2014 Volume 62 Issue 5 Pages 488-490
We developed a color difference signal method to evaluate the degree of blending powdered medicines in pharmacies. In the method, the degree of blending is expressed as the relative standard deviation of the color difference signal value (Cb or Cr) of the YCbCr color space after digital photos of the blended medicines are analyzed by image processing. While the method is effective to determine the degree of blending colored medicines, it remains unknown whether it can be applied to uncolored or white-colored medicines. To investigate this, we examined colored diluents to identify an indicator of the degree mixtures are blended. In this study, we applied this method to Pontal® and Prednisolone® powders, which were used as uncolored and white-colored medicines, respectively. Each of these medicines was blended with the colored lactose using a pestle and mortar, and then the uniformity of blending was evaluated. The degree of blending was well-monitored in both mixtures with various blending ratios (1 : 9–9 : 1), showing a sufficient uniformity at 60 rotations of the pestle. Moreover, the Cr values of the mixtures with various blending ratios were correlated with the concentration of active pharmaceutical ingredients in these medicines, which was determined using HPLC. This indicated the usefulness of the color difference signal method for the quantitative determination of medicines. Thus, we demonstrated the applicability and effectiveness of this method to check dispensing powders.
In their dispensation, powdered medicines are frequently blended with diluents or other medicines. In such a case, the uniformity of powdered medicine mixtures is essential for quality control. In pharmacies, however, this is unchecked quantitatively before pharmacists give the medicines to patients. This is due to the lack of an effective method to test the dispensing mixtures regarding the degree of blending. Therefore, we developed the color difference signal method as a nondestructive and concise method which is practicable in pharmacies.1) In brief, the method consists of three steps: 1) Digital photos of blended medicines are taken with a handy microscope. 2) The photos are analyzed using image-processing software to obtain R, G, and B values in the Red-Green-Blue (RGB) color space of all pixels. Mean values of R, G, and B are calculated, and then converted to Y, Cb, and Cr values in the YCbCr color space. 3) The degree of blending is expressed as the relative standard deviation of the color difference signal value (Cb or Cr). While the method is effective to determine the degree of blending colored medicines, it remains unknown whether it can be applied to uncolored or white-colored medicines.
In order to solve the problem in practical use, we proposed the usage of colored diluents as an indicator of the degree of blending in dispensing mixtures. Ordinarily, colored diluents are used for the trituration of powerful drugs and also as an indicator to confirm the homogeneity of blended ingredients by visual inspection. In this case, pigments are selected based on the effectiveness of the indicator as well as safety of patients. For example, food blue No. 1 is added at a concentration of 0.002% for deleterious drugs, while food red No. 3 is used at a concentration of 0.001% for powerful drugs.2)
In this study, we employed 0.001% food red No. 3-lactose powder as a diluent for uncolored and white-colored medicines. The mixtures were assessed regarding their degree of blending using the color difference signal method. Moreover, a correlation between the method and HPLC method for the quantification of active pharmaceutical ingredients (APIs) in the mixtures was investigated.
Powdered lactose (Mylan Seiyaku Co., Ltd., Tokyo, Japan) was selected as a diluent because it is generally used for dispensation. Food red No. 3 was obtained from Tokyo kasei kogyo Co., Ltd. (Tokyo, Japan) and used as a model pigment. Pontal® powder 50% (Mefenamic Acid, Daiichi-Sankyo Pharmaceutical Co., Ltd.) and Prednisolone® powder 1% (Prednisolone, Takeda Co., Ltd.) were purchased for model powdered medicines. Mefenamic acid and prednisolone were obtained as standard compounds from Wako Pure Chemical Industries, Ltd. (Osaka, Japan). Methanol and acetonitrile were of HPLC grade. All other reagents were of reagent grade.
Colored Diluent PreparationPowdered lactose and food red No. 3 were sieved using a 355-µm sieve, and then mixed using a pestle and mortar at a mass ratio of 1 : 999 until becoming homogeneous.
Measurement of Physical PropertiesThe mean diameter, angle of repose, and Carr’s index were determined according to Japanese Pharmacopeia XVI. In brief, the mean diameter was calculated from the cumulative amount remaining on sieves with 53, 75, 100, 150, 250, and 355-µm mesh sizes. The angle of repose was determined from circular cones formed on a cylinder of 20 mm in diameter. The poured density was measured with a measuring cylinder (25 mL) using 10 g of powder, and calculated by dividing the mass by the volume. Then, the measuring cylinder was tapped until a constant volume was obtained. The tapped density was calculated by dividing the mass by the constant volume. Carr’s index3) was determined as follows:
 | (1) |
The smaller Carr’s index, the better the flow properties. For example, <10 indicates excellent, 11–20 good, 21–25 fair, and >26 poor flow.
Blending PowdersManual blending was performed with a pestle and mortar according to the Japanese Guidelines for Dispensation.4) That is, the mortar (10 cm in diameter and 5 cm in depth) was held and rotated in the left hand, and the pestle (12 cm in length) was rotated in the right hand spirally in the direction opposite to that of the mortar. Then, 10 clockwise rotations followed by 10 anticlockwise rotations were repeated three times, with the mortar thus being rotated 60 times.
Mixtures of medicine and colored diluents were prepared at ratios of 1 : 9, 3 : 7, 5 : 5, 7 : 3, and 9 : 1. The total amount of the mixtures was 5 g, which was charged in the mortar. The experiments were repeated three times.
Digital Image ProcessingImage processing was carried out according to a previous report.1) In brief, digital photos of mixtures were taken with a microscope (M3, Scalar Co., Ltd., Tokyo, Japan) equipped with a lens (30N, magnification ×30), and obtained as Microsoft Bitmap files (VGA). Pixel analysis was performed by ImageJ5) (National Institute of Health, Washington, D.C., U.S.A.) to obtain color element values of R, G, and B in the RGB color space. Then, the mean values of R, G, and B were converted to Y, Cb, and Cr in the YCbCr color space.6) Y, Cb, and Cr shows brightness, color difference in red chroma, and color difference in blue chroma, respectively. The color difference signal, Cr, was used for an indicator of the degree of blending in this study.
Assessment of Degree of BlendingFive photos were taken at different positions of a sample spread on the stage area. The blending state of the mixture was assessed using the relative standard deviation of the Cr value. In general, blending proceeds with an increase in the blending time, and then uniform blending is achieved. Therefore, a smaller relative standard deviation (%) of the Cr value means a higher uniformity of blending. A value less than 6.08% is recognized as sufficient blending.7)
Determination of APIsThe concentration of APIs in the mixtures was assayed by HPLC. The HPLC system (Shimadzu, Kyoto, Japan) consists of a pump (LC-10AD), an automated sample injector (SIL-10A), and UV spectrophotometric detector (SPD-10A). Samples were analyzed on an ODS column (4.6 mmϕ×150 mm, YMC-Pack Pro C18, Japan). The wavelength and mobile phase are summarized in Table 1.
Blending of colored lactose and powdered medicines, Pontal® and Prednisolone® powders, was performed using a pestle and mortar at the weight ratio of 1 : 1. During the blending, its degree was monitored using the color difference signal method. Figure 1 shows the relative standard deviation of the Cr value as a function of the number of blending rotations.
Blending of colored lactose and powdered medicines was performed using a pestle and mortar at the weight ratio of 1 : 1. The relative standard deviation (%) of the Cr value is plotted as a function of the number of blending rotations. The values are presented as the mean±S.D. of three experiments.
For both medicines, the relative standard deviation decreased with an increasing number of blending rotations. The value for Prednisolone® tended to be lower than that for Pontal®. It was probably due to difference in particle size of powders. The particle diameters of Prednisolone® and colored lactose were similar while the diameter of Pontal® was larger than that of colored lactose as shown in Table 2. A value less than 6.08% means a sufficient blending state.7) Therefore, the blending state at 30–60 rotations was in a sufficient level. These results are consistent with a previous report using colored medicines.1)
| Medicine | Mean diameter (µm) | Angle of repose (°) | Carr’s index |
|---|---|---|---|
| Pontal® powder | 163±13 | 54.5±2.6 | 31.4±1.5 |
| Prednisolone® powder | <75* | 56.7±1.2 | 32.4±1.5 |
| Colored lactose | <75* | 58.4±0.62 | 41.3±1.7 |
Each value represents the mean±S.D. (n=3 or 5). *: More than 50% of powder passes through a 200 mesh sieve.
Thus, the degree of blending was well-monitored using colored lactose, indicating that the color difference signal method is effective for uncolored and white-colored medicines. It is important to assess the degree of blending dispensing mixtures quantitatively because visual inspection is dependent on individual pharmacists. This nondestructive method was practicable in pharmacies without expensive apparatuses.
Influence of Blending Ratio on UniformityThe blending of colored diluents at various ratios was carried out using a pestle and mortar with 60 rotations. The relative standard deviations are shown in Fig. 2.
Colored lactose and powdered medicines were blended using a pestle and mortar with 60 rotations. Each relative standard deviation (%) represents the mean+S.D. of three experiments. No significant difference (p <0.01) was detected among mixtures with various blending ratios for either medicine.
With all ratios, the values were less than 6.08%, indicating a sufficient degree of blending. There was no significant difference among mixtures with various ratios for either medicine. Using colored lactose, the medicines were well blended over a wide range of 1 : 9–9 : 1.
In general, blending is influenced by many factors. Among them, the physical properties of powders affect the miscibility.10,11) Therefore, Table 2 summarizes the physical properties of the powders used in this study.
The mean diameter of Pontal® was different from that of colored lactose. The angle of repose of these powders indicated low flowability. Carr’s index of these powders was determined to be more than 26, suggesting a cohesive property. Although these powders were not suitable for blending with each other, a sufficient degree of blending was achieved. We confirmed this by the quantification of APIs in the medicines with HPLC, as shown below.
Correlation between Color Difference Signal and HPLC MethodsUsing the mixtures of colored lactose and Pontal® or Prednisolone® powder with various blending ratios, the mass percent measured by this method was compared to that calculated based on the API content of the mixtures determined with HPLC. The relationship is shown in Fig. 3.
The mixtures of colored lactose and powdered medicines were prepared using a pestle and mortar with 60 rotations. Correlation curves are displayed as solid lines with their correlation coefficients (R 2).
With both medicines, the Cr value decreased as the mass percent of the medicines increased. An increase of the medicines means a decrease of colored lactose which is red. Therefore, the Cr value becomes smaller. A good linear correlation was observed in both cases. This indicated that the color difference signal Cr value changed quantitatively with the mass percent. Therefore, the relative standard deviation of the color difference signal represents the uniformity of the mixtures.
We applied the color difference signal method to uncolored and white-colored medicines using colored lactose. The degree of blending was well-monitored at various ratios of the medicines and colored lactose. In addition, the mass percent of the medicines in the mixture was determined quantitatively. This concise method is effective not only to check the uniformity of dispensing mixtures, but also to determine the mass percent nondestructively.
