The Most Expensive Drug and the Argument for Speed

Zolgensma costs $2.1 million for a single intravenous infusion. It is a gene therapy that delivers a working copy of the SMN1 gene to a child with spinal muscular atrophy. One dose. The gene integrates. The motor neurons that were dying begin to function. The child reaches milestones, sitting, standing, walking, that the disease would have made impossible.

In presymptomatic children treated before symptoms appear, the results are as close to a cure as any gene therapy has produced. Long-term follow-up data through 7.5 years after dosing shows 100% of presymptomatically treated children achieving all assessed motor milestones, including independent walking. These are children who, without treatment, would have lost the ability to sit, to breathe unassisted, to swallow. Most would have died before age two.

The price provoked outrage. The outcomes ended the argument.

The Disease

Spinal muscular atrophy is caused by mutations in the SMN1 gene, which produces survival motor neuron protein. Without enough SMN protein, the motor neurons in the spinal cord degenerate. The muscles they control weaken and atrophy. In the most severe form, SMA type 1, this process begins before birth and accelerates in the first months of life. Breathing muscles fail. Swallowing becomes impossible. Death follows, typically within the first year or two.

SMA is the leading genetic cause of death in infants. It affects roughly 1 in 10,000 births. Every person with SMA has a second gene, SMN2, that produces a small amount of SMN protein, roughly 10% of normal. The number of SMN2 copies a person carries determines severity: two copies typically produce type 1 (most severe), three copies type 2 (intermediate), four or more copies type 3 or 4 (milder). But 10% of normal protein production is not enough. The motor neurons die.

The biology created two therapeutic targets. Replace the missing SMN1 gene entirely (gene therapy). Or force the backup SMN2 gene to produce more functional protein (splicing modification). Both approaches now have approved drugs, and the difference in outcomes between treated and untreated children is the most dramatic demonstration of therapeutic impact in the history of genetic medicine.

The Three Drugs

Adrian Krainer, a molecular biologist at Cold Spring Harbor Laboratory, attended a National Institute of Neurological Disorders and Stroke workshop on SMA in the early 2000s. The discussion focused on the subtle difference between SMN1 and SMN2: a single nucleotide change that causes SMN2 to skip exon 7 during RNA splicing, producing a truncated, unstable protein. Krainer's laboratory specialized in RNA splicing. He recognized that an antisense oligonucleotide could mask the splicing signal and force SMN2 to include exon 7, producing full-length functional protein.

By 2003, Krainer published the proof of concept. Frank Bennett at Ionis Pharmaceuticals initiated a collaboration. By 2008, the team published the sequence of nusinersen. Clinical trials followed. In December 2016, the FDA approved nusinersen (Spinraza) as the first treatment for SMA.

Nusinersen is administered by intrathecal injection, directly into the spinal fluid, every four months after an initial loading phase. Each injection costs approximately $118,000 in the first year and $236,000 per year thereafter. The treatment continues for life. It does not replace the gene. It modifies splicing in real time. When the drug clears, the splicing reverts.

Zolgensma (onasemnogene abeparvovec), developed by AveXis and acquired by Novartis, took a different approach. An AAV9 viral vector delivers a functional copy of the SMN1 gene. The vector crosses the blood-brain barrier after intravenous infusion. The gene integrates into the motor neuron's nucleus and produces SMN protein. One dose. The gene persists. The FDA approved Zolgensma in May 2019 for children under two years of age with SMA.

Risdiplam (Evrysdi), developed by Roche/Genentech, received FDA approval in August 2020. It is a small molecule taken orally, once daily, in liquid form. Like nusinersen, it modifies SMN2 splicing to increase full-length protein production. Unlike nusinersen, it does not require spinal injections. Unlike Zolgensma, it is not a one-time treatment. It is taken every day, at home, by mouth or feeding tube.

Three drugs. Three mechanisms. Three routes of administration. Three price structures. The choice between them depends on the child's age, SMA type, SMN2 copy number, and which healthcare system the family lives in.

The Timing Argument

SMA was added to the Recommended Uniform Screening Panel in July 2018, after the Advisory Committee on Heritable Disorders in Newborns and Children reviewed the evidence and recommended its inclusion to the Secretary of Health and Human Services. The nomination had first been submitted in 2008, but the committee required evidence that disease-modifying treatments existed before recommending screening. Nusinersen's approval in 2016 provided that evidence.

The screening test detects homozygous deletion of the SMN1 gene from a dried blood spot, the same sample already collected for other newborn screening conditions. It adds no additional procedure.

The timing data from screened populations confirms what the biology predicted: treatment before symptom onset produces dramatically better outcomes than treatment after motor neurons have already been lost.

A child with two SMN2 copies (typically SMA type 1) who receives Zolgensma before symptoms appear can achieve independent walking. The same child treated after symptom onset may achieve sitting but rarely walking. The same child untreated will die.

A child with three SMN2 copies (typically SMA type 2) who receives presymptomatic treatment achieves motor milestones comparable to unaffected children. The same child treated after symptom onset may achieve walking but with significant motor limitations. Without treatment, most children with SMA type 2 never walk independently.

The variable is time. The disease does not wait. Motor neurons lost before treatment are not recovered. Every week between birth and treatment is a week of irreversible neuronal death. Newborn screening identifies affected children in the first days of life. Treatment can begin within weeks. The gap between birth and treatment, compressed to its minimum by screening, is the difference between a child who walks and a child who does not.

The Price Question

Zolgensma at $2.1 million is the most expensive single drug administration in history. The figure is real and the sticker shock is warranted. It is also, by every actuarial analysis published, less expensive than the alternative.

Nusinersen costs approximately $750,000 in the first year and $375,000 per year thereafter, for life. A child treated with nusinersen from infancy through age 40 accumulates roughly $15 million in drug costs alone, excluding the cost of intrathecal injections, hospitalizations, respiratory support, mobility equipment, personal care assistance, and lost parental productivity.

Risdiplam costs approximately $100,000 to $340,000 per year depending on the dosing weight, also for life.

A one-time gene therapy at $2.1 million, if it eliminates the need for lifelong treatment, is the least expensive option by a wide margin. The question is durability: does the gene continue to express for decades? The longest follow-up data currently extends 7.5 years. The children treated presymptomatically in the earliest trials are now in elementary school. They are walking, running, attending school without respiratory support. The gene appears to be expressing. The 20-year, 40-year, 60-year durability data does not yet exist.

The Global Access Problem

SMA affects families in every country. Zolgensma is approved in the United States, the European Union, Japan, and a growing number of other countries. Nusinersen is more widely available. Risdiplam, as an oral medication that does not require specialized administration infrastructure, has the broadest potential global reach.

Access depends entirely on the healthcare system. In the United States, Zolgensma is typically covered by insurance, often after protracted prior authorization battles. In countries with national health systems, coverage decisions are made centrally: the UK's NICE, Germany's G-BA, Japan's PMDA. In countries without established rare disease drug access programs, none of the three treatments may be available at any price.

Newborn screening for SMA is similarly uneven. By early 2022, SMA screening was available to roughly 87% of newborns in the United States. Adoption in Europe varies by country: Germany, Belgium, and the Netherlands screen; the UK added SMA to its screening program more recently. Most of Asia, Africa, and South America do not screen for SMA at the population level.

A child born with SMA in the United States in 2026 is identified within days, treated within weeks, and has a high probability of walking. A child born with the same disease, the same mutations, the same biology in a country without screening or treatment access will follow the natural history of the disease: progressive paralysis and early death. The geography of birth determines the outcome more than the biology of the disease.

What SMA Built

SMA has produced more therapeutic firsts than any other rare disease in the past decade.

Nusinersen was the first antisense oligonucleotide approved for a neurodegenerative disease. The ASO chemistry, the intrathecal delivery route, and the regulatory pathway that Krainer and Bennett established for SMA have been directly applied to ASO programs for Huntington's disease, ALS, and Alzheimer's disease. The rare disease application came first because the unmet need was most urgent and the biology was most tractable. The platform now serves conditions affecting millions.

Zolgensma was the first AAV gene therapy for a neurodegenerative disease administered intravenously. The demonstration that an AAV9 vector could cross the blood-brain barrier and deliver a gene to motor neurons after a peripheral infusion expanded the design space for gene therapy across neurology.

Risdiplam was the first oral splicing modifier approved for any disease. The proof that a small molecule could modify RNA splicing with sufficient precision and safety for chronic daily use opened a drug development pathway applicable far beyond SMA.

Each drug was developed for a rare disease. Each created a platform that the pharmaceutical industry is now applying to common diseases. The families who enrolled their children in the early SMA trials, when the drugs were unproven and the outcomes uncertain, generated the safety and efficacy data that made those platforms possible. They went first.