If you use asthma inhalers and are ever arrested for DUI, you should perhaps think twice about taking a breath test. Most inhalers operate primarily by injecting a mist containing a substantial quantity of alcohol into the lungs. As an example, one of the most commonly used inhalers, Primatene Mist, contains 34 percent alcohol. This alcohol does not pass into the blood stream, but remains in the alveolar lining of the lungs — from where it will be exhaled into the breath machine.

The problem is that "breathalyzers" are designed to assume that the breath sample contains alcohol which has been swallowed and then metabolized by the body before being diffused into the lungs. As I mentioned in an earlier post ("Breathalyzers — and Why They Don’t Work"), they are further designed to assume that there are 2100 units of alcohol in the blood for every unit measured on the breath. So the breathalyzer’s computer mistakenly multiplies the alcohol measured from the asthma inhaler 2100 times. In other words, a very tiny amount of alcohol in the lungs from the inhaler mist can have a very large effect on the machine’s reading.

Just to make things worse, scientists have found that some asthma inhalers can cause high readings on breath machines due, apparently, to the propellent gasses used in the aerosols, in particular, chlorofluorocarbons. See "Using Asthma Inhalers Can Give False Positive Results in Breath Tests", 324 British Medical Journal 756 (March, 2002). As I mentioned in another earlier post ("Why Breathalyzers Don’t Measure Alcohol"), one of the many design defects in breath machines is that they are non-specific — that is, they will falsely report thousands of different chemical compounds as being alcohol.

That’s right: Breathalyzers don’t actually measure alcohol. What they actually detect and measure is any chemical compund that contains the methyl group in its molecular structure. There are thousands of such compounds — including quite a few which can be found on the human breath. And this machine that determines a person’s guilt or innocence will "see" all of those chemicals as alcohol — and report a falsely high "blood-alcohol" concentration (BAC).

Most breath machines used in DUI cases by law enforcement today employ a technology called "infrared spectroscopy". The DUI suspect breathes through a tube connected to the machine and a breath sample is captured in a small "sample chamber" inside the machine. Then beams of infrared energy are shot through the captured breath sample. If there are any compounds containing the methyl group, they will absorb some of this energy; the more of the chemical compound in the breath sample, the more energy is absorbed. The more energy that is absorbed, the less infrared energy that reaches sensors at the other end of the sample chamber. And the less energy that is detected by the sensors, the higher the "blood-alcohol" reading.

Problem: the machine is designed to simply assume that the chemical compound absorbing the energy is alcohol. If a person has any of these other compounds on his breath, called "interferents" by the engineers, he will get a falsely high BAC test result. And if there are two or three such compounds on his breath, the machine will read a cumulative result: it will add them up and falsely report the total as the blood-alcohol level.

So what kinds of compounds may be on a person’s breath that can cause false BAC readings in a DUI case? In one study of eight men, 69 different compounds containing the methyl group were discovered. "Trace Composition of Human Respiratory Gas", 30 Archives of Environmental Health 290. In another study invoviing 28 subjects, researchers found that teh "combined expired air comprises at least 102 various organic compounds of endogenous and exogenous origin". "Characterization of Human Expired Air", 15 Journal of Chromatographic Sciences 240. And Camnadian scientists have discovered over 200 such compounds. "The Diagnostic Potential of Breath Analysis", 21(1) Clinical Chemistry 5.

What are these compounds? Are there any on my breath?

Well, for starters, diabetics with low blood sugar can have high levels of acetone — which is "seen" as alcohol by Breathalyzers. And scientific studies have found that people on diets can have reduced blood-sugar levels, causing acetone hundreds of times higher than found in normal individuals. Frank and Flores, "The Likelihood of Acetone Interference in Breath Alcohol Measurements", 3 Alcohol, Drugs and Driving 1. And there are many other so-called "interferents". See, for example, "Excretion of Low-Molecular Weight Volatile Substances in Human Breath: Focus on Endogenous Ethanol", 9 Journal of Analytical Toxicology 246.

If you are a smoker, your Breathalyzer result is likely to be higher than expected. The compound acetaldehyde — reported by the Breathalyzer as "alcohol" — is produced in the human body as a by-product in metabolizing consumed alcohol, and eventually passes into the lungs and breath. Researchers have discovered that levels of acetaldehyde in the lungs can be 30 times higher in smokers than in non-smokers. Result: higher BAC readings on the machine.

And then there are the industrial compounds: paint, glue, gasoline, thinners, and other compounds contain the methyl group. No, you don’t have to drink the stuff: simply absorbing it through your skin or inhaling the fumes can result in significant levels of the chemical in your body for hours or even days, depending upon the "half-life" of the compound. So if you’ve painted a room or siphoned some gasoline in the last day or two, don’t breath into a Breathalyzer.

Some law enforcement officials say that this is not a problem, claiming that levels of the compound would have to be at toxic levels to raise a breath test result to .08% or higher. These officials are displaying their ignorance of the science involved — specifically, of the partition ratio. This is the ratio of the compound found in the breath to that found in the blood. With ethanol, the ratio is 2100-to-1, which means that, on average, there will be 2100 units of alcohol in the blood for every unit found in the breath.

These officials are using this ratio for all compounds, but every compound has its own ratio. Toluene (found in paint, glue, thinners, cleaning solvents. etc.), for example, has a partition ratio of only 7-to-1; a far greater amount of toluene in the blood will pass into the breath, and so a much smaller amount in the body will have a far greater impact on the breath machine.

Have you ever noticed those "Warning: Microwave in Use" signs in restaurants? It’s for folks who have heart pacemakers: There is a risk that the electromagnetic interference (EMI) from the microwave will interfere with the electronic circuitry in the customer’s pacemaker and cause it to malfunction. This phenomenon, often called radio frequency interference (RFI), can be a recurring problem with any instrument containing electronic circuitry.

Now try to think of some place in your neighborhood that is chock full of electronic gizmos constantly transmitting RFI 24 hours a day. How about a police station? Powerful dispatch radio transmitters, radio transmitters in squad cars in the parking lot, walkie-talkies in every officer’s belt, cell phones, computer cathode ray tubes, microwave relays, electronic door locks, microwave ovens, fluorescent lighting — a veritable jungle of RFI. Now let’s put a Breathalyzer smack in the middle of this police station. An instrument filled with sensitive electronic circuitry that has to analyze tiny amounts of alcohol in breath to an accuracy of one tenth of a percent…..

Just a theory of some DUI defense attorney? Consider a report from the National Bureau of Standards, under contract with the National Highway Traffic Safety Administration to conduct accuracy testing on breath machines (referred to in the report as "Evidential Breath Testing" devices, or "EBTs"):

The Washington D.C. Metropolitan Police Department reported to NHTSA that EBTs were found to display erroneous BAC [blood-alcohol content] readings in the presence of electromagnetic fields from radio transmission…. Representatives of NHTSA and NBS were given a demonstration by police officers who routinely conduct breath testing using an EBT in a mobile van. One police officer operated his handheld radio within 1 foot of the EBT and demonstrated that the electromagnetic field could severely affect the analysis of alcohol samples.

The National Bureau of Standards subsequently conducted the testing and subsequently reported that

These results show that EMI is a potential problem with many of the EBT units currently in use….The states may have to take interim measures to determine the extent of their individual problems with EMI affecting EBTs.

The reaction by the federal government to this report was, perhaps, predictable. Afraid that it would undermine public confidence in law enforcement methods, the government classified the document and then buried it. However, it was later resurrected by a Minneapolis DUI law firm’s Freedom of Information Act lawsuit. Most manufacturers of breath machines today quietly offer an "RFI detector" as an option in their products. Unfortunately, these "detectors" are unreliable and, in any event, are rarely purchased by law enforcement agencies.

Note: all 50 states now make it a crime to drive with a blood-alcohol level of .08% or higher. In most cases, the only evidence of this comes from the breath machine. The breath cannot be re-analyzed. The machine cannot be cross-examined.

If you are facing drunk driving charges, you will have taken (unless you refused) a chemical test for blood alcohol concentration (BAC). In the great majority of cases, the test will be done with a breath machine. When you go to court, you will find that you have been charged with not just one, but with two crimes.

The first is the so-called "per se" offense: driving while having a BAC of .08% or greater. No one cares whether you were intoxicated or not. All of the evidence could prove that without question you were sober: the crime is your chemical composition, not your condition. And what is the sole source of evidence upon which you will be either convicted or acquitted? A machine.

The second charge you are facing is "driving under the influence of alcohol" ("DUI"), or in some states, "driving while intoxicated" ("DWI") or "operating under the influence" ("OUI"). They are basically the same thing. In each case, however, the prosecution can prove you were under the influence of alcohol by offering the results of the same breath test into evidence — and the jury will be instructed that the defendant is rebuttably presumed to be guilty unless he can prove otherwise.

That’s right: a presumption of guilt. Based upon what? Again, a machine. So it all comes down to a machine. Your innocence or guilt depends largely if not entirely upon what a machine says. Maybe we should take a closer look at this "breath machine"….

Sometimes generically referred to as "Breathalyzers" after the original Breathalyzer 900, today there are a number of makes and models manufactured by different companies. For many years, the most popular of these has been the "Intoxilyer 5000", manufactured by CMI, Inc. How reliable is this machine at measuring alcohol in a person’s blood by measuring his breath? How accurate?

Well, what do the manufacturers think? How confidant are they that these devices are reliable enough to send a man to jail? Let’s take a look at their manufacturer’s warranty. The following is from their manual’s "Statement of Warranty":

"CMI, Inc., a subsidiary of MPD, Inc., warrants that each new product will be free from defects in material and workmanship, under normal use and service, for a period of one year from the date of delivery to the first user-purchaser…."

One year? These things are warranted for only one year? Model 5000s are commonly found in service at law enforcement agencies for ten years or more. What if there’s a problem with the machine requiring repair by the manufacturer?

"Repaired components are warranted for a period of 90 days from the date of repair."

90 days? The toaster in my kitchen has a better warranty. But the "warranty" continues:

"There are no other warranties expressed or implied, including but not limited to, any implied warranties of merchantibility or fitness for a particular purpose…."

What? CMI, Inc., says this machine is not warranted for any "particular purpose" — which, for the Intoxilyzer 5000, is measuring alcohol on the breath. So they don’t guarantee that it will measure breath alcohol? And this, the law says, is "proof beyond a reasonable doubt"?

Ok, let’s take a look at another of these machines which determine guilt or innocence: the BAC DataMaster, manufactured by National Patent Analytical Sytems, Inc. Their warranty, at least, is for two years –but with that same refusal to guarantee that the thing measures breath alcohol:

"There are no other warranties expressed or implied including, but not limited to, any implied warranty of merchantability or fitness for a particular purpose."

And, as with CMI, Inc., there is the added warning that "In no event shall National Patent Analytical Systems be liable for any loss of profit or any indirect or consequential damages arising out of any such defect in material or workmanship". In other words, if you end up going to jail because of defects in our machines, you can’t sue us.

The simple fact is that, for perhaps the first time in our history, we are convicting people of crimes — beyond a resonable doubt — based entirely upon what a machine says. Are we that sure of their accuracy? Are the manufacturers?

Bryan is presently facing criminal charges for driving under the influence of alcohol. Except that he wasn’t under the influence of alcohol. He had one drink after work and was stopped at a DUI sobriety checkpoint on the way home. The officer smelled the alcohol on his breath and asked Bryan to step out of the car to take some field sobriety tests. He did fairly well on the tests but, to be sure, the officer asked him to breathe into the breath machine that had been set up at the checkpoint. The results: .09%. Bryan was arrested for DUI, handcuffed and taken to jail; his license was immediately confiscated and he was served with a notice of automatic suspension. When finally released six hours later, he was given a notice to appear in court for arraignment on drunk driving charges.

What happened? How could Bryan have only consumed one beer but registered .09% on the machine — at least four times higher than would be expected?

Well, to begin with, breath machines (commonly referred to as "Breathalyzers", although there are many competing makes and models) are notoriously inaccurate and unreliable. Calibration, maintenance, repair and use by inexperienced or poorly trained officers are always problems. And there are inherent design defects, such as being "non-specific" for alcohol — that is, they don’t actually measure alcohol; due to the nature of infrared analysis, they will report thousands of other compounds as "alcohol". Another recurring problem is "mouth alcohol".

What is "mouth alcohol" — and how could this have caused Bran’s false reading? The machine measures alcohol on the breath, and an internal computer then multiplies the reading 2100 times to get a reading of alcohol in the blood. This is because the amount of alcohol in the blood is greatly reduced as it crosses from the blood into the alveolar sacs of the lungs and into the breath; the average person has 2100 times more alcohol in his blood than in his breath (this varies widely among individuals, however, and is another inherent defect in the machines).

But what if the alcohol in the breath sample did not come from the lungs? What if the alcohol came from Bryan’s mouth or throat? Then it will not have been processed through the body, into the blood and finally out through the lungs — and it will not have been reduced 2100 times. But the machine, being a machine, will always multiply it 2100 times. Result: false high reading and Bryan is facing DUI charges.

So what was alcohol doing in Bryan’s mouth or throat?

Well, alcohol will usually stay in the tissue of the oral cavity or esophagus for about 15 minutes until it is finally diluted and flushed down into the stomach by saliva. So if Bryan had "one for the road" just before being tested, he could have a problem. Or the alcohol could have become trapped in dentures or gum cavities and lasted much longer. Bryan may have burped or belched within 15 minutes before taking the test, sending up alcohol from the beer in his stomach into his mouth and esophagus. But what actually happened was that Bryan suffers from a very common condition: GERD, or "gastroesophageal reflux disease". This causes "acid reflux", often experienced as heartburn.

Acid reflux is commonly caused by a "hiatal hernia" – damage to the pyloric valve separating the stomach from the esophagus. When the valve cannot close completely, then liquids and gasses from the stomach can rise into the throat and oral cavity, to remain there until once again flushed back down. Since a bout of acid reflux can be caused by stress, it is not unusual to find that people stopped by police officers for suspicion of DUI and subjected to field sobriety tests experience the condition.

Bryan is now ordered to breathe into the machine’s mouthpiece. With alcohol from his stomach now rising into and permeating his mouth and throat, it is mixed with the breath passing from the lungs through the throat and mouth and into the machine. Since this alcohol is being multiplied by the machine 2100 times, it takes only a tiny — invisible — amount of absorbed alcohol to cause a disproportionately high reading. In Bryan’s case, an "innocent" reading of perhaps .02% became a "guilty" .09%. And Bryan lost his driver’s license….and now has to try to prove his innocence in court.

Prove his innocence? Aren?t we presumed innocent in America? Here we have the notorious "DUI exception to the Constitution" again. Strangely, Bryan is not presumed to be innocent as we all thought: almost all state laws legally presume a person is under the influence of alcohol if if the machine’s reading is .08% or higher.

Yes, we have a system where citizens are convicted by a machine….A very fallible machine.