Effect of Fill Weight, Capsule Shell, and Sinker Design on the Dissolution Behavior of Capsule Formulations of a Weak Acid Drug
Candidate BMS-309403
ABSTRACT
Two strengths of BMS-309403 capsules were developed from a common stock granulation. Dissolution testing of the capsules was conducted utilizing the USP apparatus 2 (paddle) with a neutral pH dissolution medium. Unexpectedly, the lower- strength capsules exhibited slower dissolution than the higher-strength capsules filled with the same stock granulation. Higher variability was also observed for the lower- strength capsules. This was found to be mainly caused by a low fill weight in a relatively large size hard gelatin capsule shell. Instead of bursting open, some gelatin capsule shells softened and collapsed onto the granulation, which delayed the release of the active drug. The problem was aggravated by the use of coil sinkers which hindered the medium flow around the capsules. Switching from the gelatin capsule shells to the HPMC (hydroxypropyl methylcellulose) shells reversed the dissolution rate ranking between the two capsule strengths. However, both dissolved at a slower rate initially than the gelatin capsules due to the inherent dissolution rate of the HPMC shells at pH 6.8. Notably, the HPMC shells did not occlude the granulation as observed with the gelatin shells. The study demonstrated that the dissolution of capsule formulations in neutral pH media was significantly affected by the fill weight, sinker design, and capsule shell type. Careful selection of these parameters is essential to objectively evaluate the in vitro drug release.
Key Words: Dissolution; Capsule formulation; Gelatin capsule; HPMC capsule; BMS-309403.
INTRODUCTION
Drug product development time lines have been compressed significantly over the years. Innovative formulation approaches have been adopted to reduce the formulation development time, bulk drug substance requirement, and overall resources. One of the strategies for oral solid dosage form is to use simple hard gelatin capsule formulations to enable an early start of clinical studies. Typically, a common stock granulation or the pure drug substance is used to prepare multiple capsule strengths, and the same size/color capsule shell is used for all strengths to simplify the manufacturing process and the need for placebo/blinding.
BMS-309403, an investigational drug, is a poorly water soluble weak acid. Two strengths of capsules, 50 and 200 mg, were developed from a common stock granulation to support the early phase clinical studies. Dissolution testing of the capsules was performed to aid in the formulation development and to serve as a quality control tool. Some unexpected dissolution differences were observed between the two strengths of capsules. The aim of this study was to identify the source for this discrepancy and strategies to circumvent the problem.
The study findings are applicable to not only con- ventional capsule formulations but also liquid-filled semi-solid capsules and over-encapsulated compara- tor formulations.
MATERIALS AND METHODS
Materials
BMS-309403 was provided by Process R&D of Bristol– Myers Squibb Company in New Brunswick. The capsule formulation consisted of 55.6% w/w BMS- 309403, 20% w/w microcrystalline cellulose, 14.65% w/w lactose anhydrous, 5% w/w croscarmellose sodium (equally split between intragranular and extragranular), 2% w/w colloidal silicon dioxide, 2% w/w sodium lauryl sulfate, and 0.75% w/w magnesium stearate (equally split between intragranular and extragranular). The hard gelatin capsule shells were obtained from Capsugel, and the hard HPMC capsule shells from Shionogi Qualicaps. The coil sinkers (six spirals) and the three-prong magnetic sinkers were purchased from Zymark and from Vankel, respectively.
The dry blend of the above composition was den- sified by roller compaction (TF-Mini roller compactor from Vector Freund). The ribbons were sized through an 18-mesh screen and mixed with the extragranular disintegrant and lubricant. In this study, the finished stock granulation was manually filled into the hard shell capsules, so that no additional variables were introduced by the automatic capsule filling machines. The fill weights were 90 mg and 360 mg for the 50 mg and 200 mg strengths, respectively.
Dissolution Method
A Distek dissolution system (Model 2100A) with the USP apparatus 2 (paddle) was used for the disso- lution studies of BMS-309403 capsules. The dissolu- tion medium was 1000 mL 0.5% sodium lauryl sulfate in 0.1 M sodium phosphate buffer pH 6.8 maintained at 37°C. Unless mentioned otherwise, metal coil sinkers were used to hold the capsules at the bottom of the vessel. The paddle speed was 60 rpm. Sample analysis was carried out by an automated on-line UV system at a wavelength of 280 nm.
RESULTS AND DISCUSSION
Effect of Fill Weight and Gelatin Capsule Size
Two strengths of capsules, 50 mg and 200 mg, were needed to support Phase I clinical studies of BMS- 309403. A common stock granulation (55.6% active) was formulated for both 50 mg and 200 mg capsules, and the total fill weights were 90 mg and 360 mg, respectively. Both strengths were encapsulated in the same size 0 hard gelatin capsules to simplify the manufacturing process and requirement for placebo/blinding.
As a common practice in pharmaceutical develop- ment, only one dissolution method was developed for both capsule strengths of BMS-309403. The higher strength was assumed to represent the worst case scenario, since the final drug concentration was closer to the solubility in the dissolution medium.
The dissolution profiles of both strengths of BMS- 309403 capsules are shown in Figure 1a. The 50 mg capsules unexpectedly dissolved at a much slower rate than the 200 mg capsules. In addition, the results were more variable for the 50 mg capsules. Upon close visual examination, it was observed that the capsule shells of the 50 mg capsules had a tendency to soften and collapse inward during the first 10 min of the dissolution test. Subsequently, the collapsed gelatin shell occluded the granules and slowed down the drug release. This was, however, not observed for the 200 mg capsules, where the capsule shell readily burst open within the first 10 minutes.
Figure 1. Dissolution profiles of (a) 50 mg and 200 mg capsules in size 0 gelatin capsule shells with coil sinkers, (b) 111 mg capsules in size 0 gelatin capsule shells with coil sinkers, (c) 50 mg capsules in size 4 gelatin capsule shells with coil sinkers, and (d) 50 mg capsules in size 0 gelatin capsule shells with three-prong magnetic sinkers. The dissolution was performed in 1000 mL of 0.5% SLS in sodium phosphate buffer pH 6.8 at 37°C with a paddle speed of 60 rpm and on-line UV analysis at 280 nm. N= 6.
Since the capsules in this study were all filled manually, the granulation in both strengths of capsules was loose and not compacted. The only difference between the two strengths was the fill weight, or in other words, the fill volume in the capsule shell. Based on the bulk density of the BMS-309403 granulation, a 360 mg fill weight occupied 80% volume of the size 0 capsule body, whereas 90 mg only about 20%.
The dissolution medium, pH 6.8 phosphate buffer with 0.5% SLS, was selected based on the solubility of BMS-309403. This compound was a poorly water soluble carboxylic acid (pKa~ 5.4), and the solubility increased significantly at pH above pKa and in the presence of SLS.
This dissolution medium was, however, not optimal for the dissolution of the gelatin capsule shell. It has been reported that gelatin capsule shells dissolve at a slower rate in neutral pH media than in acidic media.[1] This is attributed to the fact that the isoelectric point of gelatin is around 5 – 7.[2] The presence of SLS in the dissolution medium (and a small amount in the capsule formulation) might also slow down the gelatin dissolution. However, this phenomenon has not been well studied, and addi- tional investigation is warranted. Nevertheless, it ap- peared that the high fill weight capsules were able to compensate for the slow dissolution of gelatin shell by the stronger disintegration of the capsule shell through bursting and rapid release of the granulation. In contrast, the low fill weight capsules did not exhibit a burst phase, and a significant portion of the granules was trapped within the softened gelatin.
Two additional capsule formulations using the same granulation were tested to confirm the hypothesis that the slow dissolution of the 50 mg capsules was primarily due to the low fill weight in a relatively large capsule shell. In the first formulation, a fill weight of 200 mg granulation (111 mg strength), about halfway between 90 mg and 360 mg, was filled into the size 0 capsule. In the second formulation, 90 mg granulation (50 mg strength) was filled into a size 4 capsule, which holds only 30% volume of a size 0 capsule. The 90 mg BMS-309403 granulation occupied 65% volume of the size 4 capsule body. The dissolution of these two formulations was performed under the identical condi- tions as described previously, and the results are shown in Figures 1b and 1c. Both formulations were found to release the active drug faster than the original 50 mg strength capsules in size 0 shells. Lower variability among the capsules was also observed for the two new formulations.
Due to the apparent dissolution problem with the gelatin capsule shells, hard HPMC capsule shells were examined as an alternative for the capsule formulation of BMS-309403. The same 50 mg and 200 mg capsules were filled into the size 0 HPMC shells, and the dissolution was
performed under identical conditions. The results are presented in Figure 2a. In this case, the dissolution rate of the 50 mg capsules was slightly faster than that of the 200 mg capsules. However, both strengths dissolved more slowly during the first 10 – 20 min than the gelatin capsules. Visually, the HPMC capsule shells were observed to swell and expand without leaking much granulation for the first 10 min. This is consistent with a recent finding that the dissolution of HPMC capsule shells slowed down significantly at pH> 5.8.[1] Notably in the current study, the HPMC shells did not collapse onto the granulation as was observed for the gelatin shells with the low-strength capsules.
Effect of Sinker Design
Various types of sinkers are available commercially. The metal coil sinkers were used in this study initially for the following two reasons. First, these sinkers conform to the USP description of ‘‘a small, loose piece of nonreac- tive material such as not more than a few turns of wire helix may be attached to dosage units that would other- wise float.’’[3] Second, these magnetic metal coil sinkers could also be easily retrieved when using the automated robotic dissolution systems.
Studies have shown that sinkers with different geometry design can significantly alter the dissolution profile depending on the capsule formulation.[4,5] It was suspected that the coil sinkers used in this study were also contributing to the slow and variable disso- lution of the lower-strength capsule in addition to the inherent problem of the gelatin shell and the low fill weight. The three-prong sinker is another common design available commercially. This type of longitudi- nal sinkers has been shown to cause less hindrance on the hydrodynamic than the helical type sinkers.[4,5] However, the three-prong sinkers typically do not have high enough density to sink the low fill weight cap- sules, such as the 50 mg BMS-309403 capsules in size 0 shells. The newly designed Vankel magnetic three- prong sinkers are dense enough to sink the low fill weight capsules, and they are also compatible with the automated robotic dissolution systems.
The magnetic three-prong sinkers were used to repeat the dissolution experiments of the 50 mg BMS- 309403 capsules in both gelatin and HPMC shells. As shown in Figures 1d and 2b, the dissolution improved significantly for both types of capsules. Occasionally, some gelatin capsule shells were still observed to soft- en and wrap around the granules. However, the occur- rence rate was much lower than when the coil sinkers were used. This experiment confirmed that the coil sinkers hindered the flow of the dissolution medium around the capsules, which contributed to the slower and more variable dissolution of the 50 mg BMS- 309403 capsules.
CONCLUSIONS
Two strengths of BMS-309403 capsules were developed from a common stock granulation. The lower-strength capsule exhibited an unexpected slow dissolution rate with higher variability among capsules. This was mainly attributed to the lower fill weight of the granulation. As a result, the gelatin capsule shell did not burst open but instead collapsed onto the granules and slowed down the drug release. This phenomenon was not observed when the HPMC capsule shells were used. However, the dissolution of the HPMC shells at pH 6.8 was inherently slower, leading to the delayed initial release of the drug for both capsule strengths. The use of the three-prong sinker instead of the coil sinker improved the dissolution rate and lowered the variability among capsules, probably due to less hindrance of the hydrodynamics around the capsules.
Although these study findings are not expected to be relevant to in vivo performance, they represent potential complications to formulation and method development for early clinical studies. Therefore, it is recommended to match the size of the gelatin capsule shell with the fill weight, especially when the pH of the dissolution medium is neutral.BMS309403 As an alternative, HPMC capsules may also be useful for low fill weight capsule formulations.