President Trump’s comments about hydroxychloroquine (HCQ), an antimalarial drug also used to treat rheumatic diseases and a possible treatment for Covid-19, have poured kerosene on an already fierce scientific debate about its efficacy. On March 21, Trump tweeted, “HYDROXYCHLOROQUINE & AZITHROMYCIN, taken together, have a real chance to be one of the biggest game changers in the history of medicine.” Trump rekindled the controversy a few weeks ago when he claimed that he had been taking HCQ as a prophylactic after he was exposed to someone with the disease.
Trump’s statements provoked accusations that he was peddling false hope, but the evidence merits consideration. Trump’s March 21 tweet pointed to a study from an infectious-disease laboratory at a university hospital in Marseilles. The laboratory’s leader, Didier Raoult, advocates a combination of HCQ and the antibiotic azithromycin for Covid-19. Further evidence for HCQ in March came from several doctors and patients who claimed that the medication had worked for them.
Most researchers were skeptical of both sorts of evidence, for good reason. Raoult’s first study was widely criticized for irregular data collection and a small sample size. His follow-ups did not include a control group, making the results hard to interpret: Raoult says that he is so certain that his treatment works that giving patients a placebo would be unethical. Meantime, most researchers discount individual doctor and patient testimonials, because even worthless treatments may seem to help a disease that usually goes away on its own.
But more evidence exists in the drug’s favor than anecdotes and the claims of one French professor. First, HCQ and the similar drug chloroquine, widely used in China to treat Covid-19, are mildly alkaline and might inhibit a stage of viral replication that involves the acidification of components of host cells. HCQ is also an “ionophore” that increases within-cell concentrations of zinc ions, which can also inhibit viral replication. (Vladimir Zelenko, a doctor in a Hasidic Jewish community in upstate New York and a prominent advocate of Raoult’s treatment, prescribes zinc supplements alongside HCQ and azithromycin.) Some studies of cells in vitro (that is, in Petri dishes) also suggest the efficacy of HCQ or chloroquine. In 2004, a study found that chloroquine protected cells in vitro from infection with the SARS coronavirus, which is similar to SARS-CoV-2, the virus that causes Covid-19. A study from China this past March also found that HCQ works in vitro against SARS-CoV-2.
These findings are not probative, because medications that work in theory and in vitro often fail in practice. But several other strands of evidence suggest that HCQ may help treat Covid-19. Raoult, for example, points out that fatality rates in Marseilles, where his protocol is in wide use, are significantly lower than elsewhere in France. This result could be due to other factors, such as Marseilles’ sunny Mediterranean climate (Vitamin D deficiency probably aggravates Covid-19), but it’s still a strong point in his favor. Several clinical trials, though not all, have shown positive results, such as two small randomized trials in China and “retrospective” studies, which analyze collections of medical records to discern the effectiveness of treatments, in China, Spain, and Korea.
HCQ seems to show promise in treating Covid-19 in its early stages. In February, doctors at a Wuhan hospital noted that none of the 80 patients taking HCQ to treat lupus had contracted Covid-19, and, conversely, none of 178 patients diagnosed with Covid-19 pneumonia had been taking HCQ—a possible coincidence, given the small numbers of patients involved, but still suggestive. In Italy, infection rates for patients with rheumatic diseases, many of whom take hydroxychloroquine, seem to be a fraction of the overall rate. The Indian Council of Medical Research has conducted trials of HCQ as a preventive for hospital employees and police officers, and found positive results for sufficiently high doses. In a hospital in Korea, hundreds of patients and staff were treated prophylactically with HCQ after they were exposed to an infected hospital worker; 14 days after exposure, none of the 211 patients and staff exposed had an active infection.
Some evidence for the efficacy of HCQ comes from a Twitter user who writes under the pseudonym “Covid19Crusher,” and who has collated data on Covid-19 cases and deaths for many nations. He notes that several nations that adopted a protocol that included early treatment with HCQ—such as Morocco, Turkey, and Russia—saw accelerated disease-recovery rates shortly afterward. Something similar seems to have happened in Italy, where doctors have reported a rapid fall in hospitalization rates in regions that adopted early HCQ treatment. “Covid19Crusher” notes that Costa Rica, which has long used a protocol including early administration of HCQ, has an order of magnitude fewer cases than its Central American neighbors.
Much of this positive evidence is circumstantial, relies on small data sets, or contradicts other studies. For example, some studies have failed to find prophylactic effects of HCQ. A recent study from Korea, for example, found no evidence that low doses of HCQ taken for autoimmune disorders worked as prophylaxis. A recent randomized controlled study, run by the University of Minnesota and published in the New England Journal of Medicine, found that HCQ prophylaxis did not reduce infection rates in American medical workers exposed to infectious Covid-19 patients. (The Minnesota study diagnosed Covid-19 in most subjects with inexact symptom checklists rather than proper virus tests, but it was otherwise well designed.)
Still, some evidence for the efficacy of HCQ exists—and has gone almost entirely ignored by most media outlets. Reports on Trump’s announcement that he was taking HCQ prophylactically, for instance, typically featured the word “unproven” or even said that there was “no evidence” that the drug was effective. One report from the BBC, for example, was headlined, “Trump says he is taking unproven drug hydroxychloroquine,” and further warned, “There is no evidence hydroxychloroquine can fight coronavirus, and regulators warn the drug may cause heart problems.” Vanity Fair likewise warned, “There’s no evidence showing that the drug is an effective preventative measure against the coronavirus, and there can be dangerous side effects.” Almost the only journalists to present evidence in favor of HCQ have been explicitly conservative, such as Fox News anchor Laura Ingraham.
Studies claiming that hydroxychloroquine is useless or even harmful, by contrast, have gained massive media attention, despite frequently severe flaws. The clearest example is a recent high-profile article published in the journal The Lancet of hospital records that concluded that hydroxychloroquine administration increased the risk of death for hospitalized Covid-19 patients by 33 percent. The study got uncritical writeups in the New York Times, Washington Post, and CNN. Both the Times and the Post used the study to level implicit criticism at Trump’s prophylactic use of HCQ, a non sequitur in any case: antiviral drugs work better the earlier they are given, and a finding of no benefit in patients who have already been hospitalized has little bearing on prophylactic use.
The Lancet study’s methodology, moreover, had several severe flaws that scientists spotted quickly. For instance, in most nations (including, for the most part, the United States), HCQ is still given only to the worst-off patients (patients in the study who received HCQ were almost three times as likely to be on ventilators). Without careful statistical controls for disease severity, this “indication bias” will tilt any comparison of patient outcomes against HCQ. The study’s controls, though, were inadequate: a rough measure of organ sepsis whose utility for predicting Covid-19 severity is dubious, and a dichotomous measure of blood oxygen saturation that lumped patients with severe oxygen shortages together with patients with only mild cases. (It has since transpired that the data set itself, purportedly assembled from records at 1,200 hospitals at unprecedented speed by a secretive enterprise called Surgisphere, is severely flawed and possibly fraudulent; most of the study’s authors have now asked for the study to be retracted.)
Similar flaws can be found in other studies that received widespread uncritical coverage. For example, a widely publicized retrospective study of patients in New York, which appeared in the New England Journal of Medicine, found that HCQ did not help patients avoid a “composite endpoint” of dying or going on a ventilator. But when these two outcomes are disaggregated, the study’s own data suggest that HCQ significantly improves patients’ chances of survival.
Though not wholly harmless, HCQ is a relatively safe medication when taken with proper precautions. The CDC’s own information sheet on the drug calls it “relatively well tolerated” and notes that the most common side effects are minor and manageable: “stomach pain, nausea, vomiting, and headache . . . [which] can often be lessened by taking hydroxychloroquine with food.” Higher doses can affect eyesight, but only after several years of use. The World Health Organization also lists chloroquine and HCQ as treatments for rheumatic diseases on the Model List of Essential Medicines, which includes drugs the WHO considers “efficacious, safe and cost-effective.”
HCQ can cause an occasionally dangerous distortion of heart rhythm called QT prolongation—exacerbated by using it in combination with azithromycin, which also prolongs QT. (One large retrospective study of almost 2 million Covid-19 patients found that the HCQ/azithromycin combination increased cardiovascular risk, but not HCQ alone, and the Minnesota prophylaxis study reported only minor side effects from HCQ alone.) But this risk is manageable. The risk factors for dangerous QT prolongation are well understood: heart rhythm can be monitored easily with a cheap home EKG, and other antibiotics such as doxycycline can substitute for azithromycin without posing risks. Millions of Americans already take HCQ at home without incident to treat autoimmune disorders or for other reasons.
Media coverage of HCQ, though, has consistently exaggerated the danger of rare side effects. One op-ed in the Washington Post by a professor of science journalism warned against taking HCQ without Trump’s advantages of “an in-house doctor and 24-hour monitoring to check for side effects—such as blindness, kidney failure, suicidal thoughts and heart attacks.” Blindness is only a danger to patients who take HCQ for several years, and though kidney impairment is a side effect of the related but more toxic drug chloroquine, HCQ seems to be harmless for kidneys and even protects against kidney damage in patients with rheumatic diseases. The suspension of a Brazilian study on chloroquine after several patients developed fatal heart arrythmias was also widely reported on, usually without mentioning that chloroquine is significantly more toxic than HCQ and that the doses used in the study, though comparable with some early Chinese protocols, far exceeded anything used in the West. (Another arm of the study testing a lower dose of chloroquine proceeded without incident, though without finding any benefit.)
Many articles about HCQ refer to an Arizona man who supposedly died after listening to Trump and self-medicating with a lethal dose of fish tank cleaner, which contains chloroquine. The relevance of this story to the risks of taking HCQ under medical supervision was always questionable, but is especially so now that it seems the man may have been murdered by his wife, who invented the fish tank cleaner story as an alibi.
Though the efficacy of HCQ in treating Covid-19 remains uncertain, available evidence suggests that the drug could have some benefit and, in any case, presents negligible risk for most. Of course, further research into HCQ and other treatments should continue.
What is not warranted, though, is the portrayal of HCQ as deadly. Exaggeration of the dangers of HCQ has likely reduced patients’ willingness to enroll in randomized controlled trials, the only definitive way of determining whether HCQ will work. Even if HCQ proves effective as a treatment, many patients frightened by the media may refuse it, costing lives that could have been saved. Politically motivated skepticism and refusal to do due diligence on studies that support preconceptions are likely to have fatal costs. One hopes that when the next emergency comes, the media set aside their hostility to a politician and honestly report on scientific information with vast consequences for public health.