Europ. J. clin. Pharmacol. 15, 35-50 (1979)

European Journal of Clinical Pharmacology © by Springer-Vertag I979

Absolute Bioavailability of Chlorthalidone in Man: A Cross-Over Study After Intravenous and Oral Administration H. L. J. Fleuren t, Th. A. Thien 2, C. P. W. Verwey-van Wissen 1, and J. M. van Rossum 1 1Department of Pharmacologyand 2Department of Internal Medicine,RadboudHospital, Universityof Nijmegen, Nijmegen, The Netherlands

Summary. Seven normal human volunteers each received a constant-rate infusion of chtorthalidone for 2 h, and the same (commonly 50 mg) single oral dose on separate occasions. The concentration of unchanged chlorthalidone was analyzed over a 100 to 220 h period in plasma, red blood cells, urine and faeces after both dosage forms. A three compartment model was required to describe the intravenous plasma concentrations in five of the subjects. A two compartment model sufficed to account for the decay of the oral plasma concentrations in all seven subjects. The mean plasma ti/2 after i.v. dosing was 36.5 h (± 10.5 SD), and the mean plasma tl/2 after oral doses was 44.1h (+ 9.6SD). The mean red blood cell concentration tl/2 after i. v. doses was 46.4 h (-2-_ 9.9 SD), and the mean red blood cell t,/2 after the oral doses was 52.7 h (_ 9.0 SD). The shorter i. v. half-live was not equally manifest in all subjects, being mainly apparent in three of them. In all cases the urinary excretion rate plots were parallel to the plasma concentration curves. As the faster decay after i.v. administration was not accompanied by increased renal clearance, the difference must have been due to non-renal mechanism. The mean total of 65.4 (_+ 8.6 SD) % of the intravenous dose was excreted in urine over infinite time, whereas the mean total excretion after the oral dose was 43.8 (+_ 8.5 SD) %. Faecal excretion ranged from 1.3-8.5% of dose in the i. v. study to 17.5-31.2% of dose in the oral study. The sum of the amounts present in urine plus faeces pointed strongly to an important metabolic route of elimination of chlorthalidone. Bioavailability estimates (F) from three sets of data were - a mean F of 0.61 from plasma concentrations,

1 Unpublisheddata, obtainedfroma representativepopulationof ca. 300.000 subjectsregisteredin the NationalHealth Service.

0.67 from urinary excretion measurements and 0.72 from the erythrocyte concentrations. Simulations with a non-linear model indicated lesser validity of the estimate from erythrocyte concentrations. It was concluded that the average of plasma and urine data, F = 0.64, yielded the best estimate of the oral availability of chlorthalidone 50 mg in man.

Key words: chlorthalidone; pharmacokinetics, oral and i. v. doses, bioavailability.

The sulfonamide-diuretic chlorthalidone (Hygroton e) is widely used in drug therapy, e. g. it was prescribed in about 70 percent of medicinal treatments of moderate hypertension in the Netherlands in 1975 and 19761 . Despite frequent use since its introduction in about 1960 [1], knowledge of the fate of this drug in the human body is limited. The concentration in plasma, red cells and urine was recently reported after oral administration of therapeutic doses [2-6] and these studies revealed a relatively long half-life and strong binding by the red cells. The binding was incorporated in a non-linear pharmacokinetic model to explain the differing elimination half-lives for decay curves of the plasma and erythrocyte concentrations [5]. However, several important pharmacokinetic parameters, e. g. clearance and volume of distribution, remain uncertain as long as the absolute bioavailability is not known. In the present crossover study, plasma and red cell concentrations and the excretion of chlorthalidone in urine and faeces were measured both after intravenous and oral administration of single doses of the drug to seven normal human subjects. 0031-6970/79/0015/0035/$03.20

36

H. L. J. Fleuren et al.: Absolute Bioavailabilityof Chlorthalidone

Table 1. Details of the Subjects

Subject

Sex

Age (years)

T. Be

M

32

69

J.v. Bo

M

29

L. De T.Ho A. KI

M M F

24 24 60

M.O. M. We

M F

28 43

Body weight (kg)

Diagnosis

Co-medication and daily doses during triala (rag)

73

ulcus

110

90

ulcus

aluminium hydroxide susp. (Aluminox) 4 x 30 ml cimetidine 5 × 200 rag; diazepare 3 × 5 mg

73 76 57

101 90 77

healthy healthy mental distress, borderline prima~ hH~othyroidism, cholelithiasis healthy spondylitis

flurazepam 1 × 30 mg

83 62

Plasma creatinine (~tmol/1)

87 77

acenocoumarol 4 × 1 mg; amoxycillin 3 × 750 mg; diazepam 3 × 5 mg; nitrazepam 5 mg; flucloxacillin 3 × 1000 mg

a not on first half day of studies

Materials and Methods

Subfects and Drug Administration Seven adults (2 females and 5 males) participated in the study after giving informed consent; personal data are given in Table 1. Three were healthy men (L. De, T. Ho, M. O.), who rested in bed at night and were mainly ambulant or sitting during the trial whilst carrying out their normal activities as students in the Authors' laboratory. The four other subjects had been hospitalized for various reasons (see Table 1). They had normal values for serum urea, creatinine, bilirubin, alkaline phosphatase, G. O. T., G. P . T . , hematocrit, hemoglobin, plasma proteins and electro-cardiogram pattern. Therefore, all subjects were considered normal as far as kidney, liver and heart function were concerned. Sterile chlorthalidone solution for intravenous administration 100 mg/1 was prepared by dissolving pure chlorthalidone (kindly provided by Ciba-Geigy, Basle, Switzerland) in 0.9% NaC1, which had been brought to p H 8.0 with a few drops of 0.1 M N a O H . After sterile filtration (Millipore, pore diameter 0.21 ~tm) a known volume was transferred to a graduated infusion flask. The drug solutions were prepared the day before the intravenous experiment and were kept in a refrigerator until used to prevent bacterial growth; chlorthalidone was found to be chemically stable in these solutions for at least four weeks, as assayed by gas chromatography [7]. Each subject received a constant-rate infusion, via a fore-arm antecubital vein, over a period of approximately 2 h (exact times in Table 3). A n elec-

tronic drop counter (Ivac 531, Ivac Corporation, San Diego, Ca., USA) was used to maintain a zero-order infusion rate. The exact volume infused was calculated from the difference between the initial and remaining volumes in the flask. Five subjects received a dose of approximately 50 rag, one subject 60 mg and another 75 mg (exact amounts in Table 3), such that the doses applied were in the range of 0.66-0.88 mg per kg body weight. The infusions started in the morning at 9.00 a. m. after an overnight fast. The subjects took nothing except water during the first four hours of the study. In the oral study commercially available tablets were used (Hygroton ®, Ciba-Geigy). For our purpose tablets of 1 0 0 m g were divided and weighed until the desired amount was obtained, the dose being equal to that for the same subject in the intravenous study (exact oral doses in Table 4). The tablets were ingested, crushed with the teeth and swallowed completely with about 150 ml tap water, on an empty stomach, in the morning at 9.00 a. m. No food or beverages except water were taken until 4 - 5 h after administration of the drug. During the first two hours, a volume of 0.9% NaC1 (w/v) equal to the volume of fluid given in the intravenous study ( 5 0 0 - 7 5 0 ml, depending on the dose) was infused in order to exclude any systematic difference in the fluid status of the subjects between the oral and intravenous experiments. The order of the administrations was such that four subjects received the oral dose first (Day 1), followed by the intravenous dose (T. Be on Day 10, J. v. Bo on Day 11, A. K1 on Day 8 and M. W e on Day 10). Three subjects received the intravenous dose first (Day 1), followed by the tablet

H. L. J, Fleuren et al.: Absolute Bioavailability of Chlorthalidone

(L. De on Day 15, T. Ho on Day 15 and M. O. on Day 36).

Sampling and Analytical Measurements Venous blood samples (ca. 7 ml) were taken from the contralateral fore-arm at the following approximate times (exact times were noted) from the start of the infusion: 0, 0.5, 1, 1.5, 1.75, 2, 2.25, 2.5, 3, 4 and 5 h. After oral administration, samples were taken at 0, 0.5, 0.75, 1, 1.5, 2, 3 and 4 h. For both modes of administration sampling was continued usually at 6, 8, 12, 24, 32, 48, 56, 72, 80, 96 and 104 h after the start of the experiment, and from subjects L. De, T. Ho and M.O. additional blood samples were obtained at 171, 195 and 219 h. Because of frequent sampling in the first 4 h after the dose, the antecubital vein was kept open by a slow infusion (0.038 ml/ rain) of physiological saline containing heparin (0.5 ml Thromboliquine, Organon, Oss, The Netherlands, in 100 ml 0.9% NaC1), maintained by a Harvard infusion pump (Harvard, Millis, Mass., USA). At the desired times the tube was disconnected and blood was collected directly from the cannula, after discarding the first I ml. The remainder of the samples was taken by venepuncture. All blood samples were heparinized and centrifuged immediately after collection, (the importance of immediate separation of plasma from red cells has been described previously, [7]), and both fractions were frozen until assayed. Urine was collected completely for 120 h in separate (usually 6 h) portions from the hospitalized subjects, and in separate 12 h portions for at least 220 h (exact times in Table 7) from the three ambulant subjects. Urine volume and pH were measured and aliquots of each portion were frozen until assayed. The reliability of the completeness of the urine collections was determined from urinary creatinine excretion, and additionally by a careful supervision and personal motivation of each subject in the trial. Faeces were collected in 24 h portions for 120 h by the hospitalized subjects, and for 220 h by the ambulant subjects. The portions were weighed, a known volume of distilled water added and the mixture thoroughly homogenized by use of a Waring Blendor. Aliquots were frozen until assayed, or were analysed directly as follows: an aliquot of about 0.1 g was weighed into an extraction tube, 2 ml S6rensen phosphate buffer pH 7.4 containing internal standard (5 ~tg of the same compound as in reference 7) was added, the mixture sonicated for 10 rain, and after brief centrifugation the supernatant was transferred to another tube. The sample was extracted and processed further as described below for the assay of

37

chlorthalidone in plasma. As a control, calibration graphs were prepared with blank faeces from each subject, which proved that the extraction recoveries were the same as those obtained with buffer solutions. By performing the measurements in duplicate, a relative standard deviation of 4% for the determination in faeces was obtained (n = 8). Chlorthalidone concentration in plasma, urine and red cells was determined by gas chromatography with nitrogen detection, as reported previously by Fleuren and van Rossum [7]. In one subject (M. We) the urine gas chromatograms showed two, large interfering peaks, which were shown to originate from the antibiotic flucloxacillin used by this subject. Due to the high aqueous solubility of the latter compound, it was possible to remove the interfering peaks by washing the methyl isobutyl ketone extracts twice with 2 ml phosphate buffer pH 7.4, which resulted in excellent purification of these urine samples.

Pharmacokinetics Preliminary graphical analysis of plasma concentration versus time curves after intravenous and oral administration showed that the data could be described by either a two - or a three - compartment open model. Therefore, the plasma concentrations were fitted to the corresponding sum of exponentials: i.v. administration I1

C =

Z

Ai" ki.-~

( l - e - k r t) for t ~< T

(1)

and C = ~ ~- 'A i " k~.T 1 (e -k~(t-T)- e -krt) for t >/T

(2)

i=d

oral administration rl

c =

E~%'(e-k~~- e-r~~).

(3)

i=l

In these general equations C represents the plasma concentration, t is the time after the start of the infusion or after the lagtime of the oral dose, T is the infusion time, the coefficients Ai are the h3q)othetical intercepts with the ordinate for an intravenous bolus injection of the same amount of drug, k i refers to the i th disposition rate constant (reciprocal time constant), where i = 1 . . . n, K a is the rate constant of first-order absorption into the central compartment and n is the number of compartments of the model employed. After graphical analysis had yielded initial estimates, final parameter values were obtained by use

l!J luotula~3dtuoo-ooatI1 oql jo ,~l.uo!aodns oql mealsuotuzp pu~3 ~leOS pzpuedxo u~3 uo q fie lsa.tj oql ~u!anp suo!le.tluo3uo3 etUseld oql luos~ado.t slosu.t ~q~L "sl0potu luotttlaedmo3-o0.tql e ptre -ons~ '.~10A!130ds~:t'ol ~u!p.~oz3e s1!J amndtuo3 moqs sqdea~ ao~aoI pue aoddn oq& "sl3~.fqns tmtunq ietu.tou o,~1 u! ~sop snouoaealu! oI~ttTs 13:tmje q 00E "~3 aoj p~A~OIIOJ mUseld u! 0uop!Ieql.~olq3 jo s3!lOu.p-,lo3etuxeqd "l "~H C(~NC, IN HGIL PLFqSMA p.o.=,

~

o

p

.o o o . o

p

C

C

. . . .

.o

.~

.= o ~ ?

o

CONC. IN MG/L PLQSMR o

p

p

p pp

p

~,.

.~? rnc_o

F~

~-~

g

~

~.

N ru

g

o

~.

C(gNC. IN MG/L pLF~SMIq c~oa

o

.

.

.

.

.

.

.

o

o

c~ ~ o c ~

C(~NC. IN MG/L PLIqSMB

o?,o

o

i

i

i

(

? i i

p i

p

p

:

l

ppp :

o

.

:

:

:

.o ; ;

.

,

~ l~

,

,

o .

,

,= o o o ~_:

o

]i~

:

;1.

=~

,

eF~

" x

rrn- ~ z ~rn

x

ii

c o ~m .

ouop!leq~aolqD jo £1!I!qgI!eAeO!~to l n l o s q v :'[e 10 uoJnoI~I ' f "-I "H

8~

H. L, J. Fleuren et al.: Absolute Bioavailabilityof Chlorthalidone

of the FARMFIT computer program 2. A relative error of 5 % was attributed to each datapoint [weighti = 1/(0.05 Ci)2], as this was the standard deviation of the assay over the entire concentration range [7]. The FARMFIT program provides a measure of the error in estimation of a parameter by dividing the asymptotic standard deviation of the estimated parameter by the computer-estimated parameter itself (multiplication by 100 gives relative error in %). The choice of the value for the tagtime after oral doses was not included in the regression analysis of the computer program for two and three compartment models. Instead a preset value was used, obtained after several prior estimations with different lagtimes, which yielded the lowest sum of weighted squared deviations, and with negligibly small improvement upon further change in this parameter. Two criteria for each individual curve were used to decide between a two and a three compartment model: 1) visual inspection of the goodness of fit; and 2) statistical comparison of fit by the f-ratio [8]. This test indicates if the total weighted sum of squared deviations (SWSD) has been reduced sufficiently (for a certain level of significance, e.g. 2.5%) in going from a two to a three compartment model, by taking into account the degrees of freedom (df) which remain after choice of the number of parameters defining the particular model (df = the number of datapoints minus the number of parameters). The biological half-life 01/2) was calculated from the smallest rate constant (ki =/3) by t,/: = 0.693//3. The volume of the central compartment (V0, the volume of distribution at steady state (Vd,s~) and the plasma clearance (k~l) were calculated for each curve directly from the coefficients and exponentials of the appropriate exponential equation by use of the following relationships [9, 10]. D

i. v. administration V~ = 2

(4) Ai

i=l ./1 D

o

2

i=t

Vd,~=

..

(5)

(2Aifki) 2 i=l

D

keel =

(6)

~Ai/kl i=1

2 FARMFIT, a non-linear curve fitting program, in use at the Computer Centre of the University of Nijrnegen. Details available on request.

39 F-D

oral administration V1 = n

(7)

(1 - ki/Ka) " Ai i=l

F"D 2

(l-kl/Ka)Ai/ki~

i=1

vd.,., = (~

keel =

(8)

(1-ki/Ka)'Ai/ki)z

F' D

,

(9)

.ii

~'~ (1-kl/K~) " Ai/kl i=l

where D is the dose, F is the available fraction, equal to the fraction of dose absorbed if there were no firstpass effect, and the other symbols are the same as in Eqs. t, 2 and 3. As is evident from Eqs. 7, 8 and 9, only values for V1/F, Vd,~s/F and k~l/F can be obtained if F is not actually known. Because graphical plots of the eD'throcycte concentration vs. time curves revealed a relatively slow and monophasic uptake of drug into the erythrocytes, both after the intravenous and the oral doses, followed by an apparently mono-exponential decay phase, red cell concentrations after both modes of administration were fitted to a one compartment model with first-order absorption; for this Eq. 3 was used with n = I, and C here representing the erythrocyte concentration of chlorthalidone, and Eq. 7 yielded the volume of distribution. In the case of a one compartment model, the choice of lagtime was made by the non-linear least squares regression analysis of the FARMFIT program. The maximum concentration reached in red cells (Cm~), and the time at which this occurred (tm~,), was calculated routinely for each experiment from the model parameters [11]. tn addition to the separate approach to plasma and erythrocyte data, computer simulations were performed according to a non-linear model, which included the concentrations in plasma and erythrocytes simultaneously, as previously described [5]. In this way the theoretical influence of variation in dose upon the elimination half-life from plasma and erythrocytes, and upon the ratio of the areas under the curves of lower and higher doses were studied to see if non-linearity in this ratio would be apparent in the dose range encountered. Estimates of bioavailability, i.e. the fraction of

40

H . L . J . Fleuren et al.: Absolute Bioavailability of Chlorthalidone

Table 2, Fit of chlorthalidone plasma concentration versus time data after intravenous administration to two and three-component

exponential equations. Statistical analysis of the weighted sums of squared deviations by the f-test

Subject

T. Be J. v. Bo L. D e T. Ho A. K1 M. O M. We

Two-compartment model (four parameters)

Three -compartment model (six parameters)

df i

SWSDi

dfj

SWSDj

15 17 19 18 15 18 16

138.3 38.0 106.7 133.1 157.5 80.7 121.3

13 15 17 16 13 16 14

117.6 23.7 37.9 36.2 82.5 23.5 54.8

f ratio a

Critical value of f (P = 0.025)

1.14 4.52 15.4 21.4 5.90 19.5 8.50

4.97 4.76 4.62 4.69 4.97 4.69 4.86

df = remaining degrees of freedom SWSD = sum of weigthed squared deviations

.f =/swsD -SWSD SWSDj

) × r

)

~ df i - dfj

dose coming into the systemic circulation, were obtained from all three sets of data, viz. plasma, urine and erythrocyte measurements. The apparent bioavailability derived from the plasma and erythrocyte concentrations was calculated as follows [12]: AUCp.o. Oi. v. tl/2i.v. apparent bioavailability = ~AUCi. v. Dp. o. tl/2p.o. '

(10)

where the area under the curve (AUC) was obtained by use of the trapezoidal rule and extrapolation to infinity from the last datapoint, using the half-lives (tl/2) obtained from the curve-fitting procedure. For reasons described below under 'Results', the renal excretion rate plots were extrapolated to infinity from the actual analytical data by use of the observed half-life of decay of red cell concentration after oral administration for each individual subject. The ratio of the total amounts of drug excreted in urine over infinite time after the two modes of administration was calculated. The renal clearance in each experiment was calculated by dividing the amount excreted in urine during the period for which plasma concentrations had been measured, by the area under the plasma concentration curve obtained by the trapezoidal rule. Results

Plasma Concentrations Following Intravenous Administration During intravenous administration of chlorthalidone, a regular rise in the plasma concentration occurred until the zero-order infusion was terminated. At that time peak concentrations between 0.44 and 0.87 ~tg/ ml were measured. Immediately thereafter a rapid fall was observed with an apparent half-life between

ca. 12 and 25 min, which passed into a slower phase of disappearance with a mean half-life of 3.6 h, until finally the elimination phase was reached with a long tv2 between ca. 25 and 50 h (mean 36.5 h). Typical plasma concentration profiles are shown in Figure 1. All datapoints during and after the infusion were fitted to the exponential equations for twoand three-compartment models. Visual inspection of the plots revealed the superiority of curve-fitting according to a three compartment model in five out of seven subjects; in one subject extension of the number of compartments to three seemed doubtful (J. v. Bo), and in another (T. Be) no visual improvement at all was obtained. The statistical criterion gave results in good agreement with the judgementby-eye, as may be seen from Table 2: the f-ratio exceeded the critical value in the same five subjects, which indicates significant improvement in fit on use of the three compartment model for them, whereas the test was negative for the data from subjects T. Be and J. v. Bo. On the basis of these criteria it was concluded that a three compartment model was necessary to describe the time course of the plasma concentrations after intravenous administration in five subjects, while a two compartment model could be used for the two other subjects. The model parameters for the seven humans are presented in Table 3, together with the calculated values for the volume of the central compartment (V1), the volume of distribution at steady state (Vd.ss) and the plasma clearance (k~el).

Plasma Concentrations Following Oral Administration The time course of plasma concentrations after oral chlorthalidone on the whole was consistent with pre-

H. L, J. Fteuren et al.: Absolute Bioavailabilityof Chlorthalidone

41

Table 3, Open pharmacokinetic model parameters describing plasma concentration of chlorthalidone after intravenous administration to seven human subjects. (Rel. errors in parentheses) Three compartments Subject

Dose (rag) Dose/weight (mg/kg) Infusion time (h) (h -~) a (h -I) 13(h -q) Half-life (h) P (mg/1) A (mg/l) B (mg/l) V I (1) Vdss (1) kcei (I/h)

Two compartments Subject

L. De

T. Ho

A. K1

M. O

M. We

T. Be

J.v. Bo

50,0 0.69 2.00 2.18 (t9.9%) 0.217 (37.6%) 0.0139 (2.6%) 49.8 1.09 (11.8%) 0.116 (36.1%) 0.106 (4.4%) 38.0 366 5.77

51.5 0.68 2.00 2,12 (15.5%) 0.133 (33.8%) 0.0157 (3.4%) 44.2 1.28 (11.0%) 0.119 (21,7%) 0.119 (6.0%) 33,9 306 5.66

50.4 0.88 1.85 1,71 (29.0%) 0.124 (58.1%) 0.020 (24.3°,/0) 34.6 1,40 (19.5%) 0.221 (26.5%) 0.142 (31.9%) 28.6 198 5.19

61.3 0.74 2.00 1.85 (18,3%) 0,244 (30,8%) 0.0149 (2.5%) 46.7 1.02 (8.9'/o) 0.154 (33,6%) 0.121 (3.9%) 47.4 388 6.55

50.7 0.82 2.00 3.59 (42.9%) 0.249 (44,0%) 0.0291 (4.9%) 23.8 2.15 (29.1%) 0.224 (41.0%) 0.179 (9.8%) 19.8 187 6,63

51.3 0.74 1.83

75.0 0.68 2.17

1.37 (14.7%) 0.0271 (6.5%) 25.5

1.45 (9.0%) 0.0225 (3.4%) 30.8

1.24 (13.9%) 0.131 (8.3%) 37.3 279 8,97

1.32 (8.0%) 0.194 (3.7%) 49.6 317 7.87

Table 4. Pharmacokinetic parameters describing plasma concentration of chlorthalidone after oral administration to seven human subjects according to a two-compartment open model. (Rel. errors in parentheses) Subject T. Be

J.v. Bo

k De

T. Ho

A. KI

M. O

M. We

B (mg/l)

50.0 0.73 0.25 5.79 (22.3%) 0.205 (12.5%) 0.0157 (7.8%) 44.2 0.177 (8.1%) 0.0555

75.0 0.68 0.35 1.22 (45.5%) 0.75 (47.1%) 0.0144 (5.8%) 48.2 1.04 (96.0%) 0.0940

50.0 0.69 0.40 1,53 (21,6%) 0.131 (46.5%) 0,0136 (5.4%) 51.1 0.099 (31.6%) 0.0895

50.4 0.66 0.60 1.51 (11.9%) 0.343 (12.9%) 0.0161 (2.2%) 42.9 0.393 (17.5%) 0.0975

50.0 0.87 0.70 2.73 (10.1%) 0.242 (16.3%) 0.0137 (3.7%) 50.5 0.198 (11.7%) 0.103

60.9 0.73 0.40 1.54 (20.9%) 0.190 (30.0%) 0,0145 (2.9%) 47.9 0.104 (28.4%) 0.0764

50.0 0.81 0.75 1.29 (39.0%) 0.87 (26.0%) 0.0293 (2.5%) 23.7 1.10 (42.8%) 0.0981

(8.6%)

(5.4%)

(9.5%)

(3.6%)

(4.5%)

(5.1%)

(3.5%)

V~/F (1) Vass/F (1) kce,/F (l/h)

221 602 11.5

152 690 10.7

280 467 6.92

126 399 7.35

176 402 6.06

366 681 10.7

110 424 13.6

Dose (mg) Dose/weight (mg/kg) Lag time (h) Ka (h -1) c~(h -1) fi (h -1) Half-life (h) A (mg/1)

vious o b s e r v a t i o n s [3, 5]. In t h e p r e s e n t s t u d y p e a k plasma concentrations of 0.14-0.268g/ml were found approximately 1-3h after the dose. The d e c l i n e in t h e c u r v e t h e r e a f t e r in all s u b j e c t s c o n sisted o f t w o p h a s e s ; t h e first s e g m e n t h a d a n a p p a r e n t half-life of ca. 2 h, a n d t h e s e c o n d p a r t h a d a m e a n tl/2 of 44 h. L i k e t h e j u d g e m e n t - b y - e y e , t h e f - t e s t i n d i c a t e d t h a t n o i m p r o v e m e n t in t h e c o m p u t e r fit for a n y s u b j e c t was o b t a i n e d w h e n t h r e e c o m p a r t -

m e n t s w e r e u s e d i n s t e a d o f two. T h e r e f o r e , all d a t a w e r e c a l c u l a t e d a c c o r d i n g to a t w o - c o m p a r t m e n t open model, assuming first-order absorption into and first-order elimination from the central compartment. The resulting pharmacokinetic parameters are s h o w n in T a b l e 4. B y c o m p a r i s o n w i t h T a b l e 3 it can b e s e e n t h a t t h e t e r m i n a l t~,~in t h r e e s u b j e c t s (T. B e , J. v. B o , A . K1) was s u b s t a n t i a l l y l o n g e r t h a n t h a t

42

H. L. J. Fleuren et al.: Absolute Bioavailability of Chlorthalidone

J,V.BO, 75 M6 CHLORTHRLIOONE & I NTRRVENSU$L% 00RRLLY

to

"i 7:

0.9 O.S

o

as

so

7s

zSo

1 o.e

0.7

J,V,BO. 75 MG CHLSBTHflLIO~NE & INTRRVE~U~LT ~RRLLY

O,s 0.s

ec 0.~ o.3

e E d

§ .,

.... j t~,oit tl,03 0,02

0,0t c

2si

50t

)

75

1,a¢0 TIME lfl M O U ~

Fig. 2. Comparison of the time course of plasma and erythrooyte concentrations of chlorthalidone after equal intravenous and oral doses in the same human subject. The measurements in both the red cells (upper graph) and in plasma (lower graph) show faster decay of the concentrations after the intravenous dose. The graphs have been taken directly from computer plots obtained by fitting the data to the appropriate pharmacokinetic equations

observed in them in the intravenous study. The values for volume of the central compartment, total volume of distribution at steady state and plasma clearance could, of course, only be obtained in a relative sense from oral administration data, i. e. divided by the bioavaitability (F). Red Cell Concentrations after Intravenous and Oral Administration The rise in erythrocyte concentration of chlorthalidone was relatively slow compared to the initial plasma concentrations, even after the intravenous dose, as maximal concentrations were not reached until ca. 6-9 h after starting the infusion.

Upon oral administration the highest values were found between approximately 8 and 16 h after the dose, as in previous results [3, 5]. The times at which maxima occurred coincided reasonably with the end of the distribution phase of the plasma concentrations. After the peak of the curves, the decline in the red cell level for the oral dosage form was monoexponential in all seven subjects. Therefore, the oral data were fitted to a one compartment model with first-order absorption; the resulting pharmacokinetic parameters are shown in Table 5 B. After the intravenous dose of chlorthalidone, substantially faster decay than after the oral dose was observed in four subjects, three of whom had also shown a more rapid fall of their plasma concentrations in the intravenous study (T. Be, J. v. Bo, A. KI). During the ca. 100 hour-period for which blood samples had been obtained from them, no appreciable deviation from a straight line was apparent when the concentrations were plotted on a semi-logarithmic scale. A typical illustration of this phenomenon is given in Figure 2, where the upper part represents the concentration versus time curve in the erythrocytes and the lower part shows concentrations in plasma, both followed in the same subject after both dosage forms. It can be seen that the tl/2 of the oral dose did not change after the first 100-h-period, judged from the concentration at 219 h (analyzed in duplicate), which was obtained from the assay for the zero-point before the next dose (t = 0 h of the intravenous dose) 3. Similarly, the late oral datapoints from the subjects A. K1, M. We and T. Be were located within a few percent of the fitted curves based upon the first 100 h only. In the intravenous study in another subject (T. Ho), from whom blood samples had been taken up to ca. 220 h, a slightly faster decay was observed in the first half of the curve, which gradually passed over to the same decline as that estimated after the oral dose. In two other subjects in whom a prolonged sampling period was employed (L. De and M. O.), the time course of the red celt concentration after the intravenous dose was apparently mono-exponential, and it remained approximately parallel to the curve of the oral study. Examples of both kinds of decay are shown in Figure 3. In the calculation of bioavailability, differences in

3 The concentrations in the second experiment were corrected for the remainder of previous dose by subtracting an amount calculated from the zero-time concentration and the half-life actually. observed during the next dose. ThLs correction was no greater. than 4.0% of the total area under the curve for subject J. v. Bo (1.3% T. Be, 6.7% A. K1, 0.4% M. We).

H. L, 3LFleuren et al.: Absolute Bioavailability of Chlorthalidone

43

CHLORTHRLIOONE ~,0, IH3flRVENSUSLY 8 1 . 3 HG ~RRLLY EO.9 MG

CNLORTHALIOONE ~,HO. tN3~RVENOUSLY 51,5 MG ORRLLT 5 0 . ~ ~G

0.7 ¸

0.3:

0

Z5

so

75

loo

125

150

175

~0o

~2S

25O

0

Z5

So

75

1o0

12~

~50

175

~OO

2~

2SO

Fig. 3. Typical red celt concentration profiles of chlorthalidone on a semi-logarithmic scale after approximately equal intravenous and oral doses in two human subjects. A slightly faster decay in the intravenousstudy, mainly in the first part of the period, was observed in the first subject (left side), whereas almost equal half-lives for both modes of administration were found in the other subject (right side)

half-life can be handled by using a correction for them, based upon constant clearance of the drug during both periods of measurements [12]. The most straightforward way would be to consider clearance to be constant over every period of measurement, even if it were anticipated that half-lives after intravenous administration would only be temporarily smaller and would change to larger values after the actual time of analysis, as suggested by subject T. Ho (see Figure 3). Erythrocyte concentrations in all subjects were therefore forced into a one-corn-

partment model, which seemed to be permissable because of the reasonable agreement between the actual data and the fitted curves, as shown in Figures 2 and 3, even with the intravenous dose in subject T. Ho, in whom the difference between the initial and final elimination rate was too small to permit use of a more complicated model. The resulting pharmacokinetic parameters are presented in Table 5. Comparison with the results after the oral dose (given in the same Table), shows that the estimated half-lives in five subjects (T. Be,

Table 5 A. Pharmacokinetic parameters describing chlorthalidone concentration in erythrocytes after intravenous administration. The data were treated according to a one-compartment open model with first-order absorption. (ReL errors in parentheses) Subject

D~,v. (mg) hag (h) Ka (h "°I) fl (h ~1) tv2 (h) B (nag/l) Vd (1) Cm~ (rag/l) tmax (h)

T. Be

J.v. Bo

L. De

T. Ho

A. KI

M. O

M. We

51,3 0.32 (6.8%°) 0.534 (8.9%) 0.0176 (4.9%) 39.3 9.25 (4.6%) 5.73 (4.3%) 7.97 6.61

75.0 0.37 (4.8%) 0.454 (9.6%) 0.0149 (5.9%) 46.7 12.66 (5.2%) 6.13 (4.7%) 10.90 7.78

50.0 0.39 (6.7%) 0.470 (10.3%) 0.0113 (4.t%) 61.4 8.61 (4.8%) 5.95 (4.6%) 7,67 8.13

51.5 0,31 (5.8%) 0.56 (6.2%) 0.0146 (1.7%) 47.5 7,44 (2.5%) 7.11 (2.4%) 6.58 5.92

50.4 0.24 (1.5.9%) 0.445 (12.0%) 0.0165 (8.1%) 42.1 16.02 (6.2%) 3.27 (5.6%) 13.6 7.69

61.3 0.33 (5.6%) 0.387 (7.4%) 0.0124 (3.5%) 55.7 8.75 (3.9%) 7.24 (3.6%) 7.56 9.19

50.7 0.34 (5.4°/,,) 0.579 (8.5%) 0.0215 (3.3%) 32,2 15.25 (4.0%) 3.45 (3.7%) 12.9 5.91

44

H.L.J. Fleuren et al.: Absolute Bioavailability of Chlorthalidone

Table 5B. Pharmacokinetic parameters describing chlorthalidone concentration in erythrocytes after oral administration. The data were treated according to a one-compartment open model with first-order absorption. (Rel. errors in parentheses) Subject

Dp.o. (mg) tlag (h) K a (h -1) /3 (h -1) tl/2 (h) B (mg/1) V J F (1) Cmax (mg/1) tmax (h)

T. Be

J.v. Bo

L. De

T. Ho

A. K1

M. O

M. We

50.0 0.34 (2.4%) 0.385 (3.9%) 0.0155 (2.2%) 44.6 7.40 (2.1%) 7.04 (1.9%) 6.21 8.69

75.0 0.51 (18.6%) 0.313 (27.7%) 0.0118 (11.2%) 58.9 8.70 (12.8%) 8.96 11.8%) 7.36 10.88

50.0 0.46 (11.3%) 0.180 (11.5%) 0.0116 (4.9%) 60.0 7.95 (6.8%) 6.72 (5.9%) 6.16 16.3

50.4 0.64 (6.9%) 0.316 (24.0%) 0.0138 (7.6%) 50.0 6.33 (12.9%) 8.32 (11.8%) 5.26 10.3

50.0 0.82 (3.4%) 0.29 (11.1%) 0.0108 (6.8%) 64.1 8.85 (5.9%) 5.87 (5.4%) 7.50 11.8

60.9 0.40 (3.9%) 0.195 (8.9%) 0.0133 (3.9%) 52.2 5.98 (6.0%) 10.9 (5.3%) 4.59 14.8

50.0 0.89 (1.4%) 0.437 (7.6%) 0.0179 (1.9%) 38.8 7.04 (3,3%) 7.41 (3.1%) 5.89 7.62

Table 6. Estimation of the oral bioavailability of chlorthalidone by comparison of the areas under the curves after intravenous and oral doses in seven normal human subjects. Upper part (A) plasma concentrations. Lower part (B) erythrocyte concentrations Subject

T. Be

p.o.

J. v. Bo

i.v. p.o. V.

L. De

p.o. V.

T. Ho

p.o. V.

A. KI

p.o. V.

M.O.

p.o. 1. V.

M. We

p.o. i.v.

Time period (t) measured (h)

Half-life (h)

AUC plasma 0-t (h • mg/l)

AUC plasma at infinity (h. mg/1)

Bioavailability estimate

104 96 104 96 219 219 195 195 192 80 219 195 80 104

44.2 25,5 48.2 30.8 51.1 49.8 42.9 44.2 50.5 34.6 47.9 46.7 23.7 23.8

3.79 5.44 5.56 8.42 6.89 8.40 6.78 8.88 7.88 8.43 5.58 8.92 3.33 7.60

4.74 5.79 7.03 9.53 7.26 8.78 7.04 9.26 8.42 9.89 5.81 9.39 3.66 7.92

0.49 0.47 0.81 0.80 0.59 0.61 0.47 mean (±SD) 0.61±0.15

Subject

T. Be J.v. Bo L. De T. Ho A. K1 M.O. M. We

p.o. i.v. p.o. i.v. p.o. i.v. p.o. i.v. p.o. i.v. p.o. i.v. p.o. i.v.

Time period (t) measured (h)

Half-life (h)

AUC erythr, O-t (h • mg/1)

AUC erythr, at infinity (h • mg/1)

Bioavailability estimate

104 96 240 104 219 219 219 219 192 80 219 195 264 104

44.6 39.3 58.9 46.7 60.0 61.4 50.0 47.5 64.1 42.1 52.2 55.7 38.8 32.2

364 407 708 635 597 671 415 474 699 678 397 618 404 596

457 509 751 817 653 734 437 496 810 947 421 686 407 678

0.81 0.73 0.91 0.86 0.57 0.66 0.51 mean (±SD) 0.72±0.15

H. L. J. Fleuren et aI.: Absolute Bioavailability of Chlorthatidone

J. v. Bo, T. Ho, A. K1, M. We) were shorter than in the intravenous study; the relatively small magnitude of the difference in two subjects (T. Be, T. Ho) was the overall result of a small but discernible change in elimination rate during the time of analysis, see e. g. Figure 3. This table also reveals that the individual values for apparent lagtime (t~g) were smaller and those for the absorption constant (K 0 were larger in the intravenous study, corresponding to the earlier times at which the maxima were reached.

Ratio of the Areas Under the Plasma and Red Cell Concentration Curves after Oral Versus Intravenous Doses of Chlorthalidone Comparison of the oral and intravenous plasma concentrations yielded a mean apparent bioavailability of 0.61 (individual values in Table 6A), whereas a mean apparent ratio of 0.72 was found from the erythrocyte concentration curves (individual values in Table 6B).

45 urinary excretion rate (mg/hr log scale) lSubject DOSe :

75 mg CHLORTHALIDONE olr~lly

Y2 = 4 9 hrs

Calculations performed on several sets of concentration-time curves showed that bioavailabifity assessments based upon red cell concentrations would yield values for the oral dose about 10% too high. As an example, when simulations with the model-fitting parameters given in a previous publication [5] were carried out (with F -- 0.25 arbitrarily, the dose thus being 50 rag), lowering the dose to three-quarters did not result in the area under the curve ratio of 0.75 expected in case of linear kinetics, but instead the result was 0.83 when calculated according to Eq. 10, while the tl/~ (32-100 h) changed only from 70 to 73 h. This non-proportional increase in area under the curve was found to be due to the strong, capacitylimited binding of chlorthalidone to erythrocytes at higher concentrations. In the converse case, plasma concentration measurements yielded smaller than proportional areas after the lower dose, but the difference between the known bioavailability (0.75) and the calculated value (0.74) was much less than that found with the erythrocyte concentrations. This was attributed to the compensatory effect of first-order elimination from the central compartment to which the plasma volume belongs. The tv~ of the plasma concentrations changed only from 42 to 44 h when the dose was reduced.

Urinary Excretion of Chlorthalidone After Intravenous and Oral Administration The urinary excretion rate versus time plots ran parallel to the decay in plasma concentrations, as long as

3

0.1-

0.05-

,,

0.01-t 0

50

100

urine f lo w ( m l / h r )

-

200

240

-

1°°1 F1 F1 urine pH

6 .

0

Compuwr Simulations of a Non-Linear Model of Chlorthafidone

: J, v. Bo,

.

.

50

.

.

t

100

i

L

,

[

200 240 time (hours)

Fig, 4. Urinary excretion rate of chlorthalidone versus time. To facilitate comparison with the plasma concentration curve of the same subject (tv2 = 48.2 h) a regression line has been drawn based upon the average 24 h excretion rate during 24-120 hours, indicated by open circles. The half-life calculated from the last point of this period and the later datapoint at 227 h 01/2 ca. 57 h) approximates to the longer tl/~ of erythrocyte concentrations after oral administration (58.9 h in this subject). In the lower graphs the urine flow and pH of the urine fractions are shown

the latter could be accurately analyzed (down to ca. 5 ng/ml using 2 mt of plasma). After that time there was a tendency for all graphs to conform gradually to the tv: of the erythrocyte concentrations observed after oral administration. Typical examples are given in the Figures 4 and 5, where the urinary excretion in two subjects was followed, in one for 120 h and from 215-239 h, and in the other for 220 h and from 292-335 h after the dose. The excretion rate during the later time in every case exceeded the value indicated by the regression line of the first period. The total amount of unchanged drug excreted at infinite time was therefore calculated by extrapolation with use of the oral half-life value after the actual period of analysis. This was preferred firstly because very precise half-life values can be obtained from renal excretion data only if the urine samples have been collected very frequently and considerable fluctuations in renal clearance within and between days do not occur. These requirements were not met in the present study, because urine was collected for 6 or 12-h-periods and some subjects showed considerable variation in excretion rate, often coinciding with changes in urine flow, as shown in Figure 4. Secondly, according to the non-linear model of chlor-

H. L. J. Fleuren et al.: Absolute Bioavailability of Chlorthalidone

46 urinary excretion rate ( rnglhr; log scale) 2-

1

Fig. 5. Urinary excretion rate of chlorthalidone versus time after a single intravenous dose of 51.5 nag in man, The excretion rate parallels the observed decay of plasma concentrations in the subject (tl/2 = 44.2 h), shown by regression of the averaged daily excretion rate values from 24-220 h (open circles). The half-life calculated from the later period of assay (292-335 h) and the last point of the first period (t,/2 ca. 52 h) appears to equal the tv~ of the red cell concentration after oral administration (tl/2 = 50 h in this subject). The total amount of drug excreted unchanged at infinite time was 72 percent of the intravenous dose

Subject: T, Ho. 24 y r , 76 kg Dose: 51,5 mg CHLORTHALIDONE intravenously

o

0.1 = 4 4 hrs

0.0

"',,..

0.003] 0.01

• ~"', , t . . . . 0 50

t .... 100

Table 8. Excretion of chlorthalidone in faeces after intravenous (i, v.) and oral (p, o.) administration of equal (usually 50rag) doses to the same human subjects

5

'I , , , , i . . . . " ' I 150 200

, , ~ , ,., 300

i 350

Subject

urine flow ( ml/hr)"

300- h

'°°l"

2 6 0 -I~. - -

urine

T. Be J. v. Bo L. De T. Ho M. O M. We

--'7__

pH.

s

~

' ' '16d

'1do' ' ' ~ 6 6

~6d time

Time period measured

Cumulative amount (% of dose)

(h)

p.o.

i.v.

120 120 220 220 220 120

17,5 21.0 24.2 20.9 31.2 n.d.

1.70 1,27 4.14 3.86 3.69 8.51

'36o (hours)

n.d. = not determined

Table 7. Urinary excretion of chlorthalidone after intravenous and oral administration of equal (usually 50 mg) doses to the same seven human subjects. Calculation of renal clearance and estimation of oral bioavailability Cumulative excretion of chlorthalidone Subject and mode of administration

Time period measured (h)

During assay amount % of (mg) dose

At infinite timea amount % of (rag) dose

Bioavailability estimate

RenaP clearance (l/h)

Urine flow (ml/h)

T. Be

119 119 119 119 220 216 220 219 95 95 296 215 119 119

22.96 30.15 26,83 44.6 22,15 31.6 24.81 35.89 15.94 32.23 26.39 37.03 12.3 24,62

25.91 32.85 32.34 48.16 23,82 33.65 25.6 37.1 21.32 39.2 26.92 38.43 13.25 25,40

0.81

5.86 5.29 4.63 4.89 3.10 3.70 3.65 3.99 2.61 3.67 4.66 4.06 3,39 3.20

73,3 123.0 55.7 73.2 65.7 56.0 66.7 77.1 65.1 81,2 83.7 66,2 103.0 97.0

p,o. l.V.

J. v. Bo

p.o. LV,

L. De

p,o, V.

T. Ho

p.o, 1, V.

A. KI

p.o. V,

M.O.

p.o. LV.

M. We

p,o, I.V,

45.9 58.8 35.8 59,5 44.3 63.2 49,2 69.7 31.9 64.0 43.3 60.4 24,6 48,6

51.8 64.0 43.1 64.2 47.6 67.3 50.9 72.0 42.3 77.7 44.2 62,7 26.5 50.1

0.67 0.71 0.71 0.54 0.71 0.53 mean (+_SD) 0.67 (+0.10)

a extrapolation to infinity was performed by use of the half-life of the erythrocyte concentration after oral administration. For further explanation see text. b very similar renal clearance values were obtained when only the first 48-h-period was used.

H. L. J. Fleuren et al.: Absolute Bioavailabitity of Chlorthalidone

thalidone disposition [5], the plasma concentration decay and therefore, under the assumption of firstorder renal elimination, the urinary" excretion rate too, should ultimately become parallel to the red celt concentration decay after a period of time long enough to reach appropriately low concentrations. The urinary excretion data of all subjects after both dosage forms of chlorthalidone have been summarized in Table 7. The actual cumulative amounts excreted during the sampling period are given, as well as the total calculated amounts, obtained by extrapolation to infinite time. After intravenous administration of chlorthalidone, a mean value of 65.4% of the dose was excreted as unchanged drug at infinite time, and a mean value of 43.8% was found after the oral dose. The mean bioavailability of the oral dose, calculated as the ratio of the total amounts excreted, was 67%. Renal clearance of chlorthalidone ranged from 43.5-97.7ml/min (Table 7). Except for a rather large difference in one subject (A. K1), the values in the entire group of subjects were not systematically higher or lower after either of the two dosage forms. Faecal Excretion

Measurable concentrations of chlorthalidone in human faeces (greater than ca. 20 ng/sample) were present during a period varying from 4-8 days after the intravenous or oral dose. The peaks in excretion mostly occured during the second day after intravenous administration, and from 1-3 days after the oral dose, and the actual excretion rate of each individual depended upon the regularity" of the defaecation pattern; individual data are shown in Table 8. During the intravenous study, total amounts of 1.3-8.5% of the dose were excreted, whereas after oral administration 17.5-31% of the dose was found in faeces.

Discussion and Conclusions As far as we know, no reports of the intravenous administration of this drug to humans have appeared in the literature. The drug was dissolved in physiological saline because no untoward effect was expected from this dosage form when administered at a sufficiently slow rate. It was possible, therefore, to employ a 2-h-period for infusion of the fluid volume required because of the limited solubility of chlorthalidone in neutral aqueous solution [1]. Upon questioning, the subjects under study stated that no effects, except upon urine production, were experienced by them during or after the infusion. With respect to the plasma concentration of

47

chlorthalidone after intravenous administration, perfect agreement was noticed between the results of the f-test and visual inspection of the goodness of fit. Both methods indicated the superiority of the three compartment model in five out of seven subjects, in whom a serious underestimate of the terminal halflife would have occurred on use of the two compartment model. Relatively large errors were found for the parameter estimates desc~bing the first and second phase of distribution after the intravenous dose (Table 3), due to the limited number of datapoints in these segments, despite a reasonable visual fit, see e. g. Figure 1. In contrast, errors in the oral absorption rate constants reflected deviation in some subjects from a purely first-order absorption process, which, however, remained the most effective approach to all the data. Good accuracy was obtained for estimates of the terminal elimination rate constants and also, therefore, of the terminal half-lives. Comparison of Tables 3 with 4 and 5 A with 5B showed that the reduction in the half-life after intravenous administration as compared to the oral dose was most pronounced in three of the seven subjects (T. Be, J. v. Bo, A. K1.). Simulation with a nonlinear model indicated that differences of such magnitude should be not expected to follow from less tight binding of the drug to red cells at the higher concentrations initially reached with the intravenous dose. Moreover, the concentration range in which the differences in half-life became evident was practically" the same for both dosage forms (see e. g. Fig. 2). It is reasonable, therefore, to regard the shorter half-lives of plasma and erythrocyte concentrations as due to enhanced clearance of chlorthalidone after intravenous administration, assuming that the volume of distribution remained the same in both trials. Inspection of Figure 4 suggested that the faster elimination could have been caused by the greater urine flow generally observed after the intravenous doses (see Table 7). Surprisingly, however, no proportional increase in time-averaged renal clearance was evident in the subjects in whom a shorter half-life was found after the intravenous dose, except in one subject (Table 7). The implication is that this difference comes about mainly because of increases in nonrenal clearance. For proof of greater plasma clearance after intravenous than after oral medication, the exact values of bioavailability F must be known. In turn F can be determhled reliably from urinary excretion data only if the balance between renal and nonrenal clearance does not change between the oral and intravenous trials. This type of vicious circle presents serious limitation in exact interpretation of data from some subjects. The disagreement for two subjects between bioavailability calculations based upon uri-

48 nary excretion data and upon plasma concentrations (Tables 6 A and 7) could be explained by increased non-renal clearance after intravenous administration, such that a greater proportion of drug was eliminated by some other route than excretion into the urine, consequently the oral bioavailability calculated from the urinary excretion data in these subjects (T. Be, J. v. Bo) was too high. On the other hand it seemed unwise to rely only upon plasma data, because the semi-logarithmic plots of the elimination phase were reasonably but not perfectly linear (see e.g. Fig.2), which indicates decreasing clearance already during the measuring period, such that the model used should be considered only approximate in these cases. In the absence of a more detailed explanation for the differences in tv2 than that given above, it was decided not to prefer either the urinary excretion or the plasma concentration data for assessment of the bioavailability of chlorthalidone; when weighing equally both sets of data, the mean absolute bioavailability of oral doses was F -- 0.64. In contrast, bioavailability estimate based upon red cell concentration data yielded excessively high values, as judged from computer simtflations. This was further supported by the observation that the sum of the apparently available dose and the amount remaining in the gastro-intestinal tract after oral administration (Tables 6B and 8) exceeded 100% in two subjects (L. De and T. Ho). In principle, decreased oral bioavailability may be caused by limited absorption, by a so-called "firstpass effect" in the liver, or by a combination of both processes. Judged from the total amount of drug recovered from the faeces after an oral dose, limited absorption must be the predominant factor here, as only minor amounts were found in the faeces after intravenous administration, which rules out the possibility of considerable transport back into the gut lumen under normal conditions. However, comparison of the amount not absorbed from the gastrointestinal tract (by subtraction of the amount in faeces after an i.v. dose from that found there after an oral dose; see Table 8) with the calculated bioavailability, which is the fraction of dose entering the systemic circulation, suggests that there is a certain degree of first-pass elimination, about 10-15% of the dose. The differences in half-life between the intravenous and oral trials in general were most pronounced for hospitalized subjects (see Tables 3, 4 and 5). Little is known of the effect of continuous bedrest per se on the pharmacokinetic behaviour of drugs. A report by Levy [13] showed that bedrest alone had a significant influence on the disposition of benzylpenicil-

H. L. J. Fleuren et al.: AbsoluteBioavailabilityof Chlorthalidone lin, although the underlying mechanism was not understood. The present data do at least suggest that subjects resting in bed were more susceptible to an increase in non-renal elimination rate. This could be due to the higher concentration of chlorthalidone initially reached after intravenous administration, which would make available a higher free drug concentration that would somehow affect its own elimination. Only in one subject (A. KI) was the shorter %2 after the intravenous dose concomitant with an increase in renal clearance (Table 7), i.e. the change was not due to a non-renal process. It may be speculated that the underlying hypothyroidism, a serious metabolic disturbance [14], made this subject unable to react to a sudden increase in drug concentration by a nonrenal mechanism, e.g. a metabolic process, and therefore a secondary change in renal clearance became apparent. The volume of distribution of the central compartment (V1), which was estimated from the intravenous study to have a mean value of 36.41 (+_ 10.4 SD), or 0.48 + 0.081/kg body wt. (detailed data in Table 3), was still larger than the plasma volume, which is indicative of rapid equilibration with wetlperfused tissue in relation to the infusion rate. Very rapid decline in plasma concentrations was evident shortly after the dose in the intravenous study (see Table 3 for individual data). This segment could possibly be explained by marked hepatic uptake of chlorthalidone, which has been observed immediately after intravenous doses of this drug in the rat [15]. After oral administration no such phase of rapid decay was seen. This is not surprising, because the decay showed a mean halMife of ca. 20 min, which is of the same order as the rate of absorption of the drug after oral administration, for which a mean halflife of ca. 19 min was found (see Table 4 for individual data). As a result, much larger apparent central volumes of distribution were found in the oral study, even if corrected for the bioavaitabitity F. Thus, the V1/F values found in the oral study cannot be directly compared with those for V~ from the intravenous trial, because different models were necessary to describe the time course of the plasma concentrations after each dosage form. The mean half-life in the seven subjects for the terminal decay phase in the intravenous and oral studies, respectively, was 36,5 h (+_ 10.5 SD) and 44.1 h (+_ 9.5 SD) for the plasma concentrations, and 46.4 h (_+ 9.9 SD), respectively 52.7 h (+_ 9.0 SD) for the erythrocyte concentrations. A pharmacokinetic model, including non-linear binding of chlorthalidone to red blood cells, which was able to account in detail for the observed differences in halflife between plasma and erythrocyte decay, has been

H. L. J. Fleuren et al.: AbsoluteBioavailabilityof Chlorthalidone extensively described in a previous publication [5]. The magnitude of the differences in the present study requires comment, as it was somewhat smaller than those previously reported: an average terminal tv2 of 40 h for the decay of plasma concentration and 60 h of erythrocyte concentrations was found after single 100 or 200 mg oral doses in ten human subjects [5]. A lower dose, usually 50 mg, was administered in the present experiments, so less time was required to reach the low concentrations at which strong binding of drug to components of the erythrocytes becomes the predominant factor governing the slow transport rate to plasma. There is good evidence that carbonic anhydrase has an important role in this strong binding [2, 16]. An apparent discrepancy should be noticed in this context between recent pharmacokinetic reports on chlorthalidone by different authors. In summary, our results showed parallel decay of plasma concentration and urinary excretion rate, which had shorter half-lives than the erythrocyte concentrations for at least 100-200 h after the dose (3, 5; this report). In contrast, others have estimated parallel decay of plasma concentration, urinary excretion rate and red blood cell4 concentration [6], and have stated that plasma concentration "mirrored" red cell concentration [4], or they have reported variable half-lives [2]. One explanation for this divergence may be that methodological precautions must be taken for the precise assay of plasma concentrations of chlorthalidone, otherwise much too low values are found, especially during the first 10-24 h after the dose [7]. If these conditions are not fulfilled [2, 4], or alternatively, if the concentrations during the first 24 h are not measured [6], the higher initial plasma concentrations, which contribute much to establishment of a faster decay phase, are ignored and a wrong impression of the half-life is obtained. Additionally, it may be argued that exact halflives obtained from urinary excretion data are difficult to assess if only 24 h specimens have been collected [6], particularly as there are considerable day to day fluctuations in urinary clearance, as can be seen by inspection of, for example, Figure 5. Nevertheless, Riess and co-workers reported a faster half-life from urinary excretion data than from whole blood levels at the higher concentrations reached after chronic administration, although they did not comment on this observation [6]. Analogous pharmacokinetics, viz. slower decay of erythrocyte con4 Actuallymeasured in whole blood, but due to the much higher concentration in red cells than in plasma, complete identity between the terminal decayof wholebloodand red cell concentrations can be assumed.

49 centrations than of those in plasma have been reported for acetazolamide, a drug which is also strongly bound to carbonic anhydrase in red cells [17], both in humans [18] and in dogs [17]. Although the full clinical significance of these phenomena remains to be established, it is conceivable that more rapid elimination at higher concentrations limits the usefulness of higher doses of chlorthalidone. It has been suggested in the literature that the lengthy biological half-life of chlorthalidone could be maintained by prolonged absorption of the drug from the gastro-intestinal tract [19]. Evidence against this possibility is afforded by the present faecal excretion data after oral administration, which showed that 80-90% of the non-absorbed fraction was excreted in the first portion of faeces in most subjects, which was commonly produced on the first or second day. It is reasonable to assume that, if further considerable amounts of non-absorbed drug were present in the gut lumen at a later time, this would also have been reflected in the faecal excretion during that period. This implies that absorption had ceased well before the end of the period in which the long tv2 values were apparent, i. e. up to ten days and later. Another contribution to a long half-life could theoretically have come from entero-hepatic circulation, as chlorthalidone is known to be excreted in the bile in rats [15]. Preliminary experiments in man in our laboratory, which revealed the biliary excretion of only a few percent of unchanged drug (Fleuren et al., in preparation), appear to exclude the quantitive importance of this possibility. In conclusion, it must be stated that the long biological half-life of chlorthalidone is caused to an important degree by the strong binding of the drug to red cells [2-6], and by the relatively slow exchange rate with this tissue (see Figs. 2 and 3), which together provide very low concentrations of unbound drug available for elimination. As the total urinary excretion of unchanged drug after intravenous administration amounted to only 65.4 + 8.6% of the dose (mean _+ SD), and as only a few percent was found in the faeces (Table 8), while an approximately equal cumulative amount in urine plus faeces was recovered after oral administration (Tables 7 and 8), strong evidence exists for another important, presumably" metabolic, route of elimination of chlorthalidone in man. Essentially the same information can be obtained by comparison of plasma clearances (Table 3) with the corresponding renal clearances (Table 7), the latter being on an average ca. 38% lower than the total clearance. In agreement with metabolic degradation of chlorthalidone is the observation that, after oral adminis-

50

tration of C14-chtorthalidone to humans, about 10% of the urinary radioactivity behaved differently on thin-layer chromatography from that of the original compound [2]. Metabolism of C14-chlorthalidone in the rat has been demonstrated by Beisenherz et at. [15], the largest part of the metabolites being excreted into the bile, A minor part of the radioactivity was tentatively attributed to the hydrolysis product of chlorthalidone, 3-(4-chloro-3-sulphamoyt benzoyl)benzoic acid. This compound was not found in human urine and bile in preliminary experiments in our laboratory (Fleuren et at., in preparation). Thus, the nature of the metabolic pattern of chlorthalidone in man has yet to be identified,

Acknowledgements. We wish to thank Mrs. E.L. Klok-Huyser for excellent work with the FARMFIT program, Miss T. C. M. Marcelis for assistance in the assay of chlorthalidone, Misses J. C. L. Benneker and C. E. M. van Heesch for preparation of the intravenous doses, Misses J.M.T. Dekkers and E . T . M . Berendschot for careful attendance on the resting subjects and Mrs. M. P. M. Klumpkens-Janssen for typing the manuscript. The work was partly supported by a grant from the Netherlands Foundation for Medical Research (FUNGO).

References 1. Stenger, E. G., Wirz, H , Ptdver, R,: Hygroton (G 33182), ein neues Salidiuretieum mit protrahierter Wirkung. Schweiz. Ivied. Wochensehr. 89, 1126-1130 (1959) 2. Beermann, B., Hellstrrm, K., Lindstrrm, B., Rosrn, A.: Binding-site interaction of chlorthalidone and acetazolamide, two drugs transported by red blood cells. Clin. Pharmacol. Ther. 17, 424--432 (1975) 3. Flenren, H.L.J., van Rossum, J.M.: Pharmacokinetics of chlorthalidone in man. Pharm. Weekbl. 110, 1262-1264 (1975) 4. Collste, P, Garle, M., Rawlins, M.D., Sjrqvist, F.: Interindividual differences in chtorthalidone concentration in plasma and red cells of man after single and multiple doses. Europ. J. clin. Pharmacol. 9, 319-325 (1976) 5. Fleuren, H. L. J., van Rossum, J, M,: Nonlinear relationship between plasma and red blood cell pharmacokinetics of chlorthalidone in man. J. Pharrnacokin. Biopharm. 5, 359-375 (1977)

H. L. J. Fleuren et al.: Absolute Bioavaitability of Chlorthalidone 6. Riess, W., Dubach, U.C., Burckhardt, D., Theobald, W., Vuillard, P., Zimmerli, M.: Pharmacokinetic studies with chlorthalidone (Hygroton®) in man. Europ. J. clin. Pharmacol. 12, 375-382 (1977) 7. Fleuren, H, L, J., van Rossum, J. M.: Determination of chlorthalidone in plasma, urine and red blood cells by' gaschromatography with nitrogen detection, J. Chromat, 152, 41-54 (1978) 8. Boxenbaum, H. G., Riegelman, S., Elashoff, R. IV[.:Statistical estimations in pharmacokinetics. J. Pharmacokin. Biopharm. 2, 123-148 (1974) 9. Wagner, J.G.: Linear pharmacokinetic equations allowing direct calculation of many needed pharmacokinetic parameters from the coefficients and exponents of polyexponential equations which have been fitted to the data. J. Pharmac0kin. Biopharm. 4, 443--467 (1976) 10. van Rossum, J. M.: Significance of pharmacokinetics for drug design and the planningof dosage regimens. In: Drug Design, Vol. I (E. J. Arifins (ed0, pp. 469-521. New York, London: Academic Press 1971 11. Gibaldi, M., Pettier, D.: Pharmacokinetics, pp. 37-38. New York: Marcel Dekker 1975 t2. Gibaldi, M., Pettier, D.: Phannacokinetics, pp. 145-149. New York: Marcel Dekker 1975 13. Levy, G.: Effect of bed rest on distribution and elimination of drugs. J. Pharm. Sci. 56, 928-929 (1967) 14. Werner, S. C., lngbar, S.H.: Tile Thyroid, 3rd ed. New York: Harper and Row 1971 15. Beisenherz, G., Koss, E W., Klatt, L., Binder, B.: Distribution of radioactivity in the tissues and excretory products of rats and rabbits following administration of C14-Hygroton®, Arch. int. Pharmacodyn. 161, 76-93 (1966) 16. Dieterle, W., Wagner, J., Faigle, J.W.: Binding of chlorthatidone (Hygroton®) to blood components in man. Europ. J. clin. Pharmacol, 10, 37-42 (1976) 17. Maren, T. H.: The binding of inhibitors to carbonic anhydrase in vivo: drugs as markers for enzyme. Biochem. Pharmacol. 9, 39-48 (1962) 18. Wallace, S.M., Shah, V.P., Riegetman, S.: GLC analysis of acetazolamide in blood, plasma and saliva following oral administration to normal subjects. J. Pharm. Sci. 66, 527-530 (1977) 19. Pulver, R., Wirz, H., Stenger, E. G.: CVoerdas Verhalten des Diureticums Hygroton (G 33182) im Stoffwechsel. Schweiz. Med. Wochenschr. 89, 1130-1133 (1959) Received: May 2, 1978 in revised form: August 8, 1978 accepted: C~tober 9, 1978 Dr. Harry L. J. Fletu'en University of N~¢'negen Department of Pharmacology Geert Grooteplein 21 N 6500 HB N~megen, The Netherlands