Rates for alcohol absorption and burn-off are an appropriate matter of inquiry and exposition in a trial for driving while intoxicated. However, using such rates to establish that the defendant was driving while intoxicated may be a problem because Texas law is stated in terms of concentration of alcohol in the blood (BAC) at time of driving rather than time of testing. Extrapolation back from the BAC at the time of testing to the BAC at the time of driving is an endeavor fraught with the danger of inappropriately bamboozling the jury into thinking that such an extrapolation can be anything close to accurate. The information which is usually available to the expert doing the extrapolation is woefully inadequate to make even an educated guess about a range of possible BACs.
According to Nichols and Whited,
The intoxication of an individual depends on the absorption, distribution, and elimination of alcohol through the body. The rate of absorption, distribution and elimination varies greatly between individuals, and can have a substantial effect on an individual's intoxication and chemical test results. Despite the substantial variability of these factors, the chemical tests used to determine the intoxication of an individual are based on the "average" person under constant conditions. The "average" person is created using a number of assumptions that do not take into account the substantial individual variation.
Donald H. Nichols & Flem K. Whited, III, Drinking/Driving Litigation: Civil and Criminal 14-2 (2d ed. 1998). (DDL). Add to this the reality that elimination begins as soon as the first alcohol reaches the liver. Edward F. Fitzgerald, Intoxication Test Evidence 2-14 (2d ed. 2000). (ITE). Thus absorption and elimination occur simultaneously until all of the ingested alcohol has been absorbed. Texas Breath Alcohol Testing Program Operator Manual 5-10 (1996). (OM).
The original work in the area of absorption and burn-off rates was done in the 1930s by E.M.P. Widmark. Unfortunately, his work has been subjected to "gross simplification." ITE at 3-8. All of Widmark's work was done on subjects who drank a metered amount of alcohol on an empty stomach. Id. at 3-2 et seq. In the original test group of 30 subjects, only 3 (10%) tested in the absorption range of 0.65 and 0.70, 19 (63%) tested under 0.65 and 8 (27%) tested over 0.65. The subjects' elimination rates varied from 0.006 to 0.04 with most being widely distributed between 0.010 and 0.020. Widmark and later researchers established that both absorption and elimination rates vary widely and are dependent on a large number of factors, including gender. Today, however, "experts" take the averages rate for males and apply them to everyone. They then testify authoritatively that "the average person" absorbs and eliminates alcohol at fixed, known rates, generally claimed to be 0.68 for absorption and .015 for elimination, and make no mention of the myriad variables known to have marked effect on those rates.
Absorption and burn-off rates are highly variable, even in a single individual. The generally accepted rate of burn-off is about one beer per hour, based on the "average man." However, the "average man," like the "average family" with 2.4 children, doesn't exist; the only rates which have relevance are the rates of the person on trial. Absorption and elimination rates are affected by a myriad of factors, including most commonly, weight, gender, physical condition, metabolic rate, time of day, when, what, and how much the defendant last ate, when, what, and how much and how the defendant drank, medications taken or not taken, point in a woman's menstrual cycle, emotional state, and whether the defendant is an alcoholic. We can extrapolate in regard to which way each known factor, in isolation, will affect absorption and burn-off, but prediction becomes more complicated each time another factor is added to the equation. Very quickly, accurate extrapolation becomes impossible. The impossibility is highlighted when the values of only a few of these many variables are known to the extrapolator. See generally, ITE, Chapter 11.
The first inquiry must be into where on the absorption/elimination curve the BAC sample falls. Id. at 2-1. Given the procedures mandated by police policy, that inquiry can almost never be accurately answered; determination by law enforcement of BAC via breath or blood is usually the result of a single sample. At least two, and preferably three, samples taken over time are absolutely necessary. Id. at 4-11 - 19. After drinking begins, the BAC will rise to some peak value, then decrease until all alcohol has been metabolized. If, at the time of testing, a person is in the absorption phase, in which the BAC is increasing, the BAC at time of driving would be lower than the tested value. The opposite is true if at the time of testing the person is in the elimination phase, in which the BAC is decreasing. Without knowing even the sample's general location on the BAC curve, all other calculations are specious sophistry. A single high reading can be useful as an indication of intoxication, e.g. when the BAC is 0.20 one hour after driving, but even then is not helpful in determining what the actual BAC was at the time of driving. Id. at 4-15 -16. Even if the position on the BAC curve at the time of testing can be determined, the task of determining BAC at the time of driving gets no easier.
According to the Texas Breath Alcohol Testing Program Operator Manual (1996), the manual for breath-test operators published by the Texas Department of Public Safety (OM), a "200-lb. man must consume twice as much alcohol as a 100-lb. man to attain the same alcohol concentration." Id. at 5-7. Even that statement is a generalization; alcohol is distributed through the body dissolved in the water contained in the body. OM at 5-5. For the above statement to be true, we must assume that the two men have identical levels of tissue water, for instance, 70%. The equation changes again if the individual is a woman; women tend to have less water and more body fat than men. Id. at 5-8. A woman is therefore likely to have a higher blood-alcohol level than a man of equal weight after consuming the same quantity of alcohol. Id. Even that may not be true if the man is obese and the woman has extremely low body fat. ITE at 2-6. Generally, the better the person's physical condition (high muscle, low body fat), the lower the peak BAC will be. Id. The rate at which a person metabolizes alcohol is affected by the health of that person, especially the health of the person's liver. Id. at 2-13. Size, shape, and capacity of the liver affect the rate of elimination. Id. at 4-40. At least seventeen shapes are considered "normal," id., and liver size is related to body weight, DDL at 14-139. If the person is an active alcoholic, the rate will be even more unpredictable; alcoholics tend to metabolize ethanol more rapidly (DDL at 14-149 - 53, OM at 5-5) and at least partially by a different chemical process than non-alcoholics (ITE at 4-5).
One of the major influences on how quickly and how high BAC rises is whether ingestion of alcohol is accompanied by ingestion of food. As noted above, the original research into BAC was done only on subjects drinking on an empty stomach, yet the rates derived from that research are now applied indiscriminately to all scenarios. The presence of food in the stomach slows the rate of gastric emptying, and that rate is known to have a dramatic effect in how much and how quickly alcohol gets into the small intestine, where the vast majority of alcohol is absorbed. Id. at 2-3. Absorption in the small intestine is quick. Id. at 2-3. DDL at 14-47 et seq. Absorption in the stomach is generally slow and inefficient, so the longer alcohol stays in the stomach, the longer it will take to reach peak BAC. Id. at 2-2. DDL at 14-70. Also, when alcohol stays in the stomach, the digestive process breaks down some of the ethanol into smaller pieces which do not cause intoxication when absorbed by the body, and so the peak BAC is lower than if the alcohol were consumed on an empty stomach. DDL at14-52, 14-70, ITE at 14-115.
A number of factors affect gastric emptying. Protein slows emptying more than sugars, which slow emptying more than carbohydrates, which slow emptying more than fats. DDL at 14-63. Large meals slow emptying more than small ones. Id. at 55. Size, shape, and position of the stomach also affect emptying, and at least nine positions and nineteen shapes are recognized as "normal." ITE at 4-2, 4-39. Because gastric emptying is controlled by the nervous system, emotions affect the rate; fear decreases it, while excitement tends to accelerate it. DDL at 14-70. Trauma may shut down emptying altogether. ITE at 2-10 - 11. The alcohol itself can alter the rate; alcohol ingested in large quantities causes the pyloric sphincter to close, preventing emptying. Id. at 14-37, 14-68 - 69, OM at 5-5. Carbonated beverages mixed with alcohol accelerate emptying. ITE at 14-69, OM at 5-5. Surgery such as stapling also alters the rate of emptying, ITE at 14-73 -74, as does gravity, id. at 14-73, some kinds of drugs, such as Tagamet, id. at 14-73, and some diseases (id. at 14-74). (1)
Which kind of alcoholic beverage was ingested and in what quantity and manner also affect BAC. Distilled liquors produce a higher BAC than beer or wine for a given amount of alcohol. DDL at14-38 et seq. Large quantities increase the time needed to reach peak BAC. Id. at 14-44 et seq. Chugging produces a higher, quicker peak BAC than an equal amount of alcohol consumed over a longer period of time. DDL at 14-88 - 89.
Other factors which may affect peak BAC and rates of absorption and elimination include altitude, DDL at 14-33, point in menstrual cycle, id. at 14-79, 14-159 et seq., oral contraceptives, id. at 14-84, ITE at 4-7, drugs, especially if they act on the stomach or circulation or are metabolized by the liver, ITE at 2-13, body temperature, DDL at 14-139 - 40, and time of day, DDL at 14-85 - 86, 14-164 - 66. Add in physiological differences, such as inherited or developed tolerances or sensitivities, and accurate prediction becomes even more problematic.
To make accurate prediction even more improbable, recent research has shown that two of the basic assumptions used in extrapolating to the time of driving, a smooth BAC curve and linear elimination rates, are not in fact true. The BAC curve is not smooth, as it is usually presumed to be, but rather is irregular and contains unpredictable spikes. DDL at 14-171. Other research has shown that elimination rates are non-linear and vary over time. The importance of such a finding is that "if elimination is nonlinear, then it is impossible to estimate someone's blood alcohol concentration at a time earlier or later than when the blood alcohol measurement is made . . . ." Id. at 14-166.
A simple mathematical formula for the number of possible combinations of variables is 2 to the power equal to the number of variables, e.g., 5 variables indicates 25, producing 32 possible results. Noted above are at least twenty variables. The number of possible results is thus 220, or 1,048,576. However, this simple equation assumes that each variable has only two values and does not consider any possible interactions between variables. The general equation is the number of possible values, "m," raised to the power of the number of variables, "n," or mn. Assuming an average of 5 values per variable and twenty variables, the number of possible results, 520, is 24,414,062. Each variable in the BAC calculation has many possible values. Many of the variables have a possibility of interacting with each other and further complicating the matter. Even assuming a relatively small number of possible values, the number of possible results increases exponentially and quickly becomes mind-boggling.
In most cases, the expert who is attempting to extrapolate BAC to the time of driving has little information about the defendant or the circumstances surrounding the ingestion of alcohol beyond the single BAC reading, the gender of the driver, and perhaps an approximate weight. Even with full information, the complexity of the interaction of the variables makes the accuracy of any claimed BAC value, or range of values, suspect. "For all these reasons the actual BAC curve which will result from the ingestion of a given amount of alcohol on a given occasion by a particular person is, at best, highly unpredictable, although many experts testify as though they can, in fact, predict 'the' BAC which will result on a given occasion." ITE at 2-7 (emphasis in original). The majority chooses not to go so far as to call attempted extrapolation from a single BAC sample back to the time of driving "junk science." I do not feel so constrained, and junk science has no place in a courtroom where the standard of proof is beyond a reasonable doubt.
With these comments, I join the majority.
Date Delivered: June 6, 2001
1. Gastric ulcers and diabetes decrease gastric emptying rates, while duodenal ulcers increase them. DDL at 14-74 - 75.