Who Believes in Hydroxychloroquine?
Whether the drug is effective for combating COVID-19 is a question that cannot be answered by science alone.
There was a time when we journalists struggled to find meaty stories to cover during the late-summer doldrums. No longer. Amid the continuing deluge of pandemic news, scientists at Oxford University recently complained that politics and misinformation were wrecking their attempts to fight COVID-19. They had planned to test whether a drug could slow or prevent infection with the coronavirus—but hospitals have pulled out of the study, threatening its completion. All because the drug is hydroxychloroquine. “I don’t think there’s been a more politicized and controversial medicine,” says Oxford scientist Nick White.
Touted by some as a game-changing treatment for COVID-19, hydroxychloroquine is seen by others as a dangerous and reckless approach. And the dispute is up to its neck in politics. Fueled by the endorsement of right-wing populists such as US president Donald Trump and Brazilian president Jair Bolsonaro, demand for hydroxychloroquine soared earlier this year. Most scientists, and opponents of Trump and Bolsonaro, insist the medicine is useless against the pandemic. Worse, they say, widespread use of hydroxychloroquine brings problems of its own, including serious side effects and threats to the supply for patients who rely on it to treat conditions such as rheumatoid arthritis and lupus.
The fight has spilled way beyond the pages of medical journals and into the mainstream debate about science, standards of evidence, and the trade-off between risks and benefits. If the stakes weren’t so high, the rivalry would make a fascinating case study. But as COVID-19 continues to rip across the globe, and as the Northern Hemisphere winter approaches, millions of people still have little to shield them from the impact of the virus except a crumbling collective will to follow social distancing guidelines and stay safe. Unless something can be done, tens more thousands, probably hundreds more thousands, will die.
For a hardcore group of believers, hydroxychloroquine is that something. It has been the drug of choice for many medical professionals in India and Brazil. A survey in March by Sermo, a social media network for physicians around the world, found that one-third of the 6,227 doctors polled had either prescribed hydroxychloroquine (or a related drug, chloroquine) to fight the virus or seen a colleague do so. More than one-third of the 2,171 doctors who responded to a question about treatment efficacy said it was the most effective therapy against COVID-19.
These numbers dropped as the year progressed. But it’s clear that many doctors are indeed prescribing hydroxychloroquine as a treatment for COVID-19. Their decisions are driven by the urge to help patients, but are perhaps influenced by business (India is a major manufacturer of the drug, for instance) and by politics (doctors in the United States were less likely to endorse its use, perhaps because Trump plunged the drug into a morass of partisan politics). These factors confound the already-difficult question of what it means to make evidence-based decisions about treatment in the face of a disease no one has seen before and for which almost all available information is preliminary. So is it possible to accurately assess hydroxychloroquine’s potential clinical value through the fog of uncertainty, urgency, interests, and politics?
The story of hydroxychloroquine began in the 1930s, as chemists worked to develop synthetic versions of the malaria-fighting drug quinine, which occurs naturally in the bark of South America’s cinchona tree. First came chloroquine, which had relatively toxic side effects, and then hydroxychloroquine, which was safer. It has been routinely used on lupus and rheumatoid arthritis patients for many years, and is widely seen as safe for those as long as the patients do not have heart arrhythmias.
In theory it should work . . .
But its potential applications to COVID-19 hark back to its original use as an antimalarial. Both chloroquine and hydroxychloroquine seem to be able to fight malaria because they raise the pH of fluids the malaria parasite relies on to break down and utilize its food, hemoglobin. In theory, the same deacidifying action should give protection against the entry and multiplication of invading viruses—making chloroquine and hydroxychloroquine attractive drugs to test against COVID-19.
Early lab experiments seemed to support that. In February, early on in the pandemic, researchers at the Wuhan Institute of Virology in China reported that chloroquine did indeed inhibit the growth of SARS-CoV-2, the coronavirus that causes COVID-19, in cell cultures. Weeks later the same team reported similar findings for hydroxychloroquine. “We predict that the drug has a good potential to combat the disease,” they wrote.
But for a drug widely described as an antiviral, hydroxychloroquine has never been found to act reliably against viruses in the human body. Both it and chloroquine have been tried, with mostly discouraging results, to treat HIV, hepatitis C, dengue fever, and chikungunya, which is another mosquito-borne fever.
Still, a team of scientists in France had had some success using hydroxychloroquine to treat the bacterial infections Q fever and Whipple’s disease. And in in vitro experiments, they found that the antibiotic azithromycin seemed to inhibit the replication of the Zika virus. So when COVID-19 started to crop up in France, the team treated patients with what seemed a promising combination: hydroxychloroquine and azithromycin. Fourteen COVID patients received hydroxychloroquine, while an additional six received azithromycin as well. Another 16 patients received neither and served as controls. Although 14 of the 16 control patients tested positive for the coronavirus by day six, only 6 of the 14 taking hydroxychloroquine did. And all 6 treated with the combination of hydroxychloroquine and azithromycin tested negative.
Critics have pointed to serious flaws in the study, starting with its small sample size, and also confounding differences between control patients and treated patients, but the doctors in charge declared it a success. The results caused a global sensation. Fox News hyped the study, triggering President Trump to send a now-notorious tweet on March 21 declaring the treatment had “a real chance to be one of the biggest game changers in the history of medicine.” Trump continued to talk up hydroxychloroquine, and two months later announced he was taking the drug prophylactically to try to reduce his chances of becoming infected.
But subsequent studies began reporting issues with using hydroxychloroquine to treat COVID-19. One influential study, published in The Lancet in May, reported an increased risk of heart arrhythmias and in-hospital mortality in COVID-19 patients taking the drug. The World Health Organization halted its trial of hydroxychloroquine in hospitalized COVID-19 patients—only to reinstate it when the Lancet paper was retracted in a major scandal over the authenticity of the unreleased patient database.
A few trials did suggest some benefit for COVID-19 patients taking hydroxychloroquine, but they had flaws. Meanwhile, other studies continued to show potential problems—or at least no benefit. On June 15, the US Food and Drug Administration yanked its emergency-authorization designation for the drug. Five days later the National Institutes of Health halted its randomized clinical trial, saying the drug had failed to reduce in-hospital mortality.
Which side are you on?
Now, and especially in the United States, doctors who prescribe hydroxychloroquine for their COVID-19 patients are often portrayed as reckless mavericks. Many appear on far-right television or radio shows, claiming their views are being suppressed. Facebook, Twitter, and YouTube did indeed take down one of the most egregious misinformation videos, in which doctors in white coats gave a press conference outside the Supreme Court in Washington, DC, claiming that hydroxychloroquine could cure COVID-19. Trump had retweeted the video.
Many involved say this rapid escalation—from potentially promising drug to utter outcast—has effectively shut down most of open, reasoned discussion over whether hydroxychloroquine could actually help fight COVID-19 in certain circumstances. In the United States, hydroxychloroquine is now so closely tied to politics, and its use so strongly associated with fringe movements, that the possibility that there may still be more to learn can no longer be entertained in polite company. So, what is there left to say?
It’s true that numerous clinical trials have not found any clear medical advantage. In the NIH’s words in June, using hydroxychloroquine against COVID-19 “does no harm, but provides no benefit.” But it’s also true that there are not yet enough data from multiple, well-designed clinical trials to rule out whether hydroxychloroquine can help in some circumstances against the disease.
One issue is that most of the clinical studies of hydroxychloroquine so far have been done in less-than-ideal and different hospital settings, where treatments for COVID-19 are complicated. There are various ways in which the drug can be administered—different doses, for example, and alone or in combination with an antibiotic such as azithromycin. It can also be given at different points in time as the infection progresses.
The end result is that published trials have used higher or lower doses of the drug; administered them both with and without antibiotics; and gave them to patients at different stages of infection, or sometimes not infected at all. Those differences make it hard to directly compare the results. So how to interpret them, and how to translate that interpretation into policy recommendations, comes down to a spectrum of judgement.
This is hardly an uncommon situation. The clinical efficacy of numerous drugs and treatments has been debated over years or even decades, covering such diverse matters as the benefits of microdoses of aspirin for stroke reduction and the effectiveness of various antidepressants. What makes the hydroxychloroquine situation different is the dual crucibles of a global pandemic and vicious partisan politics.
So, as always, doctors are left to make judgments based on imperfect evidence and their own experience. For some, that means holding off unless and until more studies are done. Peter Kremsner is a doctor and researcher at the University of Tübingen in Germany who is investigating possible clinical benefits of hydroxychloroquine against COVID-19. “It’s a drug I know well and we thought it certainly looked like a good option,” Kremsner says. Decades of use in malaria and other patients show the drug is safe and that the risks of side effects can be managed. Many of the trials for COVID-19 that suggest otherwise are giving unrealistically high doses of the drug, he says, which don’t represent proper clinical use.
Kremsner argues that the negative side effects of the drug have been exaggerated—but the lack of proven efficacy is enough of a deterrent to keep him from giving it to patients. “No, I would not prescribe it,” he says. “It shouldn’t be toxic and we know most patients tolerate it well, but we don’t know yet if it is doing any good. So I wouldn’t use it, because there is no clear indication yet that hydroxychloroquine is useful with this disease.”
Where’s the harm?
On the other side are those who say there is little harm in trying something that might work in patients who are seriously ill. “If they’re drowning, you either pull them out of the water or at least throw them a rope,” Constantine Tsamasfyros, a primary care doctor in Denver, Colorado, told a local TV station in April. He said he had used the combination of hydroxychloroquine and azithromycin to successfully treat symptoms in about a dozen COVID-19 patients. Of course, whether those patients would have recovered anyway can never be known.
And then there are those physicians who see potential promise in hydroxychloroquine but are unwilling to say so publicly, because of the stigma attached in repeating the same message as the right-wing activists. One such doctor, who works for a hospital in North Carolina and asked to remain anonymous, has not prescribed hydroxychloroquine to his COVID-19 patients because none have requested it. But he says he might do so in the future if they ask, and he keeps a supply for his own possible personal use if he were to become infected. “I just don’t think we know enough for anyone to say 100% that it works or doesn’t work,” he says.
When faced with having to make a clinical decision as fast as possible, this doctor says he draws on treatment information posted by other doctors on private medical social media channels. That’s more reliable, he says, than trying to sift through all the politicized statements and misinformation that is floating around in the public domain. On the front lines of treatment, he says, he can’t just wait for the perfect clinical trial to happen. “People in medicine talk about the gold standard of evidence,” he says. “But we just don’t have that kind of time.”
Among the few mainstream scientific voices who are willing to publicly agree with Trump and say that doctors should use hydroxychloroquine against COVID-19, one of the most prominent belongs to Harvey Risch, a cancer epidemiologist at Yale University. In late May he published a peer-reviewed article in the American Journal of Epidemiology arguing that hydroxychloroquine might work under some circumstances. And if it does, he says, then doctors should give it to people who might die if they don’t. “Tens of thousands of patients with COVID-19 are dying unnecessarily,” he says. “Fortunately, the situation can be reversed easily and quickly.”
Risch argues that hydroxychloroquine, if administered very early in the course of the disease, could significantly reduce its severity. He bases his conclusion on a number of observational studies that are statistically weak but show possible benefits. But apart from a single study with many statistical “limitations,” which tested whether the drug could prevent infection in exposed people (it did no better than placebo), published data from clinical trials have yet to address this question.
For stepping outside medical orthodoxy, Risch has been the target of criticism. A group of his Yale colleagues strongly criticized his stance in an open letter. “While minority opinions, anecdotal evidence, novel interpretations and challenges to orthodoxies in a field can be important,” they wrote, “at some point, the application of the scientific method generating evidence from multiple, well-designed clinical trials and observational studies does matter and should be heard over the noise of conspiracy theories, purported hoaxes, and the views of zealots.”
It’s hard to disagree with the principle. But words like “at some point” and “application of the scientific method” are doing a lot of work there. As the North Carolina doctor made clear, medical decisions still have to be made, lives hang in the balance, and uncertainty is inescapable. Doctors must often exercise their expert clinical judgment under such circumstances—in fact, knowledge obtained from learning-by-doing in the clinical setting that cannot be gained from the purity of double-blinded, randomized trials has historically been, and remains to this day, an important contributor to medical advance.
The big picture from most clinical trials certainly seems to show little benefit of hydroxychloroquine against COVID-19. One has to squint to see even a weak positive signal. But there’s enough uncertainty to allow scientists such as Yale’s Risch to squeeze out the case that hydroxychloroquine could still be useful for COVID-19 patients.
What makes Risch’s argument on behalf of hydroxychloroquine especially interesting is that he acknowledges how different approaches to medical science lead to different conclusions about what might work in the clinical setting. He describes a “clash in scientific worldviews” between scientists who extrapolate from animal and laboratory studies and controlled trials to a variety of real-world clinical settings, and epidemiologists who reason “on the basis of strong totality of evidence, sometimes even without RCT [randomized controlled trial] evidence.” Risch’s point is that incomplete and imperfect evidence suggesting possible benefit in a specific clinical context—patients who have yet to develop severe COVID-19 symptoms—is strong enough to justify action, and that waiting for results of clinical trials testing that particular use for hydroxychloroquine is not justified. In other contexts, this is what many public health scientists would describe as a precautionary approach.
The rejoinder from the other “scientific worldview” is there’s no point in waiting for more large trials, because if there were a benefit to using hydroxychloroquine against COVID-19, it would have been apparent by now. “When you say a study is going to have to be really big, or take too long, to prove a treatment works, that is actually an admission that the treatment barely works if at all,” says Peter Bach, a physician and health care policy specialist at Memorial Sloan Kettering Cancer Center in New York City. “When a treatment is highly effective, it is obvious even in a small, short study.”
Still, randomized controlled clinical trials to assess hydroxychloroquine and COVID-19 continue to chug along. Although the WHO and NIH trials are over, dozens of other such trials are ongoing internationally, including a 2,000-person study of health care workers in the US and trials testing the prophylactic efficacy of hydroxychloroquine in many different settings and among many types of patients. That fact alone suggests that not all medical researchers have written off hydroxychloroquine as a COVID-19 treatment, and that government bodies that oversee and approve clinical trials in various countries believe that some benefit might be possible.
News from the trenches
Whatever trials may end up saying about the efficacy of hydroxychloroquine, doctors continue to use the drug to treat COVID-19. After Risch’s first article was published, he says, he received 3,000 emails from physicians, almost all of them supportive.
Trump’s hyping of hydroxychloroquine certainly seems to have boosted its use in the United States, despite many doctors’ reluctance to buy into his pronouncements. An analysis from the US Centers for Disease Control and Prevention showed prescriptions for the drug from 48,900 retail pharmacies rocketed by 86% from February to March. Among those patients given hydroxychloroquine, the number who were simultaneously prescribed azithromycin soared more than a hundredfold, from 8,885 in February to 101,681 in March. This was at a time when COVID-19 began gripping the country, and so the increases might reflect the desperation of doctors looking for something that might work.
Nonetheless, the March survey by the Sermo network found that just 23% of US doctors believed that hydroxychloroquine might have clinical value. Outside the United States the picture was very different, with 62% of doctors in Italy and Spain, 44% in China, and 55% in the rest of the world saying that hydroxychloroquine was the most effective treatment they had seen or used at that point.
These data are from the early days of the pandemic, yet the differences are nonetheless striking. What can explain them? Did Chinese patients really do much better on hydroxychloroquine than those in the United States? Are Chinese doctors more easily fooled, more optimistic, or more sensitive to subtle improvements in patients to whom they gave the drug? Or perhaps doctors in China were simply more inclined to agree with the message sent from their political leaders that hydroxychloroquine was a useful medicine. Could US doctors have been primed to see fewer improvements in patients on the drug because of their distaste for Trump’s promotion of it, and the very public backlash it triggered from the medical establishment? None of these explanations seems satisfying, and in any case they would be difficult to test. But the differences point to how uncertainty, when titrated with different combinations of politics and culture, can lead the expert judgment of clinicians in opposing directions.
Another way that value judgments and politics have crept into the debate on hydroxychloroquine is on its risks. Scientists critical of Trump’s advocacy of the treatment are not content with merely pointing to the lack of strong evidence of efficacy. Many have also stressed that the president’s move is dangerous, because the drug carries potentially lethal side effects. Before it was retracted, the May Lancet article purporting to document hydroxychloroquine’s considerable risks was largely covered in the US media as a scientific repudiation of Trump’s claims that hydroxychloroquine is “widely used and safe.” An editorial in April in the medical journal BMJ had already countered Trump’s assertions by pointing out that wide use of the drug would expose some patients to “rare but potentially fatal harms” including acute liver failure. True—but read the cited reference and a different picture emerges. It’s a case study published because hydroxychloroquine is so widely used and thus considered to be safe—hence, the two cases of liver damage reported are noteworthy because they are very much the exception. “About 600 drugs have been suspected of possible hepatoxicity,” the paper notes, which together are responsible for 7-15% of reported cases of liver failure. Two of the chief culprits—and far more worrying for patients than hydroxychloroquine—are antidepressants and nonsteroidal anti-inflammatories such as aspirin and ibuprofen. They, of course, remain some of the most commonly taken medicines in the world.
The Food and Drug Administration explained its June 15 decision to revoke its emergency authorization for using hydroxychloroquine for COVID-19 by stating that “the known and potential benefits of [the drug] no longer outweigh the known and potential risks for the authorized use.” At that time, the agency’s Adverse Event Reporting System listed 1,064 cases of serious cardiac problems, and 200 deaths, associated with taking it. But that’s over more than 40 years and covers millions of doses taken by people for malaria and other conditions, Risch points out.
Clinicians with experience of the drug are more bullish about using it. In April, for example, rheumatologists in China, anticipating Risch’s argument, advised that hydroxychloroquine “has few side effects and should be used as an initial treatment as soon as the diagnosis of COVID-19 is made.” They noted they had successfully treated patients with autoimmune diseases with hydroxychloroquine over the course of years. And in 2013, long before the pandemic, three US rheumatologists, writing in the journal Seminars in Arthritis and Rheumatism, speculated on its possible use in “many systemic and chronic illnesses.” Risks of treatment, they said, are low. Perhaps more to the point is that risk is itself a slippery concept, highly dependent both on context and values.
None of this is meant to argue that hydroxychloroquine is effective against COVID-19. But scientists are not as immune as some might think from the pressures and biases that influence decision-making. As a science journalist for 20 years I have worked in newsrooms and seen firsthand how politics, partisanship and subjectivity can whip up a powerful prevailing wind that a particular line on a contested scientific topic is the “correct” one. That does not always mean that stories that appear are biased—the interference instead comes in which stories are selected for coverage at all. And right now, media stories that can portray Trump as wrong on hydroxychloroquine—and even better, downright reckless and dangerous on the topic—will fly more easily in most media.
As it stands now in the United States, the combination of media, politics, and science has seemingly settled the hydroxychloroquine question. It doesn’t work; it’s dangerous; if you believe in it there’s something wrong with you. Elsewhere, the situation is less clear. A recently approved randomized clinical trial in Italy will evaluate the safety and efficacy of hydroxychloroquine in preventing COVID-19 and treating the disease in its early stage; about 2,300 subjects are currently being recruited for the study, which could take up to two years to complete. If the results are positive, a trial to evaluate efficacy relative to other treatments would then be conducted, perhaps taking another three years. Annibale Biggeri is a medical statistician at the University of Florence who helped to design the trial. Which of Risch’s scientific worldviews does he line up with? Neither, it turns out: “I am agnostic about hydroxychloroquine. I am also very skeptical toward medical enthusiasm based on pharmacological or molecular arguments or personal clinical experience.”
Can hydroxychloroquine be effective against COVID-19? The likelihood is that we’ll never know. That’s not an uncommon situation in medical science. Should women get regular mammograms? Does lowering dietary salt intake reduce population-wide mortality? Can exposure to the weed-killer Roundup cause cancer in humans? After decades of research, such questions are still being debated, with scientists and decision-makers lined up on opposing sides, recommending contradictory courses of action. The question of hydroxychloroquine’s efficacy belongs to this list.
As the pandemic ravages the globe, forcing doctors to make urgent choices, decisions about the drug’s use will continue to be determined by a combination of science and politics. For the moment, this seems to be the best that evidence-based medicine can do on the front lines of the fight against COVID-19.