Relationship Between Increased Levofloxacin Use and Decreased Susceptibility of Streptococcus pneumoniae: Report From the ARREST Program

Authors: Bhavnani SM, Hammel J, Jones RN, Ambrose PG
Conference: ICAAC
Keywords: anti-infectives, antibiotic, antimicrobial, censored data, ciprofloxacin, correlation, gatifloxacin, general linear modeling, infectious disease, MIC, minimum inhibitory concentration, moxifloxacin, ofloxacin, quinolone, S.pneumoniae, sensitivity, trovafloxacin
Division: Cognigen

Abstract

Background: Increasing reports describing the emergence of quinolone nonsusceptible Streptococcus pneumoniae (SP) are of clinical concern. We examined the relationship between outpatient quinolone use and susceptibility of community-acquired SP isolates.

Methods: Using multivariable general linear modeling for censored data with backward stepwise elimination (p>0.1), 6 yrs (1997-2002) of U.S. SENTRY Program and IMS data were analyzed to determine relationships between levofloxacin (LEV) MIC and certain variables among patients with SP infection including: age, duration of hospital stay prior to isolate collection, primary diagnosis group, medical service, hospital bed count, geographical region, and study year. Since 95% of isolates were community-acquired, outpatient quinolone use for the region surrounding each hospital (Rx/100 persons for ciprofloxacin, gatifloxacin, LEV, moxifloxacin, ofloxacin & trovafloxacin) was considered.

Results: LEV MIC50, MIC90, and MIC range were (n=384 from 26 hospitals): 1.0, 1.0, and ≤0.25 to >4.0, respectively. Significant variables associated with changes in the geometric mean of the MIC included: geo region (p<0.0001), medical service (p=0.0002), study year (p=0.0006), bed count (p=0.001), primary diagnosis group (p=0.02), LEV use (p=0.02), duration of hospital stay prior to isolate collection (p=0.07) and 2 interactions, duration of hospital stay prior to isolate collection*bed count (p=0.06) and LEV use*geographical region (p=0.08). MIC increased with LEV use across all geographical regions. Within the Southwest and West, LEV use increases from ≤0.4 to >3 and ≤0.4 to 1.5-3 Rx/100 were associated with MIC increases of 54% and 126%, respectively. The model explained 43% of the MIC variance. Across hospitals, observed vs. fitted within-hospital mean MIC values were highly correlated (weighted Spearman R=0.84).

Conclusions: Increased LEV use, in addition to other variables, was associated with decreased SP susceptibility. Given an environment of increasing SP resistance, these data may be useful in better understanding factors related to quinolone resistance.

Interscience Conference on Antimicrobial Agents and Chemotherapy (ICAAC), Chicago, Illinois, September 2003

SM Bhavnani, JP Hammel, RN Jones, PG Ambrose