An Eli Lilly drug for rheumatoid arthritis carries a warning on its label saying patients with infections should not take it because it can make infections worse. Yet the National Institutes of Health is about to test it in people hospitalized with coronavirus infections.
The study, whose innovative design is meant to find out — fast — what works, began at the end of February with the antiviral drug remdesivir made by Gilead Sciences. Four hundred patients have been treated either with remdesivir or a placebo. The results are now being analyzed and will be known within a few weeks.
Then the study will move on to baricitinib, made by Eli Lilly and Company, the company said.
Jennifer Routh, a spokeswoman at the National Institute of Allergy and Infectious Diseases, confirmed that the drug would now be tested in the federal trial but said the institute could offer no further comment.
Dan Skovronsky, chief scientific officer at Lilly, explained how and why baricitinib was chosen.
In February, when the new coronavirus was emerging as a pandemic threat, a company in the United Kingdom called Benevolent AI began using its artificial intelligence system to look for approved drugs that could possibly help people with coronavirus infections. It pointed toward baricitinib precisely because it suppresses the immune system. That, the company suggested, might allow it to quell a cytokine storm, a disastrous immune system response that kills patients.
As a coronavirus infection progresses, the amount of the virus infecting cells does not appear to be extremely high. But the immune system in some people goes into overdrive, sending out vast amounts of small proteins — cytokines — that trigger inflammation. Cytokine storms can kill patients with other diseases, including flu. This immune overreaction, some scientists think, could explain why some people infected with coronavirus have only mild symptoms while others have severe or fatal illnesses.
Benevolent AI also noted another potential advantage of baricitinib, said Dr. Vincent Marconi of Emory University, a principal investigator in the federal trial. The drug might have anti-viral activity. That, plus the chance of subduing cytokine storms, Dr. Marconi said, “made a compelling case for baricitinib to be explored further in a clinical trial.”
At Lilly, executives were a bit skeptical.
“Our initial reaction was, ‘Does it make sense to immunosuppress when patients are trying to fight off an infection?’” Dr. Skovronsky said. The warning label on the drug, he added, “tempered our enthusiasm.”
As Covid-19 spread, some doctors started giving patients the drug anyway. It is a pill and there is a huge supply, making it easy for doctors to prescribe it off-label.
Dr. Skovronsky and his colleagues at Lilly were concerned.
“We are extremely cautious,” said Patrik Jonsson, Lilly’s president of biomedicines. “We cannot encourage use.”
But such warnings were not enough.
“In desperate times, doctors are trying everything,” Dr. Skovronsky said. “Various drugs are just being tried on patients in clinical trials without a control. It is really hard to interpret those kinds of data.”
The company realized it had to accept the offer to contribute its drug for the federal trial.
Dr. Andre Kalil, a principal investigator in the federal trial, urged doctors and patients to refrain from using baricitinib until the results of the federal trial are known, which should be in a matter of months.
“This is a drug that has never been used before in this situation,” he said. “That is why it needs to be tested in a randomized clinical trial. We don’t know if it will help or harm. We have so much uncertainty.”
The final design of the next phase of the federal trial is still being worked out, but the expectation is that it will include 600 to 800 patients, Dr. Marconi said. If the first phase of the study finds that remdesivir seems to help patients, half of the patients in the second phase will take remdesivir plus a placebo pill, and half will get remdesivir plus baricitinib.
If remdesivir is no better than or even worse than placebo — a very real possibility given the progress so far of a company-sponsored study in China — one group of patients will get a placebo pill and the other group will get baricitinib.
“We are looking for a strong effect,” Dr. Skovronsky said. “If it works, it will be big. If it doesn’t, we will move on.”