Blending goes hi-tech
Light-induced fluorescence, photon migration and NIR are terms that don't seem to have much in common with mixing and blending of APIs and excipients. But, as Graham Lampard discovered, they are used to ensure better blend uniformity
Light-induced fluorescence, photon migration and NIR are terms that don't seem to have much in common with mixing and blending of APIs and excipients. But, as Graham Lampard discovered, they are used to ensure better blend uniformity
Validating the uniformity of a final powder blend continues to be a highly debated topic within the pharmaceutical manufacturing community. 'Poor blending or an inability to maintain a blend, i.e. segregation, at the point it's needed is costly in terms of rejected material and defective end product.'1
Following an FDA enforcement action against Barr Laboratories 2, it appeared that the record-keeping expectations for a finished pharmaceutical manufacturer were being applied to all stages of bulk production. The FDA introduced draft guidelines entitled 'ANDAs: blend uniformity analysis', which are still going through the FDA's advisory committee for pharmaceutical science. The current status and future plans for the guidelines were discussed at last November's meeting.3 The FDA indicated that there was a strong need for routine blend uniformity testing, following process validation, to ensure content uniformity within the blend: it is insufficient to show that adequate distribution of the drug is obtained in the final product – it must be demonstrated within the blend as well.4
How can this be achieved? The Product Quality Research Institute (pqri) issued a proposal that demonstrated how both blend and dosage unit uniformity could be achieved. Its blend uniformity working group (buwg) is currently collating the responses to the proposal.5
However, the methods it suggests do not use the newer on-line, in-process measurements, so homogeneity of the final product is measured by sampling and off-line analysis of the powder. Removal of samples is currently influenced by the skill of the operator and often provides a false representation of the sample due to desegregation and disruption of the powder bed during sampling.
For this reason, new technology is being developed to fulfil and perhaps replace the conventional thieving method with more rapid and consistent spectroscopic techniques.
Prof CK Lai at the Massachusetts Institute of Technology has developed a portable system using light-induced fluorescence technology (LIF) which monitors the progress of powder homogeneity non-invasively and in real-time by irradiating samples and evaluating the emission. Many drugs fluoresce when excited so LIF could be used to monitor blend homogeneity. It is a rapid, on-line method that examines the details of blending kinetics and the effect of blending.
The portable LIF system has been successfully field-tested, and blending homogeneity was demonstrated over a range of drug concentrations from 0.01-5% (w/w). The relative standard deviations of the runs were consistently low and always below 3%. The success of this technology was verified on ~200kg industrial scale.
This technology opens the door to a wide area of applications, and as Lai pointed out: 'Non-invasive technologies are the obvious direction in which the industry is focusing. The two leading methods for the specific monitoring of blend homogeneity of an API are the LIF and near-infrared (NIR) technologies.
'However, the challenge of NIR is that it records all spectrum of the ingredients present, and due to constant background interference, there are limits to the dosage of the API that can be tracked with NIR,' he said. Detection limits for NIR need about 5% API to be present, whereas with LIF, if the compounds of interest fluoresce, detection limits of 0.1% API are fine.
Where the API of the drug does not fluoresce, it is also possible to track the blending profiles based on the emission signal from the excipient background.
Other technologies are also being developed, such as photon migration and thermal effusivity. Both of these can detect blend homogeneity in general but lack the required sensitivity.
The main problem is that these technologies track the status of the blend mixtures only at the surface rather than the entire bulk.
Lai said that a bigger problem was one of conservatism. 'This would be a monumental jump for an industry that is so highly regulated by the FDA that most managers adopt a conservative attitude towards these new technologies.
'We would envisage that the new on-line monitoring technologies would run side-by-side with traditional techniques until the technology had been verified.'
In contrast to other optical methods, frequency domain photon migration (fdpm) consists of monitoring the time-dependent propagation characteristics of multiply scattered light rather than the amount of light detected.
Dr Eva Sevick-Muraca, Texas A&M university, said: 'The technique consists of illuminating the powder bed with light that is intensely modulated and detecting the phase-delay and amplitude attenuation of the light that has successfully propagated some distance within the bed.' 6
While these new techniques are for the future, the industry must decide whether to continue with unit dose sampling, an approach that has attracted criticism for generating false results, or use a stratified sampling technique as recommended by the pqri. Whatever is decided, with the FDA insisting on tighter regulations the industry must have a decisive answer.
What is blending? - An FDA definition
The FDA defines blending as the process of combining materials within the same specification to produce a homogeneous intermediate or API. In-process mixing of fractions from single batches (e.g. collecting several centrifuge loads from a single crystallisation batch) or combining fractions from several batches for further processing is considered to be part of the production process and is not considered to be blending.1 It continues: Out-of-specification batches should not be blended with other batches for the purpose of meeting specifications. Each batch incorporated into the blend should have been manufactured using an established process and should have been individually tested and found to meet appropriate specifications prior to blending." Acceptable blending operations include: