- Email: [email protected]
Influence of water soluble fillers in hydroxypropylmethylcellulose matrices on in vitro and in vivo drug release
Journal of Controlled Release 81 (2002) 165–172 www.elsevier.com / locate / jconrel
Influence of water soluble fillers in hydroxypropylmethylcellulose matrices on in vitro and in vivo drug release Kazuhiro Sako a , *, Toyohiro Sawada a , Hiroshi Nakashima a , Shigeharu Yokohama a , Takashi Sonobe b a
Novel Pharmaceutical Laboratories, Yamanouchi Pharmaceutical Co., Ltd., 180 Ozumi, Yaizu-shi, Shizuoka 425 -0072, Japan b School of Pharmaceutical Sciences, University of Shizuoka, 52 -1 Yada, Shizuoka-shi, Shizuoka 422 -8526, Japan Received 31 October 2001; accepted 1 March 2002
Abstract The purpose of this study was to investigate the effect of fillers in gel-forming matrix on in vivo drug release after oral dosing. A further purpose was to predict the in vivo performance from in vitro dissolution test. Three controlled-release acetaminophen tablets containing hydroxypropylmethylcellulose (HPMC) with or without highly water soluble fillers, lactose or polyethylene glycol 6000 (PEG6000), were prepared. Water penetration into the matrix was enhanced by addition of fillers in the matrices, but the three tablets showed similar in vitro dissolution profiles, indicating that fillers in the HPMC matrices little affected the in vitro drug release. In contrast, the fillers in HPMC matrices did affect the in vivo performance in dogs. The absorption profile of HPMC matrix with PEG6000 was the fastest, followed by that with lactose and without water soluble filler, in that order. As the matrix with PEG6000 had a large amount of water and gelated a large portion of the matrix when in contact with water, the gel layer would be disintegrated by the gastrointestinal motility. It was found that dissolution of gel-forming HPMC matrices under mechanical stress by glass beads well correlated with the in vivo performance of the matrix, with little correlation by the conventional paddle method. 2002 Elsevier Science B.V. All rights reserved. Keywords: In vitro / in vivo relationship; Acetaminophen; Controlled-release; Colonic release; Mechanical destructive forces
1. Introduction In recent years, many drugs have been developed as oral controlled-release (CR) dosage forms in order to optimize the drug therapy and to improve the patient’s compliance. Because of their simplicity and *Corresponding author. Tel.: 181-54-627-5155; fax: 181-54621-0106. E-mail address: [email protected] (K. Sako).
cost effectiveness, hydrophilic gel-forming matrix tablets are extensively used for oral CR dosage forms. Hydration of polymer results in the formation of a gel layer that controls the release rate of drug [1]. In vitro drug release of water-soluble drugs is controlled by diffusion out of the gel layer, whereas release for poorly soluble drugs is solely by polymer dissolution [1]. In addition, effects of various factors, namely drug / polymer ratio [1–3] and polymer viscosity [2,4,5], on in vitro drug release have been
0168-3659 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0168-3659( 02 )00067-6
166
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
extensively investigated. However, in vitro drug release behavior is not always reflected in in vivo performance, because drug release from CR dosage forms is affected by various gastrointestinal (GI) factors. Among these, while the effect of pH on drug release has been extensively investigated, little is known about physical factors in GI motility such as agitation intensity [6] and mechanical destructive forces [7,8]. The purposes of adding fillers to the hydrophilic gel-forming matrix, in general, are for dilution, manufacturing processes improvement, adjusting drug release rate and so on. Investigating the effect of fillers in gel-forming matrices on in vivo drug release is important for designing matrix formulations. In this report, we prepared the hydroxypropylmethylcellulose (HPMC) matrices containing acetaminophen (AAP), as a model drug, with or without water soluble filler, and investigated the effect of the fillers of HPMC matrices on in vitro / in vivo drug release.
2. Material and methods
2.1. Materials Acetaminophen (AAP) was purchased from Yoshitomi Pharmaceutical Industries Ltd. (Japan). Hydroxypropylmethylcellulose 2208 (HPMC, viscosity of 1% aqueous solution 4100–5600 cps) was purchased from Shin-Etsu Chemical Co., Ltd. (Japan). Poly(ethylene glycol) used was of JP XIII (The Japanese Pharmacopoeia 13th edition) grades of Polyethylene Glycol 6000 (PEG6000), represented by the formula HOCH 2 (CH 2 O–CH 2 )nCH 2 OH, where n ranges from 165 to 210. Other reagents used were of analytical reagent grade.
2.2. Preparation of CR tablets
• HP-tablet: AAP, HPMC and PEG6000 were mixed (1:2:1) in a mortar, and the 400-mg mixtures were compressed directly using 9-mm diameter round-faced (9 mmR) punches. • HL-tablet: Lactose was used instead of PEG6000. The procedure was the same as for the preparation of the HP-tablet.
2.3. In vitro dissolution test The in vitro dissolution test of AAP from CR tablets was determined using Dissolution Apparatus No. 2 (paddle) of JP XIII. The dissolution medium was 500 ml of 2nd fluid (pH 6.8, 0.05 M H 2 KPO 4 and 0.0236 M NaOH) for JP XIII disintegration test. The amount of dissolved AAP was determined spectrophotometrically at 280 nm. After the last sampling at 12 h, the remaining matrices were broken by spatula to dissolve the matrices completely, in order to correct the 100% drug dissolution. When the mechanical stress was given to the matrices, the tablets were once removed from dissolution media at 1 h after starting the in vitro dissolution test and the tablets were transferred into a 50-ml test-tube with 60 g of glass beads (diameter, 4 mm; density, 2.5 g / cm 3 ) and 2 ml of water. The closed tube, with glass beads and the swollen matrix, was turned side ways and the mechanical stress was artificially given by shaking the tube horizontally at 320 strokes / min for 10 min with a shaker. After shaking, the whole content in the tube was returned into the dissolution medium, and the in vitro dissolution test was resumed.
2.4. Water penetration into matrices The tablet was immersed in JP XIII 2nd fluid (pH 6.8) at 37 8C. At 1, 2 and 3 h, the tablet was removed from the medium and the gel layer was peeled off carefully using a spatula, and the weight of the non-gelated residual core was measured.
The three AAP 100-mg CR tablets were prepared as follows:
2.5. Water retention measurement
• H-tablet: AAP and HPMC were mixed (1:2) in a mortar, and the 300-mg mixtures were compressed directly using 8.5-mm diameter roundfaced (8.5 mmR) punches.
The tablets at 3 h after the start of the water penetration test (2.4) were removed from the dissolution medium. After measuring the weight of a wetted tablet, the tablet was dried at 40 8C in a vacuum
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
dryer for 4 days. Then the weight of dried residue was measured. The water retention of the tablet was determined as the difference between wet and dry weights.
167
level–time curves (AUC) was calculated by the linear trapezoidal method. AAP absorption after oral dosing of CR tablets was calculated from oral dosing of solution and of CR tablets by the point-area deconvolution method [9].
2.6. In vivo absorption study 2.10. Statistical analysis Six male beagle dogs weighing 9.5–13.5 kg were fasted for 20 h before administration. They were allowed free access to water but food was withheld until the last blood sample was taken. One tablet containing AAP 100 mg was administered orally with 20 ml water. The interval between each administration was more than 1 week. Blood samples were collected from the forefoot vein at frequent intervals up to 12 h after dosing. Plasma samples were immediately separated and frozen at 220 8C until assay.
2.7. Plasma extraction An aqueous solution of 60 mg / ml 2-acetaminophenol was prepared for use as an internal standard. The internal standard (0.1 ml) solution and 5 ml of ethyl acetate were sequentially added to 0.5 ml of plasma in a test-tube. The tube was shaken for 10 min, and centrifuged for 5 min at 2000 rpm. The upper organic layer was transferred to a clean testtube, then evaporated. The dried residue was redissolved with 0.1 ml of mobile phase for HPLC assay.
2.8. HPLC assay of AAP AAP in plasma was determined by HPLC with UV detection according to a previously reported procedure [8]. In brief, separation on an octadecylsilane column (Nucleosil, 150 mm length34.6 mm diameter, 5 mm) was achieved at ambient temperature at a flow rate of 1 ml / min. The mobile phase contained water / acetonitrile / methanol (88:6:6). UV detection was at 254 nm.
2.9. Pharmacokinetic analysis Maximum plasma level (Cmax ) and time to maximum plasma level (T max ) were determined according to the standard procedure. The area under the plasma
Differences in each pharmacokinetic parameter were statistically evaluated by the paired t-test.
3. Results and discussion
3.1. In vitro characteristics of CR tablets The gel-forming tablets, composed of AAP and HPMC with or without water soluble filler, were used in this study. In general, pH-independent drug release is preferable for oral controlled-release (CR) formulations, so as not to be affected largely for their drug release in the GI tracts by intra- and intersubject variations of both gastric pH and GI transit time. In addition, as gastric pH in the dog shows large individual variation [10], physical factors of the GI must be investigated in pH-independent formulations. We selected AAP (pKa 9.5) as a model drug because its solubility in water is almost constant (approximately 15 mg / ml) over a physiological range of pH (1–7). Hydrophilic polymer (HPMC) and filler (PEG6000 and lactose) are also pH-independent materials. Thus, drug release from the tablets would not be affected by pH of the media. PEG6000 was selected as a very water soluble filler (1 g dissolves in 1 ml water), with lactose as a moderately water soluble one (1 g dissolves in 8 ml water). As shown in Fig. 1a, AAP dissolution profiles from H-tablet, HP-tablet and HL tablet are similar to each other for the paddle rotation speeds of 25 rpm. The drug dissolution from the three matrices demonstrated a good relation (r.0.999) to the square root of time, suggesting that the passage of drug, via diffusion, through gel layer controls the drug dissolution [2,3,5]. Even in a high agitating condition, paddle speed at 200 rpm, the three tablets demonstrated equivalent dissolution profiles, as shown in Fig. 1b: the maximum difference was 11.3% between H-tablet and HL-tablet at 12 h, and the difference at
168
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
Fig. 1. Comparison of AAP dissolution from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) using the paddle method at a rotation speed at 50 rpm (a) and 200 rpm (b) in JP XIII 2nd fluid (pH 6.8). Each result shows mean6S.D. (n53) and error bars are in the symbols.
3 h was only 4.5%. Ford et al. [3] reported that drug release rate was accelerated when lactose content increased in HPMC matrices with constant tablet mass. In this study, the amount of gel-forming polymer, HPMC, dominates the drug release and that of fillers affected it only slightly. As shown in Fig. 2, when the matrices were in
Fig. 2. Change in the portion of dried residual core after immersing in JP XIII 2nd fluid (pH 6.8). Each result shows mean6S.D. (n53).
contact with water, the portion of non-gelated residual core decreased with progress of time, the fastest in HP-tablets, followed by HL-tablet and H-tablet, in that order. Water retentions at 3 h after immersing the dissolution media were 598692, 486672 and 351658 mg / tablet in HP-tablet, HLtablet and H-tablet, respectively, indicating that highly water soluble filler in the matrices would stimulate the water penetration into the inner parts of the matrix. As a result, the dried residual part which contained the undissolved drug was larger in H-tablet (60.562.9%) than in HP-tablet (17.360.4%) after 3 h immersing. This would also affect the difference of wet strength of the matrices. The mechanical strength of HP-tablet in wet condition with a thick gel layer and a small dried residual core could be weaker than that of the H-tablet with thin gel layer and large dried residual core. We previously reported [8] that CR tablet was subjected to mechanical destructive force in the GI tract by the GI motility. We also reported that high agitation condition of in vitro dissolution test, such as paddle speed at 200 rpm, reflected the in vivo performance in dogs of low viscous HPMC (50 cps, 2% aqueous solution) matrix which did not have sufficient wet strength against GI motility. However, the difference among the three matrices demonstrated equivalence in the in
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
169
Fig. 3. Plasma levels of AAP after oral dosing of H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) to dogs. Each result shows mean6S.E. in six dogs.
Fig. 4. In vivo absorption of AAP from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) after oral dosing to dogs.
vitro dissolution profiles even in the high agitation condition of paddle speed at 200 rpm (Fig. 1b). The results demonstrate that the gel layer of high viscous HPMC could resist the high agitation by the paddle rotation, even though the matrix contained the watersoluble filler. As a result, the three tablets showed similar drug release profiles in vitro, even in high agitating conditions.
dosing with the three tablets. Fig. 4 shows AAP absorption profiles of the three tablets calculated by deconvolution. Nevertheless similar in vitro dissolution profiles were observed for the three tablets, the in vivo absorption of HP-tablet was markedly faster than those of H-tablet and HL-tablet, which little reflected the in vitro dissolution test. Fig. 4 also indicates that drug absorption of the three tablets was suppressed in the later phase, such as after 3 h post-dosing in H-tablet, while half of the drug was not yet absorbed. The result is in accordance with previous findings of insufficient drug absorption from several CR products in dogs [11–13]. As the colon arrival time of solid dosage forms was 2–3 h after dosing in fasted dogs [8,14], the initial phase would reflect the drug absorption mainly from the small intestine, and the later phase that from the colon. Our previous study [8] suggested that the cause of low AAP absorption in the colon was the suppression of drug release from CR tablets, not because of low absorption of drug in the colon. This would be due to the environment in the colon, namely the small volume of GI fluid and viscous colonic content, which would restrict fluid movement around the tablets and thereby retard drug dissolution. As a result, as shown in Table 1, AUCs of H-tablet and HL-tablet with incomplete drug release
3.2. In vivo absorption of CR tablets in dogs The three tablets were administered to dogs, in order to investigate the in vivo absorption profiles. Fig. 3 shows the mean plasma levels of AAP as a function of time in dogs. The highest plasma levels were obtained in dosing of HP-tablets. Table 1 shows the pharmacokinetic parameters after oral Table 1 Pharmacokinetic parameters after oral dosing of CR tablets at a dose of AAP 100 mg / body to fasted dogs
Each result shows the mean6S.D. of six dogs. *Significantly different (P,0.05), n.s., not significant.
170
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
before reaching the colon were significantly smaller (P,0.05) than that of HP-tablet.
3.3. In vitro /in vivo relationship Despite the equivalence in in vitro dissolution profiles of H-tablet, HL-tablet and HP-tablet (Fig. 1), the initial phase of in vivo AAP absorption markedly differed from each other (Fig. 4), with absorption rate of HP-tablet being fastest, followed by those of HL-tablet and H-tablet. In other words, a relationship was seen between higher water penetration and faster initial AAP absorption rate. The gel layer of the HP-tablet with high water retention could be disintegrated by GI destructive forces and its in vivo release rate markedly accelerated, while the H-tablet with low water penetration and thin gelled layer could disintegrate slightly in the GI tracts. Our previous study [15] demonstrated that, when polyethylene oxide was used as a gel-forming polymer, the matrix with rapid water penetration did not show the rapid initial drug absorption in dogs. The results suggest that the strength of the gel layer was different between the polymers; the gel layer of the HPMC could be weaker than that of polyethylene oxide. However, drugs release from HP-tablet may not be accelerated in human, because the mechanical destructive force in dogs would be greater than that in humans [16]. Further study will be required to clarify the behavior in human. The initial phase of in vivo drug release from HP-tablet containing PEG6000 was markedly faster, especially at around 1 h after dosing, than expected from its in vitro dissolution by the paddle method. If the mechanical stress in the GI tracts was constant, the absorption profiles of HP-tablet could be mimicked by the paddle method in high agitating condition. However, motility patterns in the GI tracts in fasted animals and humans follow complex cyclical processes and several phases of interdigestive migrating contraction (IMC). Single-unit dosage forms remain in the stomach until the phase III activity of IMC [17], which occurs about every 2 h in fasted dogs [18]. When the dosing time of tablet is not adjusted to the IMC patterns of dogs, the dosed tablets would be subjected to the intense contraction of phase III activity at irregularly, but possible at
around 1 h on average theoretically, for a half of average period of the complex cyclical process. This would affect the markedly accelerated drug absorption around 1 h after dosing. Thus, when designing an in vitro dissolution test reflecting the in vivo behavior, the intense mechanical stress at around 1 h after starting the dissolution test might be necessary. The rotating flask method with beads [19] and solubility simulator [19] are well-known dissolution tests for giving the mechanical stress to pharmaceutical formulations. But, in the preliminary test using these dissolution methods, the HPMC matrices were not subjected to damage by the beads, because the matrices floated in the test media while the beads existed on the bottom of vessels. In order to give the mechanical stress to the matrices in the in vitro dissolution test, the tablets were once removed from dissolution media and mechanical stress was given by shaking with glass beads. Fig. 5 shows the dissolution profiles of the tablets with mechanical stress. The release profile of HP-tablet was markedly accelerated by the mechanical stress. In the treatment, glass beads struck thousands of times on
Fig. 5. In vitro AAP dissolution profiles from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) using the paddle method at a rotation speed at 200 rpm in JP XIII 2nd fluid (pH 6.8) with mechanical stress by shaking with glass beads for 10 min at 1 h after starting the dissolution test. Each result shows mean6S.D. (n53).
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
swollen matrices, resulted in the gel layer of the HP-tablet almost completely disintegrating, while those of H-tablet and HL-tablet disintegrated only slightly. The experimental errors are relatively small in triplicate trials, maximum 3.2%, because the mechanical stress would be sufficient to disintegrate the whole gel layer of the HP-tablet. Therefore, good reproducibility was obtained in the drug release even with mechanical stress. As a result, the in vitro dissolution profiles differed significantly at the initial phase of in vivo drug absorption of the three matrices. Further, the patterns are similar to those of absorption profiles in dogs (Fig. 4). For example, the in vivo absorption at 3 h and in vitro dissolution with the mechanical stress demonstrates almost 1:1 relationship, while plateau relation (no relation) was observed in paddle methods, as shown in Fig. 6. The results indicated that a dissolution test with intense stress produced by shaking with glass beads would be one of the effective methods to design in vitro dissolution tests correlating with the initial phase of
171
in vivo conditions for the matrices. The results also suggest that a good in vitro / in vivo relationship could be obtained for matrix formulations by adjusting the mechanical stress of the dissolution test, even if the conventional dissolution test condition is not able to detect the difference in the formulations. The initial phase difference of the absorption profiles of the three tablets in dogs (Fig. 4) would be due to mechanical stress by GI motility mainly in the upper GI tracts, and later phase (after 3 h post dose) of H-tablet by colonic drug release. It would be difficult to mimic the later phase of the in vivo absorption profiles of H-tablet and HL-tablet, because the absorption rate was much slower than the paddle method in low agitating conditions (Fig. 1a). Fig. 2 demonstrates that the H-tablet without the water soluble filler had 60.562.9% of non-gelated residual core after 3-h immersion. As the residual core contained undissolved drug, the portion might be hardly released in the colon where little water existed. Thus, we should also consider the suppression of drug release in the colon in predicting in vivo performances.
4. Conclusions
Fig. 6. Relationship between the mean in vivo absorption in dogs at 3 h from H-tablet (circle), HL-tablet (triangle) and HP-tablet (square) and in vitro drug dissolution amount at 3 h by various dissolution test conditions (shaded symbol; paddle 25 rpm, open symbol; paddle 200 rpm and closed symbol; paddle 200 rpm with mechanical stress). Solid lines demonstrate results of linear curve fitting of each in vitro dissolution condition. Dotted line demonstrates 1:1 relationship between in vivo absorption in dogs and in vitro dissolution.
Although water-soluble filler in HPMC matrix tablets did not affect the in vitro AAP dissolution rate from the matrix, it did affect the in vivo absorption rate in the GI tracts. The in vivo disintegration was not predicted from in vitro dissolution test by the paddle method, even in the high agitation condition, such as with a paddle speed of 200 rpm. In contrast, the in vitro dissolution test with mechanical stress was able to reflect the in vivo absorption with almost 1:1 relationship between the three formulations. The difference was also observed in water penetration rate into the matrices. These results suggest that changes in the absorption rate in the GI tracts could be due to the potential intense mechanical stress. Thus, in selecting filler for matrices, predicting the effect of mechanical stress in the GI tracts by an effective method, such as the dissolution test under mechanical stress by glass beads, could be required in designing and investigating gel-forming HPMC matrices.
172
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
References [1] J.L. Ford, K. Mitchell, P. Rowe, D.J. Armstrong, P.N.C. Elliott, C. Rostron, J.E. Hogan, Mathematical modelling of drug release from hydroxypropylmethylcellulose matrices: effect of temperature, Int. J. Pharm. 71 (1991) 95–104. [2] J.L. Ford, M.H. Rubinstein, J.E. Hogan, Formulation of sustained release promethazine hydrochloride tablets using hydroxypropylmethylcellulose matrices, Int. J. Pharm. 24 (1985) 327–338. [3] J.L. Ford, M.H. Rubinstein, F. McCaul, J.E. Hogan, P.J. Edgar, Importance of drug type, tablet shape and added diluents on drug release kinetics from hydroxypropylmethylcellulose matrix tablets, Int. J. Pharm. 40 (1987) 223–234. [4] J.L. Salomen, E. Doelker, P. Burl, Sustained release of a water soluble drug from hydrophilic compressed dosage forms, Pharm. Ind. 41 (1979) 799–802. [5] J.L. Ford, M.H. Rubinstein, J.E. Hogan, Propranolol hydrochloride and aminophylline release from matrix tablet containing hydroxypropylmethylcellulose, Int. J. Pharm. 24 (1985) 339–350. [6] N. Katori, N. Aoyagi, T. Terao, Estimation of agitation intensity in the GI tract in humans and dogs based on in vitro / in vivo correlation, Pharm. Res. 12 (1995) 237–243. [7] S. Aoki, H. Ando, K. Tatsuishi, K. Uesugi, H. Ozawa, Determination of the mechanical impact force in the in vitro dissolution test and evaluation of the correlation between in vivo and in vitro release, Int. J. Pharm. 95 (1993) 67–75. [8] K. Sako, T. Mizumoto, A. Kajiyama, T. Ohmura, Influence of physical factors in gastrointestinal tract on acetaminophen release from controlled-release tablets in fasted dogs, Int. J. Pharm. 137 (1996) 225–232. [9] K. Iga, Y. Ogawa, T. Yashiki, T. Shimamoto, Estimation of drug absorption rates using a deconvolution method with non-equal sampling time, J. Pharmacokinet. Biopharm. 14 (1986) 213–225. [10] C.Y. Lui, G.M. Amidon, R.R. Berardi, D. Fleisher, C. Youngberg, J.B. Dressman, Comparison of gastrointestinal
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
[19]
pH in dogs and humans: Implications on the use of the beagle dog as a model for oral absorption in humans, J. Pharm. Sci. 75 (1986) 271–274. W.A. Cressman, D. Sumner, The dog as a quantitative model for evaluation of nondisintegrating sustained-release tablets, J. Pharm. Sci. 60 (1971) 132–134. N. Katori, K. Okudaira, N. Aoyagi, Y. Takeda, M. Uchiyama, In vitro and in vivo correlation for controlled-release formulation of d-chlorpheniramine maleate, J. Pharmacobio-Dyn. 14 (1991) 567–575. T. Uchida, M. Kawata, S. Goto, In vivo evaluation of ethyl cellulose microcapsules containing ampicillin using rabbits, dogs and humans, J. Pharmacobio-Dyn. 9 (1986) 631–637. K. Sagara, Y. Nagamatsu, I. Yamada, M. Kawata, H. Mizuta, K. Ogawa, Bioavailability of commercial sustained-release preparations of diclofenac sodium in gastro-intestinal physiology regulated-dogs, Chem. Pharm. Bull. 40 (1992) 3303– 3306. K. Sako, H. Nakashima, T. Sawada, M. Fukui, Relationship between gelation rate of controlled-release acetaminophen tablets containing polyethylene oxide and colonic drug release in dogs, Pharm. Res. 13 (1996) 594–598. H. Ogata, N. Aoyagi, K. Kaniwa, T. Shibazaki, A. Ejima, N. Takasugi, Bioavailability of nalidixic acid from uncoated tablets in human: II. Bioavailability in beagle dogs and its correlation with bioavailability in humans and in vitro dissolution rates, Int. J. Clin. Pharmacol. Ther. Toxicol. 22 (1984) 240–245. R.A. Hinder, K.A. Kelly, Canine gastric emptying of solids and liquids, Am. J. Physiol. 233 (1977) E335–E340. C.F. Code, J.A. Marlett, The interdigestive myoelectric complex of the stomach and small bowel of dogs, J. Physiol. 246 (1975) 289–309. H. Ogata, T. Shibazaki, T. Inoue, A. Ejima, Comparative studies on eight dissolution methods using 21 commercial chloramphenicol tablets and a nondisintegrating benzoic acid tablet, J. Pharm. Sci. 68 (1979) 708–712.
Influence of water soluble fillers in hydroxypropylmethylcellulose matrices on in vitro and in vivo drug release Kazuhiro Sako a , *, Toyohiro Sawada a , Hiroshi Nakashima a , Shigeharu Yokohama a , Takashi Sonobe b a
Novel Pharmaceutical Laboratories, Yamanouchi Pharmaceutical Co., Ltd., 180 Ozumi, Yaizu-shi, Shizuoka 425 -0072, Japan b School of Pharmaceutical Sciences, University of Shizuoka, 52 -1 Yada, Shizuoka-shi, Shizuoka 422 -8526, Japan Received 31 October 2001; accepted 1 March 2002
Abstract The purpose of this study was to investigate the effect of fillers in gel-forming matrix on in vivo drug release after oral dosing. A further purpose was to predict the in vivo performance from in vitro dissolution test. Three controlled-release acetaminophen tablets containing hydroxypropylmethylcellulose (HPMC) with or without highly water soluble fillers, lactose or polyethylene glycol 6000 (PEG6000), were prepared. Water penetration into the matrix was enhanced by addition of fillers in the matrices, but the three tablets showed similar in vitro dissolution profiles, indicating that fillers in the HPMC matrices little affected the in vitro drug release. In contrast, the fillers in HPMC matrices did affect the in vivo performance in dogs. The absorption profile of HPMC matrix with PEG6000 was the fastest, followed by that with lactose and without water soluble filler, in that order. As the matrix with PEG6000 had a large amount of water and gelated a large portion of the matrix when in contact with water, the gel layer would be disintegrated by the gastrointestinal motility. It was found that dissolution of gel-forming HPMC matrices under mechanical stress by glass beads well correlated with the in vivo performance of the matrix, with little correlation by the conventional paddle method. 2002 Elsevier Science B.V. All rights reserved. Keywords: In vitro / in vivo relationship; Acetaminophen; Controlled-release; Colonic release; Mechanical destructive forces
1. Introduction In recent years, many drugs have been developed as oral controlled-release (CR) dosage forms in order to optimize the drug therapy and to improve the patient’s compliance. Because of their simplicity and *Corresponding author. Tel.: 181-54-627-5155; fax: 181-54621-0106. E-mail address: [email protected] (K. Sako).
cost effectiveness, hydrophilic gel-forming matrix tablets are extensively used for oral CR dosage forms. Hydration of polymer results in the formation of a gel layer that controls the release rate of drug [1]. In vitro drug release of water-soluble drugs is controlled by diffusion out of the gel layer, whereas release for poorly soluble drugs is solely by polymer dissolution [1]. In addition, effects of various factors, namely drug / polymer ratio [1–3] and polymer viscosity [2,4,5], on in vitro drug release have been
0168-3659 / 02 / $ – see front matter 2002 Elsevier Science B.V. All rights reserved. PII: S0168-3659( 02 )00067-6
166
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
extensively investigated. However, in vitro drug release behavior is not always reflected in in vivo performance, because drug release from CR dosage forms is affected by various gastrointestinal (GI) factors. Among these, while the effect of pH on drug release has been extensively investigated, little is known about physical factors in GI motility such as agitation intensity [6] and mechanical destructive forces [7,8]. The purposes of adding fillers to the hydrophilic gel-forming matrix, in general, are for dilution, manufacturing processes improvement, adjusting drug release rate and so on. Investigating the effect of fillers in gel-forming matrices on in vivo drug release is important for designing matrix formulations. In this report, we prepared the hydroxypropylmethylcellulose (HPMC) matrices containing acetaminophen (AAP), as a model drug, with or without water soluble filler, and investigated the effect of the fillers of HPMC matrices on in vitro / in vivo drug release.
2. Material and methods
2.1. Materials Acetaminophen (AAP) was purchased from Yoshitomi Pharmaceutical Industries Ltd. (Japan). Hydroxypropylmethylcellulose 2208 (HPMC, viscosity of 1% aqueous solution 4100–5600 cps) was purchased from Shin-Etsu Chemical Co., Ltd. (Japan). Poly(ethylene glycol) used was of JP XIII (The Japanese Pharmacopoeia 13th edition) grades of Polyethylene Glycol 6000 (PEG6000), represented by the formula HOCH 2 (CH 2 O–CH 2 )nCH 2 OH, where n ranges from 165 to 210. Other reagents used were of analytical reagent grade.
2.2. Preparation of CR tablets
• HP-tablet: AAP, HPMC and PEG6000 were mixed (1:2:1) in a mortar, and the 400-mg mixtures were compressed directly using 9-mm diameter round-faced (9 mmR) punches. • HL-tablet: Lactose was used instead of PEG6000. The procedure was the same as for the preparation of the HP-tablet.
2.3. In vitro dissolution test The in vitro dissolution test of AAP from CR tablets was determined using Dissolution Apparatus No. 2 (paddle) of JP XIII. The dissolution medium was 500 ml of 2nd fluid (pH 6.8, 0.05 M H 2 KPO 4 and 0.0236 M NaOH) for JP XIII disintegration test. The amount of dissolved AAP was determined spectrophotometrically at 280 nm. After the last sampling at 12 h, the remaining matrices were broken by spatula to dissolve the matrices completely, in order to correct the 100% drug dissolution. When the mechanical stress was given to the matrices, the tablets were once removed from dissolution media at 1 h after starting the in vitro dissolution test and the tablets were transferred into a 50-ml test-tube with 60 g of glass beads (diameter, 4 mm; density, 2.5 g / cm 3 ) and 2 ml of water. The closed tube, with glass beads and the swollen matrix, was turned side ways and the mechanical stress was artificially given by shaking the tube horizontally at 320 strokes / min for 10 min with a shaker. After shaking, the whole content in the tube was returned into the dissolution medium, and the in vitro dissolution test was resumed.
2.4. Water penetration into matrices The tablet was immersed in JP XIII 2nd fluid (pH 6.8) at 37 8C. At 1, 2 and 3 h, the tablet was removed from the medium and the gel layer was peeled off carefully using a spatula, and the weight of the non-gelated residual core was measured.
The three AAP 100-mg CR tablets were prepared as follows:
2.5. Water retention measurement
• H-tablet: AAP and HPMC were mixed (1:2) in a mortar, and the 300-mg mixtures were compressed directly using 8.5-mm diameter roundfaced (8.5 mmR) punches.
The tablets at 3 h after the start of the water penetration test (2.4) were removed from the dissolution medium. After measuring the weight of a wetted tablet, the tablet was dried at 40 8C in a vacuum
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
dryer for 4 days. Then the weight of dried residue was measured. The water retention of the tablet was determined as the difference between wet and dry weights.
167
level–time curves (AUC) was calculated by the linear trapezoidal method. AAP absorption after oral dosing of CR tablets was calculated from oral dosing of solution and of CR tablets by the point-area deconvolution method [9].
2.6. In vivo absorption study 2.10. Statistical analysis Six male beagle dogs weighing 9.5–13.5 kg were fasted for 20 h before administration. They were allowed free access to water but food was withheld until the last blood sample was taken. One tablet containing AAP 100 mg was administered orally with 20 ml water. The interval between each administration was more than 1 week. Blood samples were collected from the forefoot vein at frequent intervals up to 12 h after dosing. Plasma samples were immediately separated and frozen at 220 8C until assay.
2.7. Plasma extraction An aqueous solution of 60 mg / ml 2-acetaminophenol was prepared for use as an internal standard. The internal standard (0.1 ml) solution and 5 ml of ethyl acetate were sequentially added to 0.5 ml of plasma in a test-tube. The tube was shaken for 10 min, and centrifuged for 5 min at 2000 rpm. The upper organic layer was transferred to a clean testtube, then evaporated. The dried residue was redissolved with 0.1 ml of mobile phase for HPLC assay.
2.8. HPLC assay of AAP AAP in plasma was determined by HPLC with UV detection according to a previously reported procedure [8]. In brief, separation on an octadecylsilane column (Nucleosil, 150 mm length34.6 mm diameter, 5 mm) was achieved at ambient temperature at a flow rate of 1 ml / min. The mobile phase contained water / acetonitrile / methanol (88:6:6). UV detection was at 254 nm.
2.9. Pharmacokinetic analysis Maximum plasma level (Cmax ) and time to maximum plasma level (T max ) were determined according to the standard procedure. The area under the plasma
Differences in each pharmacokinetic parameter were statistically evaluated by the paired t-test.
3. Results and discussion
3.1. In vitro characteristics of CR tablets The gel-forming tablets, composed of AAP and HPMC with or without water soluble filler, were used in this study. In general, pH-independent drug release is preferable for oral controlled-release (CR) formulations, so as not to be affected largely for their drug release in the GI tracts by intra- and intersubject variations of both gastric pH and GI transit time. In addition, as gastric pH in the dog shows large individual variation [10], physical factors of the GI must be investigated in pH-independent formulations. We selected AAP (pKa 9.5) as a model drug because its solubility in water is almost constant (approximately 15 mg / ml) over a physiological range of pH (1–7). Hydrophilic polymer (HPMC) and filler (PEG6000 and lactose) are also pH-independent materials. Thus, drug release from the tablets would not be affected by pH of the media. PEG6000 was selected as a very water soluble filler (1 g dissolves in 1 ml water), with lactose as a moderately water soluble one (1 g dissolves in 8 ml water). As shown in Fig. 1a, AAP dissolution profiles from H-tablet, HP-tablet and HL tablet are similar to each other for the paddle rotation speeds of 25 rpm. The drug dissolution from the three matrices demonstrated a good relation (r.0.999) to the square root of time, suggesting that the passage of drug, via diffusion, through gel layer controls the drug dissolution [2,3,5]. Even in a high agitating condition, paddle speed at 200 rpm, the three tablets demonstrated equivalent dissolution profiles, as shown in Fig. 1b: the maximum difference was 11.3% between H-tablet and HL-tablet at 12 h, and the difference at
168
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
Fig. 1. Comparison of AAP dissolution from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) using the paddle method at a rotation speed at 50 rpm (a) and 200 rpm (b) in JP XIII 2nd fluid (pH 6.8). Each result shows mean6S.D. (n53) and error bars are in the symbols.
3 h was only 4.5%. Ford et al. [3] reported that drug release rate was accelerated when lactose content increased in HPMC matrices with constant tablet mass. In this study, the amount of gel-forming polymer, HPMC, dominates the drug release and that of fillers affected it only slightly. As shown in Fig. 2, when the matrices were in
Fig. 2. Change in the portion of dried residual core after immersing in JP XIII 2nd fluid (pH 6.8). Each result shows mean6S.D. (n53).
contact with water, the portion of non-gelated residual core decreased with progress of time, the fastest in HP-tablets, followed by HL-tablet and H-tablet, in that order. Water retentions at 3 h after immersing the dissolution media were 598692, 486672 and 351658 mg / tablet in HP-tablet, HLtablet and H-tablet, respectively, indicating that highly water soluble filler in the matrices would stimulate the water penetration into the inner parts of the matrix. As a result, the dried residual part which contained the undissolved drug was larger in H-tablet (60.562.9%) than in HP-tablet (17.360.4%) after 3 h immersing. This would also affect the difference of wet strength of the matrices. The mechanical strength of HP-tablet in wet condition with a thick gel layer and a small dried residual core could be weaker than that of the H-tablet with thin gel layer and large dried residual core. We previously reported [8] that CR tablet was subjected to mechanical destructive force in the GI tract by the GI motility. We also reported that high agitation condition of in vitro dissolution test, such as paddle speed at 200 rpm, reflected the in vivo performance in dogs of low viscous HPMC (50 cps, 2% aqueous solution) matrix which did not have sufficient wet strength against GI motility. However, the difference among the three matrices demonstrated equivalence in the in
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
169
Fig. 3. Plasma levels of AAP after oral dosing of H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) to dogs. Each result shows mean6S.E. in six dogs.
Fig. 4. In vivo absorption of AAP from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) after oral dosing to dogs.
vitro dissolution profiles even in the high agitation condition of paddle speed at 200 rpm (Fig. 1b). The results demonstrate that the gel layer of high viscous HPMC could resist the high agitation by the paddle rotation, even though the matrix contained the watersoluble filler. As a result, the three tablets showed similar drug release profiles in vitro, even in high agitating conditions.
dosing with the three tablets. Fig. 4 shows AAP absorption profiles of the three tablets calculated by deconvolution. Nevertheless similar in vitro dissolution profiles were observed for the three tablets, the in vivo absorption of HP-tablet was markedly faster than those of H-tablet and HL-tablet, which little reflected the in vitro dissolution test. Fig. 4 also indicates that drug absorption of the three tablets was suppressed in the later phase, such as after 3 h post-dosing in H-tablet, while half of the drug was not yet absorbed. The result is in accordance with previous findings of insufficient drug absorption from several CR products in dogs [11–13]. As the colon arrival time of solid dosage forms was 2–3 h after dosing in fasted dogs [8,14], the initial phase would reflect the drug absorption mainly from the small intestine, and the later phase that from the colon. Our previous study [8] suggested that the cause of low AAP absorption in the colon was the suppression of drug release from CR tablets, not because of low absorption of drug in the colon. This would be due to the environment in the colon, namely the small volume of GI fluid and viscous colonic content, which would restrict fluid movement around the tablets and thereby retard drug dissolution. As a result, as shown in Table 1, AUCs of H-tablet and HL-tablet with incomplete drug release
3.2. In vivo absorption of CR tablets in dogs The three tablets were administered to dogs, in order to investigate the in vivo absorption profiles. Fig. 3 shows the mean plasma levels of AAP as a function of time in dogs. The highest plasma levels were obtained in dosing of HP-tablets. Table 1 shows the pharmacokinetic parameters after oral Table 1 Pharmacokinetic parameters after oral dosing of CR tablets at a dose of AAP 100 mg / body to fasted dogs
Each result shows the mean6S.D. of six dogs. *Significantly different (P,0.05), n.s., not significant.
170
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
before reaching the colon were significantly smaller (P,0.05) than that of HP-tablet.
3.3. In vitro /in vivo relationship Despite the equivalence in in vitro dissolution profiles of H-tablet, HL-tablet and HP-tablet (Fig. 1), the initial phase of in vivo AAP absorption markedly differed from each other (Fig. 4), with absorption rate of HP-tablet being fastest, followed by those of HL-tablet and H-tablet. In other words, a relationship was seen between higher water penetration and faster initial AAP absorption rate. The gel layer of the HP-tablet with high water retention could be disintegrated by GI destructive forces and its in vivo release rate markedly accelerated, while the H-tablet with low water penetration and thin gelled layer could disintegrate slightly in the GI tracts. Our previous study [15] demonstrated that, when polyethylene oxide was used as a gel-forming polymer, the matrix with rapid water penetration did not show the rapid initial drug absorption in dogs. The results suggest that the strength of the gel layer was different between the polymers; the gel layer of the HPMC could be weaker than that of polyethylene oxide. However, drugs release from HP-tablet may not be accelerated in human, because the mechanical destructive force in dogs would be greater than that in humans [16]. Further study will be required to clarify the behavior in human. The initial phase of in vivo drug release from HP-tablet containing PEG6000 was markedly faster, especially at around 1 h after dosing, than expected from its in vitro dissolution by the paddle method. If the mechanical stress in the GI tracts was constant, the absorption profiles of HP-tablet could be mimicked by the paddle method in high agitating condition. However, motility patterns in the GI tracts in fasted animals and humans follow complex cyclical processes and several phases of interdigestive migrating contraction (IMC). Single-unit dosage forms remain in the stomach until the phase III activity of IMC [17], which occurs about every 2 h in fasted dogs [18]. When the dosing time of tablet is not adjusted to the IMC patterns of dogs, the dosed tablets would be subjected to the intense contraction of phase III activity at irregularly, but possible at
around 1 h on average theoretically, for a half of average period of the complex cyclical process. This would affect the markedly accelerated drug absorption around 1 h after dosing. Thus, when designing an in vitro dissolution test reflecting the in vivo behavior, the intense mechanical stress at around 1 h after starting the dissolution test might be necessary. The rotating flask method with beads [19] and solubility simulator [19] are well-known dissolution tests for giving the mechanical stress to pharmaceutical formulations. But, in the preliminary test using these dissolution methods, the HPMC matrices were not subjected to damage by the beads, because the matrices floated in the test media while the beads existed on the bottom of vessels. In order to give the mechanical stress to the matrices in the in vitro dissolution test, the tablets were once removed from dissolution media and mechanical stress was given by shaking with glass beads. Fig. 5 shows the dissolution profiles of the tablets with mechanical stress. The release profile of HP-tablet was markedly accelerated by the mechanical stress. In the treatment, glass beads struck thousands of times on
Fig. 5. In vitro AAP dissolution profiles from H-tablet (open circle), HL-tablet (open triangle) and HP-tablet (open square) using the paddle method at a rotation speed at 200 rpm in JP XIII 2nd fluid (pH 6.8) with mechanical stress by shaking with glass beads for 10 min at 1 h after starting the dissolution test. Each result shows mean6S.D. (n53).
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
swollen matrices, resulted in the gel layer of the HP-tablet almost completely disintegrating, while those of H-tablet and HL-tablet disintegrated only slightly. The experimental errors are relatively small in triplicate trials, maximum 3.2%, because the mechanical stress would be sufficient to disintegrate the whole gel layer of the HP-tablet. Therefore, good reproducibility was obtained in the drug release even with mechanical stress. As a result, the in vitro dissolution profiles differed significantly at the initial phase of in vivo drug absorption of the three matrices. Further, the patterns are similar to those of absorption profiles in dogs (Fig. 4). For example, the in vivo absorption at 3 h and in vitro dissolution with the mechanical stress demonstrates almost 1:1 relationship, while plateau relation (no relation) was observed in paddle methods, as shown in Fig. 6. The results indicated that a dissolution test with intense stress produced by shaking with glass beads would be one of the effective methods to design in vitro dissolution tests correlating with the initial phase of
171
in vivo conditions for the matrices. The results also suggest that a good in vitro / in vivo relationship could be obtained for matrix formulations by adjusting the mechanical stress of the dissolution test, even if the conventional dissolution test condition is not able to detect the difference in the formulations. The initial phase difference of the absorption profiles of the three tablets in dogs (Fig. 4) would be due to mechanical stress by GI motility mainly in the upper GI tracts, and later phase (after 3 h post dose) of H-tablet by colonic drug release. It would be difficult to mimic the later phase of the in vivo absorption profiles of H-tablet and HL-tablet, because the absorption rate was much slower than the paddle method in low agitating conditions (Fig. 1a). Fig. 2 demonstrates that the H-tablet without the water soluble filler had 60.562.9% of non-gelated residual core after 3-h immersion. As the residual core contained undissolved drug, the portion might be hardly released in the colon where little water existed. Thus, we should also consider the suppression of drug release in the colon in predicting in vivo performances.
4. Conclusions
Fig. 6. Relationship between the mean in vivo absorption in dogs at 3 h from H-tablet (circle), HL-tablet (triangle) and HP-tablet (square) and in vitro drug dissolution amount at 3 h by various dissolution test conditions (shaded symbol; paddle 25 rpm, open symbol; paddle 200 rpm and closed symbol; paddle 200 rpm with mechanical stress). Solid lines demonstrate results of linear curve fitting of each in vitro dissolution condition. Dotted line demonstrates 1:1 relationship between in vivo absorption in dogs and in vitro dissolution.
Although water-soluble filler in HPMC matrix tablets did not affect the in vitro AAP dissolution rate from the matrix, it did affect the in vivo absorption rate in the GI tracts. The in vivo disintegration was not predicted from in vitro dissolution test by the paddle method, even in the high agitation condition, such as with a paddle speed of 200 rpm. In contrast, the in vitro dissolution test with mechanical stress was able to reflect the in vivo absorption with almost 1:1 relationship between the three formulations. The difference was also observed in water penetration rate into the matrices. These results suggest that changes in the absorption rate in the GI tracts could be due to the potential intense mechanical stress. Thus, in selecting filler for matrices, predicting the effect of mechanical stress in the GI tracts by an effective method, such as the dissolution test under mechanical stress by glass beads, could be required in designing and investigating gel-forming HPMC matrices.
172
K. Sako et al. / Journal of Controlled Release 81 (2002) 165 – 172
References [1] J.L. Ford, K. Mitchell, P. Rowe, D.J. Armstrong, P.N.C. Elliott, C. Rostron, J.E. Hogan, Mathematical modelling of drug release from hydroxypropylmethylcellulose matrices: effect of temperature, Int. J. Pharm. 71 (1991) 95–104. [2] J.L. Ford, M.H. Rubinstein, J.E. Hogan, Formulation of sustained release promethazine hydrochloride tablets using hydroxypropylmethylcellulose matrices, Int. J. Pharm. 24 (1985) 327–338. [3] J.L. Ford, M.H. Rubinstein, F. McCaul, J.E. Hogan, P.J. Edgar, Importance of drug type, tablet shape and added diluents on drug release kinetics from hydroxypropylmethylcellulose matrix tablets, Int. J. Pharm. 40 (1987) 223–234. [4] J.L. Salomen, E. Doelker, P. Burl, Sustained release of a water soluble drug from hydrophilic compressed dosage forms, Pharm. Ind. 41 (1979) 799–802. [5] J.L. Ford, M.H. Rubinstein, J.E. Hogan, Propranolol hydrochloride and aminophylline release from matrix tablet containing hydroxypropylmethylcellulose, Int. J. Pharm. 24 (1985) 339–350. [6] N. Katori, N. Aoyagi, T. Terao, Estimation of agitation intensity in the GI tract in humans and dogs based on in vitro / in vivo correlation, Pharm. Res. 12 (1995) 237–243. [7] S. Aoki, H. Ando, K. Tatsuishi, K. Uesugi, H. Ozawa, Determination of the mechanical impact force in the in vitro dissolution test and evaluation of the correlation between in vivo and in vitro release, Int. J. Pharm. 95 (1993) 67–75. [8] K. Sako, T. Mizumoto, A. Kajiyama, T. Ohmura, Influence of physical factors in gastrointestinal tract on acetaminophen release from controlled-release tablets in fasted dogs, Int. J. Pharm. 137 (1996) 225–232. [9] K. Iga, Y. Ogawa, T. Yashiki, T. Shimamoto, Estimation of drug absorption rates using a deconvolution method with non-equal sampling time, J. Pharmacokinet. Biopharm. 14 (1986) 213–225. [10] C.Y. Lui, G.M. Amidon, R.R. Berardi, D. Fleisher, C. Youngberg, J.B. Dressman, Comparison of gastrointestinal
[11]
[12]
[13]
[14]
[15]
[16]
[17] [18]
[19]
pH in dogs and humans: Implications on the use of the beagle dog as a model for oral absorption in humans, J. Pharm. Sci. 75 (1986) 271–274. W.A. Cressman, D. Sumner, The dog as a quantitative model for evaluation of nondisintegrating sustained-release tablets, J. Pharm. Sci. 60 (1971) 132–134. N. Katori, K. Okudaira, N. Aoyagi, Y. Takeda, M. Uchiyama, In vitro and in vivo correlation for controlled-release formulation of d-chlorpheniramine maleate, J. Pharmacobio-Dyn. 14 (1991) 567–575. T. Uchida, M. Kawata, S. Goto, In vivo evaluation of ethyl cellulose microcapsules containing ampicillin using rabbits, dogs and humans, J. Pharmacobio-Dyn. 9 (1986) 631–637. K. Sagara, Y. Nagamatsu, I. Yamada, M. Kawata, H. Mizuta, K. Ogawa, Bioavailability of commercial sustained-release preparations of diclofenac sodium in gastro-intestinal physiology regulated-dogs, Chem. Pharm. Bull. 40 (1992) 3303– 3306. K. Sako, H. Nakashima, T. Sawada, M. Fukui, Relationship between gelation rate of controlled-release acetaminophen tablets containing polyethylene oxide and colonic drug release in dogs, Pharm. Res. 13 (1996) 594–598. H. Ogata, N. Aoyagi, K. Kaniwa, T. Shibazaki, A. Ejima, N. Takasugi, Bioavailability of nalidixic acid from uncoated tablets in human: II. Bioavailability in beagle dogs and its correlation with bioavailability in humans and in vitro dissolution rates, Int. J. Clin. Pharmacol. Ther. Toxicol. 22 (1984) 240–245. R.A. Hinder, K.A. Kelly, Canine gastric emptying of solids and liquids, Am. J. Physiol. 233 (1977) E335–E340. C.F. Code, J.A. Marlett, The interdigestive myoelectric complex of the stomach and small bowel of dogs, J. Physiol. 246 (1975) 289–309. H. Ogata, T. Shibazaki, T. Inoue, A. Ejima, Comparative studies on eight dissolution methods using 21 commercial chloramphenicol tablets and a nondisintegrating benzoic acid tablet, J. Pharm. Sci. 68 (1979) 708–712.