Optimization and Its Applications on Pharmaceutical Conventional and Extended Release Solid Dosage Forms

. ............................. ........................... ... .................................................. I I I ACKNOWLEDGEMENTS..................... ................... ....... ..... ... ............ ...... ......... ......... VI I PREFACE .................................................................................................................. VIII LIST OF TABLES ........ ...... ........................... ..... ... ..... ........................ ..... .......... ... .. .... . XI LIST OF FIGURES .. .... .................... ....................... .................... ........................... ... .. x;v LIST OF PUBLICATIONS AND PRESENTATIONS...... ........... .................................. ~

To understand the release mechanism of drug from extended release polymeric matrix tablet, the swelling and dissolution behavior of different molecular weight PEO (polyethylene oxide) polymers rn distilled water at 37 Oc was investigated. Due to the swelling of PEO matrix discs, considerable volume expansion was observed. Molecular weight is an important determinant of PEO dissolution rate, which was inversely proportional to the molecular weight of PEO. The results supported the hypothesis that dissolution of high molecular weight PEO is controlled by the inward diffusion of water and outward diffusion of polymer through the boundary layer.
The influence of the molecular size and solubility of four tracer compounds (phenylpropanolamine HCl, theophylline, sotalol HCI and bovine serum albumin) and the effect of the tracer/PEO ratio on the dissolution rate in SIF (simulated intestinal fluid) were determined.
In the process of bioadhesion assessment, an apparatus to be equipped with Instron tensile tester was developed to evaluate quantitatively the bioadhesive properties of various bioadhesive tablets. The equipment was designed to measure the forces required to separate two parallel surfaces (tablet and membrane) in both horizontal and vertical planes. In this work, in addition to the detachment force and adhesion work, the shear force necessary for separating bioadhesive tablet and synthetic membrane or biological tissue (rabbit stomach mucosa) were also determined since the majority of gastrointestinal mocosa surface area possesses some elements of tangential shear motion. The effects of different quantities and types of bioadhesive polymer on the tablet bioadhesive capability were also determined. The results showed good agreement with some previous findings that the relative adhesion of the tablet formulations was dependent on the bioadhesive polymer content. It was also found that tablet made with sodium carboxymethycellulose (NaCMC) possessed the best bioadhesive power when compared to tablets made with polycarbophil and carbopol 974P.
( Tables   Manuscript I   Table I   Table II   Table III   Table IV   Table V   Table VI LIST        Manuscript IV            Manuscript V Table I : The Shear Force and Adhesion Works Measured Table II   Table III Appendix B Table I '                          (13 ).
The optimization process provides not only efficient use of resources, but also a method to obtain a mathematical model which can be used to characterize and optimize a formulation or process. Furthermore, by accurately defining the whole system, optimization techniques are a useful aid to process validation.
The wet granulation process has been used as an alternative for high dose and poorly compressible active ingredients. It offers several advantages over other methods, for instance, it improves flowability, resistance to segregaion and compression characteristics by increasing the particle size and cohesion (14 ). Acetaminophen is a drug which requires a high dose, thus resulting in a large tablet.
Since the drug also has poor compactability, the amount of added excipient (MCC) required to produce acceptable compaction behavior ( ( must be very completely controlled. However, during the wet granulation process, there are many process and formulation variables which will affect the physical properties of the granules and of the final tablets (15).  (2) that requires minimal familiarity with computer programming or optimization mathematics. Through the process of optimization, the researcher may discover solutions to formulation challenges which would otherwise be dismissed as unrealistic. The technique has been extended to obtain some desirable pharmaceutics and pharmacokinetic parameters by variation of formulation and process parameters (3)(4)(5)(6)(7).
Standard formulation methods often involve running a grid search about a formulation or process starting point. The initial point is either an educated guess or deduced from prior art. Such grid searches are expensive in terms of time, labor and materials, and also may result in a missed solution to the problem. One reason for the existence of serendipitous solutions is illustrated by the analogy of the effects of higher order harmonics produced by constructive from the interaction of two or more fundamental frequencies.
Similarly, either the concentrations of two or more ingredients, or the levels of two or more processing parameters may interact to produce an unanticipated result. This is sometimes refered to as synergism or potentiation, in which the effect of supposedly independent factors is many fold the sum of effects of the factors taken separately. Thus, some factors may be dis.covered to be interdependent.
Utilizing the tool of optimization, workers have developed and marketed a tablet formulation containing 800 mg of ibuprofen, a poorly compactable material-me! · tablet with a minimum of excipient (8). This made possible the manufacture of a tablet of palatable dimensions and acceptable hardness, friability and dissolution performance.
A second advantage obtained when using optimization is the substantial time and cost savings due to the inherent efficiency of a rational experimental design (9) In this present study, acetaminophen was selected as a model drug because it requires a high dose (500 mg), resulting m a correspondingly large tablet. Since this drug also has poor compactability, the amount of added excipient, microcrystalline cellulose, required to produce acceptable compaction behavior is increased (10)(11)(12)(13), the ratio of ACMP to Emcocel™ and some process variables are required to be optimized to produce an acceptable tablet volume and physical properties. Also, most commercial tablets containing 500 mg acetaminophen are produced from pre-compacted granulation by variation on a patented roller-compactor process.
This work shows the attributes of tablets made by an alternative densification process, wet granulation. 50 micron MCC, as opposed to 90 micron MCC, is similar in particle size to the ACMP powder, which is expected to improve mixing in the manufacturing process and wettability in the disintegration process.

Experimental Design
The four independent variables and their ranges selected for wet granulation process were summarized in Table I  The sets of data obtaining from the statistical analysis were then subjected to computerized regression analysis to determine the fit to a second-order model. These regression models include an intercept and marn effect terms of each independent variable, two-way interaction terms and second order effect terms as shown in Table V.
A stepwise regression procedure was used to assess all main effects, some two-way interactions and quadratic terms for usefulness in the model to obtain a more adequate regression model for each response parameter (15)(16).  Contour plots for each of response parameters were generated using selected quadratic response surface model. Figure   The optimum values obtained from the contour plots for the independent variables in order to obtain the best values for each of the four response variables are given in minutes.

RESULTS AND DISCUSSION
The optimum range of granulation time is 8 to 10 Since in vitro dissolution data may provide an indication of in VIVO bioavailability, therefore, the percentage of drug dissolved at 20 r minutes was identified as the response parameter of pnmary concern. It was maximized so as to obtain the fastest dissolution rate. As shown in Table XIII, the constraints used for obtaining the fastest dissolution were that disintegration time should be greater than 0.3 minute, the friability should be less than 0.8% and the required compression force should be less than 14 KN. Additional constraints were the experimental range limits placed on values of all independent variables. The optimum formulat~on satisfied all constraints simultaneously and provided an optimal value for the primary function, rapid dissolution.
The formulation according to the optimal solution was prepared as shown in Table XIII and tablets were manufactured on the rotary press, tablets properties were also determined. The comparison of predicted and experimental values for optimum formulation showed very good agreement and are shown in Table XIV. A model is valid despite its inexactness in representing the system, it can give a reasonable prediction of a system performance.
The optimized 500 mg ACMP tablets were compared with some commercially available 500 mg ACMP tablets in terms of dissolution, disintegration time, hardness, friability, weight and volume. As shown in Table XV, the optimized tablets made without any disintegrant exhibit satisfactory and comparable dissolution characteristics. The disintegration time of optimized tablets is even shorter than two of commercial tablets. Tylenol tablets possess the lowest friability.
Although the optimized tablets have the highest ( ( tablet weight, however, due to the higher density of the granules, the tablet volume is very similar to the commercial tablets.

CONCLUSIONS
By usmg computer optimization process, with some constraints on other tablets properties, 300 gm formulation batch containing 25% of Emcocel™, 1 % of povidone in 120 gm granulating water, granulated in 9 minutes was found to be able to produce 500 mg ACMP tablet~ which possess the best dissolution characteristic, about 95% of drug dissolved at 15 minute sampling time. These tablets also exhibit fast disintegration time, 30 seconds, even without any disintegrant in the tablet, 8 kg hardness, 0.3% friability, 9 KN required compression force for producing tablets and very comparable tablet volume with commercially available 500 mg ACMP tablets. The wet granulation process utilizing Emcocel™ as an excipient to densification seems to be a feasible alternative to the dry compaction approach for producing 500 mg acetaminophen tablet. The expensive drying process of wet granulation process may be offset by saving in starting material costs and in omission of the slugging or roller compaction steps used in dry processing.
In this high dose ACMP tablet formulation development, computer-assisted regression analysis and mathematic model can be utilized to produce accurate representation of the relationship between the independent variables and tablets response properties Term Error Tera Error          During the last two decades, several floating drug delivery systems and formulations have been developed a1mmg to achieve the same intended intragastric buoyance function (5)(6)(7)(8)(9)(10)(11)(12)(13)(14).

2.2
Types and principles of floating dosage forms Tossounian et al. (16)  Michael et al. (5) (6) impregnating the active ingredient into a body of empty globular shell or a granular lump in small size of a material having high buoyancy. They also prepared floating systems by suitably adhering a crust of coating containing a desired drug on external and/or internal surfaces of a conventional soft or hard capsules having a bulk density less than that of gastric fluid in the stomach. In another embodiment of their invention, they also plugged a flat tablet containing an active drug ingredient into a half piece of a compositive capsule and sealed with a binding agent such as ethylcellulose dissolved in 1,1, 1-trichloroethane. This half piece of capsule was coated with a crust of hydroxypropylmethylcellulose phthalate .
Urquahart and Theeuwes also introduced a floating drug delivery system comprising a reservoir containing a plurality of tiny pills (8).
In this delivery system, the tiny pills have a core of active drug ingredient which are coated with a wall formed of a drug-release rate controlling fatty acid and wax, these tiny coated pills were then dispersed throughout a hydrophilic matrix which swells considerably in contact with gastric fluid for retaining the device in the stomach.
A flexible, sheet-like, floating sustained release medicament device having a bulk density of less than one was designed by Mitra et al. (9), the device is of a multi-layer composite construction comprising at least one dry, self-supporting carrier film which is formed of one or more water insoluble polymer matrices and drug. A recent floating sustained release system was prepared by Bolton et al. (11 ), the tablets comprise a hydrocolloid gelling agent such as agar, a pharmaceutically acceptable inert oil, such as light mineral oil, the drug, theophylline and water. The final tablets possess a density less than one and therefore will remain buoyant on gastric fluid in stomach. Typically, the density of tablet 1s ranged from 0.6 to 0.95. In the preparation of this floating tablet, a solution of the hydrocolloid gelling agent in warm water and a solution of active drug ingredient, theophylline, in the selected oil were separately prepared and these two solutions were mixed and cooled but not to the point where gelation of the gelling agent takes places, the emulsions then were poured into tablet molds and left until the gel forms and drying.
Although the resulting tablet is not compressed, the inventors claimed that the final tablets hardness values are comparable to that of most commercially available tablets.
These tablets have sufficient mechanical stability to stand up to the normal stress of production, packaging and despensing. The hardness is characterized by a network of multitudinous air holes and passages. In this invention, the preferred gelling agent is agar, although the inventors claimed other gelling agents may be used. to 30 % of inert oil is necessary in the initial mixture before gelling.
Ushimaru ~also manufactured a floating sustained release delivery system consisting a substance which forms gel in water, a ( fat/oil which is solid at room temperature and drug (10). These substances were simply mixed and filled into a capsule, the capsule is heated at the temperature higher than the melting point of the fat/oil and then cooled to room temperature, the resulting product was then recovered and have a specific gravity of less than 1.0 to be able to float on the gastric fluid in the stomach and to undergo sustained release of active drug ingredient.
Another invention relates to a granule remammg m the stomach for a prolonged period of time was invented by Ichikawa and his coworkers to provide better buoyancy when compared to some floating tablets and capsules (12). The granules compnse a core containing a active ingredient, foaming layer coated on the core and an expansive film coated on the foaming layer. The foaming layer was composed of a bicarbonate or a combination of an inner layer of a bicarbonate and an outer layer of an organic acid. The expansive film was made of a polymer which allow the gastric fluid to penetrate into the inside of the granule and then expand like a ballon because of the gas evolved within the granule to thereby retain the gas within the granule for requred period of time.

Factors affecting floating capability of the dosage forms
In order to remain buoyant in the stomach for extended period of time, it is imperative for HBS dosage forms to maintain an overall bulk density lower than that of gastric fluid after they are m contact ( with gastric fluid. Ordinarily, the HBS tablets can be manufactured on conventional tabletting equipments, however, in accordance with the HBS tablets floating principals, HBS tablets can remain buoyant in stomach even their initial bulk density is greater than 1 because the buoyancy could be obtained from a combination of an increase in the bulk volume of the tablet due to the hydration and swelling of the hydrocolloid particles on the tablets surface when in contact with gastric fluid and the internal voids m the tablet center remaining dry due to the barrier formed by the hydrocolloid particles (23 ).
Therefore, it is essential that the tablet are not compressed so tightly that rapid hydration is retarded which result in not obtaining a bulk density of less than one after in contact with gastric fluids.
This critical maximum hardness will vary both with the initial density of the formulation and the size of the tablet. Some investigations concluded that the effectiveness of the intragastric buoyancy of floating systems is dependent on particular physiological condition (such as gastric emptying, pH and specific gravity of gastric fluid etc.) and dosage forms characteristics (such as bulk density of the excipients, hardness of the tablet, size, swelling and hydration degree of the final products) (16,(18)(19)(20)(21)(22)(23)(24)(25)(26).

Technologies of preparing floating dosage form
The HBS floating dosage forms were initially prepared m a capsule form, they were prepared by homogeneous mixing one or more drugs with one or more hydrophilic hydrocolloids (27, 28), if necessary, fatty material and some inert pharmaceutical excipients at the optimized percentages.
The formulations then were passed through a Fitzpatrick comminuting machine using different sizes of plate or screen at certain speed, the talc or magnesium stearate were then added to the formulation as a lubricant and blended for an additional time.
The blending and milling processes were repeated so that the formulation mixture can pass through a certain size mesh screen and then the mixture was filled into a optimal size soft or gelatine capsules.
In the preparation of formulation blends, granulation process sometime was required to prepare granules to increase floating capability and flowability of the formulation.
The HBS products were also manufactured in tablet form. Sheth
When the concentration of polymer is below the entanglement concentration, the polymer molecules do not interact strongly, and the polymer-solvent system behaves as a solution.
However, when the polymer concentration is above the entanglement concentration, the macromolecules are intertwined sufficiently to provide some degree of physical integrity and the system behaves as a viscoelastic gel. The entanglement concentration is a decreasing function of molecular weight. Therefore, a polymer of sufficiently high molecular weight will undergo a significant degree of swelling before dissolving. The swelling process will exert considerable stress on the polymer and crazing may occur at the swelling region. This phenomenon can be utilized to release the active agent at a controlled rate.
A number of workers including Lee (3)(4)(5), Hopfenberg et al. (6), Colombo et al. (7),  and Hogan ( 11) have demonstrated the potential utility of swelling-controlled systems for zero order or near-zero order release. Meanwhile, some previous contributions including those made by Good (12), Korsmeyer et al. (13), Lee (14,15), Peppas et al. (16), and Graham et al. (17,18) provided a preliminary understanding of the mechanism of solute release from swelling-controlled systems. Peppas and his coworkers ( 19) recently presented mathematical models to predict the mass of drug released and the polymer gel layer thickness as a function of time.
The recent developments and applications of swelling controlled release system have been reviewed by Ranga Rao et al. (20). In each case, the assumption is that the rate of dissolution of drug from any solid drug particle is rapid with respect to other processes, so that drug dissolution rate within the matrix is not rate limiting. For instance, in one type of system, the polymer matrix is initially glassy and crystalline, so that drug is immobilized and cannot diffuse out of the system. Upon entering a compatable medium, the polymer would swell and become rubbery Such a swelling controlled system would presumably be less dependent on agitation intensity than a different polymer system, such as one in which pure surface erosion controlled drug release.
This agitation-independent characteristic would be desirable since agitation intensity is difficult to assess in vivo.
In an erosion-controlled system, drug molecules would remain immobilized within a glassy matrix until the moment at which the surrounding matrix eroded and was dispersed into the medium. A hybrid system would be one with both swelling and erosion characteristics.
In such a case, the polymer matrix would swell to produce a gel layer, but after reaching a critical thickness, the dilute outer edge of the gel layer would begin to erode and disperse into the medium. As a result, the gel layer would reach a maximum or critical thickness, after which point the gel layer would have a constant thickness.
The result would be a constant rate of drug release controlled by diffusion through the gel layer. Clarification of the drug release mechanisms operating in a given dosage form is useful for design and application purposes.
In general, the behavior of a swellable/erodible delivery system is dependent on the relative rates of three processes: (a) the rate of water penetration into the polymer matrix (b) the dissolution rate of polymer matrix itself, and (c) the rate of drug transport through the polymer matrix. A complicating factor is seen for a matrix containing a high concentration of a low molecular weight solute. In this case, the extent of matrix swelling is initially high due to the high osmotic pressure exerted by yet-unreleased solute. Eventually as the solute is depleted, the osmotic pressure falls, so that the matrix contracts due to dominance of the elastic recovery tendancy of the entangled polymer. This also causes a net outward flux of medium as the matrix contracts, and dissolved solute or drug is carried outward at an enhanced rate greater than that due to diffusion alone. Any one or a combination of these processes may control the rate of drug delivery from a system. In the present investigation, the processes of swelling and erosion were studied for different molecular weight types of a model polymer, poly(ethylene oxide) (PEO).

Materials
The six grades of PEO used in this study are shown in Table II. The water soluble tracers used in these studies are listed in Table III. The size and shape (thin discs) of the tablets were selected to simplify interpretation of the swelling . and dissolution data, rather than to represent a tablet design suitable for human use.   (21).

Swelling Behavior of PEO Matrix Tablets
The swelling behavior of polyethylene oxide samples of different molecular weights is demonstrated in Figure 1 and 2. The swelling behavior is presented in terms of the water uptake. The tablet water uptake was calculated by subtracting the dissolved polymer weight from the initial dry polymer weight to obtain the remaining polymer weight (dry basis). Next, the remaining dry polymer weight was subtracted from the gross weight of the wet tablet sample. In the cases of PEO 1 M, PEO 4M and PEO 5M, the tablet underwent considerable swelling before dissolution. The swelling of these high molecular weight PEO tablets appears to be diffusion rate controlled.
As shown in Figures 3, 4  where: D is diffusivity, n is solution viscosity, r is molecular radius and N is Avagadro's number. Thus, this release mechanism( diffusion across a boundary layer) yields a dissolution rate that is proportional to molecular weight.
If instead, dissolution is controlled by the time it takes a polymer chain to disentangle itself from a concentrated gel, then the rate should be proportional to a higher power of molecular weight.
Ueberreiter (23) proposed an empirical relationship between molecular weight and dissolution rate: where G is the dissolution rate, Mw is the molecular weight, K and A are constants.
This is similar to the Mark-Houwink equation, The log (dissolution rate) vs. Log (Mw) of PEO is presented in Figure 12. The slope 1s -0.65 (S = 0.05). This value seems more supportive of the boundary layer mechanism than the disentanglement mechanism.
On the dissolution study of PEO 3.5K, the tablets dissolve directly, with no observable intermediate gel state, one would expect that a drug-containing matrix of this material would release the drug at a rate equal to the polymer dissolution rate. As shown in Figure 10, the dissolution of the PEO lOOK is constant until most of tablet (75% or so) is dissolved. The dissolution rate does not vary with tablet thickness. This is due to a nearly constant tablet surface area during the dissolution study. For the thickest ( 0.62 inch ) of these tablets, the initial sidewall surface area ( As= 2 11 r h) is only 20 % of the initial combined obverse and reverse tablet areas ( Aor = 2 11 r 2 ).
The weight curves of PEO 1 OOK (Fig. 1) show that tablet picks up water relatively rapidly at first, but then the rates of dissolution and water penetration come into balance. This kind of dissolution behavior is desirable for controlled release. The PEO lOOK dissolution appears to proceed by a pseudo-steady-state process in which an outer gel layer is continuously formed as water penetrates the tablet and is simultaneously eroding at the outer boundary.
Since both water penetration and dissolution are proceeding at the same rate, it does not matter which process controls release of the dMtl Adt = n Cd K t n -1 (Eqn. 2) Table IV summanzes the range of values of the diffusional exponent n, and the corresponding release mechanism.
The values of K, n and correlation coefficient (r2) obtained from various formulation of PEO lOOK and PEO 5M are given in Table V.
As shown in Table V,  Release profiles of three solutes from the matrices containing PEO lOOK or PEO 5M (50% solute/50% PEO) are shown in Figure 13 and Figure 14, respectively. As shown in Figure 10, drug-containing PEO lOOK tablets exhibited a linear release profile for approximately two hours.
Although the three solutes were released at different rates, these rates did not vary widely. The diffusion coefficient as predicted from the tracer molecular weights are quite disparate, leading to the conclusion that the observed similarity in tracer release rates is probably due to matrix erosion-controlled drug release. This would mean that the gel layer thickness is very small so that most of the mass of drug particles is released into the external medium before drug is dissolved. This is probably due to the dissolution rate was essentially controlled by the erosion of PEO lOOK matrix. Figure 14 shows that the release of PPA.HCl (smaller molecular size, high diffusivity) from PEO SM had the faster release rate while theophylline (smaller molecular size, lower solubility) exhibited an intermediate release rate, the release rate of theophylline from PEO SM was nearly constant (zero order) in the first 20 hours. The release of BSA (larger molecular size) from PEO SM appeared to be more controlled by its low diffusivity in the gel layer formed as the tablet swelled.
As shown in Table V, the kinetic constant for release, K, which incorporated the overall solute diffusion coefficient and geometric characteristic of the system correlated inversely with the solute molecular weight. K also increased with increasing total solubility of the matrix system. The K values of all three solutes in a PEO lOOK matrix are greater than those in a PEO SM matrix.

Effect of The Tracer/PEO Ratio on The Dissolution Rate
The effect of relative amount of tracer in the PEO formulation on the dissolution rate is shown in Figure 1  wt% PEO lOOK exhibited faster release than the SO wt% BSA/SO wt% PEO lOOK, this is due to the dissolution of PEO lOOK is faster than the dissolution of BSA. The dissolution rate of the system is essentially controlled by erosion of PEO lOOK matrix. However, as shown in ( Figure 16, tablets composed of 50 wt% BSA and 50 wt% PEO 5M had faster release rate than the 4 wt% BSA/96 wt% PEO 5M. This is attributed to the fact that PEO 5M matrix swells more extensively and the release of BSA is mostly controlled by its low diffusivity in the gel layer. As the relative amount of PEO 5M in the system was increased, the resistance of the gel layer to diffusion of drug was also increased.

CONCLUSIONS
These findings conclude that molecular weight is an important determinant of PEO dissolution rate, which was inversely proportional to the molecular weight of PEO and the release of drug from PEO matrix system follow some anomalous behavior where both diffusion and mechanical relaxation affect the whole process. In the case of theophylline, the release rates are nearly zero order. If a mixture of different molecular weight PEO is chosen carefully, it is quite possible to balance the reduction in resistance to diffusion of the drug, leading to drug/PEO systems which exhibit constant release rate.

One of the authors (H. R. Chueh) wishes to thank Pfizer Central
Research for the award of a summer fellowship and the opportunity of using the facilities and chemicals.
He also would like to acknowledge the support and expert advice of many Pfizer scientists, especially Dr. Richard W. Korsmeyer whose careful supervision and considerable input made this project possible. ( In vitro programmable zero-order release drug delivery system, Acta Pharm. Technol., 33 (1987)      ;: .c  In recent years, many attempts have been made to provide therapeutic dosage form .which will provide longer transit time and more efficient absorption for specific drugs which have a window effect of absorption or stability problem. The floating dosage form was designed to possess sufficient buoyancy to float on the top of stomach content and prolong the stomach residence time of the dosage form (3)(4)(5)(6)(7)(8).
Meanwhile, significant interest also has been shown in the development of oral bioadhesive systems to adhere the oral dosage form to mucosa wall of stomach or intestine to increase the residence of the drug in the GI tract (9)(10)(11)(12).
The floating and bioadhesive drug delivery systems are meant to provide the following advantages (1) increased and more effective absorption for drugs which have specific absorption sites (2) increased contact time for local activity in the stomach where such is required and (3) the ability to limit the number of dosages.
The floating dosage form is meant to remain buoyant on the gastric fluid when the stomach is full after a meal, however, as the stomach empties and the tablet is at the end of the stomach the buoyancy of the dosage form might be impeded (13). It will become increasingly possible that the dosage form will pass through the pylorus into the small intestine. Thus, the buoyant ability of a floating drug delivery system in the stomach could be limited to only three or four hours. In bioadhesive drug delivery system, it is quite likely that the system becomes dislodged from the stomach mucosa wall when the stomach is full and semi-liquid contents are churning around under the influence of peristaltic movement. Also, most of currently available oral floating and bioadhesive systems are made by wet granulation tabletting process and some other tedious and costly procedures.
In light of the above reasons and conditions, the objective of this work was to develop a novel sustained-release tablet made by direct compression process. The tablet possesses an unique combination to prolong the stomach residence time of sotalol HCl, a beta-blocker, which has high aqueous solubility and its absorption from GI tract is limited to the upper part of the small intestine.
In this present study, a computer optimization process utilizing a statistical Box-Wilson design experimental design (14,15) was employed to develop bioadhesive and floating tablet formulations and determine the effects of formulation variables on the response properties of tablets. Finally, an optimum tablet formulation was selected using the technique of response surface methodology.

EXPERIMENTAL Experimental Design
The two formulation variables and their ranges selected for optimization study were summarized in Table I  In the tablet formulation, NaCMC and HPMC were used as bioadhesive agents. When the tablet is in contact with gastric fluid, a combination of NaCMC and HPMC will also possess sufficient structure to form a gel layer and achieve an overall specific gravity formulations were compressed m a random order.

Tablet Evaluation
In Vitro Dissolution---Dissolution studies were conducted usmg the USP basket method. Six tablets were tested for each batch. The dissolution medium was 900 ml of 0.1 N HCl solution (pH 1.2) 132 equilibrated at 37 °C and stirred at 70 rpm. The samples (3 ml) were withdrawn at 0. 5, 1, 2, 4, 6, 8, 12, 20 and 24 hours, respectively. The dissolution medium volume was kept constant by adding the same volume of fresh dissolution medium kept at the temperature of 37 °C.
The dissolution samples were diluted and the concentration were determined on a Diode Array Spectrophotometer at the wavelength of 228 nm corresponding to the maximum absorbance of sotalol HCI.

Floatin~
Capability---The lag time required for the tablet to start floating on the top of basket in the dissolution study was measured.
The duration of floatation under the rotating condition of the dissolution study was also determined for all formulations.
Measurement of Bioadhesiveness The sets of data obtaining from the statistical analysis were then subjected to computerized regression analysis to determine the fit to a second-order model. These regression models include an intercept and main effect terms of each independent variable, two-way interaction terms and second order effect terms as shown in Table V.

RESULTS AND DISCUSSION
The dissolution profiles for the tablets are shown m Figure 2 to is an exponent characterizing the mechanism of release of the drugs (16).
Mt/Moo =Kt n (Eqn. 1)  The general quadratic surface model was applied to generate contour plots for each of response parameters. Figure 6 shows the effect of two formulation variables, NaCMC/HPMC and EC/polyplasdone XL, on tablet dissolution characteristic (diffusional exponent value, n). It indicates that tablets made with 105 mg NaCMC, 115 mg HPMC, 90 mg EC and 30 mg polyplasdone XL obtained the highest n value from dissolution profile which indicates sotalol HCl was released by non-Fickian behavior, a near zero-order release, and the release was partially controlled by viscoelastic relaxation of the matrix system during solvent penetration. This increase in the values of n may be attributed to the stronger hydrogen bonding between the carboxyl group on NaCMC and hydroxyl group on the nomomc gum, HPMC, leading to stronger cross-linking between the two gums. The formulation composed of higher amount of NaCMC or HPMC exhibited lower value of n which represents the drug was released by Fickian diffusion, a first-order release, the diffusional pathlength for the drug increases with time.
As shown in Figure 7, it illustrates that the detachment force required for separating tablet from stomach mucosa surface increased with increasing amount of NaCMC as the amount of EC in the formulation increased. This is due to the stronger bioadhesve capability provided by NaCMC. Tablets made with 180 mg NaCMC, 40 mg HPMC, 90 mg EC and 30 mg polypladone XL will possess the best bioadhesive power. Figure 8 demonstrates the effect of two formulation variables on the required shear force. Again, as the amount of EC and NaCMC in the formulation increased, the required shear force for sliding tablet away from stomach mucosa was also increased.
In Figure 9, the required compress10n force increased with increasing amount of NaCMC in the formulation as amount of EC increased. This is attributed to the poorer compactability of NaCMC The dissolution release characteristic represented by the diffusional exponent value, n, was indentified as the pnmary response parameter because a zero-order release was desired for this extended sotalol tablet formulation. In addition, the in vitro dissolution usually provides an indication of in vivo bioavailability.
The diffusional exponent n value was maximized so as to obtain a near zero-order release characteristic. As shown in Table XV, two constraints were applied in obtaining the highest n value, the required compression force was constrained under 14 KN and the shear force was required to be more than 1.1 N. Additional constraints were the experimental range limits placed on values of two independent variables. The optimum fromulation satisfied all constraints simultaneously and provided an optimum value for the primary concern, the highest n value.
The tablets were prepared on an instrumented B-2 rotary press according to the optimum formulation as shown in Table XIV, tablets properties were also determined. The comparison of predicted and experimental values for optimum formulation showed very good agreement and are shown in Table XVI. This reasonable prediction of the system's performance indicates the proposed model is valid.                       ., properties of a bioadhesive system is the most direct way to quantify the bioadhesive properties. The tensile, shear and peel stress are more commonly used to quantify the adhesive force of contact joints.
In tensile and shear loading, the stress is distributed uniformly over the entire joint. However, in peel loading, the stress is limited to a very fine line at the edge of the joint (2).
Several in vitro techniques have been reported to determine the bioadhesion properties of bioadhesive oral dosage forms (3). The majority of these methods measure the tensile stress between the dosage form and the membranes or biological tissues ( 4-7).
However, the tensile stress provides only a partial reflection of mucoadhesion, since mucosa surface has some elements of a shear motion (5).
In light of the above reason, an alternative technique was developed in this study to quantify the bioadhesiveness of selected oral dosage forms by measuring both detachment force and frictional force required to separate two paralell sufaces (tablet and membrane).
Among vanous available bioadhesive polymers, sodium carboxymethylcellulose, polycarbophil and carbopol 974P are more commonly used in the oral bioadhesive dosage forms for both stronger bioadhesive power and lower toxicity reasons (8). In this study, the custom-designed apparatus was also utilized to classify tablets made with these three bioadhesive polymers in terms of detachment force, shear force as well as adhesion work.

Preparation of the Bioadhesive Tablets
Tablets free of drug were prepared in duplicate manner by mixing microcrystalline cellulose (Avicel PH 101, FMC Co., Lot 14361) and sodium carboxymethylcellulose (NaCMC 7MF, Aqualon Co., Lot 67108), at four different propotions (12.5, 25, 50 and 75%) in a Turbular Mixer for 15 minutes, then compressing into tablets in a Carver press. The final tablet has a weight of 500 mg and a hardness of 4.5 kg. Tablets were compressed m a B-2 rotary press with a weight of 662 mg and a hardness of 6 Kg.

Biological Tissues
The biological tissue used was rabbit stomach mucosa. They were Lejoyeux ~. (9) reported that this last parameter gives more interesting information concerning bioadhesion than the simple maximum detachment force. They also compared the adhesive capability of pure poly(acrylic acid) (PAA) tablets and pure hydroxypropylmethylcellulose (HPMC) tablets to bovine sublingual mucosa in liquid medium containing 100 g/l NaCl.
It showed there is no difference in detachment force measurement, however, in terms of adhesion work, PAA tablets were almost three time greater than HPMC tablets.
The maximum shear force and adhesion work values measured at two different days with NaCMC tablets were summarized in Table I.
No significant differences were observed in these two measurements for both parameters indicating the good reproducibility of this adhesion assessment apparatus. Figure 3 shows a linear relationship which was obtained when adhesion forces were plotted against polymer content for NaCMC tablets. As shown in Figure 4, a linear correlation also exists between the adhesion force and adhesion work for NaCMC tablets. These results show good agreement with some previous findings observed by Ishida et al.(10). They indicated that within the range of 0 to 30 % PAA, there was a linear relationship between the adhesive properties of white or hydrophilic petrolatum ointment and PAA contents. Hassan et al.(11) also showed that Nb values (viscosity component due to bioadhesion) was proportional to the PAA concentration m the bioadhesive system. Leung and Robinson (12) observed that the tensile stress of the PAA-mucin interaction decreased as the percent composition of acrylic acid decreased.
Ponchel SiJ!l. (13) also reported a direct correlation between the work of adhesion and the quantity of the bioadhesive polymer, PAA, in the tablets. Meanwhile, Park (14) showed that the mucoadhesive property of copolymers of acrylic and acrylamide increased sharply until the acrylic acid content reached 70 %.
It is well known that the bioadhesiveness of certain polymers is very much dependent upon their ability to take up water from the medium immersed in, and thus become sticky and adhesive. The designed instrument would definitely satisfy the wetting condition necessary for such evaluations. Table II shows the comparison of work (energy) measured in dry and wet conditions. As is evident from this Table,      ...