Moisture-Activated Dry Granulation—Part I: A Guide to Excipient and Equipment Selection and Formulation Development

In the pharmaceutical industry, the three most common granulation processes for solid dosage form production are wet granulation, dry granulation (i.e., roller compaction), and direct blending. In spite of their popularity, each of the processes raise concerns as they are currently practiced.

The obligatory use of a granulating liquid during wet granulation generates large granules during the wet massing and kneading stages. The typical amount of water used in the formulation is 20–50% of the weight of the dry powder mixture. After granulation, most of the added water usually is removed by drying, followed by a granule-sizing step. In a way, the drying process cancels the water-addition step, and the sizing step shrinks the large granules formed during the process. One vexing, but thankfully infrequent, problem with the wet-granulation process is that it produces a bimodal particle-size distribution of the final granulation that may result in unsatisfactory granulation flow and compactability.

In the dry-granulation process, the powder mixture is roller compacted into ribbons that are milled into granules. Unlike the wet-granulation process, the roller-compaction process commonly obtains a final granulation with a bimodal particle-size distribution. Roller-compaction processes have the added problem of yielding material with low compactability during ribbon formation, thus resulting in soft and friable tablets.

The direct-blending process, followed by tablet compression or other downstream actions, often depends on the batch-to-batch and vendor-to-vendor consistency of drugs and excipients. The potential for segregation and inadequate material flow are risks often associated with this process.

In what is arguably the seminal paper on the moisture-activated dry-granulation (MADG) process, the authors proposed a simple, economical, and novel granulation process that uses a small amount (1–4%) of water to cause agglomeration without subsequently requiring a drying step (1). Few studies of this process appear in the literature (2–3). Given its simplicity and cost-saving potential, the authors expected that the MADG process would have been widely adopted in the pharmaceutical industry by this time. The MADG process has not caught on, however, perhaps because of its unusual simplicity coupled with uncertainty about equipment specifications and ambiguity about the manufacturing process.

This paper will explain the MADG process further and provide guidance for the selection of the excipients and equipment necessary for its successful implementation. The authors will also give instructions for the development of MADG-based formulations.

The MADG process

As its name implies, MADG is a process in which moisture is used to activate granule formation (i.e., agglomeration) without the need for applying heat to dry the granules. The formation of the moist agglomerates is followed by the stepwise addition and blending of common pharmaceutical ingredients that absorb and distribute the moisture, thereby resulting in a uniform, free-flowing, and compactible granulation. This process enables the drug to bind with the excipients after the agglomeration phase, thus resulting in small, almost spherical granules with low potential for segregation of the drug in the formulation. The intent of the MADG process is not to make large particles, but rather to agglomerate the fines and bind the drug with excipients to create free flowing, compactible, and nonaggregating granules.

The essence of the MADG process is to add enough water to achieve agglomeration without adding excess water that would require a drying step. It is equally important that only enough particle-size enlargement be achieved to ensure satisfactory granulation flow and compactability without segregation. The MADG process has the advantages of not generally requiring further size reduction and avoiding the regeneration of fines as a result of milling. And, unlike the conventional wet- and dry-granulation processes, MADG does not overdo and then undo what has been overdone.

The MADG process includes two major stages, the agglomeration stage, and the moisture distribution and absorption stage. Figure 1 shows a flow diagram of the MADG process.

Figure 1: Flow diagram of the moisture-activated dry-granulation process. (FIGURE 1 IS COURTESY OF THE AUTHORS)

Agglomeration stage. In this stage, all or part of the drug is mixed with filler(s) and an agglomerating binder to obtain a uniform mixture. During mixing, a small amount of water (1–4%) is sprayed onto the powder blend, thus moistening the binder and making it tacky. The binder functions as the drug and excipients move in the circular motion caused by the mixer impellers or blades. The resulting agglomerates are small and spherical because the amount of water used in the MADG process is much lower than that in conventional wet granulation. The agglomerates therefore cannot grow into large, wet lumps. The particle size of the agglomerates generally is in the range of 150–500 μm.

It is possible, based on the drug loading technique, to add only part of the drug to the formulation during the agglomeration stage. The remaining drug can be added after the moist agglomerates have been formed. The added drug particles adhere to the wet agglomerates and become incorporated into them.

Moisture-distribution and absorption stage. In this stage, moisture absorbents such as microcrystalline cellulose or silicon dioxide are added as mixing continues. When these agents come into contact with the moist agglomerates, they pick up moisture from the agglomerates and redistribute moisture within the mixture. The entire mixture thus becomes relatively dry. Although some of the moisture is removed from the wet agglomerates, some of these agglomerates remain almost intact, and some, usually the larger particles, may break up. This process results in a granulation with uniform particle-size distribution. The process continues with the addition of a disintegrant to the mixture, followed by blending for a few minutes. Then, during mixing, lubricant is added and blended for sufficient time to achieve adequate lubricity. This step completes the MADG granulation process.

Excluding material loading, the actual processing time for the MADG process is only 10–20 min. Even for a commercial-scale batch, the processing time is essentially the same as it would be for a laboratory- or pilot-scale batch. Beginning with the premixing of the drug and excipients, the final granulation could be ready for tablet compression, encapsulation, or powder filling in about an hour.

Excipients for the MADG process

Fillers for the MADG process during agglomeration. It is critical to select suitable excipients for a successful MADG process. Unlike the conventional wet-granulation process, which often employs microcrystalline cellulose or starch as fillers, MADG process uses nonabsorbent, easy-to-wet fillers such as lactose monohydrate and mannitol. The main reason for this selection is that microcrystalline cellulose and starch-based excipients absorb and retain a considerable amount of moisture during agglomeration. Because of this characteristic, more than the desired amount of water must be used during processing to form proper wet agglomerates. To ensure proper agglomeration, filler particles must not be too coarse or too fine. In general, coarse particles do not agglomerate easily, and fine particles require more moisture for agglomeration.

In rare cases, the drug itself could be soluble and become tacky upon moistening. Such drugs are classified as self-granulating. For these types of drugs, it is beneficial to include moisture absorbents during the agglomeration stage if a high drug-load formulation is desired in the MADG. Microcrystalline cellulose or starch products can help avoid overwetting and overgranulation of the product even when little moisture is used.

Agglomerating binders for the MADG process. The binders used in the agglomeration stage should be easily wettable and become tacky with the addition of a small amount of water. Previous studies indicate that low-viscosity polyvinylpyrrolidones (PVPs) such as PVP K-12 are ideal for this purpose. If PVP is not an acceptable choice because of formulation concerns such as chemical compatibility, binders such as hydroxypropyl cellulose (HPC), copovidone, maltodextrins, sodium carboxymethylcellulose (Na CMC), or hydroxypropyl methylcellulose (HPMC) can be used instead. The binders can be used singly or in multiple combinations to achieve the desired effects or address specific concerns.

If binders are available in various viscosity grades, it is desirable to use the ones with low viscosity because they tend not to retard tablet or capsule dissolution. However, binders with very low viscosity may not provide enough tackiness for agglomeration. In general, high-viscosity binders are often required in small amounts. The amount of binder needed does not depend on the viscosity alone; other factors such as binder mass must be considered. For example, if 5% of PVP K-12 is sufficient for one formulation, 2% of PVP K-30 may not be the correct proportion for the same formulation. Experiments have shown that about 3% or more of PVP K-30 would be required for proper agglomeration. This difference results from the fact that, in addition to binder viscosity and tackiness, the mass of the binder also plays an important role in covering and coating the blend particles that are to be agglomerated. The binders with small particle size and great surface area would be advantageous as well.

Generally, binders such as HPC, Na CMC, and HPMC require more water and longer hydration time compared with PVP or maltodextrin. On the other hand, binders such as Starch 1500 would not be suitable for the MADG process because this binder has a significant percentage of unhydrolyzed starch components that could absorb considerable amounts of water. As a result, the amount of water needed to effect agglomeration when using Starch 1500 would not be practical for the development of a typical MADG formulation. Completely hydrolyzed starch is not recommended because it does not have sufficient tackiness to cause agglomeration. In all cases, the binder chosen should have fine particles and sufficient tackiness upon moistening to cause adequate agglomeration.

Moisture absorbents for the MADG process. About 70–95% of any MADG formulation is agglomerated, and the remaining portion of excipients is added as is. In general, the nonagglomerated portion consists of moisture absorbents, disintegrants, and lubricants. It is desirable that nonagglomerated excipients be closer in particle-size distribution to the agglomerated portion of the formulation to minimize the potential for segregation.

Microcrystalline cellulose, which doubles as a filler and moisture absorbent, is available in the approximate particle size of 200 μm. Low moisture grades are also available. Avicel PH 200 LM (FMC, Philadelphia) is an excipient with low moisture content (< 1.5% by weight, as determined by loss on drying). Aeroperl 300, a moisture absorbent in the form of a non-lumpy, free-flowing granulated silica consisting of ~30-μm spherical particles is also available from Evonik Industries (Essen, Germany). Granular Aeroperl 300 has excellent moisture-absorbing capacity, and its surface area is much lower than that of the colloidal silica used as a glidant for granulation. The amount of Aeroperl 300 typically needed for the MADG formulation is small, which is advantageous from the standpoint of preventing tablet-ejection problems.

The disintegrant crospovidone is available in coarse particle-size grade from either ISP (Wayne, NJ) and BASF (Ludwigshafen, Germany). This material is not only a superdisintegrant, but is also compactible and acts as a moisture absorbent.

Overall, excipients such as Avicel PH 200 LM, Aeroperl 300, and the coarse grade of crospovidone for the nonagglomerated portion of the MADG process can significantly improve the quality of the formulation and facilitate the process. If the recommended excipients are not available, regular microcrystalline cellulose (e.g., Avicel PH101, PH102, and PH200), regular silicone dioxide, and crospovidone can be used as substitutes.

MADG formulation development

Assessment of API wettability. Drug solubility, particle-size distribution, and desired drug loading in the formulation are the primary factors to be considered for an MADG-based development. In general, a great amount of agglomerating binder and water are needed to create the agglomerates when a high drug load is desired for a drug with low solubility and small particle size. The converse is also true. Less agglomerating binder and water is required if the drug is water-soluble, the particle size is not small (e.g., > 10 μm), and the drug loading is low (e.g., < 25%). Self-granulating drugs sometimes do not require any binder and need less water to granulate.

Drug attributes such as wettability and agglomeration characteristics should be determined experimentally if they are not already known. Scientists can add water to the drug in a vial or in a small beaker using a syringe and stir the mixture with a small spatula. Generally, the drug is a suitable candidate for an MADG process if it can be wetted with 1–2% of water. If, on the other hand, the drug does not easily wet with 1–2% water, the formulation likely needs more binding material and water. Therefore, the higher the percentage of water needed to wet the drug, the more water or binder is needed for the agglomeration stage. As previously mentioned, it is difficult to develop an MADG process if a high amount of water or binder is required for the formulation.

Formulation assessment. Assessment of the formulation itself is the next task to be completed once the wettability of the drug has been established. For most drugs, a preliminary formulation-development evaluation can be initiated with a small batch. Using the starting formulation scheme provided in Table I, a 5–10-g batch can be prepared in a 20-mL scintillation vial.

Table I: MADG formulation for preliminary screening.

For nonwettable drugs or high drug-loading formulations, additional agglomerating binder (e.g., PVP) and more water during the agglomeration stage might be required. In addition, for drugs that are more difficult to granulate, mannitol (e.g., Perlitol 160 C, Roquette, France) or other wettable fillers can be used in place of lactose monohydrate to achieve the desired granulation. Conversely, small amounts of binder and water are needed if the drug is easily wettable and self-granulating. The ratio of Aeroperl 300 or other silicon-dioxide-type excipients to water should be kept to at least 1:1 by weight in the formulation. If PVP is not desirable in a given formulation, other agglomerating binders can be used, as described above.

Final formulation and optimization. Using the knowledge gained from the formulation-screening experiments described above, a large batch of several hundred grams can be manufactured with a high-shear granulator. The authors' experience has shown, however, that slightly more water is required for the experiments when a granulator is used instead of a vial. The preliminary studies enable adjustments to be made to improve formulation characteristics such as granulation and tableting, which can be further optimized as needed. Upon the successful completion of optimization exercises, the accelerated stability of the formulation can be evaluated. The scale-up and design-space studies can be conducted as needed.

Mechanism of the MADG process. The granule-formation mechanism in the MADG process is the same as that in conventional wet granulation. In both cases, it is a process of powder particle-size enlargement, often in the presence of water and binders, through wet massing and kneading. The main differences between these two granulation processes are the amount of granulating liquid used and the level of agglomeration achieved. In conventional wet granulation, substantially more water is used to create large and wet granules, and heat drying removes the excess water. This step is followed by milling to reduce the granule size. In the MADG process, only a small amount of water is used to create agglomeration. Moisture distribution and absorption steps follow, and neither heat drying nor milling is needed.

Additional considerations for the MADG process

Moisture in the MADG formulation. The amount of water used in the MADG process is part of the formula composition. This amount is a fixed value in the formula and is determined during formulation development. For example, if 2.0% (w/w) water is used, the rest of the ingredients should make up the 98.0% (w/w) of the formula. Because the MADG process does not include a heat-drying step, the water added would not be intentionally removed from the formulation.

Because moisture is added but not removed in the MADG process, what happens to the moisture and how it affects product quality might be causes for concern. To answer these questions, an MADG formulation that uses 1.5% water, 20% Avicel PH 200 LM, 1.5% Aeroperl 300, and other ingredients for a total weight of 100 g can be considered. First, 1.5 g of water is used in the agglomeration stage. During the moisture-absorbing and -distribution stage, 20.0 g of Avicel PH200 LM (with an inherent moisture level of 1.5%) can take 0.7 g of moisture, while 1.5 g of Aeroperl 300 can absorb 2.25 g of moisture from the wet agglomerates. As a result, the final granulation reaches its equilibrium moisture level, and neither Avicel PH200 LM nor Aeroperl 300 appears damp or lumpy. Such a MADG formulation would not have much more free water than that produced by a typical conventional granulation process. Even if only regular Avicel PH200 (with a moisture content of ~5%) is used without Aeroperl 300 in the same formulation, the amount of the remaining moisture (0.8 g) would be well distributed in the other formulation excipients, thus resulting in a free-flowing final granulation. Silicone dioxide in an MADG formulation sometimes may be preferred to minimize the risk of granulation caking during storage, to avoid flowability problems, and to reduce the chance of moisture-induced chemical instability. In general, unless the drug in the MADG formulation is moisture-sensitive, additional stability risks of the finished product would not be expected.

Water-delivery system. The agglomeration stage is critical in the MADG process and depends on the characteristics of the drug, type and amount of binders and fillers, and the addition of water. Because the amount of water used in the MADG process is small (e.g., 1–4%), it is important that the water be delivered accurately and distributed uniformly during the agglomeration stage. The selection of a spray system that provides accurate delivery and a well-defined spray pattern is important. A suitable spray system, Schlick MADG Spray Kit, is available for laboratory use from Orthos Liquid Systems (Buffton, SC).

The preferred mechanism to deliver water spray consistently would be an airless spray system, which enables the water to be directed onto the powder bed in a high-shear granulator. Any airless spray nozzle with a gear pump or pressure vessel, where the spray pattern can be reproduced and the exact amount of water delivered, would be adequate. Spray nozzles with an orifice of 0.1 mm or 0.15 mm can be attached to a syringe to deliver a low (5–10 mL) volume of water for small experiments.

Selection of the granulator for the MADG process. Although it has been reported that a simple planetary blender can be used for the MADG process, the authors believe that a high-shear granulator would be more suitable for the process (1). An ideal high-shear granulator has efficient impellers or blades and choppers to allow good mass movement and proper mixing. It also allows water to be sprayed only on the powder bed and not on the blades, choppers, or granulator wall. Also, the blades and bowl configuration should be such that they would not allow wet pockets or dead spots to remain after the moisture-distribution or absorption stages. Additional sifting or sizing of the granulation is required if such pockets or spots form. In other words, if good water distribution is achieved, further granulation sifting or sizing is unnecessary. Using this approach, the authors have successfully manufactured various products at batch sizes ranging from 100 g to 30 kg using equipment such as Bohle, Diosna, Fuji, Collette, or PMA Aeromatic-Fielder high-shear granulators.

Granulation sizing and milling. An optimized MADG formulation and process should not produce large lumps in the granulation that require sizing or milling. Therefore, once lubricant is blended in with the granulation, the result may be the final blend that can be directly used for tablet compression, encapsulation, or powder filling. At times, small amounts of lumps in the granulation may stem from material buildup on the blades, choppers, walls, or the bottom of the granulator during agglomeration. In such situations, it may be necessary to pass the granulation through a screen such as 10 mesh or any other suitable size. Often, sizing or sifting is needed only if the formulation or process contains imperfections.

Note

MADG process-based formulation-development studies carried out with various pharmaceutical compounds will be described in Part II of this article, which will appear in the December 2009 issue of Pharmaceutical Technology.

Ismat Ullah is president of Simple Pharma Solutions (Cranbury, NJ). Jennifer Wang* is a senior research investigator, Shih-Ying Chang is a principal scientist, Gary J. Wiley is a retired research scientist, Nemichand B. Jain is a director of biopharmaceutics research and development, and San Kiang is a research fellow, all at Bristol–Myers Squibb, 1 Squibb Dr., New Brunswick, NJ 08903, tel. 732.227.5684, jennifer.wang@bms.com

*To whom all correspondence should be addressed.

Submitted: Jan. 15, 2009. Accepted: Feb. 23, 2009.

What would you do differently? Submit your comments about this paper in the space below.

References

1. I. Ullah et al., Pharm. Technol. 11 (9), 48–54 (1987).

2. C. Chen et al., Drug Dev. Ind. Pharm. 16 (3), 379–394 (1990).

3. L.H. Christensen, H.E. Johansen, and T. Schaefer, Drug. Dev. Ind. Pharm. 20 (14), 2195–2213 (1994).