Thread by @Dereklowe

I appreciate Eric saying this after some of us sort of jumped on him about this issue last night! Paxlovid is a good example of where small-molecule synthesis falls on on the difficulty scale when compared to things like mRNA vaccines - here's a thread (1/13)

https://twitter.com/EricTopol/status/1474019309349265414

The mRNA vaccines are intrinsically hard to manufacture - RNA biotechnology on scale is practiced in only a few countries around the world, and the nanoparticle formulation is particularly challenging to scale up and to get right. (2/13)

Small-molecule synthesis (though no picnic) is easier than that - there are generic drug manufacturing facilities in many more countries, and many of these can potentially adapt to processes like the Paxlovid synthesis. That's the good part. (3/13)

But the tricky part becomes the supply of the various starting materials and reagents. We even saw that in the small molecule part of the RNA vaccine scale-up, when the cationic lipids became a production bottleneck for a while. (4/13)

A big difference is that proteins, DNA, and RNA will sort of make themselves, if you can do the cell culture and molecular biology involved (which is a big "if" on the industrial scale). We're co-opting a billion years of evolution for our own purposes there. (5/13)

Small molecules, though, come in far, far more varieties than the repetitive building blocks of proteins and oligonucleotides. The number of potential structures is effectively unlimited, and the number of potential industrial starting materials is likewise huge. (6/13)

We organic chemists have learned a lot of tricks about how to make small-molecule drugs, and the process chemists and chemical engineers have a whole specialized expertise in doing synthesis on a huge scale. Most of the time, we can come up with robust, workable routes. (7/13)

But you have to source the starting materials if you're going to do that - and the *huge* majority of such building blocks and reagents come from industrial producers in places like China and India. Very few of those things are made in the US. (8/13)

So invoking the Defense Production Act (for example) is of limited use. The chemical plants you'd want to commandeer just aren't there. A big part of industrial drug production is lining up complex supply chains and contract manufacturing around the world. (9/13)

A given drug might travel through two, three, four countries on its way to be packaged into pills, with different synthesis and manufacturing steps done in the places with the facilities and the expertise. Finding these partners is close to an art form. (10/13)

Now for Paxlovid, you can bet that Pfizer has been putting these blocks together for the last few months. They have a lot of experience doing this, but this means that the obvious suppliers of starting materials have already been lined up. (11/13)

Since there are so many potential starting materials in this business, it's easy to use up the capacity for any given one if there's a surge in demand. That slack can be picked up by other chemical suppliers, but it takes time, money, and effort. (12/13)

We'll get Paxlovid made, and made in many locations around the world (Pfizer has licensed the patent out). But it's going to be a scramble to get capacity up under these zero-to-thousands-of-tons conditions, and there are no easy ways (like the DPA) to fix that quickly. (13/13