The article mentions RNA implying that this is the PCR test not an Antibody test. False positives exist with the PCR test. Information I could find suggests the primary cause is sample contamination. So running the test multiple times through different labs and clinicians is an obvious workaround.

Surely this is not a false-positive, but rather a false-negative?

That is, the patients were observed clinically to have caught and recovered from the virus, yet the test shows no trace of the viral infection.

> Surely this is not a false-positive, but rather a false-negative?

Both are possible and either mess it all up.

This is the correct answer

More likely sample degradation as well. Rna is notorious for being extremely fragile exactly when you need it not to be and extremely robust when you don't. Academic rna labs have a reputation for being clean freaks, and there was a spate of spurious observations when rt(rt)pcr became a thing ten or so years ago... I'm still not convinced of rt/rt results performed by run of the mill labs, generally speaking.

Even a common well established test like a quick strep test has a significant (10% iirc) false negative rate. I'm sure that a rapidly developed test to an emerging target is going to have a significant error rate.

In general PCR tests are highly sensitive. Even smallest viral loads can be detected

(Edit) added in general above due to valid points in the comments below

The PCR can detect if there is virus in the swab. But the swab may not have scooped up any virus even though the person had some.

my longstanding argument against theranos, by the way (like, since 2005), was "look. if i go down to the beach with a bucket and take a bucket of water and put it in my super accurate shark detector van, it doesn't mean the beach is safe to swim at!"

To be fair at some level that's an argument against blood testing in general.

not really. it's just an admission that sampling frequency matters.

Unless you dilute the water until it becomes homeopathic anti-shark water.

That’s right!

They're highly sensitive to the presence of the target DNA or RNA sequence in the sample well. But they can't easily tell if the target sequence they amplified came from the patient's sample or from other potential sources of contamination in the lab (the test, after all, makes billions of copies of the target DNA sequence), whether from other positive tests or from positive controls.

But that doesn't explain false negatives, does it?

They typically do one swab of the nasopharynx (upper back of the throat, up above the soft palate). You might have the virus elsewhere but not there.

Saw reports that bronchoalveolar lavage (introducing and then suctioning out fluid from a branch of the lung) is better at picking up virus if it's present -- but it's much more time consuming, more invasive, and much less pleasant for the patient than sticking a swab up your nose. Not something you're going to do in a parking lot in a drive through testing center for someone only showing mild symptoms.

There are always compromises.

It depends on the time of the illness. Roughly my understanding is: 1st week - high virus load in the nasopharyngeal area. 2nd - almost no virus load in the nasopharyngeal area, high virus load in the lungs

Can't PCR tests detect damaged RNA and report that as a positive?

https://www.aim-med.org/hiv-rna-test/

if:

- you do them right

- your reagents are good

- the kit was assembled properly

- there's no contamination in the machines

- you got a good sample

- ...

The capability for a highly sensitive and accurate test in no way guarantees that a randomly sampled administration of that test is going to meet the high standard that the test may be capable of.

And if you swabbed an infected part of the body in the first place!

In 205 patients with proven COVID-19 [...] bronchoalveolar lavage fluid was positive in 14/15 (93%), sputum 75/104 (72%), nasal swabs 5/8 (63%), brush biopsy 6/13 (46%), pharyngeal swabs 126/398 (32%), feces 44/153 (29%), blood 3/307 (1%), and urine 0/72 (0%).

https://www.jwatch.org/na51116/2020/03/17/pharyngeal-and-nas...

Hol' up.

> blood 3/307 (1%)

Why isn't their blood rotten with virus?

I suspect it's because this virus is optimized to enter cells that line the lungs/respiratory tract, then also sheds back that same way. Meanwhile, cells in, and along, the bloodstream aren't infected – and are most-trafficked by the body's immune response, including virus-destroying antibodies.

(Why are rodents more prevalent in the walls/crawlspaces of a dwelling, rather than the hallways?)

But the lungs are full of blood, that's how we breath.

So it's spreading mucosally or through connective tissue?

Sure, but if an initially-infected respiratory cell releases its "baby coronaviruses" out to the respiratory linings, away from the blood, they'll find lots more of the respiratory cell surfaces they're optimized to infect, and maybe get a ride out to other hosts on sputum.

To the extent an infected cell releases its "baby coronaviruses" into the blood stream, they find a hot, chemically active environment – which is already inhospitable to their continued survival – plus a relative dearth of the respiratory cell-walls they're optimized to enter, plus a growing number of hostile antibodies.

(I don't know if a coronavirus-hijacked cell can direct its fresh viruses one way or the other, but it'd probably prefer to do that if it could, and I can understand why viruses wouldn't persist long in blood compared to respiratory-membranes.)

The coronaviruses are definitely active in mucosal surfaces, and carried by mucus. I haven't seen anything about, nor do I think any activity is necessarily implied, in other connective tissues.

As far as I know the route of SARS-Cov-2 spreading in the human body isn't known beyond the change in locality of symptomatic expression with progression of Covid-19 (upper respiratory tract to lower respiratory tract.)

Research has shown (1) the distribution of ACE2 protein, the functional receptor for SARS-Cov and SARS-Cov-2 (2), to include various human organs (oral and nasal mucosa, nasopharynx, lung, stomach, small intestine, colon, skin, lymph nodes, thymus, bone marrow, spleen, liver, kidney, and brain) along with arterial and venous endothelial cells and arterial smooth muscle cells in all organs studied. And ACE2 mRNA has been found in virtually all organs. So there is definitely the potential for cells in the circulatory system binding viruses.

What may be significant here is the concentration of ACE2 protein in cell membranes. One Chinese analysis of a group of 170 Covid-19 fatalities(3) found half of the cases suffered from hypertension. ACE inhibitors are a class of medication prescribed for lowering blood pressure and according to models they increase the concentration of ACE2 protein in lower respiratory tract cells. Given that fatality results from respiratory distress originating in the lower respiratory tract and that not all cases progress to the lower respiratory tract it seems possible that the spread of the virus within the body may be related to concentration of ACE2 protein in cells. The disease progression would also support mucosal transport of the virus within individuals the same as between them.

(1) Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis. https://www.ncbi.nlm.nih.gov/pubmed/15141377 (2) Structure, Function, and Antigenicity of the SARS-CoV-2 Spike Glycoprotein https://www.sciencedirect.com/science/article/pii/S009286742... (3) Controversial Study Links COVID-19 Severity to ACEIs and ARBs (Newsweek, March 24, 2020)

If you're looking for a chemical explaination, I don't know, but the virus can work it's way around quite fine using the lymphatic system, so there's no "fitness" reason that would make it need to hang out in the blood.

Because it’s a respiratory disease

Even if you do everything perfectly a PCR test will still not be able to tell working virus copies from shattered remains.

Or with a well-known phrase: garbage in, garbage out.

Even a 99.9% sensitivity will miss a few times when you're testing millions of people.

Suppose you get the disease and develop antibodies for it. Your body is pretty darn good at fighting the virus and won't get sick from it easily. But, you're in a place with a huge viral load exposure.

Wouldn't that imply there should be some of the virus in your lungs? Test accuracy aside, is the definition of "positive" even clear?