Schaffer Online Library of Drug Policy Sign the Resolution
Contents | Feedback | Search
DRCNet Home
| Join DRCNet
DRCNet Library | Schaffer Library | Drugs and Driving 


What About the Humble Mouthpiece? Breath Sample Modification and Implications for Breath Alcohol Analysis

Christopher M. Bell

Victoria Forensic Science Centre, Forensic Drive, Macleod, Victoria, 3085, Australia


The modifying effect of the mouthpiece used in breath alcohol analysis was examined in terms of alcohol concentration in, and temperature of, the sample delivered. With repetitive testing, the standard Disposable Products mouthpiece (generally used for evidential breath analyses in Australia) was examined using a simulated blood alcohol concentration (BAC) of 0.250 g/100mL, delivery flowrate of 12 L/min and delivery volume of 2.5 L. The measured BAC and temperature values for tests with and without a mouthpiece were signicantly different (p< 0.0001) and the differences averaged 0.0020 g/100mL and 1.0 C respectively. Using an Allplastics brand mouthpiece, the reduction in BAC and temperature were 0.0027 g/100mL and 1.8 C. Further testing using the Disposable Products mouthpiece pre-warmed (40 C) and cooled (3 C), when compared to using a mouthpiece at 20 C, did not significantly alter BAC, but increased and decreased the measured temperature by 1.0 and 0.6 C respectively (p<0.0001). The temperature reduction using a Disposable Products mouthpiece and a human subject who provided breath samples in a standardised fashion, was 1.0 C. The conclusions are that the mouthpiece modifies the breath sample delivered, lowering both BAC and temperature. The variations in BAC and temperature do not correlate with the change in alcohol air/water partition ratio of 6.7% per degree. Therefore, the correction of BAC readings to a standardised temperature (of 34.0 C) is not straightforward.


Breath alcohol analysis requires the use of an unused, sterile, disposable mouthpiece for every subject test for the reasons of hygiene and the maintenance of sample integrity. In Australasia, the general practice is to use a mouthpiece taken from a sealed, sterile container, and to store mouthpieces at ambient temperature. As a breath sample is delivered, condensation of breath moisture in the mouthpiece is normally observed. Undoubtedly, the exhaled breath sample, nominally at 34C (Dubowski and Essary, 1983), cools upon delivery through a cooler mouthpiece. The extent of sample cooling, as well as any effects of the mouthpiece on the measured alcohol concentration, have not previously been reported upon in the literature. Such information is relevant to breath alcohol testing for several reasons. Different mouthpieces are used in different jurisdictions and their effects upon a sample may be different. Any modification of the measured blood alcohol concentration (BAC, normally expressed in Australia in the units of g/100mL) by breath test must be considered when comparisons are made between laboratories and between results from breath analyses and from actual blood sampling. Secondly, because of the discussion, particularly in Europe, of the measurement of exhaled breath temperature and correction of BAC or breath alcohol concentration (BrAC) readings to a standard temperature (of 34C), the issues of sample temperature measurement and the validity of any such correction must be investigated.


The measurements in this investigation were performed with a specially modified Drger Alcotest 7110 Mark IV analyser and associated assembly that provided measurements of 4 variables (BAC, volume, temperature and pressure of the sample delivered) over the term of the delivery at a rate of 4Hz. This system and its performance are described elsewhere (Bell, and Flack, 1995). Alcohol / water solutions were used to provide a simulated BAC of 0.250 g/100mL at 34.0C in all tests. The solution concentration was verified by gas chromatography. Bath temperature was continuously measured using a certified thermometer to ensure temperature stability in these tests. The flowrate of air bubbled through the solutions was measured by a Fischer & Porter rotameter calibrated for 0-30 L/min air flowrates. All BAC and temperature readings were recorded after a sample delivery volume of 2.5 L. Repetitive testing was conducted to enable statistical comparisons and a fresh mouthpiece used for each test.

Two brands of mouthpiece were investigated. The Disposable Products mouthpiece is used for evidential breath testing in Australia, while the Allplastics product is used in New Zealand. Some physical characteristics of each brand were measured. Simulated breath tests were conducted to examine the effect on BAC and temperature for both brands of mouthpiece, as compared to not using a mouthpiece. Additional tests were performed with Disposable Products mouthpieces to examine the effects of variation in the air delivery flowrate and the initial mouthpiece temperature. The order of tests for any particular regime were in a cyclical fashion such that any variations caused by alcohol solution depletion were included in all the conditions under test.


The physical characteristics examined for each brand of mouthpiece are given in Table 1. Both brands are 2 baffle / 2 way type mouthpieces, i.e., neither is fitted with a non-return valve. Results for comparison between mouthpieces, in terms of BAC and temperature, and with reference to the control condition of no mouthpiece, are given in Table 2.

Table 1
Physical Characteristics of Mouthpieces

Brand Allplastics Disposable Products
Type 2 baffle / 2 way, flat shape 2 baffle / 2 way, barrell shape
Country of use New Zealand Australia
Mass (g) 6.8 4.1
Internal volume (mL) 9.3 2.5

Table 2
Comparison of Mouthpieces

Test* No mouthpiece** Disposable Products AllPlastics
BAC (g/100mL) temp. (C) BAC (g/100mL) temp. (C) BAC (g/100mL) temp. (C)
1 0.243 33.3 0.242 32.3 0.241 31.7
2 0.244 33.5 0.242 32.3 0.242 31.5
3 0.244 33.4 0.242 32.4 0.241 31.6
4 0.244 33.5 0.242 32.5 0.242 31.7
5 0.244 33.5 0.242 32.7 0.239 31.7
6 0.244 33.5 0.241 32.4 0.241 31.7
7 0.243 33.5 0.241 32.8 0.241 31.8
8 0.244 33.6 0.242 32.3 0.241 31.6
9 0.243 33.3 0.241 32.5 0.241 31.8
10 0.244 33.3 0.242 32.7 0.241 31.6
mean 0.2437 33.44 0.2417 32.49 0.2410 31.67
standard deviation 0.0005 0.10 0.0005 0.18 0.0008 0.09
* All tests were performed using an air flowrate of 12L/min
** Order of tests per cycle: no, Allplastics, then Disposable Products mouthpiece

The tests of simulated breath samples without using a mouthpiece provide a means by which to evaluate the simulator used. The temperature of the solution in the final chamber of the simulator was constantly at 34.0C throughout the testing. The measured temperature of the effluent gas was 0.6 C lower than this expected value. The accuracy of the temperature measurement was 0.2C (Bell, and Flack, 1995), therefore, it is presumed that a reduction in temperature occurred in the short metal delivery tube exiting the heated compartment of the simulator. For most of the testing conducted here, this phenomenon is irrelevant because the conditions of use of the simulator remain unaltered and therefore any temperature drop within it remains constant. The possible exception is with changes in air flowrate through the simulator.

The measured BAC and temperature values for tests with and without a mouthpiece were signicantly different (p< 0.0001). For the Disposable Products mouthpiece, the BAC and temperature reductions were, on average, 0.0020 g/100mL and 1.0 C respectively. For the Allplastics mouthpiece, the BAC reduction was 0.0027 g/100mL and significantly different to that measured for the Disposable Products mouthpiece (p<0.05). The temperature reduction when using the Allplastics mouthpiece was, on average, 1.8 C. The larger reduction in temperature when using the Allplastics mouthpiece was to be expected due to the larger mass and internal volume of this product.

The variations in BAC and temperature do not correlate with the change in alcohol air/water partition ratio of 6.8% per degree as reported by Dubowski (1979). Other studies (Dubowski; Schoknecht and Kophamel, 1989) indicate the blood/breath partition ratio change with temperature is of equivalent magnitude. Given such a correlation, the expected BAC readings for temperature reductions of 1.0 and 1.8 C for the instrument used in the testing would be 0.227 and 0.214 g/100mL respectively. Tests (n=10) with a human subject of zero BAC using Disposable Products mouthpieces also produced a temperature reduction of 1.0 C.

Further testing was performed with Disposable Products mouthpieces to examine the effects of changes in air delivery flowrate and the initial mouthpiece temperature. A synopsis of repetitive tests (n=12) is provided in Table 3 and t-statistics for various comparisons of averages given in Table 4. The BAC reduction caused by using a mouthpiece in these tests was larger than previously measured. The flowrate in these tests however, was 15 rather than 12 L/min, and as the results in Table 3 indicate, the BAC reduction is flowrate dependent. Two reasons arise which may explain flowrate dependence. Firstly, the simulator may not have attained complete alcohol equilibration in the dynamic nature of airflow through the bubbling vessels. This seems unlikely given that the simulator is fitted with two solution / bubbling chambers, 17m of copper air-preheating coil and that the internal airspace at 34.0C is approximately 2 L. Secondly, the increase in flowrate caused an increase in the pressure generated within the system. At the end of sample delivery, the sample residing in the chamber of the Alcotest expands to attain atmosperic pressure, thereby diluting the alcohol concentration measured by the instrument. The delivery pressure was also measured by the modified Alcotest unit and pressure corrections to translate sample delivery flowrates of 21 and 9 L/min to 15 L/min were calculated to be 0.0019 and -0.0014 g/100mL respectively. The possibility of variation in the temperature of the air exiting the simulator with changes in air flowrate passing through it was mentioned earlier, however, no significant differences were observed between measured temperatures at different air flowrates.

Table 3
Further Tests with Disposable Products Mouthpieces to Examine the Effects of Changing Flowrate and Intial Mouthpiece Temperature

Condition* Variable mean standard deviation
no mouthpiece, airflow 15L/min BAC (g/100mL) 0.2403 0.0013
temp. (C) 33.25 0.19
mouthpiece, airflow 15L/min BAC (g/100mL) 0.2376 0.0014
temp. (C) 32.36 0.12
mouthpiece, low airflow 9L/min BAC (g/100mL) 0.2407 0.0012
temp. (C) 32.38 0.11
mouthpiece, high airflow 21L/min BAC (g/100mL) 0.2334 0.0021
temp. (C) 32.47 0.14
warm mouthpiece 40C, airflow 15L/min BAC (g/100mL) 0.2378 0.0021
temp. (C) 33.33 0.22
cold mouthpiece 3C, airflow 15L/min BAC (g/100mL) 0.2368 0.0019
temp. (C) 31.76 0.17
* Tests were conducted in a cyclical fashion in the order listed in the Table (n=12)

Table 4
T-Statistics for Various Comparisons

Comparison Probability*
BAC temp.
mouthpiece vs no mouthpiece 0.0001 0.0000
low (9L/min) vs standard (15L/min) flow 0.0000 0.6198
high (21L/min) vs standard flow 0.0000 0.0606
high vs low flow 0.0000 0.1363
cold (3C) vs room temperature (20C) mouthpiece 0.3094 0.0000
warm vs room temperature mouthpiece 0.7459 0.0000
warm vs no mouthpiece 0.0027 0.3527
* Null hypothesis Ho: 1-2=0, n=12

The use of a pre-warmed (40C) or cold (3C) mouthpieces did not significantly alter the BAC reduction away from that measured when using a mouthpiece at room temperature. The temperature however, was significantly different (p<0.0001), higher and lower for warm and cold mouthpieces respectively. The measured temperature using a warm mouthpiece was not significantly different to that for not using a mouthpiece. The significance of this observation is unclear - the temperature correlation in this case may have been coincidental. Mouthpieces pre-warmed to other temperatures may modify the measured temperature accordingly. It is apparent that pre-warmed mouthpieces aid in avoiding a temperature reduction upon sample delivery, however further investigations were not possible as part of this study.

The temperature measurement assembly described by Schoknecht (1993) includes a mouthpiece receptacle that pre-heats the mouthpiece to 37C. Pre-heating the mouthpiece is certainly advantageous, yet part of the mouthpiece extends beyond the heated zone into ambient conditions. Therefore no guarantee of effective and complete pre-heating of the mouthpiece is offered, nor reported upon in the literature. As indicated above, the appropriate pre-heating temperature to avoid modification of the temperature of the sample needs further investigation before it can be considered for unambiguous breath temperature measurement.


It is clearly indicated that a small reduction in BAC occurs upon delivery of a (breath) sample through a mouthpiece, regardless of mouthpiece temperature, and that the delivery flowrate influences the final BAC result obtained, probably due (in part) to different degrees of expansion upon equalisation of pressure differences. A reduction in sample temperature also occurs, which is dependent upon the type and temperature of the mouthpiece used. The variations in BAC and temperature do not correlate with the change in alcohol air/water partition ratio of 6.8% per degree, nor with the similar change in blood/breath partition ratio for alcohol. Therefore, the correction of BAC readings to a standardised temperature (of 34.0 C) is not straightforward. It appears that use of a mouthpiece pre-heated to a temperature above exhaled breath temperature improves the accuracy of this measurement. However, if pre-heating is considered desirable, the selection of an appropriate pre-heating temperature needs further investigation, as does the proof of complete pre-heating of the mouthpiece in the practical situation where the end of the mouthpiece must be exposed to allow breath sample delivery by a subject.


This study was undertaken as part of a grant from the National Drug Crime Prevention Fund. The basic Alcotest 7110 instrument and the mouthpieces were kindly supplied by the Victoria Police Traffic Alcohol Section. Special thanks are given to Dr. S.J. Gutowski of the Victoria Forensic Science Centre and Mr. N. Farrell, for their assistance with the preparation of the paper and the statistical comparisons.


Bell, C.M. and Flack, H.J., "Development of a system for real-time breath alcohol analysis", ICADTS T95, Adelaide.

Dubowski, K.M., "Biological aspects of breath-alcohol analysis", Clin. Chem., v20, p294.

Dubowski, K.M., "Breath-alcohol simulators: Scientific basis and actual performance", J. Anal. Toxicol., (1979), v3, p177.

Dubowski, K.M., and Essary, N.A., "Studies on breath parameters in human subjects applicable to breath-alcohol analysis", Proceedings of the 9th ICADTS, - San Juan, Puerto Rico, U.S. Dept. of Transport, (1983), p393.

Schoknecht, G., "The influence of temperature on breath-alcohol analysis", Alcohol, Drugs and Traffic Safety - T92, Publishers Verlag TUV Rheinland, Cologne, (1993), p392.

Schoknecht, G., and Kophamel, B., "Das temperaturproblem bei der atemalkoholanalyse", Blutalkohol, (1989), v26, p137.