EVIDENCE LIBRARY
Can a Year-Old DUI Blood Alcohol Retest Match the Original—and Does Glass Versus Plastic Matter?
Can an archived DUI blood-alcohol sample retest lower after 12 months? A whole-blood HS-GC study compares ethanol loss in glass versus plastic tubes.
Short answer: A blood-alcohol result obtained a year after collection need not reproduce the original result, even when a sample was refrigerated and collected in a fluoride–oxalate tube. In a controlled study, ethanol concentrations in deliberately fortified whole blood gradually decreased during storage at 4 °C. After 364 days, the average loss was approximately 0.11 g/L in both glass and plastic tubes. The investigators did not find an analytically significant difference attributable to tube material under their test conditions. That finding does not supply a universal correction factor for a defendant’s retained specimen or prove that every archived sample changes by the same amount. Jones and Ericsson, Practical Laboratory Medicine (2016), DOI: 10.1016/j.plabm.2016.02.002.
Why does an old blood-alcohol retest raise a different scientific question?
An original ethanol result describes the analyte concentration in a specific whole-blood specimen when that sample was tested. A later analysis describes the concentration in whatever material remains at the later time. Both numbers can be analytically valid for their respective specimens and dates without being identical. The relevant questions are whether ethanol was preserved over the interval, what handling occurred, and whether any observed difference exceeds the laboratory method’s expected analytical variation.
These questions arise when attorneys request independent testing of retained DUI specimens, challenge a reported BAC, or evaluate a prosecution or defense expert’s interpretation of a delayed reanalysis. A later lower result might be consistent with storage-related loss; it does not automatically prove the initial result was wrong. Conversely, the mere presence of sodium fluoride is not affirmative proof of perfect long-term stability. The distinction matters more than whether the container looks like the original evidence tube.
What exactly did Jones and Ericsson test?
The investigators obtained hospital blood, mixed it to redistribute cells and plasma, and added known amounts of ethanol to create four target concentrations: 0.20, 1.00, 2.00, and 3.00 g/L. The corresponding initial measured concentrations were 0.195, 1.016, 2.020, and 3.019 g/L. Samples were dispensed into nominal 10 mL evacuated tubes filled with approximately 9 mL of blood. The paired tube types were glass and plastic, each supplied with sodium fluoride and potassium oxalate. The researchers then studied changes after refrigerated storage.
This was a laboratory stability experiment, not a prospective series of real-world impaired-driving cases. Ethanol was deliberately introduced into hospital-supplied blood. The work did not compare venous blood with breath, serum, plasma, or urine BAC measurements, and it did not follow the physiology of people drinking alcohol. It tested what happened to the measured ethanol concentration in stored whole blood under specified conditions. The complete primary article is available through the publisher DOI record.
How were samples stored and ethanol measured?
The investigators evaluated 2, 7, 14, 28, 91, 182, and 364 days after preparation. Following an initial two-day positioning comparison at room temperature, the tubes were kept upright in a refrigerator at 4 °C between scheduled analyses. Before each measurement, the blood was mixed; material was withdrawn for headspace gas chromatography (HS-GC). The study used two GC instruments and independent operators. At the study’s concentration levels, its reported analytical coefficient of variation ranged from approximately 2.9% at 0.20 g/L to 1.1% at 3.00 g/L.
That measurement architecture matters. The authors evaluated replicated measurements and specified a critical difference that a change had to exceed before it was considered analytically significant. Reanalysis was conducted at multiple intervals, with the stoppers opened to withdraw aliquots; that is not automatically equivalent to an unopened evidence tube held continuously for the same period. The initial two-day room-temperature comparison was likewise a limited experimental condition, not a validation of every shipping delay, temperature excursion, or overturned vial.
What happened after nearly twelve months?
The central forensic result comes from the article’s Table 3. All numbers below are measured whole-blood ethanol concentrations in g/L. The day-364 values represent experiment-specific group results, not a predicted concentration for any new retained specimen.
| Initial measured ethanol (g/L) | Day 364, glass (g/L) | Day 364, plastic (g/L) | Loss, glass (g/L) | Loss, plastic (g/L) |
|---|---|---|---|---|
| 0.195 | 0.111 | 0.107 | 0.084 | 0.088 |
| 1.016 | 0.910 | 0.915 | 0.106 | 0.101 |
| 2.020 | 1.891 | 1.880 | 0.129 | 0.140 |
| 3.019 | 2.896 | 2.900 | 0.123 | 0.119 |
Across those four starting levels, the mean absolute reduction after approximately one year was 0.111 g/L in glass and 0.112 g/L in plastic. The full ranges of absolute loss were 0.084–0.129 g/L in glass and 0.088–0.140 g/L in plastic. The study’s table explicitly identifies the day-364 losses as analytically significant within its measurement design, while identifying the glass-minus-plastic comparisons as not analytically significant. Those are distinct comparisons: concentration changed over time, but container material was not a demonstrated explanation for a difference.
Is ethanol loss a fixed percentage of the original BAC?
No. The authors emphasized that the absolute loss was fairly similar across the studied starting levels, whereas the percentage loss depended strongly on the denominator. At approximately 0.20 g/L initially, the year-long reduction represented about 43–45% of the starting concentration. At approximately 3.00 g/L, it represented only about 3.9–4.1%. A percentage taken from one starting level therefore cannot be applied uncritically to another concentration.
Nor may the average absolute decrease of about 0.11 g/L be added automatically to a real case’s late result to manufacture an “original BAC.” The investigation did not build or externally validate a person-specific or case-specific reconstruction equation. It evaluated a finite set of spiked blood samples with a particular fill volume, preservative composition, tube construction, analytical method, and storage schedule. Real evidence may have different preanalytical conditions and an incompletely documented history.
When did the concentration decline become analytically detectable?
The researchers reported that decreases exceeded their calculated analytical critical differences after roughly 14 days for the two lower starting concentrations (approximately 0.20 and 1.00 g/L), and after roughly 28 days for the higher starting concentrations (approximately 2.00 and 3.00 g/L). These observations do not mean that ethanol was physically unchanged before those days. They mean that, under the specific test precision and replicate structure, earlier observed differences were not large enough to meet the study’s analytical criterion.
A litigation opinion should not replace that distinction with a misleading deadline such as “fluoride preserves BAC for two weeks” or “any storage beyond four weeks invalidates a test.” Statistical detectability, biological preservation, and evidence admissibility are different issues. The experiment addressed the first two only within its defined laboratory conditions; it did not establish a judicial exclusion rule or a universal laboratory retention policy.
Did glass protect the ethanol better than plastic?
Not in this experiment. At the same initial ethanol levels and scheduled storage intervals, no analytically significant difference between glass and plastic was reported. The study’s result directly challenges the blanket argument that plastic evidence tubes must produce greater ethanol loss than glass tubes when the tubes contain the specified fluoride–oxalate additives and are handled as in this experiment.
However, failure to find a significant difference is not the same as establishing formal equivalence for every manufacturer, cap, seal, sample-to-air ratio, temperature, or duration. The study did not evaluate every commercially available polymer or every compromised stopper. An individualized interpretation should use the actual container brand, lot, cap condition, fill volume, temperature record, opening history, and contemporaneous analytical documentation—not solely the word “plastic.”
What could explain loss during storage, and what remains unproven?
Jones and Ericsson discussed possible processes associated with gradual ethanol loss during refrigerated storage, including chemical changes in blood constituents. Such discussion offers mechanistic hypotheses, not direct proof assigning each decrease to a single quantified reaction pathway in an individual sample. The experiment was not designed to isolate each potential chemical, microbiological, evaporation, or tube-access contribution through a complete mechanism-specific balance study.
The direction of change matters too. In this experiment the measured concentration declined; it does not follow that ethanol can never increase in a different specimen. For example, postmortem microbial formation and specimen contamination involve different biological and handling questions from this sterile, spiked, refrigerated blood model. This paper does not adjudicate those possibilities. It also says nothing directly about the metabolism or elimination of alcohol from a living person’s blood stream after drinking.
What evidence is needed to interpret an actual retained DUI tube?
Before arguing that an older retest should replace or discredit the original reported BAC, review the collection and storage history alongside the complete analytical packet. The following distinctions keep the expert opinion tied to testable evidence.
| Evidence or question | What this study can support | What still requires case-specific proof |
|---|---|---|
| Original whole-blood ethanol result | Measured concentrations in stored blood may later decline despite refrigeration. | Original chromatograms, calibrators, controls, replicates, uncertainty, method and sampling time. |
| Glass versus plastic collection tube | No detected material effect under the tested fluoride–oxalate conditions. | Actual tube model, additives, cap integrity, nominal volume, fill amount and sample handling. |
| Time and temperature | Changes were observed at defined intervals up to 364 days at 4 °C. | Continuous or interval-specific custody, receipt and refrigeration records, freezing or warming excursions. |
| Delayed retest | A lower later concentration can be compatible with storage-associated change. | Both analytical runs, validated method comparability, measured uncertainty, intervening openings and aliquot losses. |
| Driving impairment or collision causation | Nothing directly: the research did not assess human function or driving. | Independent clinical, behavioral, temporal and case evidence; an admissible causal method. |
For complementary specimen-first interpretation, see our published resources on specimen collection and handling, sodium fluoride in blood-alcohol evidence, and headspace gas chromatography for ethanol. These are method and evidence-context resources; they are not substitutes for the primary stability experiment.
Frequently asked questions
Can an old blood sample test lower than it did originally?
Yes. Jones and Ericsson observed declining ethanol concentrations during refrigerated storage, with analytically significant changes under their study’s replicate and method conditions. The result supports a possibility, not an exact prediction for any particular retained sample.
Does a lower retest prove the first BAC was inaccurate?
No. A later result measures the remaining ethanol at a different time and after a different handling history. Both tests can be technically valid. Determine whether the results are analytically comparable and whether sample integrity was adequately documented before interpreting the difference.
Do plastic blood tubes cause larger ethanol losses than glass tubes?
The studied plastic tubes did not. The paper found no analytically significant glass–plastic difference under its test conditions. It did not establish universal equivalence across all containers and storage environments.
Does sodium fluoride keep blood alcohol constant indefinitely?
No. Both tube types contained fluoride–oxalate additives, yet measured ethanol decreased over the year. A preservative can be relevant to stability without guaranteeing that all future readings match the initial value.
Can an expert add 0.11 g/L to every one-year-old result?
No. About 0.11 g/L was the mean observed loss in this finite experiment, not a validated correction for an individual forensic specimen. Tube conditions, timing, initial concentration, analytical method, and possible other changes must be assessed independently.
Did the study measure breath alcohol, serum alcohol or impairment?
No. It measured ethanol in prepared whole-blood samples using headspace gas chromatography. It did not test breath or serum conversion factors, psychomotor tasks, driving, or individual crash causation.
What is the most important qualification for court?
State both propositions: substantial storage-associated declines were measured in this controlled experiment, and a glass–plastic difference was not found. Then identify why the study’s sealed-tube design, repeated sampling, preservative, refrigeration, and analytical method may or may not match the particular evidence being litigated.
Bottom line: distinguish tube material from storage-related concentration change
Jones and Ericsson’s 2016 research supplies measured evidence that whole-blood ethanol can decline during extended refrigerated storage and that, under their conditions, glass and plastic fluoride–oxalate tubes did not differ detectably. It does not furnish an automatic retest correction, diagnose a person’s impairment, establish the cause of a collision, or resolve an individual chain-of-custody dispute. A reliable forensic opinion begins with the original and repeat laboratory records, their methods and uncertainty, and the actual specimen’s documented history. Primary study: Jones AW, Ericsson E. Decreases in blood ethanol concentrations during storage at 4 °C for 12 months were the same for specimens kept in glass or plastic tubes. Practical Laboratory Medicine. 2016;4:76–81. doi:10.1016/j.plabm.2016.02.002.