EVIDENCE LIBRARY
Can a Blood EtG/EtS Ratio Date the Last Drink? Study Limits
Can a blood EtG/EtS ratio date a last drink? A 26-person study found a group-mean timing fit, not a validated personal forensic clock.
Short answer: A measured blood ethyl glucuronide (EtG)-to-ethyl sulfate (EtS) ratio may provide information about elapsed time after a controlled alcohol dose, but it does not reliably timestamp a particular person’s last drink by itself. In a 26-volunteer study, investigators fitted a relationship between the group’s mean metabolite ratio and scheduled hours after alcohol administration. The favorable fit and mostly small errors reported within eight hours were not an independent, case-by-case validation of a forensic drinking-time clock. Neither metabolite concentration directly measures current ethanol concentration, intoxication, driving impairment, or the cause of an incident.
What did the blood EtG/EtS timing study actually test?
Wang and colleagues investigated whether blood concentrations of two minor ethanol metabolites could help estimate elapsed time after drinking. Their 2022 original research, “Estimating the time of last drinking from blood ethyl glucuronide and ethyl sulphate concentrations”, appeared in Scientific Reports. Twenty-six healthy young adults (14 men and 12 women, ages 22–27) participated in a single-dose protocol. After an overnight fast, each consumed 0.72 grams of ethanol per kilogram of body weight as 40% liquor with food, completing consumption within about 30 minutes.
Investigators drew serial venous blood samples before drinking and at 0.5, 1.5, 2, 3, 5, 8, 12, 24, 36, 48, and 120 hours after the start of consumption. They measured ethanol by headspace gas chromatography and measured blood EtG and EtS by liquid chromatography–tandem mass spectrometry (LC–MS/MS). That distinction matters: the laboratory quantified three different chemicals in collected blood, whereas a purported hour of last drinking was a later mathematical inference from metabolite measurements.
EtG and EtS are ethanol metabolites, not interchangeable BAC values
EtG forms through ethanol glucuronidation; EtS forms through ethanol sulfation. They are detectable markers of ethanol metabolism rather than direct measurements of the parent alcohol. Their concentrations may follow different time courses from ethanol itself. An EtG result reported in milligrams per liter cannot be substituted for a blood alcohol concentration in grams per liter. Similarly, dividing measured EtG by measured EtS produces a dimensionless ratio; it does not produce a measured time.
The relevant physical specimen was serial venous blood. This was not an oral-fluid assay, a urine detection-window experiment, a breath instrument study, or a postmortem validation series. A blood metabolite assay may help address prior exposure, while a separate ethanol test addresses contemporaneous ethanol in the tested specimen. For the collection and preservation details needed to interpret either finding, consult the site’s specimen collection framework.
What was measured, and what was fitted?
| Endpoint | Reported finding | What it represents |
|---|---|---|
| Ethanol, observed blood concentration | Mean peak 0.41 ± 0.11 mg/mL at the 1.5-hour scheduled sample | Directly analyzed parent ethanol in serial blood; 0.41 mg/mL equals 0.41 g/L, not a person’s driving performance. |
| EtG, observed blood concentration | Mean observed peak 0.29 ± 0.12 µg/mL at five hours | Measured conjugated metabolite; 0.29 µg/mL equals 0.29 mg/L. |
| EtS, observed blood concentration | Mean observed peak 0.16 ± 0.04 µg/mL at three hours | Measured conjugated metabolite; 0.16 µg/mL equals 0.16 mg/L. |
| EtG, model-derived peak time and concentration | 4.12 ± 1.07 h; 0.31 ± 0.11 mg/L | Noncompartmental pharmacokinetic summary, not an additional observed blood draw at exactly 4.12 h. |
| EtS, model-derived peak time and concentration | 3.02 ± 0.70 h; 0.17 ± 0.04 mg/L | Noncompartmental summary, not an observed individualized threshold. |
| EtG/EtS concentration ratio | Calculated from analyte concentrations | A derived biomarker ratio; any assigned drinking time comes from a separate fitted model. |
The parent-ethanol and metabolite measurements should not be conflated with their modeled pharmacokinetic parameters. For foundational distinctions among absorption, distribution, elimination, and fitted parameters, see pharmacokinetics in forensic toxicology.
How was the proposed “last drink” equation constructed?
The authors analyzed mean EtG and EtS blood concentrations at scheduled times following the single drinking episode. Their reported quadratic relationship was y = 1.646x2 − 0.9599x + 0.0878, in which x is the EtG-to-EtS concentration ratio and y is estimated elapsed time in hours under the study’s protocol. They reported R2 = 0.9904 for this fitted relationship and said the relative errors for the examined group-mean time points were mostly below 10% within eight hours.
That sounds precise, but the fit describes the data used to create the equation. The article’s method expressly used the mean concentrations across volunteers to form the ratios. A high coefficient of determination in those fitted means does not establish that a randomly selected person’s two measured concentrations can be inserted into the formula and yield a validated individual estimate with a 10% error bound. Nor does the reported within-study error establish a courtroom confidence or prediction interval for a new sample.
There is another timing issue: sampling times were measured from the start of drinking, whereas participants had as long as 30 minutes to finish. Those are not necessarily identical to the clock time of the final swallow. The scientific statement supported by this protocol is narrower than an exact “last drink at 10:47 p.m.” reconstruction.
Why does a strong laboratory correlation not establish an individual forensic clock?
A useful individual prediction method would need to characterize between-person variability, performance on observations not used to fit the relationship, analytical uncertainty, uncertain dose and drinking duration, and conditions different from this single supervised exposure. The reported group-level regression was not a separately validated algorithm for all sexes, ages, body sizes, repeated drinking episodes, food conditions, illness states, or samples taken outside the fitted interval.
The paper did not demonstrate blinded external validation, establish a validated individual prediction interval, or test a real-world case series with unknown dose and drinking history. Interpretation is especially precarious when one metabolite is close to a method’s quantification threshold: the ratio can respond sharply to a small change in its denominator. Such effects require actual method and case data rather than assumptions derived from the regression’s R2.
What does the study say about assay detection windows?
Under this particular single-dose protocol, the researchers reported blood ethanol detection over approximately three to eight hours, blood EtG over eight to twelve hours, and blood EtS over five to twelve hours. The report gave EtG and EtS blood limits of detection near 0.02 µg/mL and limits of quantitation near 0.05 µg/mL for the study methods. Detection windows are conditional on the administered amount, the time-sampling grid, specimen, instrument methods, and cutoffs; they are not universal timers for every drinking event.
Detecting EtG or EtS after blood ethanol has declined below a laboratory’s threshold can be consistent with prior exposure. It does not, without an independently validated timing model, disclose a precise end time for consumption. A nondetect also is not categorical proof that a person had not consumed ethanol earlier: testing methods have finite sensitivity and collection occurred at particular times.
Observed measurements, derived values, and unsupported conclusions
| Question | What this study supports | What it does not establish |
|---|---|---|
| Was ethanol directly present in a collected sample? | Headspace-GC can quantify ethanol in a tested blood specimen under the validated procedure. | An earlier or later unmeasured BAC, or clinical impairment at a separate event. |
| Was an ethanol metabolite measured? | LC–MS/MS measured EtG and EtS in serial venous blood after a known dose. | An exact number of drinks or their completion time in an unknown person. |
| Can an EtG/EtS ratio correlate with study time? | Yes: a quadratic curve closely fitted ratios constructed from the group means in this controlled experiment. | Out-of-sample individual accuracy or a case-specific uncertainty interval. |
| Did blood EtG/EtS measure behavior? | No task-performance outcome was tested by the ratio model. | Driving error, motor impairment, cognition, tolerance, collision causation, or legal intoxication. |
| Could another laboratory apply the curve? | The equation is research information for a narrowly defined protocol. | Transferability without matrix-specific assay validation and independent clinical/forensic validation. |
How should counsel review an EtG/EtS-based drinking-time opinion?
First determine which chemical results were directly measured, their units, the biological matrix, collection times, reporting limits, standards, controls, and whether the reported values were above validated quantification limits. Review the underlying laboratory chromatograms, calibration and quality-control records rather than relying on a ratio alone. The site’s blood-alcohol chromatography guide explains the separate question of direct ethanol measurement.
Next ask whether the proposed opinion uses this paper’s group-mean equation or a separately validated individual-level model. Identify the actual reference event: beginning of consumption, end of consumption, peak concentration, specimen collection, or the event in litigation. Require a defensible uncertainty interval applicable to the individual and to the time range being claimed. A scientific expert should state when a method offers contextual evidence but does not answer the proposed legal question.
Finally, distinguish an exposure-timing hypothesis from impairment testimony. Neither elevated EtG nor an EtG/EtS ratio tests divided attention, judgment, reaction time, or driving ability. Proving causation or functional impairment requires independent performance, clinical, toxicological, and case evidence.
What are the main limitations of the 26-person study?
The sample was small and relatively homogeneous in age; all volunteers followed one drinking protocol with food, a weight-based 0.72 g/kg dose, and tightly scheduled venous collection. The reported regression was built from mean values at the sampled intervals. The results did not independently establish reproducibility under repeated drinking patterns, variable exposures, unusual metabolism, unscheduled collection, or the forensic question of exactly when a last drink was finished. This is a research calibration exercise, not a generally validated retrospective timestamping procedure.
The article is also not a postmortem fermentation diagnostic study. A living-volunteer blood metabolite curve cannot determine whether ethanol in an autopsy specimen arose before or after death. Specimen-specific analytical and postmortem questions need their own evidence.
Frequently asked questions
Can a blood EtG/EtS ratio tell the exact time of my last drink?
No. The 2022 authors reported a strong fitted relationship for group-mean ratios after a scheduled single drinking episode. They did not validate an exact individual drinking-time clock for forensic use.
Does R2 = 0.9904 mean a 99% accurate personal estimate?
No. R2 summarizes how well the specified curve fitted the analyzed data. It is not a personal accuracy percentage, uncertainty interval, or independent prospective validation result.
Were blood EtG and EtS directly measured?
Yes. The investigators used LC–MS/MS for metabolite concentrations. The hours inferred from their ratio equation were calculated rather than directly observed analytes.
Is a modeled metabolite peak the same as an observed peak blood sample?
No. Observed peak concentrations occurred at scheduled blood draws. The reported 4.12-hour EtG and 3.02-hour EtS peak times were noncompartmental model summaries.
If BAC is negative but EtG is positive, does that establish when drinking stopped?
No. Such a pattern can support a history of exposure after parent ethanol falls below a detection threshold, but it does not on its own yield a validated clock time.
Can the ratio establish DUI impairment or accident causation?
No. The experiment did not validate the ratio against driving, divided attention, motor skills, cognitive impairment, or the cause of an event.
Does the equation transfer automatically to urine, oral fluid, or postmortem blood?
No. This study fitted ratios measured in serial venous blood from living volunteers. A different specimen matrix, collection context, or postmortem process requires separate validation.
Primary scientific source
Wang L, Zhang W, Wang R, et al. Estimating the time of last drinking from blood ethyl glucuronide and ethyl sulphate concentrations. Scientific Reports. 2022;12:14262. doi:10.1038/s41598-022-18527-8. These study-level findings describe measured metabolites and fitted relationships under an experimental protocol; none should be treated as a validated determination about an unstudied individual.