Issue link: https://www.e-digitaleditions.com/i/733311
more important to more companies in the dietary supple- ment industry. How do you know if a blend of a dietary supplement product is uniform? We can follow the acceptance criteria for generic drug products: The input of an active ingredi- ent should be in the range of 90 to 110 percent, and the RSD must be no more than 5 percent (usually 6 to 10 samples) [1]. Sampling can help you identify segregation and get an accurate picture of your blend quality. Improper sam- pling, however, can give you the false impression that your ingredients aren't blended well when they really are [2]. In this study, we implemented two sampling methods: thief sampling and discharge sampling of the CaMgZnD blend (Figure 1). Tests of these samples, which were taken at different blending times, show that the thief sampling method has a big negative impact on vitamin D blending uniformity results. Methods Equipment. CaMgZnD blending uniformity studies were conducted in a 5-cubic-foot V-type blender (Twin Dry Blender from Patterson-Kelley, East Stroudsburg, PA). The final blend was compressed using a 36-station tablet press (Libra from Kikusui, Lakewood, NJ) operat- ing at 30 rpm and equipped with oval tooling that mea- sured 0.625 by 0.356 inch. Batches and batch size. Three pilot batches of the same formula were made. The formula contained calcium carbonate, magnesium oxide, zinc sulfate (1.118 percent), vitamin D (0.039 percent), microcrystalline cellulose (MCC), croscarmellose sodium, silicon dioxide, and mag- nesium stearate. Each batch weighed 63.350 kilograms. Blending, sampling, and analyzing. All ingredients, except the magnesium stearate, were passed through a 12- mesh screen, added into the blender and blended for 10 minutes. Using a triple-cell thief, we took three samples from the left side of the blender and another three samples from the right side. Each sample weighed approximately six times what the tablets would weigh. Next, we ran the blender for 10 minutes and took another six samples. After this second sampling, we blended for yet another 10 min- utes and took six more samples. We collected 18 samples in total. Next, we passed the magnesium stearate through the 12- mesh screen into the blender and blended for 3 minutes. We collected two samples each from the bottom, middle, and top of the discharge stream. Each sample weighed approximately 300 grams, and each was analyzed for its content of calcium, magnesium, zinc, and vitamin D. Tabletting, sampling, and analyzing. The final blend was compressed into tablets that weighed 1.267 grams each. We took a composite sample from each lot and con- ducted content uniformity testing for calcium, magnesium, zinc, and vitamin D. Physical testing. We conducted a sieve analysis of the main ingredients and measured each tablet's weight, thickness, hardness, and disintegration time. Results and discussion Inputs of calcium, magnesium, zinc, and vitamin D were all in the range of 90 to 110 percent. Table 1 summarizes the RSDs of the uniformity data of the samples taken via the sampling thief. An RSD of 5 percent or less indicates that an ingredient is uniformly distributed in the blend [1]. Calcium, magne- sium, and zinc were homogenous in samples taken at all blending intervals, but the vitamin D was not uniform. The poor uniformity data for vitamin D may be caused by an electrostatic charge during sampling [3, 4]. Indeed, a blending uniformity study of a low-dose phar- maceutical ingredient (2 percent warfarin) has demon- strated that an electrostatic charge can have a negative impact on uniformity or increase the possibility of sampling error [4]. That study used calcium hydrogen phosphate with two different average particle sizes, 61 and 152 microns, as one filler and MCC with one average particle size as another. A Faraday pail was used to measure the electrostatic charge of the blends. Two blends made with D61 had a negative electrostatic charge that was three times as strong as the charge present in two blends made with D152, which caused the RSDs of the blending unifor- mity tests to be much higher (6.24 percent and 7.18 per- cent) than the FDA acceptance limit of 5 percent. Two blends made with D152 had much lower RSDs (1.37 per- cent and 2.58 percent). The larger particle size may have reduced the sampling error by suppressing the electrostatic charge [4]. Table 2 summarizes the sieve analysis data of calcium car- bonate, magnesium oxide, zinc sulfate, vitamin D, and MCC. Table 1 RSDs of uniformity data of samples taken by sampling thief (n=6) Lot number 15060200 15060201 15060202 Blending time (min) 10 20 30 10 20 30 10 20 30 Calcium 2.25% 3.76% 2.09% 3.36% 4.98% 2.60% 1.29% 1.68% 2.50% Magnesium 3.18% 3.90% 2.22% 3.29% 3.22% 1.83% 1.26% 1.03% 2.19% Zinc 2.48% 3.53% 3.90% 4.30% 3.59% 4.09% 2.22% 4.13% 2.08% Vitamin D 10.40% 17.26% 8.14% 16.00% 11.56% 7.89% 12.74% 19.21% 8.68% Figure 1 Sample thief Cavities open for sampling Cavities closed as thief is inserted or removed 18 October 2016 Tablets & Capsules