Background: Gastroenteritis accounts for >450,000 US hospitalizations annually; etiology is found in <50%. Multiplex PCR could provide rapid, simultaneous identification (ID) of multiple pathogens not suspected clinically or not detectable by standard tests (ST). Enhanced ID could impact Infection Prevention (IP) practices. There are few studies of PCR testing in children with diarrhea. Objectives were 1) to compare BioFire FilmArray¨ GI Panel (PCR) with ST for pathogen ID; 2) compare diagnostic yield of physician-selected ST vs PCR; 3) assess impact of PCR on IP.
Methods : Liquid stool specimens submitted to the clinical lab for ST from Mar '14-Mar '15 were studied. ST were performed per specific physician orders. PCR was performed for one specimen per patient per encounter; results were not available in real time. Available ST included stool culture, parasite microscopy, and enzyme immunoassays (EIA) for rotavirus, adenovirus, C. difficile (w/reflexive NAAT), Shiga-like toxin-producing E. coli (STEC), Cryptosporidium and Giardia. GI panel tests for 22 pathogens (13 bacteria, 5 viruses, 4 protozoa). Medical charts were reviewed retrospectively to assess clinical data and isolation measures.
Results : 51 patients had 53 unique stool specimens analyzed. Most were hospitalized (91%) with median age 4 y (range 10 d-17 y); 55% had chronic medical conditions. Diarrhea predominantly had community onset (CO-85%).
Median of 3 (range 1-7) ST ordered/stool specimen. Pathogens were identified in 33/53 (62%) specimens by some method. PCR identified pathogens in 62% of specimens vs 30% for ST (p <0.005) (Fig 1). PCR identified 23 additional pathogens among 17 ST-negative specimens.
Physician-selected ST missed 19 pathogens even when ST available: ST not ordered in 68% and ST false negative in 32%. (Fig 2)
Of 40 hospitalized patients with CO-diarrhea, only 85% were placed in enteric isolation. ST missed 15 pathogens that should have led to high-level enteric isolation (soap and water hand hygiene and bleach disinfection of patient room) per hospital policy. (Fig 3)
Conclusion: PCR enhanced pathogen ID >2 fold. Low yield of ST is due both to insensitivity and to inadequate physician selection of tests. Use of PCR could optimize isolation practices.
J. M. Gould, None
M. Heard, None
S. S. Long, None
A. T. Evangelista, None