Antibiotic Use, Resistance and Clostridium Difficile-Associated Diarrhea
Stuart B. Johnson BA MD
Edward Hines, Jr. VA Hospital
Hines, IL
Funding Period: September 2005 - August 2007

BACKGROUND/RATIONALE:
Clostridium difficile-associated diarrhea (CDAD), is strongly associated with health care facilities and previous antimicrobial treatment, as well as with increased hospital costs and increased length of stay. The estimated annual cost of CDAD in the US is likely to exceed $1.1 billion. Severe pseudomembranous colitis, toxic megacolon and fulminant CDAD are life-threatening complications of C. difficile.
Antibiotics (AB) are not equal in terms of CDAD risk. CDAD rates significantly decrease when clindamycin or third-generation cephalosporin use is reduced. Increased CDAD risk has been associated with fluoroquinolone (FQ) use. Large quantities of FQs used at hospitals contribute to high CDAD attributable risk. The risk of AB-associated CDAD increases when C. difficile is resistant to the AB. High-level FQ resistance isolates have been identified. The recent VISN 12 formulary change from levofloxacin to gatifloxacin to moxifloxacin afforded the opportunity to measure the impact of AB use on the emergence of nosocomial infections C. difficile. The VA is particularly well suited to study these issues because of its electronic medical records and extensive databases.

OBJECTIVE(S):
This study aimed to: (1) Determine the impact of AB formulary changes on CDAD rates, (2) Link CDAD-incidence to AB therapy using local and national VA databases to identify relative risk and attributable risk of specific ABs with the development of CDAD, (3) Show CDAD rates differ between levofloxacin period and gatifloxacin period and moxifloxacin period. HYPOTHESIS: Changes in medication formulary (levofloxacin to gatifloxacin and moxifloxacin) will be associated with increased CDAD rates at affected VA hospitals.

METHODS:
This is an observational study of the effects of an AB formulary change on rates of an unanticipated complication, CDAD. It is a retrospective review of inpatients receiving levofloxacin from Oct 01-March 04 to those who receiving gatifloxacin March 04 to Feb 06, and Moxifloxacin from March 06-Sept 06. We conducted a matched control cohort study, and performed tests for trends and multivariate logistic regression modeling, as well as calculation of antibiotic specific attack rates, to detect meaningful changes in rates in CDAD over the study period. The study was conducted at Hines VA, Jesse Brown VA and North Chicago VA.
Data sources included the VA National Patient Care Databases (NPCD), VA BIRLS Death file, VA Decision Support System (DSS) National Extracts. Microbiology data was obtained from each facility’s local medical record database - Veterans Health Information Systems Technology Architecture (VISTA).

FINDINGS/RESULTS:
CDAD infections increased markedly after the levofloxacin to gatifloxacin formulary change, and continued high during the moxifloxacin period, suggesting greater potentials of the two broader-spectrum FQs to induce CDAD. This was borne out both in analysis focused on pre-post trends as well as case-control. Thus, CDAD may be one unanticipated consequence of AB formulary changes. Presence of the new epidemic strain may have partially contributed to the increased CDAD rates.

IMPACT:
Study findings have: (1) increased knowledge about FQ potential to induce CDAD, (2) Support the need for including adverse reactions into formulary calculations, (3) Developed novel methods for merging VA databases to obtain meaningful patient-level data without the need for individual patient chart reviews.

PUBLICATIONS:
None at this time.

DRA: Health Services and Systems
DRE: Quality of Care, Epidemiology
Keywords: Patient outcomes, Pharmaceuticals, Quality assessment
MeSH Terms: none