Harry writes:

Hi team, 

Big fan of the show.  Great to hear such enthusiasm from such experts.

Here in Melbourne (Victoria, Australia; rather than Arkansas, California, Florida, Iowa or Kentucky, where there are also Melbournes), it is 13degC and drizzly. 

I’m currently working in clinical microbiology, so wanted to write in about PCR and serology testing for COVID-19.

Firstly, about PCR: 

This undoubtedly the most valuable test modality.  Whilst a RT-PCR is a remarkable technique, capable of detecting tiny amounts of RNA, this does not mean that the commonly-used test for COVID-19 is highly sensitive.  It seems that the sensitivity of a SARS-CoV-2 Real Time PCR on a single nasopharyngeal swab is about 70%.  I.e. a false-negative rate of 30%.  

One explanation seems to be that COVID-19 patients only intermittently shed in their nasopharynx.  PCR performed on sputum (when patients are producing sputum) seems more sensitive. 

An overriding difficulty with examining these new tests is the difficulty of defining a diagnostic gold standard, given this is an entirely novel disease.  I’ve heard from Chinese Infectious Diseases doctors that initially, in Wuhan and Shanghai, the gold standard rule-out of COVID-19 was 3 negative PCRs over 3 days, with one performed at a reference lab, then no SARS-CoV-2 genetic material detected by metagenomic sequencing on nasopharyngeal swabs and/or pulmonary aspirates.  

After all, the only initial diagnostic tool was metagenomic sequencing, because no PCR had been developed. 

I’ve heard that many patients were found who had a typical illness pattern but repeatedly negative PCRs, before SARS-CoV-2 nucleic acid was discovered through metagenomics.  Clearly not affordable in most health systems. 

The language used in the media is confusing regarding testing, and seems to assume that the PCR is perfect, no matter what circumstance it’s being used in.  When multiple of Donald Trump’s contacts developed COVID-19, he released a statement saying he tested negative, and had no symptoms.  This shows a fundamental misunderstanding. A negative nasopharyngeal swab PCR means precisely nothing in an asymptomatic person.  It neither rules in or out current infection, and it certainly gives no predictive value about the probability of developing illness. If Donald Trump developed symptoms the day after being tested negative, he should have been isolated and undergone another test, because he’d presumably have developed COVID-19.  Good luck to anyone who had to explain that to him. 

If a patient in New York or New Orleans now has fever, cough, loss of smell and taste with a bilateral pneumonia, then they should absolutely be diagnosed with COVID-19 regardless of any test.  The pre-test probability would be so high that, even with a negative test result, COVID-19 is the most likely diagnosis.

If testing capacity is limited, It’s therefore entirely reasonable to not test people with a typical cluster of symptoms and signs.  They should just be diagnosed and managed as COVID-19 patients, and testing resources used to discriminate other cases where the diagnosis is less clear.

On serology:  It has no role in testing for COVID-19 currently. 

It’s tempting to consider the presence of SARS-CoV-2 antibodies to be definitive evidence of past or present SARS-CoV-2 infection.  But this assumption has many holes. 

Firstly, IgM assays are notoriously non-specific.  Mostly for this reason, there’s almost no disease that can be definitively diagnosed by the presence of IgM alone.

IgG assays are typically more specific for the antigen being detected, however they can take weeks to develop.  In any one assay, there will always be a false-negative rate on an IgG test.  For whatever reason, a previously-infected person may not have antibodies against the particular antigen being measured by the assay. 

Unfortunately, currently the presence of an IgG against SARS-CoV-2 is not evidence of immunity from re-infection.  It’s not even evidence that viral shedding has resolved.  We just don’t know either of those things with enough confidence for COVID-19.

Whilst I think it’s a solid guess that past infection will give some degree of protection from future disease, we certainly cannot assume this will be complete immunity.  We therefore can’t offer any different guidance regarding PPE, removal from isolation, or return to work for anyone on the basis of a serology result.

The public health laboratory network of Australia has released a statement on point-of-care serology tests, which is freely available.  It’s pretty unambiguous.   


My question relates to antiviral therapies:

Which mechanism in the SARS-CoV-2 lifecycle do you think is most likely to yield a useful antiviral drug? 

All the best in hand washing and distancing,


Dr Harry Walker BSc(Psych) MIPH MBBS

Microbiology and Infectious Diseases Advanced Trainee

Microbiology Registrar