COVID, PCR and The Wrongness of Alex Berenson

PCR Lab. Photo by Gokulsanjeevan, CC BY-SA 4.0 <https://creativecommons.org/licenses/by-sa/4.0>, via Wikimedia Commons

So, the Prime Minister of COVID Wrongness Alex Berenson tweeted this the other day:

Wow. WOW. @cdcgov only wants to examine post-vaccine infections with a PCR threshold of 28 or under.

That standard ignores 90+% of #Covid infections. The entire epidemic would have looked very different if it had been used.

“An RT-PCR Ct value ≤28”https://t.co/4G18gK8HGg

— Alex Berenson (@AlexBerenson) April 29, 2021

Berenson has, for the last few months, been feeding a lot of anti-vaccine conspiracy theories by implying that the vaccines infect people with COVID-19 (which is physiologically impossible) or that the vaccines are killing people by the busload (which is not shown in any data whatsoever). In this case, the implication is that the CDC is ignoring infections in people who have already been vaccinated by using a stricter standard for examining post-vaccination COVID-19 infections than has been used pre-vaccination. As you can see by the numbers, a lot of people are buying into this. And this claim has, for reasons I will get into below, set the crazier conspiracy theorists off on yet another round of “COVID-19 doesn’t exist; the tests are false positives” (which Berenson does not claim, as far as I know).

The implication that this document proves that the CDC is covering up post-vaccination infections is wrong. It’s as wrong as everything the anti-vaccine brigade is claiming these days. The people making this claim either do not understand what they are talking about or are deliberately misrepresenting it. I’ll leave it to you to decide whether a Yale graduate and former writer for the NYT is this confused by the simple declarative English sentences of the very document he cites.

To explain why this tweet is so misleading and the reaction so wrong requires a bit of biochemical detail. Let’s do this thing.

The document cited in the tweet is not about diagnosing COVID-19 post-vaccination. It’s about studying post-vaccine infections for purposes of genomics. As we have discussed many times, the vaccines are not 100% preventative; no vaccine is. And there are concerns that variants and mutations may evade the vaccines. So, it is critical that, if someone develops an infection after being vaccinated, we know what strain they were infected with. What this document does is lay out the protocol for studying post-vaccination infections to get the finely detailed genetic information that will allow us to identify variants and trace infection pathways.

It says it right there on the first page:

Objective: Investigate SARS-CoV-2 infections among people who received COVID-19 vaccine to identify trends or clustering in demographic (sic), the administered vaccine, or the infecting virus.

Not “identify” SARS-CoV-2 infections; “investigate” ones that have already been identified. Where the RT-PCR value comes in is with specimen selection.

Specimen selection

o Clinical specimens for sequencing should have an RT-PCR Ct value ≤28.
o If a Ct value is not available, specimens that are positive for SARS-CoV-2 RNA or antigen by another testing modality may be sent.

The invaluable Bad COVID-19 account takes put it together in graphic form:

pic.twitter.com/xknD8GzURu

— Bad COVID-19 Takes (@BadCOVID19Takes) May 1, 2021

Polymerase chain reaction, or PCR, is one of the important scientific breakthroughs of the last 40 years. It’s a technique that allows scientists to take small amounts of DNA and amplify them up to massive amounts of DNA for analysis. Those of you who are into true crime may find this sounding familiar. One of the many uses of this technique is to take, say, small drops of blood, amplify their genetic signature up and use it to identify victims or perpetrators in crimes. It also allows genetic researchers to amplify up small bits of saliva or tissue samples to look for possible markers of disease. In this case, PCR is being used to identify the genetic sequence of a different type of murderer: the SARS-CoV-2 virus.

When you amplify up a sample of DNA, however, there is tradeoff. Each cycle of the PCR process amplifies the sample up more. But each time you amplify it, you blur the signal a bit and some amount of detail may be lost.

Imagine that you are driving down the road in a rural area and scanning for a radio station. You pick up a faint signal that sounds like AC/DC. Turning up the volume will make the music louder. But it also will make the static louder. Turn it up enough and you can rock out to “Highway to Hell” on your drive. Turn it up too much and it’s just noise.

Most researchers cut off their amplification at about 30 cycles because beyond that, you’re mostly amplifying noise and blurring the signal too much to be useful. From the very beginning of this pandemic, conspiracy nuts have been pointed at this to claim that the COVID-19 pandemic is a hoax. The COVID PCR test amplifies the sample up to 40 times¹. That, they claim, induces false positives, producing a signal that’s not there.

But this is nonsense. Amplifying a PCR signal cannot produce a false signal for SARS-CoV-2 any more than turning up your radio on static will make it play AC/DC. Amplifying may lose information. But it will not create false information.²

The reason such a high PCR count is used in diagnosis of COVID-19 is because we know what the virus looks like. We don’t need incredibly detailed information on the virus. We don’t want to know if the virus has the gene for breast cancer. We don’t want to know if it committed a crime. We just want to know if it’s there. Thankfully, the virus has broad genetic markers that are unique to the virus and do not occur in human beings. So, we can amplify it up quite a lot and still see it. The more we amplify, the lower the infection level we can detect it at and the earlier in the infection we can tell someone they have it. It’s as if you knew “Highway to Hell” so well that even hearing a single drumbeat would let you know you’d found your jam.³

However, sometimes we don’t just want to detect COVID-19. We want to trace it. We want to get richly detailed information about the specific virus in someone’s body. We actually do want to know if this is the specific virus that knocked over a liquor store. Such detailed information is how we track variants. Such detailed information allows us to track infections at the micro level, even from one person to another. And that kind of information is critical in studying post-vaccine infections. We want to know: if someone gets infected, was this the classic virus or a variant? How did they get infected? Who else might be at risk?

For that kind of research, simply detecting COVID-19 is no longer enough; we want the kind of detailed information that large amplification erases. Gene sequencing, as opposed to gene detection is far more specific and requires a far more fine-grained detail because you are fluorescing each individual base-pair. Viruses only have about 30,000 base pairs. Humans have billions. So, you can set your threshold as low as 28 because you don’t need to amplify viral DNA that much to get what you want.

Going back to the music analogy: we don’t just want to know whether our station is playing AC/DC. We want to know what specific recording of “Highway to Hell” we are listening to. So, we’re going to keep turning the dial until we get a radio station with a really good signal. We want a good signal, not a survey of exactly how many stations are playing AC/DC right now.

All this document says is that to trace the lineage of post-vaccine COVID-19 infections, they want a sample strong enough that it doesn’t have to be amplified up very much to tell that it’s COVID. That lower amplification will leave the detailed information intact that allows us to trace the virus. That’s it. The actual diagnosis of an infection, as the document specifies in its second sentence, will use the same techniques we use for everyone else.

Berenson has, since the beginning of the pandemic, been a staple of the Bad COVID-19 takes Twitter account and other debunkers. But since vaccines went global, he has been especially problematic. For example, he has been tweeting out VAERS reports as though they were gospel⁴ and speculating some deaths might be related to the vaccine. This is a perfect illustration of what he (and others) have been doing: making dubious claims or ominous implications that are contradicted by the very documents they cite.

I have said for a while now that we need to be vaccinating as fast as possible. In recent weeks, vaccination rates have been tailing off, particularly since the FDA’s ill-advised pause on the Johnson and Johnson vaccine. We do not need people spreading this kind of nonsense. All the information we have, from multiple countries, indicates that the vaccines work and are safe. Countries with the highest vaccination rates are seeing caseloads and deaths plunge. Israel, which is basically at herd immunity, is completely open. The one thing that can end the COVID restrictions that so many like Berenson have raged against is Getting. People. Vaccinated. And the thing that’s mostly standing in the way of that right now…are people like Berenson pawing through CDC document for something, anything, they can tweet to claim that the vaccines aren’t safe.

My thanks to my lovely wife, a genetic researcher and biochemist, who fact-checked this piece during breaks in literally running PCR tests.

1. Although most are detected after 30-35 cycles or less.
2. It is possible that massive amplification could erroneously bring out the genetic signature of trace contamination. For example, it could amplify up the genetic signature of skin cells from a lab technician that fell into the sample. But we don’t worry about that with COVID-19 because human DNA is very different from viral RNA and would not be mistaken for it.
3. In fact one of the early problems with PCR testing for COVID-19 was that if the test was given too early, it would give a false negative because there wasn’t enough of the viral RNA for the test to see no matter how much it was amplified.
4. As someone who has filed a VAERS report, I can tell you that these are not verified. The purpose of VAERS is to cast as broad a net as possible, not to engage in systematic research. If the same things keep popping up, that tells scientists what to look for. But you wouldn’t use it for actual research because literally anyone can put in a VAERS report.