Medical Device Link.

DESIGN

Preclinical studies

Image: Biocompatibles UK Ltd

In addition to physicomechanical issues discussed in Part I of this article,¹ the combination device poses a number of questions associated with its drug delivery properties and the subsequent biological action, all of which must be evaluated to demonstrate safety and efficacy.² Information from in vitro test methods is required to predict the product behaviour in vivo. The drug-eluting bead (DEB) is employed for intra-arterial drug delivery, whereby the DEB is delivered into the feeding artery of a tumour to block the vessel and starve the tumour. This process causes some of the tumour cells to express survival factors, which is combated by the local delivery of chemotherapeutic agent to kill the cells. For the DEB, important in vitro tests are required to evaluate

• microsphere size range that is produced
• drug dose (amount and reproducibility)
• drug elution rate (extent and reproducibility)
• presence and concentration of impurities/degradants
• catheter delivery properties/handling
• water content and related compressibility.

Of these tests, one that is extremely important is the evaluation of the in vitro drug elution rate, because it can affect the efficacy and the safety of the device. Robust test methods that can evaluate the consistency and reproducibility of this crucial parameter are, therefore, important. These are often developed as release tests and tend to be designed to artificially elute the drug over a time frame that can reasonably be fitted into a batch release schedule.³ It is important, however, that measures are available that allow for the prediction of the drug elution rate in vivo. In vitro tests that provide results that predict the outcome in vivo are powerful and studies on the in vitro–in vivo correlation (IVIVC) should be performed where possible.^3,4

Figure 1: Schematics of a T-cell apparatus (right) and how DEB embolise a vessel. (click image to enlarge)

In the development of the doxorubicin DEB, an in vitro test was developed that attempts to emulate the embolisation environment. Here, a number of critical factors need to be considered: temperature, elution media, time, volume of media and flow conditions. The technique adopted to predict the DEB drug elution rate in vivo was the T-Cell Model. Figure 1 shows a schematic of the T-cell model and a representation of beads embolising a feeding artery to a liver tumour. It can be seen that embolisation results in the reduction of flow through the artery. The T-Cell apparatus has, therefore, been designed as a restricted flow model that combines diffusion and convection components in the release and attempts to mimic the conditions experienced by the beads in vivo.^5,6

In work with DEBs, new in vivo preclinical models have had to be developed to evaluate these products.^7,8 The doxorubicin DEB is a combination product based around a proprietary hydrogel microsphere or bead that is loaded with doxorubicin. The product is designed for use as an embolisation device for the local treatment of hypervascular tumours such as primary and secondary liver cancer. These devices are delivered via microcatheters directly into the feeding artery of a tumour to block the vessel and starve the tumour cells of nutrients and oxygen. Drug is subsequently delivered from the device to attain further cell kill and help overcome the cells’ resistance mechanisms. When considering an appropriate model to study effects in vivo, a number of important considerations come to mind:

• What in vivo preclinical model should be used to demonstrate safety and efficacy?
• For how long should in vivo studies be performed to demonstrate long term safety (for example, one, three, six or twelve months)?
• What are the long term effects of implantation of the device in vivo?
• What is the effect of local delivery to the tissue from the device?
• Is sufficient therapeutic delivered to have a positive biological effect?
• What are the implications of device-based delivery in terms of dosage delivered, for example, dose per mm² or mm³ tissue?

Table I: The risks identified and the test used to evaluate the risk in vivo using the swine liver embolisation model. (click image to enlarge)

The in vivo preclinical studies on this device combination were designed by taking into account of the potential risks associated with administration. Clearly, the choice of model was limited to one that could accommodate the size of the device under investigation. Mice, for example, do not possess a vasculature suitable for embolisation because of their small size. Although rabbit models have proved possible, their small size presents some limitations in use. A more practical approach was embolisation via the porcine hepatic artery, which has the additional benefits that the hepatic vasculature of the swine is similar to that of humans. Once the potential risks of administration of the device were identified, a test/evaluation was then designed to address these. For example, doxorubicin is well known to be cardiotoxic; therefore, measures of cardiotoxicity were included in the study design. In addition, it is known that “off-target embolisation” can sometimes occur during administration of microspheres to the liver via the hepatic artery. Because of this, distant organs needed to be evaluated to determine these potential effects. Table I shows the risks identified and the test used in their evaluation.

Pharmacokinetic analysis was also conducted using this model to determine release into the plasma during the first few days post-embolisation. Figure 2 shows the IVIVC between the T-Cell and data from the in vivo model.⁴ The data clearly shows that the T-Cell correlates well with the in vivo results during the first 24 hours of release. This is a useful test because it allows evaluation of products using an in vitro test that is predictive of the extent of early systemic exposure in vivo. This tool can be used to assess the effect of potential changes to a product/process on an important product characteristic without having to repeat in vivo validations/verifications.

Shelf-life testing

Figure 2: IVIVC for T-cell release and plasma PK in a porcine heparic embolisation model. (click image to enlarge)

The design of a suitable shelf-life test for a combination product is another area of complexity. Both components need to be evaluated over time and this may be further complicated if the device is designed in different sizes. An excellent example to illustrate this point is the case of the drug-eluting stent (DES), which is employed to retain patency of the coronary artery; the drug is delivered locally from the stent surface to deal with re-narrowing of the vessel.

Stents are designed to accommodate lesions of different lengths and for use in arteries of different diameters. These are, therefore, the two factors to consider at the outset when designing shelf-life studies for a DES. The study should consider the changes in the properties of the stent (device), the drug and the properties specifically related to the combination of the two.⁹ One way to minimise the amount of testing is to “bracket” the testing so that the extremes in the device design are evaluated and support the intermediate sizes without the need to test these. This is useful for drug related properties such as drug purity and drug elution rate. For drug degradation studies, a useful first approach is to test the sample size, which provides the highest surface area, because any drug-related degradation will be influenced by this parameter. For drug elution, a bracketing approach is useful to minimise the number of product configurations, and it should involve the largest and smallest sizes. Table II illustrates the approach to bracketing.

The risk of bracketing is that any failure can significantly limit the products for marketing until the reason for the failure is identified and rectified. For this reason, shelf-life studies involving the functional tests such as stent retention, crossing profile and balloon burst pressure are generally performed on all sizes in the product range. A shelf-life study for testing the stability of a combination product may, therefore, include a significant number of tests, some of these are

• Physical–mechanical characteristics
• Handling characteristics
• Drug dose
• Drug purity and related substances
• Sterility
• Bioburden
• Endotoxins
• Rate of drug dissolution
• Packaging integrity (primary and secondary).

Table II: Proposed shelf-life bracketing approach. (click image to enlarge)

The requirements for shelf-life testing a DEB are similar in many ways to those for a DES. Because the product is available in a range of sizes, a bracketing approach is again applicable. This is generally designed taking worst case parameters into account. For example, because the product is lyophilised into a dry, free flowing powder, the drug stability can be influenced by the surface area. For this reason, the smallest bead size (the largest surface area) is included to evaluate the drug stability over time.

It can be seen, therefore, that the adoption of a bracketing approach is critical in designing shelf-life studies for combination devices. The amount of testing can be burdensome, especially considering the other factors involved such as temperature, humidity and the various time points.

Regulatory considerations

Table III: Definition of a combination product. (click image to enlarge)

The definition of a combination product has become complex (see Table III) and in an attempt to clarify matters, the United States (US) Food and Drug Administration (FDA) has established the Office of Combination Products (OCP) (www.fda.gov/oc/combination). This agency has issued guidelines to manufacturers to clarify how these products will be regulated. It assigns responsibility for the review of the product to the agency’s Center for Biologics Evaluation, the Center for Drug Evaluation and Research or the Center for Devices and Radiological Health, depending on the determination of the combination product’s primary mode of action (PMOA). This is defined as the product’s most important therapeutic action. It is in the interests of manufacturers who are considering the development of a novel combination product to engage early with regulatory bodies such as FDA to gain feedback on jurisdictional determinations, for example, which Centre will take the lead in reviewing the submission. This can be done in the US through a Request for Designation to the OCP, whereby the PMOA and hence jurisdictional responsibility is defined.

The process is less well defined in Europe, but based essentially on the same premise as in the US: PMOA dictates the manner of the review. It is a manufacturer’s responsibility to demonstrate clearly the rationale for whether the combination product should be considered a drug or a device. For a combination whereby the PMOA is a device, the device element of a submission is reviewed by the Notified Body (NB) and the drug element by the national Competent Authority (CA). Recently, there have been resource issues whereby fewer NBs and CAs are willing or able to review submissions. This is hampered by difficulties in communication between the two distinctly separate departments, which up until now have had no real reasons for cross-functional discussions. Manufacturers of combination products should be aware of this potentially costly delay to CE approval, particularly in a complex borderline case where the wrong decision on classification of the combination could be made. Mounting pressure from industry, combined with provisions for additional resources, prompted action by the medical device authority within the Irish Medicines Board.10 It launched a new service for the sector on 1 January 2008 that offers assessment of drug–device combination products.

Commercial

The drug–device combination product is provided at the point of sale as a complete entity; it will command a higher selling price as a result of the added value of the drug. There is also a subtle switch in the selling strategy for these combinations from the conventional safety and cost-effectiveness arguments of the classical device to clinical efficacy arguments. This requires a sales force with extensive knowledge of not merely the physical elements of the materials used in the device construction, but also of the chemistry and biology of the drug and its mode of action. Moreover, the call point may change for the sale force. The inclusion of a drug therapy may mean that the referring physician, who makes a judgement on the type of therapy employed, is different from the end-user of the product. This can be illustrated in the case of the DEB, which may be administered to a cancer patient by an interventional radiologist, but a medical oncologist may control the patient in the first instance and needs to be convinced of the efficacy of the treatment.

Simplifying development

It can be seen that where there is a sound rationale for local delivery, the combination product has vast potential. To simplify the development process, well known, previously approved drugs with poor toxicity profiles are often ideal candidates for local delivery. A data package will already exist for these therapeutics with their efficacy already proven, which reduces some of the risk associated with a novel combination product. Materials with appropriate physical and biological properties that can modulate the release of drugs are essential to the successful development of a combination product. Alternative product approval pathways open up where these systems allow the drug to be added at the point of use. There is, however, an unavoidable increase in the clinical and regulatory burden when a combination product is used, which should not be underestimated. Regulatory processes are different for each market and are currently lacking definition in the European Union. Because of the challenges and complexity of these products, the market size must be sufficient to justify the added expense of approving a product. The combination may offer, however, a valuable route to extending drug product lifecycles and the potential for new intellectual property.

Sean Willis is Development Director and Andrew Lewis is Research and Technology Director at Biocompatibles UK Ltd, Chapman House, Farnham Business Park, Weydon Lane, Farnham GU9 8QL, UK, tel. +44 1252 732 732, e-mail: sean.willis@biocompatibles.com, www.biocompatibles.com.