A statistically significant ethanol decrease can coexist with a result judged acceptable under a chosen analytical tolerance. Kocak and colleagues demonstrated that distinction in 80 frozen plasma specimens. Their findings should not be simplified to “freezing prevents change” or applied automatically to postmortem whole blood.
Four groups, not repeated testing of every sample
The hospital study used specimens from intoxicated drivers in Turkey during November 2013–April 2014. Twenty different specimens were assigned to each storage period: two, three, four or five months at −20°C. The same 80 specimens were not repeatedly opened and tested at every interval.
Blood was collected into sodium-fluoride/K3EDTA tubes, mixed and centrifuged. Plasma was divided into an immediately tested aliquot and a tightly stoppered polystyrene storage tube. Frozen aliquots were thawed and reanalyzed on one day. Hemolyzed, clotted, improperly filled or otherwise unsuitable collection specimens were rejected.
The measured matrix and method were specific
Ethanol was measured in plasma by an alcohol-dehydrogenase enzymatic assay on a Roche Cobas C501, in duplicate with internal controls. This was not chromatographic confirmation of ethanol in decomposed autopsy whole blood.
The study included initial concentrations above the reagent’s stated lower detection limit of 2.17 mmol/L, equivalent to 0.1 g/L. Reported intra-assay CVs were 8.27% and 2.36% at the two control levels. The article did not independently validate every possible interference associated with postmortem matrices.
All four groups showed statistically significant decreases
Median plasma ethanol changed from 31.4 to 29.1 mmol/L after two months, 34.3 to 30.5 after three, 24.3 to 19.2 after four and 39.3 to 25.8 after five. Each paired group comparison had P < 0.001.
Those medians describe different groups, with different initial concentration distributions. They do not establish a precise month-by-month loss curve for one individual specimen. The study also reported no ethanol generation in the tested specimens under these conditions.
The authors applied a separate 25% criterion
The paper compared deviations with a ±25% allowable-total-error criterion that it attributed to CLIA’88 proficiency-testing limits. This was the authors’ selected evaluation rule, not a conclusion that a 25% discrepancy is acceptable for every forensic opinion or legal purpose.
The abstract emphasizes the five-month group, but the detailed results report individual deviations exceeding the criterion in 11 of 20 five-month tubes and four of 20 four-month tubes. No such exceedances were reported in the two- and three-month groups. A broad statement that every four-month specimen remained within the chosen tolerance would omit a material result.
Why the distinction matters for interpretation
A significance test asks whether the data support a change beyond random variation under its model. An analytical tolerance asks a different question about the size of a discrepancy. Neither decides whether a disputed case result supports impairment, timing of alcohol use or cause of death.
The authors concluded that plasma could be stored for approximately three to four months under suitable conditions, while recommending laboratory-specific criteria. The individual four-month exceedances qualify that practical conclusion. The study did not demonstrate stability across other containers, temperatures, matrices or postmortem conditions.
For death investigations involving a hospital specimen, preserve its plasma designation and enzymatic method when comparing it with later whole blood. Our discussion of hospital ethanol specimens addresses additional matrix and method differences that should remain separate from storage loss.
Primary source
Kocak FE, Isiklar OO, Kocak H, Meral A. Comparison of blood ethanol stabilities in different storage periods. Biochemia Medica. 2015;25:57–63. doi:10.11613/BM.2015.006.
Leave a Reply