This study was performed to elucidate the relationship between antimicrobial use density (AUD) and Clostridium difficile infection (CDI) manifesting as antimicrobial-associated diarrhea (AAD) in hospital wards during a 4-year period. Case definition of CDI was an adult exhibiting AAD with a daily stool frequency of three or more, arising at least 48 hours after ward admission, and fecal samples testing positive for toxin (A and/or B). Metronidazole or vancomycin was orally administered as treatment. AUDs were calculated for a total of 21 antimicrobials in a span of 48 months and nine wards. We included the average value of AUDs, representing two succeeding months of sample submission into the sample information. We also entered data on the 2-year division and intensified contact precaution for statistical analysis. Of a total of 463 cases, 95 (20.5%) were CDI-positive. Multivariate regression analysis showed odds ratios [OR] of 1.739 (95% confidence interval [CI] of 1.050 - 2.881, P = 0.032) and 1.598 (95% CI of 1.006 -2.539, P = 0.047) for clindamycin and piperacillin, respectively in AUD. Thus increased ward AUDs of clindamycin and piperacillin may run the risk of CDI.