What is causing AstraZeneca’s vaccine production woes?

WITH Covid-19 vaccine manufacture making headlines for all the wrong reasons, I spoke with experts from the pharmaceutical engineering sector to try and shed some light on the production challenges facing AstraZeneca.

Tensions in Europe have risen after a production problem at a contract manufacturing facility used by AstraZeneca in Belgium has left the EU facing a shortfall of 60% fewer doses than it was expecting this quarter. With EU leaders under pressure because of the bloc’s relatively slow vaccine rollout, and fears spreading about “vaccine nationalism”, Belgian authorities sent inspectors to the factory to make sure the delay is due to production problems, a spokesperson told the Guardian.

AstraZeneca has yet to respond to a request for details on the specific production issues faced.

Remarkable achievement

The three experts I talked to all made a point of praising the quick progress AstraZeneca has made in developing and manufacturing a vaccine under the circumstances.

“Rather than seeing this on the negative side, they’ve done remarkably well,” said Peter Dodd, the Covid-19 Response Project Manager for the International Society for Pharmaceutical Engineering UK Affiliate (ISPE UK).

Tim Clayton, Director of CMC at biopharmaceutical firm Antev and a member of IChemE’s Covid-19 Response Team, echoed the sentiment: “Clearly there are going to be hiccups on the way. They’ve shortcut a development process that would normally take several years. So, I think they have done a remarkable job getting where they are. I don’t think we should underestimate that.”

“They’ve done a phenomenal job,” said Adam Hawthorne, a biochemical engineer and an ISPE UK Board Member.

Yet all agreed a setback was highly likely given that AstraZeneca has limited experience in viral vector vaccine development.

“Vaccines, amongst all the biologics I’ve worked on are the most sensitive processes,” said Dodd.

Clayton said slight variations in the likes of production conditions or feedstock variability can play havoc with yields.

“If you’re point one of a pH unit out, that can be enough to massively disrupt your productivity. Other factors can be cell culture medium, process timing, pH, carbon dioxide concentration, oxygen control and mixing time to name a few.”

“I worked with one process – if there was a slight overshoot on temperature because the PID loops weren’t correctly tuned, the cells would stop producing.”

Differences in culture, and culture

On top of this, AstraZeneca’s scale-out relies on contract manufacturers around the globe, which inherently adds further layers of variability: manufacturers with their own corporate cultures, working in different countries with their own languages and idiosyncratic production cultures, and relying on different suppliers for equipment and feedstocks further exacerbate the challenge of standardising production.

“There is simply the interpretation of what different human cultures mean by temperature,” said Dodd.  “It may seem absurd but whereas in the UK we would round temperature up, in other cultures they round down or quote many significant figures when the probe measures temperature at a point in a reactor that will have an element of variation across its entirety. That one degree can cause a very different result.”

Hawthorne said: “A lot of these companies have their own standard practices of bringing products onstream. Site-based standardisation is one thing. Company-wide standardisation is another. Standardisation between companies is a huge sectoral engagement activity for tech transfer alone.”

Clayton added: “I once worked on a project in Switzerland where we wanted a second manufacturing site. The site was only 30 km away. They spoke the same language. They had empty facilities that we could use, and the facility already had company-standard purified water and water for injection. The bioreactors were already being used and we just transferred them. It still took us a year to transfer a process we had been running for ten years in the same country.”

Priority actions

Asked what he thinks engineers will be doing to resolve these challenges, Clayton said: “I think their priorities will include trying to characterise what has been happening and then trying to compare data from the other manufacturing facilities.

“We used to run Kaizen events to try and brainstorm all the different factors and identify what might be different. You bring in as much information as possible. But it depends on how each company deals with it.”

“It’s data processing – looking for facts and continuous improvement,” Dodd said.

“But what you can’t do is change everything at the same time,” said Clayton. “You have to go through sequentially and try to find the factors that will bring you closer to that target.”

Dodd added: “A batch of vaccine product, once it’s made, has something like ten thousand pieces of data that’s gone into it and been recorded. These are not simple cake recipes. These are very complicated processes.

But on top of having all that data to select from, AstraZeneca will have a shortfall of the sort of data that would enable it to better tweak its processes had it not been rushing to develop its process amid a pandemic.

“With a normal process where you’ve had five years for clinical development, you would have much more lab data and know what the key factors are and the response.  The responses would have been characterised, which takes time.  Whereas in the hurry to develop this vaccine that data is probably rather thin, though sufficient for emergency use,” said Dodd.

The last word goes to AstraZeneca CEO Pascal Soriot who, the Wall Street Journal reports, has said the company is three months behind fixing manufacturing “glitches” in Europe, which it has already worked out in the UK and elsewhere.