High resolution and rapid separation of bacteria from blood using elasto‐inertial microfluidics

• Published in: Electrophoresis 2021-12, Vol.42 (23), p.2538-2551
• Main Authors: Narayana Iyengar, Sharath, Kumar, Tharagan, Mårtensson, Gustaf, Russom, Aman
• Format: Article
• Language: English
• Subjects: Analytical Chemistry; Bacteria; Bacteria - isolation & purification; Bacteria separation; Biochemistry; Blood cells; Blood Cells - microbiology; Bloodstream infection; Clinical Biochemistry; Dilution; E coli; Elastoinertial microfluidics; Flow velocity; Humans; Infectious diseases; Medicin och hälsovetenskap; Microfluidic Analytical Techniques; Microfluidics; Microparticles; Non‐Newtonian fluid; Separation; Sepsis; Sepsis - blood; Sepsis - microbiology; Viscoelastic flow
• ISSN: 0173-0835; 1522-2683; 1522-2683
• DOI: 10.1002/elps.202100140

Improved sample preparation has the potential to address unmet needs for fast turnaround sepsis tests. In this work, we report elasto‐inertial based rapid bacteria separation from diluted blood at high separation efficiency. In viscoelastic flows, we demonstrate novel findings where blood cells prepositioned at the outer wall entering a spiral device remain fully focused throughout the channel length while smaller bacteria migrate to the opposite wall. Initially, using microparticles, we show that particles above a certain size cut‐off remain fully focused at the outer wall while smaller particles differentially migrate toward the inner wall. We demonstrate particle separation at 1 μm resolution at a total throughput of 1 mL/min. For blood‐based experiments, a minimum of 1:2 dilution was necessary to fully focus blood cells at the outer wall. Finally, Escherichia coli spiked in diluted blood were continuously separated at a total flow rate of 1 mL/min, with efficiencies between 82 and 90% depending on the blood dilution. Using a single spiral, it takes 40 min to process 1 mL of blood at a separation efficiency of 82%. The label‐free, passive, and rapid bacteria isolation method has a great potential for speeding up downstream phenotypic and genotypic analysis.