Relative potency of vecuronium in male and female patients in Britain and Australia

British Journal of Anaesthesia 1994; 72: 190-194

Relative potency of vecuronium in male and female patients in Britain and Australia P. SEMPLE, D. A. HOPE, P. CLYBURN AND A. RODBERT

SUMMARY

KEY WORDS Neuromuscular block: vecuronium, potency.

Patients vary considerably in their response to a standard dose of a neuromuscular blocking drug. Despite this, several studies have reported that sensitivity to these drugs also varies between different populations and different geographical locations [1, 2, 3]. Patients in Paris, France, apparently require 27% more vecuronium than patients in Montreal, Canada, to achieve the same degree of block [2]. Variations of this magnitude would have clinical implications, therefore this study was undertaken to compare the potencies of vecuronium when used in British and Australian patients. PATIENTS AND METHODS

Studies of similar design were conducted at the University Hospital of Wales, Cardiff, U.K. and the Princess Alexandra Hospital, Brisbane, Australia. Ethics Committee approval was obtained in both centres and all patients gave informed consent.

PETER SEMPLE*, B.M., F.F.A.R.C.S.I., F.R.C.A.; Princess Alexandra Hospital, Brisbane, Queensland 4102, Australia. DAVID ALAN HOPE, B.M., F.R.C.A.; PAUL CLYBURN, M.B., B.S., M.R.C.P., F.R.C.A.;

University Hospital of Wales, Heath Park, Cardiff CF4 4XW. ANDREW RODBERT, M.B., B.S., D.A.; Maryborough Base Hospital,

Maryborough, Queensland 4650, Australia. Accepted for Publication: September 6, 1993. •Present address: Department of Anaesthetics, Leeds General Infirmary, Great George Street, Leeds LSI 3EX.

Downloaded from http://bja.oxfordjournals.org/ at Université Laval on July 12, 2015

We compared the potency of vecuronium when given to similar patients in Brisbane, Australia, and Cardiff, United Kingdom. Forty patients in each centre were anaesthetized using the same technique with propofol, fentanyl, nitrous oxide and vecuronium either 20 or 30 ug kg'1 by random allocation. Neuromuscular block was measured with similar Datex Relaxographs. There was no significant difference in potency between British and Australian patients. The ED^ and ED^ for a British male of average weight were 29.5 fig kg'' (95% confidence limits 27.3-32.3 fig kg-'J and 51.3 fig kg-' (44.3-63.9 fig kg-'), respectively. ED50 and EDX for Australians were 5.5% greater, with confidence limits from 4% less to 17% greater. Females were significantly more sensitive to vecuronium than males, requiring 22% less drug to achieve the same degree of neuromuscular block (confidence limits 12-32%). The results are consistent with the EDS0 being independent of body weight when the dose is expressed as ug kg-2?3, but not as fig or fig kg-'. (Br. J. Anaesth. 1994; 72: 190-1 94)

We studied 40 patients in each centre (20 male, 20 female) undergoing routine operations requiring neuromuscular block. For an SD between patients of 20 % peak twitch depression, this number of subjects gave a power of 0.9 for detecting a difference of 15 % depression, at the 5 % level of significance. Inclusion criteria were that the patients should be ASA I or II, aged 18—60 yr, of Caucasian race and should have lived all their lives in their current country of residence. Patients were excluded if they had any neuromuscular disease or any renal, hepatic or electrolyte abnormality, or if they were taking any drugs affecting neuromuscular transmission or excessive amounts of alcohol. All patients received an oral premedication of temazepam 20 mg and metoclopramide 10 mg 2 h before operation. Routine monitoring equipment was attached (non-invasive arterial pressure, ECG and pulse oximetry) and an i.v. cannula was inserted. Anaesthesia was induced with propofol 2.5 mgkg"1 and fentanyl 1 ug kg"1. A laryngeal mask airway was inserted and anaesthesia maintained with 70% nitrous oxide in oxygen administered through a Bain (coaxial D) system, with a fresh-gas flow of 10 litre min"1. Boluses of propofol 10 mg were administered as required. Manual ventilation of the lungs was performed to maintain end-tidal carbon dioxide partial pressure in the range 4.6-5.3 kPa. Neuromuscular transmission was measured by electromyography using a Datex Relaxograph (model number NMT 100) in each centre. Both machines contained the same software (Revision 01) and had been maintained according to the manufacturer's recommendations. Skin electrodes were applied over the non-dominant ulnar nerve and the adductor pollicis muscle. In each centre, patients were allocated randomly (males and females separately) to two subgroups to receive vecuronium 20 or 30 ug kg"1. Thus, in each centre, 10 subjects of each gender received each dose. The investigator was blinded to the dosage group at the time of re-

191

VECURONIUM IN AUSTRALIA AND U.K. cruitment. As the outcome measure was objective, it was not thought necessary to blind the investigator at the time of drug administration. Vecuronium was administered as a dilute solution of 50 jig ml"1 after the Relaxograph had been calibrated and train-offour stimulation was applied every 10 s. When the twitch depression (Tl %) had stabilized for three successive measurement it was recorded. Patient age, height, weight and the total amount of propofol given were recorded and the anaesthetic then proceeded as normal, including tracheal intubation when indicated. Patients were recruited from a variety of specialties in both centres (U.K./Australia): ENT five/five, oral surgery 16/10, gynaecology 10/seven, general surgery four/six, neurosurgery two/none, urology none/five, ophthalmic three/one, orthopaedic none/ six.

Country

Gender

U.K.

Male

U.K.

Female

Australia

Male

Australia

Female

Dose ("g kg"1)

Age

Height

(yr)

(m)

20 30 20 30 20 30 20 30

36.3 36.9 30.1 35.9 36.4 31.6 36.1 31.1 11.5

.80 .76 .65 .62 .81 .76

Pooled SD 5th Percentile 95th Percentile

20 59

Weight (kg) 82.6 73.5 65.6 61.9 81.1 73.0 54.2 56.3 15.1 49.0 96.5

1.59 1.64 0.107 1.52 1.87

TABLE II. Mean peak twitch depression of the eight subgroups of patients

Mean peak twitch depression (%) Dose (Ug kg"1) 20 20 Mean 30 30 Mean

Country U.K. Australia U.K. Australia

Male

Female

Mean

Pooled SD

18.1 19.4 18.8 54.3 45.5 49.9

37.3 24.2 30.8 70.0 65.3 67.6

27.71 21.8/

13.9

62.2 1 55.4 /

24.2

RESULTS

Table I shows the distribution of age, height and weight of the eight experimental subgroups. The mean peak twitch depressions of these subgroups are shown in table II. None of the variables, propofol dose, height, height2 or height3, improved the fit. Age produced a slight improvement, but the effect was not in the expected direction: older patients had less twitch depression, but this was not significant (P = 0.14). Log(dose in ug kg"1), weight and gender improved the fit in the expected directions, so were used as predictor variables, together with country of origin 100

TABLB III. Coefficients (b,), standard errors (75 degrees of freedom) and corresponding values of t and P for the fitted y = bo+bjXj+bjXj+bjXj+b^x.,, where y = the fitted ascsine (peak twitch depression); x, = 0 for U.K., 1 for Australia; x f = log(dose in figkg'*); Xj= 0 for males, 1 for females; and x 4 = body weight in kg

Predictor

Coefficient

Intercept Country (Aus-U.K.) Log(dose ug kg"1) Gender (F-M) Weight

- 2 92 -0.0537 2.33 0.249 0.00419

SE — 0.420 0.26 0.0470 - 1 .14 8 .9 < 0.0001 0.262 4 .3 < 0.0001 0.0587 2 .4 0.00177 0.020

n

co •«

tn

& 50-| T3 JZ

o

1

1 25 H Q_

10

20

30 40 Dose (ug kg"1)

50

60

70 80 90100

FIG. 1. Percent peak twitch depression plotted against dose in a logarithmic scale. • and = U.K. male; A and = U.K. female; O and = Australian male; A and = Australian female.

Downloaded from http://bja.oxfordjournals.org/ at Université Laval on July 12, 2015

Statistical analysis The ED60 and ED95, and their differences between the two countries, were determined by a form of multiple regression which was used to perform a parallel line assay [4]. This took account of the possible effects on the response to a given dose, of country of origin, gender, age, height, weight and dose of propofol (see Appendix).

TABLE I. Means ages, heights and xoeights of the eight subgroups of patients

BRITISH JOURNAL OF ANAESTHESIA

192 TABLE IV. Estimates and 95% confidence limits (CL) for the EDl0 and EDti of vecuronium in a U.K. male of average body weight (68.5 kg) and for the ratios of ED for Australians to U.K. residents and for females to males

ED M (ugkg-') ED, 6 Gig kg"') Australia: U.K. Female: male

Estimate

Lower CL

Upper CL

29.5 51.3 1.055 0.781

27.3 44.3 0.96 0.68

32.3 63.9 1.17 0.88

FIG. 2. Plot of t against n, where I is the value of Student's t for the coefficient for the dependence of arcsine (peak twitch depression) on body weight when dose is expressed in terms of Hg kg"". Note that there is no dependence when n = 0.68 and that the dependence becomes significant (P<0.05: r<2.0 or t > 2.0) at n = 0.40 and n = 0.95.

(table III). The effect of interactions between these predictors was tested: none was significant and the effects of country—log(dose) and gender-log(dose) were particularly small, thereby supporting the assumption that the regression lines for the two countries and two genders were parallel. The observed peak twitch depression and the fitted sigmoid curves for the four groups of patients are plotted against dose in |ig kg"1 on a logarithmic scale in figure 1. The ED60 and ED 96 for a British male of average weight (table IV) were 29.5 and 51.3 ug kg"1 respectively. The dose requirements were 5.5% greater for Australian subjects (not significant) and 22 % less for females (P < 0.0001). Confidence limits are given in table IV. Table III shows that, when dose was expressed in ug kg"1 (the basis on which vecuronium was administered), the degree of depression increased significantly with body weight. When the analysis was repeated with the same explanatory variables, except that dose was expressed simply as ug (i.e. Ug kg0), depression decreased significantly with body weight and there was very little change in the other coefficients. Figure 2 shows how the value of (Student's) t varied when dose was expressed as Hg kg"" using a range of values of n. It can be seen that t is zero (and therefore dependence on weight becomes zero) when dose is expressed as ug kg"068. The values of n at which t is just large enough to be

DISCUSSION

Our results show that any difference in sensitivity to vecuronium between Caucasian residents of Australia and of the United Kingdom is small: Australian residents are probably (95% confidence limits) between 17% less sensitive and 4% more sensitive than U.K. residents. We had controlled all factors known to affect neuromuscular block: the study drug was produced in Oss, Netherlands, in a single factory; the output variable, % twitch depression, was measured on identical Relaxographs in each centre; the anaesthetic technique was standardized and end-tidal carbon dioxide was kept in the normal range; other important factors, such as race and duration of residence in present country, were also controlled. Information on patient ancestry would have been interesting, but we decided that the necessary questions would be too intrusive. The possibility that racial or geographical factors affect the dose-response curve of neuromuscular blocking drugs was suggested by Katz and colleagues, who found large differences between American and English patients in their responses to suxamethonium and curare: English patients showed approximately 40% less twitch depression than American patients [1]. From the sigmoid curves in figure 1, it can be seen that, for vecuronium, this corresponds to a dose ratio of about 1.5. Despite the fact that identical drugs and instruments were not used in that study, this was a surprising result and one not easily dismissed. However, a large study of the dose-response relationship of curare by Walts and Drexler [5], failed to find any difference between groups of patients in 29 countries. Although they studied a large number of patients (306), these comprised 13 ethnic groups with as few as 10 in some. This design would make it unlikely that differences would become "significant" unless they were quite large. The report by Fiset and colleagues [2] that the EDB0 for vecuronium was 27 % greater in Paris, France than in Montreal, Canada, seems to support the findings by Katz and colleagues. Unfortunately, that study also used different "force displacement" transducers in the two centres and no mention was made of control of race or duration of residency in the country. Thus, to date, all studies reporting that there are inter-ethnic differences in sensitivity to neuromuscular blocking drugs may be criticized for failing to control all the factors involved. Equally, the study reporting no difference could, by the nature of its design, only hope to detect gross differences. Our data partly explain this problem by stating the limits between which any difference can be expected to lie (in our study populations). Although we found no appreciable difference between patients in each country, the difference between the sexes was highly significant and of clinical interest (12-32% less vecuronium required for females (table IV)). This has not been reported before for vecuronium, although pancuronium onset

Downloaded from http://bja.oxfordjournals.org/ at Université Laval on July 12, 2015

0.8

significant at the 5 % level represent estimates of the 95% confidence limits of n: 0.40 and 0.95.

VECURONIUM IN AUSTRALIA AND U.K.

APPENDIX DETAILS OF STATISTICAL ANALYSIS The dose—response curve for neuromuscular blocking drugs is generally assumed to be sigmoid. Support for linearizing this by using the probit transform of peak depression and the log transform of dose was reported as early as 1955 [7]. Subsequently, the logit transform has been used [2]. Probit, logit and arcsine transforms give similar results [8] except that, when the data include responses of 100 % (as in one of our patients) or 0 %, this leads to probits and logits of plus or minus infinity. Therefore we have used the arcsine transform, which has finite limits of 0 and s/2 at 0 and 100% response. The statistical package GLIM (generalized linear interactive modelling) [9, 10] was used to predict arcsine (peak twitch depression) from a linear function of several predictors. The main predictors were country of origin and log(dose in ng kg"1). Other explanatory variables examined were: gender, as there was a clinical impression that females were more sensitive than males; propofol dose, in case propofol enhanced relaxation [11-13]; age, on the general expectation that sensitivity to drugs may increase with age [14]; weight, as heavier patients may receive a relative overdose of a drug administered on a body weight basis; and height, as differences in build may alter the effect of the drug. In common with Tham and colleagues [15], we used only those predictors that improved the fit, and produced an effect in the expected direction. When the predictors had been decided upon, the ED M and ED t6 for U.K. males were calculated from the fined regression; the confidence limits of the differences between the two countries and the two genders were calculated using Fieller's theorem [16]. In this, context ED M and ED, 5 represent the estimated doses required to produce 50% and 95 % depression of twitch height in an average patient.

The effect of expressing the dose in terms of |ig kg *, with values of n in the range 0-1, was also examined. The raw regressions showed non-constant residual variance. The variation was found to be attributable to differences in variability between the four subgroups: U.K. male, U.K. female, Australian male, Australian female. This was corrected by weighting those subgroups in inverse proportion to the residual variance in each group. This was performed iteratively until the residua] variance stabilized. The weighted residuals then showed essentially constant variance (in a plot of weighted residuals against fitted values) and were close to Normally distributed (as shown by a "Normal scores " plot [8]). The iterative weighting was performed in every regression analysis. The use of the arcsine transform led to a finite value (ft/2) for the transformed response of the patient with 100% peak twitch depression. However, there remained the problem that the dose of 30 ug kg"1 which this patient received may have been more than enough to produce 100% depression and, therefore, that the 100% observed value in figure 1 should, perhaps, be plotted a little to the left. If so, this would imply also that the slope of the sigmoid curve was a little too gentle (ED,8 a little larger than it should be). Therefore there appears to be a case for omitting this datum. In fact, however, the 100% depression was achieved with a dose that was much smaller than predicted by any of the four fined sigmoid curves (fig. 1). Therefore, omining this result would make the slope of the sigmoid curves even less. Omission of the point increased the ED, 6 from 51.3 to 52.2 (ig kg"1, therefore all the regressions were based on the data for all 80 patients.

ACKNOWLEDGEMENTS Thanks to Professor W. W. Mapleson and Dr R. G. Newcombe for statistical help and Mr G. Smith of Medtel Australia for loan of equipment.

REFERENCES 1. Katz RL, Norman J, Seed PF, Conrad L. A comparison of the effects of suxamethonium and rubocurarine in patients in London and New York. British Journal of Anaesthesia 1969; 41: 1041-1047. 2. Fiset P, Donati F, Balendran P, Meistelman C, Lira E, Bevan DR. Vecuronium is more potent in Montreal than in Paris. Canadian Journal of Anaesthesia 1991; 38: 717-721. 3. Houghton IT, Aun CST, Oh TE. Vecuronium: an anthropometric comparison. Anaesthesia 1992; 47: 741—746. 4. Mapleson WW, Use of GLIM and the bootstrap in assessing a clinical trial of two drugs. Statistics in Medicine 1986; 5: 363-374. 5. Walts LF, Drexler H. Response to curare: a study of national groups. Anesthesia and Analgesia 1973; 52: 753-756. 6. Donati F, Bevan DR. The influence of patient's sex, age and weight on pancuronium onset time. Canadian Anaesthetists Society Journal 1986; 33: S86. 7. Mapleson WW, Mushin WW. Relaxant action in man, an experimental study; II. Results with intravenous gallamine triethiodide. Anaesthesia 1955; 10: 379-390. 8. Armitage P, Berry G. Statistical Methods in Medical Research, 2nd Edn. Oxford: Blackwell, 1987; 364-365, 416. 9. Payne CD. The GLIM System Release3.77: Manual. Oxford: Numerical Algorithms Group, 1987. 10. Healy MJR. GLIM: an Introduction. Oxford: Clarendon Press, 1988; 94-99. 11. Robertson EN, Fragen RJ, Booij LH, Van Egmond J, Crul JF. Some effects of diisopropyl pheno (ICI 35868) on the pharmacodynamics of arracurium and vecuronium in anaesthetized man. British Journal of Anaesthesia 1983; 55: 723-728. 12. DeGrood PM, Van Egmond J, Van De Wetering M, Van Beem HB, Booij LH, Crul JF. Lack of effects of emulsified propofol on vecuronium pharmacodynamics—preliminary results in man. Postgraduate Medical Journal 1985; 61 (Suppl. 3): 28-30.

Downloaded from http://bja.oxfordjournals.org/ at Université Laval on July 12, 2015

time has been reported to be shorter in females [6]. The question arises: is this a true gender difference or is there a confounding variable, associated with gender? The most likely variable is this respect is the known difference in body build between the sexes— females having a greater percentage of fat than males. This explanation is supported by anthropometric data in a recent study, which also found easier tracheal intubating conditions in women than in men [3]. Because we recorded only height and weight, our analysis of the effect of body build was necessarily limited. When gender was omitted from the predictors, body mass index (weight/height2) was a poor substitute, as were height, height2 and height3, with or without weight. It is interesting that the effect of body weight on the maximum twitch depression can be eliminated if dose is expressed in terms of ug kg"" with n equal to 0.68 (fig. 2). This is close to, and consistent with, the value of 0.67 which would be expected theoretically if peak effect depended on metabolic rate, and if that were proportional to surface area which, for a given body shape, is proportional to weight2'3—that is, weight087. However, with 95% confidence limits of 0.40-0.95, the results are also consistent with other hypotheses—but not with the hypothesis that the effect of a given dose is independent of body weight (r = —4.3, P < 0.0001), or that the requisite dose is proportional to weight (r = 2.37, P = 0.020). We conclude that any difference in sensitivity to vecuronium between Caucasian lifelong residents of the United Kingdom and of Australia is probably too small to be important, but that females are about 22% more sensitive than males.

193

194 13. McCarthy G, Mirakhur RK, Pandit SK. Lack of interaction between propofol and vecuronium. Anesthesia and Analgesia 1992; 75: 536—538. 14. O'Hara DA, Fragen RJ, Shanks CA. The effects of age on the dose-response curves for vecuronium in adults. Anesthtsiology 1985; 63: 542-544. 15. Tham EJ, Morris S, Wright EM, Campbell I A, Mapleson

BRITISH JOURNAL OF ANAESTHESIA WW. An assessment of prilocaine as a topical anaesthetic agent for fibreoptic bronchoscopy in comparison with lidocaine. Ada Anaesthesiologica Scandinavica 1993; (in press). 16. Davies OL, Goldsmith PL (eds). Statistical Methods in Research and Production, 4th Edn. London: Longman, 1977; 236.

Downloaded from http://bja.oxfordjournals.org/ at Université Laval on July 12, 2015