Mild detergent preserves bacteria for rapid sepsis susceptibility testing

Bloodstream infections are among the most dangerous conditions clinicians face, capable of spiralling into sepsis within hours and killing a significant fraction of the patients they touch. The standard diagnostic pathway, in which blood is cultured, Gram-stained, subcultured onto agar plates, identified by mass spectrometry and then subjected to overnight susceptibility testing, can take more than 48 hours to deliver actionable answers. That delay is not benign: mortality from bloodstream infections increases by roughly eight percent for every hour that effective antibiotic therapy is postponed. A team at the University of Southampton has now examined a critical but often overlooked link in the chain of rapid diagnostics: whether the chemical tricks used to strip blood cells away from bacteria leave the pathogens alive and healthy enough for fast susceptibility testing.

The problem arises because the most widely used rapid sample preparation method, the commercially available MBT Sepsityper kit, was designed for a task that does not require living bacteria. The kit uses a proprietary lysis buffer, thought to contain ionic detergents such as sodium dodecyl sulphate, to rupture red blood cells, followed by high-speed centrifugation at around 16,000 times gravity to pellet the bacteria and discard cellular debris. The resulting organisms are fed into matrix-assisted laser desorption/ionisation time-of-flight mass spectrometry, which identifies pathogens from their protein and lipid fingerprints. For that purpose, viability is irrelevant. But the new generation of rapid phenotypic antimicrobial susceptibility tests, which promise results within two to three hours, depends entirely on bacteria that are actively dividing the moment they emerge from the extraction process.

This distinction matters most for Gram-positive pathogens. Organisms such as Staphylococcus aureus, which together with other Gram-positive bacteria account for more than half of bloodstream infections, possess a thick cell wall and membrane architecture that renders them far more vulnerable to detergent damage than their Gram-negative counterparts such as Escherichia coli. Previous studies have already hinted at the problem, reporting lower identification scores for Gram-positive organisms processed with Sepsityper. The Southampton group, led by Hywel Morgan and Daniel Spencer, set out to quantify exactly how much harm five different blood lysis protocols inflict on bacterial viability during the earliest stages of growth, the window that rapid susceptibility assays cannot afford to lose.

The researchers spiked healthy human blood with laboratory strains of E. coli and S. aureus, incubated the mixtures in standard aerobic blood culture bottles until they flagged positive, and then processed each bottle with one of five extraction chemistries: the Sepsityper kit, saponin alone, an ammonium chloride and potassium bicarbonate buffer, deionised water lysis by hypoosmotic shock, and a combination of saponin with the non-ionic detergent Triton X-100 supplemented with reducing agents to suppress reactive oxygen species. Each protocol lyses red blood cells by a different mechanism, from detergent-mediated membrane permeabilisation to osmotic swelling, and each carries its own trade-off between how completely the blood cells are destroyed and how gently the bacteria are treated.

To judge the outcome, the team turned to microfluidic impedance cytometry, a label-free technique that measures the electrical properties of thousands of individual cells as they stream through a microchannel lined with microelectrodes. An alternating voltage is applied at two frequencies, 5 MHz and 40 MHz. At the lower frequency the signal reflects cell volume, while at the higher frequency it is sensitive to the cell membrane and wall. Because the channel is only about 20 by 40 micrometres, bacteria can be counted and characterised one by one, even in the presence of residual red cell ghosts and debris, provided their electrical signatures do not overlap. The technique has already been shown to detect antibiotic-induced changes in bacterial impedance within two hours, forming the basis of a rapid susceptibility method known as iFAST.

The impedance data revealed striking differences between the extraction protocols. For E. coli, which reaches very high concentrations in blood culture bottles, all five methods yielded counts in agreement with conventional colony counting, and the bacteria resumed normal growth after only a brief 30-minute lag. For S. aureus, which grows more slowly and is present at roughly tenfold lower concentrations at the time the bottle flags positive, the picture changed dramatically. Samples treated with Sepsityper produced far fewer detectable organisms immediately after extraction, and the population did not begin to grow until two and a half hours had passed. Untreated and saponin-treated samples, by contrast, showed only a one-hour delay before entering exponential growth, with doubling times of around 65 to 67 minutes that were statistically indistinguishable from each other.

Independent confirmation came from optical flow cytometry using live/dead fluorescent staining. The team labelled bacteria with SYTO 9, a green dye that enters all cells, and propidium iodide, a red dye that penetrates only cells with damaged membranes. Immediately after Sepsityper extraction, most S. aureus cells showed strong propidium iodide signal, indicating compromised membranes, while saponin-treated cells looked essentially identical to untreated controls and recovered as quickly. The likely culprit is the ionic detergent in the Sepsityper buffer, which can permeabilise bacterial membranes and may also provoke oxidative stress. Colony imaging after overnight incubation reinforced the finding: S. aureus colonies isolated with Sepsityper were significantly smaller than those from untreated samples, whereas saponin-treated colonies showed no significant difference.

Extraction efficiency told a similar story. Measured against colony counts from untreated cultures, viable E. coli recovery ranged from 37 to 54 percent across the protocols, while S. aureus recovery ranged from just 5 to 35 percent. Saponin delivered the highest impedance-based efficiency for S. aureus at 35 percent and produced the cleanest samples, with debris particles small enough to be separated from the bacteria electrically. The ammonium chloride protocol generated so much debris overlapping with the S. aureus population that accurate counting became impossible, and deionised water lysis performed worst of all, possibly because S. aureus cells adhered to red blood cells and sedimented away with them during centrifugation. For E. coli, the data even suggested that lysis and centrifugation may be unnecessary, since fast-growing Gram-negative organisms can be discriminated from blood cells directly.

Beyond viability, the study highlights a subtler advantage of impedance cytometry for the diagnostic workflow itself. Preparing the inoculum for a phenotypic susceptibility test normally relies on turbidity matching against a McFarland standard, but blood cells and debris inflate turbidity and skew the measurement. Because impedance cytometry counts each bacterium as a discrete electrical event and can distinguish bacteria from red cells, ghosts and platelets by their size and high-frequency phase, it offers a direct route to accurate inoculum preparation at the concentration recommended by EUCAST guidelines. The technique can even discriminate E. coli from S. aureus by their electrical phase, an analogue of conventional Gram staining performed without any reagents or stains.

The implications reach well beyond the laboratory bench. Any rapid susceptibility platform that promises answers within three hours of a blood bottle flagging positive will be defeated if its sample preparation quietly kills or stuns a large fraction of the pathogens, and the organisms most likely to be harmed are precisely the Gram-positive bacteria that dominate bloodstream infections. The Southampton results point to saponin, a mild natural detergent, as the chemistry of choice for preserving viable organisms, and the team now plans to integrate on-chip lysis, automated washing and size-based separation into a streamlined microfluidic sample preparation system. Clinical samples, polymicrobial cultures and a broader panel of Gram-positive pathogens remain the next hurdles, but the study makes clear that the road to same-hour antibiotic decisions runs through gentler chemistry.

Subject of Research: Rapid isolation of viable bacteria from positive blood cultures using microfluidic impedance cytometry to compare blood lysis protocols for antimicrobial susceptibility testing

Article Title: Isolating viable bacteria from positive blood cultures: impedance cytometry analysis

Article References: Isolating viable bacteria from positive blood cultures: impedance cytometry analysis. (n.d.). https://doi.org/10.1007/s10544-026-00847-5

Image Credits: AI Generated

DOI: 10.1007/s10544-026-00847-5

Keywords: bloodstream infections, sepsis, antimicrobial susceptibility testing, impedance cytometry, microfluidics, Sepsityper, saponin, Staphylococcus aureus, Escherichia coli, bacterial viability, MALDI-TOF, antimicrobial resistance