Biological Aspects of Breath-Alcohol Analysis1

Transcription

1 CLIN. CHEM. 20/2, (1974) Biological Aspects of Breath-Alcohol Analysis1 Kurt M. Dubowski We studied several relevant biological aspects of breath-alcohol, analysis in 55 healthy men and women, after alcohol ingestion and during breathalcohol analysis with a typical 4th-generation instrument. We measured breath volumes, delivery pressures, and end-expiratory temperatures, with the following findings: End-expiratory temperature, #{176}C (mean, #{176}C); forced vital Capacity, ml (mean, 4038 ml): maximum expiration after normal inspiration, ml (mean, 2730 ml); breath delivery pressure into a prototype Borg-Warner Model P-7 breath-alcohol apparatus, 8-50 inches H20 (mean, 21.3). These data and other pertinent findings from this study should assist in more rational design of breath-sampling systems in forthcoming breath-alcohol instruments and in developing valid procedures for their use. Additional Keyphrases: breath volume, delivery pressure, and end-expiratory temperature measured #{149}breath alcohol vs. blood alcohol concentrations #{149}toxicology #{149} variables in measyring alcohol in law-enforcement practice The measurement of alcohol2 in breath importantly involves various physiological and biological aspects (1, 2), and adequate information regarding them is fundamental to sound design of instruments and development of valid procedures for breath-alcohol analysis. However, in contrast to the analytical aspects of alcohol detection and quantitation, the biological factors have heretofore received only limited experimental attention under conditions applicable Departments of Medicine, Biochemistry, and Pathology, and Toxicology Laboratories, University of Oklahoma Health Sciences Center, Oklahoma City, Okla Presented in part at the 25th National Meeting of the AACC, New York, N. Y., July 15-20, Terms and nonstandard abbreviations used: The unmodified word alcohol in this article refers to ethanol. BAC, blood-alcohol concentration, which is universally expressed in applicable statutes in per cent weight/volume concentration units; % w/v X 10 = g/liter and % w/v X 10 = mg/dl (e.g., 0.10% w/v 1.0 g/ liter 100 mg/dl). Received and accepted Dec. 5, to breath-alcohol analysis for clinical and law-enforcement purposes, and in studies of human subjects under the influence of alcohol to the extent commonly encountered in such situations. Accordhgly, we have studied the significant breath-sample characteristics of volume, delivery pressure, and endexpiratory temperature in a series of healthy men and women after they had ingested alcohol and during breath-alcohol analysis with typical 4th-generation instruments. Materials and Methods Apparatus, Methods, and Procedures Blood-alcohol concentrations were determined by automated gas-chromatographic headspace analysis with the Multifract F-40 instrument (Perkin- Elmer Corp., Norwalk, Conn ), by the procedure of Machata (3), modified by adding 1 g of sodium chloride to 1 ml of whole blood before equilibration at 65 #{176}C. Breath-alcohol concentrations were monitored by measuring infrared absorbance at 3.39 tm with a Model 4001 Intoxilyzer (Omicron Systems Corp., Palo Alto, Calif ), and further quantitative breath-alcohol measurements were made by a catalytic oxidation technique with a Model P-7 H.A.L.T. breath-alcohol apparatus (Borg-Warner Corp., Alcohol Countermeasures Systems, Des Plaines, Ill ). Breath pressures were measured with a Model 2050C Magnehelic direct-reading differential pressure gauge (Dwyer Instruments, Inc., Michigan City, md ). Breath temperatures were measured with a Heath/Schlumberger Model EU /EU digital thermometer (Heath Co., Benton Harbor, Mich ) and a Model No. 705 thermolinear probe (Yellow Springs Instrument Co., Yellow Springs, Ohio 45387). Breath volumes were measured with a Drager Model Volumeter (North American Drager, Telford, Pa ). 294 CLINICAL CHEMISTRY, Vol. 20, No. 2, 1974

2 Table 1. Characteristics of Methods Employed Blood-Alcohol determination by automated GC headspaee analysis with Perkin-Elmer Multifract F-40 instrument Precision of replicate determinations (n = 29, 28, 27, 24): At 50 mg/dl BAC Mean = 50.0mg/dl SD = 0.27 mg/dl CV = 0.55% Gas-Pressure measurement by direct-reading 7llagnehelic pressure gauge Precision of replicate measurements (n = 11): At 15 in. H20 Mean = 15.0 in. H20 SD = 0.08in.H,O CV = 0.53% Gas-Temperature measurement by Heath-Schlumberger digital thermometer + YSI thermolinear probe No. 705 Precision of replicate measurements (n = 21): At 34.5 #{176}C Mean = 34.50#{176}C SD = #{176}C CV 0.02% Gas-Volume measurement by direct-reading Drager Volumeter Precision of replicate measurements of gas syringe output (n = 10): At 470 ml volume Mean = ml SD = 1.94 ml CV = 0.41% Simulator solution measurement by automated GC headspace analysis with Perkin-Elmer Multifract F40 instrument Validation of ethanol solution concentration for nominal 100 mg/dl BAC Simulator output at 34 #{176}C For 14 solutions on 14 days (expressed as corresponding nominal BAC); (n = 14): Mean = mg/di SD = 1.64mg/dl CV = 1.64% Known concentrations of alcohol in air were produced by controlled temperature equilibration at 34 tubes (Cat. No. 4726; Becton, Dickinson & Co, lected, with sterile precautions, into Vacutainer ± 0.2 #{176}C with Mark H Simulators (Smith & Wesson Rutherford, N. J ) containing 20 mg of potassium oxalate and 25 mg of sodium fluoride. After Electronics Co., Eatontown, N. J ), operated with compressed gas (N2:02, 79:21 by vol) at 15 numerical coding, the blood specimens were stored at inches of water delivery pressure. Ethanol reference solutions for equilibration with air at 34 #{176}Creplicate, by analysts who had no knowledge of the 4 #{176}C, and were analyzed for alcohol within 48 h, in were prepared by appropriate dilution from a 60.5 results of the breath-alcohol tests. g/liter stock solution of ethanol.3 These alcohol-inair reference mixtures were used to calibrate the Other Procedures H.A.L.T. breath-alcohol apparatus, and periodically to check the calibration and performance of this man was measured by continuous mass-spectromet- The composition of expired breath in a healthy device and of the Intoxilyzer apparatus. nc analysis, in the single-breath test mode, with a The in vivo studies were conducted. on healthy volunteer human subjects (40,, age years; 15 Corp., Medical Instruments, Pomona, Calif ). Model 1100 Medical Gas Analyzer (Perkin-Elmer 9, age years), in accordance with national The CO2 and 02 content of the expired breath were standards for investigations involving human recorded, in percent by volume, against time, with subjects (6). After appropriate briefing, preparation, precautions to assure constant rates of breath flow. and establishment of alcohol-free status, the subjects consumed doses of alcoholic beverages calculated to Results yield maximum BACs ranging up to 0.20% w/v.2 Performance Characteristics of Methods Breath volume, pressure, and temperature were Key characteristics of the principal methods we measured in replicate following alcohol consumption. used are summarized in Table 1, which includes data After the subjects had reached the post-absorptive on the precision, in nonbiological reference systems, state, as evidenced by frequent breath-alcohol monitoring, paired samples of breath and antecubital ve- of the methods used for breath measurements. nous whole blood were collected as nearly simultaneously as practicable, and always within 2 mm. Subjects after Alcohol Consumption Breath-Sample Characteristics in Human Blood specimens of approximately 10 ml were col- 3 Equilibration of air with a 1.21 g/liter ethanol solution maintained at 34 #{176}C will yield a gas mixture containing mg of ethanol per liter. The alcohol content of this gas mixture is equal to that of expired alveolar air from a subject with a BAC of 1.0 g/liter, according to the accepted blood/breath alcohol relation that 2.1 liters of expired alveolar air contain approximately the same quantity of alcohol as 1 milliliter of blood (4, 5). Results of the studies of breath temperature and breath volume in 55 healthy human subjects are summarized in Tables 2 and 3. All temperatures shown were recorded at the end of an expiratory vital capacity maneuver4 and are therefore end-expiratory temperatures. The breath-volume data include both the maximum forced expiratory volume, CLINICAL CHEMISTRY, Vol. 20, No.2,

3 Table 2. End-Expiratory Breath Temperatures in Human Subjects, after Alcohol Consumption, Measured at the Mouth End.explratory temperature, #{176}C 5ubjects ii Range Mean Men Women Total Table 3. Expiratory Breath Volumes in Human Subjects after Alcohol Consumption Subjects Men Women Total n Forced Range vitalcapacity, ml Mean Maximum exhalation after normal Inhalation, ml Range Mean as measured by an expiratory forced vital-capacity maneuver4 in standing subjects, and the maximum expiratory volume in the same subjects after a normal inhalation. All volumes shown are for breath at physiological temperature, saturated with water vapor, and at ambient barometric pressure.5 Table 4 summarizes the data on breath pressure and breath volume for 19 subjects during breath-alcohol analysis with a prototype version of the Borg- Warner P-7 H.A.L.T. apparatus. The breath pressures shown are peak values attained above the ambient atmospheric pressure during the sampling period, measured at the breath sample inlet of the P-7 instrument. The breath volumes are shown as the absolute volumes delivered to this breath-alcohol apparatus during the fixed 4.2 s sampling period, and as the fraction of the maximum expinatory volume after a normal inhalation in the same subjects. Performance of the Breath-Alcohol Apparatus Figure 1 shows the accuracy with which synthetic (Simulator-produced) alcohol-vapor reference specimens were analyzed with the Borg-Warner P-7 prototype apparatus. Typical precision of this apparatus in the analysis of such reference specimens is illustrated by results (indicated as the corresponding BAC) of 51 consecutive measurements of Simulatorproduced reference specimens of nominal 100 mg/dl BAC value3: range, mg/dl; mean, mg/ dl; SD, 1.19 mg/dl; and CV, 1.19%. Fieure 2 shows how the same apparatus performed in predicting the coexisting BAC by analysis of the breath. Precision of these measurements in breath is illustrated by the mean difference of +0.3 mg/dl be tween consecutive results for 19 sets of three consecutive breath specimens over a BAC range to 113 mg/ dl. The P-7 apparatus was calibrated for these measurements, as before, with reference mixtures of alcohol vapor, on the basis of the accepted blood! breath alcohol relation for expired alveolar air EQUIWL.ENT BAC of ETHANOL P0R, S a/v Fig. 1. Correlationbetween expected values and results obtained with a prototype Borg-Warner Model P-7 breathalcohol apparatus in the analysis of 220 Simulator-produced alcohol-vapor reference specimens Vapor alcohol concentrations are shown in terms of the corresponding blood alcohol concentrations Vital capacity (VC) is the volume of air that can be expelled during a maximal exhalation after a maximum inspiration: forced vital capacity (FVC) is the vital capacity performed with expiration as forceful and rapid as possible. In normal subjects, VC equals FVC. ambient barometric pressure extremes over the duration of the human subject studies were mm Hg, corrected to sea level. Breath-alcohol instruments were calibrated at the atmospheric pressures existing at the time the breath or reference samples were analyzed BAC by DIRECT BL000 ANALYSIS,S Fig. 2. Correlation between results of analyses for alcohol in 121 simultaneouslycollectedblood and breath sample pairs Breath analyses were performed with a prototype Borg-Warner Model P-7 breath-alcoholapparatuswithdirectdigital BAC readout; blood specimens were analyzed by an automated gas chromatographic headspace procedure (3) 296 CLINICAL CHEMISTRY, Vol. 20, No. 2, 1974

4 Table 4. Characteristics of Breath Samples from Human Subjects, after Alcohol Consumption and during Breath-Sampling with a Prototype Model P-7 Borg-Warner Breath-Alcohol Apparatus P-i sample volume, mu Breath pressure, in. H20 Breath sample volume, ml Maximum expiratory volume, ml Sampling Subjects n Range Mean Range Mean Range Mean Men Women Total period,4.2s Mass Spectrometric Expirogram Figure 3 shows a typical single-breath mass spectrometric expirogram for CO2 and 02. Discussion The several methods used in this study have good precision (Table 1), and the variations found in the human subject studies are, thus, not attributable to the methods of measurement. The breath temperature data shown in Table 2 are of significance in several respects. Breath temperatures as low as 31 #{176}C (leaving the mouth) have been reported in the literature dealing with breath-alcohol analysis (7, 8). Information regarding the temperature of the breath collected for alcohol analysis is required to determine the minimal temperature at which the breath collection and storage systems must be maintained to prevent condensation of water vapor and consequent alcohol loss, and to apply proper Charles law corrections to the breath volumes for calibration and analysis. Further, the vapor-pressure curve of alcohol is very steep between 30 #{176}C and 40 #{176}C, and the air/blood and air/water lol 1#{176} l lil 02 - I (l Ilol ll CO2.: 0% Gould Inc., Inst,un,snt Systems D4,lston Fig. 3. Mass spectrometric single-breath expirogram obtainelyzer with a Model 1100 Perkin-Elmer Medical Gas Ana- upper and lower curves show the breath 02 and CO2 content, respectively, (in vol per 100 volumes), during a single continuous full expiration at constant breath flow rate partition ratios of alcohol increase about 9% per degree of temperature increase over this range (9). The end-expiratory air used in breath-alcohol analysis can thus be expected to have a lower ethanol vapor pressure at a mean temperature of 34.5 #{176}C than the air in the alveoli at 37.5 #{176}C, and the vapor pressure of alcohol in breath will clearly fluctuate somewhat with changes in deep-body temperature and in breath temperature. However, the final ethanol vapor pressure attained is not solely a function of the temperature differences. The end-expiratory temperatures shown in Table 2 and the mean value of 34.5 #{176}C agree well with the findings of Harger and Forney (10),who reported a mean temperature of 34.4 #{176}C in a plastic mouthpiece at the end of an exhalation in a small series of subjects. We have observed that the rise in expiratory temperature approximately parallels the alcohol content of exhaled air during expiration, in a manner similar to the phenomenon illustrated in Figure 3. Temperature measurements with a sufficiently rapidly responding system may thus be usable to indicate when the breath-alcohol plateau has been reached and the sample for analysis should be taken. Procurement of expired true alveolar breath, in which alcohol is considered to be in equilibrium with that present in the pulmonary arterial blood (4, 5, 11), is one of the most difficult practical problems encountered in breath-alcohol analysis, especially in law-enforcement, where untrained subjects are involved and only a single sampling may be permitted. The common breath samplers shown in Figure 4 illustrate one facet of this problem of procuring expired alveolar air from untrained subjects. The samplers are designed to discard initial dead-space air, on the assumption that the remaining sample to be retained is alveolar air (12, 13). However, as Wright (14) has noted, there is a strong suggestion that much more than the conventional physiological dead space requires to be discarded before a true sample of alcoholic breath can be obtained. It seems reasonable to hypothesize that the alcohol content of breath, during continuous complete expiration, parallels the breath CO2 and 02 contents, especially the former. As illustrated in Figure 3, these become asymptotic at about 8 s dur- CLINICAL CHEMISTRY. Vol. 20, No.2,

5 BREATH#{149} BREATH I Fig.4. Breath samplers for alcohol analysis, intended to procure expired alveolar air by discard of dead-space air into discard bag, which fills first Left: Wright Pattern Sampler (12). Right: Commercial screening test sampler (13) ing a continuous expiration period of 9.2 s, or after 65% of the total breath has been exhaled, assuming a constant rate of breath flow during the expiration. Hence, 65% or more of the maximum expiratory volume must be discarded before the highest consistent breath-alcohol concentration (or partial pressure) plateau is attained in the specimen. The data in Table 3 show that this requires that a mean volume of at least 1775 ml be discarded after a normal inspiration, whereas volumes of the discard bags of current commercial breath-collection systems of the type illustrated in Figure 4 vary from 500 to 750 ml. The larger discard requirement is in agreement with our studies, which indicate that end-expiratory breath samples are in substantially better alcohol equilibrium with the blood than are so-called alveolar-air specimens obtained after discard of the dead-space air. Jones and Jones (15) recently studied the alcohol concentration in the breath of seven subjects as a function of their expired breath volume, and also concluded that it is necessary to discard more than 70% of the lung vital capacity prior to breath sampling. Forensic breath-alcohol analysts, unfortunately, commonly direct their subjects to provide a breath specimen after a maximal or near-maximal inhalation. In such situations, the data in Table 3 indicate that a mean breath volume of 2625 ml must be discarded before sampling to obtain the highest breathalcohol concentration plateau. These findings can largely explain the documented (16) inadequate performance of breath-alcohol screening test devices in which breath-collection bags of a too-limited volume are used. Similar considerations with respect to breath volumes apply to other schemes for breath sampling and collection for alcohol analysis. The practical effect of inadequate breath sampling is illustrated in Figures 1 and 2. Although the prototype Borg-Warner P-7 breath-alcohol apparatus is obviously capable of highly accurate and precise alcohol analysis in gas specimens (Figure 1), the correlation between the alcohol content of paired breath and blood specimens obtained simultaneously is clearly not as good (Figure 2). The differences in results are probably attributable to the fact that a mean of only 24% of the available breath (and never more than 51%) was supplied to the instrument, as shown in Table 4, whereas the instrument calibration (and the direct BAC readout) were based on the established blood/ breath ratio (4, 5), which applies only to expired true alveolar air, or rebreathed air. Accordingly, the instrument is undergoing redesign. Classical literature, and much of the recent breath-alcohol literature are, unfortunately, particularly confusing or misleading with respect to breath sampling procedures suitable for obtaining expired true alveolar air. In 1905, Haldane and Priestley (17) first reported that in air obtained after normal breathing by a quick forced expiration, the partial pressure of CO2 was constant after washout of the dead space, and they concluded that all such air was pure alveolar air. Their work has been widely and authoritatively cited to the effect that an expiration of 400 ml will completely wash out the respiratory dead space of a normal resting adult and that, therefore, all air expired after the first 400 ml is alveolar air (18). A typical recent report of a new instrumental breath-alcohol analysis procedure (19) includes the statement that to obtain a true alveolar air sample, the volume of expired breath must be greater than 1 liter. As Table 3 shows, these suggested volumes may be grossly inadequate to achieve a 65% discard plateau of breath composition. The data on breath pressure shown in Table 4 for the Borg-Warner P-7 apparatus are of significance in at least two respects. With a fixed breath sample path, fixed flow resistance, and a fixed sampling time, the rate of breath flow and hence the breath volume throughput are functions of the breath pressure. Further, the typical breath-delivery pressures shown in Table 4 are presumably reflected in comparable increases (above atmospheric) in the intrathoracic pressure gradients. When breath samples are brought from these pressures to atmospheric pressure, the concentration of alcohol is proportionately decreased below that in the lungs. Hence the resistance to breath flow should be kept to a minimum in designing breath-alcohol instruments. Further, when alcohol reference gas mixtures are produced by means of Simulators or equilibrators for calibration of breath-alcohol instruments, they should preferably be produced at total pressures corresponding to those to be expected during breath sampling with the instrument being calibrated. This investigation was supported in part by Grant No. GM from the NIGMS, NIH, USPHS; and by the U. S. Department of Transportation, The Insurance Institute for Highway Safety, Perkin-Elmer Corp., and Borg-Warner Corp. 298 CLINICAL CHEMISTRY, Vol. 20, No. 2, 1974

6 References 1. Mason, M. F., and Dubowski, K. M., Alcohol, traffic, and chemical testing in the United States: A r#{233}sum#{233} and some remainingproblems. Clin. Chem. 20,126(1974). 2. Dubowski, K. M., Measurement of ethyl alcohol in breath. In Laboratory Diagnosis of Diseases Caused by Toxic Agents, F. W. Sunderman and F. W. Sunderinan, Jr., Eds. Warren H. Green, Inc., St. Louis, Mo., 1970, pp Machata, G., Uber die gaschromatographische Blutalkoholbestimmung. Blutalhohol 4,252(1967). 4. Borkenstein, R. F., et al., Statement. In Proceedings of the Ad Hoc Committee on Blood/Breath Alcohol Ratio, Indiana University, Indianapolis, Jan Harger, R. N., Forney, R. B., and Baker, R. S., Estimation of the level of blood alcohol from analysis of breath. II. Use of rebreathed air. Quart. J. Stud. AIc. 17, 1(1956). 6. The Institutional Guide to DHEW Policy on Protection of Human Subjects. DHEW Publication No. (NIH) U. S. Government Printing Office, Washington, D.C., Liljestrand, G., and Linde, P., Uber die Ausscheidung des Alkohols mit der Expirationsluft. Skand. Arch. Physiol. 60, 273 (1930). 8. Borkenstein, R. F., and Smith, H. W., The Breathalyzer and its applications. Med. Sci. Law 2, 13(1961). 9. Harger, R. N., Raney, B. B., Bridwell, B. G., and Kitchel, M. F., The partition ratio of alcohol between air and water, urine and blood; estimation and identification of alcohol in those liquids from analysis of air equilibrated with them. J. Biol. Chem. 183, 197 (1950). 10. Harger, R. N., and Forney, R. B., The temperature of rebreathed air, and the arterial blood:rebreathed air alcohol ratio. In Proc. 5th mt. Conf. on Alcohol and Traffic Safety. Freiburg i/ Br., Germany, Sept. 1969, pp II Harger, R. N., Forney, R. B., and Barnes, H. B., Estimation of the level of blood alcohol from analysis of breath. J. Lab. Clin. Med. 36, 306 (1950). 12. Wright, B. M., A versatile expired air sampler. J. Physiol. 184, 66P (1966). 13. Alveolar Air Breath Alcohol System. Becton-Dickinson Div., Becton, Dickinson and Co., Rutherford, N.J., Wright, B. M., Breath alcohol analysis. In Alcohol and Road Traffic: Proc. 3rd mt. Conf. on Alcohol and Road Traffic. British Medical Association, London, England, 1963, pp Jones, A. W., and Jones, T. P., Breath alcohol analysis with the Intoximeter gas chromatograph. In Proc. OECD mt. Symp. on Countermeasures to Driver Behaviour Under the Influence of Alcohol and Other Drugs, London, England, Sept Prouty, R. W., and O Neill, B., An Evaluation of Some Qualitative Breath Screening Tests for Alcohol. Insurance Institute for Highway Safety, Washington, D.C., 1971, pp Haldane, J. S., and Priestley, J. G., The regulation of the lung ventilation. J. Physiol. 32, 225 (1905). 18. Peters, J. P., and Van Slyke, D. D., Quantitative Clinical Chemistry, II. Methods. Williams & Wilkins Co., Baltimore, Md., 1932, p Hill, R. G., Simmonds, M. A., and Straughan, D. W., Electrochemical measurements of blood alcohol levels. Nature 240, 62 (1972). CLINICAL CHEMISTRY, Vol. 20, No. 2,