The pharmaceutical industry is increasingly looking for approaches that shorten drug development timelines, especially virtual and biotech companies that rely heavily on key milestone payments. Consequently, it is imperative that new chemical entities can be quickly manufactured into clinical drug products.
Across the industry there are two principle approaches to formulation development of solid oral dosage forms for First-In-Human (FIH) studies, namely the "commercial formulation approach" and the "exploratory formulation approach."
Following the exploratory formulation approach, the criteria explored in this article must be met for the API-in-capsule prototype to be successful. This article is broadly broken-down into two areas: the direct and indirect cost savings associated with the API-in-capsule dosage form, and how systems such as the Xcelodose� 600S mitigate against risks associated with the "commercial formulation approach."
Existing Industry Approaches
The commercial formulation approach advocates a clinical formulation that will bear resemblance to commercially acceptable formulations; for a solid oral dose this will most commonly take the form of a capsule or coated tablet. The goal in developing this first-generation formulation will focus on ensuring that it is "fit for purpose" for Phase I and II clinical studies. Done well, there is a strong possibility that the development work will provide a strong starting point for development of a formulation for late-stage clinical development and registration. The challenge here is to select a dosage form that can reach late-stage development quickly and in a cost-effective manner.
The exploratory formulation approach favors the use of the simplest possible formulation, such as powder-in-capsule (PIC)1. In its simplest form, this will be API-in-capsule. Use of API-in-capsule has seen a large growth in popularity during the last five to ten years, not least because of the arrival of systems which enable automated and accurate filling of relatively large quantities of API-in-capsules, such as the Xcelodose (Capsugel, Cambridge, UK) or Powdernium� (Accelrys Inc., San Diego, CA). Directly filling API into a capsule is probably the quickest option for entering clinical trials, as this method requires little or no excipients - potentially saving three to four months of formulation development and stability testing.
Contrary to what many people believe, pursuing the commercial formulation approach for FIH studies will not necessarily place supply of drug product on the critical path of the overall development timeline. Assuming good planning by the Project Manager, a formulation can be developed and stability data generated in parallel to preclinical toxicology studies, which in general takes four to sixth months to complete.
In the context of overall development costs, direct outlays associated with pursuing the commercial formulation approach are not huge. Also, investment in this area can increase the value of the overall package if out-licensing is planned after early development. Nonetheless, in an environment where development budgets are tight and where there is pressure to minimize development spend before proof of concept, the direct and indirect cost savings associated with adopting an API-in-capsule can be attractive to many.
Is API-in-Capsule Always aSuitable Approach?
In the case of companies like ours, the suitability of API-in-capsule is decided upon during up-front assessments by a multidisciplinary project team - including CMC experts from Chemical Development, Solid State Chemistry, Formulation Development, Analytical Development and Project Management. Firstly, the multidisciplinary team scrutinizes the data to ensure that the API is suitable for an API-in-capsule presentation.
If the API possesses poor physicochemical properties - e.g. poor aqueous solubility such as Biopharmaceutics Classification System class II or IV compounds - then API-in-capsule may not be appropriate, as some formulation development work will be required to investigate the potential benefits of functional excipients, such as solubilising agents, which may be required to improve bioavailability. This is an important point, as too often organizations in a hurry can overlook the potential impact of the API properties when determining their pharmaceutical development strategy and API batch sizes. While all the information may not be available, sufficient data will have usually been generated during the lead optimization stage of discovery to enable an informed prediction on the likely formulation strategy.
Another property to consider is the bulk density of the API, as most API-in-capsule machines do not possess a 'tamping' feature. Consequently, the bulk density and the way the drug particles pack have a direct effect on the amount of API that can fit into a capsule. Therefore, the dose is essentially limited by the size of the capsule.
It is well known that micronized powders possess poor flow properties, and this can result in issues with both the API-in-capsule machines and capsule boards. For example, the high throughput unit (HTU) of the Xcelodose automatically refills the dispense head with API, but a certain degree of flowability is required so that the API can be transferred from the HTU hopper into the dispense head. Capsule boards can also be affected by poor flow properties, as the filling process can be irregular, leading to variability in dose/fill weight. Not only is the particle size of the API to be considered when determining if API-in-capsule is a suitable strategy, but the distribution can also have an influence. If the API possesses a wide particle size distribution, then it may not be suitable for the Xcelodose. Potentially small particles may pass through the dispense head into the capsule, with the larger particles retained. This could lead to having to stop and clean the dispense head at regular intervals, resulting in increased downtime. It is worth noting, however, that careful selection of the dispense head and sieving the API to narrow the particle size distribution could assist in overcoming this issue.
In summary, if the physicochemical properties of the API are good, then API-in-capsule using a manual fill, capsule boards or Xcelodose system should be possible. Considera-tion should also be given to the likely doses and the quantity of capsules required in determining whether the Xcelodose is an appropriate piece of equipment to use.
Direct Cost Savings withAPI-in-capsule using the Xcelodose
API usage: In early development, API is usually in short supply, with material being needed for toxicology and DMPK studies, analytical development, API characterization studies, and formulation development. Although API is usually relatively expensive at this stage of development, one of the benefits of the Xcelodose system is that only small quantities of API are required to determine operating parameters.
Typically ~5g of API is required to develop Xcelodose filling parameters, such as, tap frequency, pulse width, amount of slow tapping and high throughput unit settings.
Compared to usage expected by developing a dosage form using the "commercial formulation approach," 5g is accepted as a reasonable amount of API for drug product development. Furthermore, from a scheduling perspective, development of Xcelodose operating parameters can be left until immediately prior to GMP operations - meaning that the project manager has one less aspect to worry about when allocating API from earlier development batches.
Direct labor hours: Four to six weeks of development time is usually required for a "fit for purpose" formulation in early development following the commercial approach. Although the formulation scientist will not be 'hands on' for all of this time, they will be reasonably busy interpreting excipientcompatibility data, manufacturing prototype blends and stability batches of the dosage units, and writing protocols/reports. In contrast, API-in-capsule negates the initial need for lengthy excipient compatibility, blend feasibility and probe stability studies.
In order to manufacture API-in-capsule supplies for clinical use, traditional methods have included hand-filling, and occasionally the use of capsule boards. Hand-filling is certainly not rapid, and if the weight required is low - for example, tens of milligrams or less - this filling process may take up to five minutes per capsule.2Capsule boards can produce filled capsules quicker than hand-filling, but can result in large losses of API as the process is not particularly "clean." A further problem is that the use of capsule boards does not lend itself to low dose fills, as small volume dosing plates can lead to non-reproducible fills particularly for poor flowing APIs. In contrast, the Xcelodose can fill a wide range of weights, including very low weights both accurately and relatively quickly. Once a suitable set of parameters has been developed for the API in question, there is little need for direct labor.
Indirect Cost Savings withAPI-in-capsule using the Xcelodose
In addition to the direct cost savings outlined above, there are many more indirect savings that can be achieved by using an API-in-capsule approach. Some of these are highlighted below:
Simplifies analytical development: The analytical methods that have been developed for the API will most likely be suitable, with little or no analytical development. In the majority of cases, one would not expect specificity issues, making the assay method particularly straightforward for the analytical group. If the API is particularly soluble in acid conditions one could argue that the API-in-capsule approach can remove the need for a dissolution method.
Content uniformity: The API-in-capsule approach removes the need for conducting content uniformity testing. This is because a record of every fill weight, and information on the distribution of fill weights around the target fill, is available from the Xcelodose 600. Frequently, results will give an RSD of 2% to 3% and equate to a weight typically within 1% of target.3 The data recorded is fully compliant with 21CFR part 11, and is saved as a pdf file that can be printed but not edited. If capsules are hand-filled, the weight of each capsule produced must be checked to ensure that it is within limits. This is achieved by reviewing the balance printouts, which can be a very time-consuming process.
Stability: Typically for API-in-capsule drug products, our clients focus their Clinical Trial Application (CTA) stability justifications around the drug substance stability data. This data is used as primary evidence to justify shelf life for the drug product. Typically, when this strategy is employed, clients commit to a CTA that places the API-in-capsule clinical batch(es) on storage, providing updates to the CTA as data becomes available. This strategy helps accelerate the program through to clinical trial, as there is no need to wait for product stability data to become available.
Cleaning methods: The cost of replacing the contact parts on the Xcelodose is significantly cheaper than the cost required to develop a cleaning verification method. Another consideration is time; developing a cleaning verification method can take weeks of work, and that time could be spent developing the Xcelodose method, and indeed, starting manufacture of the clinical product.
Ease of containment: A lack of toxicity data for compounds entering FIH studies is common, and therefore to ensure operator safety protection, the Xcelodose can be used within the Xcelohood�. This means potentially toxic compounds can be easily contained within the Xcelohood system.
Mitigation of Risks Associated with the Commercial Formulation Approach
It is not uncommon in early development to encounter differences in particle size, hygroscopicity, polymorph content and crystallinity of the API. Such changes in the API during chemical development can result in headaches for the multidisciplinary team working on the commercial formulation ap-proach. For example, a change in flow properties can affect blending time, roller compaction settings, and compression or encapsulation parameters, resulting in lengthening timelines, more development work and therefore increased costs. Physicochemical characterization is still extremely important and API properties can have a significant effect on the processing parameters for the Xcelodose.
However, a change in API characteristics when using this approach would most likely result in only one or two days of additional time for optimization of Xcelodose processing parameters.
Blending excipients and a drug together can be difficult, especially when attempting to attain uniformity at low starting doses. Segregation of the components, sampling bias and the need for a blend uniformity method are all challenges associated with the commercial approach. All of these factors are negated with the exploratory formulation approach.
A client approached us requesting supplies for a FIH study. Initial requirements were for less than 1000 capsules of two different dosage strengths. No formulation studies were to take place, as the client wanted to save time and API for reasons outlined previously. A decision was taken to manufacture these supplies manually using an analytical balance. The client subsequently contacted us about a re-supply of 25 mg and 100 mg capsules in quantities of greater than 10,000 and 5,500, respectively. It was quickly decided that an automated process was the best solution for manufacture of re-supplies.
The API possessed poor flow properties and was treated as a potent compound (Operator Exposure Limit of between 0.1 � 10 �g/m3) due mainly to the lack of toxicological information available. This re-supply project was therefore a good fit for the Xcelodose system, in combination with the potent handling capabilities of the Xcelohood.
From the solid state characterization package, the particle size distribution of the batch of API to be used for the re-supply is shown in Table 1 (below):
This information, coupled with flowability data, suggested that the API possessed poor flow properties. Although poor flow does not impact hand-filling, it was thought that poor flow may cause issues with the Xcelodose, particularly when using the high throughput unit, where the API must flow from the hopper into the dispense head. However, with this API, it was possible to adjust the frequency of the vibrations and the angle of the hopper to allow a reasonable flow of API into the dispense head.
Overall, this project was completed within three weeks (single shift, 40-hour working week). Less than one day was required to determine suitable processing parameters for the 25 mg and 100 mg strengths.
The low-fill-weight capsules took approximately 76 hours to complete, whereas the larger-fill-weight capsules were completed in just over 36 hours.
When the initial clinical supplies were manufactured by hand-filling, the average rate of production of 25 mg and 100 mg capsules was 30 and 50 capsules per hour, respectively. Therefore, if the Xcelodose had not been available for the re-supply, manual capsule filling would have required approximately 400 hours to complete the 25 mg fill and 110 hours for the larger fill (see Figure 2).
Figure 2: Timelines using the Xcelodose� compared to hand-filling
It is also worth noting that, as commonly seen with FIH studies, there was a lack of toxicology information about the API. Therefore, in order to ensure protection of the operators, the original supplies were manufactured within a ventilated balance safety enclosure (VBSE). However, because the analytical balance was in direct contact with the VBSE, any vibration from the unit resulted in difficulties with balance taring. For the re-supply operation, the Xcelodose was used within the specially designed Xcelohood system. As the Xcelodose is not directly in contact with the Xcelohood, the micro-balance used by the Xcelodose is inherently more stable and results in much quicker taring.
There are a number of approaches for the manufacture of supplies in FIH studies, including ones we've implemented. With the increasingly favored exploratory formulation approach, possible manufacturing options include hand-filling, capsule boards and the Xcelodose. This article has explored the direct and indirect benefits of using the Xcelodose. In the case study, the direct cost savings of the Xcelodose were shown in terms of direct labor costs, with clinical supplies delivered approximately seven weeks earlier than if a hand-filling operation had been employed. In addition, timelines for this project were aided further by the fact that no cleaning method was required, meaning development of the Xcelodose process parameters could start immediately. Finally, only 5g of API was required for development activities and, coupled with minimal losses during processing, resulted in less API usage when compared to a hand-filling approach, and significantly less API usage, than would have been required for development using a commercial formulation approach.
If the exploratory formulation approach is suitable, and speed to clinical is imperative, then we have found the Xcelodose system a viable option for the manufacture of supplies for FIH studies. However, if time and API supply are not of the essence, then we will work with client partners on a strategic commercial formulation approach to develop a suitable dosage form, fit for purpose for clinical supply.
1. Hariharan M, et al., Reducing the time to develop and manufacture formulations for First Oral Dose in Humans Pharmaceutical Technology 2003
2. Bryant, s. et al., Advantages in powder-dosing technology Innovations in Pharmaceutical Technology 95-100 (2002)
3. Capsugel website. http://www.capsugel.com
Colin Lorimer, Ph.D. MPSNI, is senior formulation scientist at Almac. He has worked in the pharmaceutical industry for more than eight years and has practical experience of interpreting solid state information pertinent to formulation development. Specializing in solid oral dosage forms, he has worked on many formulations for early stage development using various technologies. He can be reached at firstname.lastname@example.org.