530 Proc. roy. Soc. Med. Volume 61 May 1968 200 blood and tissue uptake difference at this time may be postulated to be due to alteration in tissue blood flow. Ten minutes later, with vaporizer still set at '2', the marked decrease in alveolar ventilation was reflected by a fall in halothane uptake by the blood. Tissue uptake was greater than blood uptake, which was consistent with the fall in arterial halothane from 25 6 to 21-0 mg/100 ml. Despite the maximal vaporizer setting for a period of ten minutes the blood uptake was only 11-1 mg/min because the alveolar ventilation was low. The associated fall in cardiac output to 460 of the control value led to the low uptake of 8-6 mg halothane per minute. HEART RATE 160 20 20 L/nmin - ...._______. RESPIRATORY - RATE X ,v 2 -'ALVEOLAR * VENlTILATION 1.5 L/.,i. CARDIAC OUTPUT 1 14 0.5 * mg/100ml 25 -BLOOD HALOTHANE 15 - 10 ARTERt~L/ '/ TISSUE RALOTRANE MIXED VENOUS 35.1 311.6 6.611.1 36.9 212. 4 31.6 mg/min ARTERIAL HALOTHANE UPTAKE Dg/miD S IErr 2 10 20 TIME (.in.) 4 |r 30 a.e ON FF < 40 Fig 1 Halothane uptake in a 20 kg dog during closed circuit anaesthesia with vaporizer inside the circuit and spontaneous respiration saturation at a constant inspired halothane concentration prolonged administration would be required (Duncan & Raventos 1959). The results are consistent with the view that tissues of high lipid content and high blood flow take up large amounts of halothane rapidly. The low mean halothane value for omentum, despite its high lipid content, is probably due to its low blood flow. The range of variation of uptake by the different tissues may be explained by the variation in minute ventilation and body weight among the dogs. Halothane uptake in the dog: Fig 1 shows the data for uptake of halothane by a 20 kg dog during closed circuit anxsthesia. Previous studies on halothane uptake have been made during constant inspired halothane concentrations (Mapleson 1962, Eger & Guadagni 1963). During closed circuit ancesthesia with the vaporizer inside the circuit, the major factors in halothane vaporization are the vaporizer setting, alveolar ventilation and halothane temperature. From Fig 1, after ten minutes' exposure with the vaporizer set on '1', blood and tissue uptake were virtually the same. After a further ten minutes with the vaporizer set at '2', halothane uptake was high since alveolar ventilation and cardiac output were near control levels. The REFERENCES Duncan W A M & Raventos J (1959) Brit. J. Anasth. 31, 302 Eger E I It & Guadagni N P (1963) Anesthesiology 24, 299 Kety S S (1951)Pharmacol. Rev. 3, 1 King N W & Broughton G (1967) Brit. J. Ana'sth. 39, 515 Larson C P, Eger E I II & Severinghaus J W (1962) Anesthesiology 23, 349 Mapleson W W (1962) Brit. J. Anarsth. 34, 11 Nunn J F & Hill D W (1960) J. appl.Physiol. 15, 383 Wolfson B, Ciccarelli H E & Siker E S (1966) Brit. J. Anesth. 38, 591 Professor J P Payne (Research Department of Anacsthetics, Royal College of Surgeons of England, London) Alcohol in Blood and Urine The use of a gas chromatograph combined with an internal standard and an integrator offers the most convenient method yet developed for the accurate measurement of ethyl alcohol concentration in biological fluids. In practice those fluids most commonly analysed are blood and urine and a suitable gas chromatograph consists of a moving phase source, an injection port and vaporizer, a stationary phase, a detector and a recorder. The moving phase is nitrogen, although other gases such as argon can also be used. The nitrogen which passes through the instrument at a constant rate purges the injection port and vaporizer of the volatile components of the sample to be analysed and carries them on to the stationary phase, a 4 ft (I 2 m) copper column of 025 in (6 mm) internal diameter packed with porous polymer beads of 80 to 100 mesh (Poropak Q) maintained at a temperature of 171°C. The various com- is Section of Ancesthetics ponents of the sample are selectively adsorbed on the column, to be shortly released in the inverse order of their affinity for the column material. After release they pass through a hydrogen flame detector where a complex ionization process occurs. As a consequence carbon atoms are given up in proportion to the amount of organic material present. These carbon atoms are counted as they pass through the detector and appropriate signals are transmitted to the recorder where they are transcribed in the form of a peak. The area under the peak is an accurate representation of the concentration and if the signals are fed into an integrator greater precision in measurement can be achieved. Technique Before analysis, the blood or urine sample is diluted with an aqueous solution of propanol (25 mg/100 ml) in the ratio of 1: 10; 2 V1 of the diluted sample is injected into the entry port of the chromatograph. Both ethanol and propanol are eluted and appear as separate peaks within five minutes of injection. The peaks are charted and the areas under each are computed and recorded by the integrator. Since the areas are directly proportional to the concentration, the ratio of the concentration of alcohol to that of propanol can be calculated. The dilution of the sample decreases the rate at which blood solids accumulate on the column; the presence of propanol provides an internal standard and eliminates the need to measure the injected volume precisely since the important factor is the ratio of alcohol to propanol. For calibration purposes a suitable concentration of ethyl alcohol is made up by weight in water. A sample of this standard is diluted with the aqueous solution of propanol and injected into the chromatograph exactly as the blood sample. When the ratio of alcohol present to that of propanol is obtained the alcohol concentration in the unknown sample can be derived as follows: concentration of known standard x ratio of unknown sample ratio of known standard concentration of unknown sample Duplicate analysis is usual. Typical chromatograms derived from the analysis of five separate samples from the same blood specimen are shown in Fig 1. With a mean level of 262 mg/100 ml the range is from 259 to 266 mg. 531 ETH. PROP. RATIO MEAN , -2t8 231 1 07Lt CONC. OF WATER 1 0725 STANDARD 273 255 1071 252mg/100mi 232 206 1-126 265 mg/100ml 277 252 1 099 259 312 280 1i114 262 314 278 1 129 266 315 285 260 BLOOD CONC OF ALCOHOL 1 105 . i MEAN 262 Fig 1 Typical gas chromatograms showing the extraction of ethanol from water and blood. The amount of alcohol in the water was known and forms the control. The remaining peaks were derived from five separate aliquots of the same blood sample injected consecutively quate and the handling technique is straightforward. The difficulties arise when attempts are made to calculate blood values from the results obtained on urine analysis. For this purpose a conversion factor of 1-33 has been recommended in the Report of a Special Committee of the British Medical Association (1965). But although Payne et al. (1966) demonstrated good agreement when this ratio was used in a laboratory study of the uptake, distribution and elimination of alcohol in healthy volunteers, it was emphasized that such a ratio was only valid when the alcohol level in urine had passed its peak and when the bladder had been emptied within the preceding thirty minutes. Furthermore their later work (Payne et al. 1967) and that of others (Morgan 1965, Stevens et al. 1966) showed that the ratio of 1-33 was not reliable outside the laboratory. In the Belfast investigation (Morgan, personal communication) 213 of the 518 accused motorists had a urine:blood ratio greater than 1 40, the overall ratio being 1 38. In the more detailed study carried out by Payne and his colleagues the urine:blood ratio was 1 -41 and 19 ofthe 35 suspects had a ratio greater than 1-33. These findings support the contention of Froentjes (1963) who, on the basis of a ratio of 1 52 obtained from more than 7,000 paired urine and blood samples, argued that a realistic ratio would be 2:1 for medico-legal purposes. Urine Analysis The analysis of urine presents no special problems; the difficulties of collection are minimal, the volume of the sample is usually more than ade- The significance of these observations lies in the fact that under the Road Safety Act of 1967, which has made it an offence to drive with more than 80 mg alcohol per 100 ml blood, 107 mg 532 Proc. roy. Soc. Med. Volume 61 May 1968 16 Table I Details of 12 volunteers with blood alcohol levels in the 50-80 mg/100 ml range Subject Age 1 29 43 2 20 3 28 4 23 5 28 6 7 49 8 28 46 9 10 51 33 11 18 12 Weight (kg) 86 88 70 70 67 73 60 71 71 60 83 69 First urine (U,) 159 99 100 93 91 156 139 82 141 89 140 109 Time interval (min) 45 21 18 17 30 60 22 26 19 24 65 Second urine (U2) 1540 90 91 79 15 106 88 1320 1300 76 1250 86 1210 Blood (B) 71 57 50 59 61 70 80 54 71 79 73 62 U2/B 2 17 1-58 1 82 1-34 1*44 1-88 1-63 1-41 1-76 1.09 1-65 1-71 0 Urine level above the statutory limit alcohol in 100 ml urine has been taken as the equivalent of 80 mg per 100 ml in blood. This, in effect, means that despite the evidence to the contrary the ratio of 1-33 has been accepted by Parliament and that in the interests of justice accused motorists will be well advised to insist on blood analysis. It has been implied that the high urine:blood ratios are a function of relatively high alcohol levels in blood and urine and that around the critical level, i.e. 80 mg alcohol per 100 ml blood, urine analysis is marginally more helpful to the accused driver than blood analysis (Enticknap 1967 a, b). These assertions were made on the basis of studies on eight subjects but the evidence is not convincing. Scrutiny of Enticknap's data reveals that not only was there no evidence that the urine values used to determine the urine :blood ratios had been collected after the peak concentration of alcohol had been reached but also that, even when second urine samples had been collected, no corresponding blood samples appear to have been obtained. The validity of Enticknap's ratios therefore must be seriously in doubt. In our own experience high urine:blood ratios are just as common when the alcohol level in the blood is low. Of 12 volunteers who had blood alcohol levels in the 50-80 mg range no fewer than 11 had urine:blood ratios greater than 1V33 (Table 1). Moreover, in 5 of the 12 volunteers the urine alcohol level was substantially above 107 mg/100 ml although the blood level was well below the statutory limit. In all 12 subjects the urine samples had been collected after the peak concentrations had been reached and in 9 the samples had been taken within thirty minutes of the earlier collection. The implication of these observations is serious. Almost certainly innocent men and women risk conviction if urine analysis is used alone, so long as the value of 107 mg of alcohol in 100 ml urine continues to be taken as the equivalent of 80 mg in 100 ml blood. Blood Analysis The analysis of blood is more complicated than that of urine. Controversy exists about the method of collection and discrepancies between the alcohol levels in capillary and venous blood samples collected simultaneously have added to the disquiet. Although by implication preference has been expressed for capillary blood samples collected either from the finger-tip or from the ear-lobe (Road Safety Legislation 1965-6) little formal guidance has been given about the method of collection. Many doctors favour the ear-lobe because of its dependent position and its plentiful blood supply but the ear stab can be embarrassing as well as painful for the suspect, and for this reason a finger prick is sometimes preferred. Unfortunately, finger blood flow is not uniformly adequate and even when the pulp of the finger is stabbed the blood flow may be meagre. Under these circumstances the temptation to apply pressure must be resisted, as the ensuing himatocrit disturbance will almost inevitably distort the results of analysis. Once the site of puncture has been decided, selection of a suitable skin sterilizing agent is important. Although in practice the use of an alcoholic solution, such as tincture of iodine or methylated spirits, is unlikely to lead to any significant effect on the alcohol concentration in the blood, there are theoretical objections which can be avoided by the use of a nonalcoholic antiseptic. The choice of a container for the blood sample presents further difficulties. With the accepted methods of collection capillary blood is exposed to the atmosphere for a varying period of time at body temperature and the possibility of loss of alcohol by evaporation has to be considered. This risk seems to be lowest when blood is allowed to flow freely into polypropylene plastic pots. Loss can also occur by a leak past the seal or by adsorption of alcohol on the materials used. 17 Section ofAnwsthetics Clotting is a particular hazard of capillary sampling, since the sample is usually too small to allow the use of liquid anticoagulants. In practice, the containers are usually coated with a heparin or fluoride-oxalate solution and allowed to dry. This has not proved entirely satisfactory and our own experience indicates that about 7 % of samples are clotted. One further source of error remains to be considered. It has recently been shown that when alcohol is added to blood withdrawn from volunteers, the alcohol content of the plasma is about 40 % higher than that of an equal volume of red blood cells (Payne et al. 1967). Since separation of plasma from red cells is not uncommon in blood collected in capillary tubes, the need for careful mixing before analysis is obvious. This observed difference between the alcohol content of plasma and red blood cells from the same sample has been confirmed in vivo in 20 volunteers given varying quantities of alcohol to drink (Payne, Hill & Wood 1968) and raises new problems of a more fundamental nature. When the distribution of alcohol taken by mouth is in equilibrium in the body, the tissue levels, including those in the lungs and in the kidneys, reflect the plasma concentration. In particular, the tension of alcohol in the alveolar gases is directly proportional to the plasma concentration. But since substantially more alcohol is present in plasma than in red cells, it follows that, for any given plasma concentration, the whole blood concentration must vary inversely with the number of red blood cells present. On this basis it seems reasonable to question the validity of blood:breath and urine:blood ratios. The potential sources of significant error encountered in the analysis of capillary blood by gas chromatography emphasize the need for a standard protocol for the withdrawal and handling of blood for the determination of the alcohol content. Until such a protocol is established there is every justification for insisting on the use of venous blood when blood alcohol determinations are required. Dr J B Enticknap (East Ham Memorial Hospital, London) said that there was not, in his view, any conflict between his results and those of Professor Payne, as his own were specifically obtained under field conditions in which it was most unlikely either that the bladder was emptied or the postabsorptive peak reached. These conditions most closely resembled those in which an actual arrest might be made at a low blood level and accounted for his findings of urine blood ratios as low as 0-3. He also felt that with more development there was a future for quantitative breath tests. REFERENCES British Medical Association (1965) The Drinking Driver. London Enticknap J (1967a) Brit. med.J. ii, 49 (1967b) Brit. med. J. iv, 11O FroentjesW (1963) In: Alcohol and Road Traffic. Ed. J D J Havard. London; p 179 Morgan W H D (1965)J. forens. Sci. Soc. 5, 15 Payne J P, Foster D V, Hill D W & Wood D G L (1967) Brit. med. J. iii, 819 Payne J P, Hill D W & King N W (1966) Brit. med. J. i, 196 Payne J P, Hill D W & Wood D G L (1968) Nature (Lond.) 217, 963 Road Safety Act (1967) HMSO, London Road Safety Legislation 1965-6 (1965) Cmnd 2859. HMSO, London; p 12 Stevens P J, Mason J K & Bowden C H (1966) Med. Sci. Law 6,96 533 Professor Payne denied that breath analysis had any place in the determination of blood alcohol concentrations. In his opinion breath analysis provided a useful screening test but nothing more. Dr B M Wright (National Institute for Medical Research, Mill Hill, London) said that Professor Payne's view that breath analysis was of no practical value was not shared by any other worker in the field. It was based on experiments carried out in his laboratory, using a method that had not been used before for this purpose, which gave results different from those of other workers, many of whom had obtained excellent correlation between breath and blood alcohol. Professor Payne's observations about the differing alcohol content of plasma and red cells were interesting but the relevant question was the relationship between plasma and whole blood. Available evidence suggested that the alcohol content of plasma was only about 10 % more than that of whole blood. The suggestion that the plasma-red cell difference could account for variations in the breath/blood relationship was fallacious, because what mattered in this case was not the actual w/v concentration but the partial pressure of alcohol, which was uniform through the blood. There was no doubt that substantial and unexplained variations in the breath/blood ratio did occur but, nevertheless, the only published evidence (Wright, 1965, Brit. med. J. ii, 1430) suggested that in practice breath analysis with a reliable instrument such as the 'Breathalyzer' gave as accurate an estimate of blood alcohol as blood analysis under operational conditions. Professor Payne admitted that the variations in blood/breath ratios reported by his own group might not be acceptable to Dr Wright who had developed his own method of measuring alcohol in breath; nevertheless their results had never been publicly challenged. While it was true that the technique of infra-red analysis which had been used to determine the alcohol content of breath was not the usual method, it was well recognized for its reliability and accuracy; if it had not been 534 Proc. roy. Soc. Med. Volume 61 May 1968 18 previously used in alcohol studies by other workers, this was surely a reflection on them rather than on the method. There was no virtue in using a less reliable method simply because everyone else used it. Dr Wright was being naive in talking about the many workers who had obtained excellent correlation between breath and blood alcohol levels. Certainly a few groups of workers had obtained a reasonable measure of agreement but this was scarcely surprising since they all had the same built-in error. Professor Payne questioned the reliability of Dr Wright's statement that the alcohol content of plasma was only about 10 % more than that of whole blood. Published papers gave values for plasma between 15 and 25% higher than whole blood, and these were broadly in agreement with the 18 % value obtained by his own group (Payne et al. 1968, Nature (Lond.) 217, 963). Dr Wright's comment about the plasma/red cell difference implied a lack of understanding of the problem. For any given blood concentration of alcohol the plasma concentration will vary directly with the hmematocrit. Since the alveolar tension of alcohol reflects the plasma tension which in turn is directly related to the plasma concentration, it follows that different alveolar tensions can be obtained for the same blood level. (1960, J. appl. Physiol., 15, 383) who determined this in man. It was felt that using this assumption was more accurate than using minute ventilation which would have resulted in artificially higher values of halothane uptake by the blood. Dr R M A McClelland (King's College Hospital Medical School, London) asked whether Dr Cervenko's figure of 0-68 for alveolar ventilation ratio was an assumption based on human work. Dr Cervenko replied that it was. The value used to calculate alveolar ventilation from minute ventilation in the dog had been taken from Nunn & Hill Dr Gavin Robinson (Glasgow) asked if levels of halothane in skeletal muscle had been measured. Dr Cervenko said that this had not been done. The President asked why carbon tetrachloride had been used as a solvent for halothane. Dr Cervenko replied that, using the chromatographic column described, several extracting solvents had been tried. Carbon tetrachloride was found to be the one with the best characteristics, namely rapid elution with little tailing, good separation from ether and halothane, low water solubility and economy. The President asked Professor Payne whether capillary or venous blood gave the greater advantage to the accused motorist. Professor Payne replied that the choice of site would depend on a number of factors including the stage of absorption and the state of the circulation. He pointed out that, after uptake had ceased, the arterial, capillary and venous blood concentrations of alcohol were virtually the same but, while alcohol was still being absorbed, the peripheral venous blood usually contained less than capillary and arterial blood, especially if the peripheral blood flow was reduced, for example in arterial disease of the lower limbs.