Sustained multimodal antimicrobial stewardship in an Australian tertiary intensive care unit from 2008-2015: an interrupted time-series analysis

The long-term outcomes and sustainability of antimicrobial stewardship (AMS) in the intensive care unit (ICU) require evaluation. This study analysed the effect of a multimodal ICU AMS introduced in a 15-bed medical-surgical tertiary Australian adult ICU in November 2008, using interrupted time-series analysis of antibiotic usage, Gram-negative resistance and cost from November 2005 to October 2015, including national ICU average usage as a control. Overall ICU mortality, 30-day blood stream infection (BSI) mortality and length of stay (LOS) were compared over the same period. There were 2512 patients admitted to ICU before and 6435 after AMS intervention. Post-AMS there was a reduction in the trend of aminoglycoside usage both absolute from 63.3 DDD/1000 occupied bed days (OBD)/month (-1.1; 95% confidence interval [CI] -2.2, -0.1; P = 0.033) and relative to the national trend (-1.3; 95%CI -2.4, -0.3; P = 0.016). Vancomycin usage increased both absolute from 161.2 DDD/1000 OBD/month (1.8; 95%CI 0.03, 3.6; P = 0.046) and relative to the national trend (1.8; 95%CI -0.3, 3.9; P = 0.092). There were sustained post-AMS downward trends in carbapenem, antipseudomonal penicillin, third-generation cephalosporin and fluoroquinolone use that did not reach statistical significance. Post-AMS, antipseudomonal penicillin resistance declined (-12.8%; 95%CI −24.9, -0.6; P = 0.040). Antimicrobial acquisition costs declined by AUD$0.5/OBD/month (95%CI -1.1, 0.1; P = 0.096). Over the study period, severity-adjusted ICU mortality declined from 12.9% to 10.4%; risk ratio (RR) 0.92 (95%CI 0.82, 1.03) and BSI 30-day mortality from 37.9% to 26.3%; RR, 0.76 (95%CI 0.56, 1.03). Median ICU LOS for ICU survivors increased from 2.3 to 2.6 days. Multimodal AMS sustainably embedded in ICU was associated with reductions in broad-spectrum Gram-negative antibiotic use, overall antibiotic costs and Gram-negative resistance, without adverse clinical impact.