pharmacokinetics of closely related benzodiazepines

Br. J. clin. Pharmac. (1979), 8, 15S-21S

PHARMACOKINETICS OF CLOSELY RELATED BENZODIAZEPINES S.H. CURRY & R. WHELPTON Department of Pharmacology and Therapeutics, The London Hospital Medical College, Turner Street, London El 2AD, UK

1 It is commonplace for drugs to vary by only minor chemical differences. This is particularly so for those seven benzodiazepines discussed in this paper which are related both as precursors and as metabolites. However, minor chemical differences may cause major differences in physicochemical and pharmacodynamic properties. 2 Although the physicochemical differences are difficult to relate to effect, the influence of structure on absorption, distribution and elimination is of considerable importance in governing duration of effect, as shown by studies in monkeys and in man. This in turn dictates the suitability of a particular drug as a day or night sedative, as an anticonvulsant, or as an anxiolytic. 3 Structure affects the relative potency of the compounds as anticonvulsants, anxiolytics or sedatives so that judicious choice of a particular compound for a particular patient and condition will lead to improved therapy. It is fallacious to consider all benzodiazepines as similar.

Introduction THERE are many groups of drugs in which the individual compounds are clinically similar. Within such groups there may be only the most minor chemical differences. Minor chemical differences are, however, commonly exploited to increase the number of commercially available drugs, as they can lead to major pharmacokinetic, physicochemical and pharmacodynamic differences. This is so to some extent with the benzodiazepines which are of particular interest in this context because many of the compounds are so chemically similar that they are related as precursors and metabolites. This paper is concerned with some of the general principles involved in exploitation of small differences, with particular, but not exclusive, reference to the compounds in Figure 1. It is first necessary to mention some elementary pharmacokinetic principles.

clearly more complex than any one- or two-compartment model, but data are rarely adequate to demonstrate such complexity. The commonest compartmental subdivision in the multi-compartment case is into a 'central' compartment, consisting of blood plus highly vascular tissues (for example, heart, liver, lungs and glands) and a 'peripheral' compartment consisting of tissues of lesser vascularity (for example, voluntary muscle). The position with regard to brain is equivocal, as it is of high vascularity, but drug penetration is limited by the blood-brain barrier. However, as most useful centrally acting drugs cross the blood-brain barrier rapidly, the brain is generally considered as part of the central compartment. This is especially relevant to oral dosing, as the rate of absorption, not the rate of penetration of the blood-brain barrier, is usually the rate-limiting factor in transfer of drug molecules to this particular set of sites of action.

Pharmacokinetic principles

The one- and two-compartment models are represented diagrammatically in Figure 2. Intravenous dosing into a one-compartment system is followed by instantaneous (in this context 'instantaneous' obviously means 'so brief as to be of undetectable duration') mixing within the compartment, and then decay of the drug concentration within the single compartment, usually assessed by venous plasma drug concentration measurement. Decay is usually exponential. Intravenous dosing into the central compartment © Macmillan Journals Ltd. 1979

When drugs are given intravenously, their concentrations in plasma usually indicate that the body distributes them in accordance with either a onecompartment or a two-compartment model. Occasionally a three-compartment model is invoked. Drugs given orally or intramuscularly produce concentrations indicating similar behaviour, with the added influence of absorption at a finite rate. It should be emphasized at the outset that the body is 0306-5251/79/160015-07 $01.00

16S

S.H. CURRY & R. WHELPTON

CH3 N CI

Medazepam CH3

CH3

NHCH3

N-/ -6

0OH

0

Chlordiazepoxide via intermediates

1 Temazepam

Diazepam

H N-..,0

N 0

H

O0

OK

Desmethyldiazepam

Oxazepam

Cl

Oxazepam glucuronides Clorazepate dipotassium Figure

1

Chemical formulae for and metabolic relationships among seven closely related benzodiazepines.

of a two-compartment system is followed by instantaneously mixing within the central compartment, then decay of the concentration in this compartment in two phases. Blood sampling is obviously from the central compartment. The first of the decay phases is the one of faster decay. Both phases are usually exponential. The first phase relates primarily to penetration of the peripheral compartment. The second phase relates to elimination by metabolism and/or excretion from the system as a whole, although elimination actually occurs within the central compartment. Of course, elimination and

tissue penetration occur concurrently. The double decay reflects dominance by the two events at different times. Pharmacokinetically, neither phase is pure, but functionally it is convenient, even if incorrect, to think of two separate partially overlapping phases. The changeover point is obviously not a precisely defineable time. Oral dosing adds a growth phase before decay occurs in both cases. In addition, there may be a preliminary lag phase. In the single compartment case, there is then one phase of growth and one of decay. In the two-compartment case, there is one

PHARMACOKINETICS OF RELATED BENZODIAZEPINES a

1 7S

k a

Body as a single

ke

compartment

c

0

b

4-

Ila

CU c C C.)

C 0

03) 0

ka

llb Time Figure 3 Model plasma concentration graphs (semilogarithmic plots) for one- and two-compartment systems (I and 11 respectively) and intravenous and oral dosing (a and b respectively). lb

Figure 2 The one- and two- compartment pharmacokinetic models.

phase of growth and then two phases of decay. The decay phases correspond with those seen after intravenous dosing and the growth phase is usually also exponential. A special situation arises when the growth phase and the first phase of decay have similar rates of change. In this situation the data can fail to show more than two single phases, one each of growth and decay. The model drug concentration curves for the two cases and for intravenous and oral dosing, are shown in Figure 3. Note the semilogarithmic style of plotting. Standard pharmacokinetic terminology relevant to this paper includes: a, decay constant for the first decay phase (a-phase) in the two-compartment case; (, decay constant for the second decay phase (13phase) in the two-compartment case; kei, decay constant for the one-compartment case; ka, growth constant for absorption into a single compartment system, or into the central compartment of a twocompartment system; half-life (TI), time for halfchange in any exponential process; volume of distribution (Vd), theoretical volume which would be needed to accommodate the entire body content of the drug at the concentration measured in plasma, usually once equilibrium within the compartments has been reached. Pharmacokinetic, physicochemical and pharmacodynamic control of drug response The possibilities of pharmacokinetic control of drug illustrated by an examination of Figure

response are

4. This is an annotated version of Figure 3 Ilb, with semi-log coordinates, emphasizing the various phases in the curve following oral input into a two-compartment system. The effect persists for the time the concentration is above the threshold for effect. Thus, there is at first a delay before onset of effect. The effect may wear off during the a-phase and duration of effect will then reflect a relatively high rate constant (a). However, if the threshold is below the concentration at the point of the a/l inflexion, then the effect will wear off during the P-phase and

duration of effect will reflect a relatively low rate constant ((). The relative importance of the two phases may thus affect the duration of effect and this relative importance is a function of the compartment sizes, and of Vd. It is important to realize that compounds vary in their relative values of ka, a and (3, and in Tma,, the concentration at Tma,x and in the time and mean concentration of the a/p inflexion. Additionally, individual patients vary in exactly the same way, and no consistent pattern is seen. Thus, for example, a patient with a late Tma for one drug will not necessarily show a late Tma for another closely related drug. Additionally, all this presupposes that the site of action is in the central compartment. If the site of action is in the peripheral compartment, the effect will actually increase during the a-phase, and this emphasizes the need to determine the site of action within the pharmacokinetic system.

18S


T.,

Onset of effect

t__ B (function of Vd)-.

L

I

- --

- Threshold for effect (a)

X-Cangeover area a/l?

/Threshold for effect (b)

Lag (tablet dissolution)

Time Figure 4 Annotated model plasma concentration graph (semi-log plot) for a two-compartment system and oral dosing.

The rate of absorption, assessed by ka, will affect onset of effect for obvious reasons. Many factors will

additionally affect pharmacokinetic control of effect. For example, multiple dosing may increase the minimum concentration above that prevailing at the point of inflexion, so that the effect would then wear off as a function of P. A whole variety of factors affect the pharmacokinetics of a compound, such as saturation of enzymes, interfering drugs, disease states, and so on. Binding to plasma protein will promote absorption, and will affect availability for action at any particular time, but it should be noted that current beliefs dismiss any role for protein binding in controlling rates of metabolism and excretion except in very exceptional circumstances. Finally, kinetics of any active metabolites must be taken into account, and from a pharmacodynamic point of view, receptor binding, receptor sensitivity, and intrinsic activity of the drug will obviously affect response. Pharmacokinetic differences among benzodiazepines There has been no systematic examination of all the various benzodiazepines within any one group of

human subjects, although various reports on individual compounds have appeared and one paper concerned three compounds. A systematic comparison within one primate species yielded the data in Table 1. The key observations were: (1) nordiazepam is unique in the group in showing delayed absorption, in having an apparent Vd approximately equal to body water, and in failing to provide evidence of multicompartment behaviour; (2) all of the other compounds showed biphasic decay, and in the case of oxazepam there was a hint of three phases; (3) only nordiazepam occurred to any extent as a metabolite of another compound, clorazepate was quantitatively converted to nordiazepam, and diazepam was also very rapidly and substantially metabolized (approximately 75%) to this compound within 1 h of intraperitoneal dosing; (4) measurement of effects supported the idea that the site of action was in the central compartment. The values of a were around 1.0/h-', with diazepam marginally the highest and clorazepate the lowest. Values for P ranged from 0.025 to 0.24 h-', with temazepam the highest and clorazepate the lowest. The value of keI for nordiazepam was in the region 0.08 to 0.09 h-1. These data provide a model for human

PHARMACOKINETICS OF RELATED BENZODIAZEPINES

investigations, which, pooling observations from a number of studies, provide the data in Table 2. It seems that the important differences in man are analogous to, but not identical with monkeys. In humans, diazepam has quite a long half-life (>30 h) and forms a persistent active metabolite with a very long half-life (>50 h). Diazepam has one or more brief a-phases, and so gives a sharp peak of short duration after oral doses. Diazepam is the most rapidly absorbed of the compounds studied adequately. Nordiazepam administered as such shows single compartment characteristics, and a long Ti. Temazepam shows biphasic decay, with both phases having relatively short T, values. Oxazepam has two phases and is intermediate between Table 1

19S

temazepam and diazepam. Of the compounds not included in Table 2 medazepam data are inadequate for comparison. Chlordiazepoxide disappears fairly rapidly but forms the persistent nordiazepam. Chlordiazepoxide is complicated by a low Vd (0.25 1/kg) and it is possibly inactive until converted to N-desmethyldiazepam. Clorazepate is believed to be inactive as such, but it delivers nordiazepam efficiently. It is of interest to propose ways in which these pharmacokinetic properties might influence effects. Assuming similar pharmacology among these compounds, oral diazepam would clearly be suitable for inducing an effect rapidly, but it would be unsuitable if a brief effect was needed (for example,

Mean plasma concentrations (ig/ml) in the monkey (Macaca mulatta) after intraperitoneal injection.

0.5

1

Time (h) 2

4

8

24

Mean s.e. Mean s.e. Mean s.e. Mean s.e. Mean s.e.

1.81 0.70 0.55 0.13 2.61 0.67 0.71 0.23 1.38 0.42

1.10 0.31 0.33 0.07 2.72 0.38 0.46 0.13 0.87 0.19

0.80 0.20 0.11 0.02 2.11 0.29 0.17 0.05 0.39 0.13

0.30 0.11 0.04 0.01 1.84 0.24 0.04 0.01 0.22 0.06

0.05 0.02 0.03 0.01 1.40 0.21 0.02

0.02 0.01 0.00

0.05 0.02

0.01 0.01

Mean

0.99 0.19

1.16 0.37 2.05 0.38

1.08 0.27 1.57 0.20

0.57 0.09 1.27 0.21

0.51 0.11 1.06 0.13

0.09 0.01 0.25 0.05

Compound Unmetabolized drugs

Clorazepate Diazepam Nordiazepam Temazepam

Oxazepam Nordiazepam as a metabolite of other drugs After clorazepate After diazepam

s.e. Mean s.e.

1.77

0.39

0.35 0.07 0.00

Table 2 Mean pharmacokinetic data in man for various benzodiazepines (human volunteer subjects except where indicated). Inflexion No. of subjects

ka (h- 1)

Tmax (min)

Txa (h)

(a/lp)

Diazepam N-desmethyidiazepam * N-desmethyldiazepam t Temazepam

7 7 9 4

6.1 3.3 NE NE

45 80 NE 94

2-4 2-4 NE

Oxazepam

7

1.6

114

0.69 0.58 NE In range 2-3 1.63

(h)

8-10 8-24

Tip (h) 48 62 51 In range 15-20 11.0

*Given as clorazepate. tGiven as such (measurements at end of chronic medication in patients). NE, Not evaluated. Note: considerable quantities of N-desmethyldiazepam are found after administration of diazepam.

20S


sleep induction) unless it was given in doses where the effect was confined to the a-phase. Even then, buildup of nordiazepam might be a problem. Diazepam and nordiazepam, and so clorazepate, are clearly suitable where a prolonged effect is required (for example, in anxiety or in prophylaxis of convulsions), and both can probably be given once daily (this has been shown to be appropriate for clorazepate). The slow absorption of nordiazepam as such might be a problem, but this has only been observed in monkeys to date. In anxiety, combining a brief sedative effect with a prolonged anxiolytic effect in a bedtime dose may be particularly appropriate, and diazepam is clearly the best drug for doing this. If the brief sedative effect is not sought nordiazepam or clorazepate are the most suitable, as these drugs lack the sedative effect when given in anxiolytic doses. If a brief sedative effect is desired for sleep induction, with no prolonged effect, then temazepam would seem to be appropriate, but by the same token this drug would not be good for anxiety, unless anxiety was relieved purely as the result of better sleep. As an anticonvulsant, temazepam would probably require unsuitably frequent dosing. Oxazepam is probably not suitable for a short term effect, as its absorption is slow and its a-phase prolonged. In addition, this drug is of low potency. In this context, medazepam is converted to diazepam and nordiazepam and has no obvious advantages, and considerable opportunity for problems, as does chlordiazepoxide. However, medazepam, in contrast to chlordiazepoxide, is of high potency, and with its brief persistence and its own effect at a low dose, metabolite effects might well be minimal because of their own formation in quantities smaller than those arising after, say, diazepam. This might then be the best drug for a brief sedative effect on its own. Very few of these ideas have been systematically tested.

Physicochemical differences among benzodiazepines As with other properties of the benzodiazepines, isolated reports concerning physicochemical

properties of the individual compounds have appeared, but systematic comparisons of the compounds in similar conditions had been undertaken to only a limited extent. In one study, lipid solubility (octanol/water true partition coefficient) was in the rank order: medazepam >Ndesmethyldiazepam > diazepam > oxazepam > temazepam. Protein binding to 1 % bovine serum albumin was in this order except for interchange of oxazepam and temazepam, and protein binding correlated with true partition coefficient r = 0.93. Additionally, protein binding correlated with pKa r = 0.'96. This probably indicated participation of the ionized species in the binding reaction, but the exact relationship between lipid solubility, pKa and binding is obscure. The data are complicated by the fact that heptane/water partition coefficients place the lipid solubilities in the order: medazepam > diazepam > temazepam > N-desmethyldiazepam > oxazepam; this is approximately the rank order of sedative potency and probably of rate of brain penetration. Against this, the correlations between heptane/water partition coefficients and protein binding and PKa are weaker, and the approximate rank order of Vd, judging from the data in monkeys is diazepam > temazepam > oxazepam > nordiazepam (medazepam not assessed). Thus, although it seems likely that a relation exists between the physicochemical measures, binding and pharmacokinetic features of the compounds, the exact detail of the relationship has eluded detection to date and information in this category does not help in explaining pharmacological differences. Pharmacodynamic variations among benzodiazepines As already mentioned, the compounds, as sedatives, vary in potency: medazepam > diazepam > nordiazepam > temazepam > oxazepam. This information is, however, useless if the variations in potency can be overcome by giving equipotent doses, which is possible if the maximum obtainable effect is the same. In such a case, only the duration of effect is different. However, it seems that the balance of

Table 3 Physicochemical data for various benzodiazepines

Compound

Log (true partition coefficients) Octanol/water Heptane/water

Protein binding (% bound to 1% BSA)

pKa

Medazepam

3.88

3.39

95

6.2

N-desmethyldiazepam Diazepam Oxazepam Temazepam

2.83 2.79 2.29 2.19

0.06 1.68 -1.30 0.30

86 84 64 76

3.5 3.3 1.7 1.6

PHARMACOKINETICS OF RELATED BENZODIAZEPINES

various effects - sedative, anxiolytic and anticonvulsant - varies among the compounds. For example, as already mentioned, nordiazepam exerts anxiolytic effects at non-sedative doses. Additional structure-activity data include the fact that nitro-compounds (clonazepam and nitrazepam) are the most potent anticonvulsants, nitrazepam being a particularly unfortunate choice as a sleep inducer because of long-term effects. These variations are insufficiently exploited at present.

Conclusions By judicious choice, it would seem to be possible to select the most suitable benzodiazepine for day or night sedation, anxiolytic effect or prevention of convulsions by considering the relative pharmacokinetic, physicochemical and pharmacodynamic properties of the available compounds. However, it must always be remembered, that even though data may offer guidelines, the best choice may well be dependent on individual pharmacokinetic properties of each drug in each patient and that the important data may not be at hand in a particular case. Nevertheless, to claim that all benzodiazepines are the same shows an unfortunate lack of insight and careful examination of the different examples for different purposes should be undertaken.

21S

References WRETLIND, M., PILBRANDT, A., SUNDWALL, A. &

VESSMAN, J. (1977). Disposition of three benzodiazepines after single oral administration in man. Acta Pharmac. Toxic., suppl. 1, 40, 28-39. CURRY, S.H., WHELPTON, R., NICHOLSON, A.N. & WRIGHT C.M. (1977). Behavioural and pharmacokinetic studies in the monkey (Macaca mulatta) with diazepam,

nordiazepam and related, 1,4-benzodiazepines. Br. J. Pharmac., 61, 325-330. ROBIN, A., CURRY, S.H. & WHELPTON, R. (1974). Clinical and biochemical comparison of clorazepate and diazepam. Psychol. Med., 4, 388-392. FUCCELLA, L.M., TOSOLINI, G., MORO, E. & TAMASSIA, V. (1972). Study of physiological availability of temazepam in man. Int. J. clin. Pharmac. Ther. Toxic., 6, 303-309. TOGNINI, G., DE MAIO, D., ALBERTI, G.G., FRANCISCI, P.,

GOMENI, R. & SCIEGHI, G. (1975). Pharmacokinetics of N-desmethyldiazepam in patients suffering from insomnia. Br. J. clin. Pharmac., 2, 227-233. KAPLAN, S.A., JACK, M.L., ALEXANDER, K. & WEINFELD,

R.E. (1973). Pharmacokinetic profile of diazepam in man following single intravenous and oral and chronic administrations. J. Pharm. Sci., 62, 1789-1796. GARATTINI, S., MUSSINI, E. & RANDALL, L.O. (1973). (Eds.) The Benzodiazepines. New York: Raven Press. WHELPTON, R. (1978). Lipophilicity as a factor in the biochemical pharmacology of tranquillizing drugs. PhD thesis, University of London.