PHARMACEUTICA

ACTA
HELWI'IAE
ELSEVIER Pharmaceutics Acta Helvetiae 70 (1995) 307-313

In vivo evaluation of an indomethacin monolithic, extended zero-order

release hard-gelatin capsule formulation based on saturated
polyglycolysed glycerides
A.-C. Vial-Bernasconi a, P. Buri a, E. Doelker ap*, E. Beyssac b, G. Duchaix b, J.-M. Aiache b
aLaboratoire de Pharmacie Gakkique, Uniuersitk de GenPue, 1211 GenZve 4, Switzerland
b Groupement de Recherche et d’&udes Galkniques, 63000 Clermont-Ferrang France

Received 18 May 1995; accepted 5 September 1995

Abstract

The sustained release properties of an indomethacin hard-gelatin capsule formulated with saturated polyglycolysed glycerides
(Gelucire@) were demonstrated in vivo. Indomethacin was selected as a model drug with very poor solubility in water and acidic media.
It is known to exhibit high intersubject variability because of enterohepatic circulation. The formulation, which in vitro showed an
erosion-controlled release, was compared in six human volunteers in the fed state by using a randomized cross-over design, to a standard
multiple-unit diffusion-controlled pellet capsule. Close action period values (time duration with plasma levels higher than 0.5 pg/ml)
were found for the test and the reference formulation (5.2 and 5.7 h). The time to reach peak t,, appeared slightly shorter for the test
preparation (1.75 h) than for the reference formulation (2.67 h), but the difference was not statistically significant because of the high
intersubject variability (non-parametric Wilcoxon matched pair test). Again, due to the small number of subjects entered in the study
(insufficient for a real bioequivalence study) equivalence could not be accepted in terms of extent and rate of absorption based on the
decision procedures involving the 90% confidence interval and the two one-sided t-tests. The mean maximum plasma concentrations
C max were 3.35 and 2.82 pg/ml for the test and the reference formulation respectively, with the corresponding values of the area under
the plasma concentration-time curve AUC amounting to 10.14 and 11.38 pg h/ml. However, a simulation on 24 subjects (3 repetitions
of the same data) would lead to bioequivalence of the two preparations. As for other corrosion-controlled forms, drug release from the
proposed Geluciree formulation was very sensitive to hydrodynamic conditions, leading to poor in vitro-in vivo correlation, when
comparison is made with a reference formulation characterized by a diffusion-controlled release. Finally, it was concluded that
erosion-controlled release formulations are especially suitable for drugs, such as indomethacin, that have low solubility in water or acidic
media. More generally, sustained release hard gelatin capsules with thermosetting excipients is very versatile and their preparation is very
straightforward.

Keywords: Indomethacin; Pharmacokinetics, zero-order; Bioavailability; Hard-gelatin capsule; Erosion-controlled release; Saturated polyglycolysed
glycerides

1. Introduction

Numerous methods exist for manufacturing hard-gelatin

capsules with sustained-release properties, especially of the
multiple-unit type (Vial-Bernasconi and Doelker, 1988).
* Corresponding author. The techniques are usually complicated and require several

0031-6865/95/$09.50 0 1995 Elsevier Science B.V. All rights reserved

SSDIOO31-6865(95)00036-4
308 A.-C. Vial-Bernasconi et al./Pharmaceutica
steps. By contrast, monolithic systems such as lipid sys-
tems are straightforward to prepare. In the molten state,
these formulations can be directly filled into hard-gelatin
capsules, bearing in mind that capsule binding may be
necessary. The principle of thermosetting formulations has
already been studied in early papers (Cuint et al., 1978;
Frangois, 1979; Walker et al., 1980), and in particular from
our investigation on the use of saturated polyglycolysed
glycerides (Eur. Ph.) commercially available as Gelucire@
(Doelker et al., 1983; Bemasconi et al., 1985; Buri et al.,
1985;Doelker et al., 1986; Buri et al., 1988; Vial-
Bemasconi et al., 1987). Gelucire@ excipients are total or
partial glycerides (triacylglycerols) and polyethylene gly-
co1 mono- and diesters of different molecular masses. The
various grades are characterized by their hydrophile-lipo-
phile-balance (HLB) value and melting point which leads
to a specific behaviour when placed in the gastrointestinal
fluids in respect of hydrodispersibility, melting and floata-
bility (Doelker et al., 1983, Doelker et al., 1986). As for
the in vitro sustained-release characteristics, we have shown
that careful selection and proportion of the Gelucire@
grades can lead to diffusion- or erosion-controlled matrix
formulations (Buri et al., 1985). Thus, a diffusion-con-
trolled system was obtained with the Gelucire@ grade
46/07, 48/09 or 62/05 containing 660 mg lithium sulfate
(square root of time kinetics), all three Gelucire grades
remaining inert in the aqueous environment at 37°C. On
the contrary, erosion was proved to control the release of
indomethacin (75 mg) when the dispersible Gelucire@
33/01 was added to the inert Gelucire@ 46/07. Since
then, several studies on Gelucire@-based capsules have
been published, most of them concerned with in vitro
release. Very few human studies using such systems have
been reported. Djimbo et al. (1984) showed some retarda-
tion in the urinary excretion of total salicylate when ad-
ministering capsules of aspirin mixed with Gelucire@
50/13. Also with aspirin Mathis and Heimendinger (1989)
observed the same pattern using salicylate salivary data
and capsules prepared with a mixture of Gelucire 48/09
and 44/14. We demonstrated, in a comparative study with
a commercially available compressed hydrophilic tablet
matrix, that fairly uniform drug plasma levels were also
observed with two different lithium sulfate capsule formu-
lation based on Gelucire@ (Vial-Bernasconi et al., 1987;
Buri et al., 1988). Dennis et al. (1988) obtained similar
results with ketoprofen formulated with Gelucire@ 50/02
and 50/13. The in vitro release profiles of all these
formulations were either not characterized or found to be
diffusion-controlled.
The purpose of this work was to demonstrate in vivo,
by measuring plasma levels, the sustained release proper-
ties of an indomethacin capsule with erosion-controlled
Acta Helvetiae 70 (1995) 307-313
release, formulated with Gelucire@ excipients. As refer-
ence, a sustained-release multiple-unit hard-gelatin capsule
product was selected.
2. Materials and methods
2.1. Materials
The sparingly soluble model drug indomethacin used in
the preparation of the capsules (lot # 32/83M-1M) was
supplied by Selectchemie (Ziirich, Switzerland), with a
volume-surface diameter of 10 pm. The Geluciree excipi-
ents (Gattefosse, Saint-Priest, France) selected were grades
46/07 (lot # 1411) and 33/01 (lot # 3413). For each
grade the first two figures characterize the melting point
and the last two the HLB value (Waginaire and Glas,
1981).
The reference formulation was Indocid@-Retard 75 (lot
# 718481T, Merck Sharp and Dohme-Chibret, Zurich,
Switzerland). Each capsules contained 25 mg of immediate
release pellets and 50 mg of pellets coated with an inert
vinyl polymer for ensuring a slow release over approxi-
mately 12 h.
2.2. Preparation of the capsules
The nominal composition of the test capsules was the
following: indomethacin 75 mg, Gelucire@ 46/07 55 mg
and Gelucire@ 33/01 102 mg. The drug was dispersed in
the Gelucire@ excipients heated at 55°C and the mixture
was manually filled into transparent size 4 Snap-Fit TM
hard-gelatin capsules (Capsugel, Basle, Switzerland) using
a syringe. The weight and content coefficients of variation
were 1.6 and 4.7%, respectively.
We verified that the formulation could be prepared
using an industrial filling machine (Zanasi Nignis 25000,
Bologna) equipped with a volumetric dosing pump.
2.3. In vitro drug release testing
The USP paddle apparatus was used in which the
capsule was held in a spiral. Two different release media
were used, because release from both formulations proved
to be highly pH dependent. Thus, the release media were
the pH 6.2 phosphate buffer of the ‘Indomethacin Ex-
tended-release Capsules’ USP monograph and the pH 7.2
phosphate buffer of the ‘Indomethacin Capsules’ mono-
graph. For sake of validity of the comparison, the same
agitation rate (75 rpm) and volume of release medium
(1000 ml) were used in the two cases. The volume larger
than those specified in the monographs was judged neces-
 A.-C. Vial-Bernasconi et al. /Pharmaceutics
sary because the ‘sink’ conditions were not fulfilled in the
pH 6.2 phosphate buffer (solubility of about 0.3 mg/ml
vs. 1.0 mg/ml at pH 7.2). Samples were taken at various
time intervals and assayed for indomethacin spectrophoto-
metrically at 320 nm. Results are means of six replicates.
2.4. In vivo experiment
Six healthy male volunteers, after giving their informed
written consent, were entered into the study. Their age
ranged from 21 to 35 years, and their body weight from 62
to 78 kg, not deviating by more than 10% from the ideal
weight for height according to the Metropolitan Life Insur-
ance Company Bulletin 1983. Each subject was required to
be free of cardiovascular, hepatic, renal, or gastrointestinal
diseases and had no history of drug or alcohol abuse. In
addition, the blood pressure, electrocardiogram and clinical
analyses were within normal ranges and subjects were
asked to abstain from taking any other drugs for 15 days
prior to completion of the study. They were also required
to refrain from consuming alcoholic or caffeine-containing
beverages from 72 h before the administration until after
the collection of the last blood sample.
Subjects were randomly assigned to receive, in a fed
state, in a crossover design either one 75 mg sustained
release test capsule or one 75 mg reference capsule, to-
gether with 150 ml water. A light continental breakfast
was taken 1 h before administration, a standardized lunch
(1000 kcal) 4 h later, then a standardized dinner (970 kcal)
10 h after. There was a l-week interval between each
treatment period.
2.5. Blood sample collection and analysis
Blood samples (10 ml) were taken at time 0 h and at
0.33, 0.67, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10, 12, 24 and 36 h.
The blood samples were centrifuged immediately and the
plasma was stored at - 20°C until analysis. Plasma (500
~1) was assayed for indomethacin using a modification of
published HPLC methods (Yi-Hung Tsai and Shun-Ichi
Naitu, 1981; Turlier, 1985).
The apparatus used was a Waters MS10 high-pressure
liquid chromatograph equipped with a WISP 710B auto-
matic injector (Millipore Corp., Milford, USA), an Applied
Biosystems 757 UV detector (Warrington, UK) and a
Shimadzu CRGA integrator (Tokyo, Japan). A Merck
LiChrospher@ 100 R-18 (5 pm) precolumn and a Merck
LiChrospher @ 100 RP-18 (5 pm) 12.5 cm column
(Darmstadt, Germany) were used. The mobile phase con-
sisted of an acetonitrite/0.03% acetic acid (48:42 v/v)
mixture. The operating temperature was ambient, and the
flow rate was 2 ml/min with a pressure of 1200 psi. The
column effluent was monitored continuously at 254 nm
Acta Heluetiae 70 (1995) 307-313 309
with a full scale deflection of 0.01 aufs. The injected
volume was 20 ~1, and the run time was 12 min.
Phenylbutazone (Aldrich Chemicals, Steinheim, Germany)
was used as internal standard.
The procedure was the following. A 500 ~1 aliquot of
plasma was pipetted into a glass-stoppered tube, along
with 50 ~1 of a 2 mg % methanolic solution of phenylbu-
tazone, 1 ml of pH 5.0 Sorensen citrate buffer and 1 ml of
distilled water. The mixture was vortexed for 10 s and
extracted using 3 ml diethyl ether by shaking the tube for
10 min. After separation, the organic phase was transferred
to a haemolysis tube and evaporated to dryness under
vacuum. The residue was dissolved in 200 ~1 of mobile
phase through vortexing and 20 ~1 of this solution were
injected. The extraction yields were 99% for indomethacin
and 95% for phenylbutazone. The respective retention
times were 7 and 9 min. The quantification limit was 0.25
pg/ml plasma and the linearity was verified up to 4
pg/ml range.
2.6. Pharmacokinetic parameters and statistical analysis
Calculation of the pharmacokinetic parameters and the
statistical analysis were performed using a Siphar pro-
gramme (Simed, Creteil, France). The following parame-
ters were estimated from the plasma concentration-time
profile:
the lag time, tlag, i.e., the time elapsed from the inges-
tion of the form to the first measurable concentration
was evaluated directly from the plasma concentration;
the time at which the concentration is maximum, t,,,,
or time to peak, and the maximal concentration, C,,,,
were also estimated directly from the plasma concentra-
tion;
the area under the plasma concentration up to the last
measured concentration, AUC(O-24), was calculated
using the trapeizoidal rule;
the area under the plasma concentration-time curve
extrapolated to time infinity, AUC, was determined by
adding to the appropriate AUC(O-24) value the quo-
tient of the estimated last plasma concentration and the
appropriate value of k,;
the terminal elimination rate constant, k,, was calcu-
lated by least squares regression from the slope of the
linear segment of the log concentration vs time profile;
. . .
the ehmmatron half-life, t,,,, was calculated from the
quotient of 0.693 and k,;
the mean residence time, MRT, was calculated as de-
scribed by Cutler (1987);
the action duration, t,, was taken as the period over
which the plasma concentration was higher than 0.5
pg/ml.
310 A.-C. Vial-Bernasconi et al. /Pharmaceutics Acta Helvetiae 70 (1995) 307-313

Bioequivalence of the two formulations was assessed by

considering the parameters t,,,, C,,, and AUC. The
parameter t,,, was tested monoparametrically using the
Wilcoxon (1947) rank-sum test for period and treatment
effects. For C,,, and AUC the 90% confidence interval of
the ratio of the parameter of the test formulation to that of
the reference formulation was calculated (Wijnand and
Timmer, 1983) after a logarithmic transformation of the
data and analysis of variance including the factors treat-
0 2 4 6 6
ment, period and subject. The two one-sided tests proce-
Time (h)
dure of Shuirman (1987) was also performed as confirma-
Fig. 1. Release profiles of indomethacin from the test ( n ) and reference
tion. For reasons, discussed below, only a paired-sample
(0) formulations in pH 7.2 phosphate buffer.
t-test was used for the other parameters (tug, AUC(O-241,
L r,/,, MRT and t,).
however, be noted that about 90% of the total amount was
released in four hours showing that the coated pellets do
3. Results and discussion not provide strong sustained release properties. By con-
trast, the test formulation released its 75 mg indomethacin
3.1. In vitro release dose more gradually, most of the profile showing zero-order
kinetics, owing to the erosion-controlled mechanism.
The release profiles of both formulations at pH 7.2 are Comparatively, when tested in a pH 6.2 phosphate
shown in Fig. 1). buffer, the reference formulation released its content much
Release patterns differed markedly, the reference for- slower (19% after 1 h, 45% after 4 h and 62% after 8 h),
mulation showing a relatively quick release as already thus demonstrating a pH-dependent behaviour. Release
observed for instance by Turlier (1985). The fact that with from the test formulation was very slow, approximately
Indocid@-Retard 75 approximately 50% of indomethacin is 30% being released after 8 h. Again the pH dependence
released in one hour is because one-third of the 75 mg observed may be explained by the limited solubility of this
dose is intended to be immediately available. It should, acid drug (pK, = 4.5) in the pH 6.2 medium specified in

Table 1
Pharmacokinetic parameters obtained after oral administration of the test formulation (T) and the reference formulation (R) to six subjects
Subject no. Form riag (h) r,,, (h) =,,,a, AUC (O-24) AUC k, (l/h) tr/z (h) MRT (h) t, (h)
( pg/mt) ( pg h/ml) ( Kg h/ml)

T 1.50 2.00 4.79 9.79 10.77 0.285 2.43 3.86 4.57

R 1.50 2.00 2.44 9.96 12.71 0.091 7.60 7.73 6.04
T 1.33 2.00 2.27 9.71 11.64 0.181 3.82 5.87 7.46
R 0.33 1.00 4.72 12.06 16.03 0.088 7.90 8.42 7.82
T 0.66 3.00 2.12 9.95 11.10 0.234 2.96 4.89 7.87
R 2.00 5.00 2.41 9.51 10.50 0.263 2.63 5.53 5.91
T 1.00 1.00 3.45 8.79 10.10 a - - 4.31
R 1.50 2.50 1.94 7.95 9.37 0.183 3.79 5.83 5.69
T 0.33 1.00 4.75 8.01 8.03 0.504 1.37 2.07 4.00
R 1.00 2.50 2.99 9.47 12.22 0.098 7.05 7.29 4.71
T 1.50 1.50 2.75 8.78 9.,20 0.610 1.13 2.74 2.93
R 1.00 3.00 2.42 7.08 7.47 0.710 0.97 3.20 4.09

Mean T 0.89 1.75 3.35 9.17 10.14 0.362 2.34 3.88 5.19
R 1.22 2.67 2.82 9.34 11.38 0.239 4.99 6.33 5.71

SD. T 0.50 0.76 1.19 0.77 1.33 0.185 1.12 1.55 2.00
R 0.50 1.33 0.98 1.72 2.97 0.240 2.90 1.89 1.28

C.V. (o/o) T 56.2 43.3 35.5 8.3 13.6 51.3 47.7 39.9 38.6
R 40.9 49.8 34.8 18.4 26.1 100.4 58.1 29.9 22.5

a Calculated using the k, value obtained for the reference formulation.

 A.-C. Vial-Bernasconi et al. /Pharmaceutics Acta Heluetiae 70 (I 995) 307-313 311

centration was sometimes observed due to enterohepatic

cycling, so that the evaluation of the elimination constant,
the elimination half life, the extrapolated area under the
curve and the mean residence time were questionable. In
one case (subject 4, test formulation) these parameters
were considerably overestimated and were not included in
Table 1 (in fact, AUC was calculated using the k, value
found for the reference formulation). More generally, it is
to be observed that the mean plasma profile found for the
0 2 4 6 8 10 12
reference formulation is very similar to that already pub-
Time (h)
lished on this commercialized form (see, e.g., Rowe and
Fig. 2. Mean plasma level profiles after oral administration of the
Carless, 1981; Fredj et al., 1982; Yeh et al., 1982; Adams
Gelucire-based test capsule formulation ( W) and the multiple-unit refer-
ence capsule (0) (n = 6, SD.).
et al., 1982; Verbesselt et al., 1983).
Two main parameters, in addition to t,,,, are consid-
ered for bioequivalence, namely C,,. and AUC. The mean
the ‘Extended-release Indomethacin Capsules’ USP mono- C max value of 3.35 pg/ml calculated for the test formula-
graph. tion, apparently higher than that of the reference formula-
tion (2.82 pg/ml), is directly linked to the shorter t,,,
3.2. In vivo experiment value. However analysis of variance showed that this
difference was again not statistically significant ( p > 0.50).
Fig. 2 shows the mean plasma profiles obtained with Note that slightly lower C,,, values (1.16 to 2.50 pg/ml)
the test and the reference capsules administered to six have been reported in the literature for the same reference
volunteers, and Table 1 summarizes the corresponding formulation. One should, however, bear in mind that the
pharmacokinetic parameters. latter values were found with subjects in the fasting state
After oral administration of both formulations indo- and that according to various authors, concomitant inges-
methacin was detectable after a lag time of about one hour, tion of food delays and reduces the peak concentration.
which can be ascribed to both the delay for the opening of This holds true for both immediate-release formulations
the capsule and the insolubility of indomethacin in gastric (Wallush et al., 1978; Emoni et al., 1976) and for sus-
juice (about 0.01 mg/ml). Actually, the two values ob- tained-release tablets (Davis et al., 1990). According to
served for tlag (0.89 h for the test formulation and 1.22 h ANOVA results the 11% difference in the area under the
for the reference formulation) did not differ significantly plasma concentration-time curve AUC (10.1 pg h/ml for
(paired-sample t-test) because of high variability. the test formulation vs. 11.4 pg h/ml for the reference
The two mean plasma profiles are typical of sustained formulation) was not significant either (p > 0.40).
release formulations, although the peak seems to appear In fact, the equivalence of the two formulations can
slightly earlier and to be higher for the test formulation. In only be assessed by using statistical tests taking non-equiv-
both cases intersubject variability was quite high mainly alence as the null hypothesis (Steinijans and Hanschke,
because of the enterohepatic circulation of indomethacin 1991). Both the standard 90%-confidence interval test and
(Alv&n et al., 1975; Kwan et al., 1976). the two one-sided test procedures indicate that bioequiva-
The mean time to peak for the test formulation was 1.75 lence cannot be accepted in terms of C,,, and AUC. In
h, 35% shorter than that of the reference formulation (2.67 both cases the lower limit of the 90% confidence interval
h). Note that t,,, values ranging from 1.45 to 4.25 h have is outside the bioequivalence range (0.80-1.25). This is
been published for the same marketed product, whereas confirmed by the non-significant t values in the two
times to peak for immediate-release formulations are gen- one-sided tests procedure (Table 2). Here, bioequivalence
erally around 1 h. is rejected because of insufficient data to distinguish be-
Some remarks can be made when examining the indi- tween equivalence and non-equivalence .
vidual plasma profiles (not shown here). With the refer- Because of the uncertainty in the terminal phase of the
ence formulation, the profiles were quite uniform, only plasma concentration-time profiles, no special comment
subject 2 showing an early elevated peak. With the test can be made on the k,, t,,, and MRT parameters, except
formulation, peaks were more marked and for subjects 4 that no statistical differences were shown. However, note
and 6, the peaks corresponded to the first measurable that the calculated mean t,,, values (2.34 and 4.99 h for
point. Subject 2 had a very uniform profile. the test and the reference formulation, respectively) were
In the elimination phase, some increase in plasma con- well within the 2 to 11 h range reported by Gilman et al.
312 A.-C. Vial-Bernasconi et al. /Pharmaceutics Acta Heluetiae 70 (1995) 307-313

Table 2 particularly suitable for drugs which are very poorly solu-
Summary of bioequivalence attributes
ble in water or acid media.
Param- A.M.R. a 90% C.I. b G.M.R.’ 90% C.I. b tl t2 More generally, formulating with saturated polyglycol-
eter lower limit upper limit
ysed glycerides (Gelucire@) and the one-step manufacture
c max 1.305 0.727 1.178 1.908 1.708 d 0.264 d process of these hard gelatin capsules with thermosetting
AUC 0.933 0.728 0.910 1.138 1.227 d 3.027 e excipients are claimed to offer many potentialities.
a Arithmetic mean of ratios.
b Confidence interval.
’ Geometric mean of ratios. Acknowledgements
d Not significant.
e p < 0.05.
The authors are grateful to Gattefosse S.A. (J.-P.
Lafor&t) and Capsugel S.A. (Dr E.T. Cole) for supporting
this work.
(1985). The two values observed did not differ signifi-
cantly (paired-sample t test). Finally, taking 0.5 pg/ml as
the threshold concentration for pharmacological action, the
action duration of the test formulation was 5.19 h and that References
of the reference formulation 5.71 h, the latter being quite
Adams, K.R.H., Halliday, L.D.C., Sibeon, R.G., Baber, N., Littler, T. and
close to the value reported for the same product by Fredj et
Orme, M.L.E. (1982) A clinical and pharmacokinetic study of indo-
al. (1982). methacin in standard and slow release formuiations. Br. J. Clin.
Pharm. 14, 286-289.
Alvan, G., Orme, M., Bertilsson, L., Ekstrand, R. and Palmer, L. (1975)
4. Conclusion Pharmacokinetics of indomethacin. Clin. Pharmacol. Ther. 18, 364
373.
Bernasconi, A.-C., Doelker, E. and Buri, P. (1985) Diffusion and erosion
The primary objective of the present work, which was controlled drug release from lipid matrix formulations incorporated
to demonstrate in vivo sustained release properties from an into hard gelatin capsules. Proc. lnt. Symp. Control. Rel. Bioact.
erosion-controlled Gelucire’s capsule formulation, was Mater. 12, 272-273.
achieved. The time to reach maximum concentration fol- Buri, P., Vial-Bemasconi, A.-C. and Doelker, C. (1985) Modulation de la
cinttique et de la vitesse de liberation par un choix adequat de
lowing oral administration was longer than the t,,, values
I’excipient: exemple des Gelucine. Bull. Tech. Gattefosst 78, 67-73.
reported for immediate-release formulations. Values of Buri, P., Vial-Bernasconi, A.-C. and Doelker, E. (1988) Evaluation in
times to plasma peak and of action period (time duration vivo de gelules B liberation prolongee de sulfate de lithium incorpore
above therapeutic level) of the test and reference formula- dans un systeme matriciel lipidique. Bull. Tech. Gattefosst 81, 7-10.
tion were quite close. However, bioequivalence could not Cuine, A., Mathis C., Stamm, A. and Franpis D. (1978) Mise en gelules
de solutions visqueuses de principes actifs. I. Etudes preliminaires,
be statistically demonstrated when considering the C,,,
excipients. Labo-Pharma 26, 222-227.
and AUC parameters, because of high interindividual vari- Cutler, D.J. (1987) Definition of mean residence time in pharmacokinet-
ability inherent to indomethacin. In fact, when bioequiva- its. Biopharm. Drug Dispos. 8, 87-97.
lence is to be shown, more subjects have to selected and Davis, S.S., Khosla, R., Wilson, C.G., Washington, N., Leslie, ST. and
here a simulation on 24 volunteers would indicate equiva- Malkowska, S. (1990) The gastrointestinal transit of a controlled
release formulation of indomethacin. lnt. J. Pharm. 60, 191-196.
lence (3 repetitions of the same data).
Dennis, A.B., Kellaway, I.W. and Davidson, R. (1988) Investigation of in
Overall, the erosion-controlled Gelucire@ formulation vitro ageing on in vivo release liquid filled hard gelatin capsule
exhibited sustained release properties, but as could be formulation. Proceed. Intern. Symp. Control. Rel. Bioact. Mater. 15,
expected the sustained action was not as marked as is the 390- 391.
case with diffusion-controlled formulations. Furthermore, Djimbo, M., Fell, J.T., Kaus, L. and MO& A.J. (1984) Release of drugs
formulated as hard gelatin filled into hard gelatin capsules. Part 2. In
like other erosion-controlled formulations, release from the
vitro studies. J. Pharm. Belg. 39, 43-49.
proposed Gelucire@ mixture is affected by the ambient Doelker, C., Doelker, E., Buri, P., Glas, B. and Waginaire, L. (1983)
hydrodynamic conditions as previously noted (Bernasconi Incorporation de principes actifs liquides, dtliquescents ou instables
et al., 1985; Buri et al., 1985). It is therefore not surprising dans des recipients pour gelules. In Proceedings 3rd International
that poor correlation between in vitro and vivo results was Conference on Pharmaceutical Technology, Paris, APGI, Vol. V, pp.
73-82.
noticed when comparing here an erosion-controlled release
Doelker, C., Doelker, E., Buri, P. and Waginaire, L. (1986) The incorpo-
test formulation and a diffusion-controlled release refer- ration and in vitro release profiles of liquid, deliquescent or unstable
ence formulation. It can be concluded that erosion-con- drugs with fusible excipients in hard gelatin capsules. Drug Dev. lnd.
trolled formulations based on Gelucires excipients are Pharm. 12, 1553-1565.
 A.-C. Vial-Bernasconi et al. /Pharmaceutics Acta Heluetiae 70 (1995) 307-313 313

Emoni, H.W., Paulus, H., Bluestone, R., Champion, G.N. and Pearson, C. Verbesselt, R., Tjandramaga, T.B., Mullie, A. De Schepper, P.J., Cook,
(1976) Indomethacin serum concentrations in man. Effects of dosage, T. and Derauwaux, C. (1983) Comparative multiple dose kinetics of
food and antacid. Ann. Rheum. Dis. 35, 333-338. two formulations of indomethacin. Eur. J. Clin. Parmacol. 24, 563-
Fran9ois, D. (1979) Making the hard gelatin capsule with the soft centre. 568.
Manuf. Chem. 50 (3), 37-39. Vial-Bernasconi, A.-C., Aebi, A., Doelker, E., Buri, P., Schulz, P. and
Fredj, G., Renat, B, Petit-Jean, O., Perret, G. and Astier, A. (1982) Dick, P. (1987) In Evaluation of highly dosed sustained-release hard
Cinttique comparee d’une forme a lib&ration prolongte et instantante gelatin capsule formulations based on lipid matrix systems. Proceed-
d’indombtacine. Thtrapie 37, 5.57-562. ings of the Third European Congress of Biopharmaceutics and Phar-
Gilman, A.G., Goodman, L.S., Rail, T.W. and Murad, F. (1985) Good- macokinetics, Vol. I, pp. 155-165.
man and Gilman’s. The Pharmacological Basis of Therapeutics, Perg- Vial-Bemasconi, A.-C., Doelker, E. (1988) Gelules a liberation prolongte:
amon Press, New York, 7th Ed., pp. 695-697. formes divisees et monolithiques. S.T.P. Pharma 4, 397-409.
Kwan, K.C., Breault, G.O., Umbenhauer, E.R., McMahon, F.G. and Waginaire, L. and Glas, B. (1981) GClucire@, Bulletin Technique Gatte-
Duggan, D.E. (1976) Kinetics of indomethacin absorption, elimina- fosst. 74, 7-12.
tion and enterohepatic circulation in man. J. Pharmacokin. Biopharm. Walker, S.E., Gaudey, J.A., Bedford, K. and Eaver, T. (1980) The filling
4, 255-280. of molten and thixotropic formulations into hard gelatin capsules. J.
Mathis, C. and Heimendinger, J. (1989) Etude de la brodisponibilite chez Pharm. Pharmacol. 32, 389-393.
l’homme de l’acide ac6tylsalicylique a partir de gdlules avec des Wallush W.W., Nowak, A., Leopold, G. and Netter, K.J. (1978) Compar-
excipients plteux. In Proceedings of the 5th International Conference ative bioavailability: Influence of various diets on the bioavailability
on Pharmaceutical Technology, Paris, APGI, Vol. V, pp. 501-506. of indomethacin. Int. J. Clin. Pharmacol. 16, 40-44.
Rowe, J.S. and Carless, J.E. (1981) Comparison of the in vitro dissolution Wijnand, H.P. and Timmer, CT. (1983) Minicomputer programs for
behaviour of various indomethacin formulations with their in vivo bioequivalence testing of pharmaceutical drug formulations in two-way
bioavailability. J. Pharm. Pharmacol. 33, 561-564. cross-over studies. Comp. Prog. Biomed. 17, 73-88.
Steinijans, V.W. and Hanschke, D. (1991) Statistical analysis of bioe- Wilcoxon, F. (1947) Probability tables for individual comparisons by
quivalence studies. In Blume, H., Gundert-Remy, U. and Mdller, H. ranking methods. Biometrics 3, 119-122.
(Eds.), Controlled/Modified Release Products, Wissenschaftliche Yeh, K.C., Berger, E.T., Breault, G.O. and Lei, B.W. (1982) Effect of
Verlagsgesellschaft, Stuttgart, pp. 99-113. sustained release on the pharmacokinetic profile of indomethacin in
Shuirman, D.J. (1987) A comparison of the two one-sided tests procedure man. Biopharm. Drug Dispos. 3, 219-230.
and the power approach for assessing the equivalence of average Yi-Hung Tsai and Shun-Ichi Naitu (1981) Simultaneous determination of
bioavailability. J. Pharmacokin. Biopharm. 15, 657-680. indomethacin and its metabolites in rabbit plasma by high-pressure
Turlier, P. (1985) Etude biopharmaceutique de cornprimes g matrice liquid chromatography. Int. J. Pharm. 8, 203-209.
hydrophile 1 base d’indometacine. Ph.D. thesis, University of Cler-
mont I.