When AST and ALT Show No Group Difference, Can Other Blood Markers Respond to Alcohol? A 40-Person Trial

Short answer: Yes. In a 2026 randomized trial of acute alcohol consumption, researchers did not find a statistically significant alcohol-versus-water difference for the familiar serum enzymes aspartate aminotransferase (AST) or alanine aminotransferase (ALT), yet reported differences in several other serum biochemical markers and cardiovascular measurements. That contrast is scientifically important, but it is not an individual diagnostic rule. A small exploratory trial cannot establish that a particular person’s recent drinking, organ injury, intoxication level, or impairment can be determined from an isolated laboratory result.

The primary source is Li, Pan, Zhang and colleagues, Dynamic evaluation of blood marker changes induced by acute binge drinking in healthy individuals: a randomized controlled trial, Scientific Reports (2026), doi:10.1038/s41598-026-40028-1. This article evaluates the actual study design, reported numerical comparisons, and limits of forensic inference. It does not propose a new alcohol biomarker or a clinical diagnostic cutoff.

What was the question, and what did the researchers measure?

The investigators asked whether a single controlled episode of heavy drinking would produce time-dependent changes in markers associated with liver function, extracellular matrix remodeling, lipid metabolism, and cardiovascular physiology in otherwise healthy young men. They examined a panel rather than relying on AST and ALT alone. The question was about acute physiological responses in a studied group, not about identifying a concealed drinker or diagnosing alcohol-related disease in an individual.

They collected venous blood before drinking and at 1, 5, 12 and 24 hours after drinking began. Serum was separated and used for biochemical and immunoassay measurements. Separately, an instrument measured exhaled breath alcohol during the early study period. Blood-pressure readings, heart rate, and peripheral oxygen saturation were taken on a different schedule. These are not interchangeable observations: a serum triglyceride result is not a blood-ethanol assay; a breath-alcohol reading is not a direct venous whole-blood BAC; and heart rate is not a laboratory measure of alcohol exposure.

Who participated, and how strong was the comparison?

Of 105 volunteers screened, 45 healthy male volunteers were enrolled and allocated in an 8:1 ratio: 40 were assigned alcohol and five were assigned an equal-volume water comparator. Five participants allocated to alcohol withdrew after vomiting. Thus the reported completed intervention sample comprised 35 alcohol participants and five controls, a total of 40. The mean age of the enrolled volunteers was approximately 21.9 years. Participants were selected for prior experience of heavy drinking or hangover and had restrictions concerning relevant illnesses, medications, smoking and recent alcohol use.

After an overnight fast, participants ate a standardized breakfast. Approximately 30 minutes later, the alcohol group consumed 40% vodka in three portions during a short interval; the intervention dose varied by BMI category. The other group received water. The researchers described masking of staff and analysts, but alcohol and water are perceptibly different, so participants were not reliably blinded to what they consumed. The unusual 35-to-five completed-sample imbalance matters greatly. Five controls provide limited precision for estimating the natural time-related drift of multiple biomarkers and make a single unusual control observation potentially influential.

The trial is consequently better read as preliminary evidence of a measurable group-level response under a specific protocol than as a calibration study for an individual forensic or clinical determination. Vomiting-related dropout creates another selection issue: the final alcohol group excludes five volunteers who did not tolerate the intervention to completion.

What happened to AST and ALT—and what does “no difference” actually mean?

The authors reported no detected difference between study groups for AST or ALT across their reported analyses. That finding does not assert that every volunteer’s AST and ALT were within a reference interval, that neither enzyme changed at any time, or that alcohol can never affect either marker. A nonsignificant statistical comparison means this particular study did not demonstrate a group contrast according to its test and data.

These enzymes are familiar in discussions of hepatocellular injury, but the study’s primary objective concerned a wider set of markers. Their lack of a detected difference coexisted with other reported group contrasts. That combination discourages two opposite errors: declaring that a recent large alcohol dose must elevate AST/ALT, and declaring that a nondiagnostic AST/ALT finding proves no physiological alcohol-related response occurred. Neither proposition follows from this study.

An individual interpretation would additionally require that person’s result, time course, reference range, medical history, possible muscle or liver conditions, drug exposures, specimen identity, and assay performance. The controlled group’s null result supplies none of those case-specific particulars.

Which serum liver-related measurements differed?

Li and colleagues reported higher α-fucosidase (AFU; p = 0.010), albumin (ALB; p = 0.018), alkaline phosphatase (ALP; p = 0.035), and total protein (TP; p = 0.037) in alcohol-versus-control comparisons. Reported post-hoc differences included AFU and TP at five and 12 hours and ALP at 12 hours. Total bile acids (TBA) were lower in the alcohol group in a reported comparison (p = 0.026), notably at five hours (p = 0.038). The text also discusses LDH and GGT at particular sampling times, although the reported time-by-group interactions for those markers did not meet the conventional significance criterion.

These are different analytes serving different clinical purposes. Serum albumin and total protein are influenced by fluid status and protein distribution as well as disease processes. AFU and ALP are not direct ethanol measurements. TBA reflects aspects of bile-acid physiology, not a unique fingerprint of drinking. A shift in a group mean or a model coefficient is not, by itself, evidence of hepatocyte necrosis, permanent organ damage, an episode’s exact timing, or the identity of the cause in a particular patient.

Statistical significance should not be translated into a universal reference interval or a diagnostic threshold. The trial reported many outcomes and several time-specific comparisons; a p-value smaller than 0.05 does not show how accurately a marker separates exposed and unexposed people or how frequently the same pattern occurs in ordinary forensic submissions.

What did the collagen and other “fibrosis-related” tests show?

The investigators also measured serum collagen type IV (IV-C), procollagen type III N-terminal peptide (PIIINP), laminin, cholyglycine acid, and hyaluronic acid. Reported group-by-time comparisons were statistically significant for IV-C (p = 0.041) and PIIINP (p = 0.011). A post-hoc IV-C comparison at one hour was p = 0.007, whereas the reported 12-hour comparison was p = 0.052—above a conventional 0.05 threshold, not a positive result under that rule. The IV-C 24-hour area-under-the-curve contrast was p = 0.044.

Cholyglycine acid was higher at five hours in a reported comparison (p = 0.005); hyaluronic acid was lower at five hours (p = 0.045); laminin did not exhibit a detected difference. This mixed pattern matters more than any one headline. The labels “fibrosis-related” or “extracellular matrix-related” describe what biomarkers are studied in biomedical research. They do not mean that actual liver fibrosis formed during a single 24-hour experiment. Fibrosis is a tissue-level pathological process requiring appropriate evidence; fluctuating serum markers alone do not show histological scarring.

The original investigators treated the responses as preliminary and noted substantial variation for IV-C. The clinical and forensic utility of applying these measurements to individual recent-alcohol disputes was not tested.

Which lipid measurements changed, and which did not?

Serum triglycerides (TG) showed a reported group contrast (p = 0.047), with post-hoc comparisons at five hours (p = 0.004) and 12 hours (p = 0.035). Apolipoprotein A1 (APOA1) had a reported group contrast (p = 0.012), including time-specific comparisons at five hours (p = 0.026) and 12 hours (p = 0.027). The APOA1-to-APOB ratio also differed (p = 0.036), with reported differences at one, five and 12 hours. The reported 24-hour area-under-the-curve analyses were higher for TG (p = 0.044), APOA1 (p = 0.009) and the APOA1/APOB ratio (p = 0.042).

By contrast, the investigators did not report detectable alcohol-versus-control differences for total cholesterol, HDL, LDL, APOB or lipoprotein(a) in their comparisons; they also reported no between-group difference in blood glucose. The mix of detected and undetected results is an essential limitation, not a footnote.

Triglycerides and apolipoproteins are clinically familiar metabolic measures. But they are not alcohol-specific intake markers such as ethanol or certain ethanol metabolites, and they cannot establish a dose, minute of consumption, or legal blood-alcohol concentration from one measured value. Diet, timing, underlying metabolism, inter-individual variation and laboratory method all matter. This trial’s group result does not provide the likelihood ratio, sensitivity or specificity required for any individual forensic classification.

How did cardiovascular measurements behave?

The study reported a heart-rate difference after the alcohol intervention (p = 0.010), with a three-hour time-specific comparison of p = 0.021. Peripheral oxygen saturation was lower at one hour in a reported comparison (p = 0.041). Systolic and diastolic blood pressure remained higher at 24 hours relative to controls in reported comparisons (p = 0.002 and p = 0.007, respectively). A serum CK-MB time-point comparison at one hour was p = 0.027; total CK and hydroxybutyrate dehydrogenase did not show reported differences.

Those measures are distinct in kind. Blood pressure and pulse are vital signs; fingertip saturation is a physiological estimate; CK-MB is a serum enzyme fraction. None is a measurement of alcohol concentration or direct proof of myocardial infarction. Neither short-lived CK-MB differences nor blood-pressure group contrasts determine whether a particular person’s cardiovascular symptoms were caused by drinking. The trial did not assess clinical cardiac events as forensic endpoints.

Evidence table: separate detected contrasts from absent contrasts

Results have been abbreviated to preserve the distinction between differing model terms, time-specific tests and 24-hour cumulative measurements. Individual estimates, their uncertainty intervals, comparator composition and complete original figures should be examined before attempting any secondary quantitative interpretation. Reporting an observed p-value does not establish a clinically useful effect size.

Evidence table: what each measurement can and cannot answer

Why isn’t a serum biomarker a blood-alcohol concentration?

A sample can be collected from a vein without the test being a measurement of ethanol in forensic whole blood. In this experiment, venous blood was processed to obtain serum, and that serum was analyzed for proteins, enzymes, lipids and related biomarkers. The investigators separately monitored breath alcohol. A conventional forensic whole-blood ethanol determination would require a separately validated method for that analyte and matrix; the existence of venipuncture alone does not supply one. The distinction is central to specimen collection in forensic toxicology and to whole-blood ethanol measurement by headspace GC-FID.

There is no reported equation in this trial converting albumin, TG, IV-C, ALT or CK-MB into a forensic BAC. A physiological marker could change after drinking while remaining insensitive or nonspecific as an individual forensic alcohol test. In particular, a biomarker-associated p-value and an analytical measurement with traceable calibration, quantification uncertainty and a defined analyte are fundamentally different kinds of evidence.

What limitations control any courtroom use?

Very small comparator group. Only five men were analyzed as water controls; imprecision or an outlier can materially influence group contrasts. Restricted population. The study involved selected young men without major medical conditions, not women, older people, patients with liver disease or a forensic case population. Unequal attrition. Five alcohol-group participants withdrew following vomiting, raising selection concerns. Many endpoints. Numerous serum measures and time-specific analyses increase the opportunity for isolated nominally small p-values; the text does not establish a comprehensive family-wise error-controlled discovery set for all of them.

Short observation window. Measurements spanned 24 hours; they do not establish long-term fibrosis, chronic heart disease, a persistent lipid disorder or another lasting condition. Partial masking. Water and vodka are readily distinguishable, even if other staff remained masked. Protocol effects. A standardized meal, BMI-stratified alcohol dose, selected volunteers, and supervised laboratory conditions may not resemble a disputed real-world event. Method-specific evidence. Forensic use demands the actual specimen, specimen identity, analytical protocol, contemporaneous quality controls, values, and uncertainty—not simply a population trial.

Statistical nonsignificance, clinical reference-range status and forensic exculpation are three different assertions. Similarly, a statistically significant difference in a non-specific physiological variable is neither a valid measurement of alcohol dose nor an attribution of illness or behavior. See also the site’s pharmacokinetics interpretation framework for the gap between time-course observations and a particular person’s earlier state.

Frequently asked questions

1. Does a normal AST or ALT test rule out recent drinking?

No. AST and ALT are not specific recent-drinking tests. In this experiment, the authors did not detect a group difference for these enzymes despite observed differences in other variables. The study did not establish the sensitivity of AST/ALT to an individual drinking event.

2. Did the researchers prove alcohol caused liver fibrosis within 24 hours?

No. They measured serum markers associated with extracellular-matrix biology. No tissue-scarring outcome was demonstrated by the reported biomarker contrasts, and “fibrosis-related” is not interchangeable with a diagnosis of fibrosis.

3. Could a triglyceride result be used to calculate someone’s BAC?

No. The paper reported group lipid responses, not a validated triglyceride-to-ethanol concentration equation. TG is nonspecific and affected by other biological conditions.

4. Was alcohol directly quantified in a forensic whole-blood specimen?

No independent forensic whole-blood ethanol assay was presented as the basis of these biomarker findings. Blood was drawn and separated into serum for chemistry tests; breath alcohol was measured separately.

5. Does a small p-value tell us whether a marker identifies a drinker?

No. A group-comparison p-value is not sensitivity, specificity, a likelihood ratio or a validated decision threshold. Assessing an individual’s recent exposure requires a different diagnostic validation design.

6. Were all 45 randomized people included in the completed intervention analysis?

No. Forty were assigned alcohol and five water. Five alcohol participants withdrew after vomiting, leaving 35 alcohol participants and five water controls in the reported completed sample.

7. Can this study show that alcohol caused impairment or a later medical event?

No. It did not measure driving, psychomotor performance, traffic crashes, pathology-confirmed organ damage or person-specific causation. Those questions require other evidence and appropriate methods.

Bottom line for scientific and forensic interpretation

The most defensible conclusion is narrow but valuable: in this selected group of healthy young men, a single controlled alcohol intervention was associated with changes in several serum analytes and vital signs even though conventional AST and ALT comparisons did not show a detected group difference. The reverse inference is not justified. One unremarkable routine enzyme test neither proves abstinence nor excludes every acute physiological response, and one altered nonspecific biomarker does not prove drinking, BAC, impairment, lasting disease or causation.

For an actual case, first identify the physical specimen and exact analyte, inspect instrument validation and quality-control records, place each result on its own measurement timeline, and determine which proposition it can support without leaping from a group study to an individual verdict. The primary research is Li et al. (2026), DOI 10.1038/s41598-026-40028-1; the broader forensic toxicology evidence library organizes related principles by specimen, method, interpretation and limitation.