A Comparison of 3 Severity Criteria for Clostridium difficile Infection

Background: Recent studies on Clostridium difficile infection (CDI) have shown that patients with severe infection have better outcomes with vancomycin (V) than with metronidazole (M); in patients with mild disease, V and M appear to be equivalent. There are different severity criteria in the literature which have not been compared or systematically validated.

Objective: To use our patient population to assess 3 different versions of CDI severity criteria: hospital-specific guidelines at NewYork-Presbyterian Hospital (NYP); IDSA guidelines; and those proposed by Zar et al., regarding differences in M use, V use, clinical outcomes and costs.

Methods: Retrospective review of all hospitalized adult patients with laboratory-confirmed CDI at NYP from April 2009 – April 2010. The 3 severity criteria were applied to the patient population and the number of patients stratified into M and V categories for each set of criteria was assessed. For each set of criteria, drug costs were compared as if all patients stratified to the M group had received M and all patients stratified to V had received V. Finally, for patients who were treated with the appropriate drug, we assessed whether the criteria leading to the most M use led to better outcomes than the criteria that led to the most V use. Poor outcome was defined as death within 7 days of diagnosis or colectomy.

Results: Records of 366 patients with a first episode of CDI were reviewed. Median age was 67 years (range 18-99). About 24% were in an ICU at the time of diagnosis. Poor outcome, mostly death, occurred in 6.6% (n=24) patients. When stratified based on IDSA criteria, median age (years) was 63 for mild-moderate group and 70.9 for severe group (p<0.001). Median white blood cell count was 8,400 cells/mm³ for mild group and 18,300 cells/mm³ for severe group (p<0.001). The Zar criteria categorized the most patients into the severe group (185; 51%) and would lead to the most V use, followed by the IDSA criteria (163; 45%). NYP criteria categorized significantly fewer patients as severe (66; 18%) (p<0.001). When patients received appropriate therapy based on the different criteria, frequency of poor outcomes was not significantly different between the criteria (6.2-7.3%). However, poor outcomes in the M group was highest for NYP criteria (3%) but not significantly different compared to the other criteria (1.2-1.3%; p=ns). Estimated drug costs were significantly higher for IDSA and Zar compared to NYP criteria, and this remained significant with ($87,000 vs $211,000-239,000) or without ($3975 vs $6200-6750) the use of V in capsule form (p<0.001).

Conclusions: NYP guidelines recommend M for more CDI cases than do IDSA or Zar criteria. This leads to a significant drug savings and no difference in poor outcomes, although overall number of poor outcomes in this population was low. The different severity criteria for CDI should be tested and validated prospectively.