Effect of host-protein test (TRAIL/IP-10/CRP) on antibiotic prescription and emergency department or urgent care center return visits: The JUNO pilot randomized controlled trial.

Abstract

Determining etiology for adults with symptoms of lower respiratory tract infection (LRTI) is challenging. MeMed BV (MMBV), an FDA-cleared blood test, computationally integrates the levels of three host proteins to differentiate bacterial and viral infections. We evaluated MMBV's impact on safe antibiotic prescribing at emergency department/urgent care centers (ED/UC). The JUNO randomized controlled trial (RCT; NCT05762302) was a prespecified pilot phase of the JUPITER RCT. JUNO enrolled adult ED/UC patients with LRTI symptoms and clinician's consideration for antibiotic treatment. Inclusion criteria were fever within 7 days and one of cough, sputum production, dyspnea, or auscultation abnormality. Exclusion criteria were prior antibiotic use or immunosuppression. Patients were randomized to standard care (SC) or SC plus MMBV arms. JUNO's primary objective was to assess antibiotic prescription rate in the SC arm; the secondary objective was to assess JUPITER's study design. Eleven centers randomized 260 patients, with 214 included (106 SC, 108 MMBV). Median (IQR) age was 40 (28-55.8) years, 57% were female, and 78.5% were enrolled at ED. Common symptoms were cough (93.0%) and chills (70.0%). Overall, antibiotic prescription rates were 30% (95% CI 22% to 40%) and 24% (95% CI 17% to 33%) in the SC arm versus the MMBV (absolute difference of -6% [95% CI -18% to 6%]). More antibiotics were given with bacterial MMBV scores (41% vs. 78%, absolute difference 37%, 95% CI 6% to 61%; n = 40) and less with viral MMBV scores (25% vs. 12%, absolute difference -13%, 95% CI -25% to 0%; n = 144) in the SC versus MMBV arms. There was no increase in ED/UC return visits (8% vs. 3%, difference -6%, 95% CI -12% to 1%) or hospitalizations (3% vs. 0%, difference -3%, 95% CI -7% to 1%) in the SC arm versus the MMBV arm. JUNO demonstrated that JUPITER's design results in 30% antibiotic prescription rate in the SC arm. JUNO supports that MMBV optimizes antibiotic prescriptions without increasing return ED/UC visits or hospitalizations.