Anaesthesia, 1992, Volume 47, pages 303-306
The relationship between the concentration of temazepam in cerebrospinal fluid and sedation in man
G. A. OSBORNE, N. R. BADCOCK, P. M. McGRATH, W. J. RUSSELL
AND
D. B. FREWIN
Summary
Twenty-six patients received oral temazepam and subsequently spinal anaesthesia. Blood and lumbar cerebrospinal fluid temazepam levels were measured together with the degree of sedation. The plasma and cerebrospinal fluid concentrations correlated well with the temazepam dose but even better with the weight standardised dose ( r = 0.65, p = 0.0003 and r = 0.75, p = 0.00001 respectively). Both the plasma and cerebrospinal Puid concentrations of temazepam were correlated with the patient’s sedation ( r = 0.42 p = 0.037, and r = 0.46 p = 0.021 respectively), but neither was strong. Thus, although the drug Concentration at the receptor may be a major factor in producing sedation, others factors, possibly the receptor population or their responsiveness, are also important contributors. Key words
Hypnotics, benzodiazepines; temazepam. Pharmacokinetics; distribution.
Benzodiazepines are well recognised for their variable dose response and several studies have examined the relationship between plasma level and effect. Bradshaw et al. reported a modest correlation between plasma concentration and the clinical effect of lorazepam [I], while another study on temazepam from the same group in 1989, claimed a good relationship between these two parameters [2]. However, in a recent review Dingemanse et al. stated that ‘data on concentration-effect relationships of CNS active drugs are quite limited’ [3]. The physicochemical properties of benzodiazepines are suitable for studying dose-effect relationships and examining variations caused by pharmacokinetic and pharmacodynamic effects. In particular, benzodiazepines with appropriate pKa values and lipid solubilities can penetrate rapidly into the central nervous system (CNS) and reach a rapid equilibrium between blood and all regions of the CNS. The normal protein level in cerebrospinal fluid (CSF) is only about 0.010 g.IOOml-’ i.e. about 1000 times less than in blood, so CSF matches closely the unbound plasma level. Animal studies with diazepam [4] and clonazepam [5]
have shown that C S F and blood can reach a rapid equilibrium and give a close correlation between free (unbound) drug concentrations in blood and the CSF level. We have recently shown a good correlation (r = 0.81) between the measured fraction of unbound temazepam in plasma and the level of temazepam in C S F in 13 human subjects administered temazepam 20 mg, 75 to 240 min before venous blood and lumbar C S F sampling [6]. Therefore, the concentration of benzodiazepines in the CSF should be a good measure of the concentration to which the neurones within the CNS are exposed, and measurement of the CSF drug concentration should give a n estimate of the concentrationleffect relationship. To explore this relationship, we have studied patients undergoing a lumbar puncture for spinal anaesthesia. The lumbar CSF temazepam levels after oral dosing have been taken to represent the ‘biophase’ drug concentrations to which the CNS benzodiazepine receptors are exposed, and used to evaluate the drug concentration-effect relationship, independent of pharmacokinetic variation. Temazepam (3-Hydroxydiazepam) is a commonly pre-
G.A. Osborne MB, BS, PhD, FFARACS, Staff Specialist, P.M. McGrath, MB, BS, FFARACS, Senior Registrar, W.J. Russell, MB, BS, FCAnaes, DIC, PhD, FFARACS, Director R & D, Department of Anaesthesia and Intensive Care, Royal Adelaide Hospital, D.B. Frewin, MB, BS, MD, FRACP, Professor of Clinical Pharmacology and Dean, Faculty of Medicine, University of Adelaide, North Terrace, Adelaide, South Australia 5000. N.R. Badcock, BSc, Food Technol. Assoc. Certif., Principal Hospital Scientist, Department of Chemical Pathology, Adelaide Children’s Hospital, King William Road, North Adelaide, South Australia 5006. Correspondence should be addressed to Dr W.J. Russell please. Accepted 8 August 1991. 0003-2409/92/040303
+ 04 $03.00/0
@ 1992 The Association of Anaesthetists of G t Britain and Ireland
303
304
G.A. Osborne et al.
scribed premedicant for spinal anaesthesia as it has a rapid onset of action and a reasonably short elimination half-life of about 8.7 h in the older age group [7]. These favourable pharmacokinetic properties are associated with good anxiolysis and sedation during anaesthesia and surgery. In this study, drug effect was measured in terms of the degree of sedation.
Materials and methods
Results Group data The mean age of the 26 patients was 69.7 years (SD 8.4) with a range of 51 to 88 years. Their mean weight was 76.4 kg (SD 7.5) with a range of 65 to 95 kg. The mean time between oral premedication and the initial venous blood sample was 137 mins (SD 54) with a range of 70 to 250 min.
Subjects
Pharmacokinetic data
Twenty-six males scheduled to undergo urological surgery under spinal anaesthesia were premedicated with oral temazepam. The study had the approval of the ethics committee of the Royal Adelaide Hospital and informed consent was obtained from each patient. Each patient received a single dose which varied from 10 to 40 mg. Drug administration occurred 1 to 4 h before arrival in the anaesthetic room. N o other premedication was prescribed. Subjects who had CNS disease such as a history of a previous cerebrovascular accident or who were receiving CNS active drugs such as anticonvulsants or regular night sedation were not studied. Upon arrival, the degree of sedation was assessed using a six level scale (Table I). An intravenous cannula was then inserted into a forearm vein and 5 ml of venous blood were taken and placed in a tube containing lithium heparin for temazepam assay. The patient was positioned in flexion and a spinal needle was inserted. One to two ml of CSF were removed immediately before injection of the spinal anaesthetic. For all patients, the spinal fluid was free of visible blood and the CSF sample was taken within 10-20min of the venous blood sample. The blood was centrifuged within 0.5 h of collection, and the separated plasma and CSF was protected from light and stored at - 20°C until assayed.
The mean dose adjusted for body weight was 0.339mg.kg-' (SD 0.099) with a range of 0.123 to 0.597 mg.kg-'. The mean plasma level of temazepam was 1284nmol.l-' (SD 1036) with a range of 296 to 5795 nmol.1-'. The relationship between the dose and plasma level was only moderate (r = 0.48, p = 0.014) but improved when the dose was adjusted for body weight, Figure 1 (r = 0.65 p = 0.00034). The mean CSF drug concentration was 68.5 nmol.1-' (SD 38.8) with a range of 18 to 217 nmol.1-'. The relationship between the temazepam dose and the C S F level was fair (r = 0.59 p = 0.002) and improved when adjusted for body weight (r = 0.75 p = O.OOOOl), Figure 2. The mean CSF/plasma ratio was 6.16% (SD 2.42%) and varied from 3.0% to 12.5%. These values were similar to those previously observed 161.
Pharmacodynamic data The patient sedation scores ranged from 1 (alert) to 4 (very drowsy, rouses when spoken to but does not converse). N o patient reached a sedation score of 5 or 6. There was a modest correlation between the plasma temazepam level and the degree of sedation, r = 0.42, p = 0.037 (Spearman), Figure 3. However, the relationship between the CSF concentration and the degree of sedation was only slightly better r = 0.46, p = 0.021, Figure 4.
Assay methodology Plasma and CSF temazepam concentrations were measured in duplicate using capillary gas chromatography. This method has been described in a previous paper [6]. The response of the system used was linear from 3-3000nmol.l-' with a lower limit of detection of 2 nmol.1-I. The between-run coefficient of variation was 4.2% at a concentration of 700 nmo1.l-I.
Discussion This study has provided some interesting data on both the kinetics and dynamics of temazepam in man. The vari0
i
'Oo0L 5000
Statistics Analysis was performed using parametric testing for interval data such as blood levels and adjusted body dose, and nonparametric statistics for correlations with the ordinal data. The correlation between CSF drug concentration and the level of sedation was assessed by a Spearman's nonparametric correlation with replication correction. Table 1. Guideline to sedation score. 1. Alert. 2. Claims to be slightly drowsy but is not obviously so. 3. Drowsy, but rouses and answers questions. 4. Very drowsy, rouses when spoken t o but does not converse. 5. Asleep, responds to mildly painful stimuli, opens eyes. 6. Asleep, no response to mildly painful stimuli.
3000
s n
I 0
/
3 0 s
F Ternweporn dose mg. kg-'
Fig. 1. Relationship between temazepam dose in mg.kg-' and total plasma drug concentration. This relationship has a good fit (r = 0.65, p = 0.00034) with the linear equation; total plasma concentration. = 6825 x dose mg.kg-' - 1029.
CSF temazepam levels and sedation in man
< 2 0
[03 0
200
.-s
'i
O
3t
305
Q
ww
00
0
-
t
!!
c
a c
;100II. v)
-
V
0
o o o w
0.1
0.2
0.3
0.4
.
0.5
0.6
Ternazeparn dose rng k g - '
Fig. 2. Relationship between the weight standardised temazepam dose and CSF drug concentration. The relationship has a good fit (r = 0.75, p = 0.OOOOl) with the linear equation; CSF concentration. = 293.7 xdose mg.kg-' - 31.00.
ability of the CSF/plasma temazepam concentration ratio is not surprising, as significant fluctuation in this ratio has been observed for this and other benzodiazepines in the past [S,81, although little information is available for humans taking temazepam. The mean CSF/plasma concentration ratio showed a fourfold variation (with its implied equivalent alteration in the proportion of free plasma drug level). This is presumably caused by differences in drug binding to plasma proteins in each patient. On theoretical grounds this should have a major impact on drug effect. Thus, for a given dose, an alteration in plasma protein binding could be expected to cause an inverse change in the free drug level and therefore the clinical effect. The change in binding could be expected to alter the plasma concentration-effect relationships. However, variation in plasma protein binding should not affect free drug concentration-effect relationships. The present study uses the CSF level of temazepam as an index of the free drug level, which should represent the concentration at the receptor. However, there is little improvement in the correlation between the CSF concentration and the sedative effect of the drug compared with its
0
43
I00
0 0
0
I000
00
0
10 000
Total plasma concentration nmol .I-'
Fig. 3. Relationship between the total plasma temazepam concentration and sedation. Overall there is a moderate positive correlation ( r = 0.42, p = 0.037 Spearman).
0
I t 01 10
I
I
I
I
I I I I I
I
I00 CSF concentration nrnol
I
I
I
1 1 1 1 1
1000
.I-'
Fig. 4. Relationship between the C S F temazepam concentration and sedation. Overall there is a moderate positive correlation (r = 0.46, p = 0.021 Spearman).
plasma concentration and effect. It is surprising that even the CSF concentration is only a modest predictor of the sedative effect. With the correlation of 0.46 between CSF concentration and sedation, much of the sedative effect remains unaccounted for. Thus, although the concentration of temazepam may set the level of sedation in an individual, as a group, the sedative effect in these subjects is not tightly coupled t o the CSF temazepam concentration. A large proportion of the effect is being influenced by other factors. Factors which could alter the response of temazepam are, differences in the population of CNS benzodiazepine receptors or differences in receptor sensitivity between patients. Other factors which might reduce the CSF concentration-effect correlation include experimental errors in drug assay or blood and CSF sampling, limitations in the precision of the sedation measurements, a difference in C S F temazepam concentrations between the lumbar and cerebral regions, variation in C S F protein, C S F pH values, or receptor upregulation/downregulation caused by chronic use of other drugs such as alcohol. None of these possibilities seems to be substantial enough to explain the difference. Indeed, only the sedation estimate seems susceptible to significant measurement error, and variability from the imprecision of the sedative measurement must be acknowledged. However, although subjective and other observer variability may explain a difference between one level and an adjacent level, the overlap of CSF drug concentrations between level 1 and levels 3 and 4 sedation and that between level 2 and level 4 sedation is harder to explain in this way. More objective measurements of CNS impairment such as critical flicker fusion could be used to improve reliability, but the clinical effect sought is sedation, which is why in this study it was chosen as the effect parameter. The only modest correlation between CSF drug concentration and effect is a significant observation. This is in conflict with the simple notion of the amount of drug at the receptor being constant for a given effect. It suggests that there may be an important element in the variability of benzodiazepine action which is a receptor or cellular response phenomenon. In the past, variability in effect has been focused on variability in absorption, first-pass meta-
306
G.A. Osborne et al.
bolism, protein binding, and/or CNS penetration. However, the measurement of temazepam in the C S F goes beyond the point at which conventional bioavailability and other pharmacokinetic factors can influence the response. The use of CSF sampling (and the examination of CSF drug level/effect relationship) as we have employed, should detect alterations in receptor numbers or responsiveness in patients. This technique should be able to detect alterations either as a consequence of genetic factors or of acquired factors such as previous chronic benzodiazepine exposure. A decrease in receptor numbers or sensitivity should be revealed as a flatter slope on the concentration/sedation curve. There is some evidence in animal studies that chronic treatment with benzodiazepines is associated with decreased receptor binding [9]. However, there is no evidence as yet for this phenomenon in man. We feel that much of the variation we have seen in CSF drug concentration-effect is likely to be due to true pharmacodynamic variability. The CSF drug concentrationeffect model that we have used should be further exploited in studies on receptor-agonist relationships for other drugs with CNS actions to see whether the drug concentration at the receptor causes only a minor influence as it seems to d o with temazepam.
Acknowledgments Some financial support from Wyeth Pharmaceuticals (Australia) Pty. Ltd. for this study is gratefully acknowledged. The authors also wish to thank D r A. A. Somogyi for editorial advice.
References [I] BRADSHAWEG, ALI AA, MULLEYBA, RYE RM. Plasma concentrations and clinical effects of lorazepam after oral administration. Brilish Journal of Anaeslhesia 1981; 53: 517-22. [2] RATCLIFFA, INDALO AA, BRADSHAW EG, RYE RM. Premedication with temazepam in minor surgery. Anaesthesia 1989, 44: 812-15. [3] DINGEMANSE J, DANHOF M, BREIMERDD. Pharmacokineticpharmacodynamic modelling of CNS drug effects: an overview. Pharmacology and Therapeulics 1988, 38: 1-52. [4] GREENBLATT DJ, OCHSHR, LLOYDBK. Entry of diazepam and its major metabolite into cerebrospinal fluid. Psychopharmacology 1980; 7 0 89-93. [5] PARRYGJG. Concentration of clonazepam in serum and cerebrospinal fluid of the sheep. Pharmacology 1977; 15: 318-23. [6] BADCOCK NR, OSBORNE GA, NYMAN TLM, SANWMLN, RUSSELL WJ, FREWINDB. Plasma and cerebrospinal fluid concentrations of temazepam following oral drug administration. European Journal of Clinical Pharmacology 1990; 38: 153-5. [7] KLEMK, MURRAY GR, LAAKEK. Pharmacokinetics of temazepam in geriatric patients. European Journal of Clinical Pharmacology 1986; 30: 745-7. [8] STANSKI DR, G R E E N B LDJ, A ~SELWYN A, SHADER RI, FRANKE K, KOCH-WESERJ. Plasma and cerebrospinal fluid concentration of chlordiazepoxide and its metabolites in surgical patients. Clinical Pharmacology and Therapeulics 1976; 2 0 571-8. [9] ROSENBERG HC, CHIU TH. Tolerance during chronic benzodiazepine treatment associated with decreased receptor binding. European Journal of Pharmacology 198I ; 7 0 453-60.
The relationship between the concentration of temazepam in cerebrospinal fluid and sedation in man
G. A. OSBORNE, N. R. BADCOCK, P. M. McGRATH, W. J. RUSSELL
AND
D. B. FREWIN
Summary
Twenty-six patients received oral temazepam and subsequently spinal anaesthesia. Blood and lumbar cerebrospinal fluid temazepam levels were measured together with the degree of sedation. The plasma and cerebrospinal fluid concentrations correlated well with the temazepam dose but even better with the weight standardised dose ( r = 0.65, p = 0.0003 and r = 0.75, p = 0.00001 respectively). Both the plasma and cerebrospinal Puid concentrations of temazepam were correlated with the patient’s sedation ( r = 0.42 p = 0.037, and r = 0.46 p = 0.021 respectively), but neither was strong. Thus, although the drug Concentration at the receptor may be a major factor in producing sedation, others factors, possibly the receptor population or their responsiveness, are also important contributors. Key words
Hypnotics, benzodiazepines; temazepam. Pharmacokinetics; distribution.
Benzodiazepines are well recognised for their variable dose response and several studies have examined the relationship between plasma level and effect. Bradshaw et al. reported a modest correlation between plasma concentration and the clinical effect of lorazepam [I], while another study on temazepam from the same group in 1989, claimed a good relationship between these two parameters [2]. However, in a recent review Dingemanse et al. stated that ‘data on concentration-effect relationships of CNS active drugs are quite limited’ [3]. The physicochemical properties of benzodiazepines are suitable for studying dose-effect relationships and examining variations caused by pharmacokinetic and pharmacodynamic effects. In particular, benzodiazepines with appropriate pKa values and lipid solubilities can penetrate rapidly into the central nervous system (CNS) and reach a rapid equilibrium between blood and all regions of the CNS. The normal protein level in cerebrospinal fluid (CSF) is only about 0.010 g.IOOml-’ i.e. about 1000 times less than in blood, so CSF matches closely the unbound plasma level. Animal studies with diazepam [4] and clonazepam [5]
have shown that C S F and blood can reach a rapid equilibrium and give a close correlation between free (unbound) drug concentrations in blood and the CSF level. We have recently shown a good correlation (r = 0.81) between the measured fraction of unbound temazepam in plasma and the level of temazepam in C S F in 13 human subjects administered temazepam 20 mg, 75 to 240 min before venous blood and lumbar C S F sampling [6]. Therefore, the concentration of benzodiazepines in the CSF should be a good measure of the concentration to which the neurones within the CNS are exposed, and measurement of the CSF drug concentration should give a n estimate of the concentrationleffect relationship. To explore this relationship, we have studied patients undergoing a lumbar puncture for spinal anaesthesia. The lumbar CSF temazepam levels after oral dosing have been taken to represent the ‘biophase’ drug concentrations to which the CNS benzodiazepine receptors are exposed, and used to evaluate the drug concentration-effect relationship, independent of pharmacokinetic variation. Temazepam (3-Hydroxydiazepam) is a commonly pre-
G.A. Osborne MB, BS, PhD, FFARACS, Staff Specialist, P.M. McGrath, MB, BS, FFARACS, Senior Registrar, W.J. Russell, MB, BS, FCAnaes, DIC, PhD, FFARACS, Director R & D, Department of Anaesthesia and Intensive Care, Royal Adelaide Hospital, D.B. Frewin, MB, BS, MD, FRACP, Professor of Clinical Pharmacology and Dean, Faculty of Medicine, University of Adelaide, North Terrace, Adelaide, South Australia 5000. N.R. Badcock, BSc, Food Technol. Assoc. Certif., Principal Hospital Scientist, Department of Chemical Pathology, Adelaide Children’s Hospital, King William Road, North Adelaide, South Australia 5006. Correspondence should be addressed to Dr W.J. Russell please. Accepted 8 August 1991. 0003-2409/92/040303
+ 04 $03.00/0
@ 1992 The Association of Anaesthetists of G t Britain and Ireland
303
304
G.A. Osborne et al.
scribed premedicant for spinal anaesthesia as it has a rapid onset of action and a reasonably short elimination half-life of about 8.7 h in the older age group [7]. These favourable pharmacokinetic properties are associated with good anxiolysis and sedation during anaesthesia and surgery. In this study, drug effect was measured in terms of the degree of sedation.
Materials and methods
Results Group data The mean age of the 26 patients was 69.7 years (SD 8.4) with a range of 51 to 88 years. Their mean weight was 76.4 kg (SD 7.5) with a range of 65 to 95 kg. The mean time between oral premedication and the initial venous blood sample was 137 mins (SD 54) with a range of 70 to 250 min.
Subjects
Pharmacokinetic data
Twenty-six males scheduled to undergo urological surgery under spinal anaesthesia were premedicated with oral temazepam. The study had the approval of the ethics committee of the Royal Adelaide Hospital and informed consent was obtained from each patient. Each patient received a single dose which varied from 10 to 40 mg. Drug administration occurred 1 to 4 h before arrival in the anaesthetic room. N o other premedication was prescribed. Subjects who had CNS disease such as a history of a previous cerebrovascular accident or who were receiving CNS active drugs such as anticonvulsants or regular night sedation were not studied. Upon arrival, the degree of sedation was assessed using a six level scale (Table I). An intravenous cannula was then inserted into a forearm vein and 5 ml of venous blood were taken and placed in a tube containing lithium heparin for temazepam assay. The patient was positioned in flexion and a spinal needle was inserted. One to two ml of CSF were removed immediately before injection of the spinal anaesthetic. For all patients, the spinal fluid was free of visible blood and the CSF sample was taken within 10-20min of the venous blood sample. The blood was centrifuged within 0.5 h of collection, and the separated plasma and CSF was protected from light and stored at - 20°C until assayed.
The mean dose adjusted for body weight was 0.339mg.kg-' (SD 0.099) with a range of 0.123 to 0.597 mg.kg-'. The mean plasma level of temazepam was 1284nmol.l-' (SD 1036) with a range of 296 to 5795 nmol.1-'. The relationship between the dose and plasma level was only moderate (r = 0.48, p = 0.014) but improved when the dose was adjusted for body weight, Figure 1 (r = 0.65 p = 0.00034). The mean CSF drug concentration was 68.5 nmol.1-' (SD 38.8) with a range of 18 to 217 nmol.1-'. The relationship between the temazepam dose and the C S F level was fair (r = 0.59 p = 0.002) and improved when adjusted for body weight (r = 0.75 p = O.OOOOl), Figure 2. The mean CSF/plasma ratio was 6.16% (SD 2.42%) and varied from 3.0% to 12.5%. These values were similar to those previously observed 161.
Pharmacodynamic data The patient sedation scores ranged from 1 (alert) to 4 (very drowsy, rouses when spoken to but does not converse). N o patient reached a sedation score of 5 or 6. There was a modest correlation between the plasma temazepam level and the degree of sedation, r = 0.42, p = 0.037 (Spearman), Figure 3. However, the relationship between the CSF concentration and the degree of sedation was only slightly better r = 0.46, p = 0.021, Figure 4.
Assay methodology Plasma and CSF temazepam concentrations were measured in duplicate using capillary gas chromatography. This method has been described in a previous paper [6]. The response of the system used was linear from 3-3000nmol.l-' with a lower limit of detection of 2 nmol.1-I. The between-run coefficient of variation was 4.2% at a concentration of 700 nmo1.l-I.
Discussion This study has provided some interesting data on both the kinetics and dynamics of temazepam in man. The vari0
i
'Oo0L 5000
Statistics Analysis was performed using parametric testing for interval data such as blood levels and adjusted body dose, and nonparametric statistics for correlations with the ordinal data. The correlation between CSF drug concentration and the level of sedation was assessed by a Spearman's nonparametric correlation with replication correction. Table 1. Guideline to sedation score. 1. Alert. 2. Claims to be slightly drowsy but is not obviously so. 3. Drowsy, but rouses and answers questions. 4. Very drowsy, rouses when spoken t o but does not converse. 5. Asleep, responds to mildly painful stimuli, opens eyes. 6. Asleep, no response to mildly painful stimuli.
3000
s n
I 0
/
3 0 s
F Ternweporn dose mg. kg-'
Fig. 1. Relationship between temazepam dose in mg.kg-' and total plasma drug concentration. This relationship has a good fit (r = 0.65, p = 0.00034) with the linear equation; total plasma concentration. = 6825 x dose mg.kg-' - 1029.
CSF temazepam levels and sedation in man
< 2 0
[03 0
200
.-s
'i
O
3t
305
Q
ww
00
0
-
t
!!
c
a c
;100II. v)
-
V
0
o o o w
0.1
0.2
0.3
0.4
.
0.5
0.6
Ternazeparn dose rng k g - '
Fig. 2. Relationship between the weight standardised temazepam dose and CSF drug concentration. The relationship has a good fit (r = 0.75, p = 0.OOOOl) with the linear equation; CSF concentration. = 293.7 xdose mg.kg-' - 31.00.
ability of the CSF/plasma temazepam concentration ratio is not surprising, as significant fluctuation in this ratio has been observed for this and other benzodiazepines in the past [S,81, although little information is available for humans taking temazepam. The mean CSF/plasma concentration ratio showed a fourfold variation (with its implied equivalent alteration in the proportion of free plasma drug level). This is presumably caused by differences in drug binding to plasma proteins in each patient. On theoretical grounds this should have a major impact on drug effect. Thus, for a given dose, an alteration in plasma protein binding could be expected to cause an inverse change in the free drug level and therefore the clinical effect. The change in binding could be expected to alter the plasma concentration-effect relationships. However, variation in plasma protein binding should not affect free drug concentration-effect relationships. The present study uses the CSF level of temazepam as an index of the free drug level, which should represent the concentration at the receptor. However, there is little improvement in the correlation between the CSF concentration and the sedative effect of the drug compared with its
0
43
I00
0 0
0
I000
00
0
10 000
Total plasma concentration nmol .I-'
Fig. 3. Relationship between the total plasma temazepam concentration and sedation. Overall there is a moderate positive correlation ( r = 0.42, p = 0.037 Spearman).
0
I t 01 10
I
I
I
I
I I I I I
I
I00 CSF concentration nrnol
I
I
I
1 1 1 1 1
1000
.I-'
Fig. 4. Relationship between the C S F temazepam concentration and sedation. Overall there is a moderate positive correlation (r = 0.46, p = 0.021 Spearman).
plasma concentration and effect. It is surprising that even the CSF concentration is only a modest predictor of the sedative effect. With the correlation of 0.46 between CSF concentration and sedation, much of the sedative effect remains unaccounted for. Thus, although the concentration of temazepam may set the level of sedation in an individual, as a group, the sedative effect in these subjects is not tightly coupled t o the CSF temazepam concentration. A large proportion of the effect is being influenced by other factors. Factors which could alter the response of temazepam are, differences in the population of CNS benzodiazepine receptors or differences in receptor sensitivity between patients. Other factors which might reduce the CSF concentration-effect correlation include experimental errors in drug assay or blood and CSF sampling, limitations in the precision of the sedation measurements, a difference in C S F temazepam concentrations between the lumbar and cerebral regions, variation in C S F protein, C S F pH values, or receptor upregulation/downregulation caused by chronic use of other drugs such as alcohol. None of these possibilities seems to be substantial enough to explain the difference. Indeed, only the sedation estimate seems susceptible to significant measurement error, and variability from the imprecision of the sedative measurement must be acknowledged. However, although subjective and other observer variability may explain a difference between one level and an adjacent level, the overlap of CSF drug concentrations between level 1 and levels 3 and 4 sedation and that between level 2 and level 4 sedation is harder to explain in this way. More objective measurements of CNS impairment such as critical flicker fusion could be used to improve reliability, but the clinical effect sought is sedation, which is why in this study it was chosen as the effect parameter. The only modest correlation between CSF drug concentration and effect is a significant observation. This is in conflict with the simple notion of the amount of drug at the receptor being constant for a given effect. It suggests that there may be an important element in the variability of benzodiazepine action which is a receptor or cellular response phenomenon. In the past, variability in effect has been focused on variability in absorption, first-pass meta-
306
G.A. Osborne et al.
bolism, protein binding, and/or CNS penetration. However, the measurement of temazepam in the C S F goes beyond the point at which conventional bioavailability and other pharmacokinetic factors can influence the response. The use of CSF sampling (and the examination of CSF drug level/effect relationship) as we have employed, should detect alterations in receptor numbers or responsiveness in patients. This technique should be able to detect alterations either as a consequence of genetic factors or of acquired factors such as previous chronic benzodiazepine exposure. A decrease in receptor numbers or sensitivity should be revealed as a flatter slope on the concentration/sedation curve. There is some evidence in animal studies that chronic treatment with benzodiazepines is associated with decreased receptor binding [9]. However, there is no evidence as yet for this phenomenon in man. We feel that much of the variation we have seen in CSF drug concentration-effect is likely to be due to true pharmacodynamic variability. The CSF drug concentrationeffect model that we have used should be further exploited in studies on receptor-agonist relationships for other drugs with CNS actions to see whether the drug concentration at the receptor causes only a minor influence as it seems to d o with temazepam.
Acknowledgments Some financial support from Wyeth Pharmaceuticals (Australia) Pty. Ltd. for this study is gratefully acknowledged. The authors also wish to thank D r A. A. Somogyi for editorial advice.
References [I] BRADSHAWEG, ALI AA, MULLEYBA, RYE RM. Plasma concentrations and clinical effects of lorazepam after oral administration. Brilish Journal of Anaeslhesia 1981; 53: 517-22. [2] RATCLIFFA, INDALO AA, BRADSHAW EG, RYE RM. Premedication with temazepam in minor surgery. Anaesthesia 1989, 44: 812-15. [3] DINGEMANSE J, DANHOF M, BREIMERDD. Pharmacokineticpharmacodynamic modelling of CNS drug effects: an overview. Pharmacology and Therapeulics 1988, 38: 1-52. [4] GREENBLATT DJ, OCHSHR, LLOYDBK. Entry of diazepam and its major metabolite into cerebrospinal fluid. Psychopharmacology 1980; 7 0 89-93. [5] PARRYGJG. Concentration of clonazepam in serum and cerebrospinal fluid of the sheep. Pharmacology 1977; 15: 318-23. [6] BADCOCK NR, OSBORNE GA, NYMAN TLM, SANWMLN, RUSSELL WJ, FREWINDB. Plasma and cerebrospinal fluid concentrations of temazepam following oral drug administration. European Journal of Clinical Pharmacology 1990; 38: 153-5. [7] KLEMK, MURRAY GR, LAAKEK. Pharmacokinetics of temazepam in geriatric patients. European Journal of Clinical Pharmacology 1986; 30: 745-7. [8] STANSKI DR, G R E E N B LDJ, A ~SELWYN A, SHADER RI, FRANKE K, KOCH-WESERJ. Plasma and cerebrospinal fluid concentration of chlordiazepoxide and its metabolites in surgical patients. Clinical Pharmacology and Therapeulics 1976; 2 0 571-8. [9] ROSENBERG HC, CHIU TH. Tolerance during chronic benzodiazepine treatment associated with decreased receptor binding. European Journal of Pharmacology 198I ; 7 0 453-60.
