AstraZeneca Plc and the University of Oxford face mounting questions about their COVID-19 vaccine trial results after acknowledging a manufacturing error.

While an announcement earlier this week by Astra and Oxford showed their shot was 70 per cent effective on average in a late-stage study, the scant details released by the U.K. partners have sparked worries, with some expressing doubts about whether U.S. regulators would clear it.

Astra and Oxford had said their vaccine was 90 per cent effective when a half-dose was given before a full-dose booster. Two full doses showed an efficacy of 62 per cent. But the head of the U.S. vaccine program known as Operation Warp Speed said the next day that the dose showing the higher level of effectiveness was tested in a younger population.

In a statement, Oxford said a difference in manufacturing processes led to later phases of the trials having a half-dose given instead of a full one. When it was apparent that a lower dose was used, it was discussed with regulators, and an agreement was reached to push ahead with the two regimens, the statement said.

“The methods for measuring the concentration are now established and we can ensure that all batches of vaccine are now equivalent,” according to the university.

Astra shares rose 0.5 per cent Thursday after slumping 6.2 per cent earlier in the week.

“The most likely explanation for the divergent efficacy in its interim analysis is either chance or patient demographics,” Sam Fazeli, a Bloomberg Intelligence analyst, wrote in a note. “Either way, approval based on current data means people will be inoculated with a vaccine the true efficacy of which is unknown.”

A spokesman for Astra said the trials were conducted “to the highest standards” and more analysis is being done to refine the efficacy reading.