Part 3 · Treatment & care
Tofersen, gene-targeted
Tofersen is an antisense drug for ALS caused by a SOD1 variant. The left panel follows one molecule from the spinal fluid to the SOD1 message; the right panel follows the VALOR trial week by week. Choose a patient group and which arm to highlight, then press Play.
Schematic, not to scale.
Who it is labelled for, and when
Adults with ALS and a SOD1 mutation
Approved under accelerated approval "based on reduction in plasma neurofilament light chain". Continued approval may depend on confirmatory trial(s). 172
FDA accelerated approval: 25 April 2023. The sponsor named the presymptomatic phase 3 ATLAS trial as the confirmatory study. 61,149
Authorised 29 May 2024
Positive CHMP opinion 22 February 2024; authorised under exceptional circumstances, with obligations for long-term data, a study in presymptomatic carriers and a registry with yearly reports. 160
Health Canada authorised it with conditions, on 28 February 2025 according to Health Canada summaries that were not read in full. Its drug database lists it on the market from 14 May 2025. 225,227
100 mg intrathecal, by lumbar puncture
100 mg (15 mL) over 1 to 3 minutes, after about 10 mL of CSF is removed. Three loading doses 14 days apart, then one every 28 days. If the second loading dose is missed, it is given as soon as possible and the third 14 days later. 160,172
EAN 2024
"Offer tofersen as first-line treatment" for progressive SOD1-ALS (strong), discuss the serious adverse events, and weigh benefit against treatment burden carefully in people whose disease progresses slowly. 506
Pregnancy: there are no adequate human data; animal studies at about 4 to 20 times human exposure showed no embryofetal effects, and tofersen was found in mouse milk. 172 Quoted from the label as published guidance, not advice for any person. Testing for SOD1 is covered in Genes: C9orf72, SOD1 and more.
Adverse events
Most common in VALOR (tofersen n = 72 vs placebo n = 36): pain 42% vs 22%, fatigue 17% vs 6%, arthralgia 14% vs 6%, raised CSF white cells 14% vs 0%, myalgia 14% vs 6%, raised CSF protein 8% vs 3%. 172 Lumbar-puncture effects were about equal in both arms (80.6% each; procedural pain 56.9% vs 58.3%). 160
| Serious neurological event | US label 147 people, median 119.4 weeks | EMA count, any severity of 147 | Lovett 2025, serious events across the trials |
|---|---|---|---|
| Myelitis | 6 people with myelitis and/or radiculitis: 2 stopped tofersen and recovered fully, 4 recovered without stopping | 4 (2.7%) | 4 |
| Radiculitis | 4 (2.7%) | 2 | |
| Aseptic or chemical meningitis | 2 serious (1 stopped tofersen) | 6 (4.1%) | 2 |
| Papilloedema and/or raised intracranial pressure | 4 people, all resolved with standard care; none stopped | papilloedema 7 (4.8%) | 4 |
| In total | 10 people (about 7%) had 12 events; none life-threatening or fatal; 3 led to stopping; all but 1 resolved |
Sources: label 172, EMA assessment 160, trial pooled report 302; the VALOR paper gives neurological serious adverse events in 7% of tofersen recipients. 335 Lovett 2025 notes that myelitis, radiculitis, raised intracranial pressure and papilloedema had not been reported with any antisense oligonucleotide before. 302 In a German cohort, 73% (11 of 15) had raised CSF cells and 9 of 10 had antibody production inside the CNS, which the authors read as a possible autoimmune reaction. 535
Outside the trial: real-world cohorts
| Cohort | People | What it reported |
|---|---|---|
| German early access, case series | 6 | Neurofilament fell clearly, including in slower-progressing p.(Asp91Ala) carriers. 329 |
| German early access, 12 months, 10 centres | 24 | Median serum NfL 78 → 36 pg/mL; CSF pNfH 2226 → 1151 pg/mL; median ALSFRS-R 38 → 35 (0.11 points/month). Two drug-related serious adverse events with myeloradiculitis symptoms. 535 |
| Single US centre | 7 | Mean serum NfL −57.9%, CSF pNfH −67.6%; mean ALSFRS-R change +1.1. 476 |
| France (FORSLA), retrospective | 46 | Progression slowed from 0.53 to 0.22 points/month by 12 months. Against a matched historical cohort, mortality was lower (adjusted HR 0.34, 95% CI 0.12 to 0.91) but progression rate did not differ (P = 0.30). 164 |
None of these has a concurrent control group, so they describe experience in practice rather than test whether tofersen works.
Long-term follow-up
148 weeks: early vs delayed start
−9.9 vs −13.5ALSFRS-R points
Slow vital capacity −13.8% vs −18.1%; ALSAQ-5 17.0 vs 22.5. 46 people completed, with follow-up up to 192 to 276 weeks. The authors report survival longer than the expected natural history of SOD1-ALS and that every serious neurological adverse event was reversible. Survival was compared with natural history, not with a concurrent placebo group. 336
Before symptoms: the confirmatory trial
158 presymptomatic carriers of SOD1 variants enrolled. Carriers are randomised to tofersen or placebo when plasma NfL rises (at least 44 pg/mL and up by at least 10 pg/mL), and the primary endpoint is the share who develop clinically manifest ALS within 12 months. 50,123 Estimated primary completion: August 2027; no results yet. 123 More in Toward a cure.
The controversy: a surrogate endpoint
The case made for approval
- Tofersen lowered its target, CSF SOD1, and plasma NfL fell by about half or more. 172
- NfL is released when axons are damaged, so a fall is read as less ongoing injury. 191
- People who started earlier lost fewer ALSFRS-R points at 52 and at 148 weeks. 335,336
The reasons for caution
- The randomised clinical primary endpoint was missed (P = 0.97). 335
- NfL rises with nerve damage of any cause; it is not specific to ALS. 262
- The early-vs-delayed comparisons are open-label and nominal; the label says they "should be interpreted with caution". 172
- FDA approval is accelerated and the EU authorisation is under exceptional circumstances; ATLAS is meant to confirm. 123,160,172