Medical Examiner

Why We Can’t Just Let People Try Drugs That Might Treat COVID-19

The president has argued for widespread use of some promising drugs, saying, “What have you got to lose?” But patients have plenty to lose.

A person holding chloroquine.
A medical staff member at the IHU Mediterranee Infection Institute in Marseille, France, shows a packet of Nivaquine, tablets containing chloroquine.
Gerard Julien/Getty Images

There is increasing discussion about potential treatments for COVID-19, and even, in some cases, very early research into their effectiveness. One drug, the antiviral remdesivir, showed promise when tested in monkeys infected with similar coronaviruses and is being studied in clinical trials, some of which should have results to offer soon.

President Donald Trump has expressed particular enthusiasm about two other drugs: chloroquine and its sister drug hydroxychloroquine. Chloroquine is an old drug used to treat and prevent malaria; hydroxychloroquine is now a mainstay of treatment for rheumatoid arthritis and lupus. These drugs have not been studied in animals, but laboratory research has shown they have activity against related coronaviruses, and there have been some reports that people treated with these drugs in China seem to have done well. The president has argued for widespread use of these drugs, saying, “What have you got to lose?”

Unfortunately, patients treated with drugs that have not been adequately studied have plenty to lose.

First, although chloroquine and hydroxychloroquine have been used in many patients, they have not been used in critically ill patients. These patients may suffer toxic effects at doses used to treat people with chronic conditions but who are otherwise reasonably healthy. Alternatively, they might need higher doses to effectively treat this infection. We have no idea what the right dose would be for these patients. There are already reports of serious harms in people self-medicating with these drugs.

Second, many COVID-19 patients already have underlying health issues that require other medications. We have no information about how those medications might interact with chloroquine or hydroxychloroquine. Such drug-drug interactions can be fatal. These potential interactions are not something you want to play around with when the patient is already in bad shape.

Third, most patients who need treatment for COVID-19 eventually recover. The standard of care provided to these patients, while intense, saves most of them. Providing untested drugs to patients whose probability of survival may be upward of 80 percent is a risky business. (I once saw a cartoon in which a researcher was offering a patient the opportunity to participate in a clinical trial. The patient responded, “OK—as long as I can get the placebo!”)

Beyond possible harm to the patients themselves, there is one other aspect to consider with respect to the chloroquine drugs. Many patients with rheumatoid arthritis and other autoimmune diseases depend on these drugs to control their diseases.  Promotion of these drugs for the untested use in COVID-19 may create shortages and skyrocketing costs, with major impact on those who have depended on them for years. It would be one thing if we knew these drugs were truly lifesaving against COVID-19. But we don’t. So there are many people with plenty to lose if these drugs are commandeered for COVID-19.

What about remdesivir, the drug that probably has the most promising data behind it?  Remdesivir is currently under study in several controlled clinical trials in the U.S. and China. Why “controlled?”  Shouldn’t everyone be getting it?  Actually, no, for two reasons.

The first is what we just discussed—for people who have a very good chance of recovery without the drug, we have to balance potential risks against potential benefits. When people are likely to recover, we don’t want to take too many risks.

Second, given the high recovery rate, it will be very difficult to know whether the drug actually works or whether most people are just getting better, without a comparison group. Suppose the recovery rate of hospitalized patients with COVID-19 is 80 percent (it may well be higher or even a little bit lower, data coming from different hospitals is varied). Suppose we treat the next 100 such patients with a new drug, and 90 percent recover. Can we conclude that the new drug is saving some patients? We can’t—the better survival rate could be due to the fact that the patients we treated were identified earlier in their disease course, or that they had fewer underlying medical conditions, or that they were younger, or it could just be random chance. In order to have any confidence about the effects of the drug, we need to compare outcomes in people who did and did not get the drug, and who were treated at the same time, with the same background care. The optimal way to do that is in a randomized clinical trial, a number of which are currently ongoing and may actually yield reliable information quite soon.

We have been here before. I worked with Dr. Anthony Fauci at the National Institutes of Health in the early days of AIDS, when many HIV-infected people demanded access to drugs for which there was little to no good data on their effects. This population faced a very high risk of rapid death, unlike the COVID-19 patients—yet some of them recognized that the best hope to control this horrible infection was the conduct of rigorous controlled studies that would actually show what worked and what didn’t. That approach ultimately led to the discovery of drugs that changed AIDS from a death sentence to a chronic disease but one that would not necessarily shorten life.

Those hospitalized with COVID-19 cannot advocate for themselves the way AIDS patients did,  demanding good clinical trials that can identify useful treatments. It’s up to the rest of us to push back on the notion that widespread use of unproven treatments for COVID-19 is a good idea.