“Answers In Hours” - Prospective Clinical Trial Using Microbiome Metagenomics For Bile Duct Cultures
*Jennifer A Yonkus, *Emma L Whittle, *Roberto Alva-Ruiz, *Susan E Horsman, *Amro M Abdelrahman, *Grotz E Travis, *Cleary P Sean, *Rory L Smoot, David M Nagorney, *Michael L Kendrick, *Nicholas Chia, *Mark J Truty
Mayo Clinic, Rochester, MN
Background: Postoperative infection is a major source of morbidity in patients undergoing pancreatic head resection, often from organisms identified from intra-operative bile duct cultures (BDC). Data suggests tailored perioperative antimicrobial therapy based on BDC decreases infection rates. However, standard cultures (SC) take days to result. Novel metagenomic techniques using nanopore sequencing (NS) can provide robust data regarding microbial profiles in hours. The present study was designed to determine the feasibility and utility of NS for microbial profiling of intra-operative bile samples.
Methods: Patients undergoing pancreatic head resection were included. At surgery, bile duct fluid was collected and sent for both SC and NS for microbial profiling. De-identified results of SC and NS were provided to a Surgical Clinical Pharmacist for antibiotic recommendations. Time-to-results was defined as time from sample acquisition to either final SC or final NS results. Organism yields, resistance patterns, antibiotic recommendations, and costs were compared between SC and NS.
Results: Bile duct fluid was collected from 41 prospective trial cases. 22/41 (54%) samples resulted in positive SC. All positive SC had microbes detected using NS. On average, NS detected twice as many bacterial species compared to SC 10.7 vs. 4.4, p<0.05). Resistance genes were screened using NS in the 22 positive samples with mean of 6.4 genes being detected(range 0-32). Resistance genes conferred an average of 13.5 predicted resistance phenotypes which was significantly higher than average number of resistance phenotypes detected using SC(13.5vs2.7, p<0.05). Antimicrobial recommendations for NS provided coverage for SC results in 13/22 (59%) samples. Additionally, broader antimicrobial coverage (resistance patterns not identified in SC) was recommended in 3/22(14%) samples based on NS. Median time to NS results was faster(8 vs. 165 hours) than SC however did result in higher costs($118.94vs.$38.49).
Discussion: Early results from this prospective study demonstrate the feasibility of NS for rapid microbial profiling with higher organism yields and broader resistance phenotypes identified compared to SC. Given our preliminary findings, a randomized control trial is under development. As microbiome metagenomics is applicable to all body fluid samples, this technology may have major impact on diagnosis and treatment of all surgical site infections.
Back to 2021 Abstracts