Demystifying Cycle Thresholds for PCR testing.


image.png

We're starting to see a lot of really bad takes on "CT scores" or "Cycle Counts" spreading, especially since Rhode Island released all their data for March-June. I'm going to explain what these cycle counts do (and don't) mean, and show how it's being spun by the usual cast of deniers/casedemic people.

tl;dr - PCR tests have insanely low false-positive rates, on the order of 0.01%, or one result in 10,000 tests. The CT score indicates how many times the test needed to double the RNA concentration before the fluorescent markers lit up enough to be detected. At 39, 40 and up there's some chance of a false positive, 38 and under, false positives are essentially zero. CT scores in the mid 30s indicate little or no live virus, possible just leftover RNA fragments or the very early stages of infection. These are being spun as "false positives" because the patients were not infectious at the time of the test. And there are accusations that officials are "turning up the cycle count" to falsely generate positive results to support their narrative. That's utter BS. Watch out for charts of the Rhode Island data like my first image, with arrows through them showing an alleged trend of fakery, it's a wild misinterpretation.

WHAT ARE CTs or CYCLE COUNTS?

https://www.publichealthontario.ca/-/media/documents/ncov/main/2020/09/cycle-threshold-values-sars-cov2-pcr.pdf

This link is an excellent introduction. PCR works repeatedly copying a specific RNA or DNA fragment (if it's present) until it is amplified enough that fluorescent chemical tags light up enough to give a positive result. The number of cycles needed to reach that point is the CT, or Cycle Threshold.

The systems count how many cycles were needed, with each cycle boosting the weak signal to twice the level of the cycle before. If your test is positive at 20 cycles, that means you have so much SARS-CoV-2 RNA in you that amplifying just a million times made the fluorescent glow visible.

If it took 30 cycles, that means it had to be amplified a billion times. Still a lot of virus, but nothing like a score of 20.

At 35 cycles, you're talking about a 30-billion amplification, and your viral level is pretty low, or you may just have left-over RNA after the viruses are killed.

By 40 cycles, you're at a trillion-to-one amplification. A positive probably means that a trace of viral RNA was found, but at this level, even a trace contamination in the lab could yield a positive, to false positive could be an occasional issue. In any case, there's not enough there to be worried about.

The take-home message is that below 39, there is almost zero chance of a false positive, but a true positive doesn't mean you're contagious that day. A positive with a cycle count of 34, for example, might mean different things. If you're testing due to a contact trace, it likely means you're in the very, very early stages of being infected, and should definitely isolate and watch for symptoms, with follow-up testing later. If you've already been sick, this probably means you're on the mend but there's a chance you're still contagious...but most likely you aren't. Or it could conceivably mean that you're on a totally asymptomatic infection and your viral levels are staying very low, either due to an excellent immune response or a very small infectious dose.

A cycle count in the 20s says you have a lot of virus in your system, and you should definitely isolate.

And that's basically it. All of these are true positives, but the cycle count can give you more information about where you are in the infection and recovery process.

SO WHAT'S ALL THE FUSS?

COVID deniers and "casedemic" people (those who claim our COVID case counts are nothing to worry about, just an artifact of "turning up the volume" on tests) are trying to paint CTs as a trick used by the establishment to generate positive results so that it looks like an epidemic even though it really isn't.

As usual, it starts with a grain of truth. CTs above 39 or 40 are more prone to false positives. Most testing outfits will choose some level for cutoffs, the details vary with the test and process.

Test results in the 33-38 range, though, are very much true positives - SARS-CoV-2 RNA was found in your sample. But at these higher CTs (lower viral loads), it means that the amount of live virus in your system is very low, or even zero. (If you've just gotten over COVID-19, some viral fragments may remain for a while. ) So a positive PCR test doesn't mean you're necessarily contagious, just that the RNA was found so you are somewhere in the infection->peak->recovery process.

The casedemic crowd is prone to labelling these as "false positives" for rhetorical purposes. They're not false positives. In fact, as we get outbreaks under control and make testing more widely available, we hope to catch more asymptomatic people (with low viral loads) and more pre-symptomatic people where the test finds a positive days before the symptoms show up, often due to formal of informal contact tracing.

The first image above shows Rhode Island's results by day, with every one of the 5000 positive tests by CT value. This is a good thing, and frankly, I think every PCR test result should post a CT value, and states should share what they have. But it's important to keep in mind what was happening in the field with testing as these months went by.

Rhode Island was overwhelmed at the same time NYC was having its crisis. There were not enough tests to go around, and a large number of positives were from people with severe symptoms, sometimes after hospital admission. As you would expect, during the crisis phase, a large portion of the positive tests were from people with severely high viral loads - ones who had been out in the community spreading the disease before they were identified as COVID cases. That's not where you want to be if you're on top of the epidemic.

As time went on, the case load fell, the availability of tests improved, and testing was available to more people with mild symptoms. This leads to fewer extreme "red" cases with low CTs and extremely high viral loads, and a larger share of people being caught in the early stages. At the same time, people who were post-symptomatic from infections in the peak of the curve were testing to see if they were safe to return to work, etc., and those people typically test negative, or positive with a high CT / low viral load.

This is absolutely normal, it's what you expect as the first wave wanes, and it's exactly what you want to see as your testing starts catching more and more people in the early symptomatic or pre-symptomatic stages. (In the logical extreme, if we tested everybody every other day, we'd almost never see a red case.)

BUT WAIT, THERE'S MORE

Unfortunately, the data from Rhode Island is sparking bad interpretations and outright conspiracy theories. I'm linking to a very mild instance, but there are plenty out there claiming that this is proof that authorities are "turning up the volume" on tests to generate positives.

The second image shows a plot of the average CT value for Rhode Island's positive tests, by day. It is increasing, which is NORMAL and WHAT YOU WANT TO SEE if your testing program is expanding and getting more proactive. Unfortunately, he then plots it with daily death counts and concludes "As the average Ct rises past 30, deaths almost disappear". Which is technically true, but he's getting cause and effect mixed up. In fact, there's some causality in the reverse direction - high CTs can mean you're catching more people early and getting treatment or reducing spread more effectively.

Lower community spread and higher testing rates will lead to more cases being caught very early, which will raise your average CT.

Lower community spread will also reduce your death rate. That's why they move together.

Others are far worse. They're claiming that the rise in average CT is proof that the states are cooking the books by "cranking up" CTs to generate false positives. That's just nonsense.

You can't take a negative sample and "turn it up" to get a positive at 35. It just doesn't work. It just shows that more cases are being identified in pre-symptomatic or early symptom phases, rather than in the hospital. That's good news all around.

This is what you want to see; it means your wave is receding and more and more testing is being done on asymptomatic people. The more positives that are identified while not infectious, the better!

There can and should be discussions about what CT level should be used as the cutoff point. 38, 39 and 40 are all reasonable values, that keep false positives low while minimizing false negatives, too.

We also need to remember that PCR (unlike rapid antigen tests) are extremely sensitive and will pick up RNA traces when your infection is very low, and sometimes even after you've cleared the live virus.

Finally, I don't understand why CT numbers aren't routinely reported on every test result. It's a good bit of data that can be useful, it's not some sort of secret.


Comments 0