Subscribe
Freezers used to hold doses of the Pfizer coronavirus vaccine at minus-70 degrees Celsius at the Kalamazoo, Mich., facility.

Freezers used to hold doses of the Pfizer coronavirus vaccine at minus-70 degrees Celsius at the Kalamazoo, Mich., facility. (Evan Cobb/for The Washington Post)

PORTAGE, Mich. — The first attempt to produce industrial-scale quantities of the experimental vaccine that has played a central role in arresting the coronavirus pandemic in the United States was a total failure.

Operators at a Pfizer plant outside Kalamazoo hoped the trial run could provide quick validation of the company’s gamble on a newfangled mRNA technology. It also was an early test of Pfizer’s strategy of refusing government aid to develop and rapidly ramp up commercial scale production of its vaccine.

But as soon as pressure built in the array of pumps, tubes and tanks at the plant on Sept. 11, it became clear something was wrong, Pfizer team leaders told The Washington Post in the most detailed public accounting of the company’s efforts to produce mass quantities of vaccine.

When operators checked a vat at the end of the production run, most of the key ingredient — the fat molecule encasing the messenger RNA — was missing.

“Our first engineering trial ... was an absolute and utter failure,” said Pat McEvoy, Pfizer’s senior director of operations and engineering at the Kalamazoo plant.

Pfizer had evidence the mRNA vaccine could stop the virus. The vaccine is composed of a lipid nanoparticle, a sphere of fat molecules encapsulating a strand of messenger RNA, which instructs human cells to make proteins that trigger antibodies and prime the immune system against future viral invasions.

But it would do little good unless Pfizer could rapidly take the new nanoparticle technology from lab to mass production — a feat never before accomplished — making immunizations available in America and around the world.

The company and its vaccine partner BioNTech would ultimately master the job of churning out large batches of mRNA vaccine, making it the clearest winner among drug companies to emerge from the pandemic. The company is producing vaccine in greater quantities than any other company and has secured an advantage in the quest to use next generation mRNA technology for treatments of other diseases.

But the failed September test shows that success was far from a foregone conclusion. Pfizer’s ambitious production objective ultimately forced the company to accept government help to procure vital supplies, defeating its earlier efforts to avoid a closer partnership with federal health authorities.

The company also has been criticized for selling most of its early supply to the wealthiest nations, which paid the highest prices, fueling severe global inequities. Pfizer was able to fill President Biden’s order this month for 500 million doses that will be distributed to low-income countries, but advocates point out that pledge is still just a fraction of global need.

“The weight of the world was on us. We have the manufacturing capability for a solution to the pandemic, and we knew we couldn’t go fast enough,” said Chaz Calitri, Pfizer’s vice president of operations for injectable drugs in the United States and Europe.

Pfizer’s spigot opened wide in March of this year, stemming the toll of death in the United States and allowing the economy to enter a strong rebound. The breakthroughs are fueling investment in the new technology.

The company says it expects to make enough for 3 billion shots in 2021, twice as much as initial projections and enough of the two-dose immunization for 1.5 billion people. It has said it will make $26 billion in vaccine sales in 2021, which would make it the biggest-selling medicine ever.

Pfizer and other companies are building mRNA vaccine pipelines for influenza, HIV, tuberculosis, rabies, rotavirus, malaria, and Zika, according to an analysis by the investment advisory firm Berenberg Capital Markets. BioNTech and Moderna are conducting extensive work on the use of mRNA against cancer.

“This is about as disruptive as it gets when it comes to a new medicine,” said Pieter Cullis, a researcher at the University of British Columbia who helped pioneer the use of lipid nanoparticles to deliver drugs into the human body. “It’s a fantastic time.”

- - -

What looks like success now was not clear at all on March 20, 2020, when McEvoy received an email from his boss, Calitri.

Three days earlier, Pfizer had announced its partnership with the German biotech company BioNTech to develop and manufacture its experimental coronavirus vaccine. BioNTech had already identified how it would make an mRNA vaccine to fight coronavirus, but it needed a big production partner with the engineering and distribution expertise to make vaccine on a global scale.

Calitri was among Pfizer executives who picked the Kalamazoo plant and McEvoy’s team to formulate doses, fill vials, and ship boxes of vaccine packed with dry ice.

“I confirmed you a moment ago. Are you good with this?” Calitri wrote to McEvoy.

McEvoy’s simple reply came seven minutes later: “Chaz, 100 percent.”

It would require a huge commitment of resources and workers in Kalamazoo, where drug manufacturing began in the 19th century under Upjohn, a company name that is intertwined with the community’s history.

Through mergers, Pfizer in 2003 acquired UpJohn’s 1,300-acre manufacturing campus, built in 1948 south of town, out past the county airport. It has 2,800 employees and mostly churns out generic drugs and bulk pharmaceutical ingredients, from birth control injections to coral snake antivenom. It is one of the largest manufacturers of steroids in the world. The building where most of the coronavirus vaccine production happens is a quarter-mile long.

As he assembled a team to lead a Michigan vaccine workforce that now numbers 600 and is expected to reach 1,000 employees, McEvoy learned that key production steps would have to be built from scratch. Large machines to mix together lipid nanoparticles and filter the finished product did not exist.

“We built this out of the Erector Set we had,” McEvoy said.

In July, the same month Pfizer cut its first deal to sell the United States 100 million doses of vaccine for $1.95 billion, McEvoy and other Pfizer executives brought the leaders of then-President Donald Trump’s Operation Warp Speed for a factory tour. McEvoy showed Moncef Slaoui, an experienced pharma executive recruited by Trump to advise OWS, and Army Gen. Gustave Perna, who managed OWS logistics, a cavernous empty space.

Here’s where the hundreds of freezers will maintain finished vials of vaccine at minus-70 degrees Celsius, McEvoy told them. Here’s where the vaccine will be packed. Prefabricated manufacturing suites would be rolled in off trucks from a Texas contractor and filled with gleaming new vats, pipes, and pumps.

Slaoui, in an interview, said he came away impressed with Pfizer’s commitment. It was apparent Pfizer planned to use its enormous, global size, vast stores of cash, and swarms of engineers in a “bulldozer, brute-force” strategy to make billions of mRNA-ferrying nanoparticles, he said.

Pfizer ordered every special freezer it could get from supplier Thermo Fisher, and then asked Thermo Fisher to make more. It built a vast system of duct work to suck heat generated by the freezers out of the building. Because of the Antarctic chilling requirements, workers had only 46 hours to get finished vaccine into six-dose vials and then into the freezers before the vaccine could spoil. Managers choreographed a relay race, with handoffs of vaccine containers outside the biosafety airlocks.

A few false starts were inevitable. In one especially costly one, Pfizer ordered special blast freezers from Europe that it ended up never using. They’re sitting in storage.

Engineers decided time was too short to design, build and test a single, industrial-scale mixer to fuse the mRNA into the lipid nanoparticles.

So they ordered multiple small jet mixers, each one about the size of a stack of silver dollars, and arranged them in a series of eight parallel systems. Inside each little disk, two jets of solution are forced through a tunnel about the width of a human hair and sent into a head-on collision at 450 pounds per square inch of pressure.

BioNTech had provided Pfizer with the blueprints for its lab-size mixing system, said Jinne Adisoejoso, a Pfizer drug manufacturing specialist from Belgium who traveled to help the Michigan team in January.

“They gave us a drawing of what they had. This would never be able to supply the entire world. It was literally sitting on a table,” he said. “We had no expertise at that point about this equipment, and we had to develop a way to make it commercially viable.”

One of Pfizer's vaccine formulation facility rooms.

One of Pfizer's vaccine formulation facility rooms. (Evan Cobb/for The Washington Post)

Silos filled with liquid carbon dioxide for making dry ice are next to Pfizer's manufacturing facility in Kalamazoo, Mich.

Silos filled with liquid carbon dioxide for making dry ice are next to Pfizer's manufacturing facility in Kalamazoo, Mich. (Evan Cobb/for The Washington Post)

Pfizer freeze operation technicians load vaccine doses into a freezer at the Kalamazoo facility. MUST

Pfizer freeze operation technicians load vaccine doses into a freezer at the Kalamazoo facility. MUST (Evan Cobb/for The Washington Post)

Melissa French, a project manager at Pfizer, was called into work to help the company produce its own supplies of a rare but vital ingredient for a vaccine nanoparticle.

Melissa French, a project manager at Pfizer, was called into work to help the company produce its own supplies of a rare but vital ingredient for a vaccine nanoparticle. (Evan Cobb/for The Washington Post)

- - -

After the first engineering test flopped in September, Pfizer engineers traced the problem to a failed membrane that allowed the precious lipid nanoparticles to leak through during a filtering step. So Pfizer developed an integrity test for each membrane. In Puurs, Belgium, the Pfizer production team adopted the lessons learned at Kalamazoo and conducted a successful engineering run on Sept. 14.

But the cumbersome filtering machines would continue to slow down production. Pfizer added more filter area to increase throughput and double batch size, from 1.7 million to more than 3 million doses. Later, it devised a system to regenerate used filters, because of severe shortages.

As the team solved these problems, Pfizer came out with positive news from clinical trials. The company said on Nov. 9 that its vaccine was shown to be more than 90% effective at preventing COVID-19, the disease caused by the novel coronavirus.

The scientific watershed put an end to the pandemic in sight for the first time. But it put even greater pressure on manufacturing engineers to produce a lot more vaccine. Buried in the same news release of the clinical trial results, Pfizer disclosed that it had cut its projected output by the end of 2020 to 50 million from 100 million.

Pfizer said manufacturing problems and raw materials supply were not the reason. It said in a statement it slashed the goal because the clinical trial results came later than anticipated, which did not leave it enough time to ship all the promised doses by the end of the year.

“In the beginning it was cool and fun. We were in this race. It was our moonshot, and it was really exciting,” said Calitri, Pfizer’s vice president of operations.

“When we got the clinical data readout, November 9, it was amazing,” he said. “It has been a stressful job, and it got more stressful as it went along.”

Two days after the FDA’s Dec. 11 emergency use authorization, cardboard boxes packed with vaccine and dry ice began to leave the Kalamazoo loading dock, part of an initial dispatch of 2.9 million doses around the country.

“On December 13th, we were out there clapping and celebrating as the trucks were waiting for the first lot,” McEvoy said. “We came in with the senior execs that were here, and they said, ‘Pat, we have to find a way to make four times the production. And we need an answer next week.’ “

Two days later, project manager Melissa French, who started working at the plant 27 years ago when she was a biology student at Western Michigan University, was called into work on the vaccine project to help Pfizer produce its own supplies of a rare but vital ingredient for a vaccine nanoparticle.

Kalamazoo was the only site with specialized chromatography equipment that could produce the ingredient, a special fatty compound called ionizable cationic lipid, at large scale. French told her two children and two stepchildren — all under 11 — that her planned two-week Christmas break was canceled.

As she walked chemists and engineers through the complex challenge in a Dec. 15 kickoff meeting, French’s emotions surged as she read the bullet points out loud on her presentation.

“I cried a little. I choked up,” French said. “It looks boring, and then there’s a simple slide that says ‘Why is this so important? Because the world needs us, and people are dying of COVID-19.’ And to say it out loud, when we needed that vaccine, and we were all cooped up at home . . .”

French led a team of 40 people who swarmed the problem from all angles, designing and assembling the equipment in a little over a month. It required exhaustive safety reviews. Highly flammable solvents are used to make lipids, so cellphones must be encased in special covers to prevent errant sparks. By Jan. 20 they began the complex, multistep process of preparing the first batch.

The ionizable cationic lipid is the linchpin of a nanoparticle. Of four lipids required, it is used in the greatest quantities. It changes its electrical charge when it enters a human cell, opening the nanoparticle and releasing the mRNA payload. Without it, a nanoparticle vaccine will not work.

“When we put product in the tank, there was an audience,” she said. “A bunch of the chemists came, and these floor operators were saying, ‘Why do we have an audience?’ “

- - -

As the Kalamazoo plant buzzed with intense effort, Operation Warp Speed officials were growing frustrated with a lack of information from Pfizer about its manufacturing problems and supply issues. Pfizer was under no obligation to share information with the government because it did not accept U.S. research and development money.

“We didn’t have the full visibility and understanding of what challenges Pfizer was experiencing in their processing,” Robert Kadlec, the former assistant secretary for preparedness and response, who helped lead vaccine procurement in the Trump administration, said in an interview. He contrasted Pfizer with other companies, such as Moderna and Johnson & Johnson, which worked more closely with the government. “Pfizer was not in that circle of interest, originally,” Kadlec said.

In December, once Pfizer had shown its vaccine would be part of the solution to the pandemic, the Trump administration was asking the company to exercise an option for an additional 100 million doses, beyond the original 100 million-dose order.

By that time, the Trump administration (and subsequently, the Biden administration) did have something Pfizer wanted. The White House could use the Defense Production Act — a Korean War-era law giving the government broad powers to control industrial output in a national emergency — to help Pfizer get priority access to raw materials and equipment.

On Dec. 22, the Trump administration granted a DPA listing for Pfizer. Pfizer announced that it had agreed to sell the United States an additional 100 million doses of its vaccine.

“The reality is they needed help,” said Slaoui. Pfizer competitors’ use of the DPA were placing the company at a disadvantage.

“Other manufacturers were able to access ingredients and equipment at the expense of Pfizer. We were using the DPA to get lipids for Moderna and sterile bags for X, Y, Z, and Pfizer would find themselves at the bottom of the queue,” he said.

Seven weeks later, the Biden administration announced it was using the DPA to help Pfizer again, to prioritize its purchase of filling pumps and filtering machines.

Adding more filtering capacity was critical to boosting production volume, Pfizer engineers said. So was stepping up to 280 liter stainless steel vats from 140 liters. The modular manufacturing suites shipped in from Texas are allowing yet another jump in capacity, with two sets of matching machinery and 320-liter vats.

“We have found a great ally in the Biden administration,” Pfizer CEO Albert Bourla declared when President Joe Biden toured the Kalamazoo plant on Feb. 19, reading from prepared remarks while standing in front of Pfizer’s vaccine-making machines.

Biden hailed his own use of the DPA to secure pumps and filters for Pfizer. He toured the factory’s freezer farm. He watched workers fill shipments to states using dry ice from Pfizer’s new dry ice production facility. Pfizer had found it could not purchase enough dry ice from its suppliers, so it built its own, with five CO2 tanks towering just outside the plant walls.

After the “miracle of science” to develop the mRNA vaccine, Biden said, the Pfizer accomplishments in Kalamazoo were “a second miracle — a miracle of manufacturing — to produce hundreds of millions of doses.”

- - -

In Kalamazoo’s quiet downtown, mask mandates have been lifted. Restaurants are open for indoor dining, although some ask that you wear masks when you’re not at your table. The Radisson Hotel has posted signs saying masks are required, unless you’ve been vaccinated. Kalamazoo County has among Michigan’s highest rates of vaccinated people over age 65, at 82 percent.

Pfizer’s team leaders and engineers who worked on the vaccine while juggling remote schooling for their children, working over Zoom from their homes, observing state social distancing mandates and quarantine periods when they had no choice but to travel, are now emerging from pandemic restrictions as well, along with their families.

McEvoy, the director of operations, described his father, Daniel J. McEvoy, who lives in Delaware County, Ind., 160 miles south of Kalamazoo. He is a widower and suffers from chronic obstructive pulmonary disease, which put him at high risk of dying if he contracted COVID-19. At 83 years old, he qualified in Indiana’s first wave of people to get the vaccine.

“He’s isolating. My mother died the year prior, and he’s all alone,” McEvoy said. McEvoy signed his father up online to get the vaccine on Jan. 12, at a hospital two miles from his home.

“He drives in, stands in line, socially distanced. He wears his mask, and he gets up there, and they said, ‘Unfortunately, we’ve reached our quota, we’re out of the Pfizer vaccine. We can only give you Moderna.’ “

“I had explained to him to get whatever vaccine is available because if you don’t, you’re going to die.” But his father ignored his advice once he got to the front of the line, said McEvoy, his voice breaking as he recounts what his father did next.

“He whips out my card and says, ‘My son is working here making the vaccine,’ and he has the little article [with a picture] of us from a month earlier, launching the vaccine. ‘That’s him right there.’ “

His father left the hospital without getting a shot. Later that evening, the hospital asked him to return.

“He called me back and he says, ‘I got the Pfizer vaccine.’ It’s the happiest day I’ve had.”

Sign Up for Daily Headlines

Sign up to receive a daily email of today's top military news stories from Stars and Stripes and top news outlets from around the world.

Sign Up Now