The Hokey Pokey

Let us suppose it takes 1 billion UV-tagged DNA chains to trigger a positive RT-PCR test result and the polymerase chain reaction process follows a perfect doubling. If we start with a swab containing 1 viral RNA molecule we'll hit that detection threshold in 30 test cycles. If we increase the initial sample to 10 viral RNA molecules we'll hit detection at 27 test cycles. With 100 viral RNA molecules on our swab the Ct drops to 24 cycles. With 1,000 viral RNA molecules on our swab the Ct drops to 20 cycles, and with 10,000 viral RNA molecules the Ct drops to 17 cycles (slide #1).

Somebody happy to give their nose a jolly good poke and fetch out 10,000 viral fragments will yield a positive result at a Ct of 17, whereas somebody who gingerly dabbed about retrieving just 1 viral fragment will generate a positive result at a Ct of 30. These differences may arise because of viral load and they may arise simply through the random variable that is the nose poke. Because we cannot count the RNA fragments on each swab at the outset before we run the test we cannot ever be sure whether variations in cycle thresholds are down to viral load, poking technique or a mixture of both.

The chances of starting out with exactly the same number of RNA molecules on every swab taken are astronomically small as to be a decent definition of zero. At high levels of amplification we can see that small differences in the sample are greatly exaggerated, thus the same person could take a dozen swabs in quick succession and theoretically obtain a dozen different threshold values.

Quite how this pans out in practice I am not certain for I have not had the pleasure of reading a paper where this issue is explored and reported upon. If the variation within consecutive swabs within the same person is as much as ±10 cycles then what are we doing trying to compare swab results within subpopulations down to a level of ±5 cycles? Perhaps somebody could enlighten me!