……but is a means to an end, the end being the correlation of powder properties with some process of manufacture, usage, or preparation.”

Not a quote of mine unfortunately, but one of my favourites from the estimable Professor Harold Heywood, which is why I used it to begin my presentation at the USP workshop on ‘Particle Size: Detection and Measurement’ in Rockville, last week.  For those of us working at Malvern this is a statement that reminds us that our role is to serve!

More generally however, it’s a sentence that focuses attention on the crucial elements of an analysis: What are you going to use the data for, and how are you going to ensure that they are appropriate? The pharmaceutical industry measures particle size to control product quality and clinical efficacy. This means they must develop sizing methods that are reliably and sufficiently sensitive to detecting changes that have an impact on either.

Who’s counting?

I was at the USP workshop to talk about laser diffraction particle size measurement and the statistics of particle sizing, responding to the general recommendations made with USP<429> (Laser Diffraction Measurement of Particle Size) and USP<788> (Particulate Matter in injections).

I had the honour of presenting twice, with my first paper discussing how USP<429> is interpreted and applied. Taking into account recent feedback from the Particle Summit and elsewhere, I chose to take things back to basics, looking at how laser diffraction analysis delivers particle size information.

Pharma has a historical reliance on microscopy which, unlike laser diffraction, is a particle counting technique i.e. you look at each particle and then present a statistical summation that describes the sample. Laser diffraction isn’t like this. It quickly measures the whole sample at once, which is why it’s called an ensemble method. The detected light scattering pattern is de-convoluted using light scattering theory to determine the particle size distribution that produced it.

You can productively switch between particle counting techniques such as microscopy and image analysis and ensemble techniques like laser diffraction, depending on the application, but you do need to be careful how you compare data. We looked at this issue as I stepped through the process of method development for laser diffraction, comparing and contrasting the guidance provided by USP<429> and ISO13320.  The standards differ in the extent of the advice they offer on different topics, stimulating debate on possible areas of improvement for the USP. Discussions contrasting wet and dry dispersion and the role of imaging in method development were particularly well-received.

The three R’s

For my second presentation, I returned to the theme of the three R’s:  reproducibility, repeatability and robustness, this time looking at the role of statistics and sampling in particle size measurement, including:

  • Different ways of calculating average particle size
  • The role of stats in determining sampling procedures
  • Statistical process control / PAT

Another favourite quote, this time from Francis F Pitard is: “Your decisions are only as good as your samples.” Sampling is often the biggest source of error in particle size measurement so understanding how to get this part of the analytical process right is really important.

Overall the USP conference was extremely stimulating. Its very existence underlines the importance of particle size within this key industry and the audience were uniformly engaged in learned debate. Developing the very best guidance for the industry is a core goal of the USP, one that I’m sure will be aided, in this vital area, by this excellent event. Personally, met old friends and colleagues and learnt a lot about the application of the techniques I’m involved in.  After all, I’m still an apprentice in the game after 20 years at Malvern.  Not sure what I’ll do when I grow up!