Dynamics of antibiotic usage in the intensive care unit at the University Hospital of the West Indies.

Objective: To determine antibiotic usage patterns in the Intensive Care Unit (ICU) at the University Hospital of the West Indies (UHWI).
Method: A cross-sectional, analytical study of consecutive patients admitted to the ICU was conducted between July and December 2007. Exclusion criteria were HIV-positive patients, patients < 12 years and those discharged or who died within 48 hours of admission. Data were collected from medical records, stored and analysed using the SPSS Version 12.
Results: Of the 150 eligible patients, 109 had complete data (73%). Mean age was 50.8 +/- 20.7 years, with mean APACHE II score of 15.6 +/- 6.7. Forty-five patients (41.3%) received prophylactic antibiotics, most commonly ceftriaxone (31.7%) and metronidazole (19.0%). Appropriate discontinuation within 24 hours occurred in only 11.1%. Two-thirds of patients (67.9%) were treated with empiric antibiotics, most commonly piperacillin/tazobactam (32.1%), ceftazidime (27.5%) or metronidazole (27.5%). Reasons for empiric choice were primarily coverage of organisms based on presumed source of sepsis (45.6%), and broad spectrum, high-powered coverage (23.5%). Courses ranged from 1 - 42 days and were adequate based on subsequent cultures in 71% of cases. Culture reports took between 2 - 8 days with a mean of 3.7 days to become available. De-escalation was practised in only 2 of 26 (7.7%) cases and intravenous to oral switch therapy in only 3.3%. Thirty-two (29.4%) patients died, with sepsis being a cause in 12 (37.5%).
Conclusions: Improved attention to discontinuation of prophylactic antibiotics, appropriate duration of antibiotic courses and de-escalation are essential if the antibiotic practices in the ICU at the UHWI are to compare favourably with international recommendations.