Azithromycin

Azithromycin (AZI) is a semisynthetic, acid-stable erythromycin derivative with a broad spectrum of activity and improved tissue pharmacokinetic characteristics compared with erythromycin (1). Perorally administered immediate-release formulations of AZI are absorbed rapidly and pharmacokinetically exhibit a two-compartment model (therapeutic range 0.1-0.4 mg/mL) with a relatively long elimination half-life (~60 h) (2). AZI therapy with immediate-release formulations is associated with various adverse effects such as nausea, vomiting, abdominal cramping, headache, and dizziness. Conventional immediate-release formulations have a more-pronounced incidence of adverse effects as result of higher peak plasma concentrations of AZI (1). One approach to eliminating the peak plasma concentration of AZI is by collapsing the two-compartment behavior of AZI into one compartment and developing a controlled-release drug delivery system (CRDDS). The CRDDS releases AZI at a predetermined rate, eliminating the undesired peak plasma concentration; reducing or eliminating unwanted side effects; and resulting in a better dosage regimen.

Extrusion-spheronization generates multiunit matrix-based particulate systems that produce multiparticulates with spherical shape, good flow properties, low friability, and uniform packing characteristics. Pellets can accommodate high drug loads, and modified drug release can be achieved by incorporating release-modifying agents such as ethylcellulose, acrylic polymers, chitosan, and glyceryl monostearate (3-5). Moreover, pellets can be compressed to single-unit dosage forms or coated to deliver a drug at a desired rate at the same site or at various sites within the gastrointestinal tract (GIT) (6).

Extrusion-spheronization was used to develop a matrix-based controlled-release formulation of AZI based on pharmacokinetic principles. Various approaches to modifying the formulation composition or process were adopted to achieve the target drug-release profile and to test the feasibility of using various excipients for successful extrusion-spheronization. The pellets' particle size, size distribution, crushing strength, and density were evaluated, and the yield was calculated. Drug-release mechanisms and kinetics were elucidated, and the stability of a selected formulation under accelerated temperature and humidity conditions was evaluated.

Materials and methods

Materials. AZI (lot number MHZ1101011) was procured from Kopran Ltd. (Maharashtra, India). Carnauba wax, directly compressible lactose, and microcrystalline cellulose (MCC PH 101) were gift samples from Panacea Biotec Ltd. (New Delhi, India). Glyceryl monostearate, magnesium stearate, and polyvinyl pyrrolidone K30 (PVP) were purchased from a local market. Hydroxypropyl methylcellulose (HPMC) was gifted by Parke-Davis (Mumbai, India) Ltd. Methocel K4M and ethylcellulose were gifted by Colorcon Asia Pvt. Ltd. (Mumbai, India) and Dabur Research Foundation (New Delhi, India), respectively. All materials were used as received.

Theoretical design of the CRDDS. To calculate the dose and release rate of AZI, a two-compartment model was collapsed into a one-compartment model (7) using the reported pharmacokinetic properties of AZI (8). The desired maximum steady-state concentration was 0.390 mg/mL (2) and the time of drug delivery ([t.sub.del]) from the dosage forms was 14 h, which was selected on the basis of the average GI transit time. Assuming first-order release (basically exhibited by envisaged matrix system), a first-order release rate of (kr) 0.164/h was obtained.

The serum concentration versus time profile of AZI after a 500-mg intravenous administration to a healthy male population (average weight of 74 kg) is shown in Figure 1 (8). The absorption phase of the curve was generated using the required absorption rate constant of 0.164/h--that is, a first-order drug release rate from CRDDS (drug absorption becomes a function of release from the delivery system in the case of a controlled-release product) and pharmacokinetic principles (back feathering). From the generated single-dose serum concentration versus time profile of perorally administered AZI, the steady-state concentration was predicted using superposition (8, 9) and the dose was calculated to achieve the desired steady-state AZI concentration according to the following equation:

D[M.sub.desired] = (Target [concentration.sub.desired] / Target [concentration.sub.test]) x D[M.sub.test]

in which D[M.sub.desired] is the dose needed to load in the system to attain a desired peak steady-state concentration (Target [concentration.sub.desired]). D[M.sub.test] is a preliminary reasonable maintenance dose needed for the calculation test based on a conventional dose. Similarly, Target [concentration.sub.test] is a theoretical plasma concentration within the therapeutic limits initially required in the calculation.

[FIGURE 1 OMITTED]

The dose of AZI for a once-a-day CRDDS to provide a maximum desired steady-state level of 0.390 mg/mL was 506.4 mg (~500 mg), which was to be released at a controlled rate of 0.164/h according to first-order kinetics.

Preparation and characterization of pellets. Accurately weighed quantities of ingredients were thoroughly mixed in a polythene bag for 15 min (dry mixing), followed by the gradual addition of the required amount of freshly prepared binder solution to prepare a wet mass. The prepared wet mass then was extruded through a 1-mm diameter screen using a roller extruder (model 10, Caleva Process Solutions, Dorset, UK). The extrudates were spheronized at 2650 rpm in a spheronizer (model 120, Caleva) fitted with a cross-hatch cut stainless-steel friction plate (120-mm diameter, 3 x 3 [mm.sup.2] pitch, 1-mm depth). Spheronization time was varied from 2 to 5 min, depending on the formula. Prepared pellets were kept at room temperature for 12 h for drying before being packed in plastic bottles.

Preparation of tablets. Pellets were compressed using a single-punch tablet machine (Cadmach India Ltd., Ahmedabad, India) fitted with 12.75-mm round, flat-faced punches. For each tablet, 680 mg of pellets (equivalent to 250 mg of AZI) were weighed, filled into the die, and compressed. Tablet hardness was kept constant (17-19 kp) for all formulations. To study the effect of compression on the release rate, pellets were compressed at three applied pressures to generate tablets of high (23 kp), moderate (18 kp), and low hardness (12 kp). To study the effect of tablet dimension on the in vitro release of AZI, tablets of various diameter and thickness also were made.

In vitro drug release studies. Dissolution testing of various pellet formulations (pellets equivalent to 250-mg AZI) and tablet formulations was performed in a sodium phosphate buffer (pH 6.0, 900 mL) using the paddle method at 50 rpm and 37 [+ or -] 0.5 [degrees]C. At various time intervals, 2.0-mL samples were withdrawn, filtered (0.45-[micro]m syringe filter, Sartorius Corp., Goettingen, Germany), suitably diluted, and analyzed using high-performance liquid chromatography (HPLC) (10).

Release profiles comparison. To compare two formulations, dissolution data were fitted to the fit factors [f.sub.2] (similarity factor) and [f.sub.1] (dissimilarity factor) (11). The mean dissolution time (MDT) (12) also was calculated to compare the dissolution characteristics of the formulations.

Release kinetics and mechanism. To establish the drug-release kinetics and mechanism, drug-release data were fitted into various drug-release kinetic models, including zero-order, first-order, Higuchi, Hixon-Crowell, Baker-Lonsdale, Hopfenberg, Korsenmeyer, and Peppas. The criteria for selecting the most appropriate model were based on the highest coefficient of determination ([R.sup.2]), the smallest sum of squared residuals (SSQ), the standard error, and the akaike information criteria (AIC). The F-statistic was applied to determine whether correlations occurred by chance (13, 14).

Stability studies. An optimized formulation was isothermally stressed to study the stability under accelerated temperature and relative humidity (RH) conditions (40[degrees]C and 75% RH) for three months. Test samples withdrawn every month were subjected to various tests, including visual inspection for any appreciable change on the tablet surface, assay, hardness, friability, and dissolution. HPLC and differential scanning calorimetry were used as analytical tools to characterize the stability of AZI in tested formulation.

Results and discussion

Formulation optimization. Initially, a base formulation was prepared using a combination of MCC and AZI. MCC was incorporated into the base formulation for imparting moisture retaining and distribution ability to extrusion-spheronization (15). Various formulation and process modification approaches were used at various stages of formulation development to achieve the target release profile. The following describes these approaches and their effects on product quality. Ingredient compositions of the formulations are shown in Table I.

Effect of granulating fluid, binding agents. To assess the effect of binding agents on the physical properties and release rate, base formulations of AZI were prepared with various granulating liquids such as water and aqueous polymeric solutions of HPMC and PVP. Because of the water-holding capacity of the polymeric solutions, a less amount of granulating fluid (compared with water) was required to provide sufficient moisture content to the wet mass (see Table I, formulations ES01A and ES01 B). Moreover, polymeric solutions significantly improved mean pellet size and size distribution (see Table II). In the case of PVP, however, mean pellet size was greater compared with HPMC (633 versus 556 [micro]m). Examination of the bulk and tap density data indicated that the density of the pellets increased when water was replaced with a granulating agent (HPMC and PVP). Changes in density may be attributed to the formation of more-uniform spherical pellets with binding agents. Binding agents also enhanced the flow characteristics of the pellets, indicated by lower Carr's index and Hausner ratio values (see Table II).

Although using granulating fluids provided a significant advantage over using water with respect to the pellets' physical properties, little influence was observed in the respective release profiles (see Figure 2). Changing the binding agent from PVP to HPMC, however, decreased the amount of AZI released during the initial hour. Prolonged AZI release may have attributed to an increase in pellet hardness because pellets made with HPMC exhibited higher hardness compared with pellets made with PVP (16).

[FIGURE 2 OMITTED]

Effect of water-soluble excipients. Various concentrations of lactose, a water-soluble excipient, were incorporated into the base formulation to improve the quality of the pellets (see Table I). Including lactose in formulations ES02 and ES15 decreased the amount of granulation fluid to 23.01% w/w moisture content for 25% w/w lactose compared with 26.92% w/w for a formulation without lactose (see Table I). In formulation ES02 (15% w/w lactose), a higher moisture content (more than 25% w/w) from overwetting led to agglomeration. A decreased proportion of MCC (which can hold a large amount of freely mobile water in the wet stage) in a formulation reduces water-holding capacity (17, 18) and leads to fast agglomeration.

Incorporating lactose in the base formulation produced a large fraction of pellets in the lower size range, thereby reducing the mean pellet size from 633 (ES1B) to 534 [micro]m (ES15) (see Table II). The shift in the pellet-size distribution toward a lower size was attributed to the crystalline nature of the lactose, which lacked the cohesiveness and plasticity to form the pellets (19). Total yield and product yield remained unchanged. Formulation ES02 exhibited a large mean pellet size (793 [micro]m) because of uncontrolled agglomeration resulting from an overwetting of the wet mass. The Carr's index and Hausner ratio values decreased, which indicated better flow properties in the presence of lactose as a filler in the base formulation. An insignificant difference in the crushing strength of the pellets containing 15% and 25% w/w lactose indicated the pellets' strength was not affected by the presence of lactose (see Table II). In addition, incorporating water-soluble excipients increased the drug release (see Figure 3) because the high aqueous solubility of the lactose caused channels or cracks to form on the pellets' surface (20, 21)

[FIGURE 3 OMITTED]

Effect of water-insoluble, hydrophobic excipients. Using water-insoluble excipients retards the release rate (22, 23) from a matrix system because they impart hydrophobicity in the matrix. Various concentrations ethylcellulose, a water-insoluble excipient, were added to the base formulation to retard the release of AZI. Mean pellet size increased (ES03) and then decreased (ES04) when the concentration of ethylcellulose increased 0-10% and 10-25% w/w, respectively. Average pellet size was higher (720 [micro]m) at 10% concentration level because of the high degree of agglomeration resulting from the large volume of granulating liquid used in this batch. In formulation ES04, pellet fragmentation occurred at high concentrations of ethylcellulose and resulted in smaller size pellets compared with pellet sizes in ES01A. Although total and product yields were improved, spherical pellets were not produced because of the low spheronizing properties of ethylcellulose (21, 24). No significant changes were observed in bulk and tap density values, which was further substantiated by the observed Carr's index and Hausner ratio values (see Table II). AZI release was retarded as a function of an increase in concentration of ethylcellulose from 10 to 25% in base the formulation (see Figure 4). Because of the highly water soluble nature of AZI, however, an initial-burst effect was observed, which was further pronounced by the pellets' high surface area and pores created by the release of AZI (24, 25).

[FIGURE 4 OMITTED]

Effect of hydrophilic-swellable polymers. Hydrophilic swellable polymers, commonly used as matrix-forming material in CRDDS, control drug release by penetrating water into the matrix, hydrating and swelling the polymer, diffusing the dissolved drug, and the eroding the gelatinous polymeric layer (26, 27).

Preparing the wet mass with Methocel K4M was difficult because of its tendency to form a tacky mass, which affected the total yield from 52% (ES01A) to 33% (ES13) (see Table II). On the other hand, mean pellet size increased from 589 to 620 [micro]m as the concentration of Methocel K4M was increased from 10 to 15%. In addition, long dumbbell-shaped pellets were obtained instead of spherical pellets. The crushing strength of the pellets, however, improved to 15 N, which can be attributed to better cohesiveness provided by the polymer (28).

With respect to drug release, a 70% release of AZI within 1 h indicated the unsuitability of a hydrophilic swellable polymer as a retardant for highly water-soluble drugs. A synergetic effect of swelling property of Methocel K4M and a disintegrating property of MCC could be the reason for the immediate release of drug (28, 29).

Effect of waxy materials. Waxy materials are hydrophobic in nature and a potential release retardant for controlled-release matrix systems (30, 31). Two waxy materials (glyceryl monostearate and carnauba wax) with varying degrees of lipophilicity were used as formulation components on the assumption that they would make the matrix hydrophobic and retard drug release.

Glyceryl monostearate and carnauba wax were incorporated to obtain formulations ES08 and ES11 using two different processes. For ES08, AZI was mixed with a dry-powder form of glyceryl monostearate and then granulated with water. At a high concentration of glyceryl monostearate (25%), flaking occurred during spheronization, but this effect was reduced by initially melting the glyceryl monostearate in hot distilled water and then adding the drug to form the slurry before mixing with MCC (30). For ES11, AZI was dispersed in molten carnauba wax by continuous stirring and then dried. The dried mixtures were milled to obtain granules, which were mixed with other dry-powder excipients and granulated with an aqueous solution of polymeric binder (32).

Total yield from ES11 was very low (33.20%) compared with ES08 (48.90%) because of the loss of material during melt granulation (see Table II). Product yield (pellets in the size range of 355-710 [micro]m), however, was from ES11. In addition, the pellets' density and crushing strength decreased as carnauba wax was incorporated into the formulations. On the other hand, glyceryl monostearate resulted in better yield, crushing strength, and density (see Table II). These results suggest that carnauba wax had unfavorable effects on the spheronizing property of the powdered mix, findings also supported by poor pellet shape (rough, dumbbell-shape pellets). It can be attributed to the hydrophobicity and poor molding capacity of carnauba wax. Even in these formulations the initial burst effect was observed as shown in Figure 6 owing to the formation of a poor matrix (30).

[FIGURE 6 OMITTED]

Effect of polymer-excipient combinations. On the basis of the results obtained using various excipients and polymers, we concluded that only water-insoluble excipients and glyceryl monostearate retarded the release rate of AZI from CRDDS effectively, with an initial-burst release that could not be controlled by these excipients alone. Because using ethylcellulose and glyceryl monostearate resulted in good quality pellets, various combinations of ethylcellulose and glyceryl monostearate were made as retardants to overcome the problem of an initial burst.

Ethylcellulose and glyceryl monostearate were added in various proportions with an aqueous polymeric solution of HPMC as the granulating liquid. Using glyceryl monostearate in formulations ES05 and ES06 decreased the amount of granulation fluid required for extrusion-spheronization. As the concentration of glyceryl monostearate was increased from 25% (ES05) to 40% (ES06), the amount of granulating liquid required for extrusion decreased from 17.0 to 13.30 g. This decrease was attributed to the reduced amount of MCC in the formulation (glyceryl monostearate was incorporated at the expense of MCC). No significant change in the required amount of granulation liquid was observed, however, when the AZI concentration was increased from 25 (ES05) to 35% (ES07) and the ethylcellulose and glyceryl monostearate concentration was kept constant. This was attributed to the water-holding capacity of AZI, which was incorporated to compensate the amount of MCC in formulation (33). In contrast, the required amount of granulating liquid increased when magnesium stearate was used in place of glyceryl monostearate (ES 10). A possible reason for this was the presence of a higher concentration of ethylcellulose in ES10 compared with other formulations (ES05, ES06, and ES07), which has a higher moisture-holding capacity than glyceryl monostearate (34).

[FIGURE 5 OMITTED]

Results of evaluating pellets made with all combinations demonstrated that combinations of a water-insoluble excipient (ethylcellulose) and a waxy material (glyceryl monostearate) could easily be incorporated into the pellets by extrusion-spheronization. When excipients were used in various proportions, the total and product yield remain unchanged (see Table II). Mean pellet size decreased as the concentration of glyceryl monostearate increased, although most of the pellets fell in the size range of 355-710 [micro]m (see Table II). In contrast, pellet density and crushing strength decreased as glyceryl monostearate concentration increased (see Table II). The lack of glyceryl monostearate to provide sufficient cohesiveness to the wet mass may be the reason for the decrease in pellet size, density, and crushing strength (30).

For ES09, an aqueous dispersion of ethylcellulose was used as a granulating agent in place of the binder HPMC, expecting that the presence of a film-forming polymer in the matrix would help in the preparation of better quality pellets and in controlling the drug release (35). To increase the solids content in the formulation, a greater amount of aqueous dispersion was used. Total and product yields and increased significantly (see Table II) compared with other formulations. Pellet crushing strength also increased.

The release profile clearly shows that a combination of excipients produced a better matrix that had a greater retarding effect than a matrix made with ethylcellulose or glyceryl monostearate alone (ES03, ES08) (see Figure 7). A combination of both excipients increased the lipophilicity of the pellet matrix, which decreased the effective penetration and wettability of the dissolution medium. Consequently, the dissolution of the drug within the pellets was slower; hence the rate of AZI release was slower. An initial-burst release was observed in all formulations and could not be prevented, even after an aqueous dispersion of film-forming polymer was used as a granulating agent (ES09).

[FIGURE 7 OMITTED]

Pellet compression. On the basis of the pellet formulation results (see Table I) and the presence of an initial-burst release, we concluded that none of the approaches were successful to achieve the formulation with a release profile close to the theoretical calculated desired release profile. Two different approaches can be taken to prevent this initial burst: coating pellets with suitable release-retarding polymers or compressing the pellets. We decided to compress the pellets as an additional step in preparing a CRDDS for AZI using extrusion-spheronization. All pellet formulations in the 355-710 [micro]m size range were compressed with a single-punch tablet machine. MDT values confirmed that AZI release from these tablets was better controlled than corresponding release profiles obtained from pellets (see Table III). This improved control over the AZI release was attributed to the close packing of the pellets after compression and to the formation of a matrix structure having limited porosity. Compression into single-unit tablets also reduced the surface area exposed to the dissolution medium, which prevented the initial-burst release observed with the pellets.

Selecting a tablet formulation. To select a formulation having a release profile close to the desired profile, we determined the model-independent parameters [f.sub.2], [f.sub.1], and MDT and the first-order kinetic model parameters R, [R.sup.2], standard error of y-estimation (SE), AIC, SSQ, and F statistic from the AZI release profiles of various tablet formulations (14) (see Table III). Model-independent parameters showed that the [f.sub.2] values of the ES02, ES07, and ES15 tablets were greater than 50, compared with the desired-release profile. MDT values of these formulations also were close to the desired values. These results suggest similarity among the drug-release profiles from three tablet formulations and the desired-release profile of the AZI formulation. On fitting the dissolution data to the first-order kinetic equation, however, results of ES02 and ES07 tablets indicated very poor correlations, characterized by the values of [R.sub.2], SSQ, AIC, and the F-statistic. Only the ES15 formulation fit well to the first-order kinetic model and exhibited a drug-release rate (0.174/h) close to the desired release value (0.164/h) with good correlations (i.e., high R, [R.sup.2], and F statistic and low SE, SSQ, and AIC values) (see Figure 8). On the basis of these results, we selected the ES15 formulation for further development as first-order kinetic controlled-release tablets of AZI. The effect of pellet size and tablet dimensions on the drug-release characteristics were studied.

[FIGURE 8 OMITTED]

Effect of pellet size. The effect of pellet size on the tablets' drug-release properties was studied using various size fractions of pellets of ES15 (retained on 16 sieve, >1000 [micro]m; 16/22 sieve, 710-1000 [micro]m; and 22/44 sieve 355-710 [micro]m) to prepare tablets of 680 mg weight (see Table IV). For all tablets, [f.sub.2] < 50 and [f.sub.1] < 15, indicating the similarity of the release profiles (see Table IV). Release was independent of the pellet size used to prepare the tablets. During compression, some of the pellets were crushed, forming a uniform matrix (clearly seen in scanning electron micrographs). Thus, a release-retardation effect was achieved because of the close packing of the matrix, irrespective of pellet size. Similarity in terms of AZI-release properties of various-size fractions of pellets indicates that all size fractions (total yield) can be used for tablet preparation, which increases the output of the extrusion-spheronization process.

Effect of tablet structural integrity. AZI tablets from ES 15 were broken into two parts using a tablet hardness tester (Erweka, Germany). Tablets were kept between the arms of the hardness tester and a force was applied until fracture occurred, exposing radial fracture surfaces. Weights of the broken tablets were checked, and only those weighing [+ or -] 2% to the respective intact tablet weights were used for the dissolution studies. Halved tablets showed faster release during the initial 1-4 hours, compared with intact tablets of the same weight. Although the [f.sub.2] and [f.sub.1] values were 51.97 and 14.91, respectively, indicating a similarity of drug release, the halved tablets had a distinct, faster initial drug release. This release was attributed to the broken tablets' increase in exposed surface area as well as the fast dissolution of AZI that on the exposed surface of the tablets.

Effect of tablet dimensions. The effects of tablet weight, hardness, thickness, and diameter on the release rate of AZI were studied by preparing tablets (ES 15) of various dimensions (see Table IV). Pellets of ES15 were compressed using three different pressures (low, medium, and high) to produce 12.75 mm diameter tablets of low (12.8 kp), moderate (18.8 kp), and high (23.2 kp) crushing strength, with respective thicknesses of 4.92, 4.50, and 3.95 mm for 680-mg tablets. As expected, drug release decreased as tablet hardness and compression force increased (see Figure 9). It is possible that during the tableting process at a high compression force (> 17kp), some pellets undergo fragmentation and generate fines. Such fines fill the interstitial spaces between the intact pellets more effectively, which enhances the interparticulate binding between the pellets, thus forming a continuous matrix (35) (see Figure 10). Correlation of the SEM photos with the release results confirms that compressing pellets at higher pressures to produce tablets of closely packed pellets helps control the release of AZI.

[FIGURES 9-10 OMITTED]

To assess the effect of thickness on tablet weight, tablets of 4.50- and 5.90-mm thicknesses were produced by compressing 680- and 805-mg ES15 pellets to obtain tablets of the same hardness (17-19 kp) and diameter (12.75 mm). Drug-release profiles from these tablets showed that tablet thickness and weight were not the controlling factors for drug release (see Table IV). Tablets having higher thickness and weight (thus increased tablet surface area) and the same hardness showed similar release profiles. We also studied the effects of tablet diameter and weight on the release rate by making tablets of 10.05- and 12.75-mm diameter. The tablets were made by compressing 545- and 680mg ES15 pellets. Tablet hardness (17-19 kp) and thickness (4.4-4.5 mm) were kept constant. Drug release was similar for various tablet diameters of the same thickness (see Table IV).

On the basis of all of these results, we concluded that tablet dimensions did not play a significant role in controlling drug release from the AZI tablets prepared from ES15 pellets, especially when these pellets were compressed under high pressures (17-19 kp). The close packing of pellets in tablets was irrespective of tablet weight or dimensions.

Effect of media pH on drug release. Because orally administered formulations pass through various pH environments in the GIT, we performed dissolution studies in media having 1.2-6.8 pH. To simulate the environment of the GIT, a pH change dissolution method (i.e., changing the pH of the dissolution medium while running the dissolution test) was used (29). AZI release was independent the drug-release medium's pH. The [f.sub.2] value (theoretical release profile was taken as reference) was 65.36, indicating the similarity of dissolution profiles. The [f.sub.1] value, calculated to be 9.11, thereby indirectly indicating the similarity among the release profiles.

Release kinetics and mechanism elucidation. Dissolution data fitted into first-order release kinetic, Higuchi, and Baker-Lonsdale models indicated that drug release was controlled by a drug concentration--dependent diffusion process (see Table V). This result suggested that the tablets were heterogeneous granular matrix systems, which control the release by diffusion through channel or capillaries in the matrices. SEM photographs of the matrix taken 24 h after the dissolution experiment also showed that matrix was intact and that channels had formed throughout the matrix (see Figure 10). Dissolution data were also well fitted in Baker-Lonsdale model, indicating that discrete pellets were the release-controlling units. These results all indicate that retardation of drug release from the tablets could be attributed to the close packing of pellets. The close packing effectively retarded the penetration of dissolution medium into tablets and consequently prolonged the drug release from tablets, thereby circumventing the problem of the initial burst release that was observed with pellets.

Poor fitting of the dissolution data to the Hopfenberg model (confirmed by a low [R.sup.2] value) clearly ruled out the possibility of erosion being the release-controlling mechanism (37). This result was supported by the intact tablet structure remaining after 24 h. Fitting data to the Hixson-Crowell model could not be explained because the tablets' dimensions did not diminish during dissolution.

Drug-release data fit well in the Higuchi and first-order kinetic models. To further define diffusion kinetics, we analyzed dissolution data with the Peppas model. The value of the exponent n obtained after treating the data was 0.64, which suggested the drug release was non-Fickian anomalous diffusion (38).

Accelerated temperature stability studies. No changes in the DSC thermograms confirmed that the AZI tablets were unaffected after three-month storage under accelerated conditions (see Table VI). Transition temperature and heat of fusion values were close to the expected values. Moreover, there were no signs of visually distinguishable changes in appearance, texture, and color of formulation. The test product's AZI content and friability were comparable with those of the control samples and within limits ([+ or -] 10%). In all cases, [f.sub.1] <15 and [f.sub.2] >50, compared with these values for the control samples. On the basis of these results, we concluded that the formulation was stable under accelerated stability test conditions for three months.

Conclusion

Desired drug-release profiles for once-a-day, first-order kinetic CRDDS of AZI were developed on the basis of drug pharmacokinetics (collapsing principles) and a method of superposition. Preparation of a matrix systems for CRDD for AZI was successful only after the prepared pellets were compressed. These pellets were stable in under accelerated stability conditions for three months. The solubility of the drug and the pellets' large surface area resulted in an extensive burst release. In addition, a higher dose of AZI made it necessary to incorporate drug in higher proportion in the delivery system, which resulted in a higher fraction of drug available on the pellet surface contributing to a pronounced burst effect. Therefore, compression can effectively prevent an initial-burst effect. Compression pressure was a critical factor in avoiding an initial burst release. A high pressure (hardness > 17 kp) was necessary to obtain a drug release profile close to the desired release profile. The drug release data fitted to a first-order release kinetics model indicated that drug release was concentration dependent.

References

(1.) J.D. Christopher and L.B. Barradel, "Azithromycin: A Review of Its Pharmacological Properties and Use as Three-Day Therapy in Respiratory Tract Infection," Drugs 51,483-505 (1996).

(2.) W.A. Ritschel and G.L. Kearns, "Pharmacokinetic Parameters of Important Drugs," in Handbook of Basic Pharmacokinetics (American Pharmaceutical Association, Washington, DC, 1999), pp. 479-503.

(3.) S.R. Goskonda and S.M. Upadrashta, "Avicel RC-591/Chitosan Beads by Extrusion-Spheronisation Technology," Drug Dev. Ind. Pharm. 19 (8), 915-927 (1993).

(4.) D. Blanque, "Some Factors Influencing the Formation and In Vitro Drug Release from Matrix Pellets Prepared by Extrusion-Spheronization," Int. J. Pharm. 119 (2), 203-211 (1995).

(5.) S.H. Neau, M.Y. Chow, and M.J. Durrani, "Fabrication and Characterization of Extruded and Spheronized Beads Containing Carbopol 974P, NF Resin," Int. J. Pharm. 131 (1), 47-55 (1996).

(6.) A. Sood, Y. Ashokraj, and R. Panchagnula, "Use of Extrusion-Spheronization to Develop Controlled-Release Dosage Forms for Diltiazem Hydrochloride," Pharm. Technol. 28 (4), 62-83 (2004).

(7.) W.A. Ritschel, "Biopharmaceutics and Pharmacokinetic Aspects in the Design of Controlled-Release Peroral Drug Delivery Systems," Drug Dev. Ind. Pharm. 15 (6,7), 1073-1107 (1989).

(8.) G. Foulds, R.M. Shepard, and R.B. Johnson, "The Pharmacokinetics of Azithromycin in Human Serum and Tissues," J. Antimicrob. Chemother. 25 (suppl. A), 73-82 (1990).

(9.) W.A. Ritschel and G. Sathyan," Development and Pharmacokinetic Evaluation of Controlled-Release Dosage Form of Lithium Carbonate in Beagle Dogs" Lithium 1,209-214 (1990).

(10.) R. Gandhi, C.L. Kaul, and R. Panchagnula, "Validated HPLC Method for In Vitro Analysis of Azithromycin by Using Electrochemical Detection," J. Pharm. Biomed. Anal. 23 (6), 1073-1079 (2000).

(11.) J.W. Moore and H.H. Flanner, "Mathematical Comparison of Dissolution Profiles," Pharm. Technol. 20 (6), 64-74 (1996).

(12.) W.D. Lindner and B.C. Lippold, "Drug Release from Hydrocolloid Embeddings with High or Low Susceptibility to Hydrodyanamic Stress," Pharm. Res. 12 (11), 1781-1785 (1995).

(13.) A. Sood and R. Panchagnula, "Drug Release Evaluation of Diltiazem CR Preparations," Int. J. Pharm. 175 (1), 95-107 (1998).

(14.) P. Costa and J.M.S. Lobo, "Modeling and Comparison of Dissolution Profiles," Eur. J. Pharm. Sci. 13 (2), 123-133 (2001).

(15.) M. Law, D.B. Deasy, and J.P. McLaughlin, "Comparison of Two Commercial Brands of Microcrystalline Cellulose for Extrusion-Spheronization," J. Microencaps. 14 (6), 713-723 (1997).

(16.) J. Varshosaz, R.A. Kennedy, and E.M. Gipps, "Effect of Binder Level and Granulating Liquid on Phenylbutazone Pellets Prepared by Extrusion-Spheronization," Drug Dev. Ind. Pharm. 23 (6), 611-618 (1997).

(17.) K.E Fielden et al., "Thermal Studies on the Interaction of Water and Microcrystalline Cellulose," J. Pharm. Pharmacol. 40 (10), 674-678 (1988).

(18.) D. Bains, S.L. Boutell, and J.M. Newton, "The Influence of Moisture Content on the Preparation of Spherical Granules of Barium Sulphate and Microcrystalline Cellulose," Int. J. Pharm. 69 (3), 233-237 (1991).

(19.) P.J Harrison, J.M Newton, R.C. Rowe, "The Characterization of Wet Powder Masses Suitable for Extrusion-Spheronization," J. Pharm. Pharmacol. 37 (10), 686-691 (1985).

(20.) A.M. Juppo et al. "Compression of Lactose Glucose and Mannitol Granules," J. Pharm. Pharmacol. 47 (7), 543-549 (1995).

(21.) A. Sood, Y. Ashokraj, and R. Panchagnula, "Multiunit Matrix-Based Particulate System (MUMPS) for Controlled Delivery of Nifedipine: Formulation Development using Extrusion-Spheronization and In Vitro Evaluation," Pharm. Technol. 28 (11), 84-102 (2004).

(22.) S.R. Goskonda, G.A. Hileman, and S.M. Upadrashta, "Controlled Release Pellets by Extrusion-Spheronization," Int. J. Pharm. 111 (1), 89-97 (1994).

(23.) S.R. Goskonda, G.A. Hileman, and S.M. Upadrashta, "Development of Matrix Controlled Release Beads by Extrusion-Spheronization Technology Using a Statistical Screening Design," Drug Dev. Ind. Pharm. 20 (3), 279-292 (1994).

(24.) A. Sood and R. Panchagnula, "Development and In Vitro Evaluation of Matrix-Based CR Pellets for Diltiazem HCL Using Extrusion and Spheronization," J. Pharm. Pharmacol. 50 (suppl.), 181 (1998).

(25.) US Patent 6068859, "Controlled-Release Formulations of Azithromycin," Pfizer Inc., New York, NY (2000).

(26.) V. Iannuccelli et al., "Effect of Loading Method on the Drug Release from Cross-Linked Carboxymethylcellulose Beads," J. Controlled Release 23 (1), 13-20 (1993).

(27.) R. Chatlapalli and B.G. Rohera, "Physical Characterization of Hydroxypropyl methylcellulose and Hydroxyethylcelloluse and Investigation of Their Use as Pelletisation Aid," Int. J. Pharm. 161 (2), 179-193 (1998).

(28.) J. Herman et al., "The Dissolution Rate and Bioavailability of Hydrochlorthiazide in Pellet Formulations," J. Pharm. Pharmacol. 40 (3), 157-160 (1988).

(29.) P.J. Cox et al., "Development and Evaluation of Multiple-Unit Oral Sustained-Release Dosage Form for s(+)-Ibuprofen: Preparation and Release Kinetics," Int. J. Pharm. 193 (1), 73-84 (1999).

(30.) E.S. Ghali, G.H. Klinger, and J.B. Schwartz, "Thermal Treatment of Beads with Wax for Controlled Release," Drug Dev. Ind. Pharm. 15 (9), 1311-1328 (1989).

(31.) F. Zhou, C. Vervaet, and J.P. Remon, "Matrix Pellets Based on the Combination of Waxes, Starches, and Maltodextrins," Int. J. Pharm. 133 (1-2), 155-160 (1996).

(32.) K.K. Peh and K.H. Yuen, "Development and In Vitro Evaluation of a Novel Multiparticulate Matrix Controlled-Release Formulation of Theophylline," Drug Dev. Ind. Pharm. 21 (13), 1545-1555 (1995).

(33.) G.A. Stephenson et al., "Solid-State Investigations of Erythromycin Dihydrate: Structure, NMR Spectroscopy and Hygroscopicity," J. Pharm. Sci. 86 (11), 1239-1244 (1997).

(34.) P.R. Katikaneni et al., "Ethylcellulose Matrix Controlled Release Tablets of a Water-Soluble Drug," Int. J. Pharm. 123 (1), 119-125 (1995).

(35.) A.M. Dyer, K.A. Khan, and M.E. Aulton, "Effect of Polymer Loading on Drug Release from Film-Coated Ibuprofen Pellets Prepared by Extrusion-Spheronization," Drug Dev. Ind. Pharm. 21 (16), 1841-1858 (1995).

(36.) T.P. Foster and E.L. Parrott, "Effect of Processing on Release from an Inert, Heterogeneous Matrix," Drug Dev. Ind. Pharm. 16 (8), 1309-1324 (1990).

(37.) H.B. Hopfenberg, in Controlled Release Polymeric Formulations, D.R. Paul and F.W. Harris, Eds. (American Chemical Society, Washington, DC, 1976), pp. 26-31.

(38.) N.A. Peppas,"Analysis of Fickian and Non-Fickian Drug Release from Polymers," Pharm. Acta. Helv. 60, 110-111 (1985).

Rajesh Gandhi, Yasvanth Ashokraj, Chaman Lal Kaul, and Ramesh Panchagnula *

* To whom correspondence should be addressed

Rajesh Gandhi, PhD, is a senior research scientist in Ranbaxy Research Laboratories (Gurgaon, India), Yasvanth Ashokraj is a doctoral student, Chaman Lal Kaul, PhD, is the director of the National Institute of Pharmaceutical Education and Research (NIPER), and Ramesh Panchagnula, PhD, is a professor and head of the Department of Pharmaceutics, NIPER, Sector 67, S.A.S. Nagar (Punjab) 160062 India, tel. +91 172 2214682, fax +91 172 2214692, panchagnula@yahoo.com.

COPYRIGHT 2005 Advanstar Communications, Inc.
COPYRIGHT 2005 Gale Group