Idebenone

Idebenone is a man-made quinone closely related to coenzyme Q10, the natural molecule that carries electrons within mitochondria. Compared with coenzyme Q10 it has a much shorter side chain, which makes it more water-compatible and, its developers argued, better able to reach cells. It was first created by the Takeda Pharmaceutical Company and studied in the 1980s and 1990s as a possible treatment for Alzheimer's disease and age-related cognitive decline, where results were mixed; over time its main role shifted to inherited mitochondrial disorders ^[3].

Idebenone's clearest use is in Leber's hereditary optic neuropathy (LHON), a rare mitochondrial genetic disease in which faulty complex I of the respiratory chain causes rapid, often severe loss of central vision as retinal ganglion cells die. Marketed for this purpose as Raxone, idebenone is approved by the European Medicines Agency and is described as the only disease-specific treatment shown in trials to help preserve and partly recover vision in LHON, with the best results when it is started soon after symptoms begin ^[1][2]. It has also been investigated in other conditions marked by mitochondrial dysfunction and oxidative stress, including Friedreich's ataxia and Duchenne muscular dystrophy, but these programs met with less consistent success; a Friedreich's ataxia application under the name Sovrima was refused authorisation in Europe in 2008 and never marketed there ^[3]. Beyond medicine, idebenone is added to some anti-aging cosmetics as a topical antioxidant, though evidence for those cosmetic claims is limited.

In the European Union idebenone is a prescription medicine for LHON; it has not been approved for neuromuscular indications in North America. It is taken by mouth, and its absorption is poor, so it is generally given with food and in formulations designed to improve uptake; most of an absorbed dose is cleared through the kidneys after conjugation. Idebenone is usually well tolerated, with generally mild side effects, and it is also sold in some markets as a dietary supplement and nootropic, where quality and evidence vary ^[1][2].

Idebenone is generally well tolerated in clinical use, but it is not without concerns. The most commonly reported side effects are gastrointestinal, including diarrhea, nausea, and abdominal discomfort, along with headache, cough, nasopharyngitis, and back pain seen in the LHON trials. Mild, usually reversible increases in liver enzymes have been observed, so caution and monitoring are reasonable in people with hepatic or renal impairment. It can also harmlessly discolor urine a reddish-brown, which is expected rather than dangerous. There is limited data in pregnancy and breastfeeding, and interactions with drugs metabolized by the liver are theoretically possible, so anyone on other medications should coordinate with a clinician. Outside its approved orphan indication, much idebenone is sold as a research chemical or dietary supplement of variable purity, and those over-the-counter products are not held to pharmaceutical quality standards; sourcing and dosing accuracy therefore matter.

The reddish-brown urine deserves a plainer statement, because it alarms people who have not been warned. The label attributes it to coloured metabolites, calls it harmless, states it is not associated with blood in the urine, and says it needs no change of dose and no interruption of treatment.

The rarer events on the label are worth naming rather than leaving under 'well tolerated'. Nasopharyngitis and cough are listed as very common, diarrhoea and back pain as common, and a further set is listed at unknown frequency: agranulocytosis, low white cell and platelet counts, seizures, hepatitis, rash and raised nitrogen in the blood. The label also warns specifically that adverse events in patients with liver impairment have led to treatment being interrupted or stopped. One documented case fits that pattern: a child on high-dose idebenone in a Friedreich's ataxia trial developed neutropenia at six months, which resolved on stopping ^[19].

The interaction picture is more specific than 'drugs metabolised by the liver'. Idebenone is a mild inhibitor of CYP3A4, tested directly against midazolam in a dedicated study, and may also inhibit P-glycoprotein, which puts narrow-margin substrates such as dabigatran and digoxin in scope.

One efficacy signal reads as a safety signal and should not be filed as anything softer. In LEROS, eyes with the m.3460G>A mutation treated in the subacute phase ended up worse than matched untreated eyes at 24 months, a relative worsening of 0.53 logMAR, about 26 chart letters, and treatment significantly raised the proportion of eyes that worsened ^[6]. A 2024 review states the consequence directly: caution is warranted in subacute or dynamic eyes with that mutation because of possible clinical worsening ^[11]. Not every LHON genotype is a candidate.

Idebenone is a synthetic short-chain analogue of coenzyme Q10, designed as a benzoquinone antioxidant able to interact with the mitochondrial electron transport chain. Takeda developed it in Japan in the 1980s under the code CV-2619 and studied it first for Alzheimer's disease and cerebrovascular decline, where it saw use in Japan and parts of Europe as a putative neuroprotective agent. Those cognitive programmes are where its nootropic reputation comes from, and they ended badly: a 450-patient European trial reported benefit over two years ^[27], but a 536-patient one-year American trial published in 2003 was titled, plainly, 'Idebenone treatment fails to slow cognitive decline in Alzheimer's disease' ^[26]. Interest then moved to mitochondrial disease, where idebenone was investigated in Friedreich's ataxia and Duchenne muscular dystrophy under names such as Catena and Sovrima.

The regulatory history that followed is more instructive than the pharmacology, and it is routinely reported wrong. Santhera took idebenone into mitochondrial disease and submitted it in Europe twice, for two different diseases, and lost the first two attempts. The Friedreich's ataxia application, as Sovrima, drew a negative opinion from the Committee for Medicinal Products for Human Use on 24 July 2008; Santhera asked for a re-examination and the committee confirmed the refusal on 20 November 2008. Sovrima was therefore never authorised in Europe, which means it was never withdrawn either. The second attempt was for LHON, and on 21 March 2013 Santhera withdrew the application after another negative opinion; the committee's stated reason was that 'taking Raxone for six months did not lead to a significant improvement in vision compared with placebo', with patients on the drug reading three more letters on the chart than patients on placebo, and it concluded that 'the benefits of Raxone did not outweigh its risks'. That refusal turned on exactly the distinction the trial data carry: RHODOS missed its primary endpoint in the intention-to-treat population and reported its positive result in a post hoc subgroup ^[4].

The third attempt worked. Raxone was authorised in the European Union on 8 September 2015 for visual impairment in adolescent and adult patients with LHON, granted under exceptional circumstances because the applicant could not provide comprehensive efficacy and safety data under normal conditions of use ^[12]. The follow-up study that showed the six-month treatment effect persisting for a median of 30 months off drug had been published in the interim ^[5], and the confirmatory Phase 4 study, LEROS, met its primary endpoint against an external natural history cohort and was published in 2024 ^[6]. Switzerland had already authorised it in 2014; the United Kingdom followed in 2015, and Israel, South Korea, Serbia, Chile, Bahrain and Taiwan later.

Two things happened in the last year that are easy to misread as the drug being pulled. In September 2025 Raxone was removed from the European Union register of orphan medicinal products, which sounds like a withdrawal and is not: the ten-year period of market exclusivity that the orphan designation of 15 February 2007 conferred had simply run out. The marketing authorisation itself remains in force, and ChEMBL still records the molecule as approved with no withdrawn flag. The other event was in the United States, where the drug has never been approved. The FDA granted priority review in September 2025 with a decision date of 28 February 2026, and on 5 March 2026 Chiesi announced that the agency had issued a Complete Response Letter, saying it could not approve the application in its current form and asking for additional clinical data from adequate and well-controlled studies along with chemistry, manufacturing and nonclinical clarifications. No new safety concerns were identified. So a compound with a real approval in Europe for eleven years still has no approval in the United States, and the reason both times has been the same: the pivotal randomised evidence is thinner than the clinical experience.

Idebenone occupies an unusual position between an approved orphan drug and a longevity-oriented supplement. In the mitochondrial and ophthalmology fields it is taken seriously as the first and, for a time, only disease-modifying option for LHON, though its clinical benefit there is considered real but partial. Its earlier cognitive and Friedreich's ataxia programs produced mixed and often disappointing results, tempering the broader neuroprotective enthusiasm that once surrounded it. In supplement circles it is marketed as a more bioavailable, mitochondria-targeted alternative to CoQ10 and as a topical antioxidant.

The first framing has the mechanism wrong in an interesting way: the real difference from CoQ10 is that idebenone is a substrate for the enzymes NQO1 and NQO2 and CoQ10 is not ^[14], not that it penetrates tissue better, which no study has compared. The topical claim traces to a single 2005 paper that scored it above vitamin E and CoQ10 on a composite of skin assays ^[32]; an independent group testing photoprotection the following year found it ineffective against a vitamin C, vitamin E and ferulic acid combination ^[33]. Overall it is viewed as generally well tolerated but of context-dependent efficacy, and the context that decides it is how much NQO1 the target tissue has.