Journal of Analytical Toxicology, Vol. 34, June 2010
The Relationship of Normal Body Temperature,
End-Expired Breath Temperature, and BAC/BrAC Ratio
in 98 Physically Fit Human Test Subjects*
J. Mack Cowan1,†, James M. Burris2, James R. Hughes3, and Margaret Parker Cunningham4
1Breath
Alcohol Laboratory, Texas Department of Public Safety, Austin, Texas 78752; 2Crime Laboratory, Texas Department of
Public Safety, Austin, Texas 78752; 3Breath Alcohol Laboratory, Texas Department of Public Safety, Garland, Texas 75043; and
4Breath Alcohol Laboratory, Texas Department of Public Safety, Bryan, Texas 77803
Abstract
The relationship between normal body temperature, end-expired
breath temperature, and blood alcohol concentration (BAC)/breath
alcohol concentration (BrAC) ratio was studied in 98 subjects (84
men, 14 women). Subjects consumed alcohol sufficient to produce
a BrAC of at least 0.06 g/210 L 45–75 min after drinking. Breath
samples were analyzed using an Intoxilyzer 8000 specially
equipped to measure breath temperature. Venous blood samples
and body temperatures were then taken. The mean body
temperature of the men (36.6°C) was lower than the women
(37.0°C); however, their mean breath temperatures were virtually
identical (men: 34.5°C; women: 34.6°C). The BAC exceeded the
BrAC for every subject. BAC/BrAC ratios were calculated from the
BAC and BrAC analytical results. There was no difference in the
BAC/BrAC ratios for men (1:2379) and women (1:2385). The
correlation between BAC and BrAC was high (r = 0.938,
p < 0.0001), whereas the correlations between body temperature
and end-expired breath temperature, body temperature and
BAC/BrAC ratio, and breath temperature and BAC/BrAC ratio were
much lower. Neither normal body temperature nor end-expired
breath temperature was strongly associated with BAC/BrAC ratio.
Introduction
In vitro studies have confirmed that ethanol1 in blood complies with Henry’s Law (1). At a given temperature, a fixed relationship exists between the blood alcohol concentrations
(BAC) and the air that is in equilibrium with the blood for
each individual (2). It is an oversimplification to state that
breath alcohol strictly complies with Henry’s Law, as some
* The data in this paper were presented at the International Association for Chemical Testing
meeting in Little Rock, AR, April 2004.
† Author to whom correspondence should be addressed. E-mail: [email protected].
1 Ethanol and alcohol are used interchangeably in this paper. Blood analytical results are reported as
g/100 mL; breath analytical results are reported as g/210 L.
238
early researchers asserted, because breath alcohol analysis is a
dynamic process involving diffusion across the alveolar membrane as well as gas exchange in the upper-respiratory tract.
However, it is clear that temperature is a factor in breath alcohol testing (3). When blood containing alcohol enters the
lungs, normal gas exchange occurs, and a small portion of the
alcohol in the blood diffuses into the breath and is exhaled. The
concentration of alcohol exhaled is proportional to the concentration of alcohol in the blood in the lungs. The temperature of the blood in the lungs affects the concentration of alcohol in the alveolar air, but temperature is not the sole
determinant of the alcohol concentration. Theoretically, hyperthermia should increase the partial pressure of alcohol in
the vapor phase from the blood, resulting in an increase in the
breath alcohol concentration (BrAC), whereas hypothermia
should lower the partial pressure of alcohol in the vapor phase
from the blood, resulting in a decrease in the BrAC (4). In
vitro blood experiments have shown that for every 1.0°C
change in the temperature of the blood, a corresponding
change of approximately 6.5% occurs in the concentration of
alcohol in the vapor in equilibrium with the blood (1,5).
Fox and Hayward (6,7) reported that human test subjects
who have had their body temperature artificially raised have
shown an increase in the concentration of ethanol in their exhaled breath, and human test subjects who have had their
body temperature artificially lowered have shown a decrease in
the concentration of ethanol in their exhaled breath. However, little research has been conducted to determine the effect
that differences in normal human body temperature have on
BrAC. Normal body temperature has been reported to span
from 36.1 to 37.8°C (8).
This study was undertaken to examine the relationships between normal body temperature and end-expired breath temperature; venous BAC/end-expired BrAC (BAC/BrAC) ratio;
breath temperature and BAC/BrAC ratio; and body temperature
and BAC/BrAC ratio.
Reproduction (photocopying) of editorial content of this journal is prohibited without publisher’s permission.
Journal of Analytical Toxicology, Vol. 34, June 2010
Methods
Ninety-eight physically fit volunteer subjects (84 men and 14
women) gave informed consent to participate in this study. The
subjects were fed one sandwich approximately 1.5 h prior to the
administration of three equal portions of whiskey (50.5%
ethanol by volume) mixed with a carbonated beverage at 15min intervals to produce a peak BrAC of at least 0.06 g/210 L.
After a 15-min deprivation period, each subject’s BrAC was
determined through the use of an Intoxilyzer 5000 (CMI,
Owensboro, KY) until it was established that the subjects were
in the post-peak phase of alcohol metabolism, 45–75 min after
completion of drinking. These preliminary BrAC results were
not reported.
Each subject was given instructions on how to provide an adequate breath sample. The subjects were instructed to take a
breath and immediately blow into the instrument as long as
possible while maintaining a steady flow of breath. Each subject then delivered a single breath sample into an Intoxilyzer
8000 (CMI) specially equipped and calibrated at the factory to
measure the temperature of the breath sample. The Intoxilyzer
8000 is designed to analyze and report the alcohol concentration of the end-expired breath sample in g/210 L. Prior to subject testing, the alcohol calibration of the Intoxilyzer 8000 was
verified with aqueous ethanol standards used in conjunction
with a Draeger model Mark IIA Alcohol Breath Simulator (National Draeger, Durango, CO). An aqueous ethanol stock solution was prepared by adding 78 mL of absolute ethanol (AAPER
Chemical, Shelbyville, KY) to a 1-L volumetric flask and diluting to volume at room temperature with deionized water
(9). To prepare a solution that would yield a vapor concentration of 0.08 g/210 L at 34°C, 8 mL of ethanol stock solution was
diluted to 500 mL with deionized water at room temperature
and then placed into a Mark IIA Simulator. All pipettes and
glassware were Class A. A custom-made, 78-mL Class A pipette
was used to measure the absolute ethanol (Fisher Scientific,
Pittsburgh, PA).
A single venous blood sample was taken from each subject
immediately after the breath test. The blood was drawn from
the cubital vein and placed into a 6-mL gray top Vacutainer
Plus tube (Beckton, Dickinson and Co., Franklin Lakes, NJ)
Table I. Age and Weight*
Age (years)†
Mean
SD
Span
Weight (kg)‡
Mean
SD
Span
Men
Women
All Subjects
28.7
5.49
21–51
31.1
6.92
23–44
29.1
5.74
21–51
91.2
14.25
58–126
* n = 98; 84 males and 14 females.
† p = 1.455; t = 1.4674.
‡ p < 0.0001; t = 6.1933.
66.9
9.00
54–86
containing 15 mg sodium fluoride and 12 mg potassium oxalate. Prior to the blood draw, the site was swabbed with an antiseptic towelette containing benzalkonium chloride (Professional Disposables, Orangeburg, NY). The blood tubes were
refrigerated at the end of each day of testing.
The whole blood samples were analyzed by headspace gas
chromatography (GC) using a PerkinElmer (Norwalk, CT) HS
40XL automatic headspace sampler with 40 position autosampler and the Autosystem XL GC equipped with Restek
Rtx-BAC-1 and Rtx-BAC-2 (30 m × 0.32-mm i.d.) chromatographic columns with two flame ionization detectors. The GC
was calibrated, and response factors were calculated using four
standard runs of a 0.08 g/100 mL NIST traceable solution (Cerilliant, Round Rock, TX) with internal standard. Samples were
diluted with a 0.10 g/100 mL internal standard solution of
n-propanol and deionized water and with a saponin solution
containing sodium azide and sodium fluoride to eliminate the
matrix effects. Samples were divided after injection and delivered onto the two separate GC columns, thus enabling two independent analyses of the same sample (PerkinElmer). The
mean of the four BAC results (two vials per sample, two analyses per vial) was the reported value.
Immediately after the blood was drawn, the body temperature of each subject was measured by three different thermometers: oral, tympanic, and temporal. All testing was conducted in a single room that was thermostatically controlled at
about 22°C. The three thermometers, operated per the manufacturers’ instructions, used to determine the body temperature of the subjects were a BD Basal Digital oral thermometer
(Becton, Dickinson), a Braun ThermoScan IRT 3520 Type 6013
Table II. BrAC, BAC, the BAC-BrAC Difference, and
BAC/BrAC Ratio*
Men
Women
All Subjects
0.085
0.011
0.062–0.108
0.087
0.014
0.061–0.104
0.085
0.012
0.061–0.108
Mean
SD
Span
0.096
0.013
0.068–0.132
0.099
0.016
0.067–0.124
0.096
0.014
0.067–0.132
BAC-BrAC §
Mean
SD
Span
0.011
0.005
0.001–0.025
0.012
0.005
0.006–0.020
0.011
0.005
0.001–0.025
BAC/BrAC ratios #
Mean
SD
Span
1:2379
123
1:2125–2765
1:2385
95
1:2258–2535
1:2380
119
1:2125–2765
BrAC†
Mean
SD
Span
BAC ‡
* n = 98; 84 males and 14 females.
† p = 0.4519; t = 0.7553.
‡ p = 0.4075; t = 0.8319.
§ p = 0.6155; t = 0.5038.
# p = 0.8627; t = 0.1734.
239
Journal of Analytical Toxicology, Vol. 34, June 2010
tympanic thermometer (Gillette, Boston, MA), and a TemporalScanner 2000C temporal thermometer (Exergen, Watertown, MA). All three thermometers displayed temperature
measurements digitally and were reported by their manufacturer as complying with applicable ASTM standards. The displayed temperature results were recorded manually by the device operator. The mean of the three observed body
temperatures for each subject was reported as the subject’s
body temperature.
The correlation between body temperature and end-expired
breath temperature is shown in Figure 3; the correlation between body temperature and BAC/BrAC ratio is shown in
Figure 4; and the correlation between breath temperature and
BAC/BrAC ratio is shown in Figure 5. The correlation between
breath temperature and BAC/BrAC ratio for women was not
statistically significant (p = 0.0649). The other correlations
were statistically significant; however, all were much lower
than the correlation between BAC and BrAC.
Results
Discussion
In this paper, BrAC results are reported in g/210 L, and BAC
results are reported in g/100 mL of whole blood. All temperature measurements are reported in degrees Celsius. In the tables, the results are reported separately for men and women
and combined (all subjects); however, results for all subjects are
not included if the difference between the results for men and
women was statistically significant. Table I shows the demographics of the volunteer subjects, and Table II reports the
mean BrAC and BAC results used to calculate the BAC/BrAC ratios. There was a high and statistically significant correlation
between BAC and BrAC (r = 0.930, p < 0.0001), as shown in
Figure 1.
The mean end-expired breath temperature was 34.5°C, and
the difference between the end-expired breath temperature of
men and women was not statistically significant (p = 0.5765),
as shown in Table III. The mean body temperature was 36.6°C
for the men and 37.0°C for the women. The body temperature
for men was only 1% lower than the body temperature for
women; however, the difference was statistically significant (p
< 0.0001). The distributions of breath and body temperatures
are shown in Figure 2.
The percentages of men and women in the 98 physically fit
volunteer subjects (86% men and 14% women) who participated in this study closely approximate the percentages of men
(85%) and women (15%) who submitted to breath alcohol
tests after being arrested for driving while intoxicated in Texas
(unpublished statistic from the Texas Department of Public
Safety Breath Alcohol Laboratory, 2003–2008). The mean body
temperature was 36.6°C for men and 37.0°C for women and
spanned 36.0–37.6°C, as reported in Table III. This span agrees
with the 36.1–37.8°C reported by Simmers (8). The mean endexpired breath temperature was 34.5°C, which agrees with the
findings reported by Mason and Dubowski (4), Dubowski (10),
and Jones (11). The BAC exceeded the BrAC for every subject
Table III. End-Expired Breath and Body Temperature*
Breath temperature†
Mean
SD
Span
Body temperature
Oral‡
Mean
SD
Span
Tympanic§
Mean
SD
Span
Temporal#
Mean
SD
Span
Mean**
Mean
SD
Span
Figure 1. Scatter plot of BAC against BrAC. Men: r = 0.930, p < 0.0001,
n = 84; women: r = 0.969, p < 0.0001, n = 14; all Subjects: r = 0.938,
p < 0.0001, n = 98.
240
Men
Women
All Subjects
34.5
0.42
33.3–35.5
34.6
0.47
33.8–35.3
34.5
0.43
33.3–35.5
36.6
0.29
35.8–37.2
36.9
0.37
35.9–37.3
36.5
0.40
35.5–37.6
36.9
0.47
36.1–37.7
36.8
0.34
35.6–37.6
37.3
0.29
36.6–37.8
36.6
0.21
36.0–37.1
37.0
0.30
36.3–37.6
* n = 98; 84 males and 14 females.
† p = 0.5765; t = 0.5605.
‡ p = 0.0043; t = 2.9245.
§ p = 0.0011; t = 3.3752.
# p < 0.0001; t = 4.8271.
** p < 0.0001; t = 5.6988.
Journal of Analytical Toxicology, Vol. 34, June 2010
by an average of 0.011 and is reported in Table II. Consequently, the BAC/BrAC ratio for every subject exceeded 1:2100
(span 1:2125–2765). The mean BAC/BrAC ratio for men
(1:2379) and women (1:2385) were virtually identical. Similar
results have been reported by Jones and Andersson (12). The
breath temperature rose 0.84°C/°C increase in body temperature for men and 0.92°C/°C increase in body temperature for
women; however, the Pearson correlations of body temperature
against end-expired breath temperature, although statistically
significant, were lower than expected (men: r = 0.431, p <
0.0001; women: r = 0.582, p = 0.0290), as shown in Figure 3.
The temperature of the solution is a key variable in Henry’s
Law. The concentration of a volatile in the gas above a solution
increases as the temperature of the solution rises. In the measurement of BrAC, blood is the solution, ethanol in the breath
is the volatile gas, and core body temperature is the temperature
variable. Measuring the core body temperature on this number
of subjects was not practical in this study. The reported body
temperature of the subjects was the mean of three commonly
used clinical methods. If breath alcohol testing in humans
strictly complies with Henry’s Law, persons with lower body
temperatures should have consistently lower breath temperatures and higher BAC/BrAC ratios than persons whose body
temperatures are higher. That was not observed in this study.
The subject with the lowest body temperature (36.0°C) did
not have the lowest breath temperature; however, he did have
the highest BAC/BrAC ratio (1:2765). The subject with the
lowest breath temperature (33.3°C) had neither the lowest
body temperature nor the highest BAC/BrAC ratio. The subject
with the highest body temperature (37.6°C) had neither the
highest breath temperature nor the lowest BAC/BrAC ratio.
And the subject with the highest breath temperature (35.5°C)
had neither the highest body temperature nor the lowest
BAC/BrAC ratio. In fact, these three subjects had nearly identical BAC/BrAC ratios of 1:2259, 1:2263, and 1:2262, respec-
Figure 2. Distribution of end-expired breath and body temperatures (n =
98). Breath temperature: mean 34.5°C, median 34.5°C, mode 34.3°C;
body temperatures: mean 36.7°C, median 36.7°C, mode 36.7°C.
Figure 4. Scatter plot of BAC/BrAC ratio against mean body temperature.
Men: r = –0.349, p = 0.0011, n = 84; women: –0.670, p = 0.0087, n = 14.
Figure 3. Scatter plot of mean body temperature against end-expired
breath temperature. Men: r = 0.431, p < 0.0001, n = 84; women: r =
0.582, p = 0.0290, n = 14.
Figure 5. Scatter plot of BAC/BrAC ratio against end-expired breath temperature. Men: r = –0.415, p < 0.0001, n = 84; women: r = –0.506,
p = 0.0649, n = 14; all subjects: r = –0.421, p < 0.0001, n = 98.
241
Journal of Analytical Toxicology, Vol. 34, June 2010
tively. These findings strongly suggest that other factors in
addition to the temperature component of Henry’s Law are involved in producing a BrAC in humans.
The lower than expected correlation between body temperature, breath temperature, and BAC/BrAC ratio is reminiscent
of hematocrit research done by Jones. Men typically have a significantly higher mean hematocrit level (46.2%) than women
(40.6%). Jones’ in vitro experiments using a closed system of
air and blood demonstrated that the higher hematocrit levels
in blood taken from men resulted in significantly lower
BAC/BrAC ratios than in blood taken from women (5). However, Jones found no significant variation in the BAC/BrAC
ratio between men and women when in vivo experiments were
conducted. To explain the apparent contradiction Jones wrote,
“It seems that there are so many other physiological factors and
biological variations inherent in the quantitative measurement of BrAC that this hematocrit effect is completely submerged” (13). Similarly, this study suggests that there are so
many other physiological factors and biological variations inherent in the quantitative measurement of BrAC that the temperature component of Henry’s Law is completely subsumed.
Conclusions
For the physically fit subjects studied, their BrAC results
were consistently lower than their BAC results, and these results were very well-correlated. However, neither normal body
temperature nor end-expired breath temperature was strongly
associated with BAC/BrAC ratio.
Acknowledgments
A special thanks to CMI for providing the Intoxilyzer 8000,
to Toby Hall for modifying the instrument to measure breath
temperature, and to Tom Myers for making all the necessary
arrangements.
242
References
1. R.N. Harger, B.B. Raney, E.G. Bridwell, and M.F. Kitchel. The
partition ratio of alcohol between air and water, urine and
blood: estimation and identification of alcohol in these liquids
from analysis of air equilibrated with them. J. Biol. Chem.
183: 197–213 (1950).
2. American Medical Association. Committee on Medicolegal
Problems. Alcohol and the Impaired Driver: A Manual on
The Medicolegal Aspects of Chemical Tests for Intoxication:
With Supplement on Breath/Alcohol Tests. National Safety
Council, Chicago, IL, 1976, pp 96–98.
3. A.W. Jones. Physiological aspects of breath-alcohol measurement. Alcohol Drugs Driving 6(2): 1–25 (1990).
4. M.F. Mason and K.M. Dubowski. Breath as a specimen for
analysis for ethanol and other low molecular weight alcohols.
In Medicolegal Aspects of Alcohol, 3rd ed., J.C. Garriott, Ed.
Lawyers and Judges Publishing, Tucson, AZ, 1996, pp 175–
176.
5. A.W. Jones. Determination of liquid/air partition coefficients
for dilute solutions of ethanol, whole blood, and plasma.
J. Anal. Toxicol. 7: 193–197 (1983).
6. G.R. Fox and J.S. Hayward. Effect of hypothermia on breathalcohol analysis. J. Forensic Sci. 32(2): 320–325 (1987).
7. G.R. Fox and J.S. Hayward. Effect of hyperthermia on breathalcohol analysis. J. Forensic Sci. 34(4): 836–841 (1989).
8. L. Simmers. Diversified Health Occupations, 6th ed.
Thompson Delmar Learning, Albany, NY, 2003, p 351.
9. K.M. Dubowski. Breath-alcohol simulators: scientific basis
and actual performance. J. Anal. Toxicol. 3: 177–182 (1979).
10. K.M. Dubowski. Biological aspects of breath-alcohol analysis.
Clin. Chem. 20(2): 294–299 (1974).
11. A.W. Jones. Quantitative measurements of the alcohol concentration and the temperature of breath during a prolonged
exhalation. Acta. Physiol. Scand. 114: 407–412 (1982).
12. A.W. Jones and L. Andersson. Comparison of ethanol concentrations in venous blood and end-expired breath during a
controlled drinking study. Forensic Sci. Int. 132: 18–25 (2003).
13. A.W. Jones. Metabolism of alcohol in healthy men and women
and comparison of Widmark parameters and blood/breath
ratios of ethanol between the sexes. In Women, Alcohol, Drugs
and Traffic, Proceedings of an International Workshop,
M.R. Valverius, Ed. DALCTRAF, Stockholm, Sweden, 1989,
pp 169–175.
Manuscript received September 18, 2009;
revision received January 10, 2010.