Therapeutic Dose: Facilius per Partes in Cognitionem Totius Adducimur

I take part in a number of “chat sites” on pharma and biopharma. Many of the topics are rather interesting, although I don’t always contribute. I was reminded of an ongoing conversation when my wife asked a question about how detergent was made. The simplest and most common is sodium docdecyl sulfate (sodium lauryl sulfate or SLS). It is in everything from laundry detergent to toothpaste and shampoo. The synthesis is pretty simple: a fatty alcohol (mainly C-12) is reacted with chlorosulfonic acid (CSA). It is then neutralized with NaOH solution and diluted to the desired strength.

The reason this is salient is that the profit margin is hardly large (but, if you want to sell other products, the SLS must be competitive), much like generic drugs today. One way that the Japanese optimized the process was to make it continuous. They had a string of rail cars with the fatty alcohol, another string with CSA, and a third string of cars with caustic. The first and second strings of cars were reacted in a continuous stream for a set time, the caustic was added and the resultant liquid was pumped into a fourth string of (empty) rail cars. This was in the mid-1970s!

The process set me to thinking about continuous processing in the pharmaceutical arena . . . akin to imagining what a unicorn might sound like. A number of my “pharma-friends” and I have been discussing why we don’t have long continuous batches instead of discrete, smaller batches. One reason, as I see it, is the expense, time, and number of personnel needed to make a full-blown continuous process stream.

That was my argument to Ajaz Hussain during the 2002 PAT meetings in Maryland. He was all for, in essence, shutting down an entire process line, spending a lot of time and money to fully automate and validate it, then proceed full-blown to perform PAT. My view was to take small bites (my mother’s advice again) at first. It is also known as going for the low-hanging fruit. I suggested that a group/company first set up a mechanism for incoming raw material quality (as we did in 1984 at Sandoz) with something like NIR. THEN look at mixing, THEN look at drying, THEN look at granulation, THEN worry about tabletting, THEN look at the coating process.

After each success (starting with what has been proven over and over to work), management will be more likely to fund the next project, and so on. Within a reasonable period of time, why, goodness, looks like we’re doing PAT! (Remember the frog in cool water being heated slowly?)

How does this relate to continuous processing? In my view, we can explore each step of a process in the same manner as PAT. Suppose we concentrate on mixing first. We now have tools (NIR, Raman, LIF) to assure that the mix is, indeed, mixed. The logic of making one “batch” of powder after another in the same mixer, without the cleaning and cleaning validation time, seems a winner. A number of products are made as direct compression tablets and, in theory, can be continuously made by simply pouring batch after batch of mixed materials into a large hopper above the tabletting machine.

While the actual mixing and compression times, per se, are the same as before, the downtime between cleaning and verifying cleanliness is gone. From what I’ve seen, that is 75% of the time needed to produce a series of products. With the monitors developed for PAT, it seems that the uniformity of the blend and CU of the tablets could be monitored continuously, also. (For those who ask about “batch identity,” perhaps a variation in the lot number could be made, say, every twelve hours?)

It would seem that, if we concentrate on time-savings and control over each step of production, then continuous processing becomes a logical outgrowth of QbD and PAT. We will be learning at each step (as we did with PAT), just as we learned in college. We didn’t have a four-year course called “Chemistry,” did we? No, we had Intro, Organic, Analytical, Physical, Inorganic, and so forth. It can’t hurt to learn and practice one step at a time . . . just like any learning process for playing an instrument, ice skating, chemistry, or continuous processing. It comes down to really WANTING to do it.

1. We are more easily led part by part to an understanding of the whole. –Seneca