It’s not quite right to say PCRs are more sensitive. I’ve pasted a post fro @gruntbaby from another thread which was really informative. I think the conclusion is that a positive LF means you almost certainly have COVID, even if PCR negative. Particularly if you have symptoms:
I know, this is a nightmare to get your head around. I'll see if I can explain...
Yes, you're right - normally a PCR will pick up cases earlier and later than the LFT, which tends to catch cases only at their peak. PCR basically analyses RNA (genetic material), amplifying tiny amounts over numerous cycles of the machine until there's enough to measure. Whereas LFTs pick up viral proteins, rather than the genetic material.
It's exceedingly unlikely that a PCR will get a false positive. However PCRs can get false negatives, often due to poor sampling technique (you need to swab high enough and get viral material on swab).
LFTs vary by brand as you've seen above, but false positives are still very rare. False negatives are a lot more common on LFT than on PCR, (so it's not safe to get a negative LFT and assume you're clear to go and hug grandma), but if you have a positive LFT you are exceedingly likely to be infected, AND highly infectious at that moment. IIRC you are then advised to get PCR confirmation in part because a certain % of PCR results are analysed more fully to check for new and dangerous variants. The UK system is pretty good at picking these new variants up.
So normally you would get a positive LFT and then it would be confirmed by a positive PCR. What is happening now is that it is highly unusual (and suspect) for there to be so many positive LFTs followed by negative PCRs as we are seeing at present. It indicates a problem with the LFTs or a problem with the PCRs. The contextual and anecdotal evidence I've seen seems to indicate that it is less likely to be an issue with the LFTs (many of these strange cases eventually get picked up by PCR, or by PCR in a different lab, or there's a direct transmission chain to a PCR positive case).
So... it seems at the moment more likely to be a PCR issue. The most likely suspects would be:
- A problem at a PCR testing lab, e.g. a machine not properly calibrated, or set up wrong for example. Investigators should be checking all this.
- A new variant the current PCR tests aren't picking up.
It's years since I've had anything to do with PCRs, so this is conjecture but if a new variant is sufficiently different in its RNA, then the PCR machines just looking for yes/no might be missing some new variants. Another possibility is that, given lots of different PCR tests and machines are used by different labs, some labs' machines could be able to pick up the new variant, and other labs can't. A new variant being missed is still unlikely (because the PCR machines are looking for at the bit of the virus code that tends not to evolve much, rather than testing the bit that keeps changing). However there IS precedent for a new variant evading PCR and this happened with the Brittany variant. You'd only be able to pick up cases by realising there was an issue and then looking into the genetic code in more detail and discovering a new variant.
Even if it's a variant, the virus could still be producing similar enough proteins to be picked up by the LFTs - and most LFT types look for the most stable, unchanging protein bits.
So PCR and LFT are 2 different tests measuring two different things. If Covid changes enough in one way but not as much in another, that could explain why LFT is still picking it up, but PCR is not.
The situation is new and rapidly evolving, so we're in the period where the cause of this discrepancy is being investigated and guidance could change as a result (and already is changing in some places as a stop gap). If it's a lab issue they can fix that. If it's a new variant that is evading the current PCR tests then they would be likely to send out guidance to the labs and the labs will adjust their testing protocols and we can go back to relying on PCR confirmations. But all that will take time to figure out. Hence the current confusion.