Getting to the bottom of potential toxicity variability between Vax Lots

The ongoing mystery... You know what would put an end to the mystery? Transparency and good conduct.

I am trying to figure out if certain COVID-19 injectable product VAX LOTS are more toxic than others. There is a quality control document of the COVID-19 mRNA Vaccine BioNTech BNT162b2, 5’capped mRNA encoding full length SARS-CoV-2 Spike protein product (say that 5 times fast) that reports that there are ~1,545,000 doses per batch size. This is based on multiplying 2 numbers: the number of doses per vial (5) and the target drug batch size (309,000). I am 99% sure that when they write ‘batch’ they mean ‘VAX LOT’. Let’s make this assumption.

But, it is not clear, by their own admission, what variablity between batch sizes actually exists.

They were meant to address this according to the quality control assessment.

Let’s make the general assumption that we don’t know if the VAX LOTS are the same size and let’s also make the general assumption that we don’t know about potential content variability meaning that we don’t know if the vials contain the same stuff - including complete mRNA templates. This is a pretty safe assumption, also based on this quality control assessment. By the way, the mRNA variablity between vials is a huge separate stinking kettle of rotting fish. I will write something on this soon.

Let us define ‘toxicity’ as the rate of adverse event occurrance per number of doses. This is actually the crux of most misunderstandings on this subject I believe: in many instances, toxicity is being equated to a number when it should be equated to a rate.

Let us take 2 real VAX LOTS and use them in a demonstration of how certain VAX LOTs appear to be causing more death than others. Let’s use only Pfizer VAX LOTs, for simplicity’s sake: VAX LOT EN6201 - that has been associated with 114 deaths so far and VAX LOT EJ1685 that has been associated with 50 deaths so far. Let’s assume, for now, that both of these VAX LOTs have the same number of doses that have been delivered to the United States people - 1.545M, to be specific, and to keep in line with Pfizer’s quoted dose/VAX LOT number as per the quality control assessment document.

We can calculate the toxicity (the rates of death per number of doses) by dividing the respective number of deaths by the number of doses. So toxicity for EN6201 = 114/1.5M*1M = 76 deaths per million doses versus 50/1.5M*1M = 33 deaths per million doses for EJ1685.

Let’s also assume, for the second now, that one of the VAX LOTs has a lower number of doses - the one that is associated with fewer deaths (EJ1685), say. We do the same calculation. Thus, the toxicity for EN6201 = 114/1.5M*1M = 76 deaths per million doses versus 50/100,000*1M = 500 deaths per million doses for EJ1685. A visual result of both calculations can be seen in the following figure with the former calculation results on the left and the latter calculation results on the right. The toxicity for EN6201 does not change - just the scale does.

Figure 1: Comparison of variations in VAX LOT sizes as per differences in associated death counts (lot sizes equal/death counts not) (left) and neither the lot sizes nor the associated death counts are equal (right).

The take home message here is that even though VAX LOT EJ1685 is associated with fewer deaths in the VAERS dataset, there are more incidences reported when the size of the lot is lower than the Pfizer estimate.  ie: The toxicity is unknown and dependent on the number of actual doses administered.

Let’s go one more step.

Let’s assume that EVERY VAX LOT has 1.5M doses. With the brilliant help of Simit Patel and his R saaviness, I now have a version of the merged VAERS data with clean VAX LOTs. Hurrah! Thank you my new friend. So the new and ‘real’ total number of Pfizer VAX LOTs reported in VAERS is 3,315. What this means is that if each VAX LOT actually has 1.5M doses per lot, then there have been 4,972,500,000 total doses administered/delivered by Pfizer. Say what? This is not the case at all. If we seek the total number of doses either administered/delivered from Pfizer in the United States to date, we find that 310,332,802 doses have been administered and 384,326,705 have been delivered.

If we back-calculate, this gives us, as an average, 93,614 doses administered and 115,936 delivered, per VAX LOT, in reality. Now, it is far more likely that some VAX LOTs have 1.5M doses and others have far fewer. But, just for argument’s sake, the number of doses per VAX LOT cannot be 1.5M per.

Let’s find out what the dose number would have to be in order to equalize the toxicities between the two example VAX LOTs. In other words, what would the number of doses administered have to be in order for VAX LOT EJ1685 to have the same toxicity as VAX LOT EN6201? We assume here that EN6201 had 1.545M doses administered. The answer is 667,631. Sorry, I ruined the punchline. What this means is that the toxicities are the same even though the number of deaths are different and this is based on differences in dose numbers.

Figure 2: Comparison of variations in VAX LOT sizes as per differences in associated death counts where the lot sizes are equal and the death counts are not (left) and neither the lot sizes nor the associated death counts are equal (centre) and where the dose size for VAX LOT EJ1685 is adjusted to make it ‘as toxic’ as VAX LOT EN6201.

The three above scenarios demonstrate three ways that toxicities (the rate of death per N doses) can vary in the context of different absolute death counts for any 2 VAX LOTs.

There are many ways that the toxicity of the various VAX LOTs can vary and it appears to heavily depend on the number of doses per VAX LOT. Variability in toxicity is a function of the number of doses administered, and of course, the number of death reports being filed to VAERS. Even though EJ1685 has fewer associated death reports in VAERS, it could very well be far more toxic if the number of doses is less than the reported 1.5M doses per VAX LOT. And based on our calculation of the number of doses that would have had to have been administered to date if all VAX LOTs were equal at 1.5M each, we know that they are not equal at this number. It is highly likely, in my opinion, that some of the VAX LOTs have 1.5M doses, and some have fewer. It is impossible to know at this point for sure, but safe to say, this demonstration illustrates the possibility that the VAX LOTs with fewer associated deaths could indeed, be more toxic and therefore, it is wrong to assume that more N deaths = more toxicity. Again, toxicity is a function of death count and dose number.

Just as an aside, if VAX LOTs have variable mRNA content we can imagine a scenario where different VAX LOTs indeed have different associations with adverse events, both in number and type, due to the effects of protein production of complete mRNAs and incomplete ones. As is stated below in the Pfizer checkpoint document, the effects of this, was unknown and remains so. It has actually been reported that complete mRNAs leading to production of modified spike protein impairs DNA double-stranded repair enzyme function. These enzymes include BRCA1 and 53BP1 and are the key checkpoint proteins for non-homologous end joining (NHEJ) and homologous recombination (HR) repair, and both of these checkpoints were inhibited by spike.¹

One of the outcomes could be a higher yield of Severe Adverse Events. We don’t know yet.

Also, just to give a bit more merit to the idea that VAX LOTs are different content-wise, remember the VAX LOT 041L20A that was pulled due to high association with anaphylactic reactions? By the way, it’s so laughable (well, to me because I am a data nerd) how even in this online report of the incident, they mis-print the VAX LOT number (they write 41L20A - what? no 0?). This is one of the biggest problems with assessing anything VAX LOT-related from VAERS, by the way. The point I was making is that there is potential variability between lots and potentially even vials.

Bottom line: toxicity is the measure that should be used to assess potential differences in VAX LOTs and this is a function of both adverse event count and dose number per VAX LOT. We have no idea yet whether some VAX LOTs are definitively causing more (severe) adverse events - I suspect that some are - but time will tell. Actually, if the bloody manufacturers would share the complete list of VAX LOTs with true dose number as per each, we wouldn’t need our buddy time any more.

I hope this Substack article helps in spreading understanding. We all need to work together in this to reveal the true rankness of it all. Smells like something other than Teen Spirit.

1. Jiang H, Mei Y-F. SARS–CoV–2 Spike Impairs DNA Damage Repair and Inhibits V(D)J Recombination In Vitro. Viruses. 2021; 13(10):2056. https://doi.org/10.3390/v13102056.