2018 Volume 66 Issue 5 Pages 554-561
The properties of wet mass, which indicate the progress of high shear granulation processes, usually have an effect on final product properties, such as tablet dissolution. The mixer torque rheometer (MTR) is a useful tool for quantitatively measuring the ‘kneading state’ of wet mass and detecting differences in granules. However, there have been no studies of the relationship between the MTR torque and the final product properties to date. In this study, we measured the MTR torque of wet granules at different kneading states, which were prepared by changing the granulation conditions. We then evaluated the relationship between the MTR torque and the dissolution rate of the final product properties. The amperage of the high shear granulator is usually monitored during granulation, but we could not detect a difference in the kneading state through the amperage. However, using MTR torque we were able to quantify the difference of the wet mass. Moreover, MTR torque showed a high correlation with dissolution, compared with the correlations with other intermediate properties, such as granules particle size and tablet hardness. These other properties are affected by following processes and are not properties that directly relate to the kneading state. Thus, MTR torque is a property of wet mass after granulation, and it can be used to directly evaluate differences of the kneading state, and as a result, dissolution. These results indicate the importance of controlling the kneading state, i.e., the progress of granulation, and the utility of MTR for detecting differences in wet mass.
Granulation is the process by which drug particles and fillers are combined to make particles into larger granules1,2) and add various functions. The purpose of granulation is to improve handling, blend uniformity, and tablet properties such as hardness, appearance, and dissolution rate, by modifying powder properties.3–8) For granulation processes, various technologies and methods have been established by using high shear granulators, fluidized bed granulators, twin-screw granulators, foam granulators, steam granulators, and dry granulators.9)
High shear granulation is a wet granulation method, which is widely used in the pharmaceutical industry because of its ability to handle difficult formulations, including high viscosity binders and fine cohesive powders.10) The high shear granulation process consists of three steps, i.e., a blending step to blend the materials, a granulation step to add binder solution, and a wet massing step. The granules are then dried in a dryer, such as a fluidized bed dryer or circulation dryer. Various studied regarding high shear granulation have been conducted, examining the mechanism of granulation, optimization of the manufacturing conditions,11) and scale up methods.12,13)
High shear granulation processes are important and can affect product properties, such as the dissolution profile, generation of related substances, and stability.14) Therefore, there has been considerable interest in establishing a relationship between the product properties and the potential critical material attributes (p-CMAs). The p-CMAs include granule properties, such as particle size and specific volume. However, in actual pharmaceutical development, there are some cases where it is not possible to explain product properties from already-known p-CMAs. This is because the p-CMAs are the properties of granules processed by drying and dry sieving process after completed granulation. The properties are affected by these processes and might not reflect differences in the wet massing step. However, the properties of wet mass usually have an effect on final product properties. Hence, there is a need for new CMAs, which account for differences of the wet mass directly and enable prediction of the properties of the final product.
One useful method is to evaluate the kneading state of wet mass obtained during the granulation process. A hand squeeze test is typically used to evaluate the kneading state of wet mass after high shear granulation process, because no methods have been established to quantify and control the kneading state. The amperage value and power consumed by the granulator are usually monitored and used as end point determination parameters; however, it is not always possible to detect the differences in the kneading state. In addition, these parameters depend on the individual equipment and manufacturing scale; hence, it is difficult to use the amperage value and consumed power as an absolute measure of the kneading state.
On this basis, we focused on the torque of wet mass measured with a mixer torque rheometer (Mixer Torque Rheometer 3, abbreviated as MTR-3, Caleva, U.K.). The mixer torque rheometer was proposed for use in the pharmaceutical industry by Rowe and Sadeghnejad.15) in 1987, and the equipment has been developed and improved over time. Mean torque, which is measured by MTR-3, is the torque of blending the wet mass by the blade, and can be used to directly evaluate the shear stress. Rowe and Sadeghnejad plotted the profile of MTR torque from kneading microcrystalline cellulose with various grades and manufacturers, and reported that the solid-liquid-air packing state of the wet mass during granulation varied according to the manufacturer, and MTR could be used to detect differences in the materials.15) There have been some other reports concerning applications of MTR torque, such as the cross interaction of binder and other materials,16) the relationship between the torque profile and extruder,17,18) and scale up methods.19) Because MTR torque can accurately detect differences of wet granules, we believe that it could be a useful tool for quantifying the kneading state, i.e., progress of granulation, and to evaluate the effects on the properties of final products. However, there have yet to be any reports of a relationship between MTR torque and final product properties, such as dissolution.
Here, we measured the MTR torque of wet granules at different kneading states, which were prepared by changing the granulation time, amount of binder solution, and blade rotation speed in the high shear granulation process. We evaluated the relationship between the MTR torque and dissolution rate, as a final product property. We also evaluated the relationship between dissolution and the powder properties, such as particle size and specific volume, and tablet properties, such as hardness and disintegration time.
Compound A was used as a model active pharmaceutical ingredient (API). Physical properties of compound A is described in Table 1. The solubility in water of compound A is 0.73 mg/mL, and the solubility to acidic solution is almost same. Median diameter (D50) of compound A is approx. 180 µm. Lactose monohydrate (Pharmatose® 200M, DFE Pharma, Germany) and microcrystalline cellulose (Ceolus® PH-101, Asahi Kasei Co., Ltd., Tokyo, Japan) were used as fillers, hydroxypropyl cellulose (HPC-SL, Nippon Soda Co., Ltd., Tokyo, Japan) was used as a binder, and crosscarmellose sodium (Kikkolate® ND-2HS, Asahi Kasei Co., Ltd.) was used as a disintegrant. Magnesium stearate (Merck KGaA, Germany) was used as a lubricant.
| Molecular weight | Approx. 340 |
|---|---|
| Solubility to water | 0.73 g/mL |
| pKa | 6.0, 8.5 |
| Particle size D50 (Median diameter) | Approx. 180 µm |
| Loose density | 0.75 g/mL |
Tablet formulation and the batch formula are shown in Table 2. Compound A, lactose monohydrate, microcrystalline cellulose, hydroxypropyl cellulose and crosscarmellose sodium were added into a high shear granulator (VG-05, Powrex Co., Ltd., Hyogo, Japan, abbreviated as HSG) and granulation was conducted by adding water. Granulation conditions were varied according to the experimental design, described in Table 3. We used a 23 full factorial design in order to vary the experimental factors systematically. Granulation time (X1), water amount (X2), and blade speed (X3) were selected as objective variables. It has already been reported that blade rotation speed and mixing time can affect the granulation progress in high shear granulator,20–22) therefore blade rotation speed and granulation time were selected as variable parameters in this study. In addition, Rowe and Sadeghnejad reported that the solid-liquid-air packing state of the wet mass is changed by the amount of water, and it has influence on the granules particle size,15) so water amount to be added during granulation was also varied as other variable parameter. Other processing parameters were fixed in all batches. The pre-blending time in HSG before adding the water was 3 min, water addition time was 2 min, and the chopper speed was 3000 rpm. Generally, the amperage value and/or consumed power of motor to drive the blade are recorded during granulation as a guide to monitor the progress of granulation. In this high shear granulator, only amperage value could be recorded, so amperage value of motor to drive the blade was recorded. Amperage value at the end of pre-blending step were almost same for all batches, and increased during water addition. However, it did not change during wet massing. Therefore, the average of amperage value throughout granulation depends on the final point value, so the value at the end point was taken to be a representative value for each batch. After granulation, kneaded granules were dried with a circulation dryer (PH-120S-OM, Matsui MFG. Co., Ltd., Osaka, Japan) and sieved with a 1000-µm opening sieve.
| Component | Purpose | Formulation [mg/tablet] | Batch formula [g] |
|---|---|---|---|
| Compound A | Active agent | 50 | 250 |
| Lactose monohydrate | Filler | 100 | 500 |
| Microcrystalline cellulose | Filler | 28 | 140 |
| Hydroxypropyl cellulose | Binder | 10 | 50 |
| Crosscarmellose sodium | Disintegrant | 10 | 50 |
| Granules sub-total | 198 | 990 | |
| Magnesium stearate | Lubricant | 2 | 10 |
| Tablet total | 200 | 1000 |
| Batch No. | X1 | Granulation time [min] | X2 | Water amount [g] | X3 | Blade speed [rpm] |
|---|---|---|---|---|---|---|
| 1 | −1 | 1 | −1 | 300 | −1 | 200 |
| 2 | 1 | 5 | −1 | 300 | −1 | 200 |
| 3 | −1 | 1 | 1 | 500 | −1 | 200 |
| 4 | 1 | 5 | 1 | 500 | −1 | 200 |
| 5 | −1 | 1 | −1 | 300 | 1 | 600 |
| 6 | 1 | 5 | −1 | 300 | 1 | 600 |
| 7 | −1 | 1 | 1 | 500 | 1 | 600 |
| 8 | 1 | 5 | 1 | 500 | 1 | 600 |
| 9 | 0 | 3 | 0 | 400 | 0 | 400 |
| 10 | 0 | 3 | 0 | 400 | 0 | 400 |
| 11 | 0 | 3 | 0 | 400 | 0 | 400 |
Granules and magnesium stearate were blended and tablets were prepared using a single shot compression machine (Autograph AGS-20kNG, Shimadzu Corporation, Kyoto, Japan) with 8 mmϕ flat-faced round punches. The tablet weights were set to be 200 mg and the compression force were adjusted to achieve a tablet thickness of 2.9 mm.
Measurement of MTR TorqueMTR torque was measured with a mixer torque rheometer (Mixer Torque Rheometer 3, abbreviated as MTR-3, Caleva, U.K.). Figure 1 shows a photograph of the MTR-3. Wet granules were added to the mixing bowl and followed by a rotating main and auxiliary blade. The auxiliary blade rotated two times as fast as the main blade, and a strain gauge measured the forces acting on the auxiliary blade. The measuring procedure was as follows; step 1: calibration step; torque of the rotating empty bowl was measured, step 2: logging step; the wet mass, which was granulated by high shear granulator, was blended for T1 seconds to ensure a uniform wet mass and the average of the torque measured for the following T2 seconds was taken to be the mean torque. In this study, the rotation speed of the main blade was set to be 50 rpm, and 35 g of wet mass was used for measuring the torque. T1 and T2 were set to be 30 and 20 s, respectively. MTR torque measurement was conducted for three times.
The particle size distribution was measured with a Robot Sifter RSP-205 (Seishin Enterprise Co., Ltd., Tokyo, Japan). Particle size distribution was measured for one-time. The median diameter, D50 was used as a representative value. The definition of σg was as follows;
 |
A 100-mL cylinder container was filled with granules and the top of the granules was leveled off. The loose specific volume was calculated as the ratio of the volume to the weight. After the granules were added to the top container, which was mounted on the bottom container, the two containers were fixed with a vibrator and tapped for 3 min. The tapped specific volume was calculated as the ratio of the volume to the weight after tapping. Measurement of specific volume was conducted for one-time. The compressibility was calculated by the following calculation;
 |
The granules were heaped and the angle of the slope of the bed was measured with a protractor. Angle of repose was measured for three times, and the average value was used for evaluation.
Measurement of the Tablet HardnessTablet hardness was determined by diametrical compression tests, which were performed with a tablet hardness tester (Model 6D, Dr. Schleuniger Pharmaton AG, Solothum, Switzerland). Tablet hardness was measured for three times, and the average value was used for evaluation.
Measurement of the Disintegration TimeDisintegration time was measured according to Japanese Pharmacopoeia 17th Edition (JP 17th), using the disintegration tester (NT-400, Toyama Sangyo Co., Ltd., Osaka, Japan). Distilled water at 37±0.5°C was used as a medium. Disintegration time was measured using three tablets, and the average value was used for evaluation.
Measurement of the Dissolution ProfileDissolution profile of compound A from the tablets was measured by a paddle method using the dissolution tester (NTR-6100 A, Toyama Sangyo Co., Ltd., Osaka, Japan), according to JP 17th. A 900-mL portion of JP 1st solution was used as dissolution media, and the paddle rotation speed was 50 rpm. Dissolution test was conducted using six vessels. The solution was sampled for up to 60 min at 5-min intervals, and the absorbance was measured at 281 nm with a spectrometer (UV-1800, Shimadzu Corporation, Kyoto, Japan). All values were reported as the mean of six vessels.
Statistical AnalysisPrinciple component analysis (PCA) and ANOVA were performed with Unscrambler® X (CAMO Software AS, Oslo, Norway).
The amperage value of the HSG at the end point of granulation and the properties of the wet mass, granules, and tablets are shown in Table 4, and Table 5 shows the particle size distribution of granules. Table 6 shows the results of ANOVA of the process parameters and intermediate properties. Blade of high shear granulator is used to blend the wet granules in granulator, and amperage value is parameter of load in this time. MTR torque is the shear stress to blend the wet mass by the blade in the MTR, so these two parameters measure similar properties of wet mass seemingly. In addition, as another p-CMAs, which can affect dissolution, the particle size distribution (D50 and σg), loose specific volume, compressibility, and angle of repose, together with the hardness and disintegration times of the tablets were also evaluated. Granules particle size affects particle size after disintegration from tablets, so it has effect on total surface area.23) Therefore, granules particle size seemed to have an effect on the dissolution from tablets. Specific volume, compressibility, and angle of repose are parameters affecting the flowability of granules, and they can also have an effect on the compactability of the tablets. Generally, dissolution of API from tablets has a relationship with the disintegration of the tablets to primary granules and dissolution from the primary granules.23) The disintegration time reflects the time for disintegration of the tablets to small granules, and the tablet hardness has a negative correlation with disintegration time. These tablet properties were selected because of their effects on dissolution.
| Batch No. | Amperage value of HSG [A]a) | Wet mass | Granules | Tablets | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| MTR Torque [Nm] | D50 [µm] | σg [—] | Loose specific volume [mL/g] | Compressibility [%] | Angle of repose [˚] | Hardness [N] | Disintegration time [s] | |||
| Ave. | SD | |||||||||
| 1 | 4.1 | 0.114 | 0.007 | 219.3 | 2.0 | 2.02 | 19.38 | 41 | 33.7 | 63.79 |
| 2 | 4.1 | 0.116 | 0.005 | 203.4 | 2.0 | 1.86 | 16.17 | 41 | 42.3 | 69.51 |
| 3 | 4.2 | 0.447 | 0.029 | 537.1 | 2.3 | 2.38 | 17.00 | 35 | 66.7 | 185.21 |
| 4 | 4.1 | 0.677 | 0.009 | 666.0 | 3.6 | —b) | —b) | —b) | 70.3 | 209.31 |
| 5 | 3.6 | 0.178 | 0.002 | 190.4 | 1.8 | 2.07 | 21.11 | 41 | 34.0 | 91.59 |
| 6 | 3.4 | 0.103 | 0.005 | 203.7 | 1.6 | 1.81 | 18.10 | 43 | 27.7 | 81.03 |
| 7 | 4.5 | 0.707 | 0.011 | 590.2 | 3.0 | —b) | —b) | —b) | 70.0 | 173.88 |
| 8 | 4.9 | 0.756 | 0.052 | 466.6 | 2.0 | 1.79 | 13.85 | 39 | 69.0 | 157.18 |
| 9 | 3.5 | 0.376 | 0.035 | 320.7 | 1.6 | 1.91 | 14.80 | 42 | 71.3 | 391.48 |
| 10 | 3.7 | 0.419 | 0.008 | 322.8 | 1.8 | 1.85 | 13.42 | 43 | 71.7 | 535.43 |
| 11 | 3.6 | 0.416 | 0.010 | 367.2 | 1.8 | 1.75 | 10.38 | 42 | 88.0 | 477.31 |
a) Amperage value, which HSG indicated at the end point of the granulation process. b) Compressibility and angle of repose could not be measured owing to the lack of yield in these batches.
| Sieve size (µm) | Batch No. | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | |
| 710 on | 5.46 | 3.97 | 24.30 | 44.48 | 0.99 | 0.59 | 34.78 | 12.52 | 3.83 | 6.00 | 6.95 |
| 500 on | 7.49 | 6.56 | 31.23 | 26.38 | 2.97 | 2.58 | 26.68 | 32.93 | 12.33 | 13.00 | 19.83 |
| 355 on | 10.73 | 10.74 | 19.76 | 13.58 | 8.54 | 8.53 | 15.61 | 19.80 | 22.82 | 21.20 | 25.36 |
| 250 on | 18.42 | 16.70 | 13.04 | 7.09 | 19.44 | 22.03 | 10.28 | 18.58 | 33.74 | 32.00 | 30.88 |
| 180 on | 18.02 | 18.09 | 5.54 | 2.95 | 21.23 | 24.60 | 4.54 | 8.49 | 16.77 | 16.00 | 11.25 |
| 150 on | 9.31 | 10.54 | 1.97 | 1.18 | 11.90 | 13.49 | 2.18 | 2.22 | 4.85 | 5.20 | 2.66 |
| 106 on | 15.18 | 17.29 | 2.18 | 1.57 | 16.87 | 19.64 | 2.57 | 3.84 | 4.04 | 4.00 | 2.45 |
| 75 on | 9.51 | 10.74 | 1.18 | 0.79 | 12.50 | 6.15 | 1.38 | 1.21 | 1.41 | 1.80 | 0.00 |
| 75 pass | 5.88 | 5.37 | 0.80 | 1.97 | 5.56 | 2.39 | 1.98 | 0.41 | 0.21 | 0.80 | 0.62 |
| Wet mass | Granules | Tablets | ||||
|---|---|---|---|---|---|---|
| Variables | Amperage value of HSG | MTR torque | D50 | σg | Hardness | Disintegration time |
| p-Value | ||||||
| X1 | 0.935 | 0.059 | 0.989 | 0.953 | 0.935 | 0.997 |
| X2 | 0.094 | 1.22E-05 | 0.001 | 0.091 | 0.070 | 0.593 |
| X3 | 0.935 | 0.008 | 0.387 | 0.396 | 0.838 | 0.975 |
| X1X2 | 0.685 | 0.011 | 0.967 | 0.767 | 0.996 | 0.987 |
| X1X3 | 0.806 | 0.030 | 0.283 | 0.189 | 0.747 | 0.941 |
| X2X3 | 0.115 | 0.022 | 0.550 | 0.859 | 0.787 | 0.894 |
| R2 | 0.236 | 0.987 | 0.859 | 0.197 | 0.027 | −1.29117 |
From Table 5, the 75 µm pass fraction of granules is approx. 6% even in the largest batch. When only API was blended using HSG with center condition of Table 3, any significant difference in particle size of API before and after blending was not observed (data not shown). The particle size of lactose monohydrate and microcrystalline cellulose are under 75 µm and approx. 50 µm, respectively, and each ratio in formulation is 50% and 14%. Therefore, it was confirmed that the particles were bonded by collision, and the pulverization hardly occurred.
As for granule and tablet properties, they varied widely owing to the difference of the granulation conditions (In particular, D50; 190.4–666.0 µm, tablet hardness; 27.7–88.0 N and disintegration time; 63.8–535.4 s, Table 4). However, from the results of ANOVA (Table 6), only the X2 (Water amount) had significant effects on D50 among all parameters (p=0.001).
As for the properties of the wet mass, the amperage value varied from 3.4 to 4.9, and the MTR torque varied from 0.103 to 0.756 Nm for each batch, owing to the different granulation conditions (Table 4). The standard deviation (S.D.) of the MTR torque measurement (n=3) was small and good repeatability was confirmed. From the results of ANOVA, neither of process parameters affect the amperage value. Conversely, the R2 value of MTR torque was highest among all the intermediate properties, and the p-values of X2 (Water amount) and X3 (Blade speed) were both p<0.01. Figure 2 shows a ‘cause and effect diagram’ of each process parameter on the MTR torque. ‘Cause and effect diagram’ is widely used in quality engineering, which is described in order to shows a rough relationship between explanatory variables and response variables simply and conceptually. As shown in Fig. 2, the MTR torque increased with increasing of the amount of water and blade speed. These results indicate the state of kneading, i.e., granulation process, progressed with increasing of these parameters. Figure 3 shows scanning electron microscope (SEM) images of the granules from batch Nos. 1 and 8. The granulation conditions (level) of batch No. 1 were (X1/X2/X3)=(−1/−1/−1), and those of batch No. 8 were (+1/+1/+1). When comparing between granules with the same particle size, we found that smaller pores formed in granules of batch No. 8 than those of batch No. 1 and the surface of batch No. 8 was also smoother. Taken together, these results indicate that MTR torque could detect differences in the granulation conditions among various p-CMAs.
The dissolution profiles of the tablets are shown in Fig. 4. The dissolution profiles varied among the batches, and these differences might be attributed to differences in the manufacturing conditions during granulation. For evaluation time points, 30 and 60 min were selected. 30 min time points was selected because variations in dissolution of 6 vessels were confirmed in earlier points than 30 min, and 60 min time point was selected as a time point to reach almost plateau. Table 7 shows the dissolution rate and S.D. at these two points for the six vessels. The S.D. was quite small and there was little variation between the vessels.
| Study No. | Dissolution rate at 30 min [%] | Dissolution rate at 60 min [%] | ||
|---|---|---|---|---|
| Ave. | S.D. | Ave. | S.D. | |
| 1 | 22.6 | 0.5 | 26.0 | 0.6 |
| 2 | 22.5 | 1.7 | 26.5 | 1.2 |
| 3 | 26.5 | 1.7 | 32.1 | 2.4 |
| 4 | 30.5 | 0.6 | 36.4 | 1.1 |
| 5 | 22.3 | 0.9 | 25.5 | 0.8 |
| 6 | 26.9 | 1.7 | 31.2 | 1.4 |
| 7 | 32.9 | 0.9 | 38.7 | 1.2 |
| 8 | 32.3 | 2.2 | 37.8 | 2.6 |
| 9 | 29.5 | 1.8 | 34.7 | 1.5 |
| 10 | 27.6 | 2.1 | 34.5 | 1.9 |
| 11 | 29.0 | 1.0 | 35.7 | 0.9 |
Principle component analysis (PCA) was conducted to evaluate the relationship between the p-CMAs and dissolution rate. A loading plot of the PCA is shown in Fig. 5. PCA allows the results to be simplified into latent variables (Principal components, PC) that explain the main variance in the data.24) Loading of the 1st PC was 58% and the 2nd PC was 42%. These relationships could be explained with these two PCs. Generally, in PCA, items that are plotted in close proximity in a loading plot have a positive correlation, and items which plots in origin symmetry have negative correlation with each other. From Fig. 5, the MTR torque was plotted closely associated with dissolution rate at 30 and 60 min; hence, a high positive correlation between the MTR torque and dissolution was confirmed. In addition, the results of PCA indicated a positive correlation between the dissolution rate and D50 of the granules/tablet hardness.
ANOVA was conducted to evaluate the height of the correlation between the p-CMAs and dissolution. Table 8 shows the results of ANOVA. The correlation between the D50 values of granules and the dissolution rate was confirmed by a p-value less than 0.05 for the p-CMAs. However, the p-value of the MTR torque was smallest and the R2 value was largest, which suggests that MTR torque showed the highest correlation with dissolution among all the p-CMAs. The relationship between the MTR torque and dissolution rate at 30 and 60 min is shown in Fig. 6. The dissolution rate increased with increasing MTR torque, which is the same result as that from PCA, which also showed a positive correlation.
| CMA | Dissolution rate at 30 min | Dissolution rate at 60 min | |||
|---|---|---|---|---|---|
| p-Value | R2 | p-Value | R2 | ||
| Amperage value of HSG | 0.2236 | 0.16 | 0.2236 | 0.16 | |
| Wet mass | MTR torque | 0.0003 | 0.79 | 0.0003 | 0.78 |
| Granules | D50 | 0.0072 | 0.57 | 0.0065 | 0.58 |
| σg | 0.1924 | 0.18 | 0.2186 | 0.16 | |
| Tablets | Hardness | 0.0115 | 0.53 | 0.0023 | 0.66 |
| Disintegration time | 0.2483 | 0.15 | 0.1037 | 0.27 | |
The PCA and ANOVA results suggested that the MTR torque value quantified the kneading state, i.e., the progress of granulation, which had the greatest influence on dissolution. The MTR torque value increased as kneading progressed, resulting in rapid dissolution. The particle size of compound A was approx. 180 µm, which was larger than other excipients (the particle size of Pharmatose 200M is under 75 µm, and that of Ceolus PH-101 is approx. 50 µm). Therefore, the mechanism of granulation of this formulation is thought to be the layering of excipients to the particle of compound A. In addition, the solubility of compound A was low (0.73 mg/mL) and wettability was thought to be low. Therefore, API became covered with other wettable components and the wettability was improved as the granulation progressed. It resulted in dissolution improvement. From the SEM images of the granules in Fig. 3, the surface of the batch No. 8 granules was smooth; thus, the surface modification progressed further compared with that of batch No. 1. This means that in No. 8 granules, particles of compound A were fully layered with other excipients from estimated mechanism of granulation, and resulted in faster dissolution.
Whereas, the particle size increased with progressing granulation, the D50 value of the granules correlated with the dissolution rate, quadratically. From Table 4, the tablet hardness increased with increasing D50; thus, tablet hardness correlated with both dissolution. This result confirms the correlation between the particle size of the granules and dissolution; however, the particle size is a secondary evaluation, and the CMA, which directly affects the dissolution, was determined to be the kneading state, as quantitated by the MTR torque value.
However, no correlation was confirmed between the amperage value of the HSG and dissolution from the results of PCA and ANOVA. This is because the amperage value was not changed significantly by the difference of the granulation conditions and the correlation with process parameters was not confirmed in this study as shown in Tables 4 and 6. This result emphasizes the fact that no differences of the kneading state were apparent in the amperage values by statistical analysis. However, even in such a case, MTR torque can detect the difference, so the superiority of MTR torque than HSG amperage value was shown here. From these results, MTR torque is shown to be a suitable tool for quantifying and controlling the kneading state.
In this study, we focused on the torque of wet mass measurements by MTR, as a parameter to evaluate the kneading state, i.e., the progress of the granulation objectively, and we evaluated the relationship between the MTR torque and dissolution profile. Granules with a different kneading state were obtained by varying the granulation conditions of the HSG; however, the amperage value, which is usually used to determine the end point in HSG, could not detect differences of the kneading state; however, differences were apparent from MTR torque measurements.
Furthermore, we evaluated the relationship between dissolution profile and various p-CMAs, including the MTR using PCA and ANOVA. From these results, the MTR torque showed the highest correlation with dissolution compared with other p-CMAs, including particle size and tablet hardness, which are usually considered to affect dissolution. Particle size and hardness are properties of granules and tablets, and are affected by drying, sizing, and compression processes. Therefore, the properties of granules and tablets were affected by these processes, so they cannot directly evaluate the kneading state. Compared with these parameters, MTR torque is a property of wet mass after granulation, and can be directly used to evaluate the difference of the kneading state and as a result, dissolution.
Furthermore, MTR torque can be measured off-line, and can thus be used for scale-less evaluation. Therefore, MTR torque is a useful parameter for seeking appropriate conditions for HSG and for scale up, and changing equipment and manufacturing sites.
The authors declare no conflict of interest.
