Direct-from-blood RNA sequencing identifies a pathogen that would not have been identified by other common techniques

Antibiotic resistance is rising at disturbing rates and contributes to the deaths of millions of people yearly. Antibiotic resistant infections disproportionately affect those with immunocompromising conditions, chronic colonization, and frequent antibiotic use such as transplant patients or those with cystic fibrosis. However, clinicians lack the diagnostic tools to confidently diagnose and treat infections, leading to widespread use of empiric broad spectrum antimicrobials, often for prolonged duration.

Case Presentation – A 22 year-old Caucasian female with cystic fibrosis received a bilateral orthotopic lung transplantation 5 months prior to the index hospitalization. She underwent routine surveillance bronchoscopy and was admitted for post-procedure fever. A clear cause of infection was not identified by routine methods. Imaging and bronchoscopic lung biopsy did not identify an infectious agent or rejection. She was treated with a prolonged course of antimicrobials targeting known colonizing organisms from prior bronchoalveolar lavage cultures (Pseudomonas, Staphylococcus aureus, and Aspergillus). However, we identified Stenotrophomonas maltophilia in two independent whole blood samples using direct-pathogen sequencing, which was not identified by other methods.

High-throughput sequencing based microbial detection in replicate samples from blood

a Quality controlled reads from two replicates were profiled by four different taxonomic classifiers: BWA-mem against RefSeq; GOTTCHA against a bacterial database (GOTTCHA.b); GOTTCHA against a viral database (GOTTCHA.v); Kraken-mini; and MetaPhlAn. Normalized read counts mapping to the top 20 organisms by one or more tools are presented at the species level (heatmap). b The SIQ score was calculated for each replicate using reads that mapped to RefSeq using BWA-mem. Coordinates on SIQ plots are relative abundance of all species (x-axis), by significance (1 – p value, y-axis), and the size of the dot represents the SIQ score. The dashed red line indicates a p value of 0.05. Microbes detected at significant levels with a normalized abundance level > 4 are labeled: S, Stenotrophomonas maltophilia; P, Propionibacterium acnes; E, Escherichia coli. Plots were generated using R statistical programming

This case represents a common clinical conundrum: identification of infection in a high-risk, complex patient. Here, direct-pathogen sequencing identified a pathogen that would not otherwise have been identified by common techniques. Had results been clinically available, treatment could have been customized, avoiding a prolonged course of broad spectrum antimicrobials that would only exacerbate resistance. Direct-pathogen sequencing is poised to fill a diagnostic gap for pathogen identification, allowing early identification and customization of treatment in a culture-independent, pathogen-agnostic manner.