Pfizer Inc (PFE.N) on Friday said a large trial found that its COVID-19 oral antiviral treatment Paxlovid was not effective at preventing coronavirus infection in people living with someone infected with the virus.

The trial enrolled 3,000 adults who were household contacts exposed to an individual who was experiencing symptoms and had recently tested positive for COVID-19. They were either given Paxlovid for five or 10 days or a placebo.

Those who took the five-day course were found to be 32% less likely to become infected than the placebo group. That rose to 37% with 10 days of Paxlovid. However, the results were not statistically significant and thus possibly due to chance.

Pfizer said safety data in the trial was consistent with previous studies, which had shown the pills to be nearly 90% effective at preventing hospitalization or death in COVID patients at high risk of severe illness when taken for five days shortly after symptom onset.

“While we are disappointed in the outcome of this particular study, these results do not impact the strong efficacy and safety data we’ve observed in our earlier trial for the treatment of COVID-19 patients,” Pfizer Chief Executive Albert Bourla said in a statement.

Pfizer said Paxlovid, which consists of two different antiviral drugs, is currently approved or authorized for conditional or emergency use in more than 60 countries across the globe to treat high-risk COVID-19 patients.

Sales of Paxlovid, part of a class of drugs known as protease inhibitors, are projected to reach $24 billion this year, according to Wall Street forecasts compiled by Refinitiv.

Shares of Pfizer, which fell 3% in regular trading, were down another 1% at $48.53 after hours.