DRUG DISPOSITION
Clin. Pharmacokinet. 21 (3): 165-177, 1991 0312-5963/91 /0009-0165/$06.50/0 © Adis International Limited. All rights reserved. CPK1050
Clinical Pharmacokinetics of Anxiolytics and Hypnotics in the Elderly Therapeutic Considerations (Part I) I David J, Greenblatt, Jerold S. Harmatz and Richard / , Shader Division of Clinical Pharmacology, Departments of Psychiatry, Pharmacology, and Medicine, Tufts University School of Medicine and New England Medical Center Hospital, Boston, Massachusetts, USA
Contents /65
166 167 168 168 169 177
Summary
Summary I. Anxiolytic-Hypnotic Use in the Elderly 2. Scientific Bases 3. Pharmacokinetic Changes in Old Age 3. 1 Methodological Principles 3.2 Pharmacokinetics of Individual Drugs 3,3 Predicting Age-Related Changes in Pharmacokinetics
Anxiolytic and hypnotic drugs are extensively prescribed for elderly individuals throughout Western society. Old age may be associated with an altered clinical response to this class of compounds, and there is a considerable ethical and economic stake in understanding these changes so that therapy may be approached with a maximum likelihood of therapeutic benefit and a minimum risk of side effects, Old age may lead to altered pharmacokinetics of sedative-anxiolytic drugs, causing higher plasma concentrations (relative to young individuals) after single or multiple doses. By far the majority of the available scientific data refer to the benzodiazepines, which have become the most widely prescribed class of sedative-anxiolytic drugs. Although there is not complete consistency in the available data, the weight of evidence indicates that old age is associated with impaired clearance of the benzodiazepines which are biotransformed by microsomal oxidation (such as diazepam, desmethyldiazepam, desalkylflurazepam, bromazepam, alprazolam, triazolam and others). For those benzodiazepines metabolised mainly by glucuronide conjugation (oxazepam, lorazepam, temazepam) or nitroreduction (nitrazepam), there are minimal, if any, age-related decrements in clearance. Only in the case of triazolam is there direct evidence linking impaired clearance to enhanced clinical effects in the elderly, The logical suggestion that benzodiazepines biotransformed by conjugation or by nitroreduction may be safer for the elderly than those biotransformed by oxidation has not yet been directly validated. Reasonable epidemiological evidence has linked the use of long (versus short) half-life benzodiazepines (regard-
I A complete reference list will appear in Part II of this article in the next issue of the Journal.
Clin. Pharmacokinet. 21 (3) 1991
166
less of the specific metabolic pathway) with an increased incidence of adverse reactions such as confusion, falls and hip fractures in elderly persons. However, the decreased clearance and increased accumulation of the benzodiazepines in question are not clearly validated as the cause of the increased frequency of adverse reactions. Old age may also be associated with an increased intrinsic sensitivity to benzodiazepines; that is, enhanced pharmacodynamic response, relative to young individuals, at any given plasma or target organ concentration. This change in sensitivity may coexist with, or be independent of, alterations in pharmacokinetics. Altered benzodiazepine sensitivity has been documented both in the course of clinical use of benzodiazepines prior to endoscopy or cardioversion, and in placebo-controlled laboratory trials. Animal models of aging have validated an enhanced response to benzodiazepines as a consequence of impaired clearance, increased intrinsic sensitivity or both. However, many studies directly assessing benzodiazepine receptor affinity, density and function in aging animals have failed to identify significant age-related changes. Despite substantial clinical and scientific data, few specific recommendations can be made regarding needed alterations in the approach to benzodiazepine therapy in elderly persons. Similarly, methodology is not yet available to identify precisely which elderly persons are at risk of developing adverse reactions to benzodiazepines. Clinical guidelines must still be based on the cautious approach of reduced dosage and careful monitoring of treatment.
1. Anxiolytic-Hypnotic Use in the Elderly Society has a large economic and ethical stake in the appropriate use of anxiolytics and hypnotic drugs in elderly individuals. Studies of the prevalence of anxiety and sleep disorders in the general population consistently show that elderly individuals are equally or disproportionately represented (Blazer et al. 1990; Mellinger et al. 1985; Murphy et al. 1988; Regier et al. 1988; Uhlenhuth et al. 1983). Surveys of the extent of anxiolytic and hypnotic drug use by ambulatory patients also demonstrate that the elderly in general are disproportionately represented among users of such drugs in a number of Western nations (Dunbar et al. 1989; Greenblatt et al. 1975; Koenig et al. 1987; Morgan 1983, 1990; Morgan et al. 1988). Among individuals hospitalised in acute or chronic care facilities, the elderly are Qf course disproportionately present, and the use of anxiolytic and hypnotic drugs in these settings is common (Miller & Greenblatt 1976; Morgan 1983). It has recently become fashionable to equate the prescribing to, and use of psychotropic drugs by, elderly individuals with misprescribing and overuse (Shorr et al. 1990). Most reports in both the lay and scientific literature focus mainly on the potential hazards of anxiolytic and hypnotic drug use
in the elderly, with little consideration of either the potential benefits of these drugs in terms of relieving suffering and improving the quality of life, or of the balance between risk and benefit. It is possible that anxiolytics and hypnotics are underprescribed, rather than overprescribed, for the elderly (Uhlenhuth et al. 1983). Nonetheless, potential hazards from this class of drugs have been demonstrated in a number of epidemiological studies. In most cases, primary care physicians, rather than psychiatrists, do the prescribing (Beardsley et al. 1988). Among elderly individuals, there is a statistically significant association of sedative-hypnotic drug use with adverse reactions such as cognitive impairment, falls and hip fractures (Larson et al. 1987; Ray et al. 1987; Robbins et al. 1989; Sorock & Shimkin 1988; Tinetti et al. 1988). Within the class of benzodiazepines, the risk of hip fracture appears to be greater in association with long halflife, accumulating drugs than with short half-life agents (Ray et al. 1989). Studies by the Boston Collaborative Drug Surveillance Program (BCDSP), evaluating the frequency of unwanted drowsiness associated with benzodiazepine use by hospitalised medical patients, indicated an increased frequency of such adverse reactions among elderly individuals. This association was found for the anxiolytic drugs chlordiazepoxide and diazepam (BCDSP
Anxiolytics and Hypnotics in the Elderly
(973), and for the hypnotics flurazepam and nitrazepam (Greenblatt & Divoll Allen 1978; Greenblatt et al. 1977). Insofar as could be determined, the association of adverse reactions with age was not explained by factors such as gender, dosage or severity of underlying disease. The limitations of such studies must be recognised. They demonstrate association and not causality. Furthermore, selection bias cannot be ruled out: individuals whose underlying condition or degree of debility inherently increased the risk of drug therapy were precisely those patients who received medications. Taken collectively, the data indicate that the issue of appropriate sedative-anxiolytic prescribing for elderly individuals is one of large scale (Thompson et al. 1983). There is a pressing need for continuing development of scientific data on which to base an understanding of drug response and proper prescribing for elderly patients (Monamat et al. 1989).
2. Scientific Bases The fundamental clinical question can be summarised as follows (Shader et al. 1987): in utilising pharmacotherapy for the treatment of agitation, anxiety or sleep disorders, what adjustments should be made in the approach to therapy in the elderly patient such that the likelihood of therapeutic efficacy is maximised, and the likelihood of adverse reactions is minimised? Possible alterations in the approach to therapy using a specific medication might include a change in the size of each dose, the interval between doses, the duration oftherapy or the approach to initiating and terminating treatment. An altogether different choice of drug may be appropriate. Diagnostic criteria and methods of assessing efficacy and toxicity might also require alteration during treatment of aging populations (Gurian & Miner 1990; Shamoian 1990). More complex considerations include the assessment of intercurrent or coincident medical or psychiatric illness, concurrent drug therapy and the patient's diet and living situation (Cohen 1990; Salzman (990). For practical purposes, 'drug therapy' implies 'benzodiazepine therapy', since benzodiaze-
167
pines are by far the predominant class of anxiolytic and hypnotic drugs in use today. The vast majority of available scientific data refers to the benzodiazepines (Greenblatt et al. 1983b), although recently some data have become available for the nonbenzodiazepine anxiolytic buspirone. The clinical phenomenon most commonly confronting health care professionals involved in drug therapy of the elderly is that a given dose ofa given drug produces a different (or perhaps unexpected) response in an elderly person compared with a younger individual of the same gender and similar bodyweight. The most obvious explanations are pharmacokinetic and/or pharmacodynamic in nature (Greenblatt & Shader 1990). In pharmacokinetic terms reduced drug clearance (CL) is often associated with old age. By definition, reduced CL implies that, after any specific dose of a given drug, an elderly person will have a greater area under the plasma concentration-time curve (AUC) than a corresponding young 'control' subject. This in tum implies that at some (or perhaps all) times after a single dose, plasma drug concentrations are higher in the elderly. If plasma concentration is proportional to target organ concentration, which in tum is proportional to clinical response, then the elderly person may have a greater drug response than the young control person, at least at some time after a single dose. During long term dosage, steady-state plasma concentration (CSS) is inversely proportional to CL. Therefore, the elderly person receiving therapy at a specific dosing rate will have a higher css than the young control; with this comes the possibility of greater drug response in the elderly. In pharmacodynamic terms, the aging organism may have a greater drug 'sensitivity' than a young control individual. That is, any given plasma concentration, or target organ concentration, produces a greater clinical response in an elderly than in a young organism. In principle, increased pharmacodynamic sensitivity in the elderly may coexist with, or be independent of, pharmacokinetic changes. Requirements for scientific excellence differ substantially between the pharmacokinetic and
Clin. Pharmacokinet. 21 (3) 1991
168
Dose
Fig. 1. Schematic representation of the cascade of events between drug administration and pharmacodynamic effect.
pharmacodynamic dimensions. Pharmacokinetic studies have relatively tangible and objective methodological requirements. Studies of intrinsic drug sensitivity, on the other hand, must necessarily include the complex cascade of events between drug administration and response (fig. I). Not surprisingly, the methodological requirements are more stringent and difficult. Furthermore, the elucidation of mechanisms almost always requires experimental data in addition to clinical data, since many steps on the cascade are not ordinarily accessible to study in humans.
3. Pharmacokinetic Changes in Old Age 3.1 Methodological Principles Scientific data on age-related changes in pharmacokinetics and drug metabolism have been subject to continuing, extensive review (Dumas et at. 1990; Greenblatt et at. 1982a, 1983a, 1986a, 1989a; Loi & Vestal 1988; Ouslander 1981 ; Schmucker 1985a,b; Vestal 1982, 1989). Since hepatic biotransformation is the principal route of clearance for essentially every clinically relevant anxiolytic and hypnotic drug, age-related alterations in hyp-
notic drug-metabolising capacity are of most relevance. The pharmacokinetic consequences of a reduction in hepatic drug clearance (CLH) after oral dosage will depend on whether the compound is a high- or low-clearance drug (Wilkinson 1987). The majority of benzodiazepines can be characterised as low-clearance drugs, with CLH considerably less than hepatic blood flow (Q H) [Greenblatt & Shader 1987}. Under these circumstances, a reduction in CL will be evident mainly as a prolongation of elimination half-life (t 'l'~) rather than a change in peak plasma concentration (C max ) of the parent compound. A few benzodiazepines (such as triazolam and midazolam), as well as the nonbenzodiazepine buspirone, have values ofCLH which are somewhat closer to QH. For such drugs, reduced CL may also be evident as lower presystemic extraction, leading to higher plasma concentrations at all times after administration (Pond & Tozer 1984; Wilkinson 1987). Prolongation of tl;'~ may also be evident, although this is generally less dramatic than with low-clearance drugs. In clinical practice, old age per se cannot usually be separated from consequences of aging such as intercurrent disease, use of other medications, alterations in diet and body habitus, and changes in activity. These other factors may potentially be as important as or more important than age itself as a determinant of drug-metabolising capacity. The scientific study of aging itself, however, requires control or exclusion of these other factors insofar as this is possible. Therefore, the most meaningful and convincing studies of pharmacokinetics and aging involve healthy elderly populations, with the proviso that further study is necessary to elucidate the pharmacokinetic consequences of age combined with other factors. It is clear that pharmacogeriatric studies can be carried out with proper implementation of the principles of ethical research and informed consent (Abernethy & Azarnoff 1990). Increasing evidence indicates that gender as well as age may influence drug distribution and clearance. Therefore, the problem of pharmacokinetics in aging cannot be separated from the issue of pharmacokinetics and gender; i.e. a well designed study should evaluate pharmacokinetics
Anxiolytics and Hypnotics in the Elderly
in elderly persons, and in young controls, of both genders. Statistical issues also assume great importance. Greater sample sizes in each cell increase the likelihood that the study population is representative of the general population, and assure sufficient statistical power for an accurate assessment. In actual research practice, the desirability of a large sample size in each cell is always tempered by feasibility considerations of time, funds, energy and the availability of suitable study participants. 3.2 Pharmacokinetics of Individual Drugs Findings across studies and laboratories are not always completely consistent, but some general conclusions can be offered. For benzodiazepines metabolised by hepatic microsomal oxidation, old age is generally associated with reduced CL. In many studies, the age-related reduction is more evident in men than in women. For those drugs metabolised mainly by glucuronide conjugation or by nitroreduction, age-related alterations in CL are far Jess evident. Relative increases in the fraction oriotal bodyweight comprised of adipose tissue, as occur with both old age and female gender, are of considerable importance. Differences in pharmacokinetic volume of distribution (Vd) in the elderly compared with the young, and in women compared with men, may be influenced by the lipid solubility of the drug in question (Greenblatt et al. 1982a). Changes in Vd will by themselves cause reciprocal changes in t1hi3' even in the absence of a change in CL (Abernethy & Greenblatt 1982, 1986). Because old age may also alter the plasma protein binding of benzodiazepines, evaluation of agerelated changes in free fraction (fu) is generally another necessary component of such studies. 3.2.1 Chlordiazepoxide Chlordiazepoxide, available since 1960, was the first of the benzodiazepines to be used in clinical practice (Greenblatt & Shader 1974). This drug is metabolised by hepatic microsomal oxidation. Its initial metabolite, desmethylchlordiazepoxide, is pharmacologically active, as are subsequent prod-
169
ucts in the metabolic cascade (demoxepam, desmethyldiazepam) [Greenblatt et al. 1978]. Shader et al. (1977) administered single oral doses of chlordiazepoxide hydrochloride 25mg to healthy young and elderly men. Compared with young controls, chlordiazepoxide tlhi3 was significantly prolonged in the elderly (18.2 vs 10.1h), and CL was significantly reduced (0.35 vs 0.61 ml/min/ kg). Increased Vd was also observed in elderly individuals. Roberts et al. (1978) administered single intravenous doses of chlordiazepoxide hydrochloride 0.6 mg/kg to 27 healthy subjects aged 16 to 86 years. There was a significant positive correlation of t'l2i3 and Vd with age; CL likewise declined significantly. Chlordiazepoxide fu in plasma averaged 3.5%, and was not significantly related to age. Greenblatt et al. (l989b) administered doses of chlordiazepoxide hydrochloride 50mg both orally and intravenously to groups of young male, young female, elderly male and elderly female volunteers. As in the previous 2 studies, the tlhi3 was significantly prolonged in elderly vs young men (20.4 vs 7.9h), and CL significantly reduced (0.26 vs 0.59 ml/min/kg). Among women, however, age-related changes did not approach statistical significance. Chlordiazepoxide fu in plasma was significantly higher in elderly as opposed to young women (6.6 vs 5.9%); the values in elderly and young men were nearly identical (6.6 vs 6.8%). The absolute bioavailability of oral chlordiazepoxide was not significantly different from 100%, and was not influenced by either age or gender. 3.2.2 Diazepam Approximately 50% of the total clearance of diazepam in humans is accounted for by the oxidative reaction of N-demethylation, yielding desmethyldiazepam as the principal metabolite (Greenblatt et al. 1988a). The other component of CL is explained in part by hydroxylation to form temazepam, as well as by other, unidentified pathways (Klotz et al. 1980; Mandelli et al. 1978). As such, oxidation is the principal mechanism of diazepam elimination. Following single doses of diazepam, the appearance of desmethyldiazepam in systemic circulation is relatively slow, such that it
CUn. Pharmacokinet. 21 (3) 1991
170
probably contributes minimally to the pharmacodynamic action of a single dose of diazepam. During iong term therallY, however, both diazepam and desmethyldiazepam accumulate in plasma, with the CSs of the metabolite similar to or exceeding that of the parent drug (Greenblatt et al. 198Ia). Because the CL of temazepam is higher than that of either diazepam or desmethyldiazepam, temazepam CSS during long term therapy with diazepam is relatively low, as are plasma concentrations of oxazepam formed from demethyldiazepam. The first pharmacogeriatric study of diazepam in humans was that by Klotz and associates (1975), in which 20 healthy volunteers (17 men and 3 women) aged 15 to 82 years received a single intravenous dose of the drug. Its Vd and tl/211 both increased significantly with age. A key finding of this study was the relationship of CL with age: among the 20 healthy individuals, the correlation coefficient was -0.39 (p = 0.089). Since this value does not reach statistical significance at the p = 0.05 level, it was concluded that the observed decline in CL with age could have been due to chance, and therefore there was no statistical evidence for a decline in metabolising capacity with age. In the light of subsequent studies (see below), it is more reasonable to recognise that the failure to reach statistical significance was simply due to the small sample size. It should be noted that diazepam fu in this study averaged 2.6%, and was not significantly related to age. Kanto et al. (1979) next reported the effect of age on the pharmacokinetics of single intravenous doses of diazepam in 14 patients aged 32 to 78 years, receiving the drug in the context of premedication prior to general anaesthesia. Both men and women were represented in the sample. They were not healthy volunteers; many had intercurrent medical disease and were receiving other medications concurrently. In any case, diazepam CL decreased significantly with age. As in the previous study by Klotz et al. (1975), both tl/211 and Vd increased significantly with age; diazepam fu was not measured in this study. Macklon et al. (1980) administered intravenous diazepam to 19 subjects of both genders, with ages ranging from 18 to 95
years. Again, its tl/211 and Vd increased significantly with age. The fu also increased significantly with age, emphasising the need for correction of CL values for individual differences in fu. After such correction, the CL of unbound diazepam was found to decline significantly with increasing age. Greenblatt et al. (1980a) studied 4 groups of subjects: young male and female (aged 21 to 37 years), and elderly male and female (aged 61 to 84 years) [fig. 2]. The fu of diazepam was significantly higher in the elderly volunteers, in part because of lower plasma albumin levels. After correction for individual differences of fu, the Vd of unbound diazepam was higher in the elderly, but the effect of gender was more impressive than that of age. That is, women had a larger Vd than men, regardless of age. The CL of unbound diazepam was significantly lower in elderly than in young subjects regardless of gender (fig. 3), and tl/211 likewise was greater in the elderly. Ochs et al. (l98Ia) admini-
400
• = Diazepam
200
o = Desmethyldiazepam
100 50 20
::r c; .5
10 5
c:
i
Male, 24 years
2
8c: 300
8
as E as 100 ii:
'"
50
20
..,.--
.o---~--
10 6
3
p"" {"
__
-o ______
~ ~
Male, 69 years
o 1 234 5 6 789 Time (days)
Fig. 2. Plasma concentrations of diazepam and its major metabolite, desmethyldiazepam, after intravenous doses of the parent drug in representative young and elderly volunteers (from Greenblatt et al. 1980a, with permission).
171
Anxiolytics and Hypnotics in the Elderly
stered single intravenous doses of diazepam to 27 male volunteers aged 20 to 91 years. Again, tl/,{j and Vd increased significantly with age; clearance of total and unbound diazepam declined, but the correlation coefficient did not reach significance. Divoll et al. (1983) replicated the study design of Greenblatt et al. (1980a), with the addition of intramuscular and oral doses of diazepam in all subjects using a crossover design. After intravenous administration, the effects of age and gender on diazepam pharmacokinetics were very similar to those reported by Greenblatt et al. (1980a). In addition, the absolute bioavailability of oral diazepam was not significantly different from 100% complete, and was not influenced by either age or gender. The absolute bioavailability of a single intramuscular deltoid injection of diazepam averaged 101 and 97% in young and elderly men, respectively; in young and elderly women, the corresponding numbers were 77 and 86%, which may be explained by the possibility of inadvertent intraadipose as opposed to intramuscular injection in women. Only the study of MacLeod et al. (1979) failed to find any significant age-related changes in the pharmacokinetic profile of diazepam. This may represent the consequences of a small sample size (n = 10, 5 male and 5 female), together with the presence of significant intercurrent illness in several of the subjects. The weight of evidence strongly suggests that the CL of unbound diazepam declines with increasing age among healthy individuals. Vd also increases with age, probably due to the increasing fraction of total bodyweight accounted for by adipose tissue among the elderly. Similarly, diazepam Vd appears to be greater in women than in men, again presumably due to the difference in body habitus. The combination of reduced CL and increased Vd leads to a marked prolongation of diazepam tl/,{j in the elderly. In some studies, the plasma fu of this highly protein-bound drug also increases with age. Therefore, accounting for individual differences in fu is an important aspect of these pharmacokinetic analyses.
-Young
80.0
Elderly I ~
0 0
Cii 70.0 ~
c
E 60.0
;:,.
.§. ~ c 50.0
~
'"
U
U c
~ c
20.0
:::l
.8c :::l
.--.-.......
...........--30.0
Cc
•
40.0
'iii
"0
•
•••• ._........... •••
c 0
--ll---
0
c --'0"'-'---
e CcD 0
0
•
•• -"'oi!~-"
.... -.-.--....
___9. ____ cC c 0 0
••••
10.0 Male
Female
•
Male
Female
0.0 + - - - - + - - - + - - - - - 1 - - - - 1
Fig. 3. Individual and mean (± SE) values of clearance of unbound diazepam in groups of young male, young female, elderly male and elderly female volunteers; ., D = nonsmokers; e, 0 = cigarette smokers (from Greenblatt et al. 1980a, with permission).
3.2.3 Desmethyldiazepam Desmethyldiazepam is an important benzodiazepine. In many parts of the world, it is administered as the pure compound for the treatment of anxiety. Several benzodiazepines (such as clorazepate, prazepam, ketazolam and oxazolam) effectively serve as 'prod rugs' of desmethyldiazepam, being converted to it as the major active substance (Greenblatt et al. 1983a; Ochs et al. I 984a). For still other agents of this class (diazepam, chlordiazepoxide, medazepam, pinazepam and halazepam), desmethyldiazepam is a principal pharmacologically active metabolite. Desmethyldiazepam itself is biotransformed mainly by the oxidative reaction of hydroxylation, yielding oxazepam as the major metabolite. Oxazepam, in turn, is rapidly conjugated and excreted, such that plasma concentrations of intact oxazepam are much lower than those of the parent drug following administration of desmethyldiazepam to humans. Klotz and Miiller-Seydlitz (1979), in a limited study of 4 young and 4 elderly individuals, suggested a prolongation of desmethyldiazepam t'l2,3,
Clin. Pharmacokinet. 21 (3) 1991
172
30
•
~~-
20
:J"
~
10
•
c:
I
0
~c: CD 0
I
4
I
c: 0
E as
"0 N
~.
I
"iii
'"
r'~
= 20.6h
I
~
as E as
tv.
t
0
'-........,.
0.5
1i:
Male, 68 years
Male, 37 years 0.2
I
I
I
I
I
I
0
10
20
30
40
0
I
i
I
I
I
10
20
30
40
50
Time (h)
Fig. 4. Plasma drug concentrations after single oral doses of alprazolam I mg in representative young and elderly male volunteers; till = half-life (from Greenblatt et al. 1983c, with permission).
increase in Vd and reduction of CL in the elderly as opposed to the young. Divoll Allen et al. (1980) studied the pharmacokinetics of desmethyldiazepam after single oral doses of prazepam in 29 subjects aged 22 to 85 years: as with diazepam, desmethyldiazepam Vd was larger in women regardless of age, and increased with age in both men and women. However, the effect of age on the elimination and CL of desmethyldiazepam was highly gender-dependent. The tV2P increased significantly with age in elderly vs young men (128 vs 62h), but was unrelated to age in women. The CL of unbound desmethyldiazepam was unrelated to age in women, but declined with age in men (r = -0.4); the latter effect approached, but did not reach, statistical significance. Shader et al. (1981) repeated this study except that clorazepate dipotassium rather than prazepam was used as the precursor of desmethyldiazepam. The effects of age and gender on desmethlyldiazepam Vd and tV2P were essentially the same as those reported by Divoll Allen et al. (1980). In the Shader et al. (1981) study, the CL of unbound desmethyldiazepam declined significantly with age in men, but was not significantly related to age in women.
3.2.4 Desalkyljlurazepam Although not administered as such in clinical practice, desalkylflurazepam is the final and major metabolite of the benzodiazepine precursor flurazepam, and probably accounts for most of the clinical activity observed after use of the latter in humans (Miller et al. 1988). A new, recently introduced benzodiazepine, quazepam, also has desalkylflurazepam as the final metabolite (Zampaglione et al. 1985). Biotransformation of desalkylflurazepam itself appears to proceed by hydroxylation, with the hydroxylated metabolite rapidly conjugated and excreted. Using flurazepam as the precursor substance for desalkylflurazepam, Greenblatt et al. (1981 b) administered single 15mg oral doses to 26 healthy subjects aged 19 to 85 years. The tV2P was significantly prolonged in elderly vs young men (160 vs 74h), but the prolongation oftv2P in elderly vs young women (120 vs 90h) was not significant. Since desalkylflurazepam was administered as a precursor rather than the intact substance, CL could not be calculated. During multiple dosage with flurazepam 15mg nightly for 15 nights, the CSs of desalkylflurazepam was significantly higher in elderly
173
Anxiolytics and Hypnotics in the Elderly
than in young men (81 vs 53 jLg/L), but was not significantly different in elderly as opposed to young women. Using quazepam as the precursor, Hilbert et al. (1984) found a mean desalkylflurazepam t'hi3 of 190h in 10 healthy elderly volunteers (6 men and 4 women) aged 65 to 77 years. Although young control subjects were not included in this study, the authors suggest that a comparison of values in the elderly subjects with those previously established in young volunteers is consistent with a substantial prolongation of t'hi3 in the elderly. 3.2.5 Alprazolam The triazolobenzodiazepine alprazolam is currently the most widely prescribed benzodiazepine in the US. It is metabolised mainly by the oxidative reaction of hydroxylation, yielding 2 metabolites which are less pharmacologically active than the parent drug (Sethy & Harris 1982). Since the metabolites are rapidly conjugated and excreted, unconjugated concentrations of these hydroxylated products are considerably lower than those of the parent compound (Smith & Kroboth 1987). Greenblatt et al. (1983c) administered single oral doses of alprazolam I mg to young men and women aged 21 to 45, and elderly men and women (62 to 78 years). The Vd was slightly smaller in elderly than in young men, and was unrelated to age in women; t'/'11 was significantly prolonged in elderly compared with young men (19 vs h), and CL significantly reduced (0.64 vs 1.33 ml/min/kg) [figs 4, 5]. Among women, however, the slight prolongation of t'hi3 in the elderly (13.5 vs 1O.8h) did not reach significance, nor did the slight reduction in mean CL (1.07 vs 1.34 ml/min/kg). Plasma fu was unrelated to age. As with desmethyldiazepam, the age-dependent decrement in alprazolam CL appears to be gender-specific. Kroboth et al. (1990) studied the pharmacokinetics and clinical effects of alprazolam during a multiple dosage regimen administered to 16 men and 10 women aged 65 to 80 years. Although there was no concurrent study of young control subjects, the reported mean values of CL and t'hi3 in the group of elderly men (0.56 ml/min/kg; 24.0h) are similar to those reported for elderly male subjects by Greenblatt et al. (1983c),
but the mean values for elderly women (CL 0.63 ml/min/kg; t'hi3 22.4h) are substantially different. The reasons for these discrepancies are not established, but the findings emphasise the potential effect of unknown differences in study population on pharmacokinetic results. 3.2.6 Triazolam The triazolobenzodiazepine triazolam is oxidatively metabolised analogously to alprazolam (Eberts et al. 1981), with hydroxylated metabolites of essentially no clinical importance (Sethy & Harris 1982). It has become the most commonly prescribed benzodiazepine hypnotic in the US. Greenblatt et al. (1983d) administered single oral doses of triazolam 0.5mg to groups of young male, young female, elderly male and elderly female subjects. A highly significant age-related decrement in triazolam CL was observed in elderly as opposed to young volunteers, and the effect was essentially independent of gender. Apparently because triazolam has a somewhat higher value of CLH than the other benzodiazepines described above, the reduced CL in the elderly (that is, increased AUC) was evident mainly as higher plasma concentra-
Young
2.5
Elderly 1-
o
2.0
1.5
•
•
.__ ...... .......
o
--_._._.. _---0
•
B
............ _... 1.0
o
••
• •
0
.. __·0········ 0 0
0
0.5 Male
--_._ .. _.....
0
Female
0
•• • ._------------------...... --..... •• •• Male
0
.. ............... '
0 0 0
Female
O+-------+-------+-------+-----~
Fig. 5. Individual and mean (± SE) values of alprazolam clearance in groups of young male, young female, elderly male and elderly female volunteers (from Greenblatt et al. \ 983c, with permission)
174
tions at all times after dosage. Although values of tl1211 were somewhat longer in the elderly, these differences were not statistically significant. These findings were subsequently replicated by Greenblatt et al. (1991) in an entirely different set of volunteers who received oral triazolam 0.12S and 0.2Smg. After both doses oral CL was significantly lower in the elderly subjects regardless of gender. However, these findings were not replicated by Smith et al. (1983), who administered single oral doses of triazolam O.Smg to 30 men and women aged 20 to 76 years. There were no apparent effects of age on its pharmacokinetic properties. The plasma triazolam concentration determinations for the studies of Greenblatt et al. (1983d, 1991) and Smith et al. (1983) were all made in the same laboratory using the same analytical methodology (Greenblatt et al. 1981 c; Scavone et al. 1986). This suggests that the differences in pharmacokinetic outcome are attributable to variations in the study populations or in experimental design.
3.2.7 Midazolam Midazolam has a CL falling in the range of SO% of QH (Reves et al. 1985). Its 2 hydroxylated metabolites formed by microsomal oxidation are of minimal clinical importance, since their intrinsic activity and relative brain uptake are considerably less than those of the parent compound (Arendt et al. 1987). In a study of 11 young and 11 elderly women, Avram et al. (1983) found no effect of age on the pharmacokinetics of a single intravenous dose of midazolam 0.2 mg/kg. Greenblatt et al. (1984) again studied groups of young and elderly men and women, who received single intravenous (2.S to Smg) and oral (S to lOmg) doses of the drug. Among women the CL of midazolam was lower and tl1211 longer in the elderly; however, none of the differences was statistically significant. Among men, tlj211 was significantly prolonged in the elderly (S.6 vs 2.1h), and CL significantly reduced (4.4 vs 7.8 mlf min/kg); apparently as a consequence of the latter, the absolute systemic availability of oral midazolam was greater in elderly than in young men (SO vs 41 %). Although midazolam was highly protein
Clin. Pharmacokinet. 21 (3) 1991
bound in all subjects (3 to 4% unbound), fu was not influenced by age or gender. Similar findings were reported by Holazo et al. (1988), who administered single intramuscular doses ofmidazolam 7.S mg to male and female volunteers aged 20 to 7S years. Midazolam CL was significantly lower in elderly than in young men, but was not significantly related to age in women. In a population study of intravenous midazolam among 168 surgical patients, no significant relationship was observed between age and tl1211 (Kassai et al. 1988).
3.2.8 Clobazam The I,S-benzodiazepine clobazam is biotransformed by the oxidative reaction of N-demethylation. Greenblatt et al. (1981d) studied the effects of age and gender on the disposition of single oral doses of clobazam. The tl12/l was significantly longer in elderly than in young men (48 vs 17h); among women, it increased with age, but the differences were less striking (49 vs 31 h). Vd increased with age regardless of sex, and within each age group was larger in women than in men. Total CL of clobazam was not significantly related to age among women; in the male volunteers, it declined significantly with age. Clobazam fu also increased with age, but did not influence the pharmacokinetic conclusions. In a subsequent study, Greenblatt et al. (1983d) administered the drug on a multiple dose basis to a subgroup of the same volunteers. Clobazam CSs was highly consistent with the values predicted from CL observed in the single-dose study. Between young and elderly women, CSs was highly consistent with the values predicted from CL observed in the single-dose study. Between young and elderly women, css and steady-state CL of clobazam were not significantly different; among male volunteers, however, the latter was significantly reduced, and CSS significantly increased, in elderly compared with young subjects. Thus, the age- and gender-related changes in clobazam pharmacokinetics observed after single doses lead to predictable consequences when the drug is administered on a multiple-dose basis.
Anxiolytics and Hypnotics in the Elderly
3.2.9 Brotizolam Brotizolam is a thienodiazepine that is metabolised by hepatic microsomal oxidation (Bechtel et al. 1986a,b). 10chemsen et al. (l983a) found a significant prolongation of t'l211 and a reduction in total CL in elderly as opposed to young volunteers. Interpretation of these findings is complicated by the fact that all of the elderly volunteers had significant intercurrent medical illness and in most cases were receiving concurrent medications. 3.2.10 Loprazolam Loprazolam is a benzodiazepine hypnotic biotransformed by hepatic microsomal oxidation. Swift et al. (1985a) found a significantly longer t'l211 among 9 elderly volunteers compared with 10 young controls (19.8 vs 11.2h) after oralloprazolam Img. Oral CL was correspondingly reduced in the elderly. 3.2.11 Bromazepam Ochs et al. (1987) administered single oral doses of bromazepam 6mg, an oxidatively metabolised benzodiazepine (Schwartz et al. 1973), to 32 healthy young and elderly men and women. Its Vd was significantly lower in the elderly, and oral CL likewise was significantly reduced; these age-related changes were not influenced by gender. Although bromazepam fu increased significantly with age, correction of the pharmacokinetic variables for individual differences did not alter the conclusions. 3.2.12 Metaclazepam Metaclazepam is a short half-life benzodiazepine metabolised by microsomal oxidation. Molz et al. (1985) administered single oral doses of metaclazepam 10mg to elderly volunteers and 10 young controls. There were no differences between the groups in the t'hll or CL of the drug or of its principal metabolite, formed by N-demethylation. 3.2.13 Clotiazepam Clotiazepam is biotransformed by microsomal oxidation, leading to a series of hydroxylated and demethylated metabolites (Arendt et al. 1982). Ochs et al. (1984b) administered single oral doses of clotiazepam 5mg to 29 healthy volunteers aged
175
between 22 and 89 years. Among men, clotiazepam CL was negatively correlated with age (r = -0.31), but the association did not reach significance. Among women the CL was, if anything, positively correlated with age. 3.2.14 Oxazepam Unlike the benzodiazepines discussed above, the major metabolic pathway of oxazepam in humans involves glucuronide conjugation as opposed to microsomal oxidation (Greenblatt 1981). Shull et al. (1976) first reported that the pharmacokinetic parameters did not differ significantly between 8 young and 8 elderly volunteers after an oral dose of oxazepam 45mg. Greenblatt et al. (1980b) found no significant age-related changes in oxazepam t'hll or CL among 18 male volunteers aged 22 to 76 years, or among 20 female volunteers aged 28 to 84 years. The fu averaged 4.3% in men and 4.4% in women; these values were not significantly related to age. Murray et al. (1981) found no significant correlation of age with any pharmacokinetic parameters among 13 healthy volunteers who received a single dose of oxazepam 30mg. Ochs et al. ( 1981 b) studied the pharmacokinetics of single oral doses of oxazepam 30mg in 22 male and 9 female volunteers aged 22 to 86 years: neither t'h nor CL was significantly related to age in either sex. 3.2.15 Lorazepam Lorazepam, like oxazepam, is metabolised by glucuronide conjugation, with the principal metabolite (lorazepam glucuronide) recovered in urine (Greenblatt 1981; Greenblatt et al. 1976). Kraus et al. (1978) found no significant effect of age on the pharmacokinetics oflorazepam among 11 male volunteers aged 15 to 73 years. Greenblatt et al. (1979) compared the pharmacokinetics of single intravenous doses of lorazepam 2mg between 15 young (19 to 38 years) and 15 elderly subjects (60 to 84 years). The Vd was smaller in the elderly subjects (0.99 vs 1.11 Ljkg); t'hll did not differ between groups, but there was a statistically significant reduction in lorazepam CL in the elderly (0.77 vs 0.99 ml/minjkg). The findings were
176
apparently not influenced by gender. The overall mean lorazepam fu was 11.1 %, and increased slightly but significantly with age (Divoll & Greenblatt 1982). Absolute bioavailability of intramuscular and oral lorazepam was also evaluated in a subgroup of the elderly subjects, and did not differ significantly from 100% by either route. Aaltonen et al. (1982) administered intravenous lorazepam 0.03 mg/kg to 14 surgical patients aged 25 to 86 years. Age was not significantly associated with t'h/l' Vd, CL or fu in plasma. 3.2.16 Temazepam Temazepam is also biotransformed mainly by direct glucuronide conjugation (Locniskar & Greenblatt 1990; Schwarz 1979). Divoll et al. (1981) found no significant association of age with temazepam t'h/l or CL in a series of 32 healthy male and female volunteers aged 24 to 84 years. The fu increased significantly with age, but this did not influence the conclusions of the study. Smith et al. (1983) studied the pharmacokinetics of single oral doses of temazepam in 30 male and female volunteers aged 22 to 76 years. There appeared to be a trend towards reduced CL with age among women, but the association was not significant. No such association was found in men. Briggs et al. (1980) studied the pharmacokinetics and -dynamics of this drug in young and elderly subjects, but a formal pharmacokinetic comparison between groups was not possible. 3.2.17 Lormetazepam Lormetazepam is a fourth benzodiazepine biotransformed principally by glucuronide conjugation (Hiimpel et a. 1979). Hiimpel et al. (1980) found no significant age-related alterations in its pharmacokinetics between 6 young and 6 elderly volunteers. Three men and 3 women were included in each group, making the sample sizes too small to regard the conclusions as definitive. 3.2.18 Nitrazepam The benzodiazepine hypnotic nitrazepam is biotransformed mainly by nitroreduction (Kangas & Breimer 1981) although the initial metabolite, 7-
Clin. Pharmacokinet. 21 (3) 1991
amino-nitrazepam, has minimal pharmacological activity. The effect of age on its pharmacokinetics was first reported by Castleden et al. (1977), who administered single oral doses of nitrazepam 10mg to 10 elderly subjects aged 69 or older, and 10 young volunteers aged 40 or less. There were no significant differences between groups in t'l2/1 or Vd. Kangas et al. (1979) studied the pharmacokinetics of single oral doses of nitrazepam 5mg in 22 hospitalised, chronically ill geriatric subjects (all of whom were receiving other medications), in comparison with healthy young volunteers. The Vd was significantly greater in the elderly patients, and t'l2/1 was significantly prolonged. However, there was no difference between groups in nitrazepam CL; fu was not measured. Holm et al. (1982) found no significant difference in the pharmacokinetics of nitrazepam 5mg between 8 healthy elderly subjects and 8 young volunteers. 10chemsen et al. (1983b) studied the pharmacokinetics of intravenous nitrazepam 5.25mg in 9 young volunteers (22 to 49 years) and 8 healthy elderly subjects (67 to 76 years). Vd and t'h/l were both increased in elderly subjects, although the difference in t'h/l did not reach significance. However, there were no significant effects of age on the clearance of total or unbound nitrazepam. Greenblatt et al. (1985) evaluated the pharmacokinetics of single oral doses of nitrazepam 10mg in 40 healthy men and women aged 19 to 80 years: Vd was increased in elderly as opposed to young men, but did not differ significantly with age in women. Similarly t'l2/1 was prolonged in elderly vs young men, but did not change with age in women. The differences between young and elderly groups of either gender in clearance of total or unbound nitrazepam did not reach significance. 3.2.19 Buspirone Buspirone is not a benzodiazepine derivative, but its biotransformation proceeds mainly by hepatic microsomal oxidation (Goldberg 1984). After oral administration in humans, bus pirone undergoes extensive presystemic extraction leading to the appearance of a major metabolite, I-pyrimidinylpiperazine (I-PP). This metabolic product appears
Anxiolytics and Hypnotics in the Elderly
177
to be pharmacologically active, and its plasma concentrations considerably exceed those of the parent drug. Gammans et at. (1989) studied the pharmacokinetics of single and mUltiple oral doses of buspirone in 24 men and 24 women aged 20 to 77 years. There were no statistically significantly age or gender effects on the pharmacokinetics of the drug or its principal metabolite after single oral doses of buspirone 15mg, but considerable individual variation was seen, as might be expected for a drug with high presystemic extraction. The findings were similar during the multiple dose arm of the study, in which buspirone 15mg was administered every 8h for 4 days. 3.3 Predicting Age-Related Changes in Pharmacokinetics A number of studies have evaluated individual consistency in the capacity to biotransform various benzodiazepines. Among those benzodiazepines biotransformed by microsomal oxidation (for example, diazepam, desmethyldiazepam, desalkylflurazepam), there are significant within-individual correlations in metabolising capacity for the various drugs (Greenblatt et at. 1982b, 1983a,c), as
measured by t'l2{j or CL. In particular individuals who were slow metabolisers of one of these drugs were likely to be slow metabolisers of them all. The capacity to metabolise phenazone (antipyrine), the oxidatively-metabolised marker compound, was also significantly correlated with the capacity to biotransform low-clearance oxidatively metabolised benzodiazepines (Greenblatt et at. 1982b, 1983c, 1989b). However, the CL and t./, of phenazone were not correlated with the capacity to biotransform benzodiazepines metabolised mainly by conjugation (Greenblatt et at. 1982b). Although CL of the latter drugs is minimally influenced by age as such, there was still consistency in interindividual capacity (Greenblatt et at. I 980b, 1982b), which in turn was significantly correlated with the capacity to clear the conjugatively metabolised marker compound paracetamol (acetaminophen) [Greenblatt et at. 1983c). (Part II of this article, including references, will appear in the next issue of Clinical Pharmacokinetics.)
Correspondence and reprints: Dr D.J. Greenblatt, Division of Clinical Pharmacology, Box 1007, Tufts-New England Medical Center, 171 Harrison Avenue, Boston, MA 02111 , USA.
9th International Symposium on
Preparative and Industrial Chromatography - 'PREP 92' Date: 6-8 April 1992 Venue: Nancy, France For further information, please contact: Congres 'PREP-92' Societe Francaise de Chimie 1, rue Grandville - B.P. 451 F-54001 Nancy Cedex FRANCE
Clin. Pharmacokinet. 21 (3): 165-177, 1991 0312-5963/91 /0009-0165/$06.50/0 © Adis International Limited. All rights reserved. CPK1050
Clinical Pharmacokinetics of Anxiolytics and Hypnotics in the Elderly Therapeutic Considerations (Part I) I David J, Greenblatt, Jerold S. Harmatz and Richard / , Shader Division of Clinical Pharmacology, Departments of Psychiatry, Pharmacology, and Medicine, Tufts University School of Medicine and New England Medical Center Hospital, Boston, Massachusetts, USA
Contents /65
166 167 168 168 169 177
Summary
Summary I. Anxiolytic-Hypnotic Use in the Elderly 2. Scientific Bases 3. Pharmacokinetic Changes in Old Age 3. 1 Methodological Principles 3.2 Pharmacokinetics of Individual Drugs 3,3 Predicting Age-Related Changes in Pharmacokinetics
Anxiolytic and hypnotic drugs are extensively prescribed for elderly individuals throughout Western society. Old age may be associated with an altered clinical response to this class of compounds, and there is a considerable ethical and economic stake in understanding these changes so that therapy may be approached with a maximum likelihood of therapeutic benefit and a minimum risk of side effects, Old age may lead to altered pharmacokinetics of sedative-anxiolytic drugs, causing higher plasma concentrations (relative to young individuals) after single or multiple doses. By far the majority of the available scientific data refer to the benzodiazepines, which have become the most widely prescribed class of sedative-anxiolytic drugs. Although there is not complete consistency in the available data, the weight of evidence indicates that old age is associated with impaired clearance of the benzodiazepines which are biotransformed by microsomal oxidation (such as diazepam, desmethyldiazepam, desalkylflurazepam, bromazepam, alprazolam, triazolam and others). For those benzodiazepines metabolised mainly by glucuronide conjugation (oxazepam, lorazepam, temazepam) or nitroreduction (nitrazepam), there are minimal, if any, age-related decrements in clearance. Only in the case of triazolam is there direct evidence linking impaired clearance to enhanced clinical effects in the elderly, The logical suggestion that benzodiazepines biotransformed by conjugation or by nitroreduction may be safer for the elderly than those biotransformed by oxidation has not yet been directly validated. Reasonable epidemiological evidence has linked the use of long (versus short) half-life benzodiazepines (regard-
I A complete reference list will appear in Part II of this article in the next issue of the Journal.
Clin. Pharmacokinet. 21 (3) 1991
166
less of the specific metabolic pathway) with an increased incidence of adverse reactions such as confusion, falls and hip fractures in elderly persons. However, the decreased clearance and increased accumulation of the benzodiazepines in question are not clearly validated as the cause of the increased frequency of adverse reactions. Old age may also be associated with an increased intrinsic sensitivity to benzodiazepines; that is, enhanced pharmacodynamic response, relative to young individuals, at any given plasma or target organ concentration. This change in sensitivity may coexist with, or be independent of, alterations in pharmacokinetics. Altered benzodiazepine sensitivity has been documented both in the course of clinical use of benzodiazepines prior to endoscopy or cardioversion, and in placebo-controlled laboratory trials. Animal models of aging have validated an enhanced response to benzodiazepines as a consequence of impaired clearance, increased intrinsic sensitivity or both. However, many studies directly assessing benzodiazepine receptor affinity, density and function in aging animals have failed to identify significant age-related changes. Despite substantial clinical and scientific data, few specific recommendations can be made regarding needed alterations in the approach to benzodiazepine therapy in elderly persons. Similarly, methodology is not yet available to identify precisely which elderly persons are at risk of developing adverse reactions to benzodiazepines. Clinical guidelines must still be based on the cautious approach of reduced dosage and careful monitoring of treatment.
1. Anxiolytic-Hypnotic Use in the Elderly Society has a large economic and ethical stake in the appropriate use of anxiolytics and hypnotic drugs in elderly individuals. Studies of the prevalence of anxiety and sleep disorders in the general population consistently show that elderly individuals are equally or disproportionately represented (Blazer et al. 1990; Mellinger et al. 1985; Murphy et al. 1988; Regier et al. 1988; Uhlenhuth et al. 1983). Surveys of the extent of anxiolytic and hypnotic drug use by ambulatory patients also demonstrate that the elderly in general are disproportionately represented among users of such drugs in a number of Western nations (Dunbar et al. 1989; Greenblatt et al. 1975; Koenig et al. 1987; Morgan 1983, 1990; Morgan et al. 1988). Among individuals hospitalised in acute or chronic care facilities, the elderly are Qf course disproportionately present, and the use of anxiolytic and hypnotic drugs in these settings is common (Miller & Greenblatt 1976; Morgan 1983). It has recently become fashionable to equate the prescribing to, and use of psychotropic drugs by, elderly individuals with misprescribing and overuse (Shorr et al. 1990). Most reports in both the lay and scientific literature focus mainly on the potential hazards of anxiolytic and hypnotic drug use
in the elderly, with little consideration of either the potential benefits of these drugs in terms of relieving suffering and improving the quality of life, or of the balance between risk and benefit. It is possible that anxiolytics and hypnotics are underprescribed, rather than overprescribed, for the elderly (Uhlenhuth et al. 1983). Nonetheless, potential hazards from this class of drugs have been demonstrated in a number of epidemiological studies. In most cases, primary care physicians, rather than psychiatrists, do the prescribing (Beardsley et al. 1988). Among elderly individuals, there is a statistically significant association of sedative-hypnotic drug use with adverse reactions such as cognitive impairment, falls and hip fractures (Larson et al. 1987; Ray et al. 1987; Robbins et al. 1989; Sorock & Shimkin 1988; Tinetti et al. 1988). Within the class of benzodiazepines, the risk of hip fracture appears to be greater in association with long halflife, accumulating drugs than with short half-life agents (Ray et al. 1989). Studies by the Boston Collaborative Drug Surveillance Program (BCDSP), evaluating the frequency of unwanted drowsiness associated with benzodiazepine use by hospitalised medical patients, indicated an increased frequency of such adverse reactions among elderly individuals. This association was found for the anxiolytic drugs chlordiazepoxide and diazepam (BCDSP
Anxiolytics and Hypnotics in the Elderly
(973), and for the hypnotics flurazepam and nitrazepam (Greenblatt & Divoll Allen 1978; Greenblatt et al. 1977). Insofar as could be determined, the association of adverse reactions with age was not explained by factors such as gender, dosage or severity of underlying disease. The limitations of such studies must be recognised. They demonstrate association and not causality. Furthermore, selection bias cannot be ruled out: individuals whose underlying condition or degree of debility inherently increased the risk of drug therapy were precisely those patients who received medications. Taken collectively, the data indicate that the issue of appropriate sedative-anxiolytic prescribing for elderly individuals is one of large scale (Thompson et al. 1983). There is a pressing need for continuing development of scientific data on which to base an understanding of drug response and proper prescribing for elderly patients (Monamat et al. 1989).
2. Scientific Bases The fundamental clinical question can be summarised as follows (Shader et al. 1987): in utilising pharmacotherapy for the treatment of agitation, anxiety or sleep disorders, what adjustments should be made in the approach to therapy in the elderly patient such that the likelihood of therapeutic efficacy is maximised, and the likelihood of adverse reactions is minimised? Possible alterations in the approach to therapy using a specific medication might include a change in the size of each dose, the interval between doses, the duration oftherapy or the approach to initiating and terminating treatment. An altogether different choice of drug may be appropriate. Diagnostic criteria and methods of assessing efficacy and toxicity might also require alteration during treatment of aging populations (Gurian & Miner 1990; Shamoian 1990). More complex considerations include the assessment of intercurrent or coincident medical or psychiatric illness, concurrent drug therapy and the patient's diet and living situation (Cohen 1990; Salzman (990). For practical purposes, 'drug therapy' implies 'benzodiazepine therapy', since benzodiaze-
167
pines are by far the predominant class of anxiolytic and hypnotic drugs in use today. The vast majority of available scientific data refers to the benzodiazepines (Greenblatt et al. 1983b), although recently some data have become available for the nonbenzodiazepine anxiolytic buspirone. The clinical phenomenon most commonly confronting health care professionals involved in drug therapy of the elderly is that a given dose ofa given drug produces a different (or perhaps unexpected) response in an elderly person compared with a younger individual of the same gender and similar bodyweight. The most obvious explanations are pharmacokinetic and/or pharmacodynamic in nature (Greenblatt & Shader 1990). In pharmacokinetic terms reduced drug clearance (CL) is often associated with old age. By definition, reduced CL implies that, after any specific dose of a given drug, an elderly person will have a greater area under the plasma concentration-time curve (AUC) than a corresponding young 'control' subject. This in tum implies that at some (or perhaps all) times after a single dose, plasma drug concentrations are higher in the elderly. If plasma concentration is proportional to target organ concentration, which in tum is proportional to clinical response, then the elderly person may have a greater drug response than the young control person, at least at some time after a single dose. During long term dosage, steady-state plasma concentration (CSS) is inversely proportional to CL. Therefore, the elderly person receiving therapy at a specific dosing rate will have a higher css than the young control; with this comes the possibility of greater drug response in the elderly. In pharmacodynamic terms, the aging organism may have a greater drug 'sensitivity' than a young control individual. That is, any given plasma concentration, or target organ concentration, produces a greater clinical response in an elderly than in a young organism. In principle, increased pharmacodynamic sensitivity in the elderly may coexist with, or be independent of, pharmacokinetic changes. Requirements for scientific excellence differ substantially between the pharmacokinetic and
Clin. Pharmacokinet. 21 (3) 1991
168
Dose
Fig. 1. Schematic representation of the cascade of events between drug administration and pharmacodynamic effect.
pharmacodynamic dimensions. Pharmacokinetic studies have relatively tangible and objective methodological requirements. Studies of intrinsic drug sensitivity, on the other hand, must necessarily include the complex cascade of events between drug administration and response (fig. I). Not surprisingly, the methodological requirements are more stringent and difficult. Furthermore, the elucidation of mechanisms almost always requires experimental data in addition to clinical data, since many steps on the cascade are not ordinarily accessible to study in humans.
3. Pharmacokinetic Changes in Old Age 3.1 Methodological Principles Scientific data on age-related changes in pharmacokinetics and drug metabolism have been subject to continuing, extensive review (Dumas et at. 1990; Greenblatt et at. 1982a, 1983a, 1986a, 1989a; Loi & Vestal 1988; Ouslander 1981 ; Schmucker 1985a,b; Vestal 1982, 1989). Since hepatic biotransformation is the principal route of clearance for essentially every clinically relevant anxiolytic and hypnotic drug, age-related alterations in hyp-
notic drug-metabolising capacity are of most relevance. The pharmacokinetic consequences of a reduction in hepatic drug clearance (CLH) after oral dosage will depend on whether the compound is a high- or low-clearance drug (Wilkinson 1987). The majority of benzodiazepines can be characterised as low-clearance drugs, with CLH considerably less than hepatic blood flow (Q H) [Greenblatt & Shader 1987}. Under these circumstances, a reduction in CL will be evident mainly as a prolongation of elimination half-life (t 'l'~) rather than a change in peak plasma concentration (C max ) of the parent compound. A few benzodiazepines (such as triazolam and midazolam), as well as the nonbenzodiazepine buspirone, have values ofCLH which are somewhat closer to QH. For such drugs, reduced CL may also be evident as lower presystemic extraction, leading to higher plasma concentrations at all times after administration (Pond & Tozer 1984; Wilkinson 1987). Prolongation of tl;'~ may also be evident, although this is generally less dramatic than with low-clearance drugs. In clinical practice, old age per se cannot usually be separated from consequences of aging such as intercurrent disease, use of other medications, alterations in diet and body habitus, and changes in activity. These other factors may potentially be as important as or more important than age itself as a determinant of drug-metabolising capacity. The scientific study of aging itself, however, requires control or exclusion of these other factors insofar as this is possible. Therefore, the most meaningful and convincing studies of pharmacokinetics and aging involve healthy elderly populations, with the proviso that further study is necessary to elucidate the pharmacokinetic consequences of age combined with other factors. It is clear that pharmacogeriatric studies can be carried out with proper implementation of the principles of ethical research and informed consent (Abernethy & Azarnoff 1990). Increasing evidence indicates that gender as well as age may influence drug distribution and clearance. Therefore, the problem of pharmacokinetics in aging cannot be separated from the issue of pharmacokinetics and gender; i.e. a well designed study should evaluate pharmacokinetics
Anxiolytics and Hypnotics in the Elderly
in elderly persons, and in young controls, of both genders. Statistical issues also assume great importance. Greater sample sizes in each cell increase the likelihood that the study population is representative of the general population, and assure sufficient statistical power for an accurate assessment. In actual research practice, the desirability of a large sample size in each cell is always tempered by feasibility considerations of time, funds, energy and the availability of suitable study participants. 3.2 Pharmacokinetics of Individual Drugs Findings across studies and laboratories are not always completely consistent, but some general conclusions can be offered. For benzodiazepines metabolised by hepatic microsomal oxidation, old age is generally associated with reduced CL. In many studies, the age-related reduction is more evident in men than in women. For those drugs metabolised mainly by glucuronide conjugation or by nitroreduction, age-related alterations in CL are far Jess evident. Relative increases in the fraction oriotal bodyweight comprised of adipose tissue, as occur with both old age and female gender, are of considerable importance. Differences in pharmacokinetic volume of distribution (Vd) in the elderly compared with the young, and in women compared with men, may be influenced by the lipid solubility of the drug in question (Greenblatt et al. 1982a). Changes in Vd will by themselves cause reciprocal changes in t1hi3' even in the absence of a change in CL (Abernethy & Greenblatt 1982, 1986). Because old age may also alter the plasma protein binding of benzodiazepines, evaluation of agerelated changes in free fraction (fu) is generally another necessary component of such studies. 3.2.1 Chlordiazepoxide Chlordiazepoxide, available since 1960, was the first of the benzodiazepines to be used in clinical practice (Greenblatt & Shader 1974). This drug is metabolised by hepatic microsomal oxidation. Its initial metabolite, desmethylchlordiazepoxide, is pharmacologically active, as are subsequent prod-
169
ucts in the metabolic cascade (demoxepam, desmethyldiazepam) [Greenblatt et al. 1978]. Shader et al. (1977) administered single oral doses of chlordiazepoxide hydrochloride 25mg to healthy young and elderly men. Compared with young controls, chlordiazepoxide tlhi3 was significantly prolonged in the elderly (18.2 vs 10.1h), and CL was significantly reduced (0.35 vs 0.61 ml/min/ kg). Increased Vd was also observed in elderly individuals. Roberts et al. (1978) administered single intravenous doses of chlordiazepoxide hydrochloride 0.6 mg/kg to 27 healthy subjects aged 16 to 86 years. There was a significant positive correlation of t'l2i3 and Vd with age; CL likewise declined significantly. Chlordiazepoxide fu in plasma averaged 3.5%, and was not significantly related to age. Greenblatt et al. (l989b) administered doses of chlordiazepoxide hydrochloride 50mg both orally and intravenously to groups of young male, young female, elderly male and elderly female volunteers. As in the previous 2 studies, the tlhi3 was significantly prolonged in elderly vs young men (20.4 vs 7.9h), and CL significantly reduced (0.26 vs 0.59 ml/min/kg). Among women, however, age-related changes did not approach statistical significance. Chlordiazepoxide fu in plasma was significantly higher in elderly as opposed to young women (6.6 vs 5.9%); the values in elderly and young men were nearly identical (6.6 vs 6.8%). The absolute bioavailability of oral chlordiazepoxide was not significantly different from 100%, and was not influenced by either age or gender. 3.2.2 Diazepam Approximately 50% of the total clearance of diazepam in humans is accounted for by the oxidative reaction of N-demethylation, yielding desmethyldiazepam as the principal metabolite (Greenblatt et al. 1988a). The other component of CL is explained in part by hydroxylation to form temazepam, as well as by other, unidentified pathways (Klotz et al. 1980; Mandelli et al. 1978). As such, oxidation is the principal mechanism of diazepam elimination. Following single doses of diazepam, the appearance of desmethyldiazepam in systemic circulation is relatively slow, such that it
CUn. Pharmacokinet. 21 (3) 1991
170
probably contributes minimally to the pharmacodynamic action of a single dose of diazepam. During iong term therallY, however, both diazepam and desmethyldiazepam accumulate in plasma, with the CSs of the metabolite similar to or exceeding that of the parent drug (Greenblatt et al. 198Ia). Because the CL of temazepam is higher than that of either diazepam or desmethyldiazepam, temazepam CSS during long term therapy with diazepam is relatively low, as are plasma concentrations of oxazepam formed from demethyldiazepam. The first pharmacogeriatric study of diazepam in humans was that by Klotz and associates (1975), in which 20 healthy volunteers (17 men and 3 women) aged 15 to 82 years received a single intravenous dose of the drug. Its Vd and tl/211 both increased significantly with age. A key finding of this study was the relationship of CL with age: among the 20 healthy individuals, the correlation coefficient was -0.39 (p = 0.089). Since this value does not reach statistical significance at the p = 0.05 level, it was concluded that the observed decline in CL with age could have been due to chance, and therefore there was no statistical evidence for a decline in metabolising capacity with age. In the light of subsequent studies (see below), it is more reasonable to recognise that the failure to reach statistical significance was simply due to the small sample size. It should be noted that diazepam fu in this study averaged 2.6%, and was not significantly related to age. Kanto et al. (1979) next reported the effect of age on the pharmacokinetics of single intravenous doses of diazepam in 14 patients aged 32 to 78 years, receiving the drug in the context of premedication prior to general anaesthesia. Both men and women were represented in the sample. They were not healthy volunteers; many had intercurrent medical disease and were receiving other medications concurrently. In any case, diazepam CL decreased significantly with age. As in the previous study by Klotz et al. (1975), both tl/211 and Vd increased significantly with age; diazepam fu was not measured in this study. Macklon et al. (1980) administered intravenous diazepam to 19 subjects of both genders, with ages ranging from 18 to 95
years. Again, its tl/211 and Vd increased significantly with age. The fu also increased significantly with age, emphasising the need for correction of CL values for individual differences in fu. After such correction, the CL of unbound diazepam was found to decline significantly with increasing age. Greenblatt et al. (1980a) studied 4 groups of subjects: young male and female (aged 21 to 37 years), and elderly male and female (aged 61 to 84 years) [fig. 2]. The fu of diazepam was significantly higher in the elderly volunteers, in part because of lower plasma albumin levels. After correction for individual differences of fu, the Vd of unbound diazepam was higher in the elderly, but the effect of gender was more impressive than that of age. That is, women had a larger Vd than men, regardless of age. The CL of unbound diazepam was significantly lower in elderly than in young subjects regardless of gender (fig. 3), and tl/211 likewise was greater in the elderly. Ochs et al. (l98Ia) admini-
400
• = Diazepam
200
o = Desmethyldiazepam
100 50 20
::r c; .5
10 5
c:
i
Male, 24 years
2
8c: 300
8
as E as 100 ii:
'"
50
20
..,.--
.o---~--
10 6
3
p"" {"
__
-o ______
~ ~
Male, 69 years
o 1 234 5 6 789 Time (days)
Fig. 2. Plasma concentrations of diazepam and its major metabolite, desmethyldiazepam, after intravenous doses of the parent drug in representative young and elderly volunteers (from Greenblatt et al. 1980a, with permission).
171
Anxiolytics and Hypnotics in the Elderly
stered single intravenous doses of diazepam to 27 male volunteers aged 20 to 91 years. Again, tl/,{j and Vd increased significantly with age; clearance of total and unbound diazepam declined, but the correlation coefficient did not reach significance. Divoll et al. (1983) replicated the study design of Greenblatt et al. (1980a), with the addition of intramuscular and oral doses of diazepam in all subjects using a crossover design. After intravenous administration, the effects of age and gender on diazepam pharmacokinetics were very similar to those reported by Greenblatt et al. (1980a). In addition, the absolute bioavailability of oral diazepam was not significantly different from 100% complete, and was not influenced by either age or gender. The absolute bioavailability of a single intramuscular deltoid injection of diazepam averaged 101 and 97% in young and elderly men, respectively; in young and elderly women, the corresponding numbers were 77 and 86%, which may be explained by the possibility of inadvertent intraadipose as opposed to intramuscular injection in women. Only the study of MacLeod et al. (1979) failed to find any significant age-related changes in the pharmacokinetic profile of diazepam. This may represent the consequences of a small sample size (n = 10, 5 male and 5 female), together with the presence of significant intercurrent illness in several of the subjects. The weight of evidence strongly suggests that the CL of unbound diazepam declines with increasing age among healthy individuals. Vd also increases with age, probably due to the increasing fraction of total bodyweight accounted for by adipose tissue among the elderly. Similarly, diazepam Vd appears to be greater in women than in men, again presumably due to the difference in body habitus. The combination of reduced CL and increased Vd leads to a marked prolongation of diazepam tl/,{j in the elderly. In some studies, the plasma fu of this highly protein-bound drug also increases with age. Therefore, accounting for individual differences in fu is an important aspect of these pharmacokinetic analyses.
-Young
80.0
Elderly I ~
0 0
Cii 70.0 ~
c
E 60.0
;:,.
.§. ~ c 50.0
~
'"
U
U c
~ c
20.0
:::l
.8c :::l
.--.-.......
...........--30.0
Cc
•
40.0
'iii
"0
•
•••• ._........... •••
c 0
--ll---
0
c --'0"'-'---
e CcD 0
0
•
•• -"'oi!~-"
.... -.-.--....
___9. ____ cC c 0 0
••••
10.0 Male
Female
•
Male
Female
0.0 + - - - - + - - - + - - - - - 1 - - - - 1
Fig. 3. Individual and mean (± SE) values of clearance of unbound diazepam in groups of young male, young female, elderly male and elderly female volunteers; ., D = nonsmokers; e, 0 = cigarette smokers (from Greenblatt et al. 1980a, with permission).
3.2.3 Desmethyldiazepam Desmethyldiazepam is an important benzodiazepine. In many parts of the world, it is administered as the pure compound for the treatment of anxiety. Several benzodiazepines (such as clorazepate, prazepam, ketazolam and oxazolam) effectively serve as 'prod rugs' of desmethyldiazepam, being converted to it as the major active substance (Greenblatt et al. 1983a; Ochs et al. I 984a). For still other agents of this class (diazepam, chlordiazepoxide, medazepam, pinazepam and halazepam), desmethyldiazepam is a principal pharmacologically active metabolite. Desmethyldiazepam itself is biotransformed mainly by the oxidative reaction of hydroxylation, yielding oxazepam as the major metabolite. Oxazepam, in turn, is rapidly conjugated and excreted, such that plasma concentrations of intact oxazepam are much lower than those of the parent drug following administration of desmethyldiazepam to humans. Klotz and Miiller-Seydlitz (1979), in a limited study of 4 young and 4 elderly individuals, suggested a prolongation of desmethyldiazepam t'l2,3,
Clin. Pharmacokinet. 21 (3) 1991
172
30
•
~~-
20
:J"
~
10
•
c:
I
0
~c: CD 0
I
4
I
c: 0
E as
"0 N
~.
I
"iii
'"
r'~
= 20.6h
I
~
as E as
tv.
t
0
'-........,.
0.5
1i:
Male, 68 years
Male, 37 years 0.2
I
I
I
I
I
I
0
10
20
30
40
0
I
i
I
I
I
10
20
30
40
50
Time (h)
Fig. 4. Plasma drug concentrations after single oral doses of alprazolam I mg in representative young and elderly male volunteers; till = half-life (from Greenblatt et al. 1983c, with permission).
increase in Vd and reduction of CL in the elderly as opposed to the young. Divoll Allen et al. (1980) studied the pharmacokinetics of desmethyldiazepam after single oral doses of prazepam in 29 subjects aged 22 to 85 years: as with diazepam, desmethyldiazepam Vd was larger in women regardless of age, and increased with age in both men and women. However, the effect of age on the elimination and CL of desmethyldiazepam was highly gender-dependent. The tV2P increased significantly with age in elderly vs young men (128 vs 62h), but was unrelated to age in women. The CL of unbound desmethyldiazepam was unrelated to age in women, but declined with age in men (r = -0.4); the latter effect approached, but did not reach, statistical significance. Shader et al. (1981) repeated this study except that clorazepate dipotassium rather than prazepam was used as the precursor of desmethyldiazepam. The effects of age and gender on desmethlyldiazepam Vd and tV2P were essentially the same as those reported by Divoll Allen et al. (1980). In the Shader et al. (1981) study, the CL of unbound desmethyldiazepam declined significantly with age in men, but was not significantly related to age in women.
3.2.4 Desalkyljlurazepam Although not administered as such in clinical practice, desalkylflurazepam is the final and major metabolite of the benzodiazepine precursor flurazepam, and probably accounts for most of the clinical activity observed after use of the latter in humans (Miller et al. 1988). A new, recently introduced benzodiazepine, quazepam, also has desalkylflurazepam as the final metabolite (Zampaglione et al. 1985). Biotransformation of desalkylflurazepam itself appears to proceed by hydroxylation, with the hydroxylated metabolite rapidly conjugated and excreted. Using flurazepam as the precursor substance for desalkylflurazepam, Greenblatt et al. (1981 b) administered single 15mg oral doses to 26 healthy subjects aged 19 to 85 years. The tV2P was significantly prolonged in elderly vs young men (160 vs 74h), but the prolongation oftv2P in elderly vs young women (120 vs 90h) was not significant. Since desalkylflurazepam was administered as a precursor rather than the intact substance, CL could not be calculated. During multiple dosage with flurazepam 15mg nightly for 15 nights, the CSs of desalkylflurazepam was significantly higher in elderly
173
Anxiolytics and Hypnotics in the Elderly
than in young men (81 vs 53 jLg/L), but was not significantly different in elderly as opposed to young women. Using quazepam as the precursor, Hilbert et al. (1984) found a mean desalkylflurazepam t'hi3 of 190h in 10 healthy elderly volunteers (6 men and 4 women) aged 65 to 77 years. Although young control subjects were not included in this study, the authors suggest that a comparison of values in the elderly subjects with those previously established in young volunteers is consistent with a substantial prolongation of t'hi3 in the elderly. 3.2.5 Alprazolam The triazolobenzodiazepine alprazolam is currently the most widely prescribed benzodiazepine in the US. It is metabolised mainly by the oxidative reaction of hydroxylation, yielding 2 metabolites which are less pharmacologically active than the parent drug (Sethy & Harris 1982). Since the metabolites are rapidly conjugated and excreted, unconjugated concentrations of these hydroxylated products are considerably lower than those of the parent compound (Smith & Kroboth 1987). Greenblatt et al. (1983c) administered single oral doses of alprazolam I mg to young men and women aged 21 to 45, and elderly men and women (62 to 78 years). The Vd was slightly smaller in elderly than in young men, and was unrelated to age in women; t'/'11 was significantly prolonged in elderly compared with young men (19 vs h), and CL significantly reduced (0.64 vs 1.33 ml/min/kg) [figs 4, 5]. Among women, however, the slight prolongation of t'hi3 in the elderly (13.5 vs 1O.8h) did not reach significance, nor did the slight reduction in mean CL (1.07 vs 1.34 ml/min/kg). Plasma fu was unrelated to age. As with desmethyldiazepam, the age-dependent decrement in alprazolam CL appears to be gender-specific. Kroboth et al. (1990) studied the pharmacokinetics and clinical effects of alprazolam during a multiple dosage regimen administered to 16 men and 10 women aged 65 to 80 years. Although there was no concurrent study of young control subjects, the reported mean values of CL and t'hi3 in the group of elderly men (0.56 ml/min/kg; 24.0h) are similar to those reported for elderly male subjects by Greenblatt et al. (1983c),
but the mean values for elderly women (CL 0.63 ml/min/kg; t'hi3 22.4h) are substantially different. The reasons for these discrepancies are not established, but the findings emphasise the potential effect of unknown differences in study population on pharmacokinetic results. 3.2.6 Triazolam The triazolobenzodiazepine triazolam is oxidatively metabolised analogously to alprazolam (Eberts et al. 1981), with hydroxylated metabolites of essentially no clinical importance (Sethy & Harris 1982). It has become the most commonly prescribed benzodiazepine hypnotic in the US. Greenblatt et al. (1983d) administered single oral doses of triazolam 0.5mg to groups of young male, young female, elderly male and elderly female subjects. A highly significant age-related decrement in triazolam CL was observed in elderly as opposed to young volunteers, and the effect was essentially independent of gender. Apparently because triazolam has a somewhat higher value of CLH than the other benzodiazepines described above, the reduced CL in the elderly (that is, increased AUC) was evident mainly as higher plasma concentra-
Young
2.5
Elderly 1-
o
2.0
1.5
•
•
.__ ...... .......
o
--_._._.. _---0
•
B
............ _... 1.0
o
••
• •
0
.. __·0········ 0 0
0
0.5 Male
--_._ .. _.....
0
Female
0
•• • ._------------------...... --..... •• •• Male
0
.. ............... '
0 0 0
Female
O+-------+-------+-------+-----~
Fig. 5. Individual and mean (± SE) values of alprazolam clearance in groups of young male, young female, elderly male and elderly female volunteers (from Greenblatt et al. \ 983c, with permission)
174
tions at all times after dosage. Although values of tl1211 were somewhat longer in the elderly, these differences were not statistically significant. These findings were subsequently replicated by Greenblatt et al. (1991) in an entirely different set of volunteers who received oral triazolam 0.12S and 0.2Smg. After both doses oral CL was significantly lower in the elderly subjects regardless of gender. However, these findings were not replicated by Smith et al. (1983), who administered single oral doses of triazolam O.Smg to 30 men and women aged 20 to 76 years. There were no apparent effects of age on its pharmacokinetic properties. The plasma triazolam concentration determinations for the studies of Greenblatt et al. (1983d, 1991) and Smith et al. (1983) were all made in the same laboratory using the same analytical methodology (Greenblatt et al. 1981 c; Scavone et al. 1986). This suggests that the differences in pharmacokinetic outcome are attributable to variations in the study populations or in experimental design.
3.2.7 Midazolam Midazolam has a CL falling in the range of SO% of QH (Reves et al. 1985). Its 2 hydroxylated metabolites formed by microsomal oxidation are of minimal clinical importance, since their intrinsic activity and relative brain uptake are considerably less than those of the parent compound (Arendt et al. 1987). In a study of 11 young and 11 elderly women, Avram et al. (1983) found no effect of age on the pharmacokinetics of a single intravenous dose of midazolam 0.2 mg/kg. Greenblatt et al. (1984) again studied groups of young and elderly men and women, who received single intravenous (2.S to Smg) and oral (S to lOmg) doses of the drug. Among women the CL of midazolam was lower and tl1211 longer in the elderly; however, none of the differences was statistically significant. Among men, tlj211 was significantly prolonged in the elderly (S.6 vs 2.1h), and CL significantly reduced (4.4 vs 7.8 mlf min/kg); apparently as a consequence of the latter, the absolute systemic availability of oral midazolam was greater in elderly than in young men (SO vs 41 %). Although midazolam was highly protein
Clin. Pharmacokinet. 21 (3) 1991
bound in all subjects (3 to 4% unbound), fu was not influenced by age or gender. Similar findings were reported by Holazo et al. (1988), who administered single intramuscular doses ofmidazolam 7.S mg to male and female volunteers aged 20 to 7S years. Midazolam CL was significantly lower in elderly than in young men, but was not significantly related to age in women. In a population study of intravenous midazolam among 168 surgical patients, no significant relationship was observed between age and tl1211 (Kassai et al. 1988).
3.2.8 Clobazam The I,S-benzodiazepine clobazam is biotransformed by the oxidative reaction of N-demethylation. Greenblatt et al. (1981d) studied the effects of age and gender on the disposition of single oral doses of clobazam. The tl12/l was significantly longer in elderly than in young men (48 vs 17h); among women, it increased with age, but the differences were less striking (49 vs 31 h). Vd increased with age regardless of sex, and within each age group was larger in women than in men. Total CL of clobazam was not significantly related to age among women; in the male volunteers, it declined significantly with age. Clobazam fu also increased with age, but did not influence the pharmacokinetic conclusions. In a subsequent study, Greenblatt et al. (1983d) administered the drug on a multiple dose basis to a subgroup of the same volunteers. Clobazam CSs was highly consistent with the values predicted from CL observed in the single-dose study. Between young and elderly women, CSs was highly consistent with the values predicted from CL observed in the single-dose study. Between young and elderly women, css and steady-state CL of clobazam were not significantly different; among male volunteers, however, the latter was significantly reduced, and CSS significantly increased, in elderly compared with young subjects. Thus, the age- and gender-related changes in clobazam pharmacokinetics observed after single doses lead to predictable consequences when the drug is administered on a multiple-dose basis.
Anxiolytics and Hypnotics in the Elderly
3.2.9 Brotizolam Brotizolam is a thienodiazepine that is metabolised by hepatic microsomal oxidation (Bechtel et al. 1986a,b). 10chemsen et al. (l983a) found a significant prolongation of t'l211 and a reduction in total CL in elderly as opposed to young volunteers. Interpretation of these findings is complicated by the fact that all of the elderly volunteers had significant intercurrent medical illness and in most cases were receiving concurrent medications. 3.2.10 Loprazolam Loprazolam is a benzodiazepine hypnotic biotransformed by hepatic microsomal oxidation. Swift et al. (1985a) found a significantly longer t'l211 among 9 elderly volunteers compared with 10 young controls (19.8 vs 11.2h) after oralloprazolam Img. Oral CL was correspondingly reduced in the elderly. 3.2.11 Bromazepam Ochs et al. (1987) administered single oral doses of bromazepam 6mg, an oxidatively metabolised benzodiazepine (Schwartz et al. 1973), to 32 healthy young and elderly men and women. Its Vd was significantly lower in the elderly, and oral CL likewise was significantly reduced; these age-related changes were not influenced by gender. Although bromazepam fu increased significantly with age, correction of the pharmacokinetic variables for individual differences did not alter the conclusions. 3.2.12 Metaclazepam Metaclazepam is a short half-life benzodiazepine metabolised by microsomal oxidation. Molz et al. (1985) administered single oral doses of metaclazepam 10mg to elderly volunteers and 10 young controls. There were no differences between the groups in the t'hll or CL of the drug or of its principal metabolite, formed by N-demethylation. 3.2.13 Clotiazepam Clotiazepam is biotransformed by microsomal oxidation, leading to a series of hydroxylated and demethylated metabolites (Arendt et al. 1982). Ochs et al. (1984b) administered single oral doses of clotiazepam 5mg to 29 healthy volunteers aged
175
between 22 and 89 years. Among men, clotiazepam CL was negatively correlated with age (r = -0.31), but the association did not reach significance. Among women the CL was, if anything, positively correlated with age. 3.2.14 Oxazepam Unlike the benzodiazepines discussed above, the major metabolic pathway of oxazepam in humans involves glucuronide conjugation as opposed to microsomal oxidation (Greenblatt 1981). Shull et al. (1976) first reported that the pharmacokinetic parameters did not differ significantly between 8 young and 8 elderly volunteers after an oral dose of oxazepam 45mg. Greenblatt et al. (1980b) found no significant age-related changes in oxazepam t'hll or CL among 18 male volunteers aged 22 to 76 years, or among 20 female volunteers aged 28 to 84 years. The fu averaged 4.3% in men and 4.4% in women; these values were not significantly related to age. Murray et al. (1981) found no significant correlation of age with any pharmacokinetic parameters among 13 healthy volunteers who received a single dose of oxazepam 30mg. Ochs et al. ( 1981 b) studied the pharmacokinetics of single oral doses of oxazepam 30mg in 22 male and 9 female volunteers aged 22 to 86 years: neither t'h nor CL was significantly related to age in either sex. 3.2.15 Lorazepam Lorazepam, like oxazepam, is metabolised by glucuronide conjugation, with the principal metabolite (lorazepam glucuronide) recovered in urine (Greenblatt 1981; Greenblatt et al. 1976). Kraus et al. (1978) found no significant effect of age on the pharmacokinetics oflorazepam among 11 male volunteers aged 15 to 73 years. Greenblatt et al. (1979) compared the pharmacokinetics of single intravenous doses of lorazepam 2mg between 15 young (19 to 38 years) and 15 elderly subjects (60 to 84 years). The Vd was smaller in the elderly subjects (0.99 vs 1.11 Ljkg); t'hll did not differ between groups, but there was a statistically significant reduction in lorazepam CL in the elderly (0.77 vs 0.99 ml/minjkg). The findings were
176
apparently not influenced by gender. The overall mean lorazepam fu was 11.1 %, and increased slightly but significantly with age (Divoll & Greenblatt 1982). Absolute bioavailability of intramuscular and oral lorazepam was also evaluated in a subgroup of the elderly subjects, and did not differ significantly from 100% by either route. Aaltonen et al. (1982) administered intravenous lorazepam 0.03 mg/kg to 14 surgical patients aged 25 to 86 years. Age was not significantly associated with t'h/l' Vd, CL or fu in plasma. 3.2.16 Temazepam Temazepam is also biotransformed mainly by direct glucuronide conjugation (Locniskar & Greenblatt 1990; Schwarz 1979). Divoll et al. (1981) found no significant association of age with temazepam t'h/l or CL in a series of 32 healthy male and female volunteers aged 24 to 84 years. The fu increased significantly with age, but this did not influence the conclusions of the study. Smith et al. (1983) studied the pharmacokinetics of single oral doses of temazepam in 30 male and female volunteers aged 22 to 76 years. There appeared to be a trend towards reduced CL with age among women, but the association was not significant. No such association was found in men. Briggs et al. (1980) studied the pharmacokinetics and -dynamics of this drug in young and elderly subjects, but a formal pharmacokinetic comparison between groups was not possible. 3.2.17 Lormetazepam Lormetazepam is a fourth benzodiazepine biotransformed principally by glucuronide conjugation (Hiimpel et a. 1979). Hiimpel et al. (1980) found no significant age-related alterations in its pharmacokinetics between 6 young and 6 elderly volunteers. Three men and 3 women were included in each group, making the sample sizes too small to regard the conclusions as definitive. 3.2.18 Nitrazepam The benzodiazepine hypnotic nitrazepam is biotransformed mainly by nitroreduction (Kangas & Breimer 1981) although the initial metabolite, 7-
Clin. Pharmacokinet. 21 (3) 1991
amino-nitrazepam, has minimal pharmacological activity. The effect of age on its pharmacokinetics was first reported by Castleden et al. (1977), who administered single oral doses of nitrazepam 10mg to 10 elderly subjects aged 69 or older, and 10 young volunteers aged 40 or less. There were no significant differences between groups in t'l2/1 or Vd. Kangas et al. (1979) studied the pharmacokinetics of single oral doses of nitrazepam 5mg in 22 hospitalised, chronically ill geriatric subjects (all of whom were receiving other medications), in comparison with healthy young volunteers. The Vd was significantly greater in the elderly patients, and t'l2/1 was significantly prolonged. However, there was no difference between groups in nitrazepam CL; fu was not measured. Holm et al. (1982) found no significant difference in the pharmacokinetics of nitrazepam 5mg between 8 healthy elderly subjects and 8 young volunteers. 10chemsen et al. (1983b) studied the pharmacokinetics of intravenous nitrazepam 5.25mg in 9 young volunteers (22 to 49 years) and 8 healthy elderly subjects (67 to 76 years). Vd and t'h/l were both increased in elderly subjects, although the difference in t'h/l did not reach significance. However, there were no significant effects of age on the clearance of total or unbound nitrazepam. Greenblatt et al. (1985) evaluated the pharmacokinetics of single oral doses of nitrazepam 10mg in 40 healthy men and women aged 19 to 80 years: Vd was increased in elderly as opposed to young men, but did not differ significantly with age in women. Similarly t'l2/1 was prolonged in elderly vs young men, but did not change with age in women. The differences between young and elderly groups of either gender in clearance of total or unbound nitrazepam did not reach significance. 3.2.19 Buspirone Buspirone is not a benzodiazepine derivative, but its biotransformation proceeds mainly by hepatic microsomal oxidation (Goldberg 1984). After oral administration in humans, bus pirone undergoes extensive presystemic extraction leading to the appearance of a major metabolite, I-pyrimidinylpiperazine (I-PP). This metabolic product appears
Anxiolytics and Hypnotics in the Elderly
177
to be pharmacologically active, and its plasma concentrations considerably exceed those of the parent drug. Gammans et at. (1989) studied the pharmacokinetics of single and mUltiple oral doses of buspirone in 24 men and 24 women aged 20 to 77 years. There were no statistically significantly age or gender effects on the pharmacokinetics of the drug or its principal metabolite after single oral doses of buspirone 15mg, but considerable individual variation was seen, as might be expected for a drug with high presystemic extraction. The findings were similar during the multiple dose arm of the study, in which buspirone 15mg was administered every 8h for 4 days. 3.3 Predicting Age-Related Changes in Pharmacokinetics A number of studies have evaluated individual consistency in the capacity to biotransform various benzodiazepines. Among those benzodiazepines biotransformed by microsomal oxidation (for example, diazepam, desmethyldiazepam, desalkylflurazepam), there are significant within-individual correlations in metabolising capacity for the various drugs (Greenblatt et at. 1982b, 1983a,c), as
measured by t'l2{j or CL. In particular individuals who were slow metabolisers of one of these drugs were likely to be slow metabolisers of them all. The capacity to metabolise phenazone (antipyrine), the oxidatively-metabolised marker compound, was also significantly correlated with the capacity to biotransform low-clearance oxidatively metabolised benzodiazepines (Greenblatt et at. 1982b, 1983c, 1989b). However, the CL and t./, of phenazone were not correlated with the capacity to biotransform benzodiazepines metabolised mainly by conjugation (Greenblatt et at. 1982b). Although CL of the latter drugs is minimally influenced by age as such, there was still consistency in interindividual capacity (Greenblatt et at. I 980b, 1982b), which in turn was significantly correlated with the capacity to clear the conjugatively metabolised marker compound paracetamol (acetaminophen) [Greenblatt et at. 1983c). (Part II of this article, including references, will appear in the next issue of Clinical Pharmacokinetics.)
Correspondence and reprints: Dr D.J. Greenblatt, Division of Clinical Pharmacology, Box 1007, Tufts-New England Medical Center, 171 Harrison Avenue, Boston, MA 02111 , USA.
9th International Symposium on
Preparative and Industrial Chromatography - 'PREP 92' Date: 6-8 April 1992 Venue: Nancy, France For further information, please contact: Congres 'PREP-92' Societe Francaise de Chimie 1, rue Grandville - B.P. 451 F-54001 Nancy Cedex FRANCE
