Can EtG or PEth Be Positive With Negative Blood Alcohol? 1,149 Samples

Short answer: No. A nonreportable plasma ethanol result does not establish that a person has not consumed alcohol. In a retrospective laboratory study of 1,149 clinical venous blood specimens, only 95 had reportable plasma ethanol, while 454 contained reportable ethyl glucuronide (EtG) and 534 met the laboratory’s 0.050 µmol/L phosphatidylethanol (PEth) reporting threshold. These are different analytes, different analytical procedures, and different biological windows. A biomarker result can add evidence of prior alcohol exposure, but none of these tests alone proves the exact drinking time, amount, intoxication at an earlier event, intentional misreporting, or driving impairment.

This analysis is confined to Neumann, Beck and Böttcher’s original 2021 primary study, Phosphatidylethanol, ethyl glucuronide and ethanol in blood as complementary biomarkers for alcohol consumption, published in the Journal of Mass Spectrometry and Advances in the Clinical Lab. The study examined specimens sent for clinical PEth testing; it was not a timed alcohol-dose trial or a validated forensic test of whether an individual told the truth.

What did the 1,149-sample study actually find?

Researchers analyzed 1,149 venous EDTA specimens sent to one laboratory for routine PEth testing. The source population included specimens referred from addiction clinics, employment-related medical services, occupational health units and forensic hospitals. Samples arrived through postal transport one to three days after collection and were then kept at 4°C before analysis within 24 hours. These are clinical laboratory submissions, not 1,149 experimentally dosed volunteers or a probability sample of the population. The paper does not establish that all specimens came from different people.

Reportable ethanol was relatively uncommon compared with the two biochemical markers of prior exposure. That mismatch is the key result for anyone tempted to interpret a negative ethanol result as proof of abstinence. A specimen collected after alcohol has been eliminated may have no reportable ethanol even when another compound persists. Equally, a marker below a stated reporting decision level does not retrospectively prove a person drank nothing.

The denominators above are specimens, not necessarily unique individuals. The PEth results of 635 detected and 534 meeting the 0.050 µmol/L reporting level describe two different cutoffs; they should never be presented as contradictory counts or treated as interchangeable definitions of a positive result.

Did they measure ethanol, EtG and PEth in the same type of material?

All tests originated from venous EDTA blood submissions, but the assayed matrices and methods were not identical. EtG and PEth 16:0/18:1 were measured in whole blood using liquid chromatography–tandem mass spectrometry (LC–MS/MS). Ethanol was measured in separated plasma by the DRI Ethyl Alcohol enzymatic assay on a Beckman Coulter AU680 clinical chemistry analyzer. Describing that plasma ethanol value as an independently confirmed forensic whole-blood alcohol concentration would misstate the work.

The analytical ranges also differed: 1–5,000 ng/mL for blood EtG, 0.014–7.112 µmol/L for PEth, and 0.10–6.0 g/L for plasma ethanol. Concentrations above an instrument range were measured after dilution. Reported quality-control coefficients of variation and biases were below 9% for EtG, 8% for PEth and 5% for ethanol at the authors’ tested control levels. Those figures describe analytical performance of the methods in that laboratory, not sensitivity or specificity for determining when an individual last drank, whether a driver was impaired, or whether a witness was candid.

For the scientific difference between laboratory ethanol testing and forensic alcohol methods, see this site’s independently published explanation of blood alcohol testing by headspace gas chromatography.

Why can EtG or PEth be positive when plasma ethanol is negative?

Direct ethanol testing identifies the parent alcohol if its concentration remains within the assay’s reportable range at the relevant sampling and testing conditions. EtG is a metabolite formed after ethanol exposure. PEth is a phospholipid formed in the presence of ethanol and may remain measurable over a different, generally longer period. Thus the three results can legitimately disagree without laboratory error or a contradiction about whether the same individual ever drank.

In this laboratory series, 454 specimens were EtG-positive compared with only 95 plasma-ethanol-positive. Even without reconstructing every cell of the authors’ Venn diagram, those marginal counts show that at least 359 EtG-positive specimens could not also be ethanol-positive, because only 95 specimens were ethanol-positive in total. This is a simple minimum count from published totals, not a reported clinical sensitivity, a timing classifier or a claim about any named person.

Likewise, of the 534 specimens with PEth at or above 0.050 µmol/L, the study reported EtG positivity in approximately 73%. Among specimens with PEth at or above 0.30 µmol/L, about 86% were EtG-positive. These are associations within a PEth-referred clinical sample. They cannot be turned into universal probabilities that a drink occurred on a particular date, that reported abstinence is false, or that the person was intoxicated at a past encounter.

Were PEth and EtG correlated strongly enough to reconstruct a drinking history?

No. Across the study samples the coefficient of determination for measured whole-blood PEth versus EtG was only r² = 0.15. That means a simple linear association accounted for about 15% of the variation in one measured concentration relative to the other under the authors’ plotted analysis. It does not mean the markers agreed 15% of the time, that the combined panel has 85% error, or that the two biological processes were proven mathematically independent in every circumstance.

Higher plasma-ethanol categories did tend to have higher EtG concentrations. For the ethanol-negative category below 0.10 g/L, the authors gave a median EtG concentration of 16 ng/mL among the relevant plotted observations, compared with 125 ng/mL in the 0.10–0.50 g/L plasma-ethanol category and 1,815 ng/mL in the above-1.0 g/L category. These group summaries cannot be inverted into an individual’s ethanol result, ingestion time, or exact quantity of alcohol. The paper was observational, had no fixed dose, and supplied no independently established last-drink time for each sample.

What can a three-marker pattern support—and what remains unproved?

These are interpretive boundaries of the Neumann dataset. They are not diagnostic thresholds validated in that paper against continuous monitored drinking, police-grade evidential breath testing, field sobriety tests, or a known-time criminal incident.

Does a double-negative EtG/PEth result prove abstinence?

No. The authors offered strong clinical interpretive language for specimens negative for both markers. But their retrospective dataset did not independently verify that every such patient abstained through a defined period. It included no blinded behavioral reference standard, serial individually observed dose records, or population-wide calculation of the negative predictive value of the combined tests. A below-threshold result identifies an analytical reporting outcome under the laboratory’s rules; it cannot prove that no ethanol was consumed at any point in the past.

That caution is especially necessary when a laboratory’s analytical detection limit differs from its chosen reportable threshold. The authors measured PEth down to 0.014 µmol/L but routinely classified PEth at 0.050 µmol/L or above as reportable. A number below 0.050 is not necessarily evidence of zero molecules or zero prior consumption. Likewise, an EtG result below 1 ng/mL must be interpreted in relation to the assay, possible elapsed time, specimen handling and exposure history, not as a universal abstinence certificate.

How does sample handling affect interpretation?

The collected EDTA specimens traveled through postal delivery for one to three days before laboratory receipt, and the paper describes refrigerated storage after receipt. It does not furnish the collection-to-test temperature trace, preservative circumstances, and exact timeline of each individual specimen. Accordingly, the paper cannot serve as an individualized stability reconstruction or establish the ethanol concentration at the moment of venipuncture from the value obtained later.

The investigators reported a PEth storage reanalysis: among 83 ethanol-containing specimens stored under the studied 4°C conditions, the mean day-28/day-0 PEth ratio was 0.99. That specific average under controlled reanalysis conditions does not establish that every shipped specimen was stable throughout transport, that postcollection PEth formation is impossible under other conditions, or that the clinical sample tested in one case is necessarily equivalent to a properly preserved forensic exhibit. The authors themselves discussed the possibility of continued PEth formation after blood collection when ethanol is present.

A case-specific assessment therefore requires the original sample identification, specimen matrix, collection and receipt times, storage and transport records, calibrated quantitative results, reporting thresholds, QC evidence and any available confirmatory tests. For a framework, consult this site’s specimen collection and matrix guide.

Can these biomarkers establish impairment or a time-of-driving BAC?

No. None of the 1,149 submissions involved a controlled driving simulation or timed psychomotor assessment in the published dataset. Neither EtG nor PEth is a measured blood-ethanol concentration. Plasma ethanol measured later in a clinical laboratory is a different result again. The authors did not validate a mathematical transformation from EtG or PEth to an individual’s blood or breath alcohol at an earlier crash, arrest, shift, or medical encounter.

Inferring an earlier alcohol concentration would introduce additional assumptions about the exposure timeline, absorption and elimination, the specimen actually tested, laboratory timing and individual physiology. The source provides no tested person-specific equation or prediction intervals sufficient for that reconstruction. More generally, pharmacokinetic interpretation requires careful separation of measured concentration, calculated exposure and behavioral conclusions.

What are the major study limitations for expert testimony?

First, this was a single-laboratory, referral-selected clinical series: people were tested because PEth analysis had already been requested. Positivity percentages cannot be applied as prevalence rates in all motorists, all workers or all people who deny drinking. Second, the authors reported numbers of specimens, not the longitudinal independence of all 1,149 people. Third, they did not independently observe alcohol consumption or establish each subject’s last-drink time; therefore they could not calculate individual classification accuracy against a timed exposure history.

Fourth, concentrations came from different analytical matrices and instruments, and some specimens traveled through the post before testing. Fifth, neither the reported r² nor the observed frequency of concurrent positives measures a biomarker’s ability to identify impaired driving, decisional capacity, event causation or deliberate deception. Lastly, the particular PEth heavy-drinking cutoff and laboratory reporting rules are not substitutes for the governing case-specific laboratory validation, legal definitions or clinical history.

What should counsel or a forensic toxicologist request?

For a disputed alcohol-exposure opinion, the underlying record matters more than a generic label of “positive” or “negative.” Obtain original chromatographic/LC–MS/MS quantitative information for PEth and EtG where available; the plasma-ethanol method and associated raw result; reporting and quantitation limits; calibrators and QC for each assay; contemporaneous collection and receiving records; tube/additive information; storage and transport conditions; analytical dates; and any repeated or independently corroborating specimens. Obtain the person’s relevant reported drinking history and the actual dates in dispute rather than treating retrospective biomarkers as a chronometer.

The opinion should say explicitly which proposition the evidence addresses: prior exposure, recent exposure, ethanol present in plasma when tested, historical whole-blood BAC, functional impairment, or event causation. The Neumann paper provides useful observational evidence about the first three analytical categories but does not validate the latter three for a named person.

Frequently asked questions

Can blood EtG be positive after plasma ethanol is no longer reportable?

Yes. The study documented many more EtG-positive than ethanol-positive clinical blood submissions. The two substances are analytically distinct and do not have identical detection windows.

Is the study’s ethanol result a forensic whole-blood BAC?

No. Ethanol was assayed enzymatically in separated plasma. EtG and PEth were measured by LC–MS/MS in whole blood. The matrices and methods should not be conflated.

Were all 635 detectable PEth specimens “positive” under the clinical reporting threshold?

No. PEth was analytically detected in 635 specimens at or above 0.014 µmol/L, but 534 met the 0.050 µmol/L reporting threshold used for the main classification.

Does r² = 0.15 mean the PEth and EtG tests are only 15% accurate?

No. It describes the weakness of a measured concentration association, not sensitivity, specificity, agreement, or predictive accuracy.

Can a high PEth with negative EtG identify the date of the last drink?

No. Differences in detection and persistence may explain that pattern, but the study did not provide a validated clock for individual consumption dates.

Can two negative markers prove a person abstained completely?

No. Below-threshold findings do not establish continuous observed abstinence, and this study did not independently validate a universal abstinence rule.

Can these results establish intoxication, DUI impairment or why a collision occurred?

No. The authors did not test driving, incident circumstances or individual impairment, and the biomarkers cannot be converted into a historical forensic BAC from this study alone.

Primary scientific source

Neumann J, Beck O, Böttcher M. Phosphatidylethanol, ethyl glucuronide and ethanol in blood as complementary biomarkers for alcohol consumption. Journal of Mass Spectrometry and Advances in the Clinical Lab. 2021;22:3–7. doi:10.1016/j.jmsacl.2021.09.005. The design, analytical methods, counts, concentration ranges and limitations above are taken from this complete original primary research article. Interpretive cautions identify questions that the study did not test and are not claims of additional experimental findings.