When a blood-alcohol result is challenged in court, the prosecution increasingly reaches for a tidy 2023 technical note — Kosecki, Brooke & Raines, “Lack of fermentation in antemortem blood samples stored unstoppered in various locations.” The pitch is simple: researchers left blood tubes open in restrooms, locker rooms, and a jail intake area, stored them, and found no alcohol. Case closed on fermentation.

Except it isn’t. The experiment is real and competently run, but it establishes something far narrower than the way it gets cited. And on the two questions that actually decide a fermentation challenge — does sodium fluoride stop Candida, and can the organism reach blood — the paper’s own cited sources point the other way. Here is the close read, with the data.

01 / THE CLAIM vs. THE DATAWhat the experiment actually showed

Strip away the framing and one fact remains: ethanol-negative blood, drawn from three sober volunteers, left open in various rooms and then stored, did not become ethanol-positive. That’s it. From there the authors leap to a sweeping conclusion — that there is “no plausible mechanism” for routine samples to ferment and that such challenges “lack merit.”

A negative result from a system never shown capable of a positive result is uninterpretable. It proves only that blood not contaminated with a fermenting organism did not ferment — which was never in dispute.

The gap between what was tested and what was claimed is the whole story. The table below is the heart of it.

02 / SODIUM FLUORIDEThe preservative doesn’t stop Candida

The paper leans on sodium fluoride as the safeguard that makes fermentation implausible. But the classic study it cites — Blume & Lakatua (1973) — tested exactly that, and found fluoride suppressed every contaminating organism it tried except the one that matters.

First, how much ethanol these organisms make in blank blood over a single day. Candida albicans is not a minor player:

Now the decisive experiment. After 24 hours, with and without 50 mg of sodium fluoride per tube:

Chang & Kollman (1989) — another Kosecki citation, and one focused specifically on Candida in blood — reached the same place: once fermentation began in inoculated blood at room temperature, it “was not affected by the presence of sodium fluoride,” climbing toward a plateau near 0.08% w/v. Even an industrial study in the record (Arshad 2011) found sodium fluoride “not much effective against the continuous contamination load.” The authors themselves concede the fluoride literature is “inconsistent.”

03 / GLUCOSEThey removed the fuel, then reported no fire

Fermentation needs sugar. Yajima et al. (2006) showed the relationship directly: Candida produced ethanol from glucose in human blood, and the more glucose available, the more ethanol formed. With no added glucose, undiluted blood produced none at all.

The contradiction is fatal. Kosecki criticizes the earlier positive studies precisely because they used added glucose — then draws blood from three healthy, sober adults (the lowest-glucose substrate available), never measures their glucose, and treats the predictable null as proof the mechanism is implausible. Diabetic, hyperglycemic, and glucose-infused arrestees — the very people most at fermentation risk — are entirely outside the study.

04 / TEMPERATURERefrigeration did the work — not the preservative

Chang & Kollman’s headline finding was that fermentation is steeply temperature-dependent. Cold storage, not fluoride, was the effective control. That reframes the issue entirely: the protective factor a defense actually probes is the sample’s temperature history, which real cases often can’t account for.

05 / PLAUSIBILITYCandida is not a harmless speck

A pillar of the paper is that C. albicans is unlikely to reach a living person’s blood. Yet a blood–brain barrier study in the same record shows the organism adhering to, invading, surviving inside, budding within, and transcytosing human endothelial cells. It is a recognized invasive pathogen, not an inert contaminant. Combined with the existence of candidemia, indwelling lines, immunocompromise, and trauma, that defeats any categorical “it can’t get into blood” claim. The defense theory needs only possible in this case — not common.

Meanwhile the paper’s own headline probability — the oft-quoted 1-in-250,000 chance of airborne contamination — isn’t a measurement at all. It’s extrapolated from ambient mold surveys of unrelated buildings; the authors never sampled the air of the rooms they used, nor cultured the blood afterward.

06 / BOTTOM LINEWhat the paper can and cannot carry

So cite it for what it is: a narrow, negative ambient-exposure study using clean blood from three healthy volunteers, with no confirmed contamination, no positive control, and no culture data. It is useful only against an exaggerated version of the defense. It does not rebut a fact-specific challenge built on collection defects, storage failures, underfilled tubes, poor preservative mixing, contamination, glucose-rich substrate, or chain-of-custody gaps.

The literature gives the counter on every point that matters: Candida makes ethanol in blood; glucose drives it; fluoride is condition-dependent; n-propanol is no reliable exclusion; and Candida is genuinely invasive.