Phenytoin in acute optic neuritis


The original phenytoin study, published by the team at University College London, demonstrated that the use of phenytoin (- sodium channel blocker) for three months after the onset of acute optic neuritis was neuroprotective.

Challenges in Managing Progressive Multiple Sclerosis (Transcript)
Slide from Medscape education

We have now analysed the blood samples taken for neurofilament analysis and found that neurofilament heavy chain levels (as opposed to light chain levels) reduced only in those on phenytoin (see below). The lack of reduction in neurofilament light chain levels may be a fluke, or it may be that sodium channel blockers work more effectively in stabilizing the large myelinated axons, where the heavy chain is more abundant.

More work is to explore the differences and similarities between the light and heavy chain when it comes to neurodegeneration in MS, and the impact of sodium channel blockers.

Figure: No difference in serum NFL levels between those on phenytoin or placebo
Figure: A reduction in serum NfH levels in the phenytoin arm by three months compared to placebo

Many of you may be on sodium channel blockers for neuropathic pain control. To date there is positive animal data on sodium channel blockers as neuroprotectants preventing disability in mice, neurofilament heavy chain data in Secondary Progressive MS (SPMS), oxcarbazepine in SPMS and now this in Optic neuritis. It is now time to look at funding a larger Phase III study to evaluate this in Optic Neuritis and Progressive MS – anyone game for this???


Eur J Neurol . 2020 Oct 14. doi: 10.1111/ene.14591. Online ahead of print.

Neurofilament Results for the Phase II Neuroprotection Study of Phenytoin in Optic Neuritis

Rhian Raftopoulos Jens KuhleDonna Grant Simon J Hickman Daniel R AltmannDavid Leppert Kaj BlennowHenrik Zetterberg Raj Kapoor Gavin Giovannoni Sharmilee Gnanapavan

Background: A randomized trial of phenytoin in acute optic neuritis (ON) demonstrated a 30% reduction in retinal nerve fiber layer (RNFL) loss with phenytoin versus placebo. Here we present the corresponding serum neurofilament analyses.

Methods: Eighty-six acute ON cases were randomised to receive phenytoin (4-6 mg/kg/day) or placebo for 3 months, and followed up for 6 months. Serum was collected at baseline, 3, and 6 months for analysis of neurofilament heavy chain (NfH) and light chain (NfL).

Results: Sixty-four had blood sampling. Of these, 58 and 56 were available at 3 months, and 55 and 54 were available at 6 months for NfH and NfL, respectively. There was no significant correlation between serum NfH and NfL at the time points tested. For NfH, the difference in mean placebo – phenytoin was -44 pg/ml at 3 months (p=0.019), and -27 pg/ml at 6 months (p=0.234). For NfL, the difference was 1.4 pg/ml at 3 months (p=0.726), and -1.6 pg/ml at 6 months (p=0.766).

Conclusions: At 3 months, there was a reduction in NfH, but not NFL in the phenytoin versus placebo group, while differences at 6 months were not statistically significant. This suggests a potential neuroprotective role for phenytoin in acute ON, with the lower NfH at 3 months, when levels secondary to degeneration of the anterior visual pathway are still elevated, but not at 6 months, when levels have normalized.

About the author

Neuro Doc Gnanapavan


    • This is because PROXIMUS was ahead of its time…NfL was the outocome measure because then we thought it was a marker of nerve loss. However it is an excellent marker of inflammation induced nerve loss. Get rid of the inflammation with a DMT and NfL goes just like an MRI lesion goes. In PROXIMUS people were on DMT therefore the levels were reduced and so it was not possible to recruit people to the trial. So the levels were dropped to give an indication that something was going wrong and whilst an effect on Neurofilament levels were not seen there was significant clinical benefit…so maybe it did work

      • Retrospectively (the Retrospectiscope is a most useful instrument don’t you agree?) would be great to repeat the trial using nfH and perhaps a starting dose of 300mg bd (same as recommended starting dose as anticonvulsant)? The 150mg tablets probably only exist for the purposes of dose titration upwards and prescribing in children or renal impairment…

        Great idea and well done for trying but 🤔 do trials ever get a second chance?

        Coi: biased here as been happily taking oxcarbazepine 300mg bd ever since trying to join PROXIMUS trial but failing to qualify as sadly no dmt at the time

    • Interestingly we collected this piece of information on the lamotrigine trial but I don’t know what the result of it was!

      • Perhaps it might be worth asking Dr Mark Baker at Blizard Institute who is looking into this area of heat intolerance and MS

  • “A randomized trial of phenytoin in acute optic neuritis (ON) demonstrated a 30% reduction in retinal nerve fiber layer (RNFL) loss with phenytoin versus placebo.“

    30% reduction doesn’t sound that great to me. Reminiscent of the relapse reduction rate of the CRAB drugs. Surely we’d want to see a much better result for a neuroprotective drug ie 90% reduction in nerve tissue loss. Are there any more promising drugs that phenytoin to provide better neuroprotection? Would remyelination drugs do a better job at protecting exposed nerve tissue? However, the recent culling of Opicinumab by Biogen demonstrates how far we are from effective remyelination therapies!

    • Please be aware that the phenytoin was started on average 8 days into the lesion development. If it was on board from the beginning the 30% may end-up close to 90%. We don’t know.

      • Thanks. I assume phenytoin is only providing neuroprotection for nerves damaged by focal inflammation ie a lesion, rather than providing neuroprotection for nerves damaged by smouldering MS?

        • Not necessarily as the mode of action of Na channel blockers will work in all demyelinated axons. The effect seen in the original trial was on retinal nerve fibre layer thickness and macular volume. There was no effect on VEPs but this may be the result of only a short duration of treatment.

  • Did you correct for multiple comparisons? If not, then these p-values do not mean what you think they do, and the result is likely a false positive.

    • The phenytoin versus placebo outcomes don’t require correction for multiple comparisons. The remainder were exploratory analysis looking for associations between the other measures such as visual charts, VEP, optic nerve MRI that were uncorrected. It’s arguable from a statistical point of view whether blanket corrections for multiple comparisons are correct when you’re looking at a few variables unlike the ones that large data point datasets deal with. But what it showed was what we already knew from stand alone ophthalmology studies with neurofilaments in optic neuritis!

      • You report four p-values in the abstract. Did you have any reason to believe that the 3-month NfH comparison would be positive, and the other three comparisons would be negative? If you did not, then you need to correct for multiple comparisons.

        The correction for multiple comparisons would account for the probability that at least one of the four comparisons would be positive by chance.

        Here is another way of asking the question. Imagine that the six month NfL comparison was the only one that had been positive. Would you have considered the trial a success in that case?



Recent Posts

Recent Comments