- Email: [email protected]
THE VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE IN THE RABBIT
Brit. J. Anaesth. (1972), 44, 656
THE VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE IN THE RABBIT W. W. MAPLESON, P. R. ALLOTT AND A. STEWARD
SUMMARY
In the course of experiments designed to compare in-vivo and in-vitro partition coefficients for halothane, results were obtained for six tissues in each of six New Zealand White rabbits. For reasons given in reporting the results of a similar study with nitrous oxide and cyclopropane (Mapleson, Evans and Flook, 1970) it is of value to examine the variability of the results, in order to determine to what extent it is possible to estimate the partition coefficient for a particular tissue in a particular individual, without actually measuring it in-vivo. METHODS
The details of the experimental method are given in a companion paper (Steward, Mapleson and Allott, 1972) and in Allott, Steward and Mapleson (1971). In brief, the rabbits were anaesthetized in such a way as to maintain the end-tidal concentration constant at some level in the range of 0.84 to 0.92 per cent halothane. At the end of about five hours the animals were sacrificed and samples of blood, brain, heart, kidney, liver and skeletal muscle were removed for determination of their halothane content, and hence their in-vivo tissue/gas partition coefficients. Further samples of each tissue were homogenized with saline and equilibrated in a rotating tonometer with similar concentrations of halothane to yield in-vitro tissue/gas partition coefficients. The statistical techniques used to analyse the results were taken from Davies (1957).
RESULTS
The full set of observed tissue/gas and blood/ gas partition coefficients is given in table I. For reasons given in the companion paper, all the main statistical calculations were performed on the logarithms of the coefficients. A three-way cross-classification analysis of variance in the companion paper separated out the several sources of variation as follows. (i) There is no significant difference between the in-vivo and in-vitro techniques overall, but there is a significant tissue-technique interaction. As was shown in the companion paper this was due to significant differences between techniques for brain and heart with no significant differences for the other four tissues. (ii) There are significant differences between tissues, even when referred to either the tissue-technique or tissue-animal interaction (see below). Therefore it is worth calculating separate means for each tissue. (iii) There is a significant tissue-animal interaction. This means that there are real differences between animals (greater than can be attributed to experimental error) but these differences are not systematic over all the tissues. That is to say that, if a given animal has a higher-than-average coefficient for one tissue, then the coefficients for the other tissues are not, in general, equally far above the averages for W. W. MAPLESON, PH.D., F.INST.P.; P. R. ALLOTT, B.SC.;
A. STEWARD, B.SC.; Department of Anaesthetics, Welsh National School of Medicine, Cardiff, CF4 4XN.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Partition coefficients were determined for halothane in various rabbit tissues after about five hours anaesthesia. After correcting for the fall in haematocrit which occurred during the experiments the mean blood/gas coefficient was 3.94. Except in the case of skeletal muscle the tissue/blood coefficients differed significantly from unity: brain 1.55, heart 2.74, kidney 1.75 and liver 2.26. The differences between the tissues were highly significant. Differences between animals were large but mainly unsystematic. The 95% confidence limits for tissue/gas coefficients in the population were + 32% —24%. There was also variation between individual 1- to 2-ml samples of the same tissue: the 95% confidence limits for such samples were of the order of +20% of the mean for liver, and +50% of the mean for heart, kidney and muscle.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE TABLE I.
657
Tissue/gas partition coefficients for halothane at 37.5°C—individual observations.
Kidney Muscle Liver Heart Brain vivo vitro vivo vitro vivo vitro vivo vitro vivo vitro 12.3 10.4 10.0 10.1 11.4 13.8 1 3.6 4.0 6.3 5.8 8.7 6.0 _ 7.4 7.1 2 8.4 2.8 3.0 3.7 6.0 _ 7.0 4.1 7.4 3 6.8 2.8 2.6 2.7 12.0 12.5 6.7 5.9 2.8 9.1 7.6 6.5 4 3.0 4.8 3.5 10.4 11.3 11.5 5.9 6.8 3.6 5.4 8.1 7.6 3.8 _ 10.3 3.9 10.8 10.4 5.5 6.2 3.9 5 8.8 6.7 2.9 6 4.3 3.8 13.4 10.4 6.1 5.8 3.4 Brain and heart in-vivo coefficients have been corrected on the assumption that, on the average, the in-vitro coefficients provide the better estimate. If the alternative view, that the in-vivo coefficients are the more reliable, is taken then all the brain coefficients should be divided by 1.282 and all the heart coefficients should be multiplied by 1.500 (see text). Rabbit No:
Blood vivo vitro 3.6 3.8
respect of heart and brain for which there were significant differences between the results of the two techniques. In the companion paper it was shown that these differences may have been real, in which case the in-vivo coefficients are to be preferred— except that the heart samples contained some surface fat and therefore gave a coefficient higher than that to be expected for cardiac muscle. However, the companion paper also showed that the difference might instead be attributable to experimental errors in such a way that the in-vitro coefficient is to be preferred for both heart and brain. We have chosen to prefer the in-vitro coefficients. Therefore, with heart and brain, we have corrected the individual in-vivo coefficients by the mean of the vitro-vivo differences for each of these tissues. In case the invivo coefficients should be preferred the appropriate corrections are given in the footnote to table I. Details of the statistical methods used are given in TABLE II. Partition coefficients (tissue/gas and tissue/ the appendix and the results in table II. blood) for halothane at 37.5°C—means and probable limits. Table II also includes a "corrected" set of tissue/ 95% confidence limits blood coefficients the basis for which is as follows. of mean of population The haematocrit fell during the experiments perhaps (as % of mean) Tissue/gas Mean owing, at least in part, to the infusion of Hartmann's 3.72 2.83 3.24 Blood solution; the extent of the fall ranged from 1.7 to 6.98 6.09 5.31 Brain 10.79 Heart 12.35 9.44 10.5 per cent. There was a highly significant -24% +32% Kidney 7.84 5.94 6.82 (P=0.005) linear regression of blood/gas 7.83 10.29 Liver 8.98 partition coefficient (measured on samples drawn at 4.13 3.60 3.14 Muscle the end of the experiment) on this fall in haematoTissue/blood crit. The regression coefficient (-0.125/per cent 2.13 Brain 1.88 1.65 3.77 2.93 Heart 3.33 haematocrit fall) was used to calculate for each 1.87 2.43 Kidney 2.13 -20% +25% animal what the blood/gas coefficient would have 2.40 3.11 Liver 2.73 been at the beginning of the experiment, before 1.27 1.11 0.98 Muscle any haemodilution had occurred. These corrected Tissue/blood (corrected*) blood/gas coefficients (mean 3.94) should be more 1.55 1.34 1.78 Brain 3.16 2.74 2.38 Heart representative of normal values in the rabbit than 1.75 1.51 2.02 Kidney -24% +32% those which we happened to measure after varying 2.26 1.95 2.61 Liver degrees of haemodilution. They were therefore used 1.06 0.92 0.79 Muscle to calculate a corrected set of tissue/blood coeffi"Using blood/gas coefficients (mean 3.94) corrected for cients. fall in haematocrit—see text.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
those tissues. Therefore the variability of each tissue must be assessed separately. (iv) The between-animal variation is significant, even when referred to the tissue-animal interaction. This means that, despite the above remarks, a high coefficient for one tissue in an animal does to some extent imply high coefficients for the other tissues. This last finding suggests that, in an individual, tissue/gas coefficients for halothane may be predictable to a useful extent from the more easily measured blood/gas coefficient. Therefore the results are expressed in terms of tissue/blood coefficients as well as of tissue/gas coefficients (table II). In the light of the other foregoing findings the results, for each type of coefficient, are expressed in terms of the mean coefficient for each tissue with confidence limits of the mean and of the population. In calculating the means a problem arises in
658
BRITISH JOURNAL OF ANAESTHESIA
DISCUSSION
The present results. The range of variation of partition coefficients between tissues (table II) is much greater than for nitrous oxide or cyclopropane (Mapleson, Evans and Flook, 1970). The variation is by a factor of more than 3:1 instead of only 1.3:1 or 1.7:1. On the other hand, the range of variation between animals is of the same order as that found in the study of cyclopropane and nitrous oxide. This may in part be due to the use of a single strain of rabbit, New Zealand White, in the halothane study. As was the case with nitrous oxide, none of the variation in the last pair of columns in table II should be attributable to experimental error since the statistical treatment has allowed for this. The results for nitrous oxide and cyclopropane showed a small variance for blood and a large one for muscle compared to those for the other tissues; such differences were not evident in the present results except that the variance for the corrected blood/gas coefficients was small (appendix). Tissue/blood coefficients influence the speed of equilibration of a tissue with the arterial tension, higher coefficients giving slower equilibration. It is notable that all the tissue/blood coefficients in table II are significantly greater than unity except those for muscle. Since the confidence limits of the population of uncorrected tissue/blood coefficients are only slightly
closer than those for tissue/gas coefficients, the suggestion made earlier in this paper that measurement of the blood/gas coefficient in an individual might provide a worthwhile guide to the tissue/gas coefficient is not verified to any useful extent. Unlike the study with nitrous oxide and cyclopropane this investigation yielded an estimate of the variation between adjacent portions (1 to 2 ml) of apparently homogeneous tissue. The magnitude of the variation is remarkably wide (see above) and suggests that, in order to obtain a reliable coefficient for a given organ in a given animal, measurements should be made on a large sample, if not the whole organ. It is shown in the statistical appendix that this variation between samples contributed little or nothing to the variation between individuals shown in table II. Comparison with the results of other workers. Variation between individuals. There is very little information available on this source of variation. For instance, Han and Helrich (1966) give only mean values; Laasberg and Hedley-Whyte (1970) give only means and standard errors of the means, and, although Larson, Eger and Severinghaus (1962) refer to means and standard deviations, it is evident that these refer to experimental error and not to variation between individuals: samples from different individuals were first pooled before performing repeat determinations on the collective sample. On the other hand, Wortley and associates (1968) give a histogram of blood/gas solubility coefficients for 33 surgical patients and from their data we have calculated the 95 per cent confidence limits for the variation between individuals to be +17 per cent of their mean of 2.3. These limits are closer than those of —24 per cent to +32 per cent for the rabbits (table II). Therefore, as with nitrous oxide, it appears that the rabbit may be more variable than man even though, in the present study with halothane, a more homogeneous strain of rabbit was used. The complex effects of proteins on the solubility of halothane have been discussed recently by Laasberg and Hedley-Whyte (1970). Han and Helrich (1966) and Cowles, Borgstedt and Gillies (1971) give equations for the regression of halothane solubility in blood on haemoglobin concentration which are similar to each other but are somewhat in conflict with the data of Laasberg and Hedley-Whyte. If these equations are valid they could be useful in predicting blood solubility coefficients for individuals,
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
In an early series of experiments two in-vivo samples of each tissue were taken directly from the animal to the extraction phials; in the present series, for most tissues, a single large sample was crushed in liquid nitrogen and the fragments well mixed before being divided into an in-vitro and two in-vivo samples. Although care was taken to obtain apparently similar duplicate in-vivo samples in the early series, the variances between the in-vivo phials were six to nine times larger than in the present series. Since there was no systematic difference between the two series in the variance between duplicate in-vitro phials, the in-vivo between-phials variance in the early series provides an estimate of the variation between adjacent 1- to 2-ml portions of a given tissue. From these variances we estimate the 95 per cent confidence limits for a single 1- to 2-ml portion of tissue to be of the order of ± 20 per cent of the mean for liver and +50 per cent for heart, kidney and muscle.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE TABLE III. Liquid/gas and tissue/gas partition coefficients at 37 °C.
Fluid or Tissue Water
Species
0.9% NaCl solution Oil
Olive
Plasma
Man
Erythrocytes Blood
Man Man
Ox Dog
Whole Brain
Rabbit
t*4.03
Man
6.0 §5.2 4.8 *6.2 5.4 8.3 *11.0 3.6 3.5 *7.0 6.0 4.2 *9.2 8.0 7.0 *3.7 138
Ox
Rabbit Grey Matter White Matter Heart Kidney
Man Man
Rabbit Man Ox
Rabbit Liver
Man Ox
Rabbit Skeletal Muscle
Man Ox
Fat
Man
(1) (6) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1)
0.9 0.6 2.7 0.7 0.9 0.6 0.5 0.9 0.8 0.8 0.9 0.6 0.7 0.9 0.5 0.6 0.9
0.05
•Present study corrected to 37 °C. $Temperatures corrected to 37 °C. fCorrected for fall in haematocrit (see text). The references indicated by numbers in parentheses are as follows: (1) Larson, Eger and Severinghaus, 1962; (2) Duncan, 1963; (3) Laasberg and Hedley-Whyte, 1970. (4) Saidman et al., 1966; (5) Allott, Steward and Mapleson, 1971; (6) Han and Helrich, 1966; (7) Wortley et al.3 1968; (8) Lowe, 1964; (9) Cowles, Borgstedt and Gillies, 1971; (10) Lowe, 1968; (11) Lowe and Hagler, 1969; (12) Okuda, 1968.
although there would still be a considerable uncertainty if Laasberg and Hedley-Whyte are correct in attributing greater importance to the albumin-globulin ratio. Furthermore the possibility of changes of solubility coefficient consequent upon the infusion of liquids into the circulation must be borne in mind. Variation between tissues and species. More information is available about mean solubilities in various fluids and in the tissues of various species. Table III gives the means of the present results and those of other workers. Since most determinations have been made at 37 °C all data for other temperatures (including our own) have been corrected by using the temperature coefficient of —4.6%/°C derivable from Eger, Saidman and Brandstater (1965). The smaller solubility of halothane in saline and Krebs-Henseleit solution than in water is doubtless a reflection of the salting-out effect of electrolytes. There is some evidence of deviation from Henry's law: the data of Saidman and associates (1966), (ref. 4 table III) indicates an increase in the solubility of halothane in water with an increase in concentration. However, the effect appears to be small and most of the concentrations used in determining the solubilities in table III were within a narrow range. The tissue values found by Lowe (1968) for man and Lowe and Hagler (1969) for man and calf are not included in table III because the measurements were made on very small samples, up to 25 /A, drawn apparently from single individuals. Therefore, although the values generally differ from those given in table III, this is likely to be merely a reflection of variation between adjacent portions of tissue in the same individual and between different individuals. Some authors have assumed that there is no species variation in partition coefficients for particular tissues, but the data given in table III show that this assumption is not generally justified. Although the coefficients for man and calf appear to be fairly similar, as do the coefficients for whole brain in man and rabbit, the blood coefficients for dog and rabbit and the liver coefficients for the rabbit are all significantly higher than the corresponding human values. Further, the coefficient for kidney in the rabbit is very much higher than in man while that for skeletal muscle is very much lower. The high muscle/blood coefficient in man (about 3 5 compared with 0.9 in the rabbit) means that the time required for skeletal muscle to equilibrate with a constant arterial tension will be much longer in
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Krebs-Henseleit solution
Approximate Gas CoeffiConcencient tration % 0.74 (1) 1.3 _ 0.79 (2) 2.1 0.78 (3) 1 0.87 (4) 1.04 (4) 10 0.89 (10) — 0.86 (11) 0.82 (12) 0.5 0.70 (1) 1.3 0.9 §0.77 (5) 2.2 0.76 (3) 0.03 224 (1) 2.28 (3) 1.2 §3.14 (6) 3.5 §1.26 (6) 4.7 2.3 (1) 1.0 — 2.5 (2) 2.42 (3) 1.1 3.8 §2.41 (6) 0.7 2.3 (7) — 2.42 (8) 0.7 §2.56 (9) — 2.24 (10) — 2.56 (11) 1.0 2.3 (1) 2.87 (11) 0.7 §3.18 (9) 2.00 (12) 0.5
659
BRITISH JOURNAL OF ANAESTHESIA
660 man than the five hours found adequate (see companion paper) in the rabbit. CONCLUSIONS
In man and dog the uncertainty over the blood/ gas coefficient in an individual may be reduced if the haemoglobin concentration or haematocrit is known. Mean coefficients in one species are not a reliable guide to those in another.
Effects of variation between individual 1- to 2-ml samples of tissue on the estimate of between-animal variation. STATISTICAL APPENDIX In the case of brain and blood the in-vivo and in-vitro samples were removed from the animal separately; thereMissing values. fore any variation between individual samples would For the purpose of the three-way and the two-way contribute to the experimental error and hence not to breakdown-by-technique analysis of variance gaps on the the calculated between-animal variation. In the case of basic data were filled as explained in the companion liver the in-vivo and in-vitro samples came from a mixpaper. ture of fragments of a single 20-ml sample. Therefore variation between such samples would not contribute to Estimation of the means and probable limits of the experimental error but would contribute to the calculated partition coefficients. between-animal variation. However, variation between Two-way breakdown-by-technique analyses of variance 20-ml samples of an 80-ml liver would be much less than showed that the in-vivo and in-vitro results were about the +20 per cent variation found between 1- to 2-ml equally variable and therefore both were used. samples, and therefore the contribution to the betweenHierarchical analyses of variance for each tissue are animal variation, of +32% —24%, would be small. In summarized in table IV. Before performing these analyses the case of heart and kidney the whole organ was sampled the brain and the heart in-vivo coefficients were corrected and mixed so that variation between samples does not by the respective mean vitro-vivo difference as described arise. Similarly a complete muscle was sampled; therefore in the main text. The residual degrees of freedom were the between-animal variation of +32%—24% is an untherefore reduced by one in each of these two tissues. biased estimate of the variation between animals for the Application of the Box (1949) modification of Bartlett's coefficient for that particular muscle although it may be an overestimate of the variation between animals for the TABLE IV. Summary of hierarchical analyses of variance mean coefficient of all muscles. for each tissue (~ Residual Mean square Blood Brain Heart Kidney Liver Muscle
0.00118 0.00136 0.00192 0.01145 0.00191 0.00330
d.f 5 4 5
4 5 5
Between-animal Mean square d.f. 0.00586 0.00034 0.01564 0.01546 0.00981 0.01039
5 5 5 5 5 5
ACKNOWLEDGEMENTS
We are grateful to Professor William W. Mushin, Head of this Department, for many helpful discussions and for providing the conditions so necessary for the conduct of research of this nature. We are also indebted to Imperial Chemical Industries Ltd for generous supplies of halothane. The work was supported by a grant from the Medical Research Council.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Individual in-vivo partition coefficients for halothane in at least blood, kidney, liver and muscle can be reliably estimated from in-vitro measurements, provided that large samples are used. The coefficients for different tissues are very different from one another and the mean coefficient for a tissue in the rabbit population is a very important guide to the coefficient for that tissue in a particular rabbit. Variation between individuals for a given tissue is less than the variation between tissues but is still considerable except perhaps for blood. It may be marginally less in terms of tissue/blood coefficients and may be smaller in man than in the rabbit. In our experiments the variation of the blood/gas coefficient between individual rabbits was largely attributable to varying degrees of haemodilution.
(1937) test showed that the residual mean squares were homogeneous (P=0.13) but that the between-animal mean squares were not (P=0.03). The inhomogeneity is due to the low value for brain. Since a similarly low value was not found with either nitrous oxide or cyclopropane (Mapleson, Evans and Flook, 1970) and since the level of significance is not very high it seemed preferable to treat this as a chance event. Therefore, both the residual and the betweenanimal mean squares were pooled for all tissues. The variance of the mean coefficient for each tissue was then taken as the sum of, firstly, the experimentalerror variance estimate (which equals the pooled residual mean square) divided by the number of measurements used in deriving the mean and, secondly, the betweenanimal variance estimate (which equals the difference between the pooled between-animal and pooled residual mean squares divided by the mean number of measurements per animal) divided by the number of animals used. The number of degrees of freedom was taken as the sum of those of the two variance estimates. In assessing the range of values likely to be found in individuals of the species the between-animal variance estimate alone was used. This could not be done with the corrected blood/gas coefficients because the between-animal mean square was less than the residual mean square. Therefore all that can be said of the 95 per cent confidence limits of the population of corrected blood/gas coefficients is that they are close to the mean. For tissue/blood coefficients individual values were determined by dividing each tissue/gas coefficient by the corresponding blood/gas coefficient for the same technique in the same animal. Means and confidence limits were then determined as for the tissue/gas coefficients.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE REFERENCES
LA VARIABILITE DES COEFFICIENTS DE PARTITION POUR L'HALOTHANE CHEZ LE LAPIN SOMMAIRE
Les coefficients de partition pour l'halothane ont ete determines dans divers tissus de lapin apres environ cinq heures d'anesthesie Le coefficient moyen sang/gaz, apras correction de la reduction d'hematocrite survenant durant l'experimentation, etait 3,94. Les coefficients tissu/sang differaient significativement de l'unite, a l'exception du muscle squelettique: cerveau 1,55—coeur 2,74—rein 1,75 et foie 2,26. Les differences entre les tissues etaient tres significatives. Celles entre les animaux etaient grandes mais en general non-systematiques. Les limites de confiance 95 pourcent pour les coefficients tissu/gaz dans la population etaient +32 pourcent —24 pourcent. II y avait aussi une variation entre les echantillons individuels 1 et 2 ml du meme tissu: les limites de confiance 95 pourcent pour ces echantillons etaient de l'ordre de +20 pourcent de la moyenne pour le foie, et +50 pourcent de la moyenne pour le coeur, rein et muscle. DIE VARIABILITAT DER VERTEILUNGSKOEFFIZIENTEN VON HALOTHAN BEIM KANINCHEN ZUSAMMENFASSUNG
Beim Kaninchen wurden die Verteilungskoeffizienten yon Halothan in verschiedenen Geweben nach 5-sriindiger Anasthesie bestimmt. Nach Vornahme einer Korrektur fur den im Verlauf des Experimentes auftretenden Abfall des Hamatokrits ergab sich ein Koeffizient Blut/Gas von 3,94. Mit Ausnahme des Skelettmuskels unterschieden sich die Koeffizienten fur die Verteilung Gewebe/Blut signifikant von 1: Gehirn 1,55; Herz 2,74; Niere 1,75 und Leber 2,26. Die Unterschiede zwischen den Geweben waren hochsignifikant. Die Unterschiede zwischen den Tieren waren gross, jedoch insgesamt nicht systematisch. Die 95%-Vertrauensgrenzen fiir die Koeffizienten Gewebe/Gas in der Versuchseinheit waren +32% bis — 24%. Auch fiir die Einzelproben von 1 bis 2ml des gleichen Gewebes ergaben sich Abweichungen: die 95%Vertrauensgrenzen dieser Proben lagen bei der Leber im Bereich von +20 des Mittelwertes, bei Herz, Niere und Muskel im Bereich von +50% des Mittelwertes. LA VARIABILIDAD DE LOS COEFICIENTES DE PARTICION PARA HALOTANO EN EL CONEJO RESUMEN
Fueron determinados los coeficientes de partici6n para halotano en diversos tejidos de conejos despues de aproximadamente cinco horas de anestesia. Despues de corregir el descenso en el hematocrito, que ocurri6 durante los experimentos, el coeficiente medio sangre/gas fue de 3,94. Exceptuando en el caso del musculo esqueletico, los coeficientes tejido/sangre eran significativamente diferentes de la unidad: cerebro 1,55; coraz6n 2,74; ririon 1,75 e higado 2,26. Las diferencias entre los tejidos eran muy significativas. Las diferencias entre animates eran grandes, pero en su mayoria no sistematicas. Los limites del 95 por ciento de confianza para los coeficientes tejido/gas en la poblacion fueron de + 32 por ciento — 24 por ciento. Tambien hubo variation entre muestras individuales de 1 hasta 2 ml del mismo tejido: los limites del 95 por ciento de confianza para tales muestras fueron de +20 por ciento de la media para el higado y ±50 por ciento de la media para el corazdn, ririon y musculo.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Allott, P. R., Steward, A. and Mapleson, W. W. (1971). Determination of halothane in gas, blood, and tissues by chemical extraction and gas chromatography. Brit J. Anaesth., 43, 913. Bartlett, M. S. (1937). Properties of sufficiency and statistical tests. Proc. roy Soc. A., 160, 268. Box, G. E. P. (1949). A general distribution theory for a class of likelihood criteria. Biometrika, 36, 317. Cowles, A. L., Borgstedt, H. H., and Gillies, A. J. (1971). Solubilities of ethylene, cyclopropane, halothane and diethyl ether in human and dog blood at low concentrations. Anesthesiology, 35, 203. Davies, O. L. (1957). Statistical Methods in Research and Production. 3rd ed. London: Oliver & Boyd. Duncan, W. A. M. (1963). Solubility and partition coefficients; in Uptake and Distribution of Anesthetic Agents. eds. Papper, E. M., and Kitz, R. J.) pp. 17-19. New York: McGraw-Hill. Eger, E. I., n, Saidman, L. J., and Brandstater, B. (1965). Temperature dependence of halothane and cyclopropane anesthesia in dogs; correlation with some theories of anesthetic action. Anesthesiology, 26, 764. Han, Y. H., and Helrich, M. (1966). Effect of temperature on solubility of halothane in human blood and brain tissue homogenate. Anesth. Analg. Curr. Res., 45, 775. Laasberg, H. L., and Hedley-Whyte, J. (1970). Halothane solubility in blood and solutions of plasma proteins. Anesthesiology, 32, 351. Larson, C. P., Eger, E. I., n, and Severinghaus, J. W. (1962). The solubility of halothane in blood and tissue homogenates. Anesthesiology, 23, 349. Lowe, H. J., (1964). Flame ionisation detection of volatile organic anesthetics in blood, gases and tissues. Anesthesiology, 25, 808. (1968). Determination of volatile organic anaesthetics in gases, blood and tissues, in Theory and Application of Gas Chromatography in Industry and Medicine, Kroman, H. S., Bender, S. R. (Eds) pp. 194-209. New York: Grune and Stratton. Hagler, K. (1969). Determination of volatile organic anesthetics in blood, gases, tissues, and lipids: partition coefficients; in Gas Chromatography in Biology and Medicine, (ed. Porter, R.), pp. 86-103. London: Churchill. Mapleson, W. W., Evans, D. E., and Flook, V. (1970). The variability of partition coefficients for nitrous oxide and cyclopropane in the rabbit. Brit. J. Anaesth., 42, 1033. Okuda, Y. (1968). A study on passage of inhalation anesthetics from blood into cerebrospinal fluid with gas chromatography, Arch. jap. Chir., 37, 700. Saidman, L. J., Eger, E. I., n, Munson, E. S. and Severinghaus, J. W. (1966). A method for determining solubility of anesthetics utilizing the Scholander apparatus. Anesthesiology, 27, 180. Steward, A., Mapleson, W. W., and Allott, P. R. (1972). A comparison of in-vivo and in-vitro partition coefficients for halothane in the rabbit. Brit. J. Anaesth., 44, 650. Wortley, D. J., Herbert, P., Thornton, J. A., and Whelpton, D. (1968). The use of gas chromatography in the measurement of anaesthetic agents in gas and blood. Brit. J. Anaesth., 40, 624.
661
THE VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE IN THE RABBIT W. W. MAPLESON, P. R. ALLOTT AND A. STEWARD
SUMMARY
In the course of experiments designed to compare in-vivo and in-vitro partition coefficients for halothane, results were obtained for six tissues in each of six New Zealand White rabbits. For reasons given in reporting the results of a similar study with nitrous oxide and cyclopropane (Mapleson, Evans and Flook, 1970) it is of value to examine the variability of the results, in order to determine to what extent it is possible to estimate the partition coefficient for a particular tissue in a particular individual, without actually measuring it in-vivo. METHODS
The details of the experimental method are given in a companion paper (Steward, Mapleson and Allott, 1972) and in Allott, Steward and Mapleson (1971). In brief, the rabbits were anaesthetized in such a way as to maintain the end-tidal concentration constant at some level in the range of 0.84 to 0.92 per cent halothane. At the end of about five hours the animals were sacrificed and samples of blood, brain, heart, kidney, liver and skeletal muscle were removed for determination of their halothane content, and hence their in-vivo tissue/gas partition coefficients. Further samples of each tissue were homogenized with saline and equilibrated in a rotating tonometer with similar concentrations of halothane to yield in-vitro tissue/gas partition coefficients. The statistical techniques used to analyse the results were taken from Davies (1957).
RESULTS
The full set of observed tissue/gas and blood/ gas partition coefficients is given in table I. For reasons given in the companion paper, all the main statistical calculations were performed on the logarithms of the coefficients. A three-way cross-classification analysis of variance in the companion paper separated out the several sources of variation as follows. (i) There is no significant difference between the in-vivo and in-vitro techniques overall, but there is a significant tissue-technique interaction. As was shown in the companion paper this was due to significant differences between techniques for brain and heart with no significant differences for the other four tissues. (ii) There are significant differences between tissues, even when referred to either the tissue-technique or tissue-animal interaction (see below). Therefore it is worth calculating separate means for each tissue. (iii) There is a significant tissue-animal interaction. This means that there are real differences between animals (greater than can be attributed to experimental error) but these differences are not systematic over all the tissues. That is to say that, if a given animal has a higher-than-average coefficient for one tissue, then the coefficients for the other tissues are not, in general, equally far above the averages for W. W. MAPLESON, PH.D., F.INST.P.; P. R. ALLOTT, B.SC.;
A. STEWARD, B.SC.; Department of Anaesthetics, Welsh National School of Medicine, Cardiff, CF4 4XN.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Partition coefficients were determined for halothane in various rabbit tissues after about five hours anaesthesia. After correcting for the fall in haematocrit which occurred during the experiments the mean blood/gas coefficient was 3.94. Except in the case of skeletal muscle the tissue/blood coefficients differed significantly from unity: brain 1.55, heart 2.74, kidney 1.75 and liver 2.26. The differences between the tissues were highly significant. Differences between animals were large but mainly unsystematic. The 95% confidence limits for tissue/gas coefficients in the population were + 32% —24%. There was also variation between individual 1- to 2-ml samples of the same tissue: the 95% confidence limits for such samples were of the order of +20% of the mean for liver, and +50% of the mean for heart, kidney and muscle.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE TABLE I.
657
Tissue/gas partition coefficients for halothane at 37.5°C—individual observations.
Kidney Muscle Liver Heart Brain vivo vitro vivo vitro vivo vitro vivo vitro vivo vitro 12.3 10.4 10.0 10.1 11.4 13.8 1 3.6 4.0 6.3 5.8 8.7 6.0 _ 7.4 7.1 2 8.4 2.8 3.0 3.7 6.0 _ 7.0 4.1 7.4 3 6.8 2.8 2.6 2.7 12.0 12.5 6.7 5.9 2.8 9.1 7.6 6.5 4 3.0 4.8 3.5 10.4 11.3 11.5 5.9 6.8 3.6 5.4 8.1 7.6 3.8 _ 10.3 3.9 10.8 10.4 5.5 6.2 3.9 5 8.8 6.7 2.9 6 4.3 3.8 13.4 10.4 6.1 5.8 3.4 Brain and heart in-vivo coefficients have been corrected on the assumption that, on the average, the in-vitro coefficients provide the better estimate. If the alternative view, that the in-vivo coefficients are the more reliable, is taken then all the brain coefficients should be divided by 1.282 and all the heart coefficients should be multiplied by 1.500 (see text). Rabbit No:
Blood vivo vitro 3.6 3.8
respect of heart and brain for which there were significant differences between the results of the two techniques. In the companion paper it was shown that these differences may have been real, in which case the in-vivo coefficients are to be preferred— except that the heart samples contained some surface fat and therefore gave a coefficient higher than that to be expected for cardiac muscle. However, the companion paper also showed that the difference might instead be attributable to experimental errors in such a way that the in-vitro coefficient is to be preferred for both heart and brain. We have chosen to prefer the in-vitro coefficients. Therefore, with heart and brain, we have corrected the individual in-vivo coefficients by the mean of the vitro-vivo differences for each of these tissues. In case the invivo coefficients should be preferred the appropriate corrections are given in the footnote to table I. Details of the statistical methods used are given in TABLE II. Partition coefficients (tissue/gas and tissue/ the appendix and the results in table II. blood) for halothane at 37.5°C—means and probable limits. Table II also includes a "corrected" set of tissue/ 95% confidence limits blood coefficients the basis for which is as follows. of mean of population The haematocrit fell during the experiments perhaps (as % of mean) Tissue/gas Mean owing, at least in part, to the infusion of Hartmann's 3.72 2.83 3.24 Blood solution; the extent of the fall ranged from 1.7 to 6.98 6.09 5.31 Brain 10.79 Heart 12.35 9.44 10.5 per cent. There was a highly significant -24% +32% Kidney 7.84 5.94 6.82 (P=0.005) linear regression of blood/gas 7.83 10.29 Liver 8.98 partition coefficient (measured on samples drawn at 4.13 3.60 3.14 Muscle the end of the experiment) on this fall in haematoTissue/blood crit. The regression coefficient (-0.125/per cent 2.13 Brain 1.88 1.65 3.77 2.93 Heart 3.33 haematocrit fall) was used to calculate for each 1.87 2.43 Kidney 2.13 -20% +25% animal what the blood/gas coefficient would have 2.40 3.11 Liver 2.73 been at the beginning of the experiment, before 1.27 1.11 0.98 Muscle any haemodilution had occurred. These corrected Tissue/blood (corrected*) blood/gas coefficients (mean 3.94) should be more 1.55 1.34 1.78 Brain 3.16 2.74 2.38 Heart representative of normal values in the rabbit than 1.75 1.51 2.02 Kidney -24% +32% those which we happened to measure after varying 2.26 1.95 2.61 Liver degrees of haemodilution. They were therefore used 1.06 0.92 0.79 Muscle to calculate a corrected set of tissue/blood coeffi"Using blood/gas coefficients (mean 3.94) corrected for cients. fall in haematocrit—see text.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
those tissues. Therefore the variability of each tissue must be assessed separately. (iv) The between-animal variation is significant, even when referred to the tissue-animal interaction. This means that, despite the above remarks, a high coefficient for one tissue in an animal does to some extent imply high coefficients for the other tissues. This last finding suggests that, in an individual, tissue/gas coefficients for halothane may be predictable to a useful extent from the more easily measured blood/gas coefficient. Therefore the results are expressed in terms of tissue/blood coefficients as well as of tissue/gas coefficients (table II). In the light of the other foregoing findings the results, for each type of coefficient, are expressed in terms of the mean coefficient for each tissue with confidence limits of the mean and of the population. In calculating the means a problem arises in
658
BRITISH JOURNAL OF ANAESTHESIA
DISCUSSION
The present results. The range of variation of partition coefficients between tissues (table II) is much greater than for nitrous oxide or cyclopropane (Mapleson, Evans and Flook, 1970). The variation is by a factor of more than 3:1 instead of only 1.3:1 or 1.7:1. On the other hand, the range of variation between animals is of the same order as that found in the study of cyclopropane and nitrous oxide. This may in part be due to the use of a single strain of rabbit, New Zealand White, in the halothane study. As was the case with nitrous oxide, none of the variation in the last pair of columns in table II should be attributable to experimental error since the statistical treatment has allowed for this. The results for nitrous oxide and cyclopropane showed a small variance for blood and a large one for muscle compared to those for the other tissues; such differences were not evident in the present results except that the variance for the corrected blood/gas coefficients was small (appendix). Tissue/blood coefficients influence the speed of equilibration of a tissue with the arterial tension, higher coefficients giving slower equilibration. It is notable that all the tissue/blood coefficients in table II are significantly greater than unity except those for muscle. Since the confidence limits of the population of uncorrected tissue/blood coefficients are only slightly
closer than those for tissue/gas coefficients, the suggestion made earlier in this paper that measurement of the blood/gas coefficient in an individual might provide a worthwhile guide to the tissue/gas coefficient is not verified to any useful extent. Unlike the study with nitrous oxide and cyclopropane this investigation yielded an estimate of the variation between adjacent portions (1 to 2 ml) of apparently homogeneous tissue. The magnitude of the variation is remarkably wide (see above) and suggests that, in order to obtain a reliable coefficient for a given organ in a given animal, measurements should be made on a large sample, if not the whole organ. It is shown in the statistical appendix that this variation between samples contributed little or nothing to the variation between individuals shown in table II. Comparison with the results of other workers. Variation between individuals. There is very little information available on this source of variation. For instance, Han and Helrich (1966) give only mean values; Laasberg and Hedley-Whyte (1970) give only means and standard errors of the means, and, although Larson, Eger and Severinghaus (1962) refer to means and standard deviations, it is evident that these refer to experimental error and not to variation between individuals: samples from different individuals were first pooled before performing repeat determinations on the collective sample. On the other hand, Wortley and associates (1968) give a histogram of blood/gas solubility coefficients for 33 surgical patients and from their data we have calculated the 95 per cent confidence limits for the variation between individuals to be +17 per cent of their mean of 2.3. These limits are closer than those of —24 per cent to +32 per cent for the rabbits (table II). Therefore, as with nitrous oxide, it appears that the rabbit may be more variable than man even though, in the present study with halothane, a more homogeneous strain of rabbit was used. The complex effects of proteins on the solubility of halothane have been discussed recently by Laasberg and Hedley-Whyte (1970). Han and Helrich (1966) and Cowles, Borgstedt and Gillies (1971) give equations for the regression of halothane solubility in blood on haemoglobin concentration which are similar to each other but are somewhat in conflict with the data of Laasberg and Hedley-Whyte. If these equations are valid they could be useful in predicting blood solubility coefficients for individuals,
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
In an early series of experiments two in-vivo samples of each tissue were taken directly from the animal to the extraction phials; in the present series, for most tissues, a single large sample was crushed in liquid nitrogen and the fragments well mixed before being divided into an in-vitro and two in-vivo samples. Although care was taken to obtain apparently similar duplicate in-vivo samples in the early series, the variances between the in-vivo phials were six to nine times larger than in the present series. Since there was no systematic difference between the two series in the variance between duplicate in-vitro phials, the in-vivo between-phials variance in the early series provides an estimate of the variation between adjacent 1- to 2-ml portions of a given tissue. From these variances we estimate the 95 per cent confidence limits for a single 1- to 2-ml portion of tissue to be of the order of ± 20 per cent of the mean for liver and +50 per cent for heart, kidney and muscle.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE TABLE III. Liquid/gas and tissue/gas partition coefficients at 37 °C.
Fluid or Tissue Water
Species
0.9% NaCl solution Oil
Olive
Plasma
Man
Erythrocytes Blood
Man Man
Ox Dog
Whole Brain
Rabbit
t*4.03
Man
6.0 §5.2 4.8 *6.2 5.4 8.3 *11.0 3.6 3.5 *7.0 6.0 4.2 *9.2 8.0 7.0 *3.7 138
Ox
Rabbit Grey Matter White Matter Heart Kidney
Man Man
Rabbit Man Ox
Rabbit Liver
Man Ox
Rabbit Skeletal Muscle
Man Ox
Fat
Man
(1) (6) (1) (1) (1) (1) (1) (1) (1) (1) (1) (1)
0.9 0.6 2.7 0.7 0.9 0.6 0.5 0.9 0.8 0.8 0.9 0.6 0.7 0.9 0.5 0.6 0.9
0.05
•Present study corrected to 37 °C. $Temperatures corrected to 37 °C. fCorrected for fall in haematocrit (see text). The references indicated by numbers in parentheses are as follows: (1) Larson, Eger and Severinghaus, 1962; (2) Duncan, 1963; (3) Laasberg and Hedley-Whyte, 1970. (4) Saidman et al., 1966; (5) Allott, Steward and Mapleson, 1971; (6) Han and Helrich, 1966; (7) Wortley et al.3 1968; (8) Lowe, 1964; (9) Cowles, Borgstedt and Gillies, 1971; (10) Lowe, 1968; (11) Lowe and Hagler, 1969; (12) Okuda, 1968.
although there would still be a considerable uncertainty if Laasberg and Hedley-Whyte are correct in attributing greater importance to the albumin-globulin ratio. Furthermore the possibility of changes of solubility coefficient consequent upon the infusion of liquids into the circulation must be borne in mind. Variation between tissues and species. More information is available about mean solubilities in various fluids and in the tissues of various species. Table III gives the means of the present results and those of other workers. Since most determinations have been made at 37 °C all data for other temperatures (including our own) have been corrected by using the temperature coefficient of —4.6%/°C derivable from Eger, Saidman and Brandstater (1965). The smaller solubility of halothane in saline and Krebs-Henseleit solution than in water is doubtless a reflection of the salting-out effect of electrolytes. There is some evidence of deviation from Henry's law: the data of Saidman and associates (1966), (ref. 4 table III) indicates an increase in the solubility of halothane in water with an increase in concentration. However, the effect appears to be small and most of the concentrations used in determining the solubilities in table III were within a narrow range. The tissue values found by Lowe (1968) for man and Lowe and Hagler (1969) for man and calf are not included in table III because the measurements were made on very small samples, up to 25 /A, drawn apparently from single individuals. Therefore, although the values generally differ from those given in table III, this is likely to be merely a reflection of variation between adjacent portions of tissue in the same individual and between different individuals. Some authors have assumed that there is no species variation in partition coefficients for particular tissues, but the data given in table III show that this assumption is not generally justified. Although the coefficients for man and calf appear to be fairly similar, as do the coefficients for whole brain in man and rabbit, the blood coefficients for dog and rabbit and the liver coefficients for the rabbit are all significantly higher than the corresponding human values. Further, the coefficient for kidney in the rabbit is very much higher than in man while that for skeletal muscle is very much lower. The high muscle/blood coefficient in man (about 3 5 compared with 0.9 in the rabbit) means that the time required for skeletal muscle to equilibrate with a constant arterial tension will be much longer in
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Krebs-Henseleit solution
Approximate Gas CoeffiConcencient tration % 0.74 (1) 1.3 _ 0.79 (2) 2.1 0.78 (3) 1 0.87 (4) 1.04 (4) 10 0.89 (10) — 0.86 (11) 0.82 (12) 0.5 0.70 (1) 1.3 0.9 §0.77 (5) 2.2 0.76 (3) 0.03 224 (1) 2.28 (3) 1.2 §3.14 (6) 3.5 §1.26 (6) 4.7 2.3 (1) 1.0 — 2.5 (2) 2.42 (3) 1.1 3.8 §2.41 (6) 0.7 2.3 (7) — 2.42 (8) 0.7 §2.56 (9) — 2.24 (10) — 2.56 (11) 1.0 2.3 (1) 2.87 (11) 0.7 §3.18 (9) 2.00 (12) 0.5
659
BRITISH JOURNAL OF ANAESTHESIA
660 man than the five hours found adequate (see companion paper) in the rabbit. CONCLUSIONS
In man and dog the uncertainty over the blood/ gas coefficient in an individual may be reduced if the haemoglobin concentration or haematocrit is known. Mean coefficients in one species are not a reliable guide to those in another.
Effects of variation between individual 1- to 2-ml samples of tissue on the estimate of between-animal variation. STATISTICAL APPENDIX In the case of brain and blood the in-vivo and in-vitro samples were removed from the animal separately; thereMissing values. fore any variation between individual samples would For the purpose of the three-way and the two-way contribute to the experimental error and hence not to breakdown-by-technique analysis of variance gaps on the the calculated between-animal variation. In the case of basic data were filled as explained in the companion liver the in-vivo and in-vitro samples came from a mixpaper. ture of fragments of a single 20-ml sample. Therefore variation between such samples would not contribute to Estimation of the means and probable limits of the experimental error but would contribute to the calculated partition coefficients. between-animal variation. However, variation between Two-way breakdown-by-technique analyses of variance 20-ml samples of an 80-ml liver would be much less than showed that the in-vivo and in-vitro results were about the +20 per cent variation found between 1- to 2-ml equally variable and therefore both were used. samples, and therefore the contribution to the betweenHierarchical analyses of variance for each tissue are animal variation, of +32% —24%, would be small. In summarized in table IV. Before performing these analyses the case of heart and kidney the whole organ was sampled the brain and the heart in-vivo coefficients were corrected and mixed so that variation between samples does not by the respective mean vitro-vivo difference as described arise. Similarly a complete muscle was sampled; therefore in the main text. The residual degrees of freedom were the between-animal variation of +32%—24% is an untherefore reduced by one in each of these two tissues. biased estimate of the variation between animals for the Application of the Box (1949) modification of Bartlett's coefficient for that particular muscle although it may be an overestimate of the variation between animals for the TABLE IV. Summary of hierarchical analyses of variance mean coefficient of all muscles. for each tissue (~ Residual Mean square Blood Brain Heart Kidney Liver Muscle
0.00118 0.00136 0.00192 0.01145 0.00191 0.00330
d.f 5 4 5
4 5 5
Between-animal Mean square d.f. 0.00586 0.00034 0.01564 0.01546 0.00981 0.01039
5 5 5 5 5 5
ACKNOWLEDGEMENTS
We are grateful to Professor William W. Mushin, Head of this Department, for many helpful discussions and for providing the conditions so necessary for the conduct of research of this nature. We are also indebted to Imperial Chemical Industries Ltd for generous supplies of halothane. The work was supported by a grant from the Medical Research Council.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Individual in-vivo partition coefficients for halothane in at least blood, kidney, liver and muscle can be reliably estimated from in-vitro measurements, provided that large samples are used. The coefficients for different tissues are very different from one another and the mean coefficient for a tissue in the rabbit population is a very important guide to the coefficient for that tissue in a particular rabbit. Variation between individuals for a given tissue is less than the variation between tissues but is still considerable except perhaps for blood. It may be marginally less in terms of tissue/blood coefficients and may be smaller in man than in the rabbit. In our experiments the variation of the blood/gas coefficient between individual rabbits was largely attributable to varying degrees of haemodilution.
(1937) test showed that the residual mean squares were homogeneous (P=0.13) but that the between-animal mean squares were not (P=0.03). The inhomogeneity is due to the low value for brain. Since a similarly low value was not found with either nitrous oxide or cyclopropane (Mapleson, Evans and Flook, 1970) and since the level of significance is not very high it seemed preferable to treat this as a chance event. Therefore, both the residual and the betweenanimal mean squares were pooled for all tissues. The variance of the mean coefficient for each tissue was then taken as the sum of, firstly, the experimentalerror variance estimate (which equals the pooled residual mean square) divided by the number of measurements used in deriving the mean and, secondly, the betweenanimal variance estimate (which equals the difference between the pooled between-animal and pooled residual mean squares divided by the mean number of measurements per animal) divided by the number of animals used. The number of degrees of freedom was taken as the sum of those of the two variance estimates. In assessing the range of values likely to be found in individuals of the species the between-animal variance estimate alone was used. This could not be done with the corrected blood/gas coefficients because the between-animal mean square was less than the residual mean square. Therefore all that can be said of the 95 per cent confidence limits of the population of corrected blood/gas coefficients is that they are close to the mean. For tissue/blood coefficients individual values were determined by dividing each tissue/gas coefficient by the corresponding blood/gas coefficient for the same technique in the same animal. Means and confidence limits were then determined as for the tissue/gas coefficients.
VARIABILITY OF PARTITION COEFFICIENTS FOR HALOTHANE REFERENCES
LA VARIABILITE DES COEFFICIENTS DE PARTITION POUR L'HALOTHANE CHEZ LE LAPIN SOMMAIRE
Les coefficients de partition pour l'halothane ont ete determines dans divers tissus de lapin apres environ cinq heures d'anesthesie Le coefficient moyen sang/gaz, apras correction de la reduction d'hematocrite survenant durant l'experimentation, etait 3,94. Les coefficients tissu/sang differaient significativement de l'unite, a l'exception du muscle squelettique: cerveau 1,55—coeur 2,74—rein 1,75 et foie 2,26. Les differences entre les tissues etaient tres significatives. Celles entre les animaux etaient grandes mais en general non-systematiques. Les limites de confiance 95 pourcent pour les coefficients tissu/gaz dans la population etaient +32 pourcent —24 pourcent. II y avait aussi une variation entre les echantillons individuels 1 et 2 ml du meme tissu: les limites de confiance 95 pourcent pour ces echantillons etaient de l'ordre de +20 pourcent de la moyenne pour le foie, et +50 pourcent de la moyenne pour le coeur, rein et muscle. DIE VARIABILITAT DER VERTEILUNGSKOEFFIZIENTEN VON HALOTHAN BEIM KANINCHEN ZUSAMMENFASSUNG
Beim Kaninchen wurden die Verteilungskoeffizienten yon Halothan in verschiedenen Geweben nach 5-sriindiger Anasthesie bestimmt. Nach Vornahme einer Korrektur fur den im Verlauf des Experimentes auftretenden Abfall des Hamatokrits ergab sich ein Koeffizient Blut/Gas von 3,94. Mit Ausnahme des Skelettmuskels unterschieden sich die Koeffizienten fur die Verteilung Gewebe/Blut signifikant von 1: Gehirn 1,55; Herz 2,74; Niere 1,75 und Leber 2,26. Die Unterschiede zwischen den Geweben waren hochsignifikant. Die Unterschiede zwischen den Tieren waren gross, jedoch insgesamt nicht systematisch. Die 95%-Vertrauensgrenzen fiir die Koeffizienten Gewebe/Gas in der Versuchseinheit waren +32% bis — 24%. Auch fiir die Einzelproben von 1 bis 2ml des gleichen Gewebes ergaben sich Abweichungen: die 95%Vertrauensgrenzen dieser Proben lagen bei der Leber im Bereich von +20 des Mittelwertes, bei Herz, Niere und Muskel im Bereich von +50% des Mittelwertes. LA VARIABILIDAD DE LOS COEFICIENTES DE PARTICION PARA HALOTANO EN EL CONEJO RESUMEN
Fueron determinados los coeficientes de partici6n para halotano en diversos tejidos de conejos despues de aproximadamente cinco horas de anestesia. Despues de corregir el descenso en el hematocrito, que ocurri6 durante los experimentos, el coeficiente medio sangre/gas fue de 3,94. Exceptuando en el caso del musculo esqueletico, los coeficientes tejido/sangre eran significativamente diferentes de la unidad: cerebro 1,55; coraz6n 2,74; ririon 1,75 e higado 2,26. Las diferencias entre los tejidos eran muy significativas. Las diferencias entre animates eran grandes, pero en su mayoria no sistematicas. Los limites del 95 por ciento de confianza para los coeficientes tejido/gas en la poblacion fueron de + 32 por ciento — 24 por ciento. Tambien hubo variation entre muestras individuales de 1 hasta 2 ml del mismo tejido: los limites del 95 por ciento de confianza para tales muestras fueron de +20 por ciento de la media para el higado y ±50 por ciento de la media para el corazdn, ririon y musculo.
Downloaded from http://bja.oxfordjournals.org/ at New York University on May 9, 2015
Allott, P. R., Steward, A. and Mapleson, W. W. (1971). Determination of halothane in gas, blood, and tissues by chemical extraction and gas chromatography. Brit J. Anaesth., 43, 913. Bartlett, M. S. (1937). Properties of sufficiency and statistical tests. Proc. roy Soc. A., 160, 268. Box, G. E. P. (1949). A general distribution theory for a class of likelihood criteria. Biometrika, 36, 317. Cowles, A. L., Borgstedt, H. H., and Gillies, A. J. (1971). Solubilities of ethylene, cyclopropane, halothane and diethyl ether in human and dog blood at low concentrations. Anesthesiology, 35, 203. Davies, O. L. (1957). Statistical Methods in Research and Production. 3rd ed. London: Oliver & Boyd. Duncan, W. A. M. (1963). Solubility and partition coefficients; in Uptake and Distribution of Anesthetic Agents. eds. Papper, E. M., and Kitz, R. J.) pp. 17-19. New York: McGraw-Hill. Eger, E. I., n, Saidman, L. J., and Brandstater, B. (1965). Temperature dependence of halothane and cyclopropane anesthesia in dogs; correlation with some theories of anesthetic action. Anesthesiology, 26, 764. Han, Y. H., and Helrich, M. (1966). Effect of temperature on solubility of halothane in human blood and brain tissue homogenate. Anesth. Analg. Curr. Res., 45, 775. Laasberg, H. L., and Hedley-Whyte, J. (1970). Halothane solubility in blood and solutions of plasma proteins. Anesthesiology, 32, 351. Larson, C. P., Eger, E. I., n, and Severinghaus, J. W. (1962). The solubility of halothane in blood and tissue homogenates. Anesthesiology, 23, 349. Lowe, H. J., (1964). Flame ionisation detection of volatile organic anesthetics in blood, gases and tissues. Anesthesiology, 25, 808. (1968). Determination of volatile organic anaesthetics in gases, blood and tissues, in Theory and Application of Gas Chromatography in Industry and Medicine, Kroman, H. S., Bender, S. R. (Eds) pp. 194-209. New York: Grune and Stratton. Hagler, K. (1969). Determination of volatile organic anesthetics in blood, gases, tissues, and lipids: partition coefficients; in Gas Chromatography in Biology and Medicine, (ed. Porter, R.), pp. 86-103. London: Churchill. Mapleson, W. W., Evans, D. E., and Flook, V. (1970). The variability of partition coefficients for nitrous oxide and cyclopropane in the rabbit. Brit. J. Anaesth., 42, 1033. Okuda, Y. (1968). A study on passage of inhalation anesthetics from blood into cerebrospinal fluid with gas chromatography, Arch. jap. Chir., 37, 700. Saidman, L. J., Eger, E. I., n, Munson, E. S. and Severinghaus, J. W. (1966). A method for determining solubility of anesthetics utilizing the Scholander apparatus. Anesthesiology, 27, 180. Steward, A., Mapleson, W. W., and Allott, P. R. (1972). A comparison of in-vivo and in-vitro partition coefficients for halothane in the rabbit. Brit. J. Anaesth., 44, 650. Wortley, D. J., Herbert, P., Thornton, J. A., and Whelpton, D. (1968). The use of gas chromatography in the measurement of anaesthetic agents in gas and blood. Brit. J. Anaesth., 40, 624.
661