2033. Incorporating T2Candida Testing into Rational Antifungal (AF) Management: A Successful Pilot Study of Diagnostic Stewardship (DS) Directed Toward Specific Intensive Care Unit (ICU) Patients At-Risk for Sepsis due to Invasive Candidiasis (IC)

BACKGROUND: Blood cultures (BC) are ~50% sensitive for diagnosing IC. T2Candida (T2) detects five leading Candida spp. directly in blood and was ≥90%/90% sensitive/specific (S/Sp) for candidemia in clinical trials. Optimal use of T2 in clinical practice is unclear. We targeted T2 to specific ICU patients at-risk for IC, and implemented AF management algorithms developed with ICU teams. METHODS: A DS team ordered concurrent T2 and BC, and used results to guide AF in patients fulfilling pre-specified criteria for septic shock (medical ICU (MICU)), sepsis after abdominal surgery (trauma ICU), or sepsis with mechanical circulatory support (cardiothoracic ICU). We focused on groups with anticipated pre-test IC probabilities of ~3–15%. Proven IC was defined if BC+ and possible IC if BC- but a compatible clinical picture was observed. RESULTS: Seven percent (6/88) of BC in ICU patients with sepsis were Candida +. T2 and BC results are shown in the table. Using BC as gold standard, T2 S/Sp and PPV/NPV were 50%/87% and 33%/96%, respectively. Including possible IC, T2 S/Sp increased to 69%/96%, and 67% (4/6) of T2+/BC− results were likely true positive; two false-positive results were for C. parapsilosis. We focused on MICU outcomes initially since 75% (66/88) of tests were performed here. Empiric AFs were discontinued in 12 patients following a T2- result; AFs were avoided in all others. Median combined days of therapy (DOT)/month for caspofungin and fluconazole as empiric or definitive treatment prior to and after introducing DS were 26 (range: 10–53) and 15 (3–32), respectively (P = 0.0047). AF consumption was decreased 47% (figure). CONCLUSION: Targeted DS using T2 in select ICU patients with sepsis significantly reduced AF usage. 14% of patients with sepsis were diagnosed with IC using either T2+ or BC+, compared with 7% with BC+ alone, as would be expected if BC S was 50%. T2 S and T2−/BC+ results were lower and higher, respectively, than previously reported, indicating that treatment decisions should be based on results of both tests. Most T2+/BC− results were ascribed to possible IC. [Image: see text] DISCLOSURES: M. H. Nguyen, Merck: Grant Investigator, Research grant. Astellas: Grant Investigator, Research grant.