XEDS-mapping for explaining release patterns from single pellets

Abstract

A common way to formulate controlled-release (CR) pharmaceuticals is to coat pellets of active substance with a polymer film, decrease the size of the pellets and distribute them as multiple-unit dosages in capsules. To increase the understanding of the release mechanism, the pellet shape and surface structure of pellets, before and after release in microtitre plates, have been studied by scanning electron microscope and X-ray energy-dispersive spectrometry. By performing these studies we associate release profiles during the first few hours to the microscopic structure. Pellets were divided into three classes (spherical pellets, dumbbell shaped pellets and twin-pellets) according to pellet form. Cases of burst release occurred for all three shape classes due to “open-window-defects” at the surface. Areas of thinner polymer film in the neck-region of dumbbell shaped pellets broaden the range of intermediate release rates for this pellet shape. The surface of twin pellets and dumbbell shaped pellets showed more defects, which increases the release rates in comparison to spherical pellets. All pellets with high release rates revealed ruptures in the polymer film, whereas only small cracks could be traced for pellets with slow release rates. The information gained is necessary for the development of future formulations and mathematical modelling of release patterns. The pharmaceutical used as model was remoxipride coated with a polymer film of ethyl cellulose and 10 wt.% triethyl citrate.

Introduction

The purpose of this work is to describe how SEM, scanning electron microscope, in combination with XEDS, X-ray energy-dispersive spectrometry, can be used to shed light on release mechanisms as a part of developing new drug formulations or as a method to explain large unexpected variations in the release rate from single pellets.

The number of controlled release pharmaceuticals is increasing as they offer more convenience for the patients and lower risks for side effects. Usually, the release is controlled by coating a core containing active substance with a polymer film. When distributing such polymer coated pellets as multiple units the mass transfer area for release increases and the risk for dose dumping from defect pellets decreases (Aulton, 2002). This makes multiple unit dosage forms more reliable for a time controlled release than single unit dosage pharmaceuticals (Schultz and Kleinebudde, 1997).

The release behaviour of multiparticulates is determined by the individual pellets (Dappert and Thies, 1978, Jorgensen et al., 1997, Folestad et al., 2000, Sirotti et al., 2002). Since the individuals have different kinetics than the total population, single pellets have to be studied to understand the release mechanism (Dappert and Thies, 1978, Hoffmann et al., 1986a, Hoffmann et al., 1986b). Results from single pellet experiments can then be used to simulate the release behaviour on dose level (Borgquist et al., 2002, in press). There is an active research in simulations of drug release from multiparticulates, since it is assumed that simulation can save time and money in the process of developing new pharmaceuticals, by quickly predicting release patterns through a wide parameter space (Sirotti et al., 2002). Simulations also give insight into which parameters that have the greatest influence on release pattern for active feedback into modifying processes (Borgquist et al., 2002). However, most simulations and mathematical descriptions of release are based on release experiments on a complete dose level (Ragnarsson et al., 1992, Zackrisson et al., 1995, Folestad et al., 2000, Sirotti et al., 2002, Lectome et al., 2003). This makes it difficult to correlate experiments to model simulations when sample variation is unknown on the single-pellet level (Folestad et al., 2000, Lectome et al., 2003), e.g. osmotic pressure effects and ruptures in the polymer coating (Lectome et al., 2003).

There are few previous publications describing single pellet release measurements, the first reported by Hoffmann et al. (1986a). Subsequently, Jorgensen et al. (1997) measured release from single pellets in specially designed vessels, and Schultz and Kleinebudde (1997) measured release from single pellets in flow-trough cells. Previous work has also been performed using the same apparatus for single-pellet release utilised in this work (Folestad et al., 2000, Borgquist et al., 2002). A common method to study the release from single particles have been by measuring conductivity or absorption spectrometry in small cells or wells (Hoffmann et al., 1986a, Hoffmann et al., 1986b, Folestad et al., 2000). Among the parameters that have been shown to be important for determining the release are surface area, film thickness, diffusivity, coating porosity and pellet shape. The aim when producing pellets for multiple unit systems is to achieve spherical pellets, as this shape gives more free-flowing pellets during filling of capsules (Chopra et al., 2002b) and a more predictable release pattern. The filling of pellets into capsules is only possible when the aspect ratio of the pellets is below 1.2 (Podczeck et al., 1999). Thus, spherical pellets are often assumed when describing drug release from controlled release subunits (Ragnarsson et al., 1992, Sirotti et al., 2002) even though this may be far from the real picture.

Previous work on the influence of pellet shape on release behaviour, have been based on comparison of bulk release in USP-vessels. Lorck et al. (1997) assessed pellet morphology by dividing them into rough and smooth pellets, where spherical pellets were classified as smooth, and all other shapes were denoted as rough. They noticed that rough pellets had a higher release rate than smooth pellets after about 2 h of release (Lorck et al., 1997). Chopra et al. (2002a) intentionally processed batches differently to obtain shape variations, neglecting variations of shape within each batch. These investigations prove that there is a clear effect of the shape of the pellet but does not provide any detailed explanation why.

Indications of how shape influences release behaviour can be approximately predicted at dose level, but, as the dose behaviour is governed by the behaviour of the single subunits, the exact cause for unintended variations in dose behaviour must be determined and used at the subunit level (Folestad et al., 2000). According to Donbrow et al. (1988) every possible release pattern can be designed by mixing single pellets with different release behaviours.

As the pharmaceutical coating methods are moving from organic solvent based methods to water based, there is an urgent call for methods that can explain causes for variations in release pattern. Compared to organic solvent-based coatings, aqueous coatings require higher expenditure for process optimization and process validation (Lorck et al., 1997, Larsen, 2004).

Hoffman noticed in 1986 localized internal dissolution of cores coated with ethyl cellulose providing evidence of release through pores in the film (Hoffmann et al., 1986b). By investigating Roxiam^® CR with AFM, atomic force microscopy, during release, Ringqvist et al. (2003) noticed crystalline material on the surface of the pellets that instantaneously dissolved. This proved the possibility that active substance embedded in the polymer film results in pores through which release of core material occurs. The drawback of AFM in comparison to electron microscopy is that it does not provide information on variation in chemical composition. The element distribution in the material can easily be detected by analyzing the X-rays that result from the interactions of the electrons with the substrate in a SEM-instrument (Lyman et al., 1990).

The use of scanning electron microscopy to study the morphology of controlled release pellets is already described in the literature (Schultz and Kleinebudde, 1997, Chopra et al., 2002b). The advantage in comparison with light microscopy is the better resolution that SEM microscopes offers. Previously, light microscopy have been used for investigation of pellet size, coating thickness (Gunder et al., 1995, Borgquist et al., 2002) and pellet surface structures (Gunder et al., 1995, Lorck et al., 1997), but light microscopy was found to be insufficient for the features of essence in this study. Until now, X-ray emission generated in the SEM has been neglected for the use of identification and localisation of the leaking active substance. The major prerequisite for this method is that the pharmaceutical has an element specific for the active substance and not the coating, something that is true for most cases. XEDS-analyses can be used to trace active substance on the surface or in cracks where active substance is exposed.

In this study the effect of pellet shape on release profile distributions for single pellets has been investigated for the multipellet pharmaceutical Roxiam^® CR produced in pilot plant experiments on a large scale using a Wurster coater. The release profile distributions determined here have already successfully been used in a recent investigation to simulate release at dose level (Borgquist et al., 2004). The pellets in Roxiam^® CR are coated with a film of ethyl cellulose. To confirm the formation of pores through the film during release, the pellets have been investigated by SEM and XEDS before release, and after 1 h release. The release profile distribution, obtained in the conventional way, showed that 1 h was sufficient to state the release behaviour of that individual pellet. The same individual pellet was then investigated by scanning electron microscopy and XEDS. This made it possible to correlate release behaviour with defects of the polymer film.

The variability shown and discussed in this study is most probably present also in production scale as the experimental material was produced in a fairly large pilot plant unit.