The Washington PostDemocracy Dies in Darkness

New results from a once-promising therapy show the difficulty of treating covid-19

In one study, the drug, tocilizumab, reduced risk of death in ICU patients, but two others found negligible results with slightly different patients

The logo of Swiss drugmaker Roche at its headquarters in Basel, Switzerland. (Arnd Wiegmann/Reuters)
Placeholder while article actions load

An arthritis drug that showed promise to treat coronavirus infections — at least in small studies early in the pandemic — did not produce consistent results across three larger trials, according to studies published Tuesday in JAMA Internal Medicine.

The drug, named tocilizumab, blocks specific immune molecules from triggering inflammation. Clinical observations in the outbreak’s first months suggested the drug could tamp down the hyperactive, and sometimes fatal, immune reactions in critically ill patients with covid-19, the disease the novel coronavirus causes.

But those observations were limited by tiny sizes. One of the first studies to highlight tocilizumab’s potential, conducted in February, included only 21 patients. Those results have not been broadly replicated — a testament to the difficulty of gauging a treatment’s effectiveness on the fly when presented with diverse patients at different stages of illness.

Of the trio of new studies, the largest — an observational study involving about 4,000 people — compared patients who received tocilizumab within 48 hours of being admitted to an intensive care unit against those who did not. People treated with the drug were less likely to die, with a mortality rate of 28 percent versus 37 percent, the study authors found.

“That’s a pretty significant difference,” said study author Shruti Gupta, a physician at Brigham and Women’s Hospital in Boston.

The other two new tocilizumab studies were randomized, controlled trials — considered the gold standard for clinical research because patients are allocated a treatment by chance to eliminate confounding factors. One trial, which took place in 24 hospitals across Italy, compared 60 people who received the drug and 63 people who didn’t. Tocilizumab did not benefit the patients in that trial, who received oxygen but were not sick enough to be admitted to an ICU.

The second involved a similar number of patients in France, who had moderate or severe covid-19 pneumonia but did not need mechanical ventilation. The risk of death at two weeks was smaller in the patient group treated with tocilizumab, but at Day 28, the researchers observed no difference in mortality.

Brigham and Women’s Hospital physician David Leaf, an author of the observational study, said timing and disease severity may explain some of the inconsistency in the results. Leaf, Gupta and their co-authors focused on patients who received the drug relatively early.

“You want to give it before irreversible organ injury has occurred,” said Leaf, also an assistant professor at Harvard Medical School. The patient population was also more severely ill in that study than in the two randomized trials.

The reports from France and Italy provide the first available peer-reviewed results for randomized trials involving tocilizumab, said Jonathan Parr, an infectious-diseases physician at the University of North Carolina at Chapel Hill, who was not involved with the studies. Parr said the observational study was rigorous, but he gave more weight to the randomized trials.

“We have been eagerly awaiting studies like those coming out today,” he said.

Drugmaker Roche, which manufactures the drug as Actemra, had announced via a news release in July that a Phase 3 trial showed tocilizumab did not help patients with severe covid-19 pneumonia. Mortality at four weeks was not improved, either. Another randomized trial by Roche appeared to show positive results, suggesting that patients with covid-19 who took the drug were less likely to die or require mechanical ventilation than patients given a placebo.

But because the peer-reviewed results from the Roche trials aren’t yet available, “it’s difficult to assess those studies,” Parr said.

Early in the pandemic, Parr and his colleagues treated a handful of very sick covid-19 patients with tocilizumab. The doctors stopped once they began to see “different patterns than what had been described in initial case reports,” Parr said.

Parr said he does not yet plan to resume using the drug.

“My conclusion was that the results were mixed and not convincing enough for me to change my practice,” he said. “I’m going to wait until more compelling results are available.”

He should not have to wait long to judge whether new evidence is compelling. Several other clinical trials are investigating tocilizumab and related anti-inflammatory drugs.

Read more:

Steroids can save lives of patients with severe covid-19, earning WHO endorsement

These laboratory-made antibodies are a best bet for a coronavirus treatment, but there won’t be enough

‘Nobody has very clear answers for them’: Doctors search for treatments for covid-19 long-haulers