Injudicious Antimicrobial Use in Patients with Clostridium-difficile Infections at Hospital Admission

Background: The patterns of empiric antimicrobial use in patients presenting with Clostridium difficile infections (CDI) at hospital admission have not been adequately investigated.  Describing risk factors associated with injudicious antimicrobial use in CDI patients can have significant implications for antimicrobial stewardship.

Objective: to identify the prevalence and risk factors associated with injudicious antimicrobial use among patients with CDI at hospital admission. 

Methods: Study design: case-control. Study population: all patients in a university-affiliated community hospital with a positive C. difficile toxin within 48 hours of admission. Study period: 1/1/08 – 12/31/09. Definitions: Injudicious antimicrobial use (cases): receipt of antimicrobials other than metronidazole and/or oral vancomycin at hospital admission, in the absence of all of the following: 1) positive microbiology cultures; 2) pyuria; 3) pathological or radiological evidence of infection other than CDI; 4) physical signs of skin or soft tissue infection.    Controls were patients with no other antimicrobial use, or with at least one indication for other antimicrobial use, as outlined above. Severe CDI was defined as a score of ≥ 2, calculated as follows: age > 60, temperature > 38.3°C, albumin < 2.5 mg/dL, WBC count > 15,000 cells/mm3 - 1 point each; ICU hospitalization, pseudomembranous colitis - 2 points each.  Demographic and other clinical variables of interest were collected through retrospective medical records’ review. Statistics: Univariate and multivariate analysis.

Results:  Among the 88 patients with CDI at admission, 50 (57%) received other antimicrobials, in addition to CDI therapy, within the first 48 hours of hospitalization. Fluoroquinolones (21%) and piperacillin/tazobactam (21%) were most commonly used.  Injudicious antimicrobials were prescribed to a total of 30 (34%) of all CDI patients. Risk factors for injudicious use on univariate analysis were: age ≥ 80 (P = 0.02); NH residence (P < 0.01); nonverbal patient (P < 0.001); underlying dementia (P = 0.03); suspected UTI (P < 0.01); acute renal failure (P < 0.01); CDI-related ICU admission (P < 0.01); and CDI severity ≥ 2 (P = 0.02) Cases were more likely to stay in the hospital ≥ 12 days (P = 0.04), but in-hospital mortality was not significantly different between cases and controls. On multivariable analysis, suspected UTI (OR 7.0 [2.0 – 23.7], P = 0.002), acute renal failure (OR 4.5 [1.5 – 13.7], P = 0.009), and CDI severity ≥ 2 (OR 3.2 [1.0 – 10.3], P = 0.046).

Conclusions: Injudicious antimicrobial use occurred in 1/3^rd of the patients admitted with CDI, and was more likely in those with severe CDI. Clinicians should increase their diagnostic index of suspicion for CDI in elderly, frail patients who are unable to effectively verbalize symptoms. UTI may be over suspected among these, leading to unnecessary antimicrobial use.