The Problem of Mouth Alcohol in Breath Alcohol Testing
Undetected mouth alcohol can falsely elevate a breath test.
In breath alcohol testing, the term “mouth alcohol” is the name for alcohol that falsely elevates the results of a person’s breath alcohol reading.1 It can come from two sources: the unabsorbed alcohol in a person’s stomach may reflux into the oral cavity or from putting something in the mouth that contains alcohol.2
This type of breath sample contamination has been known since breath alcohol testing began. In 1867, Dr. Francis Anstie observed:
"Much caution is necessary, however, in applying this test. It must not be tried during at least the first quarter of an hour after a dose has been taken, for the mouth retains the characteristic smell, even of the most moderate dose, of any of the stronger smelling drinks, for fully this time."3
In 1927, scientist Emil Bogen observed that even small factors, such as a hiccup, could cause problems in breath alcohol testing. He made the following observation in the Journal of the American Medical Association:
"As soon as the disturbing factor of alcoholic liquor still in the mouth is removed, which occurs usually within fifteen minutes after imbibition, in the absence of hiccupping or belching, the alcohol content of 2 liters of expired air is a little greater than I cc of urine.”4
One quality assurance mechanism required to ensure a subject delivers a fair, unbiased sample to the breath analyzer is a 15-minute pre-test observation period. Other controls include duplicate sampling and slope monitoring.5
The Observation Period
During the 15 minutes preceding a breath alcohol test, the subject must be closely monitored to ensure that nothing is put into the mouth that could affect the outcome of the test results.6 Before the start of the observation period, the mouth should be inspected to ensure there is nothing in the mouth that could bias the results.
This includes anything external to the body, such as food, drink, or alcohol coming from the stomach in the form of a burp, belch, hiccup, or vomit. This is of particular concern in a forensic situation where the loss of liberty is at stake.
Drinking alcohol makes the possibility of eruptions from the stomach into the mouth more likely due to the relaxation of the esophageal sphincter. One study showed that unabsorbed alcohol stays in the stomach for several hours and can cause reflux in otherwise healthy individuals.7 Another study showed that in subjects with GERD, alcohol may remain unabsorbed in the stomach for up to 5 hours.8
Many foods and drinks contain small amounts of alcohol due to flavoring agents or fermentation.9 These amounts are so small that they will go unnoticed in daily life. But during a breath alcohol test, even these small amounts of alcohol can become significant when they are in the mouth just before the test.
In my experience, things you might not think of as containing alcohol, like a ripe banana when eaten just before a breath test, will give a positive reading on a breathalyzer due to the small amounts of alcohol produced from fermentation.
Even a warm cup of water can elevate the reading of a breath alcohol test. This is because warming the mucous membranes in the mouth will increase the amount of alcohol transferred from the liquid phase to the gas phase, according to Henry’s law.10
Duplicate Breath Sampling
Another quality assurance mechanism in breath alcohol testing is taking multiple measurements. Typically, two breath samples must meet predetermined agreement criteria to be accepted by the instrument. The most common rule for breath sample agreement in the US is that the samples must agree within 0.02 g/210L of each other.
The idea behind this quality control is that it is unlikely that the subject will reflux or place something in the mouth that would elevate both samples by the same amount. However, the scientific literature disagrees about whether the 0.02 agreement criteria is narrow enough to detect the presence of a biased breath sample due to mouth alcohol.11
Some scientists, such as A.W. Jones, have stated that mouth alcohol contamination can still be present even if the two samples even if they fall within the 0.02 g/210L criteria. The State of Washington, for example, requires breath samples to meet a 10% agreement criteria for unbiased acceptance.
In my experience conducting live drinking sessions with breath test operators, I have seen situations where subjects intentionally introduced alcohol into their mouths before a breath test, and the result did not cause a 0.02 disagreement. This lends credence to the idea that a 0.02 g/210L agreement criteria may not be a rigorous enough standard to catch all instances of mouth alcohol contamination.
Slope Monitoring
A final method of ensuring a breath sample is free of mouth alcohol is monitoring the alcohol slope as the subject provides a sample into the instrument. As the subject blows into the instrument, the slope of the alcohol curve is analyzed and compared to a predetermined set of rules.
If the profile exhibits abnormalities such as waviness or a sharp decrease in concentration, the instrument should flag the sample for mouth alcohol. However, each instrument has different algorithms, and the algorithms may even differ for the same model instrument from one jurisdiction to the next.
In my experience in conducting breath alcohol testing courses, I have noticed that the algorithm does not always flag for mouth alcohol, even when the subject has recently swished alcohol into their mouth. I have seen this both in subjects with preexisting alcohol in their system and those with zero alcohol in their system. This is consistent with the findings of other investigators.121314
When teaching about mouth alcohol, John Fusco, the president of NPAS, one of the major manufacturers of breath alcohol analyzers, said,
If you do it right, you can actually take a person who is 0.05, 0.06, 0.07, and pull that reading up to 0. 12, 0.20… The DataMaster will pick up 85-90% of the mouth alcohols. It does a pretty good job compared to everybody else. Some of the instruments that are out there are going to do as little as 50%.
Conclusion
Mouth alcohol is a persistent problem in breath alcohol testing. Considerable care must be taken to ensure fair and accurate results.
References
Caddy GR, Sobell MB, Sobell LC. Alcohol breath tests: Criterion times for avoiding contamination by “mouth alcohol”. Behavior Research Methods & Instrumentation 1978; 10: 814–818.
Jones AW. Reflections on the GERD defense. DWI Journal of Law & Science 2005; 20: 3–8.
Anstie FE. Prognosis and treatment of certain acute diseases. Lancet. 1867 Sep 28;90(2300):385–7.
Bogen E. Drunkenness—a quantitative study of acute alcoholic intoxication. JAMA. 1927;(89):1508–11.
Dubowski KM. Quality assurance in breath-alcohol analysis. J Anal Toxicol 1994; 18: 306–311.
Dubowski K. Breath as a Specimen of Analysis for Ethanol and Other Low-Molecular-Weight Alcohols. In: Garriott J (ed) Medical-legal Aspects of Alcohol. Tuscon, Arizona: Lawyers & Judges Publishing Company, 2003, pp. 177–183.
Kaufman SE, Kaye MD. Induction of gastro-oesophageal reflux by alcohol. Gut. 1978 Apr;19(4):336–8.
Booker JL, Renfroe K. The Effects of Gastroesophageal Reflux Disease on Forensic Breath Alcohol Testing. J Forensic Sci. 2015 Nov;60(6):1516–22.
Logan BK, Distefano S. Ethanol Content of Various Foods and Soft Drinks and their Potential for Interference with a Breath-Alcohol Test. Journal of Analytical Toxicology 1998; 22: 181–183.
Henry William, Banks Joseph. Experiments on the quantity of gases absorbed by water, at different temperatures, and under different pressures. Philosophical Transactions of the Royal Society of London. 1803 Jan 1;93:29–274.
Gullberg RG. Determining an Appropriate Standard for Duplicate Breath Test Agreement. Canadian Society of Forensic Science Journal. 2006 Jan 1;39(1):15–24.
Simpson CD, Kerby JA, E KS. Effects of Mouth Alcohol on Breath Alcohol Results. International Journal of Drug Testing. 3:1–14.
Gullberg RG. The inadequacy of instrumental “mouth alcohol” detection systems in forensic breath measurement. Crime Scene. 2001;27(1):9.
Hlastala MP. The slope detector does not always detect the presence of mouth alcohol. The Champion 2006; 3: 57–61.