# Can a Common Seizure Medication Protect the Optic Nerve During Optic Neuritis? A Patient-Friendly Look at the Phenytoin Trial

This phase II randomized clinical trial investigated whether phenytoin, a medication that blocks sodium channels, could protect the retinal nerve cells of people experiencing acute optic neuritis (inflammation of the optic nerve). Researchers reanalyzed optical coherence tomography scans from 80 patients and found that those treated with phenytoin for 3 months had significantly greater preservation of the macular ganglion cell-inner plexiform layer (mGCIPL)—a key marker of retinal nerve cell health—at 6 months compared with placebo. The treated group showed an average mGCIPL thickness of 73.8 μm versus 67.0 μm in the placebo group, a difference of 6.79 μm (p = 0.006). The study also revealed that mGCIPL measurements were more reliable than the traditional peripapillary retinal nerve fiber layer (pRNFL) measure for detecting neuroprotective treatment effects, and that treatment benefits were most pronounced in patients with worse vision at the start of the study.

# Can a Common Seizure Medication Protect the Optic Nerve During Optic Neuritis? A Patient-Friendly Look at the Phenytoin Trial

## Table of Contents

- Key Points
- Background: Why This Research Matters
- Study Methods: How the Research Was Conducted
- Study Procedures and Measurements
- Statistical Analysis Approach
- Who Participated in the Study
- Key Finding: Phenytoin Protected Retinal Nerve Cells
- Post Hoc Analysis: Does Baseline Vision Matter?
- Comparing Two Measurement Methods: mGCIPL vs. pRNFL
- Sensitivity Analyses: Testing the Strength of the Results
- Clinical Implications: What This Means for Patients
- Study Limitations: What This Study Couldn't Prove
- Recommendations for Patients
- Frequently Asked Questions
- Source Information

## Key Points

- In a phase II trial of 80 patients with acute optic neuritis, those given phenytoin for 3 months had significantly greater preservation of the macular ganglion cell-inner plexiform layer at 6 months than those given placebo.
- The treated group showed an average mGCIPL thickness of 73.8 μm versus 67.0 μm in the placebo group, a difference of 6.79 μm.
- In this trial, mGCIPL measurements were more reliable than the traditional peripapillary retinal nerve fiber layer for detecting neuroprotective treatment effects.
- Treatment benefits were most pronounced in patients with worse vision at the start of the study, with a threshold of about 20/120 or worse.
- The mGCIPL analysis was a post hoc secondary analysis, the final primary analysis included 67 participants, and follow-up was only 6 months.

## Background: Why This Research Matters

Multiple sclerosis (MS) is an autoimmune disease in which the body's immune system mistakenly attacks the central nervous system, including the brain, spinal cord, and optic nerves. This attack damages myelin—the protective coating around nerve fibers—and can eventually lead to permanent nerve damage and disability. Up to 70% of people with MS will experience at least one episode of acute optic neuritis (AON), a sudden inflammation of the optic nerve that can cause eye pain, blurred vision, or even temporary vision loss.

AON is clinically important for two reasons. First, it directly affects a person's vision and quality of life. Second, it serves as a "window into the brain"—because the optic nerve is an extension of the central nervous system, studying AON helps researchers understand how MS relapses damage nerves and whether treatments can protect them. The inflamed optic nerve lesion in AON is histologically (microscopically) identical to the plaques found in the brains of people with MS.

The main goal of current MS treatments is to prevent disability by calming the immune system. However, researchers are increasingly exploring two additional strategies: **remyelination** (repairing the damaged myelin) and **neuroprotection** (preventing nerve cells from dying in the first place). This study focuses on the latter—specifically, whether a medication called phenytoin can reduce the nerve damage caused by acute optic neuritis.

Phenytoin is an older medication that has been used for decades to treat epilepsy and prevent seizures. It works by blocking voltage-gated sodium channels—tiny gates on nerve cells that allow electrical signals to pass through. During acute inflammation, these channels can open excessively, allowing too much sodium to flood into nerve cells, which can trigger a chain reaction leading to cell death. Researchers hypothesized that by blocking these channels, phenytoin might "calm" the nerve cells and protect them from dying during an AON episode.

An earlier landmark phase II trial (the original phenytoin AON study) had already shown that phenytoin preserved the thickness of the peripapillary retinal nerve fiber layer (pRNFL)—a layer of nerve fibers at the back of the eye—with the phenytoin group showing 30% greater preservation compared with placebo. However, that study used only this one measurement. The current study reanalyzed those same patients using a different, newer structural marker: the **macular ganglion cell-inner plexiform layer (mGCIPL)**.

Why is this distinction important? The mGCIPL contains the actual cell bodies of retinal ganglion cells—the neurons that carry visual information from the eye to the brain. The pRNFL measures only the axons (the long projections) of those cells. Measuring the cell bodies themselves may give a more direct and accurate picture of how many nerve cells survive an inflammatory attack.

This study addressed three specific research questions:

1. Does phenytoin treatment have a neuroprotective effect on mGCIPL thickness?
1. Is mGCIPL superior to pRNFL as a marker of neuroprotection, and can this be validated using electrical activity measurements (electrophysiologic parameters)?
1. Does the severity of baseline visual impairment affect how well phenytoin works?

## Study Methods: How the Research Was Conducted

The phenytoin AON trial (registered as NCT01451593 on ClinicalTrials.gov) was a **randomized, placebo-controlled, double-blind phase II trial** conducted at two centers in the United Kingdom—one in London and one in Sheffield. "Double-blind" means that neither the patients nor the doctors knew who was receiving the real medication versus the placebo (an inactive pill), eliminating bias.

To be eligible for the trial, participants needed to meet the following criteria:

- Age between 18 and 60 years
- A clinical diagnosis of unilateral AON (affecting only one eye), confirmed by a neuro-ophthalmologist
- No previous history of clinical AON in either eye
- Visual acuity of 6/9 or worse in the affected eye (meaning vision was measurably impaired)
- Within 14 days of disease onset before randomization

Participants were randomly assigned in a 1:1 ratio to receive either oral phenytoin (at a dose of 4–6 mg per kilogram of body weight per day) or a matching placebo, which they took daily for 3 months. Randomization was carefully stratified (balanced) across several factors: time from symptom onset, study center, whether the patient had a previous multiple sclerosis diagnosis, whether they were taking disease-modifying treatments, and whether they were receiving corticosteroids. Corticosteroid use—a standard treatment for optic neuritis—was allowed at the discretion of the treating physician and was balanced between the two groups. Participants were followed for 6 months total.

This reanalysis is a post hoc (after-the-fact) secondary analysis of the original trial data. The researchers extracted mGCIPL thickness measurements from the participants' OCT scans at baseline and at 6 months, and then used statistical modeling to determine whether phenytoin had a protective effect on this specific structural measure.

## Study Procedures and Measurements

The study used three major types of measurements:

**1. Optical Coherence Tomography (OCT):** The eye scans were performed using a high-resolution spectral-domain OCT device (Spectralis, Heidelberg Engineering, Germany). This noninvasive imaging technology uses light waves to create detailed cross-sectional pictures of the retina—like an "optical ultrasound" of the eye. Two types of scans were performed:

-
**pRNFL scans:** Captured from a ring of 3.45-mm diameter around the optic nerve head, measuring the layer of nerve fibers that leave the eye.
-
**Macular volume scans:** Captured from a 20 × 20° field using 25 horizontal B-scans, covering the macula (the central part of the retina responsible for sharp, detailed vision).

The **mGCIPL average thickness** was derived from a 1-2-3 mm concentric ring grid centered on the fovea (the very center of the macula). All scans were checked for quality using the OSCAR-IB criteria, a set of standard quality control rules. Scans with signal strength below 25 were excluded, as were scans with significant segmentation failures, poor illumination, or other retinal pathology. One trained rater (S.K.) manually inspected and adjusted the automated segmentation (layer boundary detection) of every scan to ensure accuracy.

**2. Visual Acuity Testing:** Best-corrected high-contrast visual acuity was measured using retro-illuminated Early Treatment Diabetic Retinopathy Study (ETDRS) charts at a distance of 4 meters. Visual acuity was recorded as logMAR values (logarithm of the minimum angle of resolution), where lower values indicate better vision. If the patient could not identify any letters, an acuity score of 1.7 was assigned by the masked researcher.

**3. Visual Evoked Potentials (VEPs):** VEPs measure how quickly and strongly the brain responds to visual stimuli, providing a functional readout of the entire visual pathway from eye to brain. Full-field VEPs to reversing black-and-white checkerboard patterns were recorded at baseline and 6 months, following International Federation of Neurophysiology guidelines. Electrodes were placed on the scalp at standard positions (Oz as recording site, Fz as reference, Cz as ground). Two checkerboard sizes were used:

-
**Large check:** Each square subtended 1° of visual angle
-
**Small check (SC):** Each square subtended 0.25° of visual angle

The key VEP measures were the **P100 latency** (how long it takes the brain to process the visual signal, measured in milliseconds) and the **N75-P100 amplitude** (the strength or size of the brain's visual response, measured in microvolts). Longer latencies and smaller amplitudes indicate more damage to the visual pathway.

## Statistical Analysis Approach

Statistical analysis was performed on a **modified intention-to-treat population**—meaning all randomized participants who had both baseline and 6-month scans were included, regardless of whether they completed treatment as planned. Linear regression models were used to evaluate treatment effects on the 6-month affected eye mGCIPL thickness.

Researchers constructed a basic statistical model including treatment allocation, baseline mGCIPL thickness in both the affected and unaffected eyes, age, sex, and recruitment center. Then they added additional predictors one at a time to see which variables improved the model's ability to predict outcomes. The final "best fit" model included:

- Baseline unaffected eye mGCIPL thickness
- Demographic variables (age, sex)
- Recruitment center
- Electrophysiologic variables (6-month VEP amplitude and latency using small check stimuli)
- Time between symptom onset and baseline assessment
- Time between corticosteroid administration (if any) and baseline assessment
- Baseline visual acuity (logMAR) in the affected eye

Patients with unrecordable VEP responses at 6 months were excluded from the primary analysis, as their absence likely reflects severe damage that could skew results. Model performance was compared using standard metrics including the Akaike information criterion corrected (AICC), Bayesian information criterion (BIC), coefficient of determination (R²), and root mean square error (RMSE).

## Who Participated in the Study

Eighty participants were recruited into the study between February 2012 and May 2014. Thirty-nine were randomly assigned to receive phenytoin, and 41 received placebo. The mean age of participants was 33.59 years, and 70% were female—a typical demographic profile for MS-related optic neuritis, which disproportionately affects younger women.

The baseline characteristics were well balanced between the two groups. In the phenytoin group (n = 37 with complete data), the average baseline affected eye mGCIPL thickness was 92.47 μm (SD = 8.51 μm), while in the placebo group (n = 40), it was 88.83 μm (SD = 7.92 μm). Both groups had similar times between symptom onset and assessment (7.89 days for phenytoin vs. 8.38 days for placebo, p = 0.48) and similar baseline visual acuities (logMAR 1.12 vs. 1.00, approximately 20/250 to 20/200 Snellen equivalent).

Of the original 81 participants analyzed in the primary clinical trial publication, one placebo participant was excluded due to a missing 6-month OCT scan, leaving 80. Three additional patients were excluded due to poor-quality macular OCT scans (distributed across both groups). Of the remaining 77, ten more were excluded from the final statistical model because their VEP responses were unrecordable at 6 months. The final primary analysis thus included 67 participants.

For context, the study also excluded 25 potential participants with alternative diagnoses—conditions that can mimic optic neuritis but are not true demyelinating AON. These included functional visual loss (n = 4), sarcoidosis (n = 3), migraine with aura (n = 2), posterior scleritis (n = 2), Leber hereditary optic neuropathy (n = 2), compressive optic nerve lesions (n = 2), uveitis (n = 1), toxic optic neuropathy (n = 1), neuroretinitis (n = 1), central serous retinopathy (n = 1), and optic nerve drusen (n = 1). Careful diagnostic scrutiny of this type is critical because including patients with non-demyelinating conditions could dilute or confound the treatment effects.

## Key Finding: Phenytoin Protected Retinal Nerve Cells

The primary analysis estimated a significantly higher 6-month affected eye mGCIPL thickness in the phenytoin-treated group compared with the placebo group—a difference of **6.79 μm (p = 0.006, SE = 2.35 μm)**. After adjusting for other predictors, the estimated mean mGCIPL thickness was **73.8 μm (SE = 2.40 μm)** for the phenytoin group versus **67.0 μm (SE = 2.21 μm)** for the placebo group.

To put this in plain terms: patients who took phenytoin lost significantly fewer retinal nerve cell bodies during their optic neuritis episode than patients who took placebo. The difference of nearly 7 micrometers may sound tiny, but the mGCIPL is a very thin layer—thinning of even a few micrometers represents the death of thousands of nerve cells.

In addition to the treatment effect, several other predictors were significantly associated with 6-month mGCIPL thickness:

-
**Baseline affected visual acuity (logMAR):** Worse vision at baseline predicted a thinner mGCIPL at 6 months (β = −5.54 μm, p = 0.010, SE = 2.08 μm)
-
**P100 SC latency at 6 months:** Longer latency (slower signal transmission) predicted thinner mGCIPL (β = −0.32 μm, p < 0.001, SE = 0.07 μm)
-
**N75-P100 SC amplitude at 6 months:** Higher amplitude (stronger signal) predicted thicker mGCIPL (β = 0.78 μm, p = 0.013, SE = 0.30 μm)

Variables that did *not* significantly predict mGCIPL thickness included age (β = −0.03 μm, p = 0.85), sex (β = 3.05 μm, p = 0.30), center (β = 1.61 μm, p = 0.64), time from visual loss to assessment (β = 0.13 μm, p = 0.73), and time from corticosteroid prescription to assessment (β = 0.49 μm, p = 0.86).

Breaking down the VEP associations more concretely: the model predicted that **each 1-millisecond increase** in 6-month VEP P100 latency was associated with a **0.32-μm reduction** in mGCIPL thickness (p < 0.001), while each **1-microvolt increase** in VEP amplitude was associated with a **0.78-μm increase** in mGCIPL thickness (p = 0.013). These strong associations between structure (the thickness of the retinal layer) and function (how well visual signals travel to the brain) validate mGCIPL as a meaningful clinical measure.

## Post Hoc Analysis: Does Baseline Vision Matter?

A key secondary question was whether phenytoin provided more or less protection depending on how severely vision was impaired at the start of the trial. To investigate, researchers added a statistical interaction term (treatment × baseline visual acuity) to the final model, allowing them to estimate treatment effects at different levels of baseline visual acuity.

The results were revealing. While the interaction between the slopes was not statistically significant (β = 5.37 μm per logMAR unit, p = 0.230, SE = 4.42 μm per logMAR unit), the **placebo slope was significantly different from zero** (β = −7.52 μm per logMAR unit, p = 0.0062, SE = 2.64 μm per logMAR unit), while the **phenytoin slope was not** (β = −2.15 μm per logMAR unit, p = 0.5280, SE = 3.47 μm per logMAR unit). In practical terms, this means that in the placebo group, patients with worse baseline vision ended up with much thinner mGCIPL layers—greater vision loss predicted greater nerve cell death. In the phenytoin group, this relationship was largely flattened, implying that phenytoin protected the nerve cells even in patients who started with severe visual impairment.

The researchers estimated treatment effects at various baseline acuity levels, revealing a threshold effect. The estimated treatment benefits (in micrometers of preserved mGCIPL) for different baseline visual acuities were:

- logMAR 0.00 (20/20, normal vision): 1.14 μm effect (p = 0.8275, not significant)
- logMAR 0.5 (20/60): 3.82 μm effect (p = 0.2632, not significant)
- logMAR 0.6 (20/80): 4.36 μm effect (p = 0.1624, not significant)
- logMAR 0.7 (20/100): 4.90 μm effect (p = 0.0873, not significant)
- logMAR 0.75 (20/112): 5.17 μm effect (p = 0.0607, not significant)
- logMAR 0.775 (approximately 20/120): 5.30 μm effect (p = 0.0500, borderline significant)
- logMAR 0.8 (20/126): 5.43 μm effect (p = 0.0409, significant)
- logMAR 1.0 (20/200): 6.51 μm effect (p = 0.0078, significant)
- logMAR 1.5 (20/600): 9.19 μm effect (p = 0.0041, significant)
- logMAR 1.7 (counting fingers level): 10.27 μm effect (p = 0.0076, significant)

This analysis identified a **visual acuity threshold of logMAR ≥ 0.775** (approximately 20/120 Snellen or worse) above which phenytoin demonstrated significant benefits compared with placebo. In other words, the patients with the most severe vision loss at the start of their optic neuritis episode appeared to derive the greatest neuroprotective benefit from phenytoin treatment.

## Comparing Two Measurement Methods: mGCIPL vs. pRNFL

The original phenytoin trial used pRNFL thickness as its primary outcome measure. This reanalysis compared how well each structural marker performed in detecting treatment effects.

For the pRNFL analysis, an adjusted increase of **6.93 μm (p = 0.034, SE = 3.18 μm)** in 6-month affected eye pRNFL thickness was observed with phenytoin treatment compared with placebo. While this was statistically significant, the model performed less well overall. Significant predictors of pRNFL thickness included baseline affected eye logMAR VA (β = −9.82 μm, p = 0.002, SE = 3.02 μm), unaffected eye baseline pRNFL thickness (β = 0.4617 μm, p < 0.001, SE = 0.09 μm), and P100 SC latency (β = −0.26 μm, p = 0.016, SE = 0.10 μm). The pRNFL model did not find significant associations with N75-P100 SC amplitude (β = 0.51 μm, p = 0.23), age, sex, center, or timing variables—a notable weakness, as amplitude is a key measure of visual pathway function.

The comparison favored mGCIPL on several fronts:

-
**Better model fit:** The root mean square error (RMSE) for the mGCIPL model was **7.98**, compared with **11.26** for the pRNFL model—lower RMSE means more accurate predictions
-
**Stronger association with function:** mGCIPL thickness was significantly associated with *both* VEP latency and VEP amplitude, while pRNFL was associated only with latency, not amplitude
-
**Greater physiologic relevance:** mGCIPL directly measures the retinal ganglion cell bodies, which are the actual neurons that die during optic neuritis, whereas pRNFL axonal measurements can be confounded by optic nerve head swelling in the acute phase

Based on these findings, the researchers concluded that mGCIPL is a **more robust and physiologically meaningful outcome measure** for detecting neuroprotective treatment effects in acute optic neuritis trials.

## Sensitivity Analyses: Testing the Strength of the Results

The researchers performed several sensitivity analyses to confirm that their findings were not driven by statistical modeling choices or confounders:

**1. Corticosteroid use versus baseline visual acuity.** Corticosteroid use was strongly associated with worse baseline visual acuity (β = 0.7877 logMAR units, p < 0.001, SE = 0.15)—meaning patients with more severe vision loss were more likely to be treated with steroids. Although corticosteroid use was balanced between groups (p = 0.77), the researchers replaced baseline VA with corticosteroid use in the model to see if this changed the results. The original model (using baseline VA) performed better, supporting the decision to include vision severity rather than steroid use in the final model.

**2. Removing VEP variables.** When VEP variables were completely removed from the model, the treatment effect remained significant, changing only modestly from p = 0.0056 (β = 6.79 μm, SE = 2.35 μm) to p = 0.0306 (β = 6.35 μm, SE = 2.35 μm). This confirms that the treatment effect is not dependent on including electrophysiologic measurements in the model.

**3. Including absent VEP responses.** In the main analysis, patients with unrecordable VEP responses at 6 months were excluded. In a sensitivity analysis, these observations were included by assigning them standard values of 200 ms for latency and 0 μV for amplitude. The treatment effect remained significant (p = 0.0042, β = 7.30 μm, SE = 2.56 μm). However, the model with excluded absent VEPs exhibited a better fit (RMSE = 7.977 vs. 9.499), supporting the decision to exclude them from the primary analysis.

## Clinical Implications: What This Means for Patients

This study provides **Class II evidence** that phenytoin—when given within 2 weeks of symptom onset and continued for 3 months—is associated with greater preservation of mGCIPL thickness in patients with acute optic neuritis compared with placebo. This classification means the evidence comes from a well-designed randomized controlled trial but has some methodological limitations (such as being a secondary/retrospective analysis).

For patients facing acute optic neuritis, these findings are encouraging for several reasons:

-
**Neuroprotection is possible:** The study demonstrates that a relatively simple medication can reduce the structural damage caused by acute inflammation, preserving more retinal nerve cells than would otherwise survive.
-
**mGCIPL is a reliable biomarker:** Patients can take comfort that the research community now has a validated, sensitive tool for measuring nerve cell loss in the eye, which will accelerate the development of future neuroprotective treatments.
-
**Treatment timing and targeting:** The finding that phenytoin's benefits are greatest in patients with more severe baseline visual impairment (worse than approximately 20/120 Snellen) suggests that neuroprotective strategies might be most valuable for the patients at highest risk of permanent damage.

These findings also have broader implications for MS research. Because AON mimics the inflammatory demyelinating process that occurs throughout the central nervous system in MS, a treatment that protects the optic nerve may also protect other areas of the brain and spinal cord. If phenytoin—or similar sodium channel blockers—prove consistently neuroprotective, they could complement existing immunomodulatory therapies to better prevent long-term disability accumulation in MS.

The study also refines how future trials should measure outcomes. The superiority of mGCIPL over pRNFL means that future neuroprotection trials in optic neuritis should preferentially use mGCIPL as their primary outcome measure, reducing the sample size needed to detect treatment effects and speeding up the pace of clinical research.

## Study Limitations: What This Study Couldn't Prove

While this research is promising, several limitations must be acknowledged:

-
**Secondary analysis:** The mGCIPL analysis was performed after the original trial was completed (a post hoc secondary analysis), not as a prespecified primary outcome. This increases the risk of bias and is why the evidence is rated Class II rather than Class I.
-
**Exclusion of patients with unrecordable VEPs:** Ten patients with unrecordable VEP responses at 6 months were excluded from the primary analysis. If these patients represent the most severely affected cases, their exclusion may have affected the results (though the sensitivity analysis including them showed even slightly larger treatment effects).
-
**Modest sample size:** With only 67 participants in the final primary analysis, the statistical power to detect subgroup effects and interactions was limited. The treatment-by-baseline-VA interaction did not reach statistical significance (p = 0.230), so the threshold finding (logMAR ≥ 0.775) should be interpreted cautiously.
-
**Short follow-up:** The study followed patients for only 6 months. Long-term effects—including whether phenytoin's structural preservation translates into better permanent visual outcomes at 1, 2, or 5 years—remain unknown.
-
**Surrogate outcomes:** mGCIPL thickness is a structural biomarker, not a direct measure of patient experience. Although strongly associated with VEP function, the study did not demonstrate that the mGCIPL preservation translated into measurable differences in everyday visual function at 6 months (raw visual acuity data showed similar outcomes between groups).
-
**Voltage-gated sodium channel blockers in general:** This trial tested a specific drug (phenytoin). The results cannot necessarily be generalized to other sodium channel blockers, which may have different efficacy and safety profiles.
-
**Safety data not re-analyzed:** This study focused on structural outcomes; detailed adverse event data were reported in the original trial publication and are not re-presented here.

## Recommendations for Patients

For patients who have experienced, or are experiencing, optic neuritis, this study offers the following takeaway points:

1.
**Talk to your neurologist about clinical trials.** If you are diagnosed with acute optic neuritis, especially in the context of MS, ask whether any neuroprotection trials are enrolling at your center. Participation in research is one of the most powerful ways to advance treatments.
1.
**Know the window of opportunity.** This study treated patients within 14 days of symptom onset. If phenytoin or similar neuroprotectants become standard treatment, early diagnosis and prompt treatment will be critical. Seek medical attention immediately if you experience sudden vision loss or eye pain.
1.
**Understand that not all vision loss is equal.** Patients with more severe initial visual impairment appeared to benefit most from treatment. Even if vision loss is severe, that doesn't mean it's too late for neuroprotection.
1.
**Ask about OCT monitoring.** Optical coherence tomography is a painless, noninvasive imaging test that can track the health of your retinal layers over time. If your MS care team offers OCT, it provides valuable information about how your optic nerves are doing.
1.
**Keep perspective on the medication.** Phenytoin is an antiepileptic drug with potential side effects and drug interactions. It is not currently approved for treating optic neuritis, and patients should not attempt to self-medicate with it. Any use should occur within the context of a supervised clinical trial.
1.
**Stay informed about emerging therapies.** Phenytoin is one of several neuroprotective candidates being studied for optic neuritis and MS. Others include different sodium channel blockers, anti-LINGO-1 (a remyelination-promoting agent), and various other experimental therapies. The treatment landscape is evolving rapidly.

In summary, this study provides encouraging evidence that phenytoin can protect retinal nerve cells during acute optic neuritis, that the mGCIPL is a highly effective way to measure this protection, and that patients with the most severe vision loss may gain the most benefit. While phenytoin itself is unlikely to become a standard optic neuritis treatment without further trials, this research validates the whole concept of neuroprotection in AON and gives the field a better measuring tool for the next generation of therapies.

## Frequently Asked Questions

### What is acute optic neuritis and why does it matter in MS?

Acute optic neuritis is a sudden inflammation of the optic nerve that can cause eye pain, blurred vision, or temporary vision loss. Up to 70% of people with multiple sclerosis experience at least one episode. Because the optic nerve is part of the central nervous system, it offers a window into how MS relapses damage nerves.

### What did the phenytoin trial find about protecting retinal nerve cells?

In a phase II randomized trial of 80 patients with acute optic neuritis, those who took phenytoin for 3 months had significantly greater preservation of the macular ganglion cell-inner plexiform layer at 6 months than those on placebo. Average thickness was 73.8 μm versus 67.0 μm, a difference of 6.79 μm.

### Who was eligible to join the phenytoin optic neuritis trial?

Participants had to be 18 to 60 years old, have a clinical diagnosis of unilateral acute optic neuritis confirmed by a neuro-ophthalmologist, no previous optic neuritis in either eye, visual acuity of 6/9 or worse in the affected eye, and be enrolled within 14 days of symptom onset. The trial took place at two centers in the United Kingdom.

### What does the mGCIPL measurement mean for patients?

The macular ganglion cell-inner plexiform layer measures the cell bodies of retinal ganglion cells, the neurons that carry visual information to the brain. In this trial it was more reliable than the traditional peripapillary retinal nerve fiber layer for detecting neuroprotective treatment effects, and it was strongly linked to how well visual signals travelled to the brain.

### Did phenytoin help more if vision was worse at the start?

Yes. In this trial, treatment benefits were most pronounced in patients with worse vision at the start. A visual acuity threshold of about 20/120 or worse was identified, above which phenytoin showed significant benefits compared with placebo. The treatment-by-baseline-vision interaction itself was not statistically significant, so this finding should be interpreted cautiously.

### What are the limitations of this phenytoin optic neuritis research?

The mGCIPL analysis was performed after the original trial was completed, not as a prespecified primary outcome, so the evidence is rated Class II. The final primary analysis included only 67 participants, follow-up was just 6 months, and the study did not show that structural preservation translated into measurable differences in everyday visual function.

### Should patients ask their doctor about phenytoin for optic neuritis?

Phenytoin is an antiepileptic drug with potential side effects and drug interactions. It is not currently approved for treating optic neuritis, and patients should not attempt to self-medicate with it. Any use should occur within the context of a supervised clinical trial. Talk to your neurologist about whether any neuroprotection trials are enrolling.

### When should a patient with acute optic neuritis seek a second opinion about phenytoin or neuroprotection?

Phenytoin is not approved for optic neuritis, and treatment in this trial began within 14 days of symptom onset and continued for 3 months. A second opinion is reasonable when you are weighing whether to join a neuroprotection trial, whether phenytoin is appropriate given its side effects and drug interactions, or how your vision loss compares with the threshold (about 20/120 or worse) where benefits appeared. Because the evidence is Class II and follow-up lasted only 6 months, an independent review of your diagnosis, OCT imaging, and options can help. Diagnostic Detectives Network provides independent expert second opinions.

## Source Information

**Original article title:** Effects of Phenytoin on the Retinal Ganglion Cell-Inner Plexiform Layer in Acute Optic Neuritis: Analysis of a Phase II Randomized Trial.

**Authors:** Kodali S, Bianchi A, Raftopoulos R, Moccia M, Malladi AP, Yiannakas MC, Fugger M, Samson RS, Wheeler-Kingshott CAM, Koltzenburg M, Prados F, Hickman S, Kapoor R, Toosy AT.

**Journal:** Neurology®, 2025;105:e213951. doi:10.1212/WNL.0000000000213951

**Publication date:** October 21, 2025

**Trial registration:** ClinicalTrials.gov number NCT01451593 (submitted October 11, 2011; first patient enrolled February 2, 2012)

**Affiliations:** Department of Neuroinflammation, Queen Square Multiple Sclerosis Centre, UCL Queen Square Institute of Neurology, University College London; University of Palermo, Italy; Kings College Hospital NHS Foundation Trust, London; Federico II University of Naples, Italy; Universitat Oberta de Catalunya, Barcelona, Spain; and Sheffield Teaching Hospitals NHS Foundation Trust, Sheffield, UK.

**Funding source:** The Article Processing Charge was funded by UCL Library Services. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

*This patient-friendly article is based on peer-reviewed research published in Neurology® and is intended for educational purposes. It does not constitute medical advice. Patients should always consult their healthcare provider regarding treatment decisions.*


---
Publisher: Diagnostic Detectives Network (https://diagnosticdetectives.com) — independent multi-expert medical second opinions, worldwide, private-pay.
Author byline: Anton Titov, MD, PhD. Contact: https://diagnosticdetectives.com/pages/contact
Canonical page: https://diagnosticdetectives.com/products/can-a-common-seizure-medication-protect-the-optic-nerve-during-optic-neuritis-a-patient-friendly-look-at-the-phenytoin-trial