of the relationship between outpatient fluoroquinolone use and pneumococcal susceptibility is of relevance. Using both quantitative longitudinal microbiological surveillance data from patients with community-acquired pneumococcal infections and regional outpatient drug use data, the Antimicrobial Resistance Rate Epidemiology Study Team (Bhavnani et al, 2003b) carried out the analysis presented herein to determine the influence of selected patient-, institution-, and geographic region-specific factors, including local fluoroquinolone use, on the susceptibility of S. pneumoniae to levofloxacin.

2. Methodology 2.1. Data and descriptive statistics All relevant antimicrobial susceptibility, patient- and institution-specific data for S. pneumoniae blood isolates collected from US hospitals participating in the SENTRY Antimicrobial Surveillance Program (1997– 2002) were queried for analysis. Annual regional quinolone use data for the same period were obtained from Intercontinental Medical Statistics and matched using designated bMetropolitan Statistical AreasQ to the geographical region surrounding each SENTRY Program hospital. The primary outcome variable was the in vitro activity of levofloxacin against S. pneumoniae, which was measured by the reference broth microdilution minimum inhibitory concentration (MIC; National Committee for Clinical Laboratory Standards, [NCCLS], 2003). To obtain an integer value and approximately normal distribution for MIC values, a log2 transform was applied to this measure. The resulting log2 MIC was integer valued, but its model error term probabilities were assumed to be well approximated by a normal distribution with a continuity correction factor. The MIC50 (median) and MIC90

were used to further describe the susceptibility distribution of S. pneumoniae to levofloxacin. 2.2. Assessment of relationships to susceptibility of isolates Using univariable examinations and graphical representations, the apparent relationship between the MIC value of isolates and each independent variable was evaluated. Independent variables examined included age and gender of patient, duration of hospital stay before pathogen isolation, residence in the intensive care unit, primary diagnosis group, primary risk factor for infection, source of infection, and medical service category. Institution-specific characteristics included bed capacity, geographic region, and study year. Annual regional Intercontinental Medical Statistics drug-use data for ciprofloxacin, gatifloxacin, levofloxacin, moxifloxacin, ofloxacin, and trovafloxacin were each considered as independent variables. Annual drug-use data for fluoroquinolone agents commonly used to treat patients with respiratory infections (gatifloxacin, levofloxacin, moxifloxacin, ofloxacin, and trovafloxacin) were combined and also considered separately as an independent variable. Data for each agent were normalized by the population for the geographical region surrounding each hospital (Metropolitan Statistical Area) and expressed as the number of prescriptions per 100 persons. Using S-plus (version 6.0.1 for UNIX), tree-based modeling was performed as a means of describing potentially complex relationships between the independent variables and/or impressive differences in MIC values. The resulting tree models were used to identify possible interactions between independent variables to be considered for inclusion in the multivariable regression model and to suggest important breakpoints within continuous independent variables. Partitioning of the tree by at least 1 dilution that occurred within 1 tree branch, but not its complement, after the first or second split was thought to indicate a potential interaction between the sets of variables on which the splits occurred. Cost-complexity pruning was used to maintain manageable tree size; attention was directed on the early tree splits. Multivariable general linear modeling tailored to accommodate data for a dependent variable with right and left censoring was performed with backward elimination (at P N 0.1) using the LIFEREG procedure within SAS version 8.2 (SAS Institute). For the purposes of these analyses, right- and left-censored data represented MIC values above or below the upper or lower margins of the MIC range tested, respectively. For example, a MIC observation of V 0.25 is left censored at a censoring value of 0.25, and a MIC observation of N 4 is right censored at a censoring value of 4. Continuous independent variables were entered into the analysis as such or categorized into subgroups (using breakpoints to define interpretable subgroups of sufficient size) to account for potential nonlinear relationships. The principal measure of association between MIC value of levofloxacin against S. pneumoniae and each independent

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37

Fig. 1. Geographical distribution of the SENTRY Program hospitals included in the analysis (n = 26).

variable of interest was a corresponding parameter estimate (with S.E.), which described the estimated difference in the MIC50 value associated with varying levels of a categorical independent variable or with a 1-unit change in a continuous independent variable. 2.3. Assessment of model precision As explained previously (Bhavnani et al, 2003b), the maximum likelihood estimate of the model error variance was used to estimate the proportion of the variability in MIC explained by the independent variables in the final model (denoted as R 2). The average model-predicted log2 MIC value for an isolate within institutions, across all study years and for specific study years, was computed. A weighted Spearman correlation measure, the square value of which was denoted as R S2, was used to assess the strength of association between these predicted and observed institutional means of the log2 MIC. 3. Results Between 1997 and 2002, 384 nonclonal S. pneumoniae blood isolates with complete institution- and patient-specific

Fig. 2. Population-adjusted fluoroquinolone use trends from 1997 to 2002.

33

information were collected from 26 US hospitals participating in the SENTRY Program. As shown in Fig. 1, 8 and 6 of these hospitals were in the Midwest and Northeast regions of the United States, respectively, whereas the remaining regions were each represented by 4 hospitals. Similar proportions of isolates were collected from hospitals with bed capacities V 400 and 401–900, with 90% of all isolates from hospitals with bed capacities of V 900. All isolates were collected consecutively from unique patients. About 90% of isolates were collected from patients who had resided in the hospital b 1 day. The majority of patients were admitted to a medical service, with more than half having a cardiopulmonary-related primary diagnosis. The variability in observed MIC was narrow, with an estimated standard deviation of 0.59 on the log2 scale and a range of observed MIC values from V 0.25 to N 4 Ag/mL. There were 71 (18%) left-censored observations with MIC values of V 0.25 or V 0.5 Ag/mL and 1 right-censored Table 1 Parameter estimates for the final general linear model Variable

Intercept Study year 1997 1998 2000 2001 2002b Primary diagnosis Cardiopulmonary Gastrointestinal, abdominal, or hepatic Genitourinary Immunocompromised Infection Neurological Trauma Other Medical service Acute care Medicine Pediatrics Surgery Other Geographic region by levofloxacin use interactionc Prior duration of hospital stay by bed capacity interactionc

Estimate

Maximal dilution changea

P

0.4125

0.0006

0.6503

0.021

0.5223

0.0002

0.899 0 0.376 0.047 0.147 0.037 0.007 0.060

0.008 0.021 0.396 0.270 0.591 0 0.091 0 0.243 0.155 0.280 0.080

0.056

a This represents the maximal change in log2 MIC value among parameter estimates for a categorical variable. b Note that the sample size of isolates was limited for 2002 because only partial data were available at the time of the analysis. c Two-way interactions included duration of hospital stay before pathogen isolation by hospital bed capacity and geographical region by levofloxacin use. P values for main effects measuring associations within reference categories ranged from b 0.0001 to 0.07.

34

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37

observation (MIC N 4 Ag/mL). With the exception of 2002 (a study year for which only partial data were available at the time of the analysis), mean MIC values increased over the study period. Mean MIC values were 50% higher for isolates collected in 2001 as compared with those collected in 1997. Fig. 2 shows the population-adjusted fluoroquinolone use patterns during the study period among the Metropolitan Statistical Areas surrounding each hospital for all isolates. The number of prescriptions of levofloxacin per 100 people increased dramatically from 1997 to 2002, from 0.4 to 4. By comparison, the increase in use of gatifloxacin and moxifloxacin was modest, whereas use of ofloxacin and trovafloxacin declined over the study period. The final multivariable general linear model is summarized in Table 1. Seven independent variables, either alone or as 2-way interactions, were found to be significantly related to changes in MIC of S. pneumoniae against levofloxacin. Mean changes in predicted MIC values between subcategories of primary diagnosis and medical service were as high as 57% and 44%, respectively. Significant 2-way interactions included geographic region by levofloxacin use and duration of hospital stay before pathogen isolation by hospital bed capacity. Although the P value for the association between increase in levofloxacin use across geographical regions and increase in MIC was 0.078, this relationship became highly significant ( P b 0.001) when a general linear model that excluded study year was evaluated. Relationships between the use of other fluoroquinolones (either alone or as part of a 2-way interaction) and MIC failed to be significant ( P values of 0.12 – 0.82 were observed during the stepwise model building process). Furthermore, the evaluation of the com-

Fig. 4. Change in levofloxacin MIC for S. pneumoniae versus levofloxacin use by geographical region. The bolded line represents the average across all geographical regions.

bined use for the fluoroquinolone agents used to treat patients with respiratory infections also failed to be significant. The final model, which was fit to data from all isolates, explained a moderate proportion of the variability in MIC among isolates (R 2 = 0.44). However, when fit by institution as shown in Fig. 3, model-predicted mean MIC within hospitals was strongly correlated with observed mean MIC averaged over all study years (R S2 = 0.84). A high correlation was also observed when averages were computed for institution by year (R S2 = 0.78). Fig. 4 demonstrates the predicted dilution change in MIC across categorical levels of levofloxacin use for each of the geographical regions. Categories of levofloxacin use were more than 3 prescriptions per 100 people in 4 of the 5 geographic regions. As noted from the lines representing individual regions, increased levofloxacin use was generally associated with increased MIC in all regions. On average (shown with the bolded line), MIC values were ~ 60% higher in locations with high compared with low levofloxacin use. The largest MIC increases corresponding with increasing levofloxacin use were observed in the southwest (126%) and west (48%). 4. Discussion

Fig. 3. Mean model-predicted versus mean observed MIC at the institution level. Data points represent averages within individual hospitals. The size of the data point is proportional to the number of isolates collected from the hospital.

Data from this study demonstrate both a significant increase in the levofloxacin MIC for S. pneumoniae isolates over the study period and a relationship between increasing levofloxacin MIC values and increasing levofloxacin use. The finding that study year was a significant predictor of MIC was not surprising, as MIC values had increased by year over the study period. When study year was included in the final model, the association between levofloxacin use (by geographical region) and MIC achieved only borderline significance; however, when study year was excluded from the model, this association became much more impressive ( P b 0.001). The reason for this seems clear: both levofloxacin

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37

MIC values and levofloxacin use increased by year throughout the study period. Thus, the collinearity between levofloxacin use and study year prevented both variables from emerging as highly significant predictors of MIC when both were included in the final model. A similar relationship between drug use and study year was observed when assessing the impact of increasing levofloxacin use on the susceptibility of Pseudomonas aeruginosa to ciprofloxacin (Bhavnani et al, 2003a). The fact that levofloxacin use was associated with S. pneumoniae MIC values, either as a main effect or as part of an interaction with geographical region, is likewise not surprising. Several studies have reported associations between increasing use of various fluoroquinolone agents and increasing frequency of resistant isolates. However, few studies have attempted to distinguish the attributable risk of individual agents within the class. A common misconception, as evidenced by analyses in which all agents within a class are pooled, is that the pharmacokinetics and pharmacodynamics of individual agents within a class are similar. Over the last 2 decades, our knowledge of the relationship between plasma and tissue concentrations and the corresponding pharmacological effects of antimicrobial agents has dramatically improved. Using data from clinical and nonclinical models of infection, it has been demonstrated that dosing regimens can be optimized to achieve pharmacokinetic – pharmacodynamic (PK–PD) target measures associated with efficacy. In the case of S. pneumoniae, fluoroquinolone free-drug area under the concentration-time curve over a 24-hour interval (AUC0 – 24)/MIC ratios of > 30 represent the PK–PD target associated with bacterial eradication in both nonclinical models of infection and infected patients (Ambrose et al, 2001; Craig & Andes, 2000; Lacy et al, 1999; Lister, 2002; Lister & Sanders, 1999, 2001). This PK–PD target, together with pharmacokinetic data and regional MIC distributions, has been used to estimate the probability of PK–PD target attainment for standard dosing regimens of fluoroquinolones against pneumococci (Jones et al, 2003; Owens & Ambrose, 2002). Using MIC distributions for isolates collected worldwide, we have demonstrated previously that gatifloxacin was consistently more potent than levofloxacin against S. pneumoniae and had a higher probability of achieving important PK–PD target exposure ratios (Jones et al, 2003). For example, based on MIC distributions for isolates collected in North America, the probability of attaining a free-drug AUC024/MIC ratio of >30 for gatifloxacin at 400 mg/day and levofloxacin at 500 mg/day was 97.6% and 78.9%, respectively. Similar PK–PD target attainment results have been reported when standard doses of moxifloxacin, 400 mg/day, were compared with levofloxacin, 500 mg/day (99.3% versus 87.4%; Owens & Ambrose, 2002). In the present analysis, increased use of levofloxacin specifically, as opposed to other fluoroquinolones, was associated with increases in the MIC of S. pneumoniae. A failure of standard levofloxacin dosing

35

regimens of levofloxacin to achieve a high probability of PK–PD target attainment may be an important contributing factor to the results seen in this analysis. The relationship between antimicrobial use and bacterial resistance has been well documented in the literature (Chen et al, 1999; McGowan, 1983; Seppala et al, 1995). However, not all studies evaluating use and resistance have been able to demonstrate a statistically significant association. The results of studies in which no association is observed may be due to a genuine lack of association; however, it may also be due in part to small sample size, limited duration of observation, poor quality of data and/or limitations of study design, or existence of interactions with other variables not captured by the investigator. Perhaps the most important limitation of epidemiological evaluations of antimicrobial use and bacterial resistance data has been the qualitative nature of microbiological data. Such analyses usually involve relating the proportion of isolates that are susceptible or resistant to the level of antimicrobial use in the institution or region. However, the lack of a significant association between the volume of an antimicrobial agent used and the proportion of resistant isolates may not mean that a relationship is lacking. The difficulty of using ranked susceptibility data (ie, susceptible, intermediate, and resistant) is that large MIC increases are often required to detect a change in susceptibility from susceptible to resistant. Identification of factors associated with shifts in MIC, even in the lower margin of the MIC distribution (ie, within the susceptible range), is also meaningful. Analysis of quantitative MIC data is more sensitive than qualitative susceptibility data and thus increases the likelihood of identifying risk factors for the development of resistance (Bhavnani, 2003). Despite the greater sensitivity achieved when using quantitative MIC data, the censored nature of these data presents an analytical challenge. However, the use of modeling techniques tailored to accommodate censored data, as described in this report, allowed for the estimation of the likelihood of higher MIC values based on the presence of certain factors. Detection of such relationships, especially when a low burden of resistant isolates exist, represents a step forward in the effort to better understand the predictors of resistance. Modeling approaches such as these may be applied to identify subgroups of patients at increased risk for infection arising from organisms with increased MIC. The increased frequency of clinical failures associated with levofloxacin therapy (Anderson et al, 2003; Davidson et al, 2002; Kays et al, 2002; Ross et al, 2002; Urban et al, 2001; Wortmann & Bennett, 1999) has resulted in a call for the implementation of routine fluoroquinolone susceptibility testing of pneumococcal isolates. Even with increased surveillance of such an in vitro measure, however, the true impact of increasing fluoroquinolone use on the susceptibility of S. pneumoniae may go undetected. The development of bacterial resistance to fluoroquinolones is mediated by mutation of 2 intracellular

36

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37

targets (DNA gyrase and DNA topoisomerase IV). Under the current interpretative guidelines of the NCCLS (2003), pneumococcal strains with only a first-step mutation may be identified as levofloxacin susceptible because a second mutation is usually required to raise the MIC above the breakpoint for resistance. However, the probability of a second mutation is greatly increased after the development of a first mutation (Gillespie et al, 2003). Thus, although the current rates of S. pneumoniae resistance to levofloxacin appear to be generally low, the steady increases in the proportion of higher MIC values (Karlowsky et al, 2003; Thornsberry et al, 1999, 2002) and mutations in the fluoroquinolone resistance-determining regions of parC and gyrA genes (Beekmann et al, 2003; Davies et al, 2002; Lim et al, 2003) merit consideration. Beekmann et al (2003) reported a significant increase in prevalence of strains with one or both of these mutations over time: 1994 –1995, 1.6%; 1997–1998, 1.4%; 1999 – 2000, 1.0%; and 2001– 2002, 4.1%. Similarly, Davies et al (2002) estimated that 3.3% of 9438 levofloxacin-susceptible (MIC = 0.06-2) TRUST 4 surveillance isolates (1999 – 2000) had a single parC mutant. Moreover, in their examination of 82 S. pneumoniae isolates collected throughout Canada between 1993 and 1998 with a levofloxacin MIC value of 2 Ag/mL, a concentration considered susceptible by NCCLS criteria (NCCLS, 2003, 2004), Lim et al (2003) identified a total of 48 (59%) with a first-step mutation in parC. In addition to the increased prevalence in strains with mutations, examination of US data collected from the TRUST program demonstrates an increasing proportion of MIC values of 1 Ag/mL from 1997 – 1998 (14.9%) to 2001 – 2002 (53.4%) (Karlowsky et al, 2003; Thornsberry et al, 1999, 2002). Such data suggest that bacterial resistance to fluoroquinolones may be a silent, yet looming, problem and that prospective identification of risk factors associated with changes in the MIC and appropriate intervention will enhance the ongoing effort to arrest further progression of resistance. References Ambrose PG, Grasela DM, Grasela TH, Passarell J, Mayer HB, Pierce PF (2001) Pharmacodynamics of fluoroquinolones against Streptococcus pneumoniae in patients with community-acquired respiratory tract infections. Antimicrob Agents Chemother 45:2793 – 2797. Anderson KB, Tan JS, File Jr TM, DiPersio JR, Willey BM, Low DE (2003) Emergence of levofloxacin-resistant pneumococci in immunocompromised adults after therapy for community-acquired pneumonia. Clin Infect Dis 37:376 – 381. Beekmann SE, Heilmann K, Sauter C, Miller N, Richter SS, Doern GV (2003) Increasing fluoroquinolone resistance in S. pneumoniae: An association with fluoroquinolone use. Program and Abstracts of the Forty-Third Interscience Conference on Antimicrobial Agents and Chemotherapy, Chicago, IL, 2003. Abstract C2-1498. Washington, DC7 American Society for Microbiology, p 140. Bhavnani, SM (2003) Resistance surveillance, pharmacokinetics/pharmacodynamics, and simulation: Can the use of old tools be improved to impact antibiotic resistance? Program and Abstracts of the Forty-First Infectious Diseases Society of America, San Diego, CA, 2003. Speaker 1005. Alexandria, VA7 Infectious Diseases Society of America, p 33.

Bhavnani SM, Callen WA, Forrest A, Gilliland KK, Collins DA, Paladino JA, Schentag JJ (2003a) Effect of fluoroquinolone expenditures on susceptibility of Pseudomonas aeruginosa to ciprofloxacin in U.S. hospitals. Am J Health Syst Pharm 60:1962 – 1970. Bhavnani SM, Hammel JP, Forrest A, Jones RN, Ambrose PG (2003b) Relationships between patient- and institution-specific variables and decreased susceptibility of Gram-negative pathogens. Clin Infect Dis 37: 344 – 350. Brueggemann AB, Coffman SL, Rhomberg P, Huynh H, Almer L, Nilius A, Flamm R, Doern GV (2002) Fluoroquinolone resistance in Streptococcus pneumoniae in United States since 1994–1995. Antimicrob Agents Chemother 46:680 – 688. Centers for Disease Control and Prevention (1996) Defining the public health impact of drug-resistant Streptococcus pneumoniae: Report of a working group. MMWR Morb Mortal Wkly Rep 45:1 – 20. Centers for Disease Control and Prevention (2001) Resistance of Streptococcus pneumoniae to fluoroquinolones—United States, 1995– 1999. MMWR Morb Mortal Wkly Rep 50:800 – 804. Chen DK, McGeer A, de Azavedo JC, Low DE (1999) Decreased susceptibility of Streptococcus pneumoniae to fluoroquinolones in Canada. N Engl J Med 341:233 – 239. Craig WA, Andes DA (2000) Correlation of the magnitude of the AUC24/ MIC for 6 fluoroquinolones against Streptococcus pneumoniae with survival and bactericidal activity in an animal model. Abstracts of the Fortieth Interscience Conference of Antimicrobial Chemotherapy, Toronto, Canada. Abstract 289 Washington, DC7 American Society for Microbiology, p 7. Davidson R, Cavalcanti R, Brunton JL, Bast DJ, de Azavedo JC S, Kibsey P, Fleming C, Low DE (2002) Resistance to levofloxacin and failure of treatment of pneumococcal pneumonia. N Engl J Med 346:747 – 750. Davies TA, Evangelista A, Pfleger S, Bush K, Sahm DF, Goldschmidt R (2002) Prevalence of single mutations in topoisomerase type II genes among levofloxacin-susceptible clinical strains of Streptococcus pneumoniae isolated in the United States in 1992 to 1996 and 1999 to 2000. Antimicrob Agents Chemother 46:119 – 124. Gillespie SH, Voelker LL, Ambler JE, Traini C, Dickens A (2003) Fluoroquinolone resistance in Streptococcus pneumoniae: evidence that gyrA mutations arise at a lower rate and that mutation in gyrA or parC predisposes to further mutation. Microb Drug Resist 9:17 – 24. Heffelfinger JD, Dowell SF, Jorgensen JH, Klugman KP, Mabry LR, Musher DM, Plouffe JF, Rakowsky A, Schuchat A, Whitney CG (2000) Management of community-acquired pneumonia in the era of pneumococcal resistance: a report from the drug-resistant Streptococcus pneumoniae Therapeutic Working Group. Arch Intern Med 160: 1399 – 1408. Ho PL, Que TL, Tsang DN, Ng TK, Chow KH, Seto WH (1999) Emergence of fluoroquinolone resistance among multiply resistant strains of Streptococcus pneumoniae in Hong Kong. Antimicrob Agents Chemother 43:1310 – 1313. Ho PL, Tse WS, Tsang KWT, Kwok TK, Ng TK, Cheng VCC, Chan RMT (2001) Risk factors for acquisition of levofloxacin resistant Streptococcus pneumoniae: A case–control study. Clin Infect Dis 32: 701 – 707. Jones RN, Pfaller MA (2000) Macrolide and fluoroquinolone (levofloxacin) resistances among Streptococcus pneumoniae strains: Significant trends from the SENTRY Antimicrobial Surveillance Program (North America, 1997–1999). J Clin Microbiol 38:4298 – 4299. Jones RN, Rubino CM, Bhavnani SM, Ambrose PG (2003) Worldwide antimicrobial susceptibility patterns and pharmacodynamic comparisons of gatifloxacin and levofloxacin against Streptococcus pneumoniae: Report from the Antimicrobial Resistance Rate Epidemiology Study Team. Antimicrob Agents Chemother 47:292 – 296. Karlowsky JA, Thornsberry C, Jones ME, Evangelista AT, Critchley IA, Sahm DF (2003) Factors associated with relative rates of antimicrobial resistance among Streptococcus pneumoniae in the United States: Results from the TRUST Surveillance Program (1998–2002). Clin Infect Dis 36:963 – 970.

S.M. Bhavnani et al. / Diagnostic Microbiology and Infectious Disease 51 (2005) 31–37 Kays MB, Smith DW, Wack ME, Denys GA (2002) Levofloxacin treatment failure in a patient with fluoroquinolone-resistant Streptococcus pneumoniae pneumonia. Pharmacotherapy 22:395 – 399. Lacy MA, Lu W, Xu X, Tessier PR, Nicolau DP, Quintiliani R, Nightingale CH (1999) Pharmacodynamic comparisons of levofloxacin, ciprofloxacin and ampicillin against Streptococcus pneumoniae in an in vitro model of infection. Antimicrob Agents Chemother 43 672 – 677. Lim S, Blast D, McGeer A, de Azavedo J, Low DE (2003) Antimicrobial susceptibility breakpoints and first-step parC mutations in Streptococcus pneumoniae: Redefining fluoroquinolone resistance. Emerg Infect Dis 9: 833 – 837. Lister PD (2002) Pharmacodynamics of gatifloxacin against Streptococcus pneumoniae in an in vitro pharmacokinetic model: Impact of area under the curve / MIC ratios on eradication. Antimicrob Agents Chemother 46: 69 – 74. Lister PD, Sanders CC (1999) Pharmacodynamics of levofloxacin and ciprofloxacin against Streptococcus pneumoniae. J Antimicrob Chemother 43:79 – 86. Lister PD, Sanders CC (2001) Pharmacodynamics of moxifloxacin, levofloxacin, and sparfloxacin against Streptococcus pneumoniae in an in vitro pharmacodynamic model. J Antimicrob Chemother 47: 811 – 818. Mandell LA, Bartlett JG, Dowell SF, File Jr M, Musher DM, Whitney C (2003) Update of practice guidelines for the management of communityacquired pneumonia in immunocompetent adults. Clin Infect Dis 37: 1405 – 1433. Mandell LA, Marrie TJ, Grossmann RF, Chow AW, Hyland RH (2000) Canadian guidelines for the initial management of community-acquired pneumonia: An evidence-based update by the Canadian Infectious Diseases Society and the Canadian Thoracic Society. The Canadian Community-Acquired Pneumonia Working Group. Clin Infect Dis 31: 383 – 421. McGowan Jr JE (1983) Antimicrobial resistance in hospital organisms and its relation to antibiotic use. Reviews in Infectious Diseases 5:1033 – 1048. National Committee for Clinical Laboratory Standards (NCCLS) (2003) Performance standards for antimicrobial susceptibility testing ( Document M7-A6 ). Wayne, PA7 NCCLS.

37

National Committee for Clinical Laboratory Standards (NCCLS) (2004) Performance standards for antimicrobial susceptibility testing ( Document M100-S14 ). Wayne, PA7 NCCLS. Niederman MS, Mandell LA, Anzueto A, Bass JB, Broughton WA, Campbell GD, Dean N, File T, Fine MJ, Gross PA, Martinez F, Marrie TJ, Plouffe JF, Ramirez J, Sarosi GA, Torres A, Wilson R, Yu VL, American Thoracic Society (2001) Guidelines for the management of adults with community-acquired lower respiratory tract infections: Diagnosis, assessment of severity, antimicrobial therapy and prevention. Am J Respir Crit Care Med 163:1730 – 1754. Owens RC, Ambrose PG (2002). Minimum efficacy and mutation prevention exposure target attainment rates against S. pneumoniae for standard-dose levofloxacin (500 mg) and high-dose levofloxacin (750 mg) and moxifloxacin. Program and Abstracts of the Fortieth Infectious Diseases Society of America, Chicago, IL, 2002. Abstract 67. Alexandria, VA7 Infectious Diseases Society of America, p 56. Ross JJ, Worthington MG, Gorbach SL, Tillotson GS, Zhao X, Drlica K, Zhanel GG, Hoban DJ, Chan CK, Hutchinson J, Low DE, de Azavedo J, Bast D (2002) Resistance to levofloxacin and failure of treatment of pneumococcal pneumonia (letter). N Engl J Med 347:65 – 67. Seppala H, Klaukka T, Lehtonen R, Nenonen E, Huovinen P (1995) Outpatient use of erythromycin: Link to increased erythromycin resistance in group A streptococci. Clin Infect Dis 21:1378 – 1385. Thornsberry C, Jones ME, Hickey M L, Mauriz Y, Kahn J, Sahm DF (1999) Resistance surveillance of Streptococcus pneumoniae, Haemophilus influenzae and Moraxella catarrhalis isolated in the United States, 1997–1998. J Antimicrob Chemother 44:749 – 759. Thornsberry C, Sahm DF, Kelly LJ, Critchley IA, Jones ME, Evangelista AT, Karlowsky JA (2002) Regional trends in antimicrobial resistance among clinical isolates of Streptococcus pneumoniae, Haemophilus influenzae, and Moraxella catarrhalis in the United States: Results from the TRUST Surveillance Program, 1999–2000. Clin Infect Dis 34(Suppl. 1):S4 – S16. Urban C, Rahman N, Zhao X, Mariano N, Segal-Maurer S, Drlica K, Rahal JJ (2001) Fluoroquinolone-resistant Streptococcus pneumoniae associated with levofloxacin therapy. J Infect Dis 184:794 – 798. Wortmann GW, Bennett SP (1999) Fatal meningitis due to levofloxacinresistant Streptococcus pneumoniae. Clin Infect Dis 29:1599 – 1600.