The Incidence of Clostridium difficile--Associated and Non--C. difficile--Associated Diarrhea after Use of Gatifloxacin and Levofloxacin in an Acute-Care Facility

Sir—We read with great interest the recent report by Gaynes et al. [1]. We would like to share a similar but somewhat different observation from our hospital. Levofloxacin was the quinolone of choice in our hospital formulary until July 2001, when gatifloxacin was introduced. We have observed an increase in the incidence of diarrhea and in the number of requests for Clostridium difficile toxin assays after gatifloxacin became the quinolone of choice. A study was initiated with the purpose of determining whether gatifloxacin was more likely to promote C. difficile—associated diarrhea (CDAD) than was levofloxacin. Here, we briefly summarize our findings. The study was conducted in Coney Island Hospital, an acute-care community hospital of ∼350 beds located in South Brooklyn, New York. This was a retrospective study that involved chart review of the patients admitted to our inpatient department from July 1999 through June 2002. From a computer record of our inpatient pharmacy, the medical record numbers of 400 patients who received levofloxacin and of another 400 patients who received gatifloxacin were randomly selected. The computer records for each of these patients were reviewed for age, sex, date of hospital admission, length of hospital stay, date on which antibiotic (levofloxacin or gatifloxacin) therapy was started, date on which a C. difficile toxin assay was requested, result of the toxin assay, and length of therapy. The date of the C. difficile toxin assay request was defined as the date of development of diarrhea. Cases of diarrhea that developed within 48 h after admission were excluded from analysis because we did not know what the predisposing factors were. Cases of diarrhea that developed after 24 h and within 6 weeks after initiation of therapy with 1 of these 2 antibiotics were included in analysis. Inconclusive results of the C. difficile toxin assay were considered to be negative. For patients who developed CDAD, charts were reviewed for use of other antibiotics, both for concurrent use and use within the previous 6-week period. Findings were compared for statistical significance using the χ2 test. Table 1 contains data on the mean age, the ratio of male to female subjects, the average duration of therapy, and the length of hospital stay. There was no significant difference between the levofloxacin group and the gatifloxacin group with regard to these parameters. A total of 47 patients (11.75%) in the levofloxacin group and 74 patients (18.5%) in the gatifloxacin group developed diarrhea for which a toxin assay was requested; the difference was significant (P = .01). There were more patients in the gatifloxacin group than in the levofloxacin group (8 vs. 4) who had a positive result of the toxin assay, but this difference was not statistically significant (P > .05). A total of 4 patients in the gatifloxacin group had an inconclusive toxin assay result. Among the patients with proven CDAD, there were no differences in the duration of therapy, length of hospital stay, and use of other antibiotics in either group (data not shown), although the mean age of patients with CDAD in the gatifloxacin group was greater (74 vs. 56 years). Characteristics of the patients who received a course of quinolone antibiotics and who developed diarrhea. Beginning with a case report in 1989 [2], there have been a number of publications associating quinolone use with CDAD [3–5]. The study by Gaynes et al. [1] and our study are the only reports, to our knowledge, that compared 2 quinolones with regard to their association with CDAD. Both of these studies found a higher incidence of diarrhea after use of gatifloxacin than after use of levofloxacin. There are 2 main differences in the findings of these 2 studies. In our study, although a large number of patients developed diarrhea after use of gatifloxacin (18.5%) and levofloxacin (11.75%), the actual incidence of proven cases of CDAD was low (2% and 1% for gatifloxacin and levofloxacin, respectively). The difference was not statistically significant, probably because of the relatively small number of positive cases. A possible reason for the low incidence of CDAD could be the nature of the population. The study by Gaynes et al. [1] was done at a long-term care facility, where the length of stay is much longer; our patients were treated in an acute-care hospital, where the length of hospital stay is shorter (table 1). Because of the retrospective nature of our study, we do not have data on any patient who may have developed diarrhea after discharge from the hospital. An additional factor may be the more strictly followed infection-control measures in an acute-care facility, compared with a long-term care facility. The second finding of our study is that a significantly higher percentage of patients developed non-CDAD in the gatifloxacin arm than in the levofloxacin arm. The cases of diarrhea that developed in gatifloxacin recipients appeared to be self-limiting and of short duration. Thus, a significantly lower number of C. difficile toxin assays were actually performed in the gatifloxacin recipients who developed diarrhea (52 [70%] of 74) than in the levofloxacin group (42 [89%] of 47). Gatifloxacin has enhanced activity against many anaerobic bacteria [6, 7]. Whether this property of gatifloxacin is associated with development of diarrhea by a mechanism other than the overgrowth of toxin-producing C. difficile in the colon remains to be explained. We conclude that the incidence of diarrhea is significantly higher after use of gatifloxacin than after use of levofloxacin. However, most of these cases are not CDAD. Conflict of interest. All authors: No conflict.