Objectives: Over the past decade, the epidemiology of Clostridium difficile infection (CDI) has shown a remarkable increase in incidence with an associated increase in severity. This study was designed to compare the demographics, medication exposure, evaluation, treatment patterns, and outcomes of patients with CDI in two different time periods: 2006-2008 and 2009-2011. We hypothesized that mortality is decreasing with increasing appropriateness of medical management.
Methods: We retrospectively identified consecutive patients admitted to Montefiore Medical Center between 1/1/2006 and 12/31/2011 with symptomatic diarrhea and a positive C. difficile toxin assay. The cohort was subdivided into those diagnosed in 2006-2008 (CDI 06-08) and 2009-2011 (CDI 09-11). We obtained key parameters at the time of diagnosis including demographics, medication exposure, medical comorbidities, laboratory data, CDI evaluation, and various outcome measures. We created a subcohort for each time frame of patients diagnosed with severe CDI defined by white blood cell count (WBC) >15,000 cells/μl and albumin <3.0 g/dl and made the same comparisons as for the overall cohort. The two cohorts were compared using SPSS (16.0).
Results: Cohorts and the number of patients who met criteria for inclusion were as follows: CDI 06-08 (n=1189), CDI 09-11 (n=1,907), severe CDI 06-08 (n=243), and severe CDI 09-11 (n=382). CDI 09-11 patients were older (P=0.01) and had higher Charlson comorbidity scores (P=0.02) than did those in the CDI 06-08 cohort. There were no significant demographic differences in the severe cohort. For both the overall and severe cohorts, there was more macrolide exposure before diagnosis with CDI and lower rates of quinolone exposure in the more recent era. The disease process also appeared less severe in the CDI 09-11 cohort with lower peak WBC during admission and at diagnosis. Treatment patterns appeared more aggressive during the more recent time frame, with shorter durations of oral metronidazole (P<0.001), longer durations of IV metronidazole (P=0.04), more frequent use of vancomycin as the sole therapy (P<0.001), more frequent switching from metronidazole to vancomycin (P<0.001), and less frequent exposure to any metronidazole throughout treatment (P<0.001) in the overall cohort. The 30-day mortality decreased significantly in both the overall (17.1 vs. 13.1%, P<0.01) and the severe (31.3 vs. 23.3%, P<0.05) cohorts from CDI 06-08 to CDI 09-11, with mortality decreasing significantly in the 8th and 9th decades of life in the overall cohort and in the 8th, 9th, and 10th decades in the severe cohort.
Conclusions: In an urban United States population, CDI 09-11 showed changes in medication exposures, less severe disease, and more aggressive management with better outcomes and decreased mortality compared with CDI 06-08. The most important factors associated with 30-day mortality in both an overall and severe CDI population include age, WBC, and albumin level at the time of diagnosis.