That figure of 263 times greater reduction in viral load daily seemed implausibly high. If baseline reduction is 1%, then with interferon, it would be over 100% reduction, or to zero, in one day. If baseline reduction is not as high as 1%, placebo patients would need many months to recover. I'm also dubious about measuring daily reduction in viral genome count, because the underlying assumption there is that this number is constant, and it shouldn't be. As the immune system responds, it would start fairly low, but the rate would naturally increase as B-cells start pouring out antibodies and T-killer cells start multiplying.
I then searched the Lancet paper, and the number 263 does not appear in it. Typo?
The mask recommendations are missing a key thing: ratings for non-medical masks. I can buy any of a hundred "medical style" masks that look like surgical masks, but without actual regulated testing, I have no way to know their actual filtration ability, and without different testing I have no way to know how likely they are to fit tightly. Creating a rating system for this purpose would be one of the best things the FDA (or WHO) could do, in my opinion.
Hi Carl--you're right, I misread the results here. The ultimate difference by day 7 was 263 times greater, which I read as a difference in daily change. Specifically, there was a difference of 2.42 logs between the two groups at day 7; 10^2.42 is approximately 263. I gave the strict factor because I did not think most folks are used to thinking logarithmically. I'll correct the newsletter regarding when this difference appeared--nice catch. It seemed strange to me too, to be honest, but I guess the language in the paper was a little confusing.
I agree with you that a rating system for masks and mask fabrics would be great.
I don't know virology, but like any Heinlein reader[1] I have developed a decent ability to do orders-of-magnitude sanity checking of numbers. This one really seemed impossible.
The drug actually seems very promising. Very good news, if it pans out. Hey, question for our local Pharma Shill[2]: how expensive are interferon III drugs? Would this be like the monoclonal antibody things (thousands per dose) or like dexamethasone (very inexpensive) or in between (likely)?
[1]I know John through science fiction fandom. For non-fans, Robert A. Heinlein was an engineer turned SF writer, and a lot of engineer thinking can be found especially in his earlier work.
[2]"Pharma shill" is what conspiracy theory/natural medicine cultists call anyone who says a drug or vaccine actually works. Since cultists "know" that pharmaceuticals don't work, the person has to be lying, probably because they were bribed by the pharmaceutical industry. Obviously, I don't think John is lying, so that was a joke that I just ruined by explaining it.
Regarding [2], I used to call myself a "hired gun in the drug world" when I was a consultant, so I take that in the spirit you intended :) I guess now that I'm a permanent headquarters employee I'm a high-level lieutenant in a drug organization.
The order of magnitude calculations in virology are always interesting. Sometimes you do see really staggering numbers, so while this seemed odd to me it wasn't completely implausible enough for me to give it the fourth read it would have taken to catch the error. I'm glad you did, though.
Regarding the cost of delivery, I can't be *too* sure, but I do think that it would perhaps be a bit cheaper than monoclonal antibodies for a few reasons. (1) I don't think IFN lambda has substantial essential post-translational modifications, so it may be something that can be synthesized in a recombinant bacterial system rather than in a eukaryotic hybridoma system like antibodies typically require. Right there, there's a substantial savings in the complexity and finickyness of the bioengineering. (2), it might be possible to lyophilize the peg-IFN-lambda, which would help make it more shelf stable. Currently the REGN-COV2 polyclonal cocktail is delivered as a refrigerated suspension, I believe. So there's a potential logistics savings there too. (3) antibodies are usually delivered by infusion, so you need an infusion specialist or other highly skilled technical medical professional to deliver them; the IFN-lambda preparation used here was a subcutaneous injection, and might even be able to be loaded into an autoinjector, so the skill and infrastructure level required will be lower. The IFN-lambda might be something your primary care provider can easily deliver, while the monoclonal preparation might not be (this also depends on how sophisticated or PCP's office is, so I'm using a lot of qualifiers).
Anyway I do think this will be cheaper than an antibody drug for a variety of reasons, if it does pan out. I would expect it to be more on par with insulin or pegylated-IFN-alpha in its costs, and if I'm not mistaken both of those are in the hundreds of dollars per dose in the US at least. Not that that's a bargain either, but it is less.
That figure of 263 times greater reduction in viral load daily seemed implausibly high. If baseline reduction is 1%, then with interferon, it would be over 100% reduction, or to zero, in one day. If baseline reduction is not as high as 1%, placebo patients would need many months to recover. I'm also dubious about measuring daily reduction in viral genome count, because the underlying assumption there is that this number is constant, and it shouldn't be. As the immune system responds, it would start fairly low, but the rate would naturally increase as B-cells start pouring out antibodies and T-killer cells start multiplying.
I then searched the Lancet paper, and the number 263 does not appear in it. Typo?
The mask recommendations are missing a key thing: ratings for non-medical masks. I can buy any of a hundred "medical style" masks that look like surgical masks, but without actual regulated testing, I have no way to know their actual filtration ability, and without different testing I have no way to know how likely they are to fit tightly. Creating a rating system for this purpose would be one of the best things the FDA (or WHO) could do, in my opinion.
Hi Carl--you're right, I misread the results here. The ultimate difference by day 7 was 263 times greater, which I read as a difference in daily change. Specifically, there was a difference of 2.42 logs between the two groups at day 7; 10^2.42 is approximately 263. I gave the strict factor because I did not think most folks are used to thinking logarithmically. I'll correct the newsletter regarding when this difference appeared--nice catch. It seemed strange to me too, to be honest, but I guess the language in the paper was a little confusing.
I agree with you that a rating system for masks and mask fabrics would be great.
I don't know virology, but like any Heinlein reader[1] I have developed a decent ability to do orders-of-magnitude sanity checking of numbers. This one really seemed impossible.
The drug actually seems very promising. Very good news, if it pans out. Hey, question for our local Pharma Shill[2]: how expensive are interferon III drugs? Would this be like the monoclonal antibody things (thousands per dose) or like dexamethasone (very inexpensive) or in between (likely)?
[1]I know John through science fiction fandom. For non-fans, Robert A. Heinlein was an engineer turned SF writer, and a lot of engineer thinking can be found especially in his earlier work.
[2]"Pharma shill" is what conspiracy theory/natural medicine cultists call anyone who says a drug or vaccine actually works. Since cultists "know" that pharmaceuticals don't work, the person has to be lying, probably because they were bribed by the pharmaceutical industry. Obviously, I don't think John is lying, so that was a joke that I just ruined by explaining it.
Regarding [2], I used to call myself a "hired gun in the drug world" when I was a consultant, so I take that in the spirit you intended :) I guess now that I'm a permanent headquarters employee I'm a high-level lieutenant in a drug organization.
The order of magnitude calculations in virology are always interesting. Sometimes you do see really staggering numbers, so while this seemed odd to me it wasn't completely implausible enough for me to give it the fourth read it would have taken to catch the error. I'm glad you did, though.
Regarding the cost of delivery, I can't be *too* sure, but I do think that it would perhaps be a bit cheaper than monoclonal antibodies for a few reasons. (1) I don't think IFN lambda has substantial essential post-translational modifications, so it may be something that can be synthesized in a recombinant bacterial system rather than in a eukaryotic hybridoma system like antibodies typically require. Right there, there's a substantial savings in the complexity and finickyness of the bioengineering. (2), it might be possible to lyophilize the peg-IFN-lambda, which would help make it more shelf stable. Currently the REGN-COV2 polyclonal cocktail is delivered as a refrigerated suspension, I believe. So there's a potential logistics savings there too. (3) antibodies are usually delivered by infusion, so you need an infusion specialist or other highly skilled technical medical professional to deliver them; the IFN-lambda preparation used here was a subcutaneous injection, and might even be able to be loaded into an autoinjector, so the skill and infrastructure level required will be lower. The IFN-lambda might be something your primary care provider can easily deliver, while the monoclonal preparation might not be (this also depends on how sophisticated or PCP's office is, so I'm using a lot of qualifiers).
Anyway I do think this will be cheaper than an antibody drug for a variety of reasons, if it does pan out. I would expect it to be more on par with insulin or pegylated-IFN-alpha in its costs, and if I'm not mistaken both of those are in the hundreds of dollars per dose in the US at least. Not that that's a bargain either, but it is less.