More importance has recently been attached to the pharmacokinetics of ethanol during ethanol monitoring, which is a non-invasive method for indicating and quantifying the absorption of irrigating fluid in patients undergoing endoscopic procedures such as transurethral resection of the prostate [1,2] and transcervical resection of the endometrium . By adding a tracer amount of ethanol to the irrigating fluid, the volume of fluid absorbed can be estimated from the amount of ethanol measured in the patient's exhaled breath. The ethanol data allow determination of the volume of irrigant absorbed by using a nomogram. The 90% prediction interval is about ±300 mL around the regression line (Fig. 1), which might be clinically acceptable but still represents a considerable uncertainty.
The purpose of the present study was to evaluate the importance of central haemodynamic changes on the results of the breath analysis. We hypothesized that changes in the cardiac output (CO) would be associated with changes in the results of the breath test because ethanol would be distributed to the tissues at different rates. The investigation was performed using a Doppler-ultrasound technique and resting male volunteers.
Subjects and methods
Nine healthy male volunteers (age 34 ± 7 years, body weight 82 ± 10 kg; mean ±SD) underwent two experiments consisting of an intravenous (i.v.) infusion of ethanol containing irrigating fluid. After approval by the Local Ethics Committee, each volunteer gave his informed consent. The subjects ate only a light meal before the experiments. Cannulae were inserted into the antecubital veins of both arms and irrigating fluid consisting of glycine 1.5% and ethanol 1% or mannitol 3% and ethanol 1% (Baxter Medical AB, Kista, Sweden), which are the two most widely used ethanol containing irrigating solutions, was infused into one vein at a constant rate of 15 mL kg−1 over 20 min with the aid of infusion pumps (IVAC 560, San Diego, CA, USA).
The central haemodynamic reponses were assessed with a transthoracic ultrasound-Doppler device (CFM 700, VingMed Co., Horten, Norway) applied in a left parasternal position, using a combined two-dimensional and 3-MHz Doppler transducer. Stroke volume was calculated as the product of the velocity-time integral and the aortic cross-sectional area. CO was obtained from the product of stroke volume (SV) and heart rate, obtained using the ultrasound-Doppler device. The values reported here are the mean of six measurements made by one clinical physiologist. Half of these were determined immediately before the infusion began and the others just after it was completed. The coefficient of variation of a single measurement is 4% .
Pharmacokinetics of ethanol
The concentration of ethanol in the expired breath was measured with an Alcolmeter S-D2 (Lions Laboratories, Glamorgan, Wales) at 0, 2, 6, 12, 20, 26, 32, 40, 50 and 90 min. The Alcolmeter was calibrated with an alcohol-in-gas standard before each experiment.
The data were analysed using a pharmacokinetic model that allows separate evaluation of the distribution and the elimination parts of the breath ethanol concentration profile. The ethanol levels were assumed to equal the sum of a first-order distribution function and a zero-order saturated Michaelis–Menten elimination function. The following equation was used to express the ethanol concentration (C) at any time (t) after a bolus injection of ethanol : Equation 1 where α is the first-order exponential distribution rate constant, and β is the zero-order elimination rate constant. A and B are the concentrations of ethanol obtained when the first-order and the zero-order functions are extrapolated back to t = 0 (Fig. 2). The data were fitted to this equation by using the MPK non-linear regression program for digital computers (McPherson Scientific, Rosanna, Australia) after considering the effect of an i.v. infusion on the model parameters as follows : Equation 2 where T is the infusion time. Ethanol values below 0.05 mg L−1 were discarded to avoid including the unsaturated Michaelis–Menten elimination kinetics in the calculations. No weighting factor was used as the error in the ethanol analysis is fairly constant within the concentration range studied.
The central distribution volume (Vdc) and the total distribution volume for ethanol in the body (Vdextrap) were calculated from the dose of ethanol (Xd) and the blood/breath partition coefficient (2300) for ethanol. Thus: Vdc = 2300 Xd/(A + B) and Vdextrap = 2300 Xd/B.
Of the 18 infusions performed, the 13 experiments were included in which all the pharmacokinetic parameters (A, α, B and β) could be calculated with reasonable confidence, in this study with an SD of ≤20% of the estimate.
The results are expressed as means with the standard deviation (SD). The area under the curve (AUC) was calculated by the linear trapezoid method and extrapolated to infinity. The relations between the variables were evaluated by multiple and simple linear regression and were considered statistically significant at P<0.05.
The data on the pharmacokinetics of ethanol (Table 1) were compared with the measurements of SV (88 ± 16 mL), CO (6.10 ± 1.0 L min−1) and heart rate (67 ± 7 beats per min) obtained using the ultrasound-Doppler.
The maximum ethanol concentration (Cmax), which was determined at the end of each infusion, increased with SV and with CO (Fig. 3, upper). The coefficient of determination (r2) indicated that CO could account for 53% of the variation in Cmax. Multiple regression analysis was then used to test for the relation between the pharmacokinetic model parameters A, α, B and β and each of the haemodynamic variables. There was an inverse correlation between the distribution rate constant (α) and both CO and SV (Fig. 3, lower). The coefficient of determination (r2) indicated that CO could account for between 50% and 60% of the variation in the kinetic parameters used to describe the concentration-time profile of ethanol.
The same significant relations were noted after corrections for body size, viz. Cmax was divided by the actual dose given and SV and CO were divided by body surface area (stroke volume index was 43 ± 6 mL beat−1 m−2 and cardiac index 3.0 ± 0.5 L min−1 m−2).
The values of A and AUC also increased with CO, but only the latter relation attained statistical significance (r = 0.66, P<0.02). No correlations were found between pharmacokinetic variables and the heart rate and the choice of irrigating fluid (glycine or mannitol).
Finally, pooled data was analysed . The mean concentration-time profiles of the patients having a CO of <6L were compared with those with a CO of ≥6L. The pharmacokinetic model parameters A, α, B and β could be determined with an average SD of 9%, and only the values of α were different (0.31 min−1 for CO<6L, and 0.23 min−1 for CO ≥ 6L).
The present study confirms that central haemodynamic changes are relevant to the ethanol concentration in the expired breath following an i.v. infusion of irrigating fluid containing a tracer amount of ethanol. A hyperkinetic circulation was associated with a higher Cmax and a lower value of the distribution rate constant (α). As the half-life of the distribution phase is determined by ln2/α, the distribution of ethanol is slower when cardiac output is high.
The presented relations were somewhat unexpected, as a high cardiac output transports the ethanol more rapidly into the tissues. The distribution of alfentanil, for example, is faster when the cardiac output is high . However, tissues reached by blood containing ethanol become saturated very quickly, and the concentration gradient between Vdc and Vdextrap is between well perfused and poorly perfused anatomical regions rather than across the cell membranes .
The present results are best understood by assuming that a high cardiac output reduces the time for the blood to reach and to return from well perfused tissues. The ethanol content of the venous blood that returns to the heart is then more similar to that in the blood that just left the lungs. This acts to prolong the apparent rate of distribution to more poorly perfused tissues. Hence, Cmax and T1/2α increase. A high cardiac output also reduces the transit time for blood in the capillaries which, despite the ease by which ethanol saturates the tissues, decreases the extraction ratio for ethanol. The same effect is induced by local vasodilatation, which promotes a marked distribution function by increasing A and/or T1/2α .
An alternative explanation is that a high cardiac output is associated with a low V˙/Q˙ ratio, which gives higher breath ethanol concentrations for the same blood ethanol concentration. However, the blood/breath ratio for ethanol increased when females were given isoflurane anaesthesia, which lowers cardiac output, when compared with the same females in the awake state .
The ultrasound-Doppler used in the present study has certain shortcomings. It is based on the determination of the aortic cross-sectional area, which varies through the cardiac cycle. To control this error, the aortic annulus is used, as its dimension appears to be more stable than the aorta itself. Another problem is that the anatomical cross-sectional area is larger than the physiological area. However, this has little practical consequence as the ultrasound-Doppler measurements show excellent correlations with invasive measurements of cardiac output .
When ethanol is used to monitor fluid absorption during surgery, the absorption volume and the rate at which absorption takes place are regarded as the most important factors determining the result of the breath test . However, as shown in the present study some of the variation in the breath ethanol level can also be accounted for by haemodynamic factors. The aim of monitoring is to enhance patient safety by allowing decisions made by the operating team to be based on updated information about intraoperative absorption events. The risk of post-operative symptoms increases progressively with increasing amounts of absorption [1,2], and therefore the monitoring also makes it possible to choose the most suitable level of care for the follow-up period.
The pharmacokinetics of ethanol is often thought to be adequately described by a one-compartment open model with Michaelis–Menten elimination kinetics , but an exponential distribution function clearly becomes important when alcohol is given rapidly [12,13]. The analysis performed here distinguished the distribution and the elimination parts of the ethanol concentration-time profile . The model used assumes at all times a fully saturated kinetic system for the elimination of ethanol from the blood. This is not unreasonable considering that the metabolism of ethanol occurs at a maximum rate when the breath ethanol concentration is only 0.05 mg L−1.
The principle of 'the breathalyser' is that ethanol concentrations in the exhaled gas reflect the concentrations in the blood: 0.1 mg L−1 in the expired air corresponds to 0.23 g L−1 in the blood. This blood/breath ratio was used to convert the kinetic parameters A and B to the corresponding blood ethanol concentration. This allowed meaningful volumes of distribution to be calculated (Table 1). For example, these calculations confirm that Vdextrap amounts to 52% of the body weight, which corresponds well with the total body water of middle-aged men.
Other parameters differ slightly from previous analyses of ethanol kinetics using this model. The central distribution volume Vdc was about 8 L, which is smaller than previously reported [9,10]. The distribution halflife (T1/2α) of 2.4 min should be compared with the 4.9 min found in other middle-aged men . Finally, the terminal elimination rate (β) was slightly higher than in previous studies with awake volunteers [5,9,10]. These differences might be explained by the fact that ethanol was infused with a fairly large amount of fluid. Furthermore, the lack of an overnight fast results in a higher β, probably because food increases the splanchnic blood flow .
In conclusion, we found that the maximum breath ethanol level increased with CO during an i.v. infusion of irrigating fluid containing ethanol.
Dr Ivar Randmaa, Department of Clinical Physiology, Huddinge University Hospital, performed the measurements of cardiac output.
1 Hahn RG. The transurethral resection syndrome. Acta Anaesthesiol Scand
2 Hahn RG. Ethanol monitoring
of irrigating fluid absorption. Eur J Anaesth
3 Olsson J, Hahn RG. Ethanol monitoring
of irrigating fluid absorption in transcervical resection of the endometrium. Acta Anaesthesiol Scand
4 Janerot Sjöberg B, Wranne B. Cardiac output
determined by ultrasound-Doppler: clinical applications. Clin Phys
5 Hahn RG, Norberg A, Jones AW. Rate of distribution of ethanol into the total body water. Am J Therapeut
6 Wagner JG. Linear pharmacokinetic equations allowing direct calculation of many needed pharmacokinetic parameters from the coefficients and exponents of polyexponential equations which have been fitted to the data. J Pharmacokin Biopharm
7 Martin E, Moll W, Schmid P, Dettli L. The pharmacokinetics
of alcohol in human breath, venous and arterial blood after oral ingestion. Eur J Clin Pharmacol
8 Henthorn TK, Krejcie TC, Avram MJ. The relationship between alfentanil distribution kinetics and cardiac output
. Clin Pharmacol Therapeut
9 Norberg A˙, Jones AW, Hahn RG. Pharmacokinetics
of ethanol in arterial and venous blood and in the breath during vasconstriction and vasodilation. Am J Therapeut
10 Olsson J, Hahn RG. Analysis of ethanol in the expired air during low-flow isoflurane anaesthesia. Br J Anaesth
11 Wilkinson PK, Sedman AJ, Sakmar E, Earhart RH, Weidler DJ, Wagner JG. Blood ethanol concentrations during and following constant-rate infusion of alcohol. Clin Pharmacol Therapeut
12 Vestal RE, McGuire EA, Tobin JD, Andres R, Norris AH, Mezey E. Aging and ethanol metabolism. Clin Pharmacol Therapeut
13 Rangno RE, Kreeft JH, Sitar DS. Ethanol 'dose-dependent' elimination; Michaelis-Menten V classical kinetic analysis. Br J Clin Pharmac
14 Hahn RG, Norberg A˙, Gabrielsson J, Danielsson A, Jones AW. Eating a meal increases the clearance of ethanol given by intravenous infusion. Alcohol Alcoholism