A single-cell analysis platform identified bloodstream pathogens and measured their antimicrobial susceptibility within hours in laboratory and clinical-sample testing, according to a study published in Science Advances.
The system combined bacterial enrichment, fluorescence-based identification, and phenotypic antimicrobial susceptibility testing (AST). Complete results were produced in 6.75 to 17 hours when the workflow was tested using whole blood spiked with clinical bacterial isolates.
The first stage, called sedimentation-assisted tandem rocking and enrichment for analysis and monitoring (STREAM), mixed 10 mL of whole blood with culture broth and dextran. Gentle rocking caused red blood cells to aggregate and settle while bacteria remained in an upper, plasma-like layer and multiplied.
Researchers then identified individual bacteria using molecularly barcoded sequential fluorescence in situ hybridization. Fluorescent probes targeting bacterial 16S ribosomal RNA were applied in up to three rounds, allowing organisms to be classified by Gram status and identified at the family, genus, or species level.
The identification panel was assessed using 21 bacterial pathogens associated with bloodstream infection. In 104 positive blood culture samples, results were concordant with clinical laboratory identification in 96 percent of cases and were available within 1.2 to 2.5 hours. Four organisms were identified only at a broader taxonomic level, but none were misidentified.
For susceptibility testing, individual bacteria were immobilized in an agarose gel and exposed to different antimicrobial concentrations. Time-lapse imaging measured changes in bacterial number, morphology, and membrane integrity over 1 hour.
Across 219 drug-dose combinations, the single-cell assay achieved 98 percent essential agreement and 94 percent categorical agreement with clinical laboratory results. The very major error rate, representing resistant isolates incorrectly classified as susceptible, was 4 percent. The minor error rate was 2 percent, and no major errors were reported.
The full workflow was evaluated directly from human blood spiked with Klebsiella pneumoniae or Staphylococcus aureus at concentrations of 0.1 to 10 colony-forming units per milliliter. At 1 colony-forming unit per milliliter, identification and AST took 6.75 hours for K pneumoniae and 14 hours for S aureus. At 0.1 colony-forming unit per milliliter, testing took 8.75 and 17 hours, respectively.
These direct-from-blood experiments used spiked samples rather than blood collected from patients with bloodstream infections. Clinical testing was limited to previously positive blood cultures and their isolates, meaning the complete workflow has not yet been validated prospectively in patients with suspected sepsis.
Other limitations included the potential effects of prior antimicrobial exposure on bacterial recovery and growth. Faster-growing organisms could obscure additional pathogens in polymicrobial infections, while intracellular, nonculturable, anaerobic, or fastidious organisms may require different processing conditions. The multistep workflow also requires specialized imaging and further automation.
Newsletters
Receive the latest pathologist news, personalities, education, and career development – weekly to your inbox.
