Over the past decade, trends in antimicrobial resistance, epidemiology, and drug development have occurred that affect both the empiric and definite selection of antimicrobials in the septic patient. The rapid spread of highly pathogenic community-associated methicillin resistant Staph aureus (MRSA) requires clinicians to consider the inclusion of empiric coverage for MRSA even in community- acquired sepsis. Moreover, vancomycin appears to be losing its effectiveness, and while a number of new agents with broad gram positive activity have been licensed, none have emerged as clearly superior. An alarming increase in the number of hospital-acquired infections due to multi-drug resistant gram negative bacteria has also occurred, and few new gram negative drugs are in development. Clinicians, faced with Pseudomonas aeruginosa or Acinetobacter baumanii isolates resistant to all commonly used drugs, must resort to toxic older drugs such as colistin or therapy combining drugs not effective as monotherapy. Based on a desire to limit overall antimicrobial use, a re-evaluation of older data in both the neutropenic and non-neutropenic host has called into question the common practice of using combination therapy for some gram negative infections. An emerging consensus advocates emphasizing local unit specific antimicrobial sensitivity data in selecting empiric therapy and determining if combination therapy is required. New antifungal drugs and a better understanding of the risk factors for infection with Candida spp. has altered the approach to empiric and definitive treatment of Candida infections in the septic patient.