On June 5, researchers at the University of Oxford quietly made a change to a late-stage clinical trial of their Covid-19 vaccine.

In an amendment noted in a document marked "confidential", they said they were adding a new group of participants.

The adjustment might seem minor in a large-scale study. But it masked a mistake that would have potentially far-reaching consequences: Many of the United Kingdom trial subjects had inadvertently been given only about a half dose of the vaccine.

The new volunteers would now receive the correct dose. The trial continued.

Much was riding on the Oxford vaccine, a British-led endeavour also involving UK drugs firm AstraZeneca.

British Prime Minister Boris Johnson's government was desperate for a success story after its early mishandling of the pandemic contributed to one of the world's highest death tolls from Covid-19 — around 65,000 by mid-December. The government has secured 100 million doses.

On Nov 23, Oxford and AstraZeneca delivered positive news. They announced that the regimen of a half dose followed by a full dose booster appeared to be 90 per cent effective in preventing Covid-19.

Two full doses scored 62pc. Oxford researchers have said they aren't certain why the half-dose regimen was much more effective.

Johnson called the vaccine team and tweeted his thanks “for their brilliant work”. He went on, “These results are incredibly encouraging and a major step forward in our fight against Covid-19.”

Oxford and AstraZeneca are now hoping for quick authorisation by Britain's regulator. But questions about the trial and the results won't go away.

Some experts say the dosing discrepancy raises doubts about the robustness of the study's findings. And they worry about another acknowledged peculiarity of the study: The half-dose regimen wasn't tested on anyone over 55 — the group considered at high risk from Covid-19.

In contrast, a vaccine produced by Pfizer/BioNTech was tested on thousands of people over 65, with an efficacy of 94pc.

John Moore, a professor of microbiology and immunology at Weill Cornell Medical College in New York, said there needed to be a better understanding of how the Oxford trial unfolded.

“When you get corporate and academic scientists saying different things, it doesn't give you the impression of confidence in what they're doing,” he told Reuters.

“Was the dosing thing a mistake or not?”

Now a Reuters review of hundreds of pages of clinical trial records, as well as interviews with scientists and industry figures, provides the most detailed account to date of what went wrong with the dosing in the Oxford/AstraZeneca vaccine study.

The review found that Oxford researchers were responsible for what their own clinical trial documents called “a potency miscalculation”.

For Oxford and AstraZeneca, the stakes could not be higher. They hope to produce up to three billion doses of the low-cost vaccine by the end of next year, enough to inoculate much of the world, including many of its poorest inhabitants.

For months, scientists at Oxford have been busily promoting the experimental vaccine's prospects in bullish terms — beginning even before the first human test subjects were injected with the experimental vaccine.

In an interview that appeared on April 11 in Britain's The Times newspaper, Sarah Gilbert, one of the vaccine's chief researchers at Oxford, said she was 80pc certain her team would be able to produce a successful vaccine, possibly as early as September. That was 12 days before a clinical trial to test its safety began.

Oxford didn't answer detailed questions for this story, but provided a statement saying the trials have been “conducted under the strict national, ethical and regulatory requirements”.

It added that “all trial protocols and trial amendments have been subject to review and approval by the relevant authorities". "All safety data have been reviewed regularly” by regulators.

A spokesman for AstraZeneca referred questions about the UK clinical trial to Oxford, which sponsored it. A spokeswoman for Britain's regulator, the Medicines and Healthcare products Regulatory Agency (MHRA), declined to answer questions about the Oxford/AstraZeneca dosing issue.

“Our rolling review is ongoing,” she said, “so this information is currently commercially confidential.” The UK's Department of Health and Social Care declined to comment.

An exact science

Vaccines are the world's best hope of ending a pandemic that has claimed more than 1.7 million lives globally. More than 60 Covid-19 vaccine candidates are currently undergoing human trials, according to the World Health Organisation (WHO).

Vaccines made by Pfizer/BioNTech and Moderna have already been authorised for use in some countries, including the United States.

They use a relatively new technology called messenger RNA (mRNA) that instructs human cells to produce antigens, a type of protein, that stimulate the immune system.

Late-stage clinical trials showed both vaccines appeared to be at least 94pc effective. Until the UK approved the Pfizer/BioNTech vaccine on Dec 2, no RNA vaccine had ever received authorisation for general use.

The Oxford/AstraZeneca Covid-19 vaccine employs a more established technique called viral vector. The vaccine uses a bioengineered version of a harmless common-cold virus found in chimpanzees to instruct human cells to make antigens.

On Dec 8, Oxford published an interim analysis of its trial results and more than 1,100 pages of supplementary documents in the scientific journal The Lancet.

These show that measuring the concentration of viral materials can be tricky, and they shed light on the chain of events leading up to the dosing discrepancy.

Viral vectors are typically produced in bioreactors holding up to 2,000 litres, which are then filtered and purified into a concentrated batch of active substance of just a few liters.

“This is not about measuring pencils, pens, bricks, or fixed objects of a given size,” said Lucio Rovati, chief executive of Rottapharm, an Italian biotech company that is trying to develop a different type of vaccine using a genetic fragment.

“It's about live biological products.”

According to the Oxford documents, in May researchers received a vaccine delivery from Italy's IRBM/Advent one of the contract manufacturers Oxford enlisted to complement its own vaccine production. The late-stage trial of the Oxford vaccine was about to begin.

The shipment, batch K.0011, had undergone the Italian company's quality check using an established genetic test — quantitative PCR, or qPCR — to determine viral matter per millilitre.

Oxford ran its own analysis for good measure. The university had been using a different method known as spectrophotometry, which measures viral material in liquids based on how much ultraviolet light the viral matter absorbs.

Oxford's measurement showed that the batch was more potent than the Italian manufacturer had found, the documents show. Oxford trusted its own result and wanted to remain consistent with a measuring tool it had used throughout an earlier trial phase.

So it asked Britain's drugs regulator for permission to reduce the volume of vaccine injected into trial participants from the K.0011 batch. Permission was granted.

“The decisions about dosing were all done in discussion with the regulator. So when we started the trial, we had some discrepancies in the measurement of the concentration of virus in the vaccine,” Andrew Pollard, the Oxford trial's chief investigator, told Reuters.

A spokeswoman for the regulator declined to discuss when it first became aware of the dosing issue.

'Serendipity'

Trial participants who received shots from the Italian batch displayed milder than normal expected side effects, such as fever and fatigue.

AstraZeneca Executive Vice President Mene Pangalos said the dose was measured incorrectly.

“It ended up being half the dose,” he told Reuters. He called the mistake “serendipity,” given that data analysis later indicated the half dose, followed by a full dose booster shot, was much more effective than two full doses.

He also recently told the BBC: “There is no doubt I think that we would have run the study a little bit differently if we had been doing it from scratch. But ultimately, it is what it is.”

IRBM/Advent told Reuters there was no manufacturing issue with the batch.

The company said in a written statement that the measuring mishap was “the result of a change in the testing method” used to confirm the potency of the dose “once the material had been shipped".

The documents published in The Lancet confirm that the error lay with the Oxford researchers. A common emulsifier, polysorbate 80, used in vaccines to facilitate mixing, had interfered with the ultraviolet-light meter that measures the quantity of viral material, according to the documents.

As a result, the vaccine's viral concentration was overstated and Oxford ended up administering half doses of vaccine, believing they were full doses.

The documents don't give a detailed timeline, nor do they show that Oxford informed AstraZeneca at the time. But they do show that Oxford contacted the British regulator again, this time seeking approval to change its measuring method to the one used by the Italians, and to figure out how to proceed with a late-stage trial that had begun with participants receiving the wrong dose.

The documents don't provide full details of the communication between Oxford and the regulator.

In early June, the regulator gave the green light to continue injecting people with the half-dose, in an effort to keep the trial as large as possible and to try to speed up the results, according to the analysis published in The Lancet.

The regulator also ruled that Oxford must add another test group to receive the full dose, in accordance with its original plan to test the safety and efficacy of full doses.

“We went back and discussed it again with the regulators and agreed with them,” said Pollard, the Oxford study's chief investigator.

Ultimately, 1,367 trial participants — none of them over 55 — received the half-dose/full-dose regimen. Two full doses were given to 4,440 adult participants, from all age groups.

The MHRA, Britain's regulator, is expected to decide soon whether to approve the vaccine.

The agency is headed by June Raine, a doctor who trained in general medicine at Oxford. The university's website shows she has made donations, given talks and performed volunteer work for the university's Somerville College, which she attended.

The MHRA said that before any decision on the Oxford/AstraZeneca vaccine is made, Raine “will ensure, for complete transparency” her interactions with Oxford as an alumnus are declared.

It added that “none of these ties are of a nature that could give rise to a conflict requiring recusal”.

Contradictions

Deep within the more than 1,100 pages of supplemental appendices published in The Lancet appeared a description of the dosing discrepancy “a potency miscalculation”.

That admission is contained in a “Statistical Analysis Plan” by Oxford and AstraZeneca dated Nov 17.

Six days later, Oxford and AstraZeneca announced the interim results of their clinical trials in the UK and in Brazil. “Oxford University breakthrough on global Covid-19 vaccine,” was the headline of an Oxford press release.

AstraZeneca's news release was more muted. “Two different dosing regimens demonstrated efficacy with one showing a better profile,” it stated.

In interviews about the results with Reuters and the New York Times, AstraZeneca's Pangalos spoke of “serendipity,” a “useful mistake” and a “dosing error”.

But the firm's chief executive officer, Pascal Soriot, told Bloomberg: “People call it a mistake it was not a mistake.” A spokesman for AstraZeneca declined to comment on the statements.

Meanwhile, the two scientists leading Oxford's development of the vaccine Sarah Gilbert and Adrian Hill suggested that the half-dose was not administered by mistake.

They didn't provide evidence.

Gilbert, an Oxford vaccinology professor, said it was normal for researchers to look at different dose levels during vaccine trials. “It wasn't a mix-up in dosing,” she told the Financial Times in an article published on Nov 27.

A few days later, Hill told Reuters it was a conscious decision by researchers to administer a lower dose. “There had been some confusion suggesting that we didn't know we were giving a half dose when we gave it that is really not true,” he said.

Gilbert and Hill together have about a 10pc stake in a private biotech firm called Vaccitech that was spun out of Oxford University, according to a filing with Companies House, the UK's companies registry, dated Oct 29.

According to a spokeswoman for Vaccitech, the company transferred its rights to the vaccine to Oxford University's research commercialisation arm in exchange for a share of the revenue. “If the vaccine is successful then all shareholders and investors in the company could potentially indirectly benefit,” she wrote in an email.

Hill and Gilbert didn't respond to detailed questions for this article.

The conflicting explanations of what went wrong have drawn criticism from some experts.

“Personally, I can say that I think their vaccine is much better than their communication,” said Guido Rasi, who until last month was executive director of the European Medicines Agency, the European Union's regulator. He said the agency eventually will evaluate the trial data.

'Ahead of the world'

For months, the Oxford/AstraZeneca vaccine was described by officials and in the media as the front-runner in the global race to produce a Covid-19 vaccine.

Britain's health minister, Matt Hancock, told a press conference in April that Britain was “at the forefront of the global effort” to find a vaccine.

On June 26, the WHO's chief scientist, Soumya Swaminathan, said at a press conference that Oxford's vaccine was probably the world's leading candidate.

Five days later, the then head of Britain's vaccine procurement programme, Kate Bingham, told a parliamentary committee: “Oxford is ahead of the world in that it is the most advanced vaccine anywhere.”

Some Oxford scientists did little to dampen the enthusiasm. Asked at the same science and technology parliamentary committee on July 1 whether the world would have to struggle through the coming winter without a vaccine, Gilbert said, “I hope we can improve on those timelines and come to your rescue.”

At the end of July, she alluded to competing vaccine efforts.

In an interview on the Royal Society of Biology's website, she said of the Oxford/AstraZeneca vaccine: “If this doesn't work, I don't think anything will work.”

Her main partner on the project, Hill, was equally bullish. On May 15, he told Reuters the Oxford/AstraZeneca candidate is “almost certainly the best single dose rapid-response vaccine”.

He dismissed as “total unknowns” and a “wild card” the vaccines using mRNA technology, such as Pfizer/BioNTech and Moderna, which have since published results showing they were both at least 94pc effective at preventing Covid-19.

The Pfizer/BioNTech vaccine is already being distributed in the millions in the United States and the UK.

“Why would you take a vaccine technology that is new, unproven, maybe quick to manufacture, but expensive to manufacture — and has never been scaled up and has never been shown to protect against anything in humans, and prioritise that in a global emergency?” he asked.

“It's very odd.”

Ian Jones, a professor of virology at Britain's University of Reading, told Reuters that the plethora of upbeat statements hasn't benefited the Oxford/AstraZeneca vaccine candidate.

“I don't want to take away from the fact everybody has worked very hard and (the vaccine) is fundamentally safe and sound,” he said.

“But reporting has always had a slightly nationalistic tone, which I don't think has been helpful.”