Coronavirus March 2020—Part 4 The Truth about Hydroxychloroquine

At ImagineMD, we’ve started to receive a flurry of queries about a recently published study suggesting that an anti-malarial drug, hydroxychloroquine, used in combination with a commonly-prescribed antibiotic, azithromycin, can treat COVID-19. We’re writing this piece not only to answer questions about that study, but also to help people understand how to digest all the information—and misinformation—they’re encountering as the COVID-19 pandemic unfolds here in the U.S.

The Study

For those not interested in understanding the design of the study, please feel free to skip down below to our conclusions. For those who want to understand how we arrived at our conclusions, read on:

The study was carried out in France. It didn’t have a control arm other than untreated patients at other centers and cases at the treatment center who refused the drug intervention. This means that there was no randomization scheme used to determine who received the drug intervention and who didn’t, which significantly reduces the reliability of the results. The sample size was also extremely small (36 patients), drastically limiting the validity and generalizability of the results. Most importantly, the end point of the study was not clinical improvement or reduced mortality. It was the presence or absence of virus in the mouth and nose on day #6 of the study. Unfortunately, this end point doesn’t tell us what we really want to know, namely, whether the medications affect the severity of illness, time to recovery, or rates of death for COVID-19.

Patients with evidence of SARS-CoV-2 in their mouths or noses (as diagnosed by the presence of viral RNA and actual virus in cultures) were given hydroxychloroquine 200 mg three times daily. Twenty-six patients were initially started on hydroxychloroquine, but six of them stopped the drug before the end of the study. Importantly, three of these patients did so because they were transferred to the ICU (one of whom later died and who, incidentally, had cleared the virus from his/her mouth before passing away) and three of whom did so for various other reasons including intolerable side effects. Six of the remaining twenty patients who got hydroxychloroquine also received azithromycin (in the standard Z-pack dose and schedule).

The results showed that at the end of the trial, on day #6 of drug therapy, 70 percent of hydroxychloroquine-treated patients no longer had viral replication in their mouths and noses compared to only 12.5 percent of controls. One hundred percent of patients who were treated with both hydroxychloroquine and azithromycin no longer had viral replication in their mouths and noses by day #6 of the trial.

However, while earlier reduction of viral replication in the mouths and noses of SARS-CoV-2-infected patients given the drugs may have implications for reducing spread of the disease, we can draw no conclusions about the effects of these drugs on the course of the disease and clinical outcomes, including the risk of dying from the disease.

BOTTOM LINE: It’s not yet time to start taking hydroxychloroquine and azithromycin for COVID-19.

the thinking

We’re all desperately hoping for an effective treatment or even a cure for COVID-19. This has biased us all in favor of accepting of any treatment proffered by any source that seems to have any credibility at all, especially when it’s reinforced by massive social proof (meaning, our friends, family, and even government officials touting it as a good idea and promoting it on social media and in the news). But if we allow our hopes and fears to guide our decision-making rather than our reason employed in the service of a skeptical examination of the evidence, we might put ourselves at even greater risk than we are from COVID-19.

We cannot caution strongly enough against the idea of using unproven treatments for COVID-19 (or any disease for that matter). It’s only natural to feel like doing something is better than doing nothing when the stakes are high. But the history of medicine is littered with promising treatments that, based on their mechanism of action, should have worked but that not only didn’t work but that in fact caused harm (for example, hormone replacement therapy given to post-menopausal women as a way to prevent heart attacks). Until a randomized, blinded, placebo-controlled trial is carried out on a proposed treatment that measures endpoints that actually matter (that is, clinical outcome and mortality), we simply won’t know if a treatment is worth trying. (This argument may not apply, it could be argued, to the compassionate use of treatments with no clear proof that we nevertheless think might work—but compassionate use is use, by definition, in the sickest of the sick who stand a good chance of dying anyway.)

Studies are the only way to know not only if a treatment works, but also who it works for. Hydroxychloroquine and azithromycin very well might work, but if they do, maybe they only work in the sickest of the sick. Or the opposite, only for people with mild disease. And maybe they don’t work to prevent COVID-19 at all. But, if based on premature, inaccurate reports of efficacy, huge numbers of people begin acquiring these medicines and hoarding them, the people whom they might ultimately be found to benefit in well-designed clinical trials might not be able to get them. And what if people who have other infections for which these medicines are already known to be effective can’t get them either?

There’s also the question of dosing. What if 200 mg three times daily of hydroxychloroquine does nothing but 400 mg three times daily reduces mortality? What if starting the drugs early in the course of the disease reduces mortality but starting it late does nothing? Even if the drugs do work, without knowing when and with whom to use them, or how much to use, people might not get the full benefit—or any benefit at all.

Another trial looking at the HIV drugs lopinavir-ritonavir was better designed and included the right end points, time to clinical improvement and rates of death as measured between patients treated with the drugs and patients not treated with the drugs. Sadly, the drugs were found to be ineffective at improving either of these end points. There was a difference in both end points for patients given the drugs early in the course of COVID-19 infections (within 12 days of the start of symptoms), but importantly it wasn’t a statistically significant difference, which means the difference that was observed was more likely due to chance than to the drugs actually working. The authors called for a larger study to make sure chance was indeed the reason.

This, unfortunately, is how science is done: slowly and ploddingly. But it’s how we have to do it. People cannot escape their biases, and biases confound results. Unless we follow well-established protocols that eliminate the effects of those biases, we won’t be able to count on the results. We won’t be able to make sure our interventions are helping rather than hurting people.

We all want answers. We all want effective interventions. Before we commit our resources to a given intervention, let’s make sure that intervention really does work.


We’ve been receiving many calls about how to handle practical matters also, like ordering in food, shopping, and doing other necessary life errands. Here are some practical tips:

  1. Take a disinfectant wipe with you each time you go out. This will remind you not to touch your face and will enable you to avoid touching surfaces, like door handles, that might contain virus.
  2. When you get home, before touching anything in your house, wash your hands. Then use disinfectant wipes to wipe down your keys, phone, phone case, and door handles. Periodically disinfect your computer keyboard, faucet handles, light switches, and countertops.
  3. Ordering take-out food is probably safe, but it helps to confirm with the restaurant that their food handlers are being especially careful. When bringing home take-out food, wipe down all containers. Consider avoiding foods that aren’t cooked (like salads and uncooked vegetables). Wipe down all grocery items you’ve bought from the grocery store.

We know people are scared. Just remember that we, your physicians, are here to help educate and guide you through any healthcare issues you might face or decisions you might need to make.

  1. For previous posts related to COVID-19, see:
    1. Coronavirus February 2020—Part 1 What We Know So Far
    2. Coronavirus March 2020—Part 2 Measures to Protect Yourself
    3. Supporting Employee Health During the Coronavirus Pandemic
    4. Coronavirus March 2020—Part 3 Symptoms and Risks

[jetpack_subscription_form title=” subscribe_text=’Sign up to get notified when a new blog post has been published.’ subscribe_button=’Sign Me Up’ show_subscribers_total=’0′]

Leave A Comment

Your email address will not be published. Required fields are marked *

  • Great summary. One additional concern is that partial treatment may lead to the development of a mutated virus which would complicate things even further.

  • Thank you for this, I love the practical advice! With other respiratory illnesses, I typically gargle with salt water once a day. If I were to contract COVID-19, are there any contraindications to me doing this, as I usually would?

    Linda: No contraindications, but no evidence that it will help either.