It’s been a while since I’ve written about hydroxychloroquine (HCQ), the anti-malaria drug with immune modulating properties that make it useful for also treating some autoimmune diseases like rheumatoid arthritis, that was repurposed early in the pandemic to treat COVID-19 based on little to no evidence other than theoretical efficacy based on cell culture experiments with SARS-CoV-2, the coronavirus that causes COVID-19 and dodgy reports out of China about its use in Wuhan during the initial outbreak. Although I did mention it a couple of times this year—for instance, in the context of criticizing formerly respected epidemiologist turned COVID-19 crank Harvey Risch—I haven’t done a post primarily about the repurposing of this drug in quite a long time, in particular because (1) we knew as early as July 2020 that HCQ doesn’t work against COVID-19 and (2) it was long ago supplanted by the deworming drug ivermectin as the favorite cure for COVID that “They” don’t want you to know about and are therefore “suppressing.”
So it was with interest that I came across an article late last week with the headline Nearly 17,000 people may have died from hydroxychloroquine: study. Unfortunately, I didn’t come across it until two or three days after the study itself had been floating around social media. Eventually I wondered whether I had missed the opportunity to strike while the iron is hot, so to speak, given that the discussion of the study had already died down. Still, it intrigued me; so even though it’s nearly a week later, I figured, WTF? I also thought it would be a good idea to take a look at a study that seemingly supports what we were warning about nearly four years ago, namely that, contrary to what HCQ advocates were claiming, no drug is harmless, making the promotion of a drug like HCQ for an indication for which there was little evidence for efficacy not a harm-free decision. I also found it fitting that the study came out of France, given that it was a French researcher, Raoult Didier, who was one of originators and foremost promoters of HCQ as a part of a COVID-19 cure. (He also included the antibiotic azithromycin, while Vladimir Zelenko added zinc and other supplements to the mix.)
Before I get into the weeds with the study, I think it worth discussing briefly how HCQ was first repurposed as a potential antiviral treatment for COVID-19, because not only was HCQ initially recommended as a valid potential standard of care in those early chaotic days in the spring of 2020, but it even received an emergency use authorization (EUA) from the FDA based on the flimsiest of evidence.
If you go way, way back into the mists of time in the earliest days of the pandemic during the breakout in Wuhan, you’ll find that this initial recommendation was based on an observation of 80 patients full of confirmation bias. In brief, Chinese doctors in Wuhan noted that none of these patients with lupus erythematosis became ill with COVID-19 and hypothesized that the chloroquine or hydroxychloroquine that they were taking might be the reason. (These drugs are also mildly immunosuppressive, hence their use to treat autoimmune diseases.) Of course, during a pandemic, people who are immunosuppressed are the very people who most rigorously obey orders to practice social distancing and self-quarantine and thereby protect themselves from infection as much as they can. Be that as it may, the Chinese doctors started using the antimalarial drugs, and anecdotal evidence of success was reported, leading to randomized clinical trials that were announced by the Chinese government to have been “promising.” None of this stopped China from incorporating these drugs into its recommended regimen. The World Health Organization followed suit, as did several countries, and thus was born a new de facto standard of care for COVID-19 based on, in essence, no evidence other than some in vitro evidence that the drugs inhibit replication of SARS-CoV-2, the virus that causes COVID-19, anecdotes, and incredibly weak clinical trial evidence. Ultimately, decent randomized clinical trials started, and by late summer 2020 evidence had accumulated that HCQ didn’t work, a state that more recent clinical trials failed to reverse.
Another thing that’s important to note (and that fed the conspiracy theories that “They” were “suppressing HCQ as a COVID cure) is that there was a scandal over a 2020 study that had purported to find that HCQ was killing people because of its side effect of producing dangerous cardiac arrhythmias. I initially wrote about the study done by a company called Surgisphere with some—although not nearly enough—skepticism about its methods and ability to actually carry out such a large study, leading me to castigate myself later for perhaps having been too credulous about a study whose findings fit in with my then-current beliefs about hydroxychloroquine. Basically, Surgisphere was a big scam and couldn’t possibly have done the study—studies, actually, as there were several—that it claimed to have done. Naturally, HCQ cultists immediately pointed to the fake study by Surgisphere as fake and, more importantly to them, as “evidence” that “They” were out to discredit HCQ as a COVID-19 treatment, no matter what lengths “They” had to go to in order to accomplish their nefarious aim. (Unsurprisingly, antivax and COVID-19 conspiracist reactions to this new French study often cite Surgisphere.)
Now, on to the study, Deaths induced by compassionate use of hydroxychloroquine during the first COVID-19 wave: an estimate, published in Biomedicine & Pharmacotherapy.
The TL;DR summary of my conclusion about the study is that I don’t agree with Mollie James, a critical care doctor who went COVID-19 crank and now promotes ivermectin—and do agree with Greg Tucker-Kellogg:
Indeed. As you will see, the extrapolation used in this paper is problematic, but it is nonetheless likely true that rampant HCQ use probably did lead to a substantial number of deaths early in the pandemic. I would also add that the cult of HCQ, in which many later decided not to get vaccinated against COVID-19 because they mistakenly believed that HCQ and ivermectin were safe, inexpensive, and highly effective treatments, likely killed many more who might not have died if they had been vaccinated. That study doesn’t address that, as its authors only studied the time period from March to July 2020—there were no vaccines then and HCQ was also still a standard of care in many hospitals—and only tried to estimate the number of deaths then that could be attributed to HCQ.
So how did they do it? First, they used a systematic review of HCQ use and the mortality of hospitalized COVID-19 patients:
We estimated the worldwide in-hospital mortality attributable to HCQ use by combining the mortality rate, HCQ exposure, number of hospitalised patients, and the increased relative risk of death with HCQ. The mortality rate in hospitalised patients for each country was calculated using pooled prevalence estimated by a meta-analysis of published cohorts. The HCQ exposure was estimated using median and extreme estimates from the same systematic review. The number of hospitalised patients during the first wave was extracted from dedicated databases. The systematic review included 44 cohort studies (Belgium: k = 1, France: k = 2, Italy: k = 12, Spain: k = 6, Turkey: k = 3, USA: k = 20). HCQ prescription rates varied greatly from one country to another (range 16–84%). Overall, using median estimates of HCQ use in each country, we estimated that 16,990 HCQ-related in-hospital deaths (range 6267–19256) occurred in the countries with available data. The median number of HCQ-related deaths in Belgium, Turkey, France, Italy, Spain, and the USA was 240 (range not estimable), 95 (range 92–128), 199 (range not estimable), 1822 (range 1170–2063), 1895 (range 1475–2094) and 12739 (3244− 15570), respectively.
The strategy was summarized with this flowchart:
The authors based their study on this rationale:
HCQ, a 4-aminoquinoline drug used to treat malaria and autoimmune rheumatic diseases, inhibits viral replication by increasing the pH of the endosome used for cell entry. In addition, HCQ may interfere with angiotensin-converting–enzyme 2 glycosylation, which is the cellular receptor of SARS-CoV-2 , . During the first wave of the pandemics, off-label use of HCQ has been proposed as a treatment option for COVID-19. Subsequent studies documented however an unfavourable risk-benefit balance, including the RECOVERY trial that showed a significant increase in cardiac mortality as well as a trend for increased all-cause mortality risk with HCQ . In a meta-analysis of 14 trials testing HCQ in hospitalised patients with various doses, HCQ was associated with an 11% (95%CI 2–20%) increase in all-cause mortality .
The main objective of our study was to estimate the mortality attributed to compassionate use of HCQ in the context of COVID-19 before the publication of reliable randomised controlled trials (RCTs).
That’s where the cutoff of July 17, 2020 comes from, by the way. That’s when the first reliable RCTs showing that HCQ was, at least, ineffective in decreasing the chance of dying for hospitalized patients with COVID-19.
The main outcome studied was excess deaths among patients hospitalized with COVID-19 during that horrific first wave of the pandemic from March to July 2020 that could be attributed to the compassionate use of HCQ. Then they used this method to estimate excess deaths attributable to HCQ:
We calculated the number of deaths in hospitalised patients (Nhospitalised patients) by multiplying the number of hospitalised patients receiving HCQ by the mortality rate of each country. The number of deaths (Ndeaths) related to HCQ exposure was obtained by multiplying (i) the previously estimated number of in-hospital deaths per country and (ii) the OR of HCQ-related mortality. The main source of uncertainty included in the model was the range of HCQ exposure, which was estimated using the median and extreme values (minimum, maximum) reported in the included cohort studies:
Ndeath = Nhospitalised patients × mortality rate × HCQ exposuremedian, min, max × ORHCQ-mortality
We performed statistical analysis with R (meta-package, R Language and Environment for Statistical Computing, Vienna, Austria).
From the meta-analysis they found that the risk of death in COVID-19 patients treated with HCQ was 1.11, or an 11% increased risk of death, and they used this median estimate and the estimated rate of use of HCQ in each country (e.g., the percentage of hospitalized patients who received HCQ) to come up with a grand total estimate of 16,990:
The median number of HCQ-related deaths in Belgium, Turkey, France, Italy, Spain, and the USA was 240 (range not estimable), 95 (range 92–128), 199 (range not estimable), 1822 (range 1170–2063), 1895 (range 1475–2094) and 12,739 (range 3244–15570), respectively. Overall, using median estimates of HCQ use in each country, we estimated that 16,990 HCQ-related in-hospital deaths (range 6420–20294) occurred in the countries with available data.
The range, of course, is what I always look at, and in this case, it’s as few as 6,420 excess deaths attributable to HCQ to as many as 20,294. Where I agree again with Dr. Tucker-Kellogg:
And Dr. Jon:
In fact, the authors argue:
The main finding of the present study is that HCQ might have been associated with an excess of 16990 deaths during the first wave of the COVID-19 pandemic in the six countries for which data were available. Given that reliable data on hospitalizations, HCQ use and in-hospital mortality for most countries, these numbers likely represent the tip of the iceberg only thus largely underestimating the number of HCQ-related deaths worldwide.
Possibly, but it could have overestimated the number of HCQ-related deaths among COVID-19 patients; there’s no way of knowing. Indeed, the authors themselves conceded in the discussion that their “results might be overestimated by a factor 5 (i.e. the actual number of deaths related to HCQ would be ≈3000 deaths) or underestimated by a factor 2 (i.e. the actual number of deaths related to HCQ would be ≈30000 deaths), adding that the “effect of HCQ on mortality was the main source of uncertainty for the proposed estimates.”
I tend to agree that the study likely does underestimate HCQ-associated deaths, but I also have to concede that the imprecision in the estimates doesn’t allow me to assume that. Even if the study does overestimate deaths due to the rampant use of HCQ, the authors make some important observations in the discussion, which is why I quote this passage fairly extensively:
In the absence of restriction, the number of expected HCQ-related deaths is likely to be directly related to the promotion of its prescription by scientists, physicians and health agencies. In February and March 2020, the use of this treatment was widely promoted based on preliminary reports suggesting a potential efficacy against COVID-19 . For instance, the use of HCQ markedly increased from mid-March to mid-April 2020 ,  in France before a temporary recommendation supporting its use by the State Council was rapidly rejected . Similarly, the Food and Drug Administration (FDA) granted a temporary emergency use authorisation for HCQ on March 28th 2020, which was finally revoked on June 15th 2020 . In India, HCQ was also prescribed as a curative treatment to patients with COVID-19 and as a prophylactic treatment for front-line workers based on public authority guidance . Conversely, the British government promoted HCQ use only within clinical trials, explaining the absence of cohort studies reporting the use of HCQ in the United Kingdom in the present study . Consistently, a cohort of a multinational network showed a wide variation in the use of HCQ between countries, with 85% in Spain, 14% in the USA and less than 2% in China . The rush to administer this treatment caused supply shortages in community pharmacies, forcing the implementation of dispensing restrictions . Finally, the results of observational studies and randomized trials in May and June 2020, respectively, convincingly demonstrated that HCQ was ineffective and led to an increase in adverse events , , , , .
Our study showed a high proportion of prescriptions for HCQ even in countries which have restricted its use . This result argues in favour of tightly regulating access to off-label prescriptions during future pandemics . As an example, the issuance of the emergency use authorization for HCQ by the FDA, which permits distribution of the drug from the national stockpile for the treatment of hospitalized patients with COVID-19, was widely misinterpreted as an FDA approval for this indication .
Indeed, I note that at the hospitals where I have privileges, in March 2020 HCQ was included in the standard protocol to treat COVID-19 patients, based on the dodgy evidence from China that I described above. This was true at a most big hospital systems in the Detroit area, one of which even did a not-very-good non-randomized retrospective observational study full of confounders that was widely cited as evidence that HCQ worked. It wasn’t good evidence in support of HCQ.
There was, of course, wide variation regionally in the US and internationally in how much HCQ was promoted and accepted as a valid treatment for COVID-19 in those early months of the pandemic. For instance, the authors cite a study that found that in France “HCQ prescription was highly heterogeneous and largely influenced by several factors, the most important being the presence of an established departmental procedure supporting its prescription.” I also find it interesting that only 2% of patients in China were treated with HCQ, even though it was Chinese physicians and China that, based on the anecdotal observations in Wuhan that I described and some quick and dirty (mostly dirty) small RCTs, foisted HCQ onto the world as a potentially valid treatment in February 2020.
Regardless of whether this study over- or underestimated the number of deaths attributable to HCQ during the early months of the pandemic, I reiterate that it remains good evidence that, at the very least, a significant number of people died during that time because of the rampant injudicious use of HCQ to treat COVID-19 despite a lack of high quality evidence, which is why I agree with the authors when the write:
More importantly, this study illustrates the limitations of treatment-effect extrapolation from chronic to severe conditions without accurate data and the need to produce quickly high-level evidence from RCTs in case of emergent diseases.
Indeed. It also demonstrates the need to actually look at the source of evidence in favor of a treatment being quickly implemented because “we don’t have anything else,” which was basically the rationale for supporting such dubious evidence for HCQ as a treatment for COVID-19 in February 2020, when desperation, more than scientific evidence, seemed to be driving the adoption of treatments like HCQ. The same applies for the scientific plausibility of a treatment like HCQ, which requires a much higher concentration in cell culture to inhibit SARs-CoV-2 replication than is safely achievable in the bloodstream of humans, which was the same sort of preclinical evidence that drove the rapid promotion of ivermectin as a treatment for COVID-19, despite unfavorable preclinical pharmacology.
Finally, there is the later downstream effects of the low-evidence rapid adoption of treatments like HCQ and ivermectin for a pandemic disease like COVID-19, which is the encouragement of conspiracy theorists to view the later, almost inevitable, high quality randomized studies that fail to find a benefit as an example of “The Man” suppressing a “cheap and effective cure” for the disease. It’s too late to undo that, no matter how many studies conclude that HCQ is ineffective against COVID-19 or even, like this one, suggest that it harmed large numbers of people.