Novartis AG is facing a new long-term safety question around its spinal muscular atrophy gene therapy franchise after the United States Food and Drug Administration said it is investigating a serious risk of neoplasia, including malignancy, associated with Zolgensma. The regulator disclosed on November 6 that molecular analysis in two tumour cases demonstrated integration of Zolgensma adeno-associated virus vector DNA into the patients’ genomes, with the treatment construct positioned upstream of or within the GLI1 oncogene.
The FDA described the findings as confirmed reports of product-related insertional oncogenesis involving GLI1, moving part of the safety concern beyond the theoretical integration risk already contained in Zolgensma prescribing information. Novartis separately said it has received nine reports of tumours in patients previously treated with Zolgensma since the therapy was first approved in 2019, although those reports should not be interpreted as nine confirmed treatment-caused cancers.
The distinction is central to understanding the regulatory development. The FDA has identified molecular evidence linking vector integration to two tumours and is investigating another GLI1-associated case, but available information does not establish that every reported tumour following Zolgensma was caused by the gene therapy. The agency also said the overall benefits of both Zolgensma and the related Itvisma continue to outweigh their potential risks for their approved uses while it evaluates whether further regulatory action is warranted.
Why do the GLI1 findings change the Zolgensma safety discussion beyond a theoretical warning?
Zolgensma is an adeno-associated virus vector-based gene therapy designed to deliver a functional copy of the survival motor neuron 1 gene to patients with spinal muscular atrophy caused by bi-allelic SMN1 mutations. The United States indication covers pediatric patients younger than two years, and the treatment is administered as a one-time intravenous infusion.
AAV vectors have generally been developed with the expectation that much of the therapeutic genetic material remains episomal rather than becoming permanently integrated into the host genome. Integration can nevertheless occur at low frequency, which is why Zolgensma’s existing prescribing information already warns of a theoretical risk of tumorigenicity if AAV vector DNA integrates into genomic DNA.
The November 6 communication changes the nature of that risk discussion because the FDA is no longer dealing solely with a theoretical mechanism. In two patients who developed tumours, molecular analysis found Zolgensma vector DNA integrated into the host genome, with the treatment construct’s promoter positioned upstream of or within GLI1.
GLI1 is an oncogene involved in cellular growth signalling. Abnormal activation or overexpression can contribute to tumour development, making an integration event capable of increasing GLI1 activity biologically important. The FDA said three Zolgensma tumour cases involved the GLI1 oncogene, although the mechanism responsible for GLI1 overexpression in the third case remains under investigation.
The molecular findings do not establish a predictable incidence of cancer after Zolgensma, but they provide stronger evidence that the integration mechanism described in the product label can occur in a clinically consequential setting.
What do nine reported tumours among more than 6,000 Zolgensma-treated patients actually tell us?
Novartis has said more than 6,000 patients have received Zolgensma through clinical trials, managed-access programmes and commercial treatment since launch. The company has received nine reports of tumours among previously treated patients, with two so far showing evidence of vector DNA integration into patient DNA.
It would be misleading to divide nine by the treated population and present the result as the incidence of Zolgensma-associated cancer. Post-marketing adverse-event reports are not a controlled epidemiological dataset, and several important pieces of information remain incomplete, including tumour type, timing, background risk factors, duration of follow-up and molecular analysis for most reported cases.
Spinal muscular atrophy patients treated during infancy can also live for many years after receiving a single gene therapy dose, which makes long-term surveillance particularly important. A safety event arising years later may not have appeared during the relatively limited follow-up available when the therapy was initially approved.
The two molecularly linked cases therefore matter more scientifically than the raw count of nine reports. They provide mechanistic evidence that the therapeutic vector became integrated at an oncogenic locus in tumour tissue, while the remaining reports require further investigation before their relationship to treatment can be established.
Novartis has said the two patients with demonstrated vector integration underwent treatment for their tumours and continue to be monitored. The company also said it notified health authorities and is continuing to work with regulators as the cases are evaluated.
Why is the FDA extending the safety concern to Itvisma when no tumours have been reported after Itvisma?
The FDA has determined that the potential risk of neoplasia is relevant not only to Zolgensma but also to Itvisma because the two Novartis products contain the same vector and gene construct. Itvisma, or onasemnogene abeparvovec-brve, was approved in the United States in November 2025 for adults and children aged two years and older with spinal muscular atrophy and a confirmed SMN1 mutation.
The important difference is administration. Zolgensma is given intravenously to children younger than two years, while Itvisma is administered intrathecally into the spinal canal and extends the onasemnogene abeparvovec franchise to older patients.
The FDA said no cases of neoplasia have been reported following Itvisma administration to date. Its decision to apply the risk assessment to both therapies is therefore based on the shared vector and construct rather than observed Itvisma tumour cases.
This distinction is commercially and clinically important because Itvisma has become a new growth component for Novartis. The company reported $404 million in third-quarter 2026 sales for the combined Zolgensma Group, up 35% at constant currencies, with nine-month sales reaching $1.071 billion. Novartis attributed the latest growth partly to Itvisma launch momentum while saying Zolgensma continued to lead in newly diagnosed SMA populations.
Any regulatory response that affects both products could therefore influence a growing franchise rather than only an older pediatric therapy. The scale of that impact cannot yet be determined because the FDA has not announced a label change, restriction, boxed warning, additional testing requirement or other completed regulatory action related to the new tumour evidence.
What regulatory action could the FDA consider after confirming insertional oncogenesis?
The FDA said it is expeditiously evaluating whether the existing Zolgensma and Itvisma prescribing information and long-term follow-up programmes are adequate to investigate and mitigate the identified risk. That wording indicates an active regulatory review, not a predetermined decision.
Both labels already describe a theoretical risk of tumorigenicity caused by AAV vector DNA integration. Zolgensma also carries a separate boxed warning for serious liver injury and acute liver failure, including fatal cases, illustrating that its risk-management framework has evolved as post-marketing evidence accumulated.
The new question is whether a warning framed around theoretical tumorigenicity remains adequate after molecular evidence of product-related insertional oncogenesis has been identified. Potential regulatory responses could range from revised warning language and strengthened long-term monitoring requirements to additional post-marketing studies, although the FDA has not said which measures, if any, it will require.
The regulator specifically said patients and clinical-trial participants receiving these products should be monitored for tumour formation. Novartis is already conducting long-term follow-up safety studies, and those programmes may now become increasingly important for estimating timing, tumour characteristics and whether identifiable clinical or biological factors modify risk.
Any regulatory decision will also need to account for the severity of untreated spinal muscular atrophy. The FDA has explicitly maintained that the benefits of Zolgensma and Itvisma continue to outweigh their potential risks in their approved populations, which is materially different from concluding that the products should no longer be used.
Why does spinal muscular atrophy make the benefit-risk assessment unusually difficult?
Spinal muscular atrophy is a genetic neurodegenerative disorder caused by insufficient survival motor neuron protein, leading to progressive loss of motor neurons, muscle weakness and, in severe forms, respiratory failure and death. Treatment delivered early in life can substantially change the disease trajectory because motor neurons lost before therapy cannot simply be restored later.
Zolgensma was therefore developed for an unusually high-stakes clinical setting. Its one-time gene-replacement approach can provide a functioning SMN1 gene before extensive irreversible motor-neuron loss occurs, and the therapy has become an important option alongside other disease-modifying SMA treatments.
A possible long-term malignancy risk must be interpreted against that immediate and potentially life-altering benefit. The appropriate comparison is not between gene therapy and no medical risk, but between different therapeutic strategies for a severe inherited disease with major consequences if treatment is delayed or insufficient.
At the same time, successful SMA treatment has changed the safety horizon. Children who might historically have faced severely shortened survival can now live substantially longer, increasing the importance of understanding adverse effects that may emerge years after a one-time genetic intervention.
That creates a characteristic gene-therapy challenge: a treatment administered once may create benefit for years, but its biological consequences also require surveillance over an unusually long period.
Does the Zolgensma finding raise questions for the wider AAV gene therapy industry?
The FDA is explicitly looking beyond Novartis. Its November communication said it is closely monitoring neoplasia risk in other similar AAV vector-based gene therapy programmes and cited two published 2026 reports of AAV integration-driven oncogenesis involving separate investigational programmes.
One involved a GLI1 integration event associated with the FBX-101 programme, while another involved integration affecting the PLAG1 oncogene in the RGX-111 programme. Those cases involve different diseases, vectors, administration approaches and development programmes, so they should not be treated as evidence that all AAV gene therapies carry the same magnitude of cancer risk.
They do, however, make insertional oncogenesis increasingly difficult to treat solely as a theoretical class concern. Regulators and developers may now face greater pressure to characterize integration profiles, investigate tumour tissue when malignancies arise and maintain surveillance extending many years beyond dosing.
The industry implications could include greater scrutiny of vector design, promoter selection, dose, biodistribution and tissue-specific exposure. Developers may also need stronger plans for obtaining tumour samples and genomic sequencing if malignancies develop during long-term follow-up.
None of these considerations establishes that AAV gene therapy as a class has an unacceptable benefit-risk profile. The FDA continues to approve and regulate AAV products across serious diseases, but the emerging molecular evidence may alter how regulators quantify and monitor a risk that historically was considered uncommon and largely theoretical.
How important is the Zolgensma franchise to Novartis after the Itvisma launch?
The combined Zolgensma Group generated $1.071 billion in sales during the first nine months of 2026, up 13% at constant currencies, and third-quarter revenue accelerated to $404 million. Novartis said more than 6,000 patients have been treated with Zolgensma globally since launch.
The business is meaningful but not large enough to determine Novartis’ overall financial performance by itself. Third-quarter group net sales reached $14.52 billion, while major franchises including Kisqali, Kesimpta, Pluvicto and Scemblix are growing rapidly and provide substantial portfolio diversification.
For the SMA franchise, however, the safety investigation arrives at an important moment because Itvisma had begun expanding the eligible population beyond the under-two age limit associated with intravenous Zolgensma. A risk assessment applicable to both therapies could influence prescribing discussions, long-term monitoring requirements and the trajectory of that expansion.
Novartis American depositary receipts finished November 6 at roughly $162.6, modestly higher for the session despite the FDA communication. The movement should not be interpreted as evidence that investors consider the issue immaterial because the stock reflects Novartis’ much broader product portfolio, pipeline and financial outlook.
The commercially relevant milestone will instead be the regulator’s next formal action, particularly whether existing prescribing information is judged sufficient or whether stronger warnings and follow-up requirements are imposed.
What remains unknown after the FDA’s Zolgensma malignancy warning?
The largest unanswered question is the true magnitude of risk. Nine post-marketing tumour reports do not establish incidence, and only a subset has undergone the molecular analysis needed to assess whether vector integration contributed to tumour development.
Timing is another important uncertainty. Public information remains limited regarding exactly when each tumour developed after treatment, whether cases share particular anatomical or pathological characteristics, and whether patient age, dose, vector distribution or other factors influence susceptibility.
The third question concerns Itvisma. The FDA considers the risk biologically applicable because the product uses the same vector and construct, yet no neoplasia cases have been reported after Itvisma treatment. Its more recent approval and shorter post-marketing exposure mean absence of reported cases cannot yet establish absence of long-term risk.
Regulators will also need to determine whether the observed GLI1 integration events represent rare stochastic events or reveal a reproducible pattern that can be investigated through additional genomic analysis. That distinction could influence whether risk-management measures remain focused on surveillance or lead to broader questions about vector and construct design.
For Novartis, the immediate challenge is therefore not proving that every reported tumour is unrelated or assuming that all nine were caused by Zolgensma. It is helping regulators establish the boundaries of a safety signal that now contains direct molecular evidence in two patients.
The November 6 FDA communication marks an important change in that assessment. Zolgensma’s label has long acknowledged theoretical tumorigenicity from AAV integration. Investigators have now documented vector integration at the GLI1 oncogene in tumour tissue, providing evidence that the mechanism can occur in treated patients.
Whether that finding remains an exceptionally rare long-term complication or leads to a material reassessment of AAV gene therapy risk will depend on molecular testing of additional cases, long-term surveillance and the FDA’s regulatory review. For a therapy capable of fundamentally altering the course of a devastating genetic disease after one dose, the answer will matter well beyond Novartis and spinal muscular atrophy.
