Thread by @CillianDeGascun

A short thread on PCR #SARSCoV2 There’s been a lot of discussion recently about the use of PCR (Polymerase Chain Reaction) for the diagnosis of #SARSCoV2 infection, and the interpretation of its results. 1/n

Hopefully, this thread can provide some background and context for those who are not familiar with the technique. PCR has become the cornerstone of molecular diagnostics in virology, and is, simply put, a technique for amplifying DNA (and indirectly RNA). 2/n

It can be used for straightforward detection purposes i.e. to provide a qualitative yes/no answer. Alternatively, it can be used to provide a quantitative result, i.e. to tell us how much of something is present. 3/n

In this latter quantitative format (qPCR), it is regularly used to measure and monitor viral loads (i.e. the number of virus copies, typically per millilitre (ml)) in individuals infected with chronic viral illnesses, such as HIV or HCV. 4/n

When PCR is used for the diagnosis of respiratory virus infections (including #SARSCoV2 & #influenza), we do not typically quantify the amount of virus present, as the specimen type – a swab from the nose, throat, or nasopharynx – is not standardised or consistent. 5/n

By this, I mean that the quality of every swab sample will be slightly different and will contain a different amount of cellular material from the patient, and therefore a different amount of virus (if the virus is present). 6/n

In contrast, one ml of blood or urine for example from individual A has the same known volume and quantity of material (in general terms) as one ml of blood/urine from individuals B, C, and D. 7/n

A thermostable (not damaged by heat) polymerase enzyme drives a PCR. This polymerase enzyme will synthesise/make a complementary sequence to any single strand of DNA if we provide the raw materials and the correct temperature for it to do so. 8/n

There are three steps to PCR: denaturation; annealing; and extension. Denaturation refers to double-stranded DNA breaking up into single strands. Annealing refers to the binding of pathogen-specific primers and a fluorescent probe to that single DNA strand. 9/n

(Primers are small pieces of complementary DNA to the sequence we’re trying to amplify: the primers & probe are specific for the virus or other target DNA that we’re trying to detect. This is why PCR itself is so specific, and one reason false positive results are uncommon.) 10/n

Extension refers to the polymerase enzyme making the complementary strand to that single strand. These three steps constitute a single cycle of the PCR, at the end of which we have twice as much DNA as we started with. 11/n

However, in order to amplify the DNA to a level we can detect, we usually run the PCR for 35-45 cycles. This allows us to detect very small amounts of DNA in the original patient specimen. 12/n

If we develop our own (in-house or laboratory-developed) PCR test, then we can decide how many cycles we want to use. However, if we are using commercially-manufactured CE-marked kits, then the manufacturer decides how many cycles we must use. 13/n

As the amount of DNA produced increases, this is (usually) linked to the generation of fluorescence, which we can measure. The more fluorescence we see, the more DNA that was present in the original specimen. 14n

Of course, if there is no virus in the original sample, then nothing happens in the PCR: there is no amplification and no fluorescence. However, we also add an internal control to the sample so we can be confident that the PCR has worked properly. 15/n

If there is virus present, we can identify positive samples based on the amount of fluorescence produced & at which Cycle (from 1 to 45) the fluorescence becomes detectable, i.e. when the amount of fluorescence crosses a pre-determined baseline threshold for that PCR test. 16/n

This is what the Ct (cycle threshold) value means. This is the cycle at which the fluorescence crosses the baseline threshold. The lower the Ct value, the earlier the fluorescence crossed the threshold, & the greater amount of virus that was present in the patient specimen. 17/n

As a general rule (and I stress general) a Ct value of 30 equates with a virus load of around 1000 (3 logs) copies. Given the amount of DNA doubles with each PCR cycle, this means that every 3.3 cycles, the amount of DNA present changes by a factor of 10 (roughly). 18/n

This means that in a perfectly efficient PCR (which probably does not happen that often with human specimens), Ct values in the high 30s (Ct of 35 and above) suggest that there were very few copies of virus in the specimen processed. 19/n

These very high Ct values (i.e. low viral loads) have raised questions about the infectivity of these individuals. However, a single high Ct value in isolation should not be considered definitive proof of anything for the following reason. 20/n

A high Ct value (indicating a low virus burden) does not and cannot distinguish between a virus load on the way up (e.g. in a pre-symptomatic individual), a virus load on the way down (e.g. in a recovering individual), a poorly taken specimen, and a false positive result. 21/n

As such, any suggestion that we should diagnose & contact trace only those with Ct values of 30 & below constitutes too great a risk at this point in time. Yes, those with higher viral loads are likely to be more infectious, but respiratory specimens are not standardised. 22/n

In addition, testing the same sample on different assays/platforms can give different Ct values reflecting differences in the targets detected & the chemistry of the test used. As such, repeat samples from an infected person should be tested in the same lab each time. 23/n

Whilst PCR is exquisitely sensitive and remains the current gold-standard for diagnosing #SARSCoV 2 infection, it has some limitations. For example, it does not distinguish between viable virus and non-infectious RNA. 24/n

This can be challenging for the management of some individuals infected with #SARSCoV2, as RNA can remain detectable – albeit at low levels – for a number of weeks, and sometimes months, following the original infection. 25/n

In addition, PCR assays may generate ‘non-specific’ or ‘false positive’ results. These are uncommon, occurring at a rate of somewhere in the region of 1-3%. However, the larger the number of tests that is performed, the larger the number of false positive results. 26/n

These ‘false positive’ results do not mean that the test is bad: they are a feature of essentially all tests that we use in medicine on a daily basis. However, it does mean that we need to consider every test result in the context of the individual who has been tested. 27/n

In addition, as the prevalence of infection in the community declines, the likelihood of any individual result being falsely positive increases. This positive predictive value is the probability that individuals with a positive test truly are infected with #SARSCoV2. 28/28