In brief
What kind of pain this is
Familiar pain has a source: a wound, inflammation, a strain. The nerve reports it faithfully, and a painkiller works on the source.
Hence its recognisable oddities:
▸burning, as though the skin had been scalded; ▸electric shocks, sudden and brief; ▸numbness mixed with pain — an area that simultaneously feels nothing and hurts; ▸pain from light touch — when a sheet or a sleeve becomes unbearable.
The two commonest causes are shingles, which leaves postherpetic neuralgia behind it, and diabetes, in which the long nerves of the feet and lower legs are damaged over years.
Ordinary painkillers work poorly here: there is no inflammation for them to act on. What helps are drugs that calm an over-excited nerve — and that is where this class comes in.
The class's idea: not to mask the pain but to quieten the excited wire
In a nerve ending there are calcium channels: when calcium enters through them, transmitter substances are released and the signal travels on. In a damaged nerve this machinery is disordered — there are more channels, and the signal fires on any pretext or none.
The channel has an accessory part, the α2δ subunit. Pregabalin and gabapentin bind precisely to it and reduce the excessive release of transmitters. This is not pain relief in the usual sense but turning down over-sensitive wiring.
It works, at a known price: drowsiness, dizziness, oedema, weight gain. For a drug taken for years, that price is not small.
The drug's idea: hit the right half of the target
Here the subtlety begins. The α2δ subunit exists in two forms:
▸α2δ-1 — linked to pain relief; ▸α2δ-2 — linked to dizziness and drowsiness.
Pregabalin binds both. Hence the logical move: build a molecule that prefers the first.
Mirogabalin is designed that way, and it has a second feature more important than the preference itself: it dissociates from α2δ-1 slowly and from α2δ-2 quickly. It lingers where it is wanted and does not linger where it is a nuisance.
The comparison that fits: if a remote has two buttons under one finger — 'turn down the pain' and 'turn down consciousness' — you can make the finger thinner, press mostly the first, and rest on it longer.
An elegant idea. Now for what came of it in practice.
What the trials showed
| Postherpetic neuralgia, 14 weeks | Difference from placebo |
|---|---|
| Mirogabalin 15 mg/day | *−0.41* points |
| Mirogabalin 20 mg/day | *−0.47* points |
| Mirogabalin 30 mg/day | *−0.77* points |
| Significance | In all three groups |
| Pooled analysis, 14 weeks | Difference from placebo |
|---|---|
| 15 mg once daily | *No* significant advantage |
| 10 mg twice daily (20 mg/day) | *−0.31* (95% CI −0.55 to −0.08) |
| 15 mg twice daily (30 mg/day) | *−0.63* (95% CI −0.86 to −0.40) |
The price that selectivity did not abolish
And now the crux — the very thing the drug was built for. Molecular selectivity promised less drowsiness and dizziness. Here is what actually happened:
| Adverse event | Placebo | 15 mg/day | 20 mg/day | 30 mg/day |
|---|---|---|---|---|
| Drowsiness | 3.8% | 10.8% | 14.5% | *19.1%* |
| Dizziness | 2.7% | 5.7% | 9.1% | *13.1%* |
This table has to be read alongside the previous one. The greatest relief and the greatest drowsiness sit at the same dose. One patient in five on 30 mg a day was drowsy, and one in eight dizzy. Precision for α2δ-1 explains the molecular mechanism, but it did not remove the class's side effects — at best it shifted the ratio.
Discontinuations for these events were uncommon: 0.6–1.9% depending on dose in the 14-week trials. But over a year in diabetic neuropathic pain, 13.1% stopped because of adverse events — one patient in eight. Of 214 enrolled, 172 (80.4%) reached the end, with pain scores declining across all 52 weeks.
How to read 0.77 points properly
The figure looks discouragingly small, and this is the place where it is easy to err in both directions.
This is not the total relief the patient feels. What is measured is the difference between groups, and in pain trials improvement on placebo is substantial: people report less pain on a dummy too. A patient on the drug feels both their own improvement and the placebo component — and only what remains on top reaches the table.
What cannot be claimed yet
▸Superiority over pregabalin is not demonstrated. No head-to-head comparisons of adequate quality and size exist. Comparing results across separate trials is not valid — patients, scales and placebo-response levels all differ. ▸Only Asian populations were studied. All pivotal trials ran in Japan, South Korea, Taiwan, Singapore, Malaysia and Thailand. Differences in drug metabolism between populations are well known, and results cannot simply be carried over. ▸There is no FDA or EMA approval. That is no verdict on the molecule, but it does mean the drug has not passed the independent assessment of those regulators. ▸The observation horizon is one year. For pain treated over decades, that is not much. ▸Effects on function and quality of life — sleep, capacity for work, mood — were not primary endpoints; how far 0.6 points of relief changes a life does not follow from these data.
What to understand before starting
▸Expectations must be realistic. The aim of this class is to make pain bearable, not to remove it. Expecting 'it will stop hurting' leads to disappointment and to endless dose escalation. ▸The dose is raised gradually — precisely because drowsiness and dizziness are dose-dependent and usually ease as tolerance develops. ▸No driving in the first weeks, until tolerability is clear. ▸Absence of response is a reason to stop, not to endure side effects for nothing. ▸Abrupt withdrawal is undesirable across the class — tapering is discussed with the physician.
Bottom line
▸The same class as pregabalin, but more precise for α2δ-1 and slower to leave it. ▸Pain relief is demonstrated and rises with dose: 0.41–0.77 points out of ten against placebo. ▸The lowest dose failed in the pooled analysis — only 20 and 30 mg daily reached significance. ▸Selectivity did not abolish the price: drowsiness 19.1% and dizziness 13.1% at the highest dose, against 3.8% and 2.7% on placebo. ▸Over a year of treatment one patient in eight stopped because of adverse events. ▸Available only in Japan and parts of Asia; nobody has shown superiority over pregabalin.
References
1. Kato J, et al. Mirogabalin for the management of postherpetic neuralgia: a randomized, double-blind, placebo-controlled phase 3 study in Asian patients. Pain. 2019;160(5):1175–1185. PMID 30913164
2. Kato J, et al. Safety and Efficacy of Mirogabalin for Peripheral Neuropathic Pain: Pooled Analysis of Two Pivotal Phase III Studies. Clin Ther. 2021;43(5):822–835.e16. PMID 34059327
3. Baba M, et al. Long-term safety and efficacy of mirogabalin in Asian patients with diabetic peripheral neuropathic pain. J Diabetes Investig. 2020;11(3):693–698. PMID 31722446
Key facts
- Mirogabalin (brand name Tarlige, Daiichi Sankyo) is a ligand of the α2δ subunit of voltage-dependent calcium channels — the same class as pregabalin and gabapentin.
- The drug is approved in Japan and several Asian countries; there is no FDA or European Medicines Agency approval, so it is not used in the United States or Europe.
- The α2δ subunit exists in two forms: α2δ-1 is linked to pain relief and α2δ-2 to dizziness and drowsiness; mirogabalin targets the first more precisely and stays bound to it longer.
- In the postherpetic neuralgia trial in 763 patients, the difference from the average pain reduction on placebo was 0.41, 0.47 and 0.77 points on a ten-point scale for daily doses of 15, 20 and 30 mg — all significant.
- In the pooled analysis of the two pivotal studies (1587 patients with diabetic neuropathic pain and postherpetic neuralgia), only the two higher doses significantly beat placebo: −0.31 and −0.63 points.
- The 15 mg daily dose showed no significant advantage over placebo in the pooled analysis.
- Selectivity did not remove the class's side effects: at the highest dose drowsiness occurred in 19.1% versus 3.8% on placebo, and dizziness in 13.1% versus 2.7%.
- In a one-year study in diabetic neuropathic pain, 172 of 214 enrolled patients (80.4%) completed it, while 13.1% discontinued because of adverse events.
- All pivotal trials were conducted in Asia — Japan, South Korea, Taiwan, Singapore, Malaysia and Thailand; no data exist for other populations.
- No adequate head-to-head comparison with pregabalin has been carried out, so there is no basis for claiming superiority over it.
