Evaluation of Clostridium difficile-associated diarrhea with a drug formulary change in preferred fluoroquinolones.

Abstract

Recent publications report that gatifloxacin might be associated with a greater incidence of Clostridium difficile-associated diarrhea (C. difficile, CDAD) than are other fluoroquinolones. We performed a drug use evaluation to examine this issue after adding gatifloxacin to the formulary and changing from levofloxacin to gatifloxacin as the preferred oral fluoroquinolone in 10 Department of Veterans Affairs (VA) medical centers in the northeastern United States. To estimate (1) the overall incidence of CDAD before and after the change from levofloxacin to gatifloxacin as the preferred oral fluoroquinolone and (2) the incidence rates for ciprofloxacin, levofloxacin, and gatifloxacin separately. Using the VA's Pharmacy Benefits Management database, the total number of days of antibiotic therapy was determined for all inpatients and outpatients of the 10 medical centers who filled at least 1 antibiotic prescription between July 1, 2003, and June 30, 2004. This time frame was chosen because it included 6 months before and 6 months after the change in the preferred oral fluoroquinolone from levofloxacin to gatifloxacin for the VA health system on January 1, 2004. For the same study period and medical centers, the electronic medical records of all inpatients and outpatients with an entry for a positive C. difficile toxin were reviewed. Positive toxins that occurred within 6 weeks of a previous positive result in the same patient were excluded. Exact Poisson tests were used to compare the incidence rates of CDAD (number of CDAD cases per 1,000 days of antibiotic treatment) for antibiotics overall, the fluoroquinolones as a group, non-fluoroquinolone antibiotics, and the individual fluoroquinolones, com-paring the 6-month time periods before (pre-change) versus after (post-change) the addition of gatifloxacin as the preferred oral fluoroquinolone. Of 505 cases of CDAD in the 12-month study period, 413 (81.7%) were associated with antibiotic use in the previous 6 weeks. Among anti-biotic users, incidence rates of CDAD were 166 per 72,114 days of antibiotic therapy in the pre-change period (2.3 cases per 1,000 days of antibiotics) versus 247 per 72,354 days in the post-change period (3.4 cases per 1,000 days of antibiotics, P < 0.001). Fluoroquinolones accounted for 54.8% of the CDAD cases in the pre-change period and 67.2% in the post-change period, representing a 22.6% relative increase in the percentage of CDAD cases that were associated with fluoroquinolone use. The CDAD incidence rates per 1,000 days of fluoroquinolone therapy were 3.7 in the pre-change period versus 7.0 in the post-change period (P < 0.001). Among fluoroqui-nolone users, gatifloxacin accounted for none of the cases of CDAD in the pre-change period when it was nonformulary and 65.1% of the cases in the post-change period, for an incidence rate of 7.6 (108 per 14,239 days). The CDAD incidence rates per 1,000 antibiotic days for patients treated with ciprofloxacin were 4.6 (24 per 5,260 days) in the pre-change period and 7.4 (40 per 5,429 days) in the post-change period, a nonsignificant trend (P = 0.079). The incidence rate of CDAD for levofloxacin increased significantly from 3.9 (75 per 19,417 days) in the pre-change period to 10.7 (44 per 4,108 days) in the post-change period (P < 0.001). The incidence rates of CDAD in the post-change period did not differ significantly for ciprofloxacin, levofloxacin, and gatifloxacin (P = 0.119). There was an increase in the incidence of CDAD among all antibiotic users and fluoroquinolone users, but not among users of non-fluoroquinolone antibiotics, in the period following the formulary change from levofloxacin to gatifloxacin as the preferred fluoroquinolone. However, rates of CDAD among the 3 fluoroquinolone antibiotics in the post-change period were not significantly different, and levofloxacin was the only fluoroquinolone that was associated with a significant increase in the rate of CDAD between the pre-change and post-change periods. These findings suggest that the increase in the CDAD incidence rate was probably not attributable to the addition of gatifloxacin to the formulary.