approaches being pursued will be presented from a comparative design point of view and will cover both opportunities and challenges that are on the road ahead.

13.2  General Design Considerations

13.2.1  Administration Site

To consider a refilling system in the eye, an obvious requirement is that some element of the device must be easily accessible to implement the refill. Either the refill port is built into the device and the whole device or port can be sufficiently visualized to be reached with a needle, or the port is connected to a cannula or channel which is fed from the device to an accessible visible region. The possible ocular port locations that are within visually observable domains are illustrated in Fig. 13.1. Anterior spots which can be considered are subconjunctival, sub-Tenon’s space, intracorneal, intracameral, and intracapsular. While all those positions are adaptable to housing not only the port, but the device itself, it does not limit the imagination to consider other device locations that are linked to the port through a fluid channel or pathway. Certainly the device location could be designed to be proximal to the intended target tissue and thus implanted in sites such as intravitreal or subretinal, for example. In the case of separated port and device locations, this is likely to involve greater complexity designing how the channel may have to traverse through other tissues to reach the implanted device.

The location of the main body of the device is an initial factor which governs overall sizing of the device. For example, with intravitreal implanted devices, the placement to avoid interference in the visual path is critical. Devices in the vitreous which are anchored at the pars plana usually are restricted to no more than about 6 mm of length in order to avoid being in the line of sight. While the diameter or width can vary up to several millimeters, the desire to conduct smaller surgical incisions would suggest designs with diameters of no more than 1 or 2 mm. However, a cylindrical device with diameter of 2 mm and length of 6 mm can only accommodate 0.0188 cm3 of volume (i.e., 18.8 mL). This limitation highlights a second factor which governs feasibility of the size, that is, the reservoir volume needed to accommodate sufficient drug concentration over the desired delivery period. Using Tables 13.1 and 13.2 in concert, an understanding of the minimum delivery chamber size can be garnered based on the daily drug potency requirement, the drug concentration, and the desired delivery period. As can be deduced from the tables, small-sized reservoirs are possible if the required in vivo potency is high or if the drug can be formulated at high concentration. In certain cases, such as with proteins, high concentrations can lead to instability. Therefore, shortening the refill duration or using a design with the reservoir in a different anatomic location may offer other options. In this regard, the subconjunctival and sub-Tenon’s spaces provide much greater capacity for a larger device. In these regions, the device height will be flattened to fit under the tissue, however the device body can cover a much larger surface area, thus accommodating significantly greater volumes (a coin-shaped device with diameter of 1.26 cm and height of 4 mm will accommodate approximately 0.5 mL of volume).

13.2.2  Body Design

Selected components for the main body of the device should possess a number of important features. These include: (a) long-term biocompatibility if in contact with tissue, (b) chemical compatibility with the active ingredient or excipients if in direct contact or flow path, (c) low extractable or leachable impurities into the drug product,

(d) material stability following sterilization, (e) stability to environmental influences such as light and oxidation if device parts are externally exposed, (f) stability to pressure or externally applied physical forces such as digital manipulation, and (g) other functional utility as applicable. Among the potential durable materials which may meet part or all of these requirements include but are not limited to metals or alloys such as titanium, tantalum, niobium and nitinol, plastics or polymers such as polyimide, polyetheretherketone (PEEK), parylene, polytetrafluoroethylene (PFTE), polypropylene, polyethylene vinyl acetate, and polyethylene terephthalate, elastomers and sealants such as silicone, medical grade epoxy and glass ionomer and finally, various ceramics such as aluminum and titanium oxides.

Selection of the materials is usually made based on the particular function within the device or location within the tissue. Protective encasements of sensitive electronics are best provided by nonmalleable inert materials such as metals or hard plastics while the more elastic or flexible components are usually relegated to spots requiring dynamic valves or alloplastic conformity with tissue morphology. For the latter functions, silicones are often a first choice because of their diverse range of durometers, tensile strengths, and elastic modulus.

It is important to understand the chemical and physical properties, stability, and functionality of the materials following the chosen sterilization method. Sterility by terminal methods will be the expected first approach by the regulatory agencies. If acceptable validated methods such as 25 kGy of irradiation are not viable from a functional or material stability standpoint, other methods or approaches will need to be validated to show sterility through the entire device, especially those components in direct contact with the active agent. Inertness to effects of radiation, thermal stress (dry heat or steam), and chemical penetration (i.e., ethylene oxide) vary by polymer. For example, where PFTE has excellent thermal and chemical inertness it is dramatically affected by gamma irradiation. In contrast, polyimides and parylenes have much greater resistance to irradiation effects.

13.2.3  Port Design

The operation of a system that allows for a liquid refill must be constructed to allow for introduction of a needle or cannula without backflow or reflux. In addition, the port must withstand multiple piercings and be able to reseal consistently over time. Thus, resistance to coring phenomenon should be included as a design factor. Furthermore, the design consideration for the selection of port material must account

Fig. 13.2  Elastomer formations to facilitate resealing after puncture (a) webbing structure. Reprinted from Dalton (1989) and (b) preslitted depression. Reprinted from Levy (2004)

for the frequency of reinjection, the age of the patient, the in vivo life of the total device, and overall resistance to biodegradation. The historical development of injection ports comes mainly from the development of septums in general laboratory operations, particularly in chromatography vial applications. The examination of self-sealing elastomers focused on capability of punctured septums to resist evaporation of volatile solvents (Adler 1964). Such studies evaluated elastomers such as chloroprene, isoprene, isobutylene, silicone, polyurethane, vinylidine fluoride/ hexafluoropropylene, and chlorinated polyethylene. In common practice, silicone elastomers offer a good combination of resealing capability along with resistance to coring. Coatings on the silicone such as PFTE can add further chemical inertness, a property exploited in current septum designs for laboratory applications. But while PFTE is highly inert, by itself it does not possess resealing capability. As such, there is continuing work on development of inert co-polymers with PFTE such as perfluoro (alkyl vinyl ethers) that have low levels of extractables but which can reseal after puncture (Sassa et al. 2009). In addition to the biomaterial properties affecting the sealing characteristics of elastomers, there also have been design variations in the formation of the elastomers such as webbing or preslitted depressions which facilitate the reseal (Fig. 13.2).

13.2.4  Vacuum and Pressure

As most pump devices are going to include some form of check valve system on the output side to prevent reflux of bodily fluid into the device, the internal refill chamber functions as a closed system during operation. As fluid is pumped out of the chamber, without some form of concurrent gas or fluid replacement, the creation of a vacuum ensues which can lead to collapse of the chamber, depending on its flexibility or construction. In addition, the force required to pump fluid out of the device increases as the vacuum pressure increases within the chamber. Design elements that have been used to deal with this issue are counterbalance with a concurrent gradient of pressure applied external to the chamber (gas or fluid driven) or via