Why Hydroxychloroquine didn’t work

I got a little bit of pushback for my last post saying that Hydroxychloroquine doesn’t work.

There was a lot of initial hope behind the drug. We’ve known since 2005 that it inhibits the spread of the original SARS virus from cell to cell, so it seemed like a natural choice for use against SARS-CoV2. Chinese scientists were testing it as early as January 2020.

Over time, bigger studies were conducted that showed little benefit.

Last time, I only cited 2 studies to make my point. I could point to more. I cited this study from Oxford, showing that hydroxychloroquine did not help hospitalized patients.

Some people asked if we needed to use the drug in combination with azithromycin. Here’s a study where that combination of drugs did not help hospitalized patients.

Some people asked if we need to give the drug earlier, before patients ended up in the hospital. Here’s a study from Spain, where hydroxychloroquine was used for early treatment. There was little to no benefit. Viral load didn’t change. 6% of treated patients ended up in the hospital, compared to 7% of untreated. That difference was not statistically significant.

Those patients were given hydroxychloroquine within 5 days of the start of covid symptoms. Maybe that’s not early enough? Here’s a study done in the US and Canada, the majority of patients were enrolled within 1 day of symptoms starting. The study found a 12% improvement on covid symptoms when patients were given hydroxychloroquine vs. placebo. That result was not statistically significant.

We’ve tried giving it even earlier, before symptoms start. Three large studies tried giving hydroxychloroquine to people who’d recently been exposed to covid, to see if it prevented them from getting sick. This failed to prevent covid, in one trial in the US. A similar study in Spain also failed. A third study from the US and Canada failed.

I’m not sure if anyone has tested hydroxychloroquine in combination with azithromycin to prevent covid. This combination of drugs was tested in monkeys, it did not prevent them from getting sick.

It’s a difficult topic to debate because the drug’s supporters don’t give up. If you prove that the drug doesn’t work for late treatment, they’ll say it works for early treatment. If you prove it doesn’t work for prevention, they’ll say it’s not for prevention, only for treatment. They’ll say the dose was too low, or too high. They’ll say that it only works if you add azithromycin. Once you’ve disproved all those theories, they’ll say it only works if you add zinc.

That’s been studied, as well. Chloroquine doesn’t draw zinc into cells, it’s not a “zinc ionophore”. Hydroxychloroquine combined with zinc didn’t help hospitalized patients.

I have not found a study that combines hydroxychloroquine, azithromycin, and zinc all at the same time, and only looks at patients given early treatment. Supporters could fall back on this idea, if they wanted some idea to cling to.

One blogger brilliantly called Hydroxychloroquine the “Black Knight of treatments for covid-19”, referring to the famous Monty Python scene. No matter how many times you cut off a limb, he won’t give up.

There are hundreds of studies that have been done for treatment. Some had a mildly positive effect, some had a mildly negative effect. If you just naively average every study together, there is a mildly positive effect. So, why do I think it doesn’t work?

There’s a strong and simple theory for why the drug fails. Covid-19 can enter cells via 2 different pathways. Hydroxychloroquine only blocks one of them. In lung cells, it doesn’t block the more important pathway. Using the drug is like trying to stop theft by locking your house’s back door but not the front.

We can see both mechanisms of entry in a paper called Hydroxychloroquine-mediated inhibition of SARS-CoV-2 entry is attenuated by TMPRSS2:

SARS-CoV-2 can enter the cells via cell pockets called endosomes. Inside these pockets, it enters the cell with the help of an enzyme called Cathepsin L. Cathepsin L needs an acidic environment to work. Hydroxychloroquine makes the endosomes less acidic, so it blocks this pathway.

Another way covid can enter cells is by binding to the ACE2 receptor on the cell’s surface. Before this happens, the cell cuts covid’s spike protein in half with a protease called TMPRSS2.

A new article in Nature gives some great diagrams of this pathway:

After the spike protein is cut, the virus can attach to the cell wall:

After that, the attached proteins trigger the virus to fuse with the host cell:

Unfortunately, TMPRSS2+ACE2 entry is the dominant pathway that covid uses to enter lung cells. Some other cells, like kidney cells, might rely more on the endosomal pathway.

So, I suppose there’s a chance that Hydroxychloroquine might protect some of your cells from damage from covid. It’s even possible that some early treatment studies showed a positive effect because of this. But the drug won’t prevent you from catching covid, because the virus still infects cells in your airway. And the drug won’t protect your lungs. Most covid deaths come from pneumonia or acute respiratory distress.

If you’re following along so far you might be wondering, “What happens if we block both pathways for cell entry?”

Several groups have tried this approach, in vitro. Combining a Cathepsin inhibitor with a TMPRSS2 inhibitor works to prevent the virus’ entry into lung cells.

An effective treatment might look like:
Hydroxychloroquine, 200–400 mg, twice per day
Camostat Mesylate 100–200 mg, three times per day

Camostat is a fascinating drug that also shows promise against treating the flu. Right now, the drug is only available in Japan.

Instead of Camostat, you may be able to use Bromhexine, which is an over the counter drug in many countries in Europe and Asia. The dose would be somewhere between 8–32 mg, three times per day.

I’ve only seen one study that tried this on patients . A group in Iran took 80 patients, the treatment group got Bromhexine as a TMPRSS2 inhibitor along with Hydroxychloroquine. The control group got only Hydroxychloroquine.

No one died in the treatment group. 5 people died in the control group. The treated patients also fared better with most symptoms:

The treatment groups aren’t perfectly matched, the gender ratios aren’t the same, but the randomization isn’t that bad:

These are suspiciously strong results. With only one small trial, we could be seeing random noise, publication bias, data errors, or even fraud.

I might be making the same mistake as everyone else saying “rerun the trials with zinc this time”. But there is a good theory for why this treatment might work. There is one experiment in favor. It seems worth repeating the experiment on more people.

This study is from July 2020. Well written articles also described the cell entry process in July 2020.

How did people respond? Republicans kept saying that Hydroxychloroquine works, Democrats kept calling them idiots. And no bothered replicating the study on blocking both pathways. I think there’s one attempt in progress that still hasn’t finished.

I said it last time, but I’ll say it again: the culture war is killing people.

Okay, if you’re following along, you might also ask: “Since TMPRSS2/ACE2 is the dominant pathway for infecting lung cells, what happens if we just use a TMPRSS2 inhibitor by itself?”

I’ve only seen a few studies. An early treatment study of Camostat says that the drug worked. It cut the duration of covid symptoms in half, from 9 days down to 4. Only 1 treated patient ended up on oxygen, compared to 6 in the placebo group. Those are great results. But the study doesn’t look entirely convincing to me. Instead of looking at all 300 patients, they’re only focusing on 100 that were “medication compliant”. Was there actually a big issue with 2/3rds of the patients not taking their medicine? Or did the researchers divide up the patients somehow to find a signal in the noise?

Another study of Camostat gave good results, but it was observational, not randomized. A randomized trial showed that Camostat failed to help hospitalized patients. A randomized trial of Bromhexine also failed to help hospitalized patients.

So, I’m not sure. There’s some promising data here, but the studies aren’t very large or convincing. It looks like it’s much better to treat patients early, waiting until hospitalization doesn’t work. Maybe these drugs would work better with Hydroxychloroquine. Maybe they don’t work at all. Between those studies, they’ve only been tested on about 900 patients. We tested Hydroxychloroquine on 400,000 patients, many of those after we knew why the drug doesn’t work.

It would be nice to work out the best course of action for early treatment of covid. I’m hoping I won’t need any such treatment, because I’m vaccinated. But some people are still getting sick. We don’t know how long immunity lasts or how quickly covid will mutate. The threat could be gone soon after we hit herd immunity or the virus could continue to infect people, year after year.

There are a variety of over the counter drugs and supplements that might help. If you get covid, it might help to take aspirin, zinc, vitamin C and D, famotidine, melatonin, or quercetin. All are cheap and legal and might help. There’s some weak evidence they do.

Bromhexine is an over the counter drug in Europe and Asia, it’s also cheap and safe. Ambroxol is an equivalent that’s available in other countries. You can sometimes find these on Ebay or Etsy if they’re not available in your country. These drugs might also help for early treatment, but I’m not sure. They might only work in combination with Hydroxychloroquine or other endosomal entry inhibitors, and those are prescription drugs.

There are also a bunch of prescription drugs that might treat covid. Most people have moved on from arguing about Hydroxychloroquine to arguing about Ivermectin. I don’t know if Ivermectin works or not, I’ve seen mixed evidence. But we might also see results from Camostat, or Fluvoxamine, or Metformin, or Calcifediol, or Colchicine, or Cyproheptadine, or androgen blockers like Proxalutamide, or nasal sprays like Iota-Carrageenan or Taffix. I’m not recommending any of these drugs, only saying there’s a lot more options we should be testing.

The conspiracy theorists are right in one sense. We’re under-funding research into generic drugs and treatments. More profitable, patented drugs like Remdesivir can get through the testing process more easily. To fix that, we need more public funding of trials for repurposed generic drugs. But we also need to cast a wider net for options that might work, rather than focus only on Hydroxychloroquine and Ivermectin. It doesn’t help anyone to stubbornly fight for a year over whichever drug is fashionable.