Thread by @hypoautonomic

What will you do with a negative Abbott BinaxNOW result when you suspect COVID-19?

This test catches some infections and might be a net benefit but what if your test is negative? Will you be confident that your close contacts are safe? Maybe. Maybe not.
1/

BinaxNow is a rapid antigen test, giving a result in 20 minutes. That is its main selling point. Waiting for a PCR result to come back is not just a personal agony, but a threat to containment of transmission. A day of delay is a potential day the virus has had to spread. 2/

Suppose last night you get a scratchy throat, and today your body aches, you feel hot and cold. Several days prior you had contacts with numerous vaccinated people and then traveled with an n95. You remember taking off your mask on the plane to eat... 3/

You take both BinaxNOW tests that come in the 24 dollar box, and they both read negative. You remember one of your vaccinated contacts also had a sore throat and got negative BinaxNOW results just before your contact with them... 4/

How confident can you be in this negative result? This is the question that sears as you feel both hot and cold. Was SARS-CoV-2 spreading through your contacts or was it a flu? You read the Binax instructions and the worry begins:
fda.gov/media/141570/d…
5/

It states that a negative result is only presumptive right in the instructions of BinaxNOW. This is a clear warning to not have confidence in a negative result. Why? 6/

You learn one reason why BinaxNOW cannot rule out an infection is because it is not sensitive enough to do so. Infectious virus can be recovered from samples that return negative rapid antigen results. If your swab can infect cells, who can say that you are not infectious? 7/

A CDC study finds infectious virus in samples with negative BinaxNOW results. Blue dots in the antigen-negative columns are culture-positive. Culturing has its own sensitivity issues, so this may not fully capture all false negatives. cdc.gov/mmwr/volumes/7… 8/

Virus can be isolated from samples with a large range of measured CTs. The CT depends on numerous factors like the design of the PCR assay. There is no global CT cutoff that tells you not to worry about samples above X CT's.
virologyj.biomedcentral.com/articles/10.11…
ncbi.nlm.nih.gov/pmc/articles/P… 9/

In other words, an 18 CT may be certainly be full of massive amounts of infectious virus, and a 35 CT or above might not have any infectious virus at all, but 26 or 29? What can we say? nature.com/articles/s4159…
journals.asm.org/doi/10.1128/JC… 10/

Another study with the same theme. Negative BinaxNOW results leave you in uncertainty.
"Our comparison supports immediate isolation for BinaxNOW-positive persons and confirmatory testing for negative persons." 11/
wwwnc.cdc.gov/eid/article/27…

Min. infective dose? unknown. "An accurate quantitative estimate of the infective dose of SARS-CoV-2 in humans is not currently feasible and needs further research. Our review suggests that it is small, perhaps about 100 particles."
cambridge.org/core/journals/…

12/

It is not reassuring to learn there are other studies of BinaxNOW that show better sensitivity It's a different PCR and CTs from different studies cannot be compared directly. STD curves are required to be quantitative.
academic.oup.com/cid/article/73…
13/

academic.oup.com/cid/article/73…
It is even less reassuring to learn the author receives "nonfinancial support" from the makers of the test they are evaluating. What does that even mean? What can you even do with this information??? 14/

You are now days removes from your onset of symptoms, perhaps days away from your peak viral load. That's how viral kinetics might work, but you are 2x vaccinated and a week after your booster. If you were infectious, when is that window? 15/

Is the negative test going to reliably tell you that COVID didn't just spread through your contacts? The instructions warn against using the test this way. Does anyone buying one off the shelve today know this? Did they read that line in the booklet in every Binax box?? 16/

It is strange to see scientists use raw CTs as precise cutoffs, use 'moderately infectious' as if it were a defined category, promote a test that catches 'highly infectious' as a stand-alone key to safety. 17/

As turkey day ends, How many people are buying these off the shelf to use them to rule out infection? Consider how different that application is compared to a screening tool applied by professionals on a routine basis to quickly identify people with high viral loads?
18/

You isolate, book a PCR, call your doctor, and wonder if you wasted your money... 19/19

BinaxNOW according to the label: Even if we were optimistic and accepted their stated sensitivity: 91.7% with CI from 73% to 98.9%. That's not very good. Would you risk using it to prove you were not infectious? With that much variation?!
20/

To use this as a public health tool they have to be in people's noses and interpreted correctly. We could utilize the knowledge we gain from the positive results and reduce transmission- no question. But over TEN dollars on a shelf and used by amateurs? False negatives? 21/

Another paper on BinaxNOW :
1.) intro
2.) Discordant results occur as low as a CT of 20, which could be in the ballpark of like 10^5 or 10^6 viral genomes/ mL!
3.) The discussion shows 53% sensitivity in symptom-less. hope they aren't infectious?
journals.asm.org/doi/pdf/10.112… 22/

This paper explores the range and variation in CT values from all sorts of SARS-CoV-2 PCR tests.
Look how bizarre it would be to declare a CT cutoff in the 20's !!! A cutoff at above 30-something? Okay, now I am listening.
elifesciences.org/articles/64683 23/

But if CT values vary this much over time, for whatever reason, and it is a proxy for viral load, then couldn't the performance of rapid antigen tests might also vary over time? This is sad. Any data on omicron? and rapids yet? elifesciences.org/articles/64683
24/

That reminds me of the people who have grown fond of using a confident CT cutoff in the 20's... This interests me. Michael Mina does this. Does this seem legitimate after reading this thread? A proxy is now a confident cutoff? 25/

So what has Michael been up to as a friend of the rapid antigen test?
Well he called the FDA corrupt and then deleted the tweet, lol. He lectured the FDA and CDC about making money off the pandemic, a principle @michaelmina_lab stands by, which is why he is now a CSO of..
26/

eMed! Your place to buy 150 dollars worth of public health tools that you can for sure interpret with confidence as it relates to your unique clinical context and ever-shifting sands of the novel coronavirus. Well, I guess you do need to call someone about ur test 27/

Three cheers for the public health salesman, revealing to us a privatization of public health phenomenon when the need for testing in the US stays high and unmet. Infections remain rampant as they have been and we have made testing nice and profitable. 28/

Is this a testimonial or an advertisement? CSO of Emed is not the only conflict of interest that michael has. 29/

@DrMattBinnicker Summarizes the problems with attempting to use these methods to infer infectivity. journals.asm.org/doi/epub/10.11…
30/

IDSA and AMP remind us that CT values from qualitative PCR assays "do not reliably correspond to specific RNA concentrations" and that CTs across platforms are not comparable.
idsociety.org/globalassets/i…
31/

This paper demonstrates the variation across different qualitative PCR platforms. "Quantitative RT-PCR is entirely different from qualitative RT-PCR. Ct value itself cannot be directly interpreted as viral load without a standard curve"".
thelancet.com/journals/lanin…
32/

Mina has also promoted a "test to stay" strategy, where an exposed person tests instead of quarantine. What if someone is infectious before the rapid turns positive? This paper shows the sensitivity swinging up right as culture turns + 😨academic.oup.com/jid/article/22…
33/

When Omicron emerged, Mina reassured us with confidence but without evidence that rapid would work great on Omicron. There are mutations in the nucleocapsid gene. Can one predict binding of antibodies to the protein based on that? No. Caution to the wind.
34/

And then preliminary data came. rapid antigen tests may be having problems detecting Omicron; it might even be shed from saliva first. Many rapid antigen tests are not validated for use on saliva.
medrxiv.org/content/10.110…
medrxiv.org/content/10.110…
35/

The dream of widespread testing to find infectious people could still be alive, even with false negatives. Specificity is good. consistently good. But it is not a public health tool if it's a product, out of stock, hoarded by those who can afford it.
36/

Mina frequently lobbies to deregulate rapid antigen tests, telling us it's the answer to access and affordability, as if lower standards and the market can really be the solution. Excuse me but I have overdosed on this hopium.
37/

@threadreaderapp unroll