Wednesday, August 12, 2020

The Great Debunk

Nightmare, by Henry Fuseli – featuring a demonic and apelike incubus 

Treatment update

Injected interferon-beta-1a (Rebif): The National Institutes of Health have begun a randomized trial comparing hospitalized 1000-plus COVID-19 patients to receive either remdesivir alone or remdesivir plus injected interferon-beta-1a. This is great, but it won’t tell us anything about the benefits of remdesivir, which have remained murky. Hopefully thINSERM DisCoVeRy Trial, which has been going on since March, will eventually come up with some data.

Inhaled interferon-beta: Under the unsexy name of SNG001, inhaled interferon is already making fortunes. The shares of the company that concocted it, Synairgen, have soared from £35 to £206 at the first hint their drug might do some good. 

Tocilizumab (Actrema): Another promising treatment strikes out. Roche’s own global, randomised, double-blind, placebo-controlled phase III trial among patients hospitalized with COVID-19 found no evidence that the drug improved clinical status or prevented death. We still need to see all the data in a published study, but if the manufacturer feels obliged to announce such a discouraging picture we should believe it. Their drug had also done badly in a previous trial in Italy. Roche say a trial in patients with more severe disease will go forward, but it’s hard to be optimistic.

Favipiravir (FabiFlu): The results of a small randomized trial of this antiviral on mild-to-moderate COVID-19 patients are said to show faster improvement in patients given the drug. Unfortunately the study is not only very small but is open-label, without a placebo control, so patients and doctors all knew who was receiving what, which is a severe limitation. As so often, all we have to go on is the manufacturer’s enthusiastic announcement, not any data that’s published or even submitted for publication.

Hanuman: An Indian Member of Parliament predicted the monkey god would end the pandemic if enough people prayed to him consistently until August 5th. He must not have met his recruitment goal, since India has been accumulating new cases at a rate that rivals the United States – around 60,000 a day. The author of the linked article included a subtitle explaining helpfully that “The Bhopal MP's claim has no scientific backing.”

Gasoline: The  Philippine President Rodrigo Duterte proposed the public disinfect face masks with petrol. During a speech, Duterte said if you can’t afford Lysol just “Drench it in gasoline or diesel, and that son of a bitch Covid won’t stand a chance.” A health official came forward to disavow the suggestion 24 hours later, hopefully before too many people asphyxiated themselves. (Dozens of  Americans, remember, drank bleach on the advice of their president.)

Hydroxychloroquine: That study from Detroit I tore apart a couple of posts back has now been blasted by many other scientists. They point out various important flaws but, interestingly, nobody except me seems to have picked up what I consider the fatal one, i.e. that patients who received hydroxychloroquine were apparently treated later, when treatment was already better overall. I wanted to send my own Letter to the Editor but when I found out the journal would charge me $600 for the privilege my enthusiasm faded. 

Monoclonal antibodies: Lilly is about to start placebo-controlled trials in hospitalized and non-hospitalized COVID-19 patients of an antibody they’re calling LY-CoV555, isolated from the blood of a coronavirus survivor in Washington State, that for some reason is believed to be particularly potent.

Vaccines: I find it somewhat bizarre that the Moderna’s new Phase 3 trial of its mRNA-1273 COVID-19 vaccine – that’s the phase that tells you whether the vaccine actually prevents disease – is recruiting its 30,000 American volunteers randomly via the internet, instead of going into hospitals, nursing homes, prisons, ICE detention centers, factories, and other hot spots to target high-risk groups. I filled in their online questionnaire, hoping in vain to find out what kind of volunteers they were looking for. But I did find at one study site’s website: “Researchers are particularly interested in recruiting those at high risk of COVID-19 infection . . . and adults who are at high risk for severe COVID-19.” At least that. If you think you are at serious risk of contracting severe COVID-19, consider signing up.

Sputnik V vaccine: The Russians, on the other hand, claim they’re going to start vaccinating millions within weeks with their Sputnik V anti-COVID-19 vaccine, without bothering to put it through any Phase 3 trials at all. We’re not even sure it’s been tested in monkeys. As I mentioned a while back, the Chinese are already doing the same with members of the People’s Liberation Army. Pretty nervy – no way I’d let anybody shoot me up with a vaccine before we know it works.

Putin claims one of his daughters has been vaccinated—the endocrinologist or the rock-n-roll dancer?

CoronaVac: The Chinese company SinoVac seems to have rebaptized its PiCoVacc candidate vaccine CoronaVac in time to report Phase 2 results: the two-dose vaccine does elicit an antibody response in human volunteers, with relatively few side effects.

Moderna monkey business: In important vaccine news, Moderna has now caught up with the OxfordSinovac, and Janssen groups by testing its candidate vaccine, currently called mRNA-1273, on 24 rhesus macaque monkeys. Eight received two low doses of vaccine a month apart, 8 received high doses, and 8 were controls. The low-dose monkeys developed a modest amount of antibodies, fewer than human COVID-19  patients do. But high-dose monkeys produced huge amounts of antibodies, and a vigorous T cell response (the other major branch of the immune system). The researchers then planted the novel coronavirus in the animals’ noses and windpipes. Both doses of the vaccine effectively blocked the virus from multiplying in the lower respiratory tract, and the higher dose blocked it in the nose as well, suggesting the vaccinated animals might never become contagious. And while lung biopsies showed that control animals all developed big-time pneumonia, the low-dose vaccine reduced it to mild, and high-dose vaccine prevented inflammation altogether. 

Another vaccine, Ad26 from the Harvard-Janssen group, has also been shown to reduce the amount of virus in monkeys’ lung fluid, and vaccinated animals had “only mild symptoms,” but from what I can tell by their first and second published reports the benefit was neither as powerful nor as well documented as the one from Moderna – they didn’t, for example, do lung biopsies. Sinovac’s PiCoVacc (now CoronaVac?) and ChAdOx1, the Oxford vaccine, also don’t seem to have done as well in monkeys as the Moderna vaccine – many animals did get coronavirus pneumonia, though it was less severe. So score one for Moderna.

Baby superspreaders?

Is it true that – contrary to all previous evidence – “young children spread COVID-19 more efficiently than adults”? I’ve looked very carefully at the 2 studies cited by the article by that title, which has gone viral enough to worry many of my friends and my blog/Facebook followers, and I have concluded the answer is no.  

In the first study, Chicago researchers measured the amount of viral RNA present in the noses of preschoolers who were sick with COVID-19 and found it was at least as high as in adults (this is not a new finding, btw). However: 1) there is no clear relationship between infectiousness and the quantity of viral RNA in the nose – RNA detected by swabs is often just fragments, incapable of transmitting disease; 2) those RNA-carrying kiddies are unlikely to be putting much of anybody at risk, because they are as rare as hen’s teeth and thus. The reason is that the few preschoolers who do get infected with the novel coronavirus very rarely get sick – only about 1 in 100 baby viral carriers ever have symptoms – and the researchers only tested kids who had symptoms.

The second new study describes old results: based on contact tracing in Trento, Northern Italy, early in the pandemic, children under 14 supposedly passed on COVID-19 to 22% of their contacts – mainly parents and siblings – a higher percentage than any other age group. There were only 14 children among nearly 1500 cases, and 8 of them got at least one other person sick. But this report has a huge problem: the kids, and their contacts, may not have had COVID-19 at all. Thirty percent of the index cases and 45% of the secondary cases were never swabbed for testing. This is a gigantic drawback, because the vast majority of children with symptoms that suggest COVID-19 turn out not to have it – not a single one of the 119 symptomatic school contacts who underwent confirmatory testing in one study. So we have a very weak study. It is also an outlier, contradicting the very low contagiousness of children found by studies that did, on the contrary, confirm diagnoses with PCR swab testing: in schools in IrelandSingapore, and Australia, in households in GuangzhouKorea, and New York, and several others included in a literature review,  

I have to admit, though, that after sort-of agreeing with President Trump on the subject of school reopening, I am getting cold feet. The reason is the vast difference in the stage and severity of the epidemic between countries that have successfully reopened or never closed their schools (early stage, mild severity), and the dreadful current conditions in the United States of America. When 20% of COVID-19 swabs are coming up positive, as for example in Mississippi, the chance that presymptomatic teachers and adolescents will introduce the disease into schools, and thence to classmates and their families, is so great that it would be dangerous to reopen. I agree with Dr. Birx – hand-picked by Trump for her pliancy but now on his shit list for gently championing science – that a 5% positivity threshhold makes sense.


Georgia high school where 9 students and staff came up positive

Float through the air with the greatest of ease?

Like the Daring Young Man on the Flying Trapeze, the novel coronavirus travels from the near corner to the far one, where it lurks for hours waiting to drift into your nose.


The latest offering from the aerosol lobby: an op-ed piece in the New York Times from Linsey Marr, an engineer who’s published similar articles before.

She brings up the usual two examples of putative aerosol transmission. First the infamous Washington State choir rehearsal that caused 52 cases and 2 deaths. Let’s look more closely at this one. Not only was the COVID-19 carrier – patient, rather, since he already had symptoms – singing at the top of his lungs all over various configurations of fellow-choristers seated shoulder to shoulder, but also the whole bunch hung out during (snack break) and after the rehearsal (stacking chairs and “congregating”) in a poorly ventilated indoor space. Plenty of opportunity for the index patient to pass the virus on via droplets and contaminated surfaces, without the need to invoke aerosols.

Then there’s that restaurant in Guangzhou, China. If you look carefully at the original report all the fuss turns out to have been about exactly one patron who was seated more than 6 feet away from a presymptomatic case – about 9 feetaway – and later became ill and infected a family caretaker who had also been at the restaurant. I’m not entirely convinced the infamous air-conditioner was at fault, since it was in fact blowing air away from not toward that customer, and since no viral RNA was detected anywhere in the air con system. A single case is not much to hang a theory on, especially when the disease was already spreading in Guangzhou and when we don’t know whether the two patrons, say, went to the bathroom at the same time.

The New York Times op-ed author himself doesn’t seem impressed with those examples. First he admits that “some scientists” think “If SARS-CoV-2 were primarily spread by aerosols, there would be more evidence of long-range transmission.” And then, buried in the depths of his article, he goes on to say, "I agree that long-range transmission by aerosols probably is not significant,” explaining that aerosols are probably transmitting disease only within 3 to 6 feet. But we’re already keeping that distance based on droplet transmission! It is precisely the fear that virus is being wafted far away that has everybody agitated about aerosols. I wish Prof. Marr had stated that opinion up front and prominently.

“Some scientists” presumably include a Harvard group who published a careful, thorough, practical, and readable review of the evidence that concluded, “There are no perfect experimental data proving or disproving droplet vs aerosol-based transmission of SARS-CoV-2. The balance of evidence, however, seems inconsistent with aerosol-based transmission of SARS-CoV-2.” 

new article in The Atlantic, under the excellent title “We Need to Talk About Ventilation,” emphasizes the vital importance of keeping doors and windows open and – correctly, in my view – debunks the use of masks out-of-doors. Unfortunately it also perpetuates the notion that viral particles are “floating around the room.” I’m also not impressed with the revisionist computer modelling about the Diamond Princess outbreak which features airborne transmission – the gross errors already described on board, from unchecked indoor partying to infected food staff carrying meals from cabin to cabin to the utter lack of either contact isolation or face masks, seem to me to explain spread of the virus without invoking aerosols.


Paranoia is reasonable

An essay from “a senior British doctor” has been making the rounds, claiming there’s been so much progress in treatment of COVID-19, and the chance of survival is so much greater, that we no longer need to be “unreasonably paranoid.” Whoever the original author of this piece – its inaccuracy and misuse of scientific language suggest that no British doctor, senior or junior, is to blame – please do not believe it. Yes, we’re better now at treating COVID-19, but paranoia is still the proper stance. Before running through the article’s points one by one I’ll give you my takehome message in one graph:

I.e. in a country with some of the best hospitals on the planet (about other aspects of the American health care system, don’t get me started) people are dying – and how!

One multinational study has examined the impact of improved COVID-19 treatment, and found that the death rate of patients in intensive care units did in fact fall between late March and late May when current guidelines were codified – but only from 60% to 42%. 

(You may remember the crazy-high mortality rate of 88% reported for the very sickest COVID-19 patients, the ones requiring mechanical ventilation, at the beginning of the New York City epidemic. It made a huge media splash, making all subsequent statistics look like vast improvements. But that figure was quickly revealed to have been miscalculated, and the authors corrected it within days to 24.5%. Of course it’s the 88% that stuck, though.)

Now, here are the author’s five supposed great leaps forward in COVID-19 therapy:

1)    Using aspirin and heparin blood thinners. Yes, the disease involves widespread clotting, and yes, blood thinners are now given to all hospitalized patients. But the novel coronavirus is so powerful that clots develop nonetheless.

2)    Measuring oxygen saturation in COVID-19 outpatients. This might help a bit, but the only evidence we’ve got is one small, uncontrolled study whose winning result was that 4 patients whose oxygen gadgets showed low levels at home went to the Emergency Room and were admitted to the ICU.

3)    Favipiravir and remdesivir. These antivirals both give at most minor benefit. 

4)    Steroids against “cytokine storm.” Dexamethasone does decrease the risk of death in very sick patients, but only from about 40% to about 30% – still far from zero.

5)    Proning. Front-line physicians think it probably does help some, and one small study suggested improved oxygen saturation in COVID-19 patients who lay face down instead of face up. No controlled trials have been performed, though, and there is no evidence of an impact on mortality.


99% Lunacy

A blog post called Lockdown Lunacy by a gentleman named J. B. Handley, who was previously known chiefly for promulgating the myth that childhood vaccines cause autism, has landed in my inbasket. It may have landed in yours too, so I thought I’d give it a look.

I will run through his 16 supposed facts, omitting those that are accurate (kids very rarely pass COVID-19 to adults, outdoor air is highly protective), those that are repeats, and positions on masks and distancing that we can charitably hope he will have discarded given all the new evidence since he posted in late May.


1)    “Fact” #1: COVID-19 is no more deadly than the flu. 

Echoing Trump’s “99% Harmless,” and just as wrong. But worth debunking in detail.

We can now estimate COVID-19 mortality rates fairly well, on the basis of an excellent CDC study that ran tests on left-over blood that had been drawn for routine purposes. It showed that about 10 times as many people have been exposed to the novel coronavirus as get counted as cases. Combining that with the usual estimate that COVID-19 deaths are undercounted by 30%, I have calculated a mortality rate of 0.6% among all people who have been exposed to covid-19. This is 6 times higher than the usually cited mortality rate for seasonal influenza.

But there’s more, if you will bear with some wonkishness. I have now dug more deeply into that seasonal influenza death rate of 0.1%. To my amazement, it turns out to be based not on antibody tests but on estimates pulled out of a hat. Scientists started with the reported number of influenza hospitalizations, then guessed how many more people had gotten sick. I.e. the often-quoted influenza death rate does not even claim to be among all the people who were exposed to influenza, only among those with symptomatic flu. Other researchers have, however, performed antibody studies that can help obtain a more accurate denominator. Those show that between 65% and 85% of people infected with the flu never develop symptoms – a higher rate than for the novel coronavirus, by the way. This means all the comparisons between COVID-19 and flu have been using a mortality rate for influenza that is overestimated by a good 70%, and that the COVID-19 death rate is not six times higher but at least 10 times higher than the death rate from the seasonal flu.

In any case, as one astute observer has pointed out, there is only one real answer to the question “How deadly is COVID-19?” As per my graph above under “Reasonable paranoia”: Deadly enough.


2)    “Fact” #2: The risk of dying from COVID-19 is much higher than the average IFR [influenza fatality rate] for older people and those with co-morbidities, and much lower than the average IFR for younger healthy people

Most COVID-19 fatalities are in elderly persons with preexisting conditions. But between 20% and 30% of COVID-19 deaths in the USA are in people under age 65. As of today about 168,000 deaths in the US have been attributed directly to covid-19, so the disease has killed about 33,000-50,000 Americans under 65. Reassuring calculations to the contraryby John Ionnidis – he of the notorious Stanford study – are misleading, as is unfortunately usual for him nowadays.

It is worth mentioning that death is not the only outcome worth taking into consideration. Even when it doesn’t kill a young person, COVID-19 can be a bitch of a disease, making life miserable for months after a “mild” case (such as the one that affected CNN’s Chris Cuomo).


3)    “Fact” #3: People infected with COVID-19 who are asymptomatic (which is most people) do NOT spread COVID-19

Many “asymptomatic” people are actually “presymptomatic,” i.e. are going to get sick within days. And that’s precisely when their ability to transmit the disease hits its peak – during the 2 or 3 days before and the day of their first symptoms. This accounts for a majority of COVID-19 cases. After that patients become much less contagious, to the point that a week or so into their clinical illness patients can probably be considered safe to share your space. The latest confirmation that presymptomatic and some asymptomatic people are at least as infectious as patients with COVID-19 symptoms comes from a recent study from Korea that has been highly publicized but only confirms what we already knew.

(On this point Handley is not only wrong but also surely a liar: he cites an off-the-cuff statement by a WHO official who realized the very next day that she had misspoken and took it back.)

The truly asymptomatic, who are infected but never become ill, can also transmit the disease during the week or two they are swab-positive for virus, though they shed the virus more briefly. Regardless of symptoms, people infected with COVID-19 can remain swab-positive for months after they have actually stopped being contagious, because (as I already mentioned) positive swabs often reflect viral fragments that aren’t capable of transmitting disease.


4)    “Fact” #4: Emerging science shows no spread of COVID-19 in the community (shopping, restaurants, barbers, etc.)

Transmission in the community doesn’t happen? Yes, it’s rare between shoppers in stores, where people pass each other momentarily, but not rare for the unfortunates who, say, spend all day working the cash register. Restaurants are, on the contrary, known to be major culprits.

The more people ate in indoor restaurants, the more COVID-19 cases 3 weeks later

I don’t know whether anybody’s studied barber shops and beauty salons, but they were closed during lockdowns, and they have by and large reopened cautiously, taking precautions to protect clients and staff, so I’d be surprised if they caused many cases. 


Well, some are more cautious than others...

5)    “Fact” #8: The idea of locking down an entire society had never been done and has no supportable science, only theoretical modeling

“Fact” #10: The data shows that lockdowns have NOT had an impact on the course of the disease.

Let the data speak for themselves – the EU really locked down, the US did not:

6)    “Fact” #11: Florida locked down late, opened early, and is doing fine, despite predictions of doom

Oh boy did he flub that one.

7)    “Fact” #16: All these phased re-openings are utter nonsense with no science to support them

I'll answer – and close this post – with another picture that’s worth at least 1000 words:


  1. In case you missed it, the New York Review of Books reviews Timothy Snyderìs Our Malady: Lessons in Liberty from a Hospital Diary. While I'm at it, compliments for your wonderful book Dottoressa.

    1. Thanks for recommending it. I actually read the excellent excerpt from the book just today in the NYRB daily. I'm always particularly sad to see examples of dire medical incompetence in countries like the US and Germany. (I happen to have dealt with one just last week, involving a relative in the US.) I can also, on a slightly different topic, recommend the beautiful, touching piece by Roger Cohen in today's NY Times, about his own ongoing bout of COVID-19. ...and thanks for the kind words about my book!