Zolgensma: The Gene Therapy That Changed How We Treat Spinal Muscular Atrophy

The Disease That Steals Movement

Spinal muscular atrophy (SMA) is a devastating genetic disorder that attacks the nerve cells responsible for voluntary muscle movement. It is caused by mutations in the SMN1 gene on chromosome 5, which provides instructions for making the survival motor neuron (SMN) protein. Without sufficient levels of this protein, motor neurons in the spinal cord degenerate and die, leading to progressive muscle weakness that can affect breathing, swallowing, and movement.

SMA is the most common genetic cause of death in infants. It affects roughly 1 in 10,000 live births worldwide, and approximately 1 in 50 people are carriers of the mutated gene. The disease follows an autosomal recessive inheritance pattern, meaning a child must inherit a defective copy of SMN1 from both parents to develop the condition.

The SMN1 and SMN2 Genes

Understanding SMA requires understanding a quirk of human genetics. Humans carry two related genes: SMN1 and SMN2. Both can produce SMN protein, but they differ in a critical way. SMN1 produces full-length, fully functional SMN protein. SMN2, due to a single nucleotide difference in exon 7, primarily produces a truncated, unstable version of the protein. Only about 10-15% of the protein made from SMN2 is full-length and functional.

In SMA patients, SMN1 is deleted or nonfunctional. They rely entirely on their SMN2 copies for whatever SMN protein they can produce. The number of SMN2 copies a patient carries is the single most important modifier of disease severity — more copies mean more residual SMN protein, and generally a milder form of the disease.

Types of SMA

SMA is classified into four main types based on age of onset and the highest motor milestone achieved:

Type I (Werdnig-Hoffmann disease) is the most severe and most common form, accounting for about 60% of all SMA cases. Symptoms appear before 6 months of age. Affected infants never achieve the ability to sit independently. They develop profound muscle weakness, difficulty breathing, and trouble swallowing. Before the era of disease-modifying therapies, the natural history of Type I SMA was grim: the majority of children died or required permanent ventilatory support before their second birthday. These infants typically carry only two copies of SMN2.

Type II (Dubowitz disease) has an onset between 6 and 18 months. Children can sit independently but never walk unaided. Respiratory complications are common, and scoliosis often develops. Life expectancy extends into early adulthood with supportive care, though quality of life is significantly reduced. Patients usually carry three copies of SMN2.

Type III (Kugelberg-Welander disease) appears after 18 months of age. Individuals can walk at some point but may lose that ability over time. This form progresses more slowly, and life expectancy is near-normal. Patients typically have three or four SMN2 copies.

Type IV is the adult-onset form, with symptoms beginning after age 21. It is the rarest and mildest variant. Patients usually have four or more SMN2 copies and experience gradual motor decline but maintain independence for most of their lives.

Before Zolgensma: A Landscape of Limited Options

For decades, SMA management was purely supportive. Families facing a Type I diagnosis were told there was nothing that could change the trajectory of the disease. Physical therapy, respiratory support, nutritional management, and eventually palliative care were the standard of care.

The first hint of change came in December 2016, when the FDA approved Spinraza (nusinersen), developed by Biogen and Ionis Pharmaceuticals. Spinraza is an antisense oligonucleotide (ASO) — a short synthetic strand of modified RNA that binds to the pre-mRNA produced by the SMN2 gene and modifies its splicing. By altering how exon 7 is included during mRNA processing, Spinraza coaxes SMN2 into producing more full-length, functional SMN protein.

Spinraza represented a genuine breakthrough for the SMA community. For the first time, a drug could modify the course of the disease rather than merely manage symptoms. But it came with significant practical challenges. Spinraza is administered via intrathecal injection — a lumbar puncture into the spinal fluid — every four months after an initial loading phase. For infants and children with spinal deformities or surgical hardware, these procedures can be technically difficult and require sedation or anesthesia. The treatment is lifelong, the injections are repeated indefinitely, and the annual cost exceeds $300,000 after the first year.

Into this environment, a radically different approach was taking shape.

How Zolgensma Works

Zolgensma (onasemnogene abeparvovec-xioi) takes a fundamentally different approach to treating SMA. Rather than modifying splicing of the backup SMN2 gene, it delivers a brand-new, fully functional copy of the SMN1 gene directly into the patient's cells. It is, in the most literal sense, gene replacement therapy.

The AAV9 Vector

The delivery vehicle for Zolgensma is an adeno-associated virus serotype 9 (AAV9). AAV9 was chosen for several key properties. First, it is highly efficient at crossing the blood-brain barrier and infecting motor neurons in the spinal cord — the precise cells that degenerate in SMA. Second, AAV vectors in general have a strong safety track record in gene therapy. They are derived from a non-pathogenic virus, do not integrate into the host genome in most cases (the therapeutic DNA persists as an episomal element in the nucleus), and provoke a relatively manageable immune response.

The AAV9 capsid is loaded with a transgene cassette containing a codon-optimized human SMN1 gene driven by a strong constitutive promoter (the chicken beta-actin hybrid promoter, or CB promoter). Once the viral vector infects a motor neuron, the transgene is delivered to the cell nucleus, where it begins producing full-length SMN protein.

One-Time Intravenous Infusion

One of Zolgensma's most distinctive features is its route of administration for young patients: a single, one-time intravenous (IV) infusion. The entire dose is delivered over approximately 60 minutes. There are no repeat injections, no ongoing lumbar punctures, no maintenance dosing. In theory, because the transgene DNA persists in non-dividing cells like motor neurons, a single treatment could provide lasting benefit for the life of the patient.

The approved dosing for Zolgensma (for the IV formulation) is 1.1 x 10^14 vector genomes per kilogram of body weight. This is an extraordinarily high dose by gene therapy standards, which is part of why the treatment is restricted to young children — the total number of viral particles needed scales directly with body weight, and treating larger patients IV would require impractically large (and potentially more toxic) doses.

Patients receive corticosteroids before and after the infusion to manage the immune response to the viral vector, typically prednisolone at 1 mg/kg/day for at least 30 days, with a slow taper guided by liver function tests.

Clinical Trials: Unprecedented Results

The START Trial (AVXS-101)

The pivotal clinical evidence for Zolgensma came from the phase I START trial, conducted at Nationwide Children's Hospital in Columbus, Ohio, and led by Dr. Jerry Mendell. The trial enrolled 15 infants with Type I SMA, all younger than 6 months at dosing and carrying two copies of SMN2.

The results, published in the New England Journal of Medicine in 2017, were extraordinary by any measure. In the high-dose cohort (the dose that became the commercial product), all 12 infants were alive and free of permanent ventilation at 20 months of age — an outcome achieved by fewer than 8% of untreated Type I SMA patients in natural history studies. Eleven of 12 could sit independently for at least 30 seconds. Two patients could walk independently.

These outcomes were not incremental improvements. They represented a near-total rewriting of the natural history of the most severe form of the disease. For a condition where the expected outcome was death or ventilator dependence by age 2, having children sit, reach, and even walk was transformative.

The STR1VE Trials

The confirmatory phase III STR1VE-US trial enrolled 22 patients with Type I SMA. At 18 months of age, 59% of patients (13 of 22) could sit independently for at least 30 seconds — compared to 0% in natural history. All patients survived to 14 months of age without permanent ventilation. The STR1VE-EU trial produced similar results in a European cohort.

Long-Term Follow-Up

Long-term data from the original START trial has continued to be encouraging. Follow-up data published through 2023 and beyond has shown that treatment effects are durable, with patients maintaining motor milestones years after a single infusion. Some patients from the original trial, now approaching 8-10 years post-treatment, continue to show stable or improving motor function. This durability supports the hypothesis that the episomal transgene in non-dividing motor neurons provides lasting expression.

However, researchers are transparent that truly "lifelong" durability remains unproven. Motor neurons do not divide, which theoretically supports permanent transgene expression, but there are open questions about whether expression might wane over decades. The oldest treated patients are still children, so the long-term picture continues to evolve.

FDA Approval and the $2.1 Million Price Tag

The FDA approved Zolgensma on May 24, 2019, for the treatment of pediatric patients under 2 years of age with SMA. It was the most expensive drug in the world at the time of launch, with a list price of $2.125 million for a single dose.

Novartis's Pricing Rationale

Novartis Gene Therapies (formerly AveXis, which Novartis acquired in 2018 for $8.7 billion) defended the price using value-based arguments. The company pointed out that the lifetime cost of treating an SMA patient with Spinraza — at roughly $750,000 in the first year and $375,000 annually thereafter — would exceed $4 million over a decade, and potentially much more over a patient's lifetime. Against that benchmark, a one-time treatment at $2.1 million could represent a cost saving for the healthcare system, especially if the benefits proved durable.

Novartis also offered outcomes-based contracts to payers, agreeing to rebates if Zolgensma failed to meet specified clinical milestones. The company introduced installment payment options, allowing the cost to be spread over five years, and established a patient access program to help families navigate insurance coverage.

The pricing debate touched on deeper questions about how society values curative therapies for rare diseases. Traditional drug pricing models, built around chronic medications taken over many years, struggle to accommodate one-time treatments that deliver their entire value in a single dose. Zolgensma became a case study in the economics of gene therapy and remains central to ongoing policy discussions about orphan drug pricing and value assessment.

Commercial Performance

Despite the sticker shock, Zolgensma has been commercially successful. The drug generated approximately $1.4 billion in revenue in 2022 and similar figures in subsequent years. Cumulative global revenue surpassed $6.4 billion by the end of 2025, making it one of the most commercially successful gene therapies ever launched. Novartis has secured reimbursement in over 40 countries, and the drug has been administered to more than 3,500 patients worldwide.

The Critical Role of Newborn Screening

One of the most important lessons from Zolgensma's clinical program is that timing matters enormously. Motor neurons, once lost, cannot be replaced. Gene therapy can preserve existing motor neurons by providing them with the SMN protein they need to survive, but it cannot resurrect neurons that have already degenerated. The implication is straightforward: earlier treatment yields better outcomes.

This insight has driven a major push for universal newborn screening for SMA. In the United States, SMA was added to the Recommended Uniform Screening Panel (RUSP) in 2018, and by 2023, all 50 states had implemented SMA screening in their newborn screening programs. Many other countries, including Germany, Taiwan, Australia, Belgium, and parts of Italy and Japan, have followed suit or are in the process of implementing screening.

Newborn screening uses a DNA-based test to detect homozygous deletion of SMN1 exon 7 from a dried blood spot — the same heel-prick blood sample already collected from every newborn for other screening tests. The test is highly accurate and can identify affected infants within days of birth, before symptoms appear.

The clinical data strongly supports presymptomatic treatment. The SPR1NT trial studied Zolgensma in presymptomatic infants identified through newborn screening or family history. Infants with two SMN2 copies who were treated before symptoms appeared achieved motor milestones within normal developmental windows — sitting, standing, and in many cases walking independently and on time. These results were dramatically better than outcomes in symptomatic patients, underscoring a narrow but critical treatment window.

The emerging standard of care is clear: identify SMA at birth, confirm with genetic testing, determine SMN2 copy number, and treat as early as possible.

Zolgensma vs. Spinraza vs. Evrysdi: A Three-Drug Landscape

Since August 2020, SMA patients have had three disease-modifying therapies available, each with a distinct mechanism of action, route of administration, and clinical profile.

Spinraza (Nusinersen)

Spinraza, as noted earlier, is an antisense oligonucleotide that modifies SMN2 splicing to increase full-length SMN protein production. It is administered intrathecally (into the spinal fluid) via lumbar puncture. After four loading doses in the first two months, patients receive maintenance doses every four months, indefinitely.

Advantages: Extensive long-term safety and efficacy data (approved since 2016); approved for all ages and SMA types; direct CNS delivery ensures drug reaches motor neurons.

Disadvantages: Requires repeated intrathecal injections under sedation/anesthesia; technically challenging in patients with scoliosis or spinal fusion; ongoing treatment burden; high cumulative cost over a lifetime.

Evrysdi (Risdiplam)

Evrysdi, developed by Roche/Genentech, was approved by the FDA in August 2020. It is a small-molecule splicing modifier that, like Spinraza, increases the proportion of full-length SMN protein produced from SMN2. However, Evrysdi is taken orally — as a once-daily liquid administered by mouth or feeding tube — making it the most convenient option from an administration standpoint.

Advantages: Oral administration at home; no injections or hospital procedures; approved for patients 2 months of age and older; systemic distribution (reaches peripheral tissues in addition to CNS).

Disadvantages: Requires daily dosing for life; less long-term data compared to Spinraza or Zolgensma; systemic distribution means drug is present throughout the body (potential for off-target effects, though clinical safety data has been reassuring).

Comparing Outcomes

Head-to-head trials between these three therapies have not been conducted, and direct comparisons of clinical trial results are complicated by differences in patient populations, endpoints, and trial designs. However, real-world evidence registries and retrospective analyses suggest several patterns:

• Presymptomatic treatment with any of the three therapies produces the best outcomes, reinforcing the importance of newborn screening
• Zolgensma appears to produce the most dramatic motor milestone achievements in symptomatic Type I patients, particularly when administered early
• Combination therapy (Zolgensma plus Spinraza or Evrysdi) is increasingly being used in clinical practice, though evidence supporting this approach remains limited and insurance coverage is inconsistent
• All three therapies have shifted the natural history of SMA substantially, transforming a fatal disease into a manageable chronic condition for many patients

The choice between treatments often depends on patient age, disease severity, family preference, access, and insurance coverage. Many clinicians now view these therapies as complementary rather than competing.

Itvisma: Extending Gene Therapy to Older Patients

One of the key limitations of intravenous Zolgensma is that it is approved only for children under 2 years of age. The dose required for IV delivery scales with body weight, and treating older, heavier patients with the IV formulation would require impractically large doses with unacceptable toxicity risks.

To address this gap, Novartis developed Itvisma (onasemnogene abeparvovec-jkil), an intrathecal formulation of the same AAV9-SMN1 gene therapy. Itvisma is delivered directly into the cerebrospinal fluid via lumbar puncture, bypassing the need for systemic distribution and allowing a much smaller total dose to reach the motor neurons.

The FDA approved Itvisma in 2025 for the treatment of SMA in patients 2 to 18 years of age with bi-allelic mutations in the SMN1 gene. This represented a significant expansion of the gene therapy option to older children and adolescents who were previously ineligible.

Clinical trial data for Itvisma showed improvements in motor function scores in patients with Type II and Type III SMA, though the magnitude of benefit was generally more modest than what was observed in presymptomatic infants treated IV with Zolgensma. This is consistent with the principle that earlier treatment — before extensive motor neuron loss — produces better outcomes. Still, for older patients with SMA who had no gene therapy option, Itvisma provides a meaningful new treatment.

The intrathecal delivery approach also raises new questions about immunogenicity, since many older patients may have pre-existing antibodies to AAV9 from natural exposure to adeno-associated viruses. Anti-AAV9 antibody testing is recommended before treatment, and patients with elevated titers may not be eligible.

Safety Concerns

While Zolgensma has demonstrated transformative efficacy, its safety profile demands careful attention and monitoring.

Hepatotoxicity

The most well-characterized serious risk of Zolgensma is liver toxicity. The AAV9 vector has significant tropism for the liver — in addition to motor neurons — and the high dose of viral particles delivered IV can trigger a substantial immune response targeting transduced liver cells. Elevated liver enzymes (transaminases) are observed in a majority of treated patients, and in rare cases, severe liver injury has occurred.

Serious hepatotoxicity, including acute liver failure and fatal cases, has been reported in the post-marketing setting. The FDA issued a boxed warning for Zolgensma regarding the risk of acute serious liver injury. Current protocols require close monitoring of liver function for at least three months after infusion, with liver function tests checked weekly initially. The corticosteroid regimen administered alongside the gene therapy is a critical part of managing this risk, and premature tapering of steroids has been associated with hepatic flares.

Thrombotic Microangiopathy (TMA)

Cases of thrombotic microangiopathy, including thrombotic thrombocytopenic purpura (TTP)-like events, have been reported following Zolgensma administration. TMA is a condition in which small blood clots form in blood vessels throughout the body, leading to low platelet counts, anemia, and potential organ damage. The mechanism is not fully understood, but it is believed to be related to the immune response triggered by the viral vector. Monitoring of platelet counts and renal function is now standard post-treatment protocol.

Dorsal Root Ganglia Toxicity

Preclinical studies in non-human primates and some clinical observations have raised concerns about inflammation and neuronal degeneration in the dorsal root ganglia (DRG) — clusters of sensory nerve cell bodies located along the spinal cord. DRG toxicity has been observed with high-dose AAV-based gene therapies more broadly and is an area of active investigation. In clinical practice, most patients treated with Zolgensma have not shown clinically significant sensory symptoms, but this remains an area of surveillance.

Data Integrity Controversy

In 2019, shortly after Zolgensma's approval, Novartis disclosed that AveXis employees had manipulated data in animal testing used to support the regulatory filing. The FDA investigated and ultimately concluded that the manipulated data did not affect the overall benefit-risk assessment of the drug, but the episode damaged trust and led to calls for greater scrutiny of gene therapy manufacturing data. Novartis took corrective actions, and AveXis employees involved were fired.

Long-Term Outcomes and Open Questions

As of 2026, the longest-followed Zolgensma patients are approximately nine years post-treatment. The data so far is encouraging. Children treated in the original START trial have maintained their motor milestones, and many continue to gain new skills. Respiratory function has remained stable in most patients. The trajectory of these early-treated children bears little resemblance to the natural history of untreated Type I SMA.

However, important questions remain unanswered:

Durability beyond a decade: Will the transgene continue to express SMN protein for 20, 30, or 50 years? No one knows yet. The episomal DNA should persist in non-dividing neurons, but biological surprises are always possible.

Re-dosing: If expression does wane, can patients be re-dosed? The immune response to AAV9 after a first dose makes IV re-administration extremely difficult — the body's neutralizing antibodies would likely destroy the vector before it could reach target cells. The intrathecal route might offer an alternative, and immunosuppressive strategies to enable re-dosing are under investigation.

Combination therapy: Many patients are now receiving Zolgensma plus an SMN2-modifying therapy (Spinraza or Evrysdi). Is this additive? Is it necessary? Does the benefit justify the added cost and burden? Clinical data is accumulating, but definitive answers are not yet available.

Late effects: Could there be adverse effects that emerge only years after treatment? Long-term follow-up studies are ongoing, including the LT-001 and LT-002 registries, which continue to track patients from the original clinical trials.

Manufacturing and access: Gene therapy manufacturing remains complex and expensive. Each batch of AAV9 vector requires large-scale bioreactor production, extensive quality testing, and cold-chain logistics. Scaling production to meet global demand — particularly in low- and middle-income countries where SMA is equally prevalent — remains a significant challenge.

The Broader Significance of Zolgensma

Zolgensma's story extends well beyond SMA. It served as proof of concept that a single IV infusion of an AAV-based gene therapy could fundamentally alter the course of a genetic disease. It demonstrated that gene therapy could be administered systemically (not just locally or ex vivo) with transformative clinical benefit. And it forced the healthcare system to confront the pricing and reimbursement challenges that will accompany an entire generation of gene therapies in development.

Since Zolgensma's approval, the gene therapy pipeline has expanded dramatically. AAV-based therapies are now approved or in late-stage development for hemophilia A, hemophilia B, Duchenne muscular dystrophy, aromatic L-amino acid decarboxylase (AADC) deficiency, and numerous other conditions. The lessons learned from Zolgensma's manufacturing, clinical development, safety monitoring, and market access have informed all of these programs.

For the SMA community, the impact has been most direct and most profound. Children who would have died in infancy are now attending school, playing with siblings, and reaching milestones that their families once believed impossible. The challenge going forward is ensuring that these therapies reach every child who needs them — regardless of geography, income, or timing of diagnosis.

Conclusion

Zolgensma represents a landmark in modern medicine: a one-time gene therapy that transforms the prognosis of the most common genetic killer of infants. Its clinical success validated decades of research in AAV gene therapy and motor neuron biology. Its $2.1 million price tag forced a global reckoning with how we value and pay for curative treatments. And its limitations — including safety risks, age restrictions, and unanswered questions about durability — remind us that gene therapy, for all its promise, is still a technology in its relative infancy.

What is no longer in question is whether gene therapy works for SMA. The answer, written in the milestones of children who can sit, stand, and walk, is unambiguous. The next chapter is about making it work better, making it safer, and making it available to every child born with this disease, everywhere in the world.

---

Sources

1. Mendell, J.R., et al. "Single-Dose Gene-Replacement Therapy for Spinal Muscular Atrophy." New England Journal of Medicine, 377, 1713-1722 (2017). DOI: 10.1056/NEJMoa1706198

2. Day, J.W., et al. "Onasemnogene abeparvovec gene therapy for symptomatic infantile-onset spinal muscular atrophy in patients with two copies of SMN2 (STR1VE): an open-label, single-arm, multicentre, phase 3 trial." The Lancet Neurology, 20(4), 284-293 (2021). DOI: 10.1016/S1474-4422(21)00001-6

3. Strauss, K.A., et al. "Onasemnogene abeparvovec for presymptomatic infants with three copies of SMN2 at risk for spinal muscular atrophy: the Phase III SPR1NT trial." Nature Medicine, 28, 1390-1397 (2022). DOI: 10.1038/s41591-022-01867-3

4. FDA. "Zolgensma (onasemnogene abeparvovec-xioi) Prescribing Information." U.S. Food and Drug Administration. Link

5. FDA. "FDA Approves Itvisma for Treatment of Spinal Muscular Atrophy in Patients Aged 2-18 Years." FDA Press Release, 2025. Link

6. Novartis. "Zolgensma Annual Report Data and Revenue Disclosures." Novartis AG Financial Reports, 2020-2025.

7. Finkel, R.S., et al. "Nusinersen versus Sham Control in Infantile-Onset Spinal Muscular Atrophy." New England Journal of Medicine, 377, 1723-1732 (2017). DOI: 10.1056/NEJMoa1702752

8. Baranello, G., et al. "Risdiplam in Type 1 Spinal Muscular Atrophy." New England Journal of Medicine, 384, 915-923 (2021). DOI: 10.1056/NEJMoa2009965

9. Curry, M., et al. "Newborn screening for spinal muscular atrophy: a review of current practice and future directions." International Journal of Neonatal Screening, 8(3), 44 (2022).

10. Feldman, A.G., et al. "Subacute Liver Failure Following Gene Replacement Therapy for Spinal Muscular Atrophy Type 1." Journal of Pediatrics, 225, 252-258 (2020). DOI: 10.1016/j.jpeds.2020.05.044