Ebola Update:

–No new cases.

Coronavirus Update:

–A bit to talk about this week. First, top-line numbers. This part of the week is the critical point for any Thanksgiving bump here in the States. Since most patients will have symptoms 2-5 days after exposure, plus 48-72 hour turnaround time allowance for testing, and we should be seeing how many, if any, got some ‘Rona with their turkey.

And sure enough, we are seeing an increase. For example, for Indiana, this is back to at least a plateau following a week of dropping cases leading into Thanksgiving. Texas, California Arizona and New York (just to name a couple big population states) are all seeing rising new cases. On the Johns Hopkins map, most of the nation is green today–I would expect it to be at least light pink in the next couple days.

So the next big test will be the next couple of weeks. Those will be the late breaking Thanksgiving cases, but more will be the knock on cases from people infected by people who get infected at Thanksgiving. The next ring of Thanksgiving infections, if you will. This will tell us how much immunity now exists already with two big waves having hit thus far. If the new cases in the next couple weeks are merely plateaus or less in most places, Q1 2021 should be pretty close to herd immunity for many places stateside. If we get LOTS of new peaks in the next couple weeks from knock on spread, those dire predictions of a “dark winter” look a little more probable and we are definitely on the late end of the back of the envelope UFC estimates of time to herd immunity. This will vary by location. I expect places already hard hit by the first two waves will plateau or less. What this will do is tell us where the virus has not penetrated enough yet and is still finding susceptible new hosts.

The silver lining is that despite two waves and what I expect to be sustained, relatively widespread moderate-to-high transmission for the next couple weeks is that reinfection is still very, very rare. A study released out of Japan this week confirms this. They were using a pretty sensitive assay for neutralizing SARS-CoV-2 antibodies, or antibodies that block viral replication. Of 376 confirmed positive patients they tested, 280 had only mild or no symptoms. Despite this, 97% of them had neutralizing antibodies to the virus 6 months later. Of the patients with moderate or severe disease, 100% had neutralizing antibodies 6 months later. And this isn’t even measuring T-cell response durability, which, again, we have good reason to suspect is as important, if not more important, for immunity to the virus.

–Globally, some countries are still surfing the crest of the second wave (Germany, Brazil, Russia, Turkey) while others are coming down (UK, India).

–Treatments continue to roll out, and we continue to wait for publication of vaccine trial results, although it seems approval of several seems close (I believe the UK has already approved at least one). Vaccines appear most likely to be made available first to frontline healthcare workers and nursing home workers. This makes it sense– it should help expand effective hospital capacity by preventing staffing shortages.

Again, effective hospital capacity is the main public health threat of COVID. The “bed’s taken” overwhelming of the local hospital system.

We’ll get back to that momentarily too.

I would -guess- the next major group to get the roll out will be people with known at risk conditions, but that is totally a guess right now.

–In other news, the WHO has adjusted guidelines to a 10 day quarantine following positive diagnosis. This fits with data from Korea showing that after symptoms disappear, the chances that virus still detected in the patient is “live” and capable of infecting others is dismally small. They also believe that a 10 day quarantine is more likely to be adhered to than a 14 day quarantine. They’re probably right. Your employer (really, their lawyers) may have some different ideas about an appropriate quarantine though.

–Dr. Fauci announced that school closings are probably counterproductive, and, as we have long been discussing here, kids are not at significant risk from COVID. With appropriate precautions to keep them from spreading the virus to slightly more susceptible parents or worse, much more susceptible grandparents, Dr. Fauci appears to be advocating for increased in person schooling. I know my kids school, and the public school districts around here, are still all virtual until January 15th though. So your actual mileage on this latest about face may vary.

–Which brings me to the meat this week. It has been a big week for readers reaching out with questions on corona coverage they are encountering out there. I encourage it–please, by all means keep it coming, because sometimes truth has to know it needs to put its pants on and start running to catch up to some less than accurate things out there.

–So let’s start with something I have seen a couple times myself, but unattributed, and thought was odd, but it was clear that people quoting the statistic had never done PCR before, so I shrugged it off as ignorance.

The idea here is that “there are studies out there that show that a positive SARS-CoV-2 PCR result with a Ct of >35 has a 97% chance of being a false positive!” A quick Google search today pulled up front page results for variations of that statement with ledes like “the vaunted PCR test is wrong!”

Without getting too far into the weeds, Ct is inversely correlated to the copies of the target present in the initial sample. If there is a lot of virus on the swab, it should be a LOW Ct value. If there is not a lot of virus, it should be a very high Ct value. A Ct value of >35 implies there may not be much virus present–but there are a multitude of reasons for that. As we have mentioned, you are swabbing snot or saliva, which does not distribute the virus as uniformly as blood, and by random chance didn’t get much virus. There may be an inhibitor of PCR present, and the reaction is running slow. But this is a qualitative assay.

As I have stressed over and over again, the mainline PCR assays for SARS-CoV-2, including the one we run, are only for detecting the presence of the virus. We don’t routinely report Ct values because the sample type, for reasons above and previously discussed, introduces a little too much randomness into quantitation of viral load based on those Ct values. This is why most drugs that have reported looking at viral load changes based on Ct values from RT-PCR have struggled to do so, and why you should pay attention to one that manages to eke out a “win” there because it may be dropping the average viral load enough to overcome all that randomness.

Now, everybody waving this “97% false positive with Ct>35!” flag is correct that MOST labs sign out positive PCR results with a Ct > 35.

Where they lose the script is the source of that statistic and how generalizable it is to all the various PCR assays looking at SARS-CoV-2.

Not all of them are the same. We use the CDC’s assay, which is two primer and probe sets for different portions of the SARS-CoV-2 “N” gene. Another popular target for other assays is the “E” gene, which may or may not be getting run with one or more other targets on the viral genome. Aside from differences in equipment, reagents, temperature cycling protocols, differences in primers and probes and what you are targeting will all make what happens in one form of PCR assay different from what happens in another, especially when it comes to Ct value, and even when looking at the same target.

Same target, same sample, on different equipment can and most likely will have some differences in Ct value.

Just as an example, I mentioned we look at two targets on the gene. The N1 primer probe set runs a little “hotter” than the N2 primer probe set in general–to the tune of about 0.5-1.5 Cts on average.

The main point to remember is there is not one ur-PCR SARS-CoV-2 test that rules them all. On the list of EUA approved assays at the FDA’s website is a huge variation in equipment, reagents, cycling parameters, and primer/probe sequences, all of which making a given Ct value for one assay not quite the same as a Ct you might get on a different assay from that list.

A 35 for one may not be a 35 for another.

To make an analogy, it’s like saying a Ford F-150 gets crappy gas mileage, so a Tesla must too. They’re both cars, right? But they don’t run the same kind of engine, have totally different chassis, tires, drives and are used in different ways (a Model S ain’t hauling all your buddy’s stuff when he moves).

Got all that? Good.

So the specific reader question came as part of a “letter to the editor” about one specific published PCR assay that they are asking to be retracted because they do not believe that assay was sufficiently validated. Fair enough–most of that letter is between its authors and the papers authors. The problem is that this letter to the editor is being generalized to all PCR assays, because if you don’t know that there is not one PCR assay that everybody uses, you can easily interpret their criticisms as being of PCR in general instead of this one specific published assay (which is all they are criticizing). That said, the letter writers in their zeal to point out some method limitations quote the origin of the “97% false positive rate for Ct>35.”

So I was finally able to find it!

It’s here.

It’s a “correspondence”, and not a full publication either–probably because it would not have survived peer review with some serious additional work. The gist here is that for a PCR assay looking at the E gene specifically, and on the equipment and method that the letter writers used, a sample with a Ct value of >35 would not successfully culture in 97% of samples with a Ct value of >35. From this they conclude, like previous Korean publications we have talked about, that patients who have recovered from coronavirus, but have persistent high Ct value PCR positives, only rarely have “live” virus.

That’s not a false positive, like everyone touting this claim states, including the authors of the letter to the editor above. While the assay they are criticizing uses the E gene as one of its 3 PCR targets, this “correspondence” showing 97% of samples with Ct values of >35 using their exact method (which the 3 PCR target assay does not do) did not successfully culture used only the E gene. The 3 PCR target assay may have different Ct values since the E gene is working “with” 2 other gene targets, and may be reacting more or less efficiently. It may have better correlation to culture overall. Just because the F-150 has poor gas mileage, doesn’t mean the Tesla does.

The bigger problem for the “correspondence” submission is this is correlation to culture and not clinical results. We have no idea where in the clinical symptom spectrum their samples came from. For example, we have had people pop who were clearly abusing our assay as a screening assay (did not have symptoms when tested) with Cts of about 35 or better call and gripe about how our test was clearly wrong–only to develop a fever within 48 hours and lose their sense of smell and taste, with repeat tests in that span with Cts much lower (implying higher viral load). In those patients, the PCR result, even with a Ct of >35, was clinically correct–those patients were very early into a true positive clinical case of COVID.

How many of those you get depends entirely on who you are testing. If you are getting mostly patients looking for that negative PCR test to get back to work after being positive for weeks, yeah, a Ct>35 is a true positive for the virus being there (it’s a qualitative test, remember) but probably not for particularly infectious virus based on studies similar to this commentary out of Korea where they looked exclusively at this patient population. Presumably, based on their conclusion that their results corroborate the Korean studies, this is the population that the “commentary” that kicked off the “97% false positive with Ct >35” was studying.

On the other hand, if your assay is being generally abused for screening of asymptomatics, many, if not a majority, of your low positives (with Ct >35 and of course, depending on your assay) will be true positives. True positive detection of SARS-CoV-2, and true positive for an infection that is juuuuust getting started.

That absolutely happens.

And we in the lab have no way to know if the patient is just getting over an infection or potentially just getting started. It’s far better practice of medicine to call a positive (because the virus IS there–the PCR has found it, and is VERY sensitive and specific for it) and get that patient a 10-14 day quarantine than ignore a Ct positive and send them merrily back to their job at the nursing home working the bariatric and diabetic wards with a “negative” that is really an infection just getting started.

But the main point is that “commentary’s” unfortunate “97% false positive with Ct > 35” is taken totally out of context by Bob The Internet Guy who has never done by PCR, but has read about it killing time during quarantine, and assumes it applies to ALL forms of PCR for SARS-CoV-2.

As we mentioned, does not generalize.

Same for the assay being criticized in the “letter to the editor.” Our assay and theirs does not use the same targets on SARS-CoV-2, does not use the same protocol or reagents, and does not use the same equipment.

The primers and probes we use (cannot speak to the ones in the paper they are writing the letter about) are absolutely specific for SARS-CoV-2.  We did test against full length viral genome (even though the inactivated virus we got sent second hand in March did not keep for long), we tested again with a full length viral genome control construct, and with also full length viral genome PLUS closely related coronarviruses sent to us by the FDA as part of their proficiency kit in a series of standard dilutions as a condition of our EUA. We detect the lowest concentration of the FDA standard full length viral genome without any cross reactivity to other coronaviruses we have tested (at any point) at Cts of ~35-37.

Also worth mentioning that we have sequenced a lot of the PCR (as many as we have consent for), and we find SARS-CoV-2 sequence quite specifically when the PCR says it’s there–even at Cts > 35 with our assay.

Further, our primers and probes, tested against the “real thing” of other closely related coronaviruses has NOT shown false positive detection of those viruses when we have tested their full length genomes sent to us.  Only SARS-CoV-2.  I know all of the commercially available PCR tests have had to do this as well–you can probably find the results for many in their FDA EUAs at the FDA website.  Although frankly with so many using the same target sequences (we’re not the only one using the CDC’s primer/probe set), there may be less demand to show the same primer probe set is as specific as it has been for everyone else so far.

An F-150 ain’t the same vehicle as a Tesla. What’s true for one may not be true for the other–and when it comes to PCR testing, you really do have choices in make and model. And that choice may matter.

Make sure the complaint about the F-150 really does also apply to ALL other cars before you make it.

–Did you hear about the paper from Johns Hopkins where someone there went looking for excess deaths due to COVID and couldn’t find them, and then the entire publication was immediately disappeared down the memory hole?

Yeah, I heard about it too, and a reader found their way to the Wayback Machine to find the write up of that data in the Johns Hopkins student newspaper.

Because it wasn’t an actual paper, but an article about a slide deck from a webinar by the associate director of a Johns Hopkins economics program who wanted to find if there were excess deaths due to COVID, found some data, and did some manipulations of it.

It hasn’t been published via peer review. Further, the newspaper did take it down and ultimately retracted the article. In their retraction notice, they took issue with the article summarizing the webinar as the data presented percentages of deaths in each age group, rather than raw numbers (which are a little more meaningful in this context) and attributed a decline in deaths from things like cardiovascular disease and other known COVID risk factors to those being called COVID deaths instead.

“Aha!” went Bob the Internet Guy. “Ever since Plandemic I knew they were lying to me about the severity of this disease! And paying the hospitals to cover it up by changing the death certificates!”

Yeah, hold up there, Bob the Internet Guy. The relative decline in deaths from things like cardiovascular disease is easily attributable to them dying from COVID as they are more likely to do. In short, COVID got to them before the heart attack did. We also covered the whole “how to fill out a death certificate” portion before. Several times. All of you should know better than this by now.

And raw numbers are more meaningful here. Another reader asked if I had seen those yet–I had not. But I found them!

Great resource is the CDC here, who tracks weekly morbidity and mortality for all causes in the US and has done so for years. They have set up a special dashboard in the time of COVID. You can find that here.

But the money graph is this one:

The blue bars are total deaths. The yellow line is the smoothed average where total deaths are enough standard deviations above their average to now be “excess.” Notice from April to now that the blue bars are well above the yellow line in 2020. This has earned them red crosses, marking time periods of excess raw deaths of all causes. Note the peaks appropriately lag the major “waves” of coronavirus in the US thus far (remember–deaths will lag positive confirmed case numbers) AND track with what we have said about the severity thus far. It was worse mortality wise in the first wave. This second wave is more infectious, but broadly less severe.

And before you ask, the 2018 flu season was unusually severe. THAT is what a bad flu year will do for you. Notice 2020 exceeds this is raw numbers and staying power. You can also find other raw data of excess deaths here.

This is an exceptionally bad flu. I am not, and have not been, making that up.

Now, we don’t have clarity on how many of the blue bars in excess are due directly to coronavirus and how many are due to knock on effects of locally overwhelmed hospitals. Remember how I told you to remember that the risk of the virus is all cause mortality due to “bed’s taken”?

There it is in plain, raw numbers.

Sorry, Bob the Internet Guy. The virus is very real. No, it is not a direct threat to most people. But there are categories of people (we know them well) who ARE at an elevated real risk of dying. They are catching it. And some of them are dying. It’s also putting enough people into the hospitals to threaten increases in all cause mortality–that will be happening on the local level. Just because it is not happening to you, or around you, doesn’t mean it’s not happening!

For example, there are hospitals in Indiana getting pretty close to full. There are hospitals in Wisconsin that we just mentioned before triggering overflow plans.

I said wayyyyyyyy back in the beginning of these that balancing the negative effects of a total lockdown (which are severe, especially the longer it goes) versus direct effects of virus would be tricky and would not be a decision I want to make. You will be damned if you do and damned if you don’t. A good summary of the human costs to prolonged and draconian lockdown are at the back end of this article.

But I know people who have had colleagues in the ICU with COVID–just recently. I know of several practices in the area that have had physicians wind up in the ICU too. Fortunately all of them have survived. I know some area school employees who have not. One in 13 long term care or nursing home patients in the US have died this year–my understanding from my buddies in geriatrics are that this is unusually high. Anecdote, sure, but it matches the CDC and Our World In Data trends above.

Again, don’t get it twisted. I do stress to you that most people will get through SARS-CoV-2 quite well, and your individual risk is not particularly bad. For most age groups, it’s not even a severe flu. But in aggregate, the virus is a real problem. The lockdowns at the beginning of the year were probably justified given the unknowns, the need to build to our current testing capacity, and the hospital overflow that has allowed most places to weather relatively well.

Remember, in general, the wiser question to ask is not “why is Bad Event X so bad” but “why wasn’t it worse?”

The balance between cost of lockdown, and how long to maintain them, and when can really go back, is very much a political decision. The HUGE local variation you are seeing is to be expected then. And yes, there is little in the Lancet publication we published a while ago to suggest it -stops- the virus. But it did buy important time. Again, I agree with the WHO that full lockdowns over entire large nations and continents are likely overkill at this point–but more local lockdowns to pause spread and keep a local hospital system from being overwhelmed does still make sense if you are locally reaching that point.

You can get some data to that effect to help that decision, sure. Data like beds available vs. beds occupied currently and what the positivity rate for coronavirus is right now for example might help. But when the rubber meets the road, there is still going to be a LOT, and I mean a LOT, of feelingsball to making the decision to lockdown, even locally.

It’s A. Political. Decision.

And I am not a politician.

Bob the Internet Guy has an opinion on the politics of lockdowns, and Bob seems to be goalseeking whatever wispy fragment of scientific veneer he can find to support his opinion. But Bob is quite often ignorant, ignoring or dismissing (often via ad hominem) anything that doesn’t agree with his opinion. This is classic confirmation bias. This is also leading to very confused readers stumbling on these–again, keep these coming. If something doesn’t sound quite right, happy to help sort it out if I can.

–Thank God this wasn’t the Black Death. Can you imagine the updates I would be having to do if it was? Yeesh.

–Your chances of catching Ebola are equivalent to Michigan being healthy enough to play Ohio State in football next week.

–Your chances of catching coronavirus remain exceptionally good in many places in the world. Continue to be as prudent as your risk tolerance allows.

<Paladin>