Physiological aspects of geriatrie pharmacology

A number offactors which might be expected to affect drug disposition in the elderly

are listed in Table 1.Elevated gastric pH , delayed gastric emptying, reduced intestinal

blood flow as weil as other age-related alterations in gastro intestinal physiology

could affect drug absorption after oral administration. The available data are limited,

but do not indicate a significant alteration ofabsorption in the elderl y.

By contrast, drug distribution has been shown to be affected by age. Body composition is an important determinant of drug distribut ion and differs with age. Total

body water, both in absolute terms and as a percentage of body weight , has been

shown to decline by 10 to 15% between the ages of20 and 90 years. Lean body mass in

proportion to body weight also is diminished and seems to be due to a relative

increa se of 10 to 20% in body fat with age. Drugs that are distributed mainly in bod y

DR UG MET ABOLlS M STUDIES IN TH E ELD ERLY

Table 1 Factors which might affect drug disposition in the elderly.

109

Effect

Absorption

Distribution

Elimination

I) Hepatic metabolism

2) Renal excretion

Altered ph ysiology

t Gastric pH

LG.I. blood flow

LCells in G.I. mucosa

LG.1. motility

LTotal body water

LLean body mass

t Body fat

LSerum albumin

LHepatic blood flow

LHepatic mass

LEnzyme activity

LRenal plasma flow

LGlomeru lar filtration rate

LTubular function

water or lean body mass might have higher blood levels , particularly if the dose is

based on total body weight or surface area. Ethanol, for example, distributes in body

water, and higher peak ethanol levels were observed in older subjects without a

difference in rates ofmetabolism (Vestal, McGuire , Tobin, Andres, Norris & Mezey,

1977). The age dependent increase in the volume ofdistribution for diazepam, a very

lipid soluble drug, found by Klotz, Avant, Ho yumpa, Schenker & Wilkinson (1975)

could be due to an increase in body fat.

Many drugs are bound to albumin in the plasma. The reduction in serum albumin

concentration with age means fewer available binding sites. For example, Miller,

Adir & Vestal (1978) reported an ll % decrease in serum albumin concentration and a

25% increase in the unbound fraction of tolbutamide in an older group of healthy

subjects compared to a younger group. Since more unbound drug would be available

for distribution and metabolism , it was suggested that age differences in the pharmacokinetics oftolbutamide might be found. This may be the case for ph en ytoin. In one

study (Hayes, Langman & Short, 1975) total plasma c1earance was negatively

correlated with serum albumin and was greater in old than young subjects. Thus, age

differences in plasma bind ing ha ve been found for some drugs studied, but not all.

Unfortuna tely , the results of some studies are conflicting, pos sibl y du e to differences

in subject selection and th e techniques used.

Removal of drugs from the body occurs principally by two routes: (I) liver

metabolism to less active or inactive metabolites wh ich are usually excreted by the

kidney, or (2) excretion ofunchanged drug by the kidney. Both ofthese processes may

be altered in the elderly.

Studies in experimental an imals have shown reduced liver microsomal drug

metabolizing activity and alterations in liver microsomal enzyme induction with

ageing. Although there are no similar direct studies ofthe effects of age on liver drug

metabolizing enzyme acti vit y in man, it is known that liver mass and liver blood flow

decline with age. There is also evidence that in man the ageing process ma y result in

alteration of the intrinsic capacity of the liver for the biotransform ation of some

drugs. Studies with antipyrine , for example, have consistently reported a prolonged

half-life and reduced metabolic c1earance in older subjects (O'Malley, Crooks, Duke

& Ste venson, 1971; Vestal, Norris, Tobin, Cohen, Shock & Andres, 1975; Swift ,

Homeida, Halliwell & Roberts, 1978; Wood, Vestal, Wilkinson, Branch & Shand,

1979). In one study this was found even when c1earance was adjusted for an estimate

110 R. E. VESTAL

ofliver mass (Swift et a/., 1978). However , in other studies the age differences seemed

to be mostly explained by a decreased sensitiv ity ofthe elderly to enzyme induction

by environmental factors, such as cigarette smoking (Vestal et a/., 1975; Wood et a/.,

1979). Theophylline is a clinically important drug for which areduction of dosage in

the elderly is recommended because of what appears to be an age-related decline in

biotransformation by the liver (Jusko, Koup, Vance, Sehentag & Kur itzky, 1977).

Drugs requiring adjustment in dosage when given to the elderly have recently been

summarized by O'Malley et al. (1980). Not all drugs which have been studied show

an effect of age on half-life or clearance and some studies conducted by different

investigators with different groups of subjects, but using the same drug, have given

differing results (Vestal, 1979). Once again, this is probably due to differences in

methodology and study populations.

The decline in renal function with advancing age has been very weil documented.

For drugs or active metabolites which are primarily excreted by the kidney , this

means that clearance may be impaired in older patients and plasma levels may be

higher after doses which are standard for younger patients. The dosage of drugs such

as the aminoglycoside ant ibiotics, digoxin and lithium need to be reduced in the

presence of renal insufficiency, However , biological variation is such that we must

still use plasma levels to guide therapy with drugs which have serious toxic side

effects.

Methodological problems

Proper evaluation of the literature in geriatric clinical pharmacology and proper

design offuture studies requires familiarity with the strategies and pitfalls peculiar to

research in clinical gerontology (Rowe, 1977; Rowe & Troen, 1980).

Subject selection

One of the most important considerations governing the conduct and evaluation of

studies in geriatric clinical pharmacology is subject selection . If the purpose of the

study is to identify a possible effect of age on drug metabolism or drug response ,

every effort must be made to choose subjects who are free of diseases and drugs which

may interfere with this goal. Usually, this requires a careful history and physical

examination along with appropriate laboratory tests to exclude significant occult

disease, such as renal insufficiency, hepatic dysfunction, or anaemia. Since renal

function declines normally with age, age-adjusted criteria should be used to evaluate

renal function (Rowe, Andres , Tobin, Norris & Shock, 1976). Diet , habits, and

unusual occupational exposures should also be sought since these and possibly other

environmental influences are known to affect drug metabolism in the elderly (Vestal,

1978). On the other hand , one might prefer to take the view that studies should be

conducted in the population for which the drugs are intended, namely elderly

patients with one or more diseases of interest. In this case one is conducting a study of

age and disease and interpretation of the results becomes more difficult. In either

case the study population(s) should be carefully described along with the selection

criteria.

The importance of this distinction is illustrated in a study of Swift et al. (1978) in

which antipyrine disposition and liver size was examined in a group ofyoung healthy

subjects , a group of healthy elderly subjects, and a group of elderly hospitalized

patients (Table 2). This latter group were undergoing rehabilitation for locomotor

disorders such as hemiplegia or osteoarthritis and did not have other systemic

disease. All subjects were free of medications known to influence hepatic enzymes

DRUG METABOLISM STUDIES IN THE ELDERLY 111

Table2 Comparison of antipyrine disposition in young and elderly normal subjects and

elderlyhospitalizedpatients(rnodified fromSwiftet al., 1978).

Subjects Age range Antipyrine disposition

(years)

TI/, Volumeofdistribution Clearance

(mi minrl litre

(h) (litres) (mi kg') (mi min-I) liver volume'")

Young

normal 20-29 11.8±2.9 41.6±8.7 624±8.7 41~8±9 .7 33.3± lO.1

(n= 15)

Elderly

16.7±4.9b 32.8±5.8b 24.5±6.4b normal 75-86 566±51a 24.1±7.2c

(n= 11)

Elderly

500±92b hospitalized 70-89 lO.4±3.le 28.0±4.lb.

d 33.7± n.o« 42.7± 17.JC

(n= 10

Data shownare mean ± s.d. a p < 0.05 compared to youngnormalsubjects.

b p < 0.01 compared to youngnormal subjects.

cP < 0.001 compared to youngnormal subjects.

d P < 0.05 compared to elderlynormal subjects. e P < 0.01 compared to elderlynormal subjects.

either at the time ofstudy or for at least one month previously. However, nine ofthe

ten hospitalized patients were tak ing drugs of some type . All subjects were nonsmokers and had normal routine liver function tests and blood count. The results

demonstrated a significant difference in plasma half-life, volume of distribution and

plasma clearance, both in ml min-I and in ml min-I litre liver volume,-I between the

hospitalized elderly and the normal elderly. The higher antipyrine clearance in the

hospitalized group was unexplained.

Differences in subject selection may contribute to the conflicting observation

regarding the response of the elderly to oral anticoagulants. O'Malley, Stevenson,

Ward, Wood & Crooks (1977) retrospectively reviewed warfarin therapy in 177

hospitalized patients (age 30-83 years) . Although 27 percent were taking drugs

known to interact and 34 percent were taking drugs which might interact with

warfarin, patients were not excluded because of concomitant drug therapy for the

analysis of possible etTects of age. The overall trend was for the dose to fall with age

after the fifth decade such that in patients aged over 70 years the mean daily dose was

40 percent lower than patients aged 40-50 years (P < 0.001). The Thrombotest was

34 percent lower than in the fifth decade (P < 0.001), indicating greater anticoagulant effect despite the lower dose. These fmdings indicated a marked increase in

sensitivity to warfarin in the aged . Support for this conclusion comes from several

other studies (Shepherd, Hewick, Moreland & Stevenson, 1977; Husted & Andreasen,

1977; Routledge, Chapman, Davies & Rawlins, 1979). In these studies the etTect of

age on warfarin sensitivity was observed in patients taking no drugs at all or taking

drugs which are known not to interact with warfa rin, and there was no evidence of

significant pharmacokinetic age differences. In contrast, Hotraphinyo, Triggs,

Maybloon & Maclaine-Cross (1978) and Jones, Baran & Reidenberg (1980) do not

report a significant ditTerence in da ily maintenance dose between young and old

groups, but no exclusions were made because of liver disease, cardiac failure or

concurrent dr ügtherapy. However, it may weil be, as suggested by Jones et al. (1980),

that age is only a weak determinant ofwarfarin sensitivity compared to other factors.

It is also possible that intensive screening of the population may result in a select

group of elderly super-performers whose data do not reflect the influence of agerelated changes. For example, in conjunction with a study of tolbutamide

112 R.E. VESTAL

metabolism (Miller, Adir & Vestal, 1977,1978), measurements ofplasma glucose and

insulin were made during the first 30 min following intravenous tolbutamide (I g

70 kg-I body weight). Subjects were all completely free of medications, were

nonsmokers, and had a normal screening examination according to the criteria listed

in the previous section. Preliminary analysis has revealed that there was no

significant difference between the younger (age 23-46 years, n = 24) and the older

group (age 61-87 years, n = 19) in degree of obesity, maximum hypoglycaemic

response, or in the insulin response even when corrected for unbound plasma

tolbutamide which was increased in the older subjects. These results are quite

different from those reported by Swerdloff, Pozefsky, Tobin & Andres (1967) which

clearly demonstrate an age-related decline in glucose response to intravenous

tolbutamide in a larger sampie of subjects (n = 100) who were drawn from the same

population as the smaller study and who met criteria designed to exclude individuals

with overt carbohydrate intolerance. The explanation for this discrepancy is lacking

but may be due to differences in subject selection and sampie size.

Cross-sectional and longitudinal studies

Two general study designs are available for use in clinical gerontology (Rowe &

Troen, 1980). In longitudinal studies, serial prospective measurements are obtained

in one group of subjects at specified intervals, and the slopes for these variables or the

age-related changes, are determined. Thus, the rate of ageing is assessed in each

individual subject. Most longitudinal studies follow subjects concurrently in several

age cohorts throughout the adult age range and slopes for different ages can be

compared. Such studies are expensive and difftcult to perform because they require

follow-up of a stable population over a long period of time . They mayaiso be

complicated by methodological drift due to subtle changes in laboratory techniques

or equipment over several years. No longitudinal studies of drug metabolism have

been performed.

In the cross-sectional design, groups of various ages are studied . Only age

differences or effects 01 age, as opposed to age changes or effects 01ageing, can be

determined from this type ofstudy. This distinction is important when describing the

results of cross-sectional studies, since they may not reflect true age-related changes.

For example, it is important to remember that subjects over age 75 years come from a

cohort that has experienced at least 75% mortality. If the variable under study is

related to survival, a cross-sectional study will seem to show age differences that are

due to the progressive loss of individuals with high values rather than ageing

(Figure I). This phenomenon ofselective mortality can be avoided ifthe longitudinal

design is used. Thus, although there is evidence from cross-sectional studies to

suggest that advanced age is associated with impaired drug metabolism, it is possible

that this apparent difference between young and old individuals is the result of

selective mortality rather than ageing per se.

Protocol selection

Since the effects of age on drug metabolism may be difftcult to predict, conclusions

which seem to be contradictory may in fact prove to be quite compatible when

differences in the study protocol are considered. Studies of the effect of age on the

pharmacokinetics ofpropranolol are illustrative.

Castleden, Kayne & Parsons (1975) reported a 4-fold increase in plasma

propranolol levels after a single 40 mg oral dose in nine elderly as compared to nine

young subjects. It was proposed that the higher plasma levels in the elderly were the

result ofreduced hepatic extraction and metabolism, particularly a reduced first pass

DRUG MET ABOLlSM STUDIES IN THE ELDERLY 113

effect. A subsequent study provided additional support for this hypothesis (Castieden

& George, 1979). Allowing for the difference in doses, (40 mg orally and 0.15 mg kg-I

intravenously), first pass extraction and metabolism in eight elderly was 45.4 ± 8.0%

which was significantly less than in seven young subjects (69.9 ± 4.5% , P < 0.05).

Systemic clearance in the elderly was significantly less than in the young

(7.8 ± 1.3 ml kg-Imin-I vs 13.2 ± 1.4 ml kg! min,-I P < 0.02). Thus, the elderly (age

79 ± 3 years), only one ofwhom wasa cigarette smoker, achieved peak plasma levels

more than twice those found in the young (age 29 ± 2 years) after single dose

administration. None ofthe young subjects were smokers. The difference was about

4-fold after a multiple dose regimen totaling 160 mg daily. These data differ from our

own results (Table 3). Intrinsic clearance was not significantly different in the

X

<l>

:0

.Q

o---~.--.,(.o.. ... ..... ---<).lwc=:------ ...... ...... '""00:::)0:;,.,-..--

...... -------------_....:..::...~~

---0

Age

Figure1 Influence of selective mortality on age trends in cross-sectional data. In this figure

developed by Dr Reubin Andres ofthe National Institute on Agingthe X denotes a risk factor

associated with mortality. The circles indicate mean cross-sectional values and show an

apparent declinein X withage. This declinein X resultsfroma progressivelossof subjects with

high X levels in older age groups. That the levelof X is not influenced by age is shown by the

horizontallines. (Reproduced by permission; Rowe, 1977).

younger compared to the older subjects as total groups, but the results tor older

smokers were only 54 percent of the values for younger smokers (P < 0.025). In

addition, younger smokers had a significantly higher intrinsic clearance than did

younger nonsmokers (P < 0.005). Thus, our data would seem to indicate that

cigarette smoking is a more important variable than age, since in nonsmokers there

was no difference between the two age groups. However, the doses used in our studies

were higher than those used in the Castleden & George (1979) study (240 mg daily vs.

160 mg daily) . Conceivably, the threshold for maximum hepatic extraction may be

age-related. A reduced threshold in the elderly could result in saturation of the

extraction process at lower doses than in the young. The higher dose chosen for our

study may have obscured such a thresho1d difference. In this regard , Walle, Conradi,

Walle, Fagan & Gaffney (1978) have reported that between-patient variation in

plasma propranolol levels decrease from 3-fold at low doses to l.3-fold at higher

doses .