EVIDENCE LIBRARY
Can Alanine-Fortified Tomatoes Lower Breath Alcohol After the Same Drinking Dose?
A 15-man crossover trial found lower measured breath alcohol with an alanine-tomato formulation. Learn why this does not prove lower blood BAC or sobriety.
Short answer: In one small, randomized crossover study, 15 healthy men had lower measured breath alcohol concentrations after consuming a specially prepared alanine-fortified tomato product with a substantial alcohol dose and a meal, compared with a calorie-matched sugar control. The difference concerned the breath-alcohol time course in a controlled protocol—not a demonstrated reduction in directly measured blood ethanol, a validated treatment for intoxication, proof of sobriety, or protection against impaired driving.
The underlying research is Oshima, Shiiya, Tokumaru, and Kanda (2016), Nutrition & Metabolism, doi:10.1186/s12986-016-0087-9. Its unusual value in forensic toxicology is that it isolates a specific food formulation within the same participants and documents a change in a measured breath result. Its equally important limitation is that the study cannot be used as a person-specific rule about blood alcohol, the speed of elimination, or functional impairment.
What exactly was tested?
Investigators recruited healthy Japanese men with the ALDH2*1/*1 genotype who could tolerate the planned drinking challenge. The protocol was a single-center, randomized, open-label crossover: each participant took both the tomato formulation and the comparator, in randomized order, separated by four weeks. Of 21 randomized, 20 completed both testing periods; the reported main analyses used 15 participants because four did not finish the alcoholic beverage and one vomited after drinking.
On test days participants received a meal supplying micronutrients and calories, and shochu providing 1.2 grams of ethanol per kilogram of body weight. The active food was 40 g of freeze-dried alanine-fortified tomato material, prepared from tomato paste and alanine and described as equivalent to approximately 400 g raw tomatoes plus 10.4 g alanine. The control was 38 g table sugar; their assigned test-food energy values were similar, approximately 150 versus 152 kcal. This was not a comparison of eating an ordinary tomato versus skipping dinner.
The 15 analyzed participants consumed an average of 73.6 ± 8.4 g ethanol in each experiment. Investigators sampled breath serially and collected venous blood for selected metabolites and other biochemical variables. Participants also scored subjective stimulant and sedative feelings using visual analog scales.
What was measured, and what was not?
| Measure | How the study obtained it | Forensic boundary |
|---|---|---|
| Breath alcohol concentration (BrAC) | AlcoQuant 6020 plus breath analyzer at scheduled times from 18:00 through the following morning | Direct instrument finding for the breath specimen; not an independently assayed whole-blood ethanol concentration |
| Breath alcohol exposure over time | Area under the measured BrAC–time curve, calculated for 18:00–00:00 | Derived summary of repeated breath readings, not a direct measure of total absorbed dose or clearance |
| Blood acetaldehyde, acetate and other markers | Venous specimens with analytical and clinical laboratory methods | Biochemical measurements; neither acetaldehyde nor acetate is interchangeable with a blood ethanol assay |
| Next-morning sedative sensations | Participants’ self-ratings on seven subjective scales | Reported experience, not a standardized driving, balance, reaction-time or other performance test |
| Gastric emptying or individual elimination rate | Not directly measured as validated endpoints | Proposed physiological mechanism and clearance explanation were not separately established in each participant |
The distinction between direct analysis and interpretation also matters when reviewing breath alcohol measurement or using pharmacokinetic assumptions to reconstruct an earlier alcohol concentration.
How much did the breath alcohol results differ?
Compared with the sugar condition, the alanine-fortified tomato condition produced significantly lower BrAC at 20:00, 21:00, 22:00, 23:00 and 00:00. The conditions did not differ significantly at 19:00. At 18:00, before alcohol intake, and at 06:00 the next morning, breath readings were below the device’s detection limit in both conditions. A below-detection-limit result should not be described as proof of zero ethanol molecules in the body.
Across the specified 18:00–00:00 observation period, the reported mean breath-concentration area under the curve was 2.1 ± 0.5 mg/L·h with the tomato preparation and 2.5 ± 0.5 mg/L·h with sugar, a statistically significant group difference. These units describe integrated breath concentration over time; they do not report a venous blood-alcohol AUC.
| Reported finding | Alanine-fortified tomato | Sugar control | Interpretation |
|---|---|---|---|
| Participants in analyzed crossover dataset | 15 men | The same 15 men | Within-person design, small selected cohort |
| Ethanol challenge | 1.2 g/kg | 1.2 g/kg | Assigned dose comparable between periods |
| BrAC AUC, 18:00–00:00 | 2.1 ± 0.5 mg/L·h | 2.5 ± 0.5 mg/L·h | Lower integrated breath result under the tested condition |
| BrAC at 19:00 | No significant condition difference | Effect was not uniform at every measured time | |
| BrAC at 20:00–00:00 | Lower in tomato condition at the five scheduled sampling points | Time-specific group comparisons, not a prediction equation for an individual | |
| Measured blood ethanol concentration | Not reported as a directly assayed endpoint | Not reported as a directly assayed endpoint | Do not relabel breath findings as blood BAC measurements |
In the discussion, the authors also provided approximate peak blood-alcohol estimates derived from breath measurements: 0.10% in the active condition and 0.12% in the comparator. These were not measured venous whole-blood ethanol values. They should not be presented as validation of a fixed blood-to-breath conversion for every person.
Did the researchers show that the food speeds ethanol elimination?
No direct individual elimination-rate experiment was performed. The study demonstrated differences in serial breath alcohol values under two feeding conditions. The authors suggested that the product’s dietary components might delay gastric emptying, change the entry rate of ethanol into circulation, and perhaps influence metabolism. However, they did not directly measure gastric emptying in this human crossover, nor did they separately establish a validated elimination slope unaffected by ongoing absorption.
A lower curve during the absorption and distribution period can result from differences in delivery of ethanol to the circulation. That observation alone cannot establish that the liver cleared ethanol faster, that a particular participant avoided a particular BAC, or that a retrospective calculation should use a special metabolism rate. The separation among absorption, distribution and elimination is explained in pharmacokinetics in forensic toxicology.
Were blood metabolites and other laboratory results consistent with the breath findings?
Blood acetaldehyde concentrations were similar between conditions. Blood acetate was higher in the active condition at one reported sampling time, but not uniformly across all blood draws. Serum alanine was higher as expected with supplementation; urinary pH differed at selected intervals; plasma glucose did not differ significantly between conditions at the reported sampling times. These findings are part of a physiological description, not proof that an individually measured blood-ethanol concentration declined at a known greater rate.
In particular, blood acetate is not blood ethanol. Nor does a change in urine pH directly measure a driver’s motor abilities or a specimen’s ethanol content. Interpretation must stay tied to the analyte and matrix actually measured, as emphasized in the site’s alcohol and volatile analysis overview.
Did fewer participants report hangover-like symptoms?
At 06:00, the sum of seven self-rated sedative symptoms was significantly lower in the active-food condition. Among those seven items, the separate ratings for feeling “down,” “heavy head” and “sluggish” differed significantly. Stimulant ratings did not differ significantly between conditions at the reported time points.
The authors also described six of 20 crossover participants reporting a hangover in the control period and one of 20 in the tomato period. That descriptive yes/no count uses the broader 20-person crossover group, not the 15-person main analytical dataset. It is not equivalent to a validated test of next-day cognition, divided attention, reaction time, vehicle handling or legal fitness to drive.
Knowing that a participant reported fewer symptoms is useful for describing subjective experience. It does not establish objective performance restoration. A person may report feeling better and still perform poorly; conversely a person may feel unwell without a measurable deficit on a particular task. The paper did not test those alternatives.
Why the design prevents a general antidote or DUI claim
The trial was non-blinded, meaning participants knew which food they received; the authors acknowledged they could not construct a credible placebo matching the tomato product. Expectation effects therefore remain plausible for self-reported outcomes. The sample was also small, included only healthy Japanese men selected for a particular ALDH2 genotype and tolerance of a high alcohol dose, and excluded five participants from the principal results.
Participants consumed a special freeze-dried formula with a full meal. The result cannot automatically be extended to ordinary tomato juice, tomatoes eaten separately, supplements marketed as hangover cures, women, patients with liver disease, people with different ALDH2 variants, or different drinking patterns. The intervention was not validated as a clinical reversal treatment or as a way to meet a legal breath-alcohol threshold.
Most importantly, this experiment did not test driving or objective psychomotor impairment. An observed difference in BrAC curves, even if reproducible, cannot identify whether one defendant had a particular degree of functional impairment at a different time and under different circumstances. No study arm establishes a dependable waiting interval after drinking.
How should a forensic expert use this paper?
Its legitimate use is as evidence that an otherwise comparable alcohol dose can produce different measured breath-concentration profiles when the associated food formulation changes. That observation is relevant when someone proposes an oversimplified absorption curve, assumes food is irrelevant, or treats a modeled past BAC as a direct observation. It is not a substitute for case-specific timing, drinking and meal history, instrument records, measured specimens, and any relevant impairment evidence.
Expert review should distinguish: the administered ethanol dose; the actual test food and timing; serial breath measurements and unit conventions; separately assayed blood analytes; uncertainty of breath-to-blood estimates; and functional performance if it was independently observed or tested. No single component supplies all the others.
Frequently asked questions
Did the tomatoes make the alcohol disappear?
No. The study found lower breath-alcohol readings over selected times. It did not establish disappearance of absorbed ethanol or a quantitative increase in individual hepatic elimination.
Were blood alcohol concentrations actually measured?
Venous blood was collected for biochemical measurements, but the reported ethanol exposure comparison was based on serial breath readings. Breath-derived estimates of BAC are not direct whole-blood ethanol measurements.
Can someone drink the same amount and have different breath readings because of food?
This crossover observed a different average BrAC course under one specific food formulation and meal. The magnitude and timing cannot be predicted for an individual from the study alone.
Does lower breath alcohol prove lower impairment?
No. The researchers did not measure objective psychomotor performance, driving, or case-specific functional impairment. Subjective symptoms and measured BrAC address different endpoints.
Does this prove that tomatoes are a hangover treatment?
No. The non-blinded small trial reported lower selected subjective scores, but it did not establish a general treatment effect in broader populations or an objective recovery of performance.
Does the paper justify changing a forensic alcohol-elimination rate?
No. Gastric emptying and individual ethanol clearance were not directly validated as distinct causal mechanisms. An absorption-period concentration difference is not an elimination-rate coefficient.
What is the most defensible legal-scientific takeaway?
Keep the exact specimen, method, time, dose, meal, and inferential step separate. A controlled group finding can support consideration of food-dependent variability, but not a definitive statement about a person’s historical BAC, impairment, or causation.
Primary scientific source
Oshima S, Shiiya S, Tokumaru Y, Kanda T. Alanine-fortified tomatoes relieve the acute alcohol-induced adverse effects in healthy men: a randomized cross-over study. Nutrition & Metabolism. 2016;13:25. https://doi.org/10.1186/s12986-016-0087-9.
This educational research discussion describes one limited human crossover experiment. It is not clinical treatment advice or a case-specific impairment opinion.