Application of Multiplex Polymerase Chain Reaction for Pathogen Identification and Antibiotic Use in Children With Respiratory Infections in a PICU

OBJECTIVES: To compare the pathogen identification rate and use of antibiotics before and after the implementation of multiplex polymerase chain reaction testing in children with respiratory infections in a PICU. DESIGN: Single-center, pre-post study. SETTING: PICU of Osaka Women's and Children's Hospital, Osaka, Japan. PATIENTS: Consecutive children with respiratory infections who were admitted to the PICU between December 2017 and November 2018 (premultiplex polymerase chain reaction period) and between March 2019 and February 2020 (postmultiplex polymerase chain reaction period). INTERVENTIONS: Conventional rapid antigen tests and bacterial culture tests were performed throughout the study period. Multiplex polymerase chain reaction testing using the FilmArray respiratory panel (BioFire Diagnostics, Salt Lake City, UT) was conducted to detect 17 viruses and three bacterial pathogens. During the postmultiplex polymerase chain reaction period, we did not recommend prescribing antibiotics for stable children, depending on the virus species and laboratory test results. MEASUREMENTS AND MAIN RESULTS: Ninety-six and 85 children were enrolled during the pre- and postmultiplex polymerase chain reaction periods, respectively. Rapid antigen tests identified pathogens in 22% of the children (n = 21) during the premultiplex polymerase chain reaction period, whereas rapid antigen tests and/or multiplex polymerase chain reaction testing identified pathogens in 67% of the children (n = 57) during the postmultiplex polymerase chain reaction period (p < 0.001). The most commonly identified pathogen using multiplex polymerase chain reaction testing was human rhino/enterovirus. Bacterial pathogens were identified in 50% of the children (n = 48) and 60% of the children (n = 51) during the pre- and postmultiplex polymerase chain reaction periods (p = 0.18). There were no differences in antibiotic use (84% vs 75%; p = 0.14), broad-spectrum antibiotic use (33% vs 34%; p = 0.91), or the duration of antibiotic use within 14 days of admission (6.0 vs 7.0 d; p = 0.45) between the pre- and postmultiplex polymerase chain reaction periods. CONCLUSIONS: Although the pathogen identification rate, especially for viral pathogens, increased using multiplex polymerase chain reaction testing, antibiotic use did not reduce in children with respiratory infections in the PICU. Definitive identification of bacterial pathogens and implementation of evidence-based antimicrobial stewardship programs employing multiplex polymerase chain reaction testing are warranted.