Safety and Efficacy of Intravenous Onasemnogene Abeparvovec Gene Therapy in Patients With Spinal Muscular Atrophy Type 1: Interim Analysis From LT-001

Major Points and Findings:

This paper does not concern stem cells. It reports the longest follow-up published so far of onasemnogene abeparvovec (Zolgensma), the one-time gene replacement therapy given by intravenous infusion. We include it because families often ask whether the gains from the approved treatments last and what happens years later.

Thirteen children from the original START trial were followed for up to ten years after a single infusion. All thirteen were alive at the last contact, twelve of thirteen were free of permanent ventilation, and every child in the therapeutic-dose group still had the highest motor milestone they had reached in the original trial. Untreated type 1 disease had a median survival of 6 to 8 months, so the result is important. The study was very small, unblinded and industry-funded, with no comparison group, and most of the children were also taking nusinersen or risdiplam.

Aim:

To report long-term safety (the primary outcome) and survival free of permanent ventilation in patients with symptomatic SMA type 1 who received onasemnogene abeparvovec in the phase 1 START trial, with follow-up extending up to ten years after dosing.

Methods:

LT-001 (NCT03421977) is a long-term follow-up of START (NCT02122952), run at Nationwide Children’s Hospital in Columbus, Ohio. The first child was dosed on 13 May 2014. The present report is a planned interim analysis with a data cut-off of 1 July 2024.

Thirteen of the fifteen START patients enrolled, three from the low-dose group (6.7 × 10¹³ vg/kg) and ten from the proposed therapeutic dose group (equivalent to 1.1 × 10¹⁴ vg/kg). All had symptomatic SMA type 1, biallelic SMN1 mutations or deletions and two SMN2 copies. Mean age at dosing was 6.3 months in the low-dose group and 2.8 months at the therapeutic dose (range 0.9 to 7.1 months). At entry to LT-001 all thirteen had hypotonia, twelve had limb weakness, ten had had pneumonia or a respiratory syndrome, nine had swallowing or feeding difficulties, seven were on ventilatory support and eight on feeding support.

Follow-up: The protocol specifies five years of annual in-person visits, then annual telephone contact for up to ten more years, with records requested from local doctors. Mean follow-up was 9.9 years in the low-dose group and 8.3 years at the therapeutic dose. Only serious adverse events and predefined adverse events of special interest were collected. Five therapeutic-dose patients discontinued, two lost to follow-up and three who withdrew consent. All thirteen had at least one protocol deviation, most commonly an assessment that was not carried out.

Most patients started another SMA drug after START. All three in the low-dose group and seven of ten at the therapeutic dose received nusinersen and/or risdiplam. Risdiplam was still being taken by all but one of them at the data cut-off. The reasons for adding a second treatment were not recorded.

Results:

Survival and ventilation: All 13 patients (100%) were alive at the data cut-off or at their last visit. Twelve of 13 (92%) were free of permanent ventilatory support. The one exception, in the low-dose group, has had a tracheostomy since day 507. Only five patients (all therapeutic dose) needed no ventilatory support at all. Eight were using bilevel positive airway pressure at night or intermittently, and eleven of thirteen used cough assist.

All three low-dose patients needed non-oral feeding for 76% to 100% of daily intake and all had a gastrostomy with Nissen fundoplication. Six of ten therapeutic-dose patients needed feeding support and four remained independent of it.

Motor milestones: All ten therapeutic-dose patients kept the highest milestone they had achieved in START. Two achieved a new milestone during long-term follow-up (standing with assistance), and neither was on add-on therapy. Across the therapeutic-dose group the highest milestones were sitting without support in five, standing with assistance in three, and walking alone in two. In the low-dose group, one patient sat with support and two demonstrated no motor milestones at all.

Safety: Serious adverse events occurred in 11 of 13 patients (85%), most often acute respiratory failure, dehydration and pneumonia. None were judged related to the gene therapy, and none led to discontinuation or death. Adverse events of special interest occurred in 4 of 13 patients (31%). These were transient thrombocytopenia categories (gastrointestinal haemorrhage and procedural haemorrhage, one each, both recovered, neither with a low platelet count), cardiac events (one cardiac arrest in the low-dose group and one episode of breathlessness, both during acute respiratory failure and both resolved within 1 and 6 days), and new neurological findings in two patients (anal incontinence, muscle weakness). There were no cases of liver toxicity, thrombotic microangiopathy, dorsal root ganglion toxicity, new cancers, or new blood or autoimmune disorders. One patient had a liver enzyme (ALT) above twice the upper limit of normal at year 2, which normalised. Most patients had at least one blood pressure, pulse or respiratory rate reading flagged as potentially clinically significant, but no adverse events followed from them, and heart tracings and echocardiograms showed nothing of concern.

Conclusions:

A single intravenous dose of onasemnogene abeparvovec was followed by survival and preserved motor function up to ten years later in this small group, with no late safety signals attributable to the treatment. The limitations the authors list are a small sample, patients lost to follow-up, no blinding, no comparator arm, collection of serious events only, and the confounding effect of add-on nusinersen or risdiplam in most patients. The reasons for adding those drugs were never recorded.

On combination treatment, they state that there is no evidence from trials or real-world studies that combining treatments works better than gene therapy alone. The observational reports suggesting it are limited by non-randomised designs and mixed patient groups.

Background Information:

The authors note that motor neuron loss in SMA is irreversible and rapid in newborns with one or two SMN2 copies, so pre-symptomatic treatment remains the goal. Their finding is that children who were already symptomatic can still benefit. It does not imply that delaying treatment is harmless.

The ventilation and feeding figures should be read alongside the survival figure. Ten years on, most of these children still use night-time ventilation or cough assist, and most still need some feeding support. Gene therapy changed the course of a fatal disease, but it did not restore normal health.

The work was funded by Novartis Pharma AG, the manufacturer, and five of the eight authors are Novartis employees or hold company stock. The funder was involved in design, data collection, analysis, interpretation and writing.