The COVID-19 pandemic has been a golden opportunity for quackery and antivaccine conspiracy theories. This is not unexpected, given the size and the current lack of specific therapies for the disease other than supportive care and the current death toll. However, there is another assault on evidence- and science-based medicine that doesn’t come from quacks (although quacks have definitely joined in on the attack). Rather, it comes from desperate and/or opportunistic doctors who, either through a terrified imperative to “do something, anything” in the face of dying patients whom they can’t save or from self-aggrandizing doctors like Dr. Mehmet Oz, Didier Raoult, and others, who, drunk with the arrogance of ignorance and loving the attention, promote unproven treatments with no evidence an irresponsible attitude of, “What harm could it do?” (Answer: A lot. The specific drugs, chloroquine and hydroxychloroquine, have toxicities that have been well-known since the 1960s.) Add to that politicians like Donald Trump, full of magical thinking and possessed of a long history of selling snake oil himself, glommed onto these drugs as the solution to the pandemic before clinical trials showed any benefit.
I’ve written about how incredibly weak the evidence base supporting the use of chloroquine or hydroxychloroquine (including the combination of hydroxychloroquine and azithromycin) is twice already. Given that there have been developments in the two weeks since last I wrote about this topic, I thought that an update would be in order, as there are new studies. I’ll start with the “miracle cure” testimonials that have been used to argue that hydroxychloroquine ± azithromycine “saved” patients. (I know I’ve discussed a couple of these before.) Then I’ll move on to Didier Raoult’s latest study. Finally, I’ll discuss other evidence, including evidence of drug toxicity without evidence of benefit.
Familiar-sounding “miracle cure” testimonials
I’ve alluded to the sorts of anecdotes that have spread on social media and promoted by President Trump and his allies and how much they resemble alternative cancer cure testimonials from the patients of quacks like Dr. Stanislaw Burzynski. I’m referring, of course, to the stories of people like Rio Giardinieri and Jim Santilli, both of whom report their symptoms having remarkably improved after one dose of hydroxychloroquine.
The testimonial by a local Michigan State Representative Karen Whitsett takes the sort of testimonials that have been circulating one step further:
A Democratic state representative from Detroit is crediting hydroxychloroquine — and Republican President Donald Trump who touted the drug — for saving her in her battle with the coronavirus.
State Rep. Karen Whitsett, who learned Monday she has tested positive for COVID-19, said she started taking hydroxychloroquine on March 31, prescribed by her doctor, after both she and her husband sought treatment for a range of symptoms on March 18.
“It was less than two hours” before she started to feel relief, said Whitsett, who had experienced shortness of breath, swollen lymph nodes, and what felt like a sinus infection. She is still experiencing headaches, she said.
Unsurprisingly, the next day, Trump was touting her story:
Although he didn’t specifically name her, President Donald Trump praised a Detroit Democratic state representative during his daily briefing Tuesday for singling him out as she fought a coronavirus infection.
State Rep. Karen Whitsett, D-Detroit, who tested positive for coronavirus, said Monday that the drug hydroxychloroquine, and Trump’s continued trumpeting of the drug, helped save her life.
“She thought she was dead,” Trump said Tuesday. “I think she’ll be voting for me now.”
Whitsett, who learned Monday she had tested positive for COVID-19, said she started taking hydroxychloroquine on March 31, prescribed by her doctor, after both she and her husband sought treatment for a range of symptoms on March 18, twelve days. After taking the drug, she started feeling better within a couple of hours.
A relative of mine who forwarded the link to me noted that there were…inconsistencies…with Rep. Whitsett’s story. Supposedly, from what I can tell, she first started experiencing symptoms on March 18 and started taking hydroxychloroquine on March 31, with immediate relief. Then there’s this:
Whitsett said she was familiar with “the wonders” of hydroxychloroquine from an earlier bout with Lyme disease, but does not believe she would have thought to ask for it, or her doctor would have prescribed it, had Trump not been touting it as a possible treatment for COVID-19.
I note first that it shouldn’t be too surprising that Rep. Whitsett got better in roughly twelve days, given that that’s well within the parameters of the typical time course of symptomatic COVID-19 illness for people who recover. I also note that chronic Lyme disease is, of course, a fake diagnosis beloved of quacks, which to me throws up a big red flag regarding her story. Be that as it may, as I pointed out before, these sorts of stories actually make me think that hydroxychloroquine probably isn’t responsible for the recovery of the patients. If you know how these drugs work, you know that it’s very unlikely that a single dose of the drug would work so dramatically within 2-8 hours, and these stories all have a rather odd similarity to them. Unfortunately, the media gobble these stories up, because they’re compelling human interest stories and are hopeful at a time when most of the other news about the pandemic is profoundly depressing and alarming. Worse, these stories feed the narrative that I discussed before: that those doctors citing the lack of evidence for these drugs, and urging caution, are leading to the unnecessary deaths of COVID-19 patients.
Unsurprisingly, Rep. Whitsett is, unfortunately, feeding that narrative by starting with a valid criticism of our response to the pandemic thus far (the relative lack of testing for COVID-19 compared to what we need in order to be able to control the pandemic while loosening the restrictions of the current general lockdown that’s been imposed on huge swaths of the US and how this response demonstrates major disparities in care related to race) to point out that the only reason she was likely tested for COVID-19 is because she’s a state representative and to accuse state health authorities of, in essence, letting her family members die:
Now, Whitsett says her cousin, Cheryl Fowler, is asking for hydroxychloroquine but has been unable to get it at Henry Ford Hospital.
“They will not give it to her,” she said. In order to change that, “I need to talk to the president of the United States.”
Gambrell said he believes doctors plan to give his mother hydroxychloroquine on Thursday, but they did not want to give it to her Wednesday because of other medication she is taking for pneumonia.
Fowler needs the drug immediately, Whitsett said. “They want to wait until the person needs a ventilator,” she said. “By the time they’re ventilated, it’s too late.”
The most recent narrative regarding hydroxychloroquine is that it has to be started early in the course of COVID-19 infection or it will be ineffective. Unfortunately, there currently is no good evidence to support this latest narrative. Of course, nothing will convince patients relating these anecdotes that hydroxychloroquine very likely didn’t save them:
“Whether hydroxychloroquine works or not is a question that still needs an answer. At this point anyone who says the drug helped them is simply relating an anecdote. I have also seen patients on hydroxychloroquine who have died. I totally understand the need for hope but this is not a miracle drug,” Local 4′s Doctor Frank McGeorge said.
Whitsett believes it worked for her.
Just like cancer patients believe that Stanislaw Burzynski’s quackery or various other cancer quackery “worked for them.” Humans, being pattern-seeking animals, are very prone to latching on to correlations and imputing causation to them. Meanwhile, right wing Fox News pundits are Tweeting things like this:
An old pal in Los Angeles (no Trump fan) just texted me that seven friends of hers (here and in Sweden) ALL took hydroxychloroquine, all recovered quickly, zero side effects. Called it a “godsend.”
— Laura Ingraham (@IngrahamAngle) April 10, 2020
Her anecdote, even if verified, doesn’t mean that hydroxychloroquine was why they got better. The vast majority of patients with COVID-19 recover uneventfully. Only a minority develop severe disease, and only a minority of patients taking hydroxychloroquine suffer significant side effects, the minority suffering life-threatening side effects being even smaller. Without a control group and a large number of patients you simply can’t tell if an intervention for COVID-19 “worked” or, more specifically, if hydroxychloroquine is safe in this setting.
As I’ve said before, the COVID-19 pandemic has indeed revealed major disparities in care and outcomes among minority populations, but attempting to restrict the off-label use of hydroxychloroquine or chloroquine in order to prevent hoarding and preserve the supply for patients who need it is not one of them. It’s also supremely irresponsible of journalists to publish these stories, even if they interview a medical expert to caution that, barring good evidence from randomized trials, there’s no way of knowing if the treatment was responsible for the patient’s recovery. The reason is that the cautionary statements from health experts about hydroxychloroquine ± azithromycin are always lost in the hopeful message of a patient who recovered after taking the drug.
Indeed, even when the patient gets worse after taking hydroxychloroquine, the narrative is that the drug saved him, as long as he ultimately gets better. Take the case of Billy Saracino. One week after developing COVID-19 symptoms:
The fateful day came on March 20, when Saracino woke up in the middle of the night uncontrollably coughing. His wife immediately called Bock.
At his doctor’s urging, Saracino swallowed three pills of Hydroxychloroquine (at 200 mg each), and another three the following morning before he was admitted to Hackensack University Medical Center.
“I had a feeling [Billy’s case] was turning bad,” said Bock, “so we started him before he went into the hospital.”
On March 21, with a fast heart rate and difficulty breathing, Saracino went to the emergency room:
Saracino was immediately taken for a chest X-ray, which showed bilateral viral pneumonia. Then, he was admitted to the ER — which, unlike the triage area — was completely full, he said.
“My heart rate wasn’t good and my oxygen was low,” Saracino said. “My doctor was on the phone with the doctors and went through all my symptoms. He said ‘You cannot let him out.'”
Saracino spent three nights in the hospital: Saturday, March 21 through Monday, March 23. He had chills, body aches and a terrible migraine. He would spent two hours sleeping, and then another two awake.
But then, exactly 48 hours after taking the initial dose of Hydroxychloroquine, Saracino says he started to feel better.
“I was sweating, I had to change clothes,” Saracino said. “Then I started walking around a little bit.”
“It’s no question,” Bock said. “I think Hydroxychloroquine really helped him through this.”
Every patient that Bock put Hydroxychloroquine on recovered except for one, whose condition worsened rapidly over the course of several hours, before the medication had time to work, the doctor said.
So let me get this straight. Saracino got worse for two days while taking the drug, and then he got better? Truly, no matter what the clinical course (other than death), whenever a patient receives hydroxychloroquine or the hydroxychloroquine/azithromycin combination, it must have been the drugs that were responsible for his recovery, even if he kept getting worse for two days after starting the drug before turning the corner. Seriously, you could just as plausibly blame hydroxychloroquine for Saracino’s clinical deterioration on March 20! Don’t believe me? Read the case of Carl Shuck, who was treated with hydroxychloroquine, azithromycin, and zinc but was still on a ventilator for two weeks before recovering. Read the case of Sophie Avouris, a 102 year old woman who was treated with hydroxychloroquine and the azithromycin for a week and recovered.
You get the idea. Basically, if you receive these drugs and recover, no matter how much time it takes, obviously the drugs cured you. Such is the narrative the media is feeding the public, “miracle recoveries” that might or might not have had anything to do with hydroxychloroquine with or without azithromycin.
Didier Raoult publishes another study
Unchastened by the fierce criticism for his first two studies, last week Didier Raoult published another study. The first of his trio of scientific crap was an unrandomized study that claimed that the combination of hydroxychloroquine and azithromycin eliminated the virus completely compared to controls or hydroxychloroquine alone. Its flaws were so numerous and manifest as to make the study painfully uninformative. Indeed, based on the shenanigans Raoult engaged in with the data, plus his history of fabricating data, I now suspect, but obviously cannot prove, scientific fraud. Even the International Society of Antimicrobial Chemotherapy (ISAC), which publishes the journal that published this study, backed away, even going so far as to state that the article “does not meet the Society’s expected standard, especially relating to the lack of better explanations of the inclusion criteria and the triage of patients to ensure patient safety.” Muddying the waters is a second statement issued a week later, apparently at the behest of Elsevier, that seems to soften the criticism and points out that the journal’s peer review process had been followed, even as it suggests a possible joint investigation with Elsevier.
The second study in the trio was a single arm case series that consisted of a huge percentage of COVID-19 patients treated with azithromycin and hydroxychloroquine who were not very ill at all. Only 15% of them even had fever, even though data from around the world show that 90% of symptomatic COVID-19 patients have fever during the course of their illness. Basically, the vast majority of these patients were patients who would have been sent home and told to self-quarantine, with instructions to call or go to the emergency room if they get significantly worse and especially if they start experiencing shortness of breath. Raoult claimed that the combination of azithromycin and hydroxychloroquine was very effective in clearing the SARS-CoV-2 virus, but there’s no way of knowing if these patients wouldn’t have cleared the virus anyway. It, too, was a singularly uninformative study.
So what about the third study? As of Sunday, Raoult has published only an abstract and a data table for 1,061 patients treated with his azithromycin/hydroxychloroquine combination, with no full manuscript yet. Of course, this dataset suffers from even more issues than his previous papers, as there are no methods to look at, and we can’t look at his study design in depth or his inclusion/exclusion criteria. Taking the results at face value, between March 3 and April 9, Raoult’s group tested 59,665 specimens from 38,167 patients for COVID-19 by PCR. There were 3,165 patients who tested positive, of which 1,061 met Raoult’s unstated inclusion criteria and were treated with his azithromycin/hydroxychloroquine combination. The mean age of the patients was 43.6 years old, and 492 (46.4%) were male. The results further state that a “good clinical outcome and virological cure” were observed in 973 patients within 10 days (91.7%), with prolonged viral carriage in 47 patients (4.4%), with viral clearance in all but one by day 15. They further observed:
A poor outcome was observed for 46 patients (4.3%); 10 were transferred to intensive care units, 5 patients died (0.47%) (74-95 years old) and 31 required 10 days of hospitalization or more. Among this group, 25 patients are now cured and 16 are still hospitalized (98% of patients cured so far). Poor clinical outcome was significantly associated to older age (OR 1.11), initial higher severity (OR 10.05) and low hydroxychloroquine serum concentration. In addition, both poor clinical and virological outcomes were associated to the use of selective beta-blocking agents and angiotensin II receptor blockers (P<0.05). Mortality was significantly lower in patients who had received > 3 days of HCQ-AZ than in patients treated with other regimens both at IHU and in all Marseille public hospitals (p< 10-2).
This study suffers from the same issues that Raoult’s much smaller study. The overwhelming majority of patients (1,008, or 95%) had a low national early warning score (NEWS), with 97.4% of the patients with a “good clinical outcome” (not defined). The table shows patients taking beta blockers and angiotensin II receptor blockers having “poor virological outcome,” while patients taking metformin (used to treat type II diabetes), beta blockers (used for hypertension and angina, as well as certain cardiac rhythm disturbances), dihydropyridine derivatives (a class of calcium channel blockers, often used for hypertension, cardiac rhythm disturbances, and angina), angiotensin II receptor blockers, HMG CoA reductase inhibitors (statins), and diuretics (used to treat hypertension and heart failure) were more likely to have “poor clinical outcomes.” Of course, it is most likely the diagnoses for which these drugs are used, rather than the drugs themselves, that are responsible for the poor outcomes in COVID-19 patients, and there’s no indication that Raoult’s group tried to do any sort of analysis to tease out whether these associations were with the drugs themselves or with the conditions being treated with the drugs.
The bottom line is that this study is also uninformative. For one thing, a 0.5% fatality rate is at the low end of current estimates and might even be the “true” case fatality rate when all cases of mild and asymptomatic infections are included in the denominator. This study certainly doesn’t show that Raoult’s treatment was responsible for the low mortality rate. When the overwhelming percentage of the patients you treat have mild or asymptomatic disease, then of course should expect to observe that the fatality rate among these patients will be low, no matter what you do. The reason Didier Raoult refuses to do a truly randomized trial is because he’s now a true believer in his combination, which has led him to conclude that it’s unethical to randomize COVID-19 patients to a placebo control group. Worse, French President Emannuel Macron has started emulating our President Donald Trump and become a booster for Raoult’s bad science. Whether it’s out of a misplaced nationalistic pride or other reason Macron has been talking up Raoult’s results, even though Raoult has produced no data sufficiently compelling to know if his combination is safe and effective or not.
Worse, there appear to be some…discrepancies…in the reporting of Raoult’s data on his website. There’s an excellent Twitter thread by Lu Chen, a scientist at the NIH’s National Center for Advancing Translational Sciences functional genomics library:
Once again, data is broken, and I smelled their anxiety several days ago. https://t.co/sKQe0tB5DY
— Lu Chen (@houndcl) April 9, 2020
She reported possible discrepancies between what Raoult published, and what he had previous posted on the IHU COVID-19 website:
(2/n) I closely monitor IHU’s COVID-19 website since the birth on 3/26. https://t.co/1ea5nQKR37
This website reports the number of COVID positives & death in APHM/IHU, and num of positives & death with >3d treatment. Some major discrepancies have been found.
website data: pic.twitter.com/NODsEQVZRX
— Lu Chen (@houndcl) April 9, 2020
Indeed, she notes some suspicious numbers:
3/30: number of patients with “H+A” treatment -8
4/4: no update
4/5: total death + 1, BUT “H+A” death +2. (NOBEL PRIZE PLEASE!)
4/5-4/6: >10k increase in total tests, >2k increase in total positive.
4/8: Total tests -10k, total positives -1.7k
— Lu Chen (@houndcl) April 9, 2020
And a “massive data loss” after April 6, with data only through March 31 being reported”
I had a gut feeling on 4/6 that they will take “some” action since the death in H+A start to increase dramatically.
What a coincidence!
-No data on 4/7
-Massive data loss on 4/8
-Rush a preprint out on 4/9 which reports the data till 3/31
— Lu Chen (@houndcl) April 9, 2020
Now, as a response in the Twitter thread mentions, there could be an innocent explanation, but I agree 100% with Lu Chen that the data management is a disaster and Raoult’s group has no idea what it’s doing analyzing clinical trial data. Basically, I don’t trust any data coming out of that group:
The data management is a disaster. No matter combo work or not, Raoult's team doesn't have a good system to handle the clinical data from thousands of patients properly.
Cannot wait to see the data from the other 1k patients who already received 3d treatment.
— Lu Chen (@houndcl) April 9, 2020
Also, it’s important to emphasize that the patients in this trial were drawn from the general population volunteering to be tested for COVID-19, which likely explains why 95% had mild or asymptomatic disease. This trial says basically nothing about whether Raoult’s combination works.
Raoul Didier: Brave maverick doctor
I sometimes like to refer to doctors who think they have The Answer to a particular clinical problem, even when the evidence that their proposed treatment is an “answer” to anything is incredibly weak or even nonexistent, as “brave maverick doctors.” They like to see themselves as bucking the system. Because, prior to his promotion of hydroxychloroquine and azithromycin being echoed by President Trump, celebrity doctors like Dr. Oz and Dr. Drew, and a wide variety of right wing pundits, he was unknown outside of France, other than in some international scientific circles. That’s why this post was of great interest to me, as it described his history. And what a dubious history it is, at least with respect to COVID-19. On January 23, 2020, Raoult was saying things like:
You know, this is a crazy world. The fact is that some people have died from the coronavirus in China, you know, I don’t feel so concerned. It’s true to say that the world has gone completely crazy, that is to say that if something is happening where there are three Chinese people who are dying and that becomes a worldwide alert, the WHO gets involved, it’s on the radio and on television. […] In other words, there is no longer any sense of lucidity.
Whenever there is a disease in the world we wonder whether we are going to have the same thing in France. It has become utterly insane. [ …] I don’t know, people don’t seem to have anything to worry about, so they try and find something to be afraid of in China, just because they are unable to face what they might be afraid of if staying in France. That’s it, this is not serious.
Eventually, he discovered hydroxychloroquine. One thing that bothers me is how few people know the origin of the idea that hydroxychloroquine might be effective against COVID-19. The hypothesis that antimalarial drugs might be effective treatments for COVID-19 originated in Wuhan, China during the early phase of the pandemic in January. There, Chinese researchers reported that none of their 80 patients with lupus erythematosus who were taking hydroxychloroquine went on to become infected with SARS-CoV-2. As a result of that and old evidence of antiviral activity for the drugs, they became interested in using these antimalarial drugs to treat COVID-19. (Never mind that immunosuppressed patients are exactly the patients most likely to assiduously follow the recommendations of public health authorities during a pandemic.) A number of clinical trials were registered, and, based on anecdotal reports and small clinical trials (nearly all of which are as yet unpublished), in February the Chinese government published an expert consensus recommending CQ or HCQ for patients with COVID-19. Soon after, a number of nations followed suit. Apparently Raoult found out about this, and suddenly, at the end of February, he posted a video entitled Coronavirus: End of the game, later renamed “Coronavirus: A way out of the crisis?”
In the video, Raoult said:
- “From all respiratory infections, it is probably the easiest to treat.””
- “There’s no need to get excited.”
- “Be careful, there will soon be no more chloroquine in drugstores.”
The article lists numerous other examples of Raoult downplaying the severity of COVID-19, comparing it to seasonal influenza, and denouncing government actions to stop its spread as an overreaction. There’s also a recounting of his history of abusive behavior towards his underlings, charges of sexual harassment leveled at him, his history of being a climate science denier, and mismanagement at his institute. Meanwhile, evolutionary biologist P.Z. Myers has dissected in detail Raoult’s painful misunderstanding of evolutionary biology, while others have described a history of data fabrication by his group. Oh, and did I mention that he might be an antivaxxer? In 2018, he published a book entitled La vérité sur les vaccins. The title and the blurb for the book sure sound at least borderline antivaccine to me. The very best thing that can be said on this score is that Raoult strikes me as someone who exaggerates the risks of vaccines and buys into too much false balance with claims made by antivaxxers.
In other words, Didier Raoult is a dubious source.
Conflicting, but mostly negative, evidence
So, the evidence base supporting hydroxychloroquine comes mostly from Didier Raoult, but not entirely. The most recent Chinese randomized trial was published online as a preprint (meaning that it hasn’t undergone peer review yet). It’s a larger (but still small) randomized trial of hydroxychloroquine given for five days at a dose of 400 mg/d. They reported that time to clinical recovery, body temperature recovery time, and cough remission time were significantly shortened in the treated group and that the proportion of patients with improved pneumonia was higher in the hydroxychloroquine group (80.6%) compared to controls (54.2%). They also noted that all four patients who progressed to severe illness were in the control group, while two patients in the hydroxychloroquine group experienced minor side effects (one had a rash, the other headache). There are some problems, however. First, this is not placebo-controlled. The authors claim that neither “the research performers nor the patients were aware of the treatment assignments,” but how could that be? The patients could know if they were receiving hydroxychloroquine or not, as could the nurses and the physicians treating them. That could subtly have affected their assessments of the patients, and it could have affected how patients reported their symptoms. (The study relied on patient reports of how severe their coughs were, for instance.) True, body temperature is an objective endpoint, but most other endpoints examined has at least some degree of subjectivity. Worse, the statistics were dodgy on the paper. For instance, I see what looks like Student’s t-test being used on a case of multiple comparisons, and one of the p-values (Table 2) is 0.0476. Let’s just say that whenever a p-value is that close to 0.05 buy still “significant,” my skeptical antennae start twitching wondering if the data underwent some “massaging” to make sure the p-value was significant.
These Tweets also point out some other flaws:
Details do not match the original plan, to recruit 300 cases in three groups. Intended outcomes: RNA clearance and T cell recovery time.https://t.co/mkfJy3Po8A
Earlier press release is different again:https://t.co/sEEXIIeSrR
— Smut Clyde, X-Ray Haruspex (@SmutClyde) March 30, 2020
— Leonid Schneider (@schneiderleonid) March 31, 2020
They obviously had recruiting problems, given they only found 62 patients with an initial target of 300.
Original plan: (a) control group, (b) 100mg 2x/day, (c) 200mg 2x/day.
100mg 2x/day is underdosed given the IC50. So dropping that group to increase N/group makes sense.
— ice9 (@__ice9) April 1, 2020
They write "significantly improved" but the significant is just a clinical assessment. There is not a single significance test in the paper. I think they had to stop the study when they ran out of new patients and simply published results so far. Underpowered results.
— JHanisch / better 6' apart than 6' under (@jhan2qt) March 31, 2020
Basically, the best you can say about this study is that it is mildly promising, but so badly flawed that no firm conclusion can be drawn. Rigorous double-blind, randomized, placebo-controlled trials are desperately needed.
It’s also not as though there weren’t evidence of harm, either. For example:
France reported 43 cases of heart incidents linked to treating coronavirus patients with hydroxychloroquine, the malaria drug President Donald Trump has repeatedly touted as a potential “game changer.”
As the U.S. stockpiles as many as 29 million doses of the drug, which is also used to treat lupus and rheumatoid arthritis, the data on adverse reactions from France’s drug safety agency highlights the risk of using unproven treatments to stem a pandemic that’s killed more than 100,000 people worldwide.
Meanwhile, Sweden is reporting a similar experience, with several hospitals no longer routinely using these antimalarial drugs, due to adverse events.
In addition, there are studies that are concerning about the potential risks of widely prescribing these drugs. Most are not peer reviewed and thus published on a preprint server, such as this one from NYU that observed QT-segment prolongation in adult patients treated with hydroxychloroquine. Last week, another preprint was published from a multinational team out of Brazil. This was a parallel, double-blinded, randomized, phase IIb clinical trial that studied 81 COVID-19 patients. There were two groups, patients receiving high dose (600 mg twice a day over ten days) versus low dose (450 mg twice a day on the first day, once a day thereafter for the next four) chloroquine plus ceftriaxone, a cephalosporin antibiotic. There was no placebo control because Brazil had mandated chloroquine or hydroxychloroquine for COVID-19 patients. The study reports that the high dose patients showed more severe QT prolongation and that there was a trend towards higher mortality compared to the low dose. In summary:
The high dose CQ arm presented more QTc>500ms (25%), and a trend toward higher lethality (17%) than the lower dosage. Fatality rate was 13.5% (95%CI=6.9–23.0%), overlapping with the CI of historical data from similar patients not using CQ (95%CI=14.5-19.2%). In 14 patients with paired samples, respiratory secretion at day 4 was negative in only one patient.
To be honest, this “trend” had a p-value of 0.35; so I find it hard to call it even a trend. However, it is very clear that, at best there is no difference between the groups, and at worst higher doses of the drug are associated with a higher risk of death. Even so, the results were alarming enough that the Brazilian researchers stopped using high dose chloroquine, moved the patients in the high dose arm to the low dose arm, and are now studying only the low dose protocol, a reasonable precaution given that their preliminary data on viral clearance in respiratory secretions in confirmed cases suggest little or no effect of the drug, even at high dosage. However, in the absence of placebo controls, I really have a hard time figuring out how useful this study will end up being in determining if chloroquine is effective.
Bringing it on home
Finally, experience in hospitals in my neck of the woods suggest that hydroxychloroquine might actually be harmful. One is a study from my own institution that is currently undergoing peer review that suggests that reported worse outcomes in patients treated with hydroxychloroquine. It has somehow leaked prematurely on social media. I know about it because an associate dean sent it to a number of faculty. I didn’t share it, although I did mention publicly that I knew about it and had read it. All this leaves me, as it left others who knew about the study, torn on how to discuss it, given that it’s not yet peer reviewed or published, and that it hasn’t shown up on a preprint server yet. What I think I will do is simply to mention that it is a quasi-randomized study reporting on 63 hospitalized patients in affiliated institutions with COVID-19 that found, at the very least, no difference in outcomes, but with some parameters suggesting a worse outcome compared to supportive care alone. When the manuscript is published, either after peer review or on a preprint server, I’ll add the relevant links and then decide whether to beef up the discussion in this post or to do a new post about the study, either here or on my not-so-secret other blog.
That aside, my medical alma mater, the University of Michigan, recently published a blog post about its experience with COVID-19, including with the use of hydroxychloroquine to treat it. Basically, doctors at U. of M. have decided to stop routinely using hydroxychloroquine to treat COVID-19, because they have seen no evidence of efficacy but have seen toxicity:
Michigan Medicine just changed its guidelines on prescribing these drugs. There’s a lot of excitement about whether these already existing options for malaria could be useful in COVID-19, but so far that excitement hasn’t materialized in patient care data.
“Our infectious disease division and our antimicrobial pharmacists have reviewed all the available data and we found no convincing evidence that these drugs were effective in treating people with COVID-19,” Kaul says.
That’s consistent with what’s been observed firsthand in Michigan Medicine’s hospitals.
“We haven’t seen any clear evidence of benefit so we aren’t going to use hydroxychloroquine routinely anymore,” Chopra says. “We were initially recommending it to both inpatients and outpatients, but we’re no longer doing that routinely. That’s based upon the fact that we’ve been prescribing hydroxychloroquine for a few weeks, did not see therapeutic benefit, but did see adverse effects.”
Those side effects Chopra has seen in his patients include liver function toxicity, nausea and vomiting.
I’ll conclude by simply saying this. Taken together, the evidence that hydroxychloroquine, chloroquine, or the hydroxychloroquine/azithromycin combination is an effective treatment for COVID-19 is getting weaker with every publication, to the point where I am 95% sure now that they either don’t work or that any benefit they might have will be outweighed by side effects. Could I still be wrong? Sure, and I’d be happy to be wrong given the desperate need for effective treatments against COVID-19. But I don’t think I am. In the end, the hype over these drugs and particularly the rush by doctors on the flimsiest of evidence to make hydroxychloroquine, in essence, a standard of care, have shown just how weak most doctors’ allegiance to evidence-based medicine is and represent one of the most massive failures of EBM that I can recall.