Vincent Gaynor remembers, almost to the minute, when he realized his part in birthing the breakthrough gene therapy Zolgensma had ended and the forces that turned it into the world’s most expensive drug had taken over.

It was May 2014. He and his wife were sitting in the cafeteria at Nationwide Children’s Hospital in Columbus, Ohio.

Elsewhere in the hospital, an infant — patient No. 1 in a landmark clinical trial — was receiving an IV infusion that, if it worked, would fix the genetic mutation that caused spinal muscular atrophy, a rare, incurable disease. At the time, children born with the most severe form of SMA swiftly lost their ability to move, to swallow, to breathe. Depending on the disease’s progression, most didn’t live to their second birthdays.

The Gaynors’ daughter Sophia had been diagnosed with SMA five years earlier. Since then, they’d raced to fund research to save her. Their charity, Sophia’s Cure, was covering a substantial portion of the costs of the trial.

They’d helped raise about $2 million for a program at Nationwide run by Brian Kaspar, a leading researcher. Gaynor, a New York City construction worker, had forged a tight bond with Kaspar, speaking frequently with him by phone, sometimes deep into the night.

But their relationship had started to fray when — with success in sight — Kaspar became part owner of AveXis, a biotech startup that had snapped up the rights to his SMA drug. Billions of dollars were at stake.

When Kaspar walked into the cafeteria that day, Gaynor said, the scientist didn’t acknowledge him or his wife before sitting down a short distance away. Neither did the man with him, the startup’s CEO.

“It was like they didn’t know us,” Gaynor recalled.

When Zolgensma hit the market five years later, it was hailed as a miracle drug. Some babies treated with it grew up able to run and play. It helped reduce U.S. death rates from SMA, long the leading genetic cause of infant mortality, by two-thirds.

That leap forward came at a sky-high price: more than $2 million per dose, making Zolgensma then the costliest one-time treatment ever.

How did a drug rooted, like many, in seed money from the U.S. government — that is, American taxpayers — and spurred by the grassroots fundraising of desperate parents, end up with such a price tag?

The story of Zolgensma lays bare a confounding reality about modern drug development, in which revolutionary new treatments are becoming available only to be priced out of reach for many. It’s a story that upends commonly held conceptions that high drug prices reflect huge industry investments in innovation. Most of all, it’s a story that prompts, again and again, an increasingly urgent question: Do medical advances really have to be this expensive?

ProPublica traced Zolgensma’s journey from lab to market, from the supporters there at the beginning to the hired guns brought in at the end to construct a rationale for its unprecedented price.

We found that taxpayers and private charities like Sophia’s Cure subsidized much of the science that yielded Zolgensma, providing research grants and opening the door to federal tax credits and other benefits that sped its path to approval.

Yet that support came with no conditions — financial or otherwise — for the for-profit companies that brought the drug over the finish line, particularly when it came to pricing.

Once Zolgensma’s potential was clear, early champions like the Gaynors were left behind as the private sector rushed in. AveXis’ top executives and venture-capital backers made tens or hundreds of millions of dollars apiece when the startup was swallowed by the pharmaceutical giant Novartis AG in 2018.

Wall Street analysts predicted Novartis’ new prize drug would be the first therapy to smash the million-dollar-a-dose mark. The Swiss colossus crafted a sophisticated campaign to justify more than double that amount, enlisting a team of respected academics, data-modelers and pricing strategists to help make its case.

“This was a case where the charities and the government did everything to get this thing commercialized, and then it just became an opportunity for a bunch of people to make transformative, generational wealth,” said James Love, director of the public advocacy group Knowledge Ecology International.

In a statement to ProPublica, Novartis said Zolgensma’s price reflects its benefits to children with SMA and to society more broadly.

“Zolgensma is consistently priced based on the value it provides to patients, caregivers and health systems,” the company said, adding that the drug may reduce the burden of SMA by replacing “repeat, lifelong therapies with a single treatment.”

Zolgensma’s price quickly became the standard for gene therapies. Nine of them cost more than $2 million. A tenth, approved in November, is predicted to run about $3.8 million, just shy of the most expensive, also approved last year, which costs $4.25 million a dose.

“Drug companies charge whatever they think they can get away with,” said David Mitchell, the founder of Patients For Affordable Drugs. “And every time the benchmark moves up, they think, ‘Well, we can get away with more.’”

Parents of children with SMA say their concerns about costs pale in comparison to the hope offered by such cutting-edge therapies. “I mean, it’s a child’s life,” said Hailey Weihs, who battled her health plan to get Zolgensma for her daughter. “Anybody would want that for their own child.”

The seven-figure costs of Zolgensma and other gene therapies add to the nation’s ballooning bill for prescription drugs, absorbed by all Americans in the form of rising insurance premiums and taxes for public programs like Medicaid.

Breakthroughs like Zolgensma are often framed as wins for all: Patients get life-saving new therapies. Drug companies and biotech investors make enough money to incentivize even more breakthroughs.

But not everyone’s a winner, Gaynor noted.

No one wanted Zolgensma to succeed more than he did, or understands better what it has meant for families like his. Yet his years behind the scenes of the drug’s development left him and his family disillusioned.

“I learned it’s all about money,” Gaynor said. “It’s not about saving people.”

When Vincent and Catherine Gaynor started their married life in 2006, they knew one thing for certain: They wanted children.

They learned well into Catherine’s 2008 pregnancy that they were both carriers for SMA, meaning there was a 25% chance their child would be born with the muscle-wasting disorder.

They were concerned but clung to the larger chance the baby would be born healthy.

When Sophia was born in late February 2009, at first they just marveled at her sweet disposition and bright, expressive eyes. How she loved being snuggled. How she sighed after she burped.

But it didn’t take long for Vincent, who’d grown up with younger siblings, to sense something was off. Sophia didn’t lift her legs. They flopped outward like a frog’s when he changed her diaper.

Their pediatrician assured them Sophia was fine. Then a different doctor suggested testing her for SMA. While they waited for the results, the family went to a nearby park, and Catherine pushed Sophia’s stroller around a pond. “I remember walking behind her with the video camera and knowing in my heart this was the last day we were all going to be happy,” Vincent recalled.

After Sophia’s diagnosis, Catherine quit her office job to care for the baby full time. Vincent started gulping down studies and going to conferences, desperate to find a way to save his daughter.

At the time, there were no treatments to slow or stop SMA. By the time Sophia was 4 months old, she needed a machine to help her breathe overnight. At 6 months, she could no longer take a bottle and needed a feeding tube. Each time she lost ground, their urgency to find a treatment grew.

The Gaynors didn’t have deep pockets or wealthy friends. He was a steamfitter with Local 638, from a family of steamfitters. They began raising small amounts of money by hosting golf tournaments and throwing Zumba parties. As the volume of donations grew, they founded Sophia’s Cure, emerging as serious players in the small world of SMA charities.

Vincent met Brian Kaspar at a cocktail hour for high-yield fundraisers. Kaspar was among the small group of top researchers working to find treatments for SMA, competing fiercely for recognition and funds. (Kaspar declined an interview request from ProPublica and didn’t respond to written questions.)

Because his drug was a gene therapy, public grant money and private philanthropy played an especially central role, with the National Institutes of Health alone putting over $450 million into science related to SMA. Drug companies at the time approached these treatments with more skepticism, waiting longer to invest and letting universities and academic hospitals do the heavy lifting, said Ameet Sarpatwari, an assistant professor at Harvard Medical School who studies the pharmaceutical industry.

Drug companies sponsored only 40% of the U.S. gene therapy trials active in January 2019, according to a study Sarpatwari co-authored.

“The narrative of industry is, ‘We’re doing the hard, expensive part of drug development,’ and, at least for cell and gene therapies, the most risky part is actually being done by public or federally supported labs,” Sarpatwari said, calling Zolgensma a “poster child” for the study’s findings.

By the time of the cocktail party, Kaspar had turned early research into a promising drug therapy that he was beginning to test on animals — the precursor to a human trial. Gaynor remembered him as humble and almost classically nerdy, happy to spend hours on the phone explaining how motor neurons work.

More established SMA charities tended to hedge their bets, spreading money around to multiple programs. But Sophia was already around 18 months old, and Gaynor had no time for that. In September 2010, when Sophia’s Cure won a $250,000 grant from the Pepsi Refresh Project by amassing votes online, he steered the money to Kaspar’s program. The following June, the charity signed an agreement promising Kaspar up to $1 million more, for which it had launched a drive to recruit 200 people to raise $5,000 apiece.

As the money flowed in, Gaynor and Kaspar became close friends. The Gaynors stayed overnight at Kaspar’s house on their drive to an annual charity event. Kaspar did a Q&A for the Sophia’s Cure YouTube channel from the Gaynors’ dining room and proofread posts Vincent wrote for the charity’s website.

Gaynor said they often talked about how getting the drug through the development process would require way more money and muscle than the various SMA charities could muster. Kaspar shared his conversations with venture capital firms and even asked Gaynor to talk to a potential investor.

Yet Gaynor said he was blindsided when Kaspar told him he’d formed a relationship with a Dallas startup called BioLife Cell Bank that had been focused on stem cell research.

The CEO, John Carbona, then 54, had run medical device and equipment companies, but he had no background in drug development. In an interview, Carbona told ProPublica that he took the reins at BioLife in the aftermath of his mother’s death, determined to do something “significant” to fulfill her hopes for him. After an associate’s twins were born with SMA, he said he became convinced that Kaspar’s gene therapy was the answer.

Carbona remade BioLife into AveXis: Av for adeno-associated virus serotype 9, the engine of Kaspar’s drug; ve for vector; X for the DNA helix; and Is for Isis, the goddess of children, nature and magic.

Still, for much of the next year and a half, money from charities and more than $2.5 million from the National Institutes of Health remained Kaspar’s bread and butter. In late 2012, Sophia’s Cure agreed to provide another $550,000 for a Phase 1 clinical trial. The Nationwide Children’s Hospital Foundation, an affiliate of the hospital, agreed to match it.

Kaspar singled out Sophia’s Cure for the extent of its support in a Nationwide press release.

“Sophia’s Cure Foundation has been the lead funder of this program and their incredible investment in this lab has accelerated our program by many years,” he said.

The trial protocol called for Kaspar’s therapy to be tested on infants up to 9 months old. It was a pragmatic decision: The company had limited funds and capacity to produce the test doses, which would be smaller for children who weighed less. Plus, the youngest children were likely to show the most dramatic results since they’d be treated before SMA inflicted its worst damage.

That left out Sophia, as well as most of the kids whose parents made up Gaynor’s fundraising network.

Gaynor’s dream of saving his daughter had tapered into determination to stop the disease’s progression and preserve the strength she had left. Sophia could no longer move her whole hand, but she could still tap with her right pointer finger. She could use an eye-gaze computer to click open screens and attend school remotely. She could communicate a bit, blinking once for yes and twice for no.

Early on, Gaynor said, Kaspar had promised a trial for older kids. But Gaynor felt Kaspar’s commitment wavered as his ties to AveXis grew and his reliance on funding from Sophia’s Cure diminished.

Carbona struck a deal with Nationwide Children’s in late 2013, getting AveXis the exclusive right to develop an SMA treatment using the hospital’s inventions, including Kaspar’s, in exchange for stock. A few months later, Kaspar signed a contract that granted him an even larger stake in the company. The company also landed its first major investor, Paul Manning of PBM Capital.

Over this period, Gaynor said, the phone calls and updates from Kaspar slowed. The Gaynors were invited to Nationwide Children’s for the start of the clinical trial by the family of the child receiving the first dose.

After the initial awkwardness in the cafeteria, the Gaynors said, Kaspar and Carbona eventually came over and sat with them. Carbona remembers it differently, saying that he recalled seeing the Gaynors that day and the mood was friendly, even celebratory.

Tension surfaced two months later when they all converged in Lancaster, Wisconsin, for Avery’s Race, an annual SMA fundraiser benefiting Sophia’s Cure.

The event brought together dozens of families from across the country for an awareness walk, an auction and a rubber ducky race in a nearby creek. In the finale, parents posed questions to Kaspar, Gaynor and Carbona, almost all of them about the clinical trial.

In video footage captured by a documentary filmmaker, Catherine Gaynor asked bluntly whether testing the drug only on infants meant the FDA would approve the treatment only for the youngest patients while “everyone else is left hanging out to dry.”

Kaspar acknowledged this was possible. He described expanding the treatment to older children as “step two” but made clear that funds for testing would have to come from Sophia’s Cure.

That’s what the money raised at Avery’s Race would support, Vincent Gaynor said, adding pointedly that his nonprofit would focus on the work others would avoid “because it’s not going to push stock prices up.”

Neither Kaspar nor Carbona responded directly to the dig. Carbona, noting the company had other funding needs, said they would expand testing when they had proof the drug worked.

By early 2015, AveXis had raised millions from deep-pocketed biotech investors, adding members of several venture-capital funds to its board. Their participation would be critical in bringing the drug to market, paying for licenses to patented technology needed to make and administer it, for example. It also meant that Zolgensma had to do more than save lives — its promise had to make AveXis’ investors a profit.

Almost immediately, Carbona said, the board pushed to take the company public.

“I mean, they all have their hearts in the right place, but they’re being run by people who are looking for a return on investment,” he said.

As AveXis moved toward an initial public offering, some on the board questioned whether Carbona should continue running it, he said. He’d been accused years earlier of fraud and breach of fiduciary duty by a former employer, who won a $2.2 million court judgment against him. Carbona had denied any wrongdoing and the judgment was reversed in part and reduced on appeal, but the case left lasting damage. “It hurt me immensely,” he said.

Later that year, the board replaced Carbona with a new chief executive, Sean P. Nolan, who had a decadeslong record at pharmaceutical and biotech companies.

In September, a company representative offered the Gaynors a meeting with Nolan, saying Kaspar had stressed how instrumental Sophia’s Cure had been to the work on the drug. The Gaynors traveled into Manhattan for the meeting at a hotel bar. They told Nolan about their concerns, including that older kids wouldn’t have access to Kaspar’s drug since it hadn’t been tested on them. They said Nolan was cordial but never followed up. (Nolan didn’t respond to emailed questions from ProPublica.)

Early the following year, AveXis went public. Nolan celebrated by ringing the NASDAQ opening bell as Kaspar, other company executives and members of the board whooped and clapped.

The IPO and subsequent stock sales raised hundreds of millions of dollars, but little of the money went toward additional trials on Zolgensma, an analysis by KEI, the public advocacy group, concluded.

The drug’s trials were small, often involving two dozen patients or fewer. AveXis, and later Novartis, spent less than $12 million up to the point of the drug’s approval — surprisingly little — to prove the therapy was safe and effective, the group estimated, based on information obtained through Freedom of Information Act requests, from studies and in Securities and Exchange Commission filings. (Novartis did not respond to questions from ProPublica about trial costs.)

The companies spent more than 10 times that amount to license intellectual property from others, KEI found. It’s not the clinical trials, Love, the director, said, that “makes developing gene therapies more expensive than it has to be.”

By the time of AveXis’ IPO, the Gaynors had decided to wind down Sophia’s Cure and step back from the SMA community. In 2015, Sophia began having seizures that became more frequent over time. She was 6 years old and growing weaker. Her SMA had progressed too far for Kaspar’s drug to help her.

Vincent’s sense of failure was crushing. In September 2016, after years of pent-up anger, he took a last stab at getting Kaspar and AveXis to acknowledge that the charity and its donors had essentially been a partner in developing Zolgensma.

Sophia’s Cure sued Kaspar, Carbona, Nolan, AveXis, Nationwide Children’s Hospital and its affiliated research institute and foundation for breach of contract. They’d relied on the charity’s money to advance the treatment, the lawsuit alleged, then violated the terms of donation agreements by cutting it out of credit and ownership rights once the drug was headed for success. The suit sought damages of $500 million.

Many larger disease foundations have launched venture philanthropy programs that invest in biotech companies and projects, getting royalties and other financial considerations if their gifts help fund new treatments. In court filings, Nationwide Children’s called the notion that the tiny Sophia’s Cure had any right to the drug “simply not true, or even plausible,” and AveXis called it “wholly unsupported.”

Carbona said he was “disappointed and surprised” by the lawsuit. Nationwide didn’t respond to questions about the matter.

In November 2017, as the litigation went on, the results of the clinical trial that the charity helped fund were published.

They were remarkable. At 20 months, all 15 children who’d been treated remained alive, and none relied on a ventilator to breathe. Eleven of 12 infants who received a higher dose of the therapy were able to sit unassisted, speak and be fed orally. Two could walk on their own.

Based on preliminary trial data, the FDA had designated Zolgensma a breakthrough therapy, one of three special designations that helped it race from human trials to regulatory approval in five years. Once the full trial results came out, AveXis became a red-hot acquisition target.

In April 2018, Novartis beat out another bidder, agreeing to buy the company for $8.7 billion.

The sale delivered massive windfalls to those with the biggest stakes in AveXis.

Kaspar alone took in more than $400 million. He swapped his longtime family home in New Albany, Ohio, for a 9-acre estate in San Diego County, California, that had been listed for just over $8 million. It featured a dine-in stone wine cellar, a horse ring and stables.

Nolan, who’d led AveXis for less than three years, walked away with over $190 million; according to a financial filing, his payout included a golden parachute worth almost $65 million. Manning, the startup’s first big investor, made more than $315 million, multiplying his original investment by about 60. (Manning didn’t respond to calls or emailed questions from ProPublica.)

Carbona, too, made a bundle — he declined to say how much. Since he’d already left the company, his payout wasn’t disclosed in SEC filings. “It didn’t matter,” he said of the money. The 20-hour days he’d put into AveXis had helped advance a lifesaving drug. “This was a significant impact on humanity.”

After watching AveXis’ executives and investors cash in, the Gaynors were dealt another painful setback. In early 2019, a U.S. district court judge in Ohio dismissed Sophia’s Cure’s lawsuit against all parties, concluding there had been no breach of contract.

Their last hope for recognition of the charity’s role in bringing Zolgensma to the world was extinguished.

Once Novartis acquired AveXis, it turned to setting a price for its much-anticipated gene therapy.

Unlike other nations, the United States allows companies to charge whatever they want for new drugs. This often means Americans pay the world’s highest prices, particularly during the period when only the original manufacturer can market a drug. Research by PhRMA, the trade group for drug companies, suggests unfettered pricing buys Americans faster access, as long as insurers will pay: New medicines most often launch first in the U.S.

Novartis’ deliberations took place at the end of a decade in which launch prices of new drugs had risen exponentially, drawing ire from patient advocacy groups and Congress. The median annual launch price for a new drug jumped from about $2,000 in 2008 to about $180,000 in 2021, one study found.

In part, the increase reflected that a growing proportion of new drugs treated rare diseases. Drug companies have argued these therapies should cost more because their markets are smaller, making it harder to recoup expenses.

Cell and gene therapies also drove prices higher. The first three such treatments were approved in 2017, launching at prices of $370,000 or more. Luxturna, a gene therapy for a rare disorder that causes vision loss, costs $425,000 per eye.

Industry insiders assumed Zolgensma would cost more than Luxturna. But how much?

How drug companies pick prices for their products is among their most closely held secrets.

Beyond its statement, Novartis didn’t respond to questions from ProPublica about how it set or justified Zolgensma’s price. We reached out to more than three dozen people who were at the company or consulted for it at the time; most didn’t respond or declined to comment. A couple said they were bound by nondisclosure agreements.

The most visible portion of Novartis’ work was an effort to put a dollar value on how much Zolgensma would extend and improve SMA patients’ lives and offset the costs of caring for them.

This approach, known as value-based pricing, was originally championed by insurers and consumer watchdogs hoping to rein in drug prices. Other nations use economic assessments to decide whether to cover drugs and at what price, often paying far less than the U.S. for the same treatments.

But pharmaceutical companies have learned to use these techniques to their advantage.

Novartis brought together experts from academia and top consulting firms to work with its internal health economics team to publish research framing Zolgensma as a good value even at a high price.

One of the academics was Daniel Malone, then a professor at the University of Arizona’s College of Pharmacy. The target audience was mainly insurers, he said in an interview.

“We’re trying to influence the thousands of pharmacy and therapeutics committees around the country that are going to be looking at this therapy and whether they are going to provide it,” he said.

At the company’s direction, Malone said, their model mainly compared Zolgensma to the only other SMA treatment then on the market, a chronic treatment called Spinraza. It, too, was pricey, costing $750,000 in the first year and $375,000 every year after; over a decade, the tally would come to more than $4 million. (This was hypothetical; the FDA had approved Spinraza in December 2016, so no one had ever taken it for that long.)

A paper Malone co-authored concluded that Zolgensma, at prices up to $5 million, was a better buy than its rival, delivering more therapeutic benefit at a similar cost.

Company executives publicly floated multimillion-dollar prices for Zolgensma using data points from Malone and others.

“Four million dollars is a significant amount of money,” Dave Lennon, then president of Novartis’ AveXis unit, told Wall Street analysts on a call in November 2018. But “we’ve shown through other studies that we are cost-effective in the range of $4 million to $5 million.”

Such talk normalized “prices that would’ve been inconceivable a generation ago,” said Peter Maybarduk, director of access to medicines at the nonprofit consumer advocacy group Public Citizen. “It has a desensitizing effect.”

Novartis’ team of experts also helped the company prepare for Zolgensma’s evaluation by the Institute for Clinical and Economic Review, a nonprofit that assesses whether drugs are priced fairly.

Unlike agencies in Europe that do similar evaluations to set drug prices for national health systems, ICER’s recommendations aren’t binding, but they’ve become increasingly influential among public and private payers when it comes to coverage decisions.

Dr. Steven D. Pearson, the nonprofit’s founder, said that as ICER began its review, he was aware that investors were pushing for a big number.

“There was what I would call pressure from Wall Street,” he said. “This was going to set a precedent. Investors wanted to see a high price here.”

At first, it looked like ICER would resist. Its December 2018 draft report said Zolgensma would be overpriced at $2 million.

Novartis pushed back. Another consultant, University of Washington professor emeritus Louis Garrison, submitted public comments echoing a forthcoming AveXis-sponsored journal article he’d co-authored. It argued that drugs like Zolgensma, which treat rare, catastrophic conditions, deserved higher prices, in part to “incentivize appropriate risk taking and investments” by their developers.

Garrison said AveXis reviewed the article prior to publication, but he had the final say on its content. “I thought I could make a value-based argument that they would welcome and that I believe in,” he said. He said he was not directly involved in the company’s pricing decision.

Nonetheless, ICER’s final report in April 2019 concluded Zolgensma would need to be priced under $900,000 to be cost-effective, though it acknowledged the drug was still being tested on infants who hadn’t yet shown symptoms of SMA. If they also benefited, the report suggested the drug’s value might increase.

On May 24, the FDA approved Zolgensma to treat children under 2 with all forms of SMA.

Novartis finally revealed the treatment’s U.S. launch price, $2.125 million, framing this as a 50% discount on Spinraza and what the company’s research showed the gene therapy was worth.

It also pocketed yet another taxpayer-funded benefit: a voucher from the Food and Drug Administration redeemable for accelerated review of another drug. Such vouchers — designed to encourage companies to invest in pediatric rare-disease treatments — can be sold, typically bringing prices of around $100 million apiece.

That same day, ICER released an update. New data showing Zolgensma’s substantial benefits for presymptomatic children made the drug cost-effective at prices up to $1.9 million by one benchmark and up to $2.1 million by another, it said.

Pearson acknowledged the scale and timing of the switch were unusual, but said it was driven by the data, not outside pressure. “We weren’t trying to fit into somebody’s preexpectation of where the number would be, believe me,” he said.

He immediately caught flak from insurers.

“I got a lot of phone calls saying, ‘Why on earth did you say $2.1 million was a fair price? How could that possibly be the case? We’re going to get swamped with this,’” he recalled.

The Gaynors, linking to news coverage on Zolgensma’s launch, wrote on the Sophia’s Cure Facebook page that they were “ecstatic” for children newly born with SMA, but that helping create the world’s most expensive drug “is certainly not what we had in mind.”

Malone said he thought it was mostly the potential for blowback that had prevented Novartis from demanding even more for Zolgensma. He’d recommended charging the full $5 million.

“Obviously it didn’t stick,” he said. “They decided not to price the product there, I think, because of the political backlash they would’ve gotten being the first out of the gate at that price point.”

In the months after Zolgensma hit the market in the U.S., parents of children with SMA frequently ran into resistance from health insurers that refused to pay for it.

Between late 2019 and mid-2022, Chicago attorney Eamon Kelly represented at least seven parents battling health plans across the country, helping them appeal denied claims or representing them at state Medicaid hearings.

Hailey Weihs came to Kelly when her insurer, a Medicaid-managed care plan in Texas, wouldn’t pay for Zolgensma for her infant daughter Aniya. As the coverage dispute dragged on, Aniya developed tongue tremors and lost the ability to bear weight on her legs.

Kelly won the case, as he had all the others, but Aniya’s five-month wait to get the drug was terrifying. “Every day kids with this disease lose motor neurons,” Weihs said. “When you lose them, you cannot get them back.”

Now state Medicaid programs and most employer health plans cover Zolgensma, but they often limit which patients get access. Some require doctors to get approval in advance before providing the treatment or impose restrictions on who’s eligible that go beyond what’s on the drug’s label, such as requiring an SMA specialist to prescribe it.

Though fewer than 300 American children are born each year with SMA, treatments for the disease annually rank among the top 20 drug classes for Medicaid spending. From 2019 through 2022, Medicaid spent $309 million on 208 Zolgensma claims, an average of almost $1.5 million per claim. (Under federal law, Medicaid doesn’t pay list price for drugs, getting substantial rebates; other payers also negotiate discounts.)

Globally, more than 4,000 children have been treated with Zolgensma, Novartis said. The drug topped $1 billion in annual sales in its second full year on the market. Through 2024, the company had reported over $6.4 billion in revenue from Zolgensma sales.

Novartis is working to expand use of the drug in older children, in part by seeking approval for a second version of the drug, administered by spinal injection, for children with less severe SMA.

“We are unwavering in our commitment to the SMA community and will continue to advance efforts to ensure access to Zolgensma for SMA patients who may benefit from this transformative, one-time gene therapy,” the company said in its statement.

Still, more than five years after Zolgensma’s approval in the U.S., the drug remains out of reach for children in many low- and middle-income countries.

Love, KEI’s director, said he’s heard from families in countries like India and South Africa, where it’s a struggle to obtain not only Zolgensma, but also other SMA treatments available in the U.S.

“It’s maddening to me,” he said.

After setting aside their charity work, the Gaynors refocused their energy on Sophia and her two younger siblings, who don’t have SMA.

They’ve taken the clan to Disney World and to the Bahamas to swim with dolphins. Their youngest, who’s 8, lies beside Sophia on her bed and watches movies with her.

Now 15, Sophia had her longest-ever hospitalization in early 2024 when a virus caused her blood sugar to plummet and triggered frequent seizures. She didn’t wake up for two weeks. Since then, she’s been weaker, her affect flatter.

Her parents say they don’t think about the future. “Our focus is that she’s happy, that there’s love all around her,” Catherine said. “It’s just day to day.”

The Gaynors have taken solace in the idea that, through Sophia’s Cure, their daughter has made a difference for all the children with SMA who came after her. “That was kind of our consolation prize,” Catherine said.

One of those kids turned out to be her cousin, Vincent’s sister’s son, who was diagnosed with SMA in 2023 and then treated with Zolgensma. He walked at 10 months and now races around. “That helped me, in part, feel better about what we did,” Vincent said.

He still bristles at the drug’s price, which he blames on the payouts hauled in by those at AveXis and now Novartis.

“All those people, they all came in at the 12th hour once the trial was funded and you had the breakthrough,” he said. “Once it was taken from us, it was all about greed.”

The night the EMTs carried Elizabeth Nakagawa from her home, bleeding and in pain, the tarp they’d wrapped her in reminded her of a body bag.

Nakagawa, 39, is a Coast Guard commander: stoic, methodical, an engineer by trade. But as they maneuvered her past her young daughters’ bedroom, down the narrow steps and into the ambulance, she felt a stab of fear. She might never see her girls again.

Then came a blast of anger. She’d been treated for a miscarriage before. She knew her life never should have been in danger.

Earlier that day, April 3, 2023, Nakagawa had been scheduled to have a surgical procedure called a D&C, or dilation and curettage, to remove fetal tissue after losing a very wanted pregnancy. But that morning, she was told the surgery had been canceled because Tricare, the military’s health insurance plan, refused to pay for it.

While her doctor appealed, Nakagawa waited. Then the cramps and bleeding began.

In recent months, ProPublica and other media outlets have told the stories of women who died or nearly died when state abortion restrictions imposed after the Supreme Court’s 2022 Dobbs decision impeded them from getting critical care.

But long before Roe v. Wade was overturned, military service members and their families have faced strict limits on abortion services, which are commonly used to resolve miscarriages.

Under a decades-old federal law, the military is prohibited from paying for abortions except in cases of rape, incest or to save the life of the mother. This applies even to service members based in states where abortion is legal; Nakagawa lives in Sonoma County, California.

There’s also no exception for catastrophic or fatal fetal anomalies. In such cases, service members either have to pay out of pocket for abortions or carry to term fetuses that won’t survive outside the womb.

Tricare does allow abortions in cases like Nakagawa’s, in which the fetus has no heartbeat. But even then, some doctors who treat military service members say that Tricare requires more documentation and takes longer to approve these procedures than other insurers, putting women at risk.

“There definitely have been cases where our Tricare patients have required emergency services, emergency D&C procedures, blood transfusions, things that have been critical to lifesaving care because their procedure had yet to occur,” said Dr. Lauren Robertson, an OB-GYN who has served military members and their spouses in San Diego for more than a decade.

“It just feels very unnecessary.”

Since the Dobbs decision, abortion care for service members seems to be coming under heightened scrutiny, said retired Rear Adm. Dana Thomas, who was until recently the Coast Guard’s chief medical officer and advocated for Nakagawa.

“Trust me, post Roe v. Wade, I’m sure people felt there was much more of a spotlight,” Thomas said. “I think they were more guarded after June of ’22.”

After being rushed to the emergency room, Nakagawa hemorrhaged for four more hours before doctors performed the surgery Tricare had refused to authorize. Later, Tricare and Defense Department officials would all agree that Nakagawa should have been treated as her doctor recommended, and she said they told her they’d taken steps to prevent future mistakes.

But her experience, which doctors say nearly cost Nakagawa her life, laid bare the challenges service members have long faced in obtaining reproductive health care. And it raises questions about whether the Supreme Court’s ruling has created a chilling effect that has further complicated access to these procedures.

Officials at the Defense Health Agency, which runs the military health system, including Tricare, did not respond to specific questions from ProPublica, but they provided a statement saying its policies haven’t budged.

“There have not been any changes to Tricare coverage or documentation requirements for medically necessary care of D&Cs following the Supreme Court’s Dobbs decision,” the statement said. “Medically necessary care was, and continues to be, covered.”

The agency declined to answer questions about Nakagawa, saying that “as a matter of practice” it doesn’t discuss individual beneficiaries’ care. (ProPublica is involved in an unrelated public records lawsuit with the DHA.)

As a senior officer, Nakagawa felt duty-bound to press for answers about what happened to her.

“The abortion policy, in theory, is supposed to protect life, and in my case it did the opposite,” Nakagawa said. “It almost led to my children not having a mother.”

After the Supreme Court upended Roe, the Biden administration took steps to reassure service members that their access to reproductive health care would remain unaffected by a wave of state abortion bans.

An October 2022 memo from Defense Secretary Lloyd Austin pledged to facilitate leave for service members seeking abortions that were not covered by Tricare, and to pay for travel if care wasn’t available nearby. It also emphasized that these procedures would be “consistent with applicable federal law.”

The statute barring the Defense Department from paying for most abortions goes back to 1985 and mirrors language in what’s called the Hyde Amendment. Named for its author, Henry Hyde, a Republican representative from Illinois, Congress has attached the amendment to spending bills since the late 1970s to prohibit the use of federal funds on abortion.

With Congress in control of military spending, abortion care is highly politicized, said Kyleanne Hunter, a Marine Corps combat veteran and senior political scientist at RAND Corp. “There’s been a lot of backlash and a lot of scrutiny and a lot of congressional disapproval as to how the DOD has engaged with abortion care, D&C care and the like.”

About 9.5 million people, including active-duty service members and their families as well as military retirees and their dependents, rely on Tricare for health services. Women make up a growing portion of the active-duty force, more than 17%. They also leave the military at higher rates. Research by RAND and others suggests the military’s reproductive health policies may make it harder to recruit and retain them.

Dr. Toni Marengo, a former Navy OB-GYN, said she left the service in part because she felt unable to provide patients with appropriate care. Many of them only discovered how sharply Tricare’s policies curtailed access to procedures like D&Cs when they needed them.

“It was like living in a pre-Roe world,” said Marengo, now chief medical officer at Planned Parenthood of the Pacific Southwest.

The effects have been felt for decades. In 1994, Maureen Griffin and her then-husband, a captain in the Air National Guard, ended her pregnancy after learning their baby had anencephaly, a fatal birth defect. They found out the military considered her induced labor an abortion when she got a phone call from a bill collector for the hospital seeking thousands of dollars in reimbursement for the procedure.

“I said: ‘We have full coverage. My husband’s in the military.’ And they said, ‘They don’t pay for abortions,’” Griffin recalled. “We were completely blindsided. I mean, no one called it an abortion. It was a horrible tragedy.”

Griffin, then known as Maureen Britell, was so outraged that she sued the Defense Department, winning a judgment in federal district court in 2002. Two years later, an appeals court reversed the decision, upholding the Defense Department’s refusal to cover abortions in such circumstances.

Twenty-five years after Griffin’s pregnancy, Samantha Babcock spent the equivalent of seven paychecks to fly home from her husband’s Air Force base in Okinawa, Japan, for an abortion Tricare wouldn’t cover.

She was five months pregnant when doctors told her that her fetus had multiple abnormalities and wasn’t viable.

The grief was crushing. Then she found out that, by law, the military couldn’t perform or pay for a surgery called a D&E, or dilation and evacuation, which her military doctors agreed was the safest option. She and her family paid $14,000 — most of it for plane tickets — with help from a GoFundMe so that she could go home to Portland, Oregon, to get the procedure.

She still can’t believe such a step was necessary.

“I assumed Tricare had my best interest at heart,” she said. “If the condition was fatal, why wouldn’t they help me?”

Babcock said her specialist told her that the military would pay to transfer her temporarily to Hawaii for more testing. They also offered to move her family to a location where they would have access to specialty care for the baby in the unlikely chance she survived outside of the womb.

For Babcock, that was untenable. “I did not want to keep growing a baby that wouldn’t live,” she said.

In August 2022, Thomas, the Coast Guard’s chief medical officer, was galvanized into action when a service member sought help to end her pregnancy after receiving a diagnosis similar to Babcock’s.

Doctors had recommended that, like Babcock, she have a D&E. Because the fetus still had a heartbeat, Tricare would not approve the procedure.

Thomas called Tricare daily trying to find a solution, then elevated the case to leaders at the DHA, which sets policy for the health plan. “We have to do something,” she told them.

Tricare stuck to its denial even after the service member’s doctor appended a note explain­ing that continuing the pregnancy would endanger the patient’s life.

That was the first case Thomas had taken on after Dobbs.

The second was Nakagawa’s.

Nakagawa and her husband, Matt, met a couple years after earning engineering degrees from the Coast Guard Academy in the mid-2000s. Their path to a family was long and bittersweet. In 2015, they suffered a miscarriage. A year later, their first daughter was born. Then came a second miscarriage, followed by the birth of a healthy girl.

For the next three years, they tried for another child. Then Nakagawa got pregnant in 2021, only to learn at 10 weeks there was no fetal heartbeat. She waited, hoping to miscarry naturally, then tried abortion pills.

When a follow-up exam showed she hadn’t passed all the fetal tissue, her OB-GYN scheduled a D&C. The procedure was approved by Tricare, and she had the surgery soon afterward.

By early 2023, Nakagawa had risen to become chief of engineering at the Coast Guard’s training center in Petaluma, California, and her husband had left the service and was supporting her as a stay-at-home dad. They were thrilled to learn she was pregnant, only to have their joy turn to devastation when two ultrasounds showed that once again, her fetus had no heartbeat.

This time, since abortion pills hadn’t worked in 2021, Nakagawa and her OB-GYN agreed the best course would be to schedule a D&C as soon as possible. Her doctor’s office scheduled the procedure for Monday, April 3, and requested approval in advance, or prior authorization, from Tricare.

Then, five hours before the procedure was scheduled to begin, the office told Nakagawa the surgery was canceled — Tricare had refused to cover it.

In its denial letter, Health Net Federal Services, the contractor that administered claims for Tricare’s western region, said the services requested were “not a covered benefit.”

The insurer’s letter also said it had requested additional information from Nakagawa’s doctor, but that the information had not been sent. (Health Net declined to answer questions from ProPublica about Nakagawa even though she waived her right to privacy.)

Her doctor maintains that wasn’t the case. She declined to be identified, citing concerns about safety.

“Tricare has always been difficult to work with for coverage of women’s health care — they require records more than other insurances — this often creates a delay in care,” the doctor said via text.

The office staff appealed the denial, telling Nakagawa they’d provided documentation of the ultrasounds showing no fetal heartbeat. The staff also told her a Tricare medical director wasn’t available to review it that day and that it might take an additional three to five days to get a response.

Nakagawa called Tricare for answers herself, only to be told her options were to wait or pay out of pocket — not only for the surgery but for any follow-up care, including mental health counseling.

“It was surreal. I was angry and shaking,” Nakagawa said. She couldn’t understand why Tricare had approved her D&C in 2021 under similar circumstances, then denied the same care two years later.

Overwhelmed by emotion, she climbed into bed and cried herself to sleep.

At about 5 p.m., her doctor provided a prescription for abortion pills as a backup plan. But before Nakagawa could pick it up, she started to miscarry.

The first signs were mild cramping and spotting. Soon after, the fetus passed. Nakagawa yelled for her husband and sobbed. They consoled themselves with the thought that they’d made it through the hardest part.

“At least this is over,” Nakagawa recalled saying. “At least God’s giving us a break for once.”

Then the hemorrhaging started — fist-sized clots of blood that soaked through sanitary pads in minutes. Nakagawa lay in the fetal position on towels, in so much pain she couldn’t sit up.

Around 9 p.m., her husband called the doctor, who recommended they go to the emergency room.

At the hospital, she was given fluids, a clotting drug and a transfusion, but her bleeding continued.

After four hours, doctors decided her condition was critical and they needed to intervene. They performed a D&C to remove the remaining tissue.

Nakagawa’s recovery took more than a week. Lying on her couch, unable to walk, she was determined to ensure other service women would get the care she was denied. Taking a risk, she banged out a long email to Thomas, who had a reputation for being approachable.

“I feel compelled to report a traumatic experience I went through that will undoubtedly impact more women in the CG and DOD if the TRICARE policy is not changed,” the email began. “The summary is that I nearly lost my life last week due largely to a TRICARE policy regarding miscarriages and abortions.”

Thomas connected Nakagawa to the Defense Health Agency’s chief medical officer, Dr. Paul Cordts, who called her personally a month after her emergency surgery.

Nakagawa said that Cordts seemed apologetic and even angry on her behalf. “This shouldn’t have happened to you,” she recalled him saying, adding that he’d get to the bottom of what went wrong. (Cordts didn’t respond to emails from ProPublica.)

Two days later, a new record appeared in Nakagawa’s Tricare file: a letter approving the scheduled D&C she’d never received and no longer needed. “Please contact the provider to schedule your appointment(s),” it said.

Cordts also arranged for Col. John Verghese, Tricare’s chief of clinical oversight and integration, to look into her case. Nakagawa said she had two calls with Verghese, who looped in a senior official at Health Net, the Tricare contractor that had dealt with the request to cover her D&C.

In one, she said, Verghese acknowledged Tricare had become more conservative in reviewing requests for D&Cs, requiring more documentation to justify approving these procedures. (Verghese, who has retired, declined to answer questions from ProPublica about the case.)

He admitted that until her case, Tricare hadn’t understood that delaying or denying care could put women at risk, she said. This infuriated Nakagawa.

“I just said, ‘Well, maybe you didn’t realize there would be physical negative consequences, but you had to know there would be mental and emotional consequences to making women carry around their [dead] fetuses’” after a miscarriage.

Verghese quickly apologized, she said.

On the final call, Nakagawa said that Verghese and the Health Net official told her that from now on, they would no longer require doctors to submit proof of no fetal heart tones to get approvals for D&Cs and would speed up reviews of appeals.

In its statement to ProPublica, the DHA maintained that Tricare’s coverage and documentation requirements for D&Cs have not changed.

Nakagawa is one of few women in senior leadership within Coast Guard civil engineering. She remains committed to serving in the military. But she worries about the impact the Defense Department’s reproductive health policies could have on service members and their spouses and daughters. Junior members especially might be less able to advocate for themselves, she said.

“At the very least, this policy will likely encourage women, like myself, to work for a company that has insurance that will cover these procedures,” she said, “At the worst, it will lead to service members or their dependents losing their lives.”