Zolgensma tumour risk moved from a label warning to a reported safety signal on 5 November 2026. Novartis said it had received nine tumour reports in treated patients since the 2019 approval. In two cases, the company said, there is evidence that vector DNA integrated into the patient’s own DNA. The UK National Multidisciplinary Team for Zolgensma describes two tumours as specifically related to treatment. The US label says a causal relationship has not been established based on tumour analyses.
This article sets out what is confirmed, what rests on secondary reporting, and what is unknown. Novartis has not published full case details. Several points therefore depend on trade and advocacy-group accounts, and each is flagged.
What did Novartis disclose on 5 November 2026?
Novartis gave its statement to the spinal muscular atrophy (SMA) community through SMA Europe, which published the company text. Novartis said more than 6,000 patients have been treated across trials, managed access and commercial use. It said it had received nine reports of tumours in previously treated patients since 2019.
Two of those cases show evidence of vector DNA integration, according to the company. Both patients were treated and remain under monitoring. Novartis said it notified health authorities in each case. It also said the potential tumour risk from vector integration is already in the prescribing information.
STAT News reported that Novartis told neurologists earlier in the week. STAT described seven of the nine reports as previously unknown. BioSpace reported that seven reports have not been causally linked to the therapy. The coverage read here does not explain how the seven relate to the two integration cases.
Tumour types, patient ages and the interval from dosing were not given in any source read. A search on 6 November 2026 found no published breakdown.
Why does vector integration matter for a one-time therapy?
Zolgensma is given as a single infusion. The US label indicates it for paediatric patients under two years of age with SMA and bi-allelic SMN1 mutations. The UK summary of product characteristics lists two groups of patients with 5q SMA. Both have a bi-allelic SMN1 mutation, and one group has up to three SMN2 copies.
A one-time dose means any effect of the vector persists with the child for life. The label says random integration of vector DNA has been reported with Zolgensma. It also cites reports with other AAV gene therapies. It calls the clinical relevance of individual events unknown.
This is why Zolgensma tumour risk is watched closely by clinicians and families. Zolgensma was approved in 2019, so the earliest commercial recipients are several years past dosing. None of the sources read gives a time window in which such events would be expected.
Does Zolgensma cause tumours? What the evidence supports
The evidence on Zolgensma tumour risk is partial. The label says the clinical relevance of individual integration events is unknown. It adds that such events could potentially contribute to a risk of tumorigenicity. That is a statement of possibility, not a finding of cause.
The two tumours UK clinicians link to treatment
SMA UK’s Zolgensma page carries an update from the National Multidisciplinary Team for Zolgensma, dated 4 November 2026. It said pathology testing had been completed on the tumours of three of the nine children. Two tumours were described as “specifically related” to Zolgensma treatment, and one “appears to be unrelated”.
Novartis describes two cases as showing evidence of vector DNA integration. Neither statement says whether these are the same two children. The update did not give tumour types or timing. The pathology itself has not been published.
The seven reports without a causal link
Pathology results for the remaining six children were not reported in the update. The lack of a causal link at this stage is not evidence that none exists. It can reflect limited information. The US label notes that in some cases limited information was available.
European Biotechnology Magazine noted that no reliable incidence rate can be calculated. Nine reports among more than 6,000 patients cannot show a rate. Reports of this kind have no comparison group. They also do not say how many tumours would occur in untreated children with SMA.
What do the US and UK labels say about Zolgensma tumour risk?
The US prescribing information was revised in June 2026. Section 5.6 is titled AAV Vector Integration and Risk of Tumorigenicity. It describes the vector as a recombinant, non-replicating AAV9 whose DNA “persists largely in episomal form”. It then states there is “a theoretical risk of tumorigenicity”.
The same section says random integration of vector DNA has been reported with Zolgensma. The label continues: “Cases of tumor have been reported in patients who received ZOLGENSMA post-approval.”
The boxed warning concerns serious liver injury and acute liver failure. It does not concern tumours.
The UK position reads differently. The emc summary of product characteristics was revised on 5 October 2026. Section 4.4 also describes a theoretical risk of tumourigenicity. It then says: “So far, no cases of malignancies associated with onasemnogene abeparvovec treatment have been reported.”
That sentence predates the 5 November disclosure. The page read on 6 November 2026 still showed the 5 October revision. Section 5.3 states that genotoxicity and carcinogenicity studies have not been conducted. The product carries a black triangle for additional monitoring, and its first GB authorisation date is 1 January 2021.
Itvisma is the intrathecal onasemnogene abeparvovec product for patients aged two years and over. Its US label, revised June 2026, carries the same theoretical-risk wording in section 5.6. Section 6.2 of that label, as read on 6 November 2026, lists no tumours.
What did the 2023 case report show?
A tumour in a Zolgensma-treated child was published before this disclosure. Retson and colleagues described a low-grade epithelioid neoplasm of the spinal cord in Molecular Therapy in 2023. The child was treated at about two months of age, before SMA symptoms appeared. Tumour-related symptoms began at 16 months.
The authors detected vector nucleic acid in many, but not all, tumour cells. Their test could not distinguish integrated from episomal vector DNA. Integration site analysis found no high-confidence sites. The authors called that result inconclusive because little tissue and DNA remained.
The paper did not establish a causal link. The authors called for further research and for tumour surveillance protocols in patients with SMA.
What have regulators said about Zolgensma tumour risk?
SMA Europe said the relevant regulatory authorities are actively involved and assessing the evidence. It did not name them. The National Multidisciplinary Team said further information is awaited from the regulatory authorities. A search on 6 November 2026 found no FDA, European Medicines Agency or MHRA statement on the reports.
For UK families, the team said no changes to routine care or follow-up are recommended. It added that “no additional scans, tests or monitoring are needed at this time”. SMA Europe said it had urged Novartis to communicate publicly in good time. It advised families with individual concerns to speak to their treating healthcare team. Novartis said it continually evaluates new safety information and assesses all reported cases.
What would settle the question, and when?
Three developments would move the answer. None has a published date.
- Pathology and integration results for the six cases not yet tested.
- Tumour types, patient ages and the interval from dosing for all nine reports.
- Any regulator assessment of the signal, including a change to a label.
Until then, the supportable answer on Zolgensma tumour risk is narrow. UK clinicians describe two tumours as treatment-related, and Novartis reports integration evidence in two cases. How integration relates to those tumours has not been published. For the other reports, causality is unknown. The label has warned of a theoretical risk since before this disclosure.
Readers following another gene therapy can read this analysis of uniQure’s AMT-130 four-year data. Regulatory outcomes for approved products are logged on the FDA drug approval decisions tracker.
