COVID Treatment's Positive Initial Test Results Contain Troubling Wrinkles Upon Examination

Positive results from trials are a hopeful step for a COVID-19 monoclonal antibody cocktail, but some elements of the data cast some doubt on the efficacy and practicality of the treatment among the initial cohort measured. The results, announced Tuesday by pharmaceutical company Regeneron, did suggest that their drug reduced viral loads and performed well on safety metrics.

“We are highly encouraged by the robust and consistent nature of these initial data, as well as the emerging well-tolerated safety profile, and we have begun discussing our findings with regulatory authorities while continuing our ongoing trials. In addition to having positive implications for REGN-COV2 trials and those of other antibody therapies, these data also support the promise of vaccines targeting the SARS-CoV-2 spike protein,” said George D. Yancopoulos, Regeneron’s president and chief scientific officer.

Even their most positive test group, however, did not reach the standard benchmark for statistical significance when measuring symptom alleviation. Their results for the reduction of viral load did meet or exceed the level of statistical significance. Alexandra Bowie, Regeneron's senior director of corporate communications, said in a statement that this data was prescriptive, only looking at the first 275 trial patients and not yet meant to demonstrate statistical significance. Regeneron intends to confirm these results in their next cohort.

The drug also lacked a linear relationship between dosage and reduced recovery time, with high dose patients taking longer to recover than low dose patients. A press release said with equal certainty that REGN-COV2 both reduced viral load and alleviated symptoms.

Monoclonal antibody treatments like Regeneron’s REGN-COV2 can be “eye-wateringly expensive,” especially given that they were tested on non-hospitalized patients who would likely recover fine on their own. Some investors were already commenting on this on public trading forums.

This is would be initially offset by a deal between Regeneron and the government, using $450 million in taxpayer money to provide the drugs to those who need them without charging the specific patient.

Antibody infusions are intended to be administered to patients already infected with the disease and help the body recover faster, as opposed to vaccines, which help the immune system prevent patients from becoming infected in the first place.

The drug was especially promising in patients whose immune systems had not mounted a response to the disease. Without treatment, those patients have a higher viral load to start with and almost double the average recovery time of their counterparts.

Regeneron has not yet released data on the drug's effect on hospitalized patients, meaning that the gravest COVID-19 cases were not measured in this initial data. Patients that do not need hospitalization typically recover well at home over a period of one to two weeks. Jefferies analyst Michael Yee speculated in a research note that “it is unclear, though possible, that the Regeneron cocktail could work in a hospital setting where the patient is already severely ill and has a high viral load,” Reuters reports.

Of the 2,000 patients that participated, four ‘infusion reactions’ were recorded, with three of them labeled as ‘serious adverse events.’ Two of those serious events were in placebo patients, one in a low dose patient, and none among the high dose patients. Patients were sorted evenly into those categories at the start of the trial. These are good results for safety trials, which was one of the main focuses of the original trial outline.

The efficacy of the drug was less certain. Neither the high nor low dose groups reached a p-value of 0.05, the standard benchmark for statistical significance, when measured on median time to symptom alleviation in these preliminary results. The closer to 0 a p-value is the more certain the results, with 0.05 indicating that there is a one in twenty chance that the results are a result of randomness. The results had a p-value of 0.22 among high dose patients and 0.09 in low dose patients, or a one in five and one in ten chance respectively that the results were caused by randomness. Regeneron hopes to solidify its numbers soon with a larger group than this initial batch.

The increased certainty of low dosage groups touches on another wrinkle: the relationship between recovery time and dosage was not linear. High dosage patients on average needed 8 days to alleviate symptoms, but low dosage patients only needed 6 days. While perhaps good news for producers, this could suggest that the results were not as reliable as Regeneron’s characterization. The median recovery time for placebo patients was thirteen days.

An additional hitch: production infrastructure available for this type of treatment can’t match the pace of the virus. The Washington Post reports that to make enough of the drug to treat everyone infected would require switching every producer over to COVID-19 treatments for a year, an impossible request given that most are now occupied making treatments for cancer or autoimmune diseases.

Howard Levine, national leader of BDO’s BioProcess Technology Group, told the Post: “I pray to God these therapies work. But in the short term, there is going to be a problem, I think, in making enough to treat every single patient.”

Regeneron plans to address production needs by partnering with Roche to more than triple their output capacity.

(Updated, Oct. 1, 2 p.m.)