Ebola Update—  Now up to 12 cases in the latest outbreak, but still confined around Mbandaka.  Sequencing suggests this is a brand new outbreak, and not spread from the larger outbreak from the past year on the other side of the country.  They do have about 289 or so known contacts.  They have only been able to follow up with 88% of them, which is probably insufficient.  There is no indication that the known contacts in the current outbreak in Mbandaka have received the vaccine.  They should definitely be getting the vaccine, as well as anyone they know, to contain this outbreak on the beach head.  WHO seems to believe that the confluence of COVID-19 in the DRC, and two Ebola outbreaks, will stretch the DRC’s ability to respond.  They are probably right.  But confining Ebola to Mbandaka in the coming few weeks will be critical.  If the virus makes it way down the Congo to some of the megacities downstream, you have the recipe for an enormous humanitarian disaster.

Coronavirus Update—
Again, I’m going to be kind of choosy this week.

—It has been the best of times, it has been the worst of times.  It has been a week of conflicting messages pretty much everywhere.  I will give you my take, as always, but no guarantee my take is right.

—So, is there a second wave going on right now?  You are undoubtedly seeing some of the same headlines I am in.  They are stressing that many states (and kind of depends on who you read and when) are seeing an uptick in new cases.  Meanwhile, on the same news page, you’ll read about states continuing to press ahead with re-opening plans.  What the hell is going on?

First off, there is no nationwide increase in the rate of new cases.  In fact, across the US, the rate of new cases is falling:

https://coronavirus.jhu.edu/data/new-cases (accessed 11 June 2020)

(for those of you joining us from the UK and South Africa, data for your countries is also on that page—the UK is over the major peak, much like the US, and South Africa is just starting to come off the high water mark of cases)

Individual states though are seeing different results—and even differences within those states.

So here is Indiana, which is continuing to decline:

https://coronavirus.jhu.edu/data/new-cases-50-states/indiana (accessed 11 June 2020)

Meanwhile, here are states that seem to keep getting mentioned by name in all of these reports:

First off, Texas, which every article seems to mention was one of the first to start re-opening:

https://coronavirus.jhu.edu/data/new-cases-50-states/texas (accessed 11 June 2020)

Texas is showing slow, steady growth, following temporary dips April 26th and the end of May (which helped the re-open decision there).  I would argue this is consistent with a “slow burn” across the Lone Star State, with no clear association to re-opening, Memorial Day, or protests.  Most of the new cases in the last week or so have been in Houston or the surrounding area.

In fact, here’s Georgia, the first state to re-open:

https://coronavirus.jhu.edu/data/new-cases-50-states/georgia (accessed 11 June 2020)

That is at least a leveling off, and maybe the beginnings of a new slight uptick.  Georgia began re-opening on April 24th.  As you can see, cases were actually taking a slight dip up to mid May (the big spike just to the left of the pink shaded region at right), fell quickly again, and started a slight march up after Memorial Day.  That -may- argue for a Memorial Day induced increase, but tough to say for certain.

Arizona has taken off like a rocket, and is the new hotspot at the moment in the US:

https://coronavirus.jhu.edu/data/new-cases-50-states/arizona (accessed 11 June 2020)

Arizona was never really settling down though.  Other states that never really settled, but have shown strong consistent growth in new cases include both Carolinas and California.  California’s new cases in the last several weeks seem to be clustered in LA County and Alameda County (where Tesla’s Fremont plant re-opened despite the county telling them not to).  So even within states with steady growth, it is not necessarily uniformly distributed.  These states have not yet seen their “peak” to fall from for this to be a “second wave” or conclusively a blowout from various re-openings, Memorial Day weekends, protests etc.  That they seem to be warm weather states that are popular vacation destinations is the best argument for a Memorial Day effect, but more details are necessary.  I mean, are all the Houston cases everyone who popped over to Galveston for Memorial Day at the beach?  I dunno.

Other states have seen a distinct second wave, like Alaska:

https://coronavirus.jhu.edu/data/new-cases-50-states/alaska (accessed 11 June 2020)

Iran’s curve looks a lot like this.  In Alaska, though, even that second wave (which is probably the best evidence of any kind of Memorial Day effect out there) is unevenly distributed.  There are some rural counties in Alaska that have yet to record their first case.  

So I think the best we can say so far is that the pandemic is starting to differentiate a bit to a more regional level.  Many states, including former hot spots NY, NJ and Illinois show strong declines in new cases, and no evidence of a Memorial Day Effect (give protests another couple weeks—end of June, like we said last time).  Other states never got flat, and show a continued rise.  States like Arizona and Arkansas are in a sharp exponential peak (by report skewing heavily to large rural families living under the same roof).  Others have evidence of a distinct peak and a probable Memorial Day effect (like Alaska).  

In total, it suggests to me that the pandemic is maturing (in the US at least) from all states being bombarded equally to a more variable, state level experience.  Expect different states, and even different regions within states, to start to go their own way a little more, and anybody making broad national claims one direction or another is not fully consistent with the data.  A national re-lockdown makes no sense in NY, NJ, and Indiana, for instance, but continued lockdown in the Carolinas, or even more aggressive measures in places like Arizona and Arkansas to manage their spikes, is a little easier to argue.  

Who knew the virus was such a strong believer in Federalism?

—It was the best of times, it was the worst of times.  It was a time when asymptomatic, but PCR positive patients were “very rarely” infectious.  It was a time when asymptomatic, but PCR positive patients could still be very infectious—at least by the next press conference.  Yes, that was your week in WHO press conferences, which is “mixed messaging” to say the least.  

So what the hell?

Alright, so here’s what’s going on.  Right now, estimates range from 2% to 41% of the population may, or may have been, infected, but asymptomatically.  The WHO is -guessing- the real value is closer to 16% (which is close to my back of the envelope of “multiply known cases by 5-10”).  There are some models that suggest up to 40% of cases may be spread by people who are asymptomatic.  Only some.  Other models show lower rates.  Depends on who is doing the calculations, how, and where.

As we have said before, every model is wrong, but some are useful.  Some will be more useful than others.

What the same WHO spokesperson said, literally the day before, was that actual, bona fide studies tracing contacts of people who were asymptomatic but popped positive for SARS-CoV-2 showed few, if any, of those contacts went on to get infected.  Hence, the WHO spokesperson said “asymptomatic spread was very rare.”

Again, and I cannot stress this enough, what the WHO says is sometimes unduly influenced by who is paying their bills.  That poor spokesperson was dragged out again the following day to “clarify” that they were speaking only to those studies, and was not making an official endorsement by the WHO.

In my opinion, and again just my opinion, the studies following patients without symptoms but positive tests, checking the people they were around to see if they became positive or symptomatic, are far more powerful evidence than models of transmission.  One is a projection of how someone thinks the virus might act; the other is following the virus to see how it actually acts in the real world.  

There are only three scenarios (generally speaking) where someone can be asymptomatic, but positive for SARS-CoV-2 by highly sensitive PCR:

1)  They just got the virus.  Give them a few days, and they will be symptomatic.

2)  They got the virus awhile ago, are getting better, and you are detecting virus in its dying days.  Think back to the South Korean studies we mentioned a couple reports ago, where they found that virus detected in this class of patients appeared to be inactive virus particles.

3)  They got the virus, but will never get symptomatic from it, and are one of the many people we know that this seems to happen with.

Of those three scenarios, those in number 2) are almost certainly not infectious.  Scenario 3) might be infectious, but the studies cited by the WHO spokesperson suggest most of them are not.  Scenario 1) is the most plausibly infectious.  It all comes down to when they got tested in the disease course, which is tough to know.

But this does have implications for “return to work.”  The growing body of evidence (above and beyond mere models) is that asymptomatic, but PCR positive patients, are not likely infectious.  To rule out scenario 1) though, you probably still want them home for 2-5 days to make sure they don’t develop symptoms.  But if they have been positive for a few days, without symptoms, they may not be that much of a risk to others.  We’ll see how the evidence continues to develop though.

—Another popular question this week was news reports that 23andMe had found links between your ABO blood type and increased risk of severe COVID from SARS-CoV-2.  This was initially reported as a higher risk for Type A blood to be on a vent, but that mostly comes from Italian and Spanish studies in small numbers of patients that found this link (about 50% higher risk).  What 23andMe released, first off, is preliminary and frankly, not especially impressive.  They do show that Type O was less likely to be COVID positive than the other blood types, but this is not peer reviewed, and even they stress this is preliminary data.  Here’s the link to 23andMe’s data:

https://blog.23andme.com/23andme-research/23andme-finds-evidence-that-blood-type-plays-a-role-in-covid-19/

I think it is worth noting that these kinds of studies, where you do massive sequencing and look at a bunch of different genes for association with a particular condition, also known as genome wide association studies (GWAS), are really best thought of as hypothesis generating.  Unless you are looking at simply enormous numbers of people (millions, really) in the study, it is way too easy for studies like this to find “red herrings”—gene associations that occurred by chance alone, since you were looking at so many genes at one time.

In fact, GWAS get their own explanatory box for why they are likely to turn up red herrings in one of the most cited publications in all of the medical literature, John Ioannidis’ “Why Most Published Research Findings Are False”: 
https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.0020124

If you are one of the readers in medicine, that paper is a MUST read.  For those of you that are not, the less technical pointers to keep in mind any time you read about a study in the WSJ, NYT etc. are the bold corollaries (my parenthetical notes in italics):

Corollary 1: The smaller the studies conducted in a scientific field, the less likely the research findings are to be true

Corollary 2: The smaller the effect sizes in a scientific field, the less likely the research findings are to be true

Corollary 3: The greater the number and the lesser the selection of tested relationships in a scientific field, the less likely the research findings are to be true (this is where GWAS runs into trouble)

Corollary 4: The greater the flexibility in designs, definitions, outcomes, and analytical modes in a scientific field, the less likely the research findings are to be true

Corollary 5: The greater the financial and other interests and prejudices in a scientific field, the less likely the research findings are to be true

Corollary 6: The hotter a scientific field (with more scientific teams involved), the less likely the research findings are to be true (recall COVID-19 research had two papers retracted in just the last week as everyone and their brother piles in to publish—and think of how many COVID models have come and gone)

So, bearing in mind that we have a GWAS study in 23andMe and some other similar hypothesis generating studies in China, Spain and Italy suggesting ABO blood type (note, NOT the “+” part, which is a different blood group), the first question we should ask is if there is a snowball’s chance in hell that gene might have something to do with COVID based on what we know of the gene.

So, is it possible that ABO blood type could have something to do with SARS-CoV-2 infection?

To answer that, let’s first talk about how you become Type A, Type AB, Type B or Type O blood.  All of those phenotypes result from slight differences in one gene, the aptly named ABO gene.  That gene encodes for an enzyme, specifically a glycosyltransferase (technically a much longer name, but don’t worry about that).  That’s science-ese for “sticks (transfers) sugar groups (glycosyl-) to other stuff”.  To get Type A blood, you have a version of the enzyme that attaches one particular sugar group to stuff.  Type B attaches a slightly different version of the sugar group to stuff.  Since you get one ABO gene from your mother, and one from your father, you can get a Type A from one parent and a Type B version from another to be Type AB blood.  In this case, you will attach both kinds of sugar groups to stuff.  Type O is simply a defective ABO gene—no sugar groups get attached to stuff by it.  If you get an A from one parent, an O from another, you will have Type A blood, because you can still attach “A” type sugar groups to stuff.

Graphically, that looks like this:

https://upload.wikimedia.org/wikipedia/commons/a/a4/Sugars_that_form_the_H%2C_A_and_B_antigens.png

Now the sugar groups will wind up on your red blood cells, giving you your blood type.  But, any time the ABO glycosyltransferase finds the H-antigen in that diagram (a specific group of sugars), it will stick the A or B or nothing to it as your blood type dictates.  Including to things that are NOT necessarily on blood cells.

In fact, many different proteins will have sugar groups like this hanging off of them.  We call this glycosylation, and it helps these proteins to hold specific structures or do specific things.

Like all coronaviruses in its family, SARS-CoV-2 uses the ACE2 receptor to get into target cells.

Is the ACE2 receptor a glycosylated protein?

Turns out, yes!

The ACE2 receptor is known to be glycosylated at several different amino acids that are important for SARS-CoV-2 (and its family)’s spike protein to bind and thus attack the target cell.  The spike protein -itself- is also glycosylated, and may be worked on by human glycosyltransferases if SARS-CoV-2 successfully jacks a human cell.  If any of the strings of sugars attached to ACE2 or the SARS spike protein ever show the top pattern in that diagram (H-antigen, or a 1,3 linkage of galactose and fucose stuck on a n-Acetylglucosamine), the ABO blood group gene may come into play.  It is possible that this pattern is present on the ACE2 receptor and/or spike protein of SARS-CoV-2 at the parts where the spike and the ACE2 receptor meet.  If the A-antigen, the N-acetylgalactosamine (yellow square), being added is helping the spike protein of the virus attach to ACE2, that might explain the genetic association from Spain and Italy suggesting a higher rate of more severe disease.  Alternatively, the galactose added instead of the N-acetylgalactosamine (a yellow circle instead of a yellow square) by the B-antigen enzyme in people with Type B blood group may actively block the spike and ACE2 interaction, or be less “sticky” for the virus, so the virus has more trouble getting into their cells.  Since Type O blood type is simply the lack of the enzymes that add a yellow square or yellow circle in the diagram, the yellow square of type A blood helping the virus attach is a little more likely, since type O blood group does not have the yellow circle of Type B to actively block either if that is what is going on.  -If- (and I stress that is a HUGE “if”, like a yuge, beautiful “if”, the best “if” ever) Type O is protective, then, it may be because the ABO gene is tagging the ACE2 receptor and/or the spike protein of SARS, and either the A type sugar attachment or B type sugar attachment makes it a little easier for the virus to invade a cell. 

This may mean Type A people get higher viral load on average versus Type B, and Type A/B get more viral load on average than Type O, or the infection persists a little longer.  In either case, the immune system is a little more likely to get revved up.  This would be a nice, consistent story that would back up the Chinese, Spanish, Italian and 23andMe GWAS studies.
Rarely is biology that nice and consistent though.  But, could be.  Worth someone taking a look.

That should be easy enough to test in vitro that someone is probably already doing it.  Or at least they should consider doing it.
Alternatively, this finding may be a total statistical artifact, and no real association with Type ABO blood group actually exists (and none of Type A antigen or Type B antigen is added to the glycosylation chains of the spike protein OR the ACE2 receptor).  If that’s the case, it’s quick to find, and a quick and useful disproving of the GWAS generated hypothesis.

And that’s how science works.

—Lastly, this week’s think piece:  https://www.nytimes.com/2020/06/08/business/hospitals-bailouts-ceo-pay.html?smid=tw-shareI swear we will get to the “social issue” I keep kicking down the road, but this is all of the same piece.  All of it.  What institutions, what “leaders”, still have your consent?

—Your chances of catching Ebola this week are equivalent to the chances that Prince Andrew consents to talk to the FBI about his Jeffrey Epstein connections, as formally requested by the US this week.

—Your chances of catching coronavirus are equivalent to the chances you looked up your blood type when reading about the 23andMe section.

<Paladin>