Reinforcing the last post on Hydroxychloroquine and other putative treatments for COVID-19

I just (31 March 2020) got this bulletin from the College of Physicians and Surgeons of Alberta, jointly with the Alberta College of Pharmacists.

As you will note, there’s been a rush on these meds, and it’s threatening the health of those who need them for proven, legitimate reasons.

Have the French cured COVID-19? Not so fast

There has recently been suggestion that a combination of two medications (azithromycin, a macrolide antibiotic) and hydroxychloroquine (a drug which targets malaria parasites, and which is also used in rheumatic disease like rheumatoid arthritis and lupus) may help COVID-19.

There was a recent French study that’s getting a lot of enthusiasm in some quarters. (President Trump was pushing the idea that these were helpful several days ago, before the new study. One man took a form of chloroquine used to disinfect fish tanks, and it killed him, while also putting his wife into critical condition in the ICU.)

It will be great if these help, but I would suggest caution and modesty in interpreting the results. Why? Several reasons.

Not a randomized trial

The “study” was a case series–it was several patients treated with the meds. But, there was no control group–there was no placebo control. And there was no blinding–the doctors knew who was getting the treatment (i.e., everybody.)

We have learned through sad experience that our clinical experience can mislead us. (There’s a joke I tell med students: “What’s the definition of clinical experience?” Answer: It’s making the same mistake over and over with increasing levels of confidence.)

Every patient treated also had water during the treatment period. But we don’t turn around and assume that water made them better.

You need randomized-controled trials (RCT) to know if treatments work. Physicians bled patients for over two thousand years, and they (and their patients!) were convinced it helped. They had mountains of clinical experience. But they were, in the main, very wrong. (There are a few things that bleeding might help temporarily, but they’re few and far between.)

Not yet published or peer reviewed

The findings have not been published yet, and peer review has not vetted them. Peer review is important for weeding out bad work, and spotting potential errors. It probably will be published, but it will be interesting to see how much peer review makes them dial back their enthusiasm to what can be justified by the science (i.e., not much).

The patients tested were not very sick

One article on the trial reported this:

Dr Philippe Gautret, who was part of the team behind Raoult’s latest findings, admitted that they only used the combination of drugs on “patients who had not been showing signs of being seriously ill after admission” to the hospital.

So…color me unimpressed. These people weren’t terribly ill. They weren’t the ones on respirators likely to die. If you keep the sickest people out of the trial, of course it’s going to look like the medication helps.

Better studies have been negative so far

A small randomized-controlled trial on hydroxychloroquine showed no significant difference in treatment or placebo.

It was a small study and wasn’t powered to detect all effects–so there might still be benefit. But thus far, it is not a “game changer” or overwhelming.


One week after hospitalization, 86.7% of patients in the experimental group and 93.3% of patients in the usual care group tested negative. This difference was not statistically significant

This isn’t a statistically significant difference, but more people tested negative for virus in the non-treatment group than the treatment group. People treated also took four days to clear virus on average; the non-treatment group took only two days. Again, not statistically significant, but not suggesting even a “trend” toward significance.

But, don’t worry–the WHO is doing actual randomized controlled trials (though without placebo and blinding) world-wide on several treatments. That’s the best way to get good data in a hurry.

We’ve tried it in other viruses

Part of the reason hydroxychloroquine is an idea is that it causes problems for the virus in vitro. (That’s Latin for “in glass”–it means in a test tube or petri dish.) We want it to work in vivo–“in life,” or in a living body.

However, it has consistently failed to have the same effect when it’s put into actual, you know, humans. We don’t care about killing viruses in a test tube; we want to kill it in human bodies:

Studies in cell culture have suggested chloroquine can cripple the virus, but the doses needed are usually high and could cause severe toxicity. “Researchers have tried this drug on virus after virus, and it never works out in humans,” says Susanne Herold, an expert on pulmonary infections at the University of Giessen.

Think of it this way–if you pour straight bleach into a test tube of COVID-19 virus, I guarantee you that it will kill all the virus. However, drinking bleach to cure the virus is not going to kill the virus in you, and would probably have several other nasty toxicity-related side effects as well.

(Don’t worry, CBS News has your back with a great article called “Coronovirus cannot be cured by drinking bleach or snorting cocaine.”)

That this warning is necessary makes you think that perhaps our species is too dumb to survive. DON’T DRINK BLEACH. It essentially melts through your throat and esophagus. It’s a good way to die very, very painfully over several days.

“The clinical notes of the young woman commenced with the memorable words ‘Overdose of bleach’, raising the interesting medical question as to the correct therapeutic dose of bleach.”

— Theodore Dalymple, If Symptoms Persist (Monday Books, 2010).

But, at any rate, test tube does not equal real life.

The research folks aren’t impressed

From the above article:

Professor Francois Balloux of University College, London, tried to dampen talk that the drug could be a silver bullet.

“No, (this is) not ‘huge’ I’m afraid,” he said on Twitter.

“This is an observational study (i.e. not controlled) following 80 patients with fairly mild symptoms. The majority of patients recover form #COVID19 infection, with or without #Hchloroquine and #Azithromycin treatment.”

Treatments have side effects

Very few treatments in medicine have no risk of side effects. Anything that can cause good effects can probably cause bad effects.

Hydroxychloroquine is an immune suppressant — that’s why we use it in rheumatic disease. So it could be that suppressing the immune system is a bad idea in an infectious disease. Or, it could be a good thing if part of the damage is from an over-reaction by the immune system (e.g., cytokine storm).

Hydroxychloroquine also has significant potential heart toxicity. We have seen cases in COVID-19 (even without hydroxychloroquine on board) of relatively young people starting to improve, only to suddenly crash due to an inflammation of heart muscle. I’ve read anecdotal reports from ICU physicians who describe young patients who are sometimes left with a very poorly-functioning heart (ejection fraction < 10%; normal is 50-60%) afterwards.

A history of heart disease is also correlated with poor outcomes.

So introducing a cardio-toxic drug is not without its potential downsides.

Ignore people telling you to get this

There are restrictions in Alberta on prescribing these medications. If you think you have COVID-19, your family doc or whoever is not able to prescribe this combination (nor should they).

There is a limited supply of hydroxychloroquine (Plaquenil is the trade name). It is an important drug in the treatment of lupus and rheumatoid arthritis. A “run” on this drug will threaten the health of those patients.

The only way you’re going to get this experimental treatment is in the context of a clinical trial, or perhaps if you’re hospitalized. So don’t run out and try to stock up. Here’s the Alberta College of Pharmacists:

This week, pharmacists have identified to ACP an increased demand for some drugs (e.g., Kaletra®, hydroxychloroquine) due to reports of them being prescribed as treatments for COVID-19. The prescribing and dispensing of drugs used to treat COVID-19 for the purpose of stockpiling for personal use is not appropriate.

Information we have received demonstrates the diligence of many pharmacists in assessing the appropriateness of drug therapy, and we commend them for this. We have heard stories of pharmacists receiving prescriptions for these drugs for groups of family members, for personal family members, and in other instances from specialties where these drugs are not normally used. Thank you to all of you who have been diligent in your assessments and have intervened by not dispensing these prescriptions.

ACP recognizes and appreciates the extraordinary efforts of pharmacists and pharmacy technicians during these unprecedented and challenging times. Your diligence will support appropriate use, improved health, and the continued availability of these drugs for those who need them most.

Are these docs heroes?

Not really. They’re doing bad science, and they’re sensationalizing things. As one expert noted:

Statistician Tim Morris of the university’s clinical trials unit was even more scathing.

“If hydroxychloroquine turns out to be useful,” he tweeted, “it’s a shame that this group will be praised as heroes and prophets instead of held to account for the misinformation and self-promotion they’ve been churning out at a critical time.”

The rules of science and medicine don’t just disappear in a crisis. If anything, we have to adhere to them more strictly, since in emotional and stressful times, we’re more likely to go with our gut rather than our brain. And our gut has proven to be a poor predictor of what works.

As Nobel Prize-winning physicist Richard Feynmann wisely observed:

The first principle is that you must not fool yourself—and you are the easiest person to fool.  So you have to be very careful about that.  After you’ve not fooled yourself, it’s easy not to fool other scientists.  You just have to be honest in a conventional way after that….I’m talking about a specific, extra type of integrity that is not lying, but bending over backwards to show how you’re maybe wrong, that you ought to do when acting as a scientist.  And this is our responsibility as scientists, certainly to other scientists, and I think to laymen.[1]

If that’s true of scientists, it’s doubly true of medical scientists in a pandemic.

[1] Richard Feynman, “Cargo Cult Science,” Caltech Commencement Address, 1974,

No, COVID-19 is not a bioweapon

A common conspiracy theory (or worry) is that COVID-19 has been somehow human engineered as a bioweapon or the like.

The theory varies as to who did this (e.g., China, the US, someone else) and whether the release of the virus was accidental or intentional.

If it was accidental, that doesn’t make much sense. As we’ll see below, to create this virus and hide your tracks would be a huge undertaking–probably Nobel-prize level stuff. So it would be strange for people capable of engineering the virus as elegantly to be so clumsy as to then release it by accident into the wild.

If it was intentional, these are the most incompetent bioterrorists ever. (Which is again strange given the sophistication of what they’d have had to do with COVID-19.)

Theoretical reasons to doubt conspiracy

In the first place, human-to-human diseases make lousy bioweapons, because you can’t control them. There’s too much risk they’ll rebound and get your group. (The exception would be if you had a vaccine for them. But there is no COVID-19 vaccine, unless the bioterrorists have also concocted that. But that would be a huge undertaking as well.)

Better bioweapons are things like anthrax, which can spread to humans but you don’t usually get human-to-human spread.

In the second place, the deployment and pattern is all wrong for a bioweapon. If you want to attack China, don’t start in Wuhan province. You’d want a simultaneous explosion of cases everywhere. So, why not start in (to pick three examples): Beijing, Shanghai, and Hong Kong all together?

If you’re targeting the world at large or the United States, again Wuhan province makes little sense. Just spread it in half-a-dozen airports around the world, and watch the chaos. (Say, Heathrow, Beijing, Tokyo, JFK, LA-X, and Charles de Gaulle.)

Third, it’s a fairly crappy virus for this. Death rates may be high in the peak as resources are overwhelmed, but the vast majority of people will do fine. So it will produce some chaos and some suffering, but nothing fundamentally destabilizing in the long-term.

Most importantly

So, those are all theoretical reasons. The best reason, however, comes from the fact that we’ve sequenced the virus. We know its entire RNA code.

If someone had engineered the virus, it would leave tell-tale signs by the techniques used to tweak the genome. As a recent article in Nature put it:

It is improbable that SARS-CoV-2 emerged through laboratory manipulation of a related SARS-CoV-like coronavirus. As noted above, the RBD of SARS-CoV-2 is optimized for binding to human ACE2 with an efficient solution different from those previously predicted.

Furthermore, if genetic manipulation had been performed, one of the several reverse-genetic systems available for betacoronaviruses would probably have been used. However, the genetic data irrefutably show that SARS-CoV-2 is not derived from any previously used virus backbone

So, the part that makes it bind to humans is a completely novel solution–no point in reinventing the wheel if you’re making a bioweapon. But, Mother Nature is smarter and more creative than you are. Evolution in action.

And, this doesn’t show any link to “any previously used virus backbone.” So it would have to be an utterly novel technique to create this. Way more work than it would need to be.

So where did it come from?

COVID-19 is almost certainly a zoonotic infection–that’s an infection that jumps the species barrier to us. Most of history’s huge pandemics have been that kind of infection. For example:

  • influenza (birds and pigs)
  • measles (probably from rinderpest, a cattle disease)
  • TB (cattle)
  • HIV (monkeys)
  • smallpox (cattle)
  • bubonic plague (fleas on brown rats)

These cause pandemics because no human has immunity when the jump is made.

Such infections usually weaken (attenuate) over time as there is selection for strains that are less severe.

(It is easier to spread a disease that makes you mildly or not at all sick for an extended period while you spread it. Disease strains that kill fast and hard tend to be selected against because they aren’t able to spread as well. This is, for example, Ebola’s one saving grace–you get really sick really fast, and so you don’t have time to spread it. Heaven help us if we ever get an Ebola strain with a latency period of weeks to months before symptoms.

Even in recent memory, we’ve seen this. HIV used to be far more deadly early on in the 1980s than it is now. That’s because it has attenuated via this kind of selection. Even if we had no HIV medications, people would still live longer before getting sick and dying.)


Pandemics are scary. And it is perhaps more scary that they can simply arise out of no where, with no one causing them or engineering them. But that’s been the pattern throughout human history. If anything, we’ve enjoyed a “holiday from history” on that front.

COVID-19 should remind us that the planet is interconnected; health problems in one area become health problems everywhere.

Dishonest or corrupt governments (I’m looking at you, Communist Party of China) pose a risk to their own citizens and everyone else.

And, one must be prepared for such things before they start. Hopefully, we’ll learn some of those lessons. But, like generals, politicians always seem to prep for the last war. And, in the west they have great difficulty looking beyond a horizon of one election (so 2-5 years in North America). That’s a lousy timeframe for medical planning in general, and pandemics in particular.

And honestly, that worries me far more than COVID-19.

Pandemics and math–your worst nightmares, combined

As people struggle with self-isolation, it’s worth reviewing why we’re doing all this.

Humans aren’t very good at assessing mathematical risk. Our brains aren’t built for it. (See, you were right in high school!)

It takes disciplined effort and training to think statistically or mathematically. These things usually do not match our intuitions.


As we discussed in a previous post, the R0 of COVID-19 is probably between 2-2.5. We’ll use 2.3 for ease.

What this means is that each person infected with COVID-19 will, on average, spread it to 2.3 other people during the course of the illness.

All this presumes that no one in the population is vaccinated or otherwise immune.

Well, we have no vaccine. And no one has seen this virus before, so no one is immune. (Over months, most of us will be exposed and become immune–then the same R0 won’t apply, and the epidemic will burn itself out.)

COVID-19 versus influenza

It is common to compare COVID-19 to influenza, and some even complain that we don’t do this kind of hard-core quarantine for influenza.

We do it for a few reasons:

  1. the entire population is not usually vulnerable to seasonal influenza; there is some immunity;
  2. the death rate of COVID-19 seems to be much higher than seasonal influenza (probably about 10x higher);
  3. The R0 for influenza is 1.3.

Is this really a big deal? Is the difference between 1.3 (influenza) and 2.3 (COVID-19) that significant?


Brace yourselves, math coming

Influenza example

For an R-nought of 1.3, let’s assume I’m the only person with the disease. How many people will I infect? R0 doesn’t tell us at what rate the disease spreads; it only tells you that over the course of my illness, I’ll give it to 1.3 people on average. (You can’t infect 0.3 of a person obviously; some will infect a few more, some a few less.)

Let’s follow this chain out ten places (me infecting 1.3 people, and each of them infecting 1.3 people, and so on, for 10 “layers”).

That is:

# infected = 1.3 x 1.3 x 1.3 x 1.3 x 1.3 x 1.3 x 1.3 x 1.3 x 1.3 x 1.3

A short hand way to say this is with (shudder!) exponents. Specifically:

Or, “1.3 to the 10th power.” = 13.79, round up to say 14 people.

So I am responsible for 14 infections of influenza if I don’t self-isolate.

COVID-19 example

Now, do the same thing for COVID-19, with an R-nought of 2.3.

Again, let’s assume I’m the only person with the disease. How many people will I infect, again over a chain of 10? (Me and my infections, plus all the infections they cause, 10 times total.

So, that’s 2.3 multiplied by itself ten times, or:

Or, “2.3 to the 10th power” = 4143 (rounded to nearest patient).

So, with the slightly higher COVID-19 R-nought, I am responsible for 296 times more infections (29,600%).

If you assume the higher R-nought is the right number (2.5) then that jumps to 9537 people–681 times more infections (68,100%).

Small shifts in R-nought make a huge difference.

Using graphs

Some people like graphs more. Here’s those two values, plotted on a graph.

The red line is for R-nought=1.3. Notice that at 10 “cycles” out on the horizontal axis, it hits 14.

The purple line is for R-nought = 2.3. Notice that within 4 cycles, you’ve already infected more people than 10 cycles of 1.3. And, it just explodes upward–that purple line won’t reach 10 until it is up beyond 4000 on the vertical axis.

Now this is an over-simplification. There are lots of other factors. And, as more people are infected, this becomes less the case.

But it gives you a back-of-the-envelope sense of the type of problem we’re facing.

Why this matters

So, this explains why we’re self-isolating. It’s to try to keep that R-nought value down. Self-isolating reduces the number of people you will infect (hopefully to zero, but certainly lower).

If we don’t control R-nought, then all these people will explode with illness at once. And there’s no way we can cope with even the small fraction who get seriously ill. We have to slow it down so R-noughts are more like influenza (which will strain us, but which we can and do handle every year).

This also explains why “cheating” on your social isolation is such a bad idea. (“I’m just going to see my friends for a minute; we’re just going to run over to grandma and grandpa’s house once.”)

No. Just, no.

All you need to do is infect one more person than you otherwise would, and you’ve completely negated any benefit from your isolation. If you infect 2 people instead of 1 person, or 3 instead of 2, you’ve increased the disease massively if everyone else is behaving as you are.

(Don’t use this as an excuse to chuck your isolation if you’ve already cheated–remember, these numbers are averages, and so some people will do more than 2.3. Don’t be that person! Don’t make previous mistakes worse.)

But, these facts should make an overwhelming problem seem manageable. All you have to do is avoid infecting someone. Even if you avoid only one infection, that has a massive add-on effect, and you’ve made a real contribution. By contrast, if you don’t avoid one infection, you’ve caused massive amounts of damage.

Some infections will have bigger impacts

Plus, if you infect someone in a critical position–like a doctor or nurse–then your impact becomes even greater because that person doesn’t just become another sick case and potential hospital user. It also takes them out of the caring for others.

(This is why I decided I was keeping my kids home from school even before the province announced it. I didn’t want them to spread it, but more importantly I knew that if they spread it to me, that would have a higher “cost” to others.)


So, don’t cheat. Don’t make exceptions. Ignore your kids when they complain. Don’t think you’re special. You’re not.

Math plays no favourites.

A question about contagiousness

Someone asked me a really good question recently. To paraphrase:

The R0 (R-nought) is the average number of patients to whom a given infected person spreads the illness.

(If R0 = 2, then each infected person on average infects 2 other people over the course of their illness.)

The R0 of COVID-19 is between 2-2.5.[1] (Note, this number is in some flux, and could be higher or lower.)

The R0 of polio was worse (R0 = 5-7), as was SARS (R0 = 2-5).[2]

So why didn’t we all self-isolate when SARS was a thing? Why didn’t everyone self-isolate for polio back in the 1950s?

This is a great question because it reveals an understanding of what’s going on, and it is also a great question because the answer reveals a few non-obvious things.


In the case of SARS, we did slam down the self-isolation protocols. This happened quickly and aggressively. This allowed public health to isolate any carriers and any contacts. Because it was done early and thoroughly, a major outbreak was prevented. This is what ideally should have happened if the Chinese Communist government hadn’t destroyed data, lied, jailed those who tried to raise the alarm, etc.

[Incidentally, I wouldn’t trust any data coming out of China on their current infections or lack thereof. I simply don’t think they can be trusted, and no one you can trust is there watching and reporting back.]


The case of polio is more interesting.

Polio is usually spread by what is called the “fecal-oral” route–i.e., the virus is shed in your stool, and you somehow get stool in your mouth. Infected and contaminated water supplies would be the most common way for this to happen.

Even though it has been identified for thousands of years, the major outbreaks of polio are relatively recent. The major outbreaks happened in the late 19th and early 20th century. Why was this?

Polio has the interesting phenomenon of being less likely to cause paralytic polio (which occurs in about 1% of those infected) when you catch it at a younger age–below 6 months old.

So, the older you are when you first encounter polio, the more likely you are to get seriously ill.

Quarantine card from the polio epidemics. Violation of the quarantine order carried a fine of $100 (over $2000 in modern money given inflation). From: Wikipedia.

View Post

What was happening in the late 19th and early 20th century? For the first time in history, large populations were given access to relatively reliable, fairly clean water supplies. Fecal-oral spread of polio virus from birth was a much less likely possibility. (It was probably virtually inevitable before that, at least to anyone who lived in a city. And, for rural people, they probably had relatively limited contact without outsiders, so they were unlikely to be exposed either.)

Remember that polio is spread by fecal-oral route. So, for most of history, you had a very high chance of encountering polio virus when you were very young, and so less likely to get ill. Once exposed, you were immune.

This is a good example of the law of unintended consequences. Who would have thought that there would be a downside to clean water supplies. But there was–an explosion in symptomatic polio in kids 5 and up.

So, during the polio epidemics we did isolate people dramatically–we isolated children at risk. Most adults would have already been exposed as infants/children, and so were immune. It was only the kids that were at serious risks. Polio peaks in summer months, and those who grew up in the 1950s probably have vivid memories of being kept indoors all summer, swimming pools being closed, etc.

Happily, the Salk vaccine of 1955 made all this unnecessary (unless the anti-vaxxers manage to bring it back before we achieve world-wide eradication).


So, to answer the questions the short answers are:

  1. We did isolate severely with SARS–we just caught it early, and so didn’t require population-wide lockdowns. It was controlled at a more local level.
  2. With polio, we did isolate the vulnerable. But, unlike with COVID-19, the whole population was never vulnerable to polio before the vaccine. Most adults were probably exposed and immune (even if they didn’t realize it). So locking down adults would have served no purpose–it was only kids who were at risk.



Want to help with COVID-19? We need blood!

Blood products have a limited shelf life. Canadian Blood Services really needs more donations.

We’ve already received new protocols that control our use of things like “O negative” blood (the universal donor) given the new realities.

Look at it this way–the need for blood products is going to be the same as it always is, since other diseases aren’t taking a holiday (though maybe we’ll get less trauma if people are out and about on highways less).

Add to that all the needs of COVID-19. There are a few restrictions, of course:

  • Anyone asked by public health to place themselves under observation or self-quarantine are not allowed to donate for 14 days from the date of their last contact with a case or suspected case of COVID-19.
  • Anyone with a case or suspected case in their household cannot donate for 14 days after the infected person’s recovery.
  • Anyone with a confirmed case of COVID-19 are not allowed to donate for 56 days after full recovery from the infection.

See this link to learn more and help out.

Thoughts on COVID-19 testing

As Alberta residents now know, the only place to access COVID-19 testing is via 811 and then assessment centers or other designated public health locations.

People are finding this frustrating, because the waits are long, and even when they get through they may not meet testing criteria.

I think there’s some confusion about why exactly we test.

There’s an old joke in medicine: “Don’t order a test if it won’t change what you do.”

For most people, that’s the situation with COVID-19. If you are positive and you have symptoms what are we going to have you do? Go home and self-isolate for 14 days.

If you have symptoms and you aren’t tested, what are we going to do? Send you home to isolate for 14 days.

If you’ve travelled since 12 March 2020, you’re to go home and isolate for 14 days, symptoms or no symptoms, positive or negative swab or not.

So in most cases, the swab status doesn’t change what YOU do (or what we do for you) at all.

So why swab at all?

INITIALLY, swabbing at risk people was important because it’s a public health tool. We have to track what’s out in the community and what’s happening. If we find a case, we want to pounce on it, isolate all their contacts, and try to prevent the outbreak. (This strategy worked very well for a coronavirus related to COVID-19 — SARS. Problem is that the containment strategy didn’t work. This honestly isn’t surprising. It’s a tall order, esp in a free society. The time to stop it was China, but the Chinese Communist government made a number of bad choices that made that impossible until it was too late.)

When you’re trying to contain, swabbing is important, because you want a bead on what the rate of cases is out in the community. Why? Because public health and government need to know when to pull the trigger on very disruptive and costly isolation measures–just like we’re in now.

Once you’re to that phase, testing really has much less helpfulness. We’ve already locked down. We should be social distancing already. Knowing there are more cases out there won’t change that. (More data is always nice, but you get diminishing returns in terms of what you do.)

At present, medically, the time when being COVID-19 positive will really matter is if you’re sick enough to be in hospital. That will affect what treatment you might get, and where you might be located.

But, if you’re not sick enough to be in hospital, the swab won’t change much.

So, if 811 tells you not to test–don’t stress. Shelter at home, self-isolate, just like everyone else is. If you get sick enough to need testing, you’ll get it. You don’t need to race around trying to find someone to swab you.

In an ideal world, we’d swab everyone whenever. But that costs money, and scarce supplies, and scarce testing time and personnel. In most cases, those aren’t worth the benefit.