The rationale rests on a well-documented metabolic anomaly in ALS. Weight loss is common and is a prominent negative prognostic factor, and body mass index decline can begin up to ten years before disease onset. Studies using indirect calorimetry have shown that a significant share of patients with sporadic ALS exhibit an increased resting energy expenditure, a state known as hypermetabolism, which appears to persist through much of the disease course. In the trial cohort itself, 86.5 percent of participants had a resting energy expenditure higher than their calculated value, and 70.3 percent exceeded it by more than ten percent, the threshold commonly used to define hypermetabolism. Mitochondrial dysfunction, particularly impairment of complex I of the respiratory chain, is one of the two major hypotheses proposed to explain this excess energy expenditure, alongside hypothalamic atrophy or dysfunction.

Because fatty acids cannot be metabolised by motor neurons, as they are unable to cross the blood-brain barrier in the required form, the researchers reasoned that ketone bodies might constitute an optimal alternative energy substrate. Ketones cross into the central nervous system via the MCT1 transporter and yield a significantly higher amount of ATP per mole than glucose. Preclinical work lent support to the idea: in the SOD1-G93A mouse model of ALS, animals fed a ketogenic diet gained body weight, showed a significantly slower decline of motor function and maintained more motor neurons in the ventral horn of the spinal cord than controls. Isolated mitochondria from those mice produced more ATP when D-β-3-hydroxybutyrate was added, and in cultured neurons with pharmacologically blocked complex I, D-β-hydroxybutyrate restored function of the complex and improved ATP production.

A ketogenic diet itself is difficult to establish in ALS patients, however, because it demands long-term changes in eating habits that are hindered by dysphagia, rapidly declining general condition and reduced life expectancy, and restrictive food choice may even aggravate weight loss. Direct application of ketone bodies offers a practicable alternative. Before this trial, the literature search conducted by the team up to October 2025 identified no randomised, placebo-controlled trials of ketogenic interventions in ALS, only reviews, one case report, two cohort studies, two animal studies and one open-label trial that was terminated after a single patient was enrolled.

The trial enrolled adults with possible, probable or definite ALS according to revised El Escorial criteria who had lost at least 0.33 points per month on the revised ALS Functional Rating Scale since symptom onset, a threshold designed to exclude very slow progressors whose neurofilament levels would be unlikely to change measurably within six months. Participants were randomly assigned 1:1 to receive 20 millilitres, containing 10 grams, of the β-hydroxybutyrate ester three times daily, or a matching placebo, both as add-on to standard treatment including riluzole and high-caloric nutrition. Randomisation was stratified by disease progression speed and by the presence or absence of recent weight loss, and both patients and site personnel remained masked to allocation, with ketone measurements withheld from the study team until the end of the trial.

The primary endpoint was the change in serum neurofilament light chain (NfL) over six months, a biomarker of axonal damage that correlates strongly with disease progression and survival and responds earlier than clinical scales. In the unadjusted analysis, patients on the ketone ester showed a median NfL decline of 0.5 picograms per millilitre per month while levels remained stable in the placebo group at 0.0, a median difference of 1.3 picograms per millilitre per month that fell short of statistical significance at p = 0.094. After adjusting for baseline NfL, age, body mass index and slow vital capacity, however, the group effect became statistically significant, with a mean difference of 4.2 picograms per millilitre per month favouring the ketone ester (p = 0.014). A multiple-imputation sensitivity analysis preserved the adjusted difference, and the per-protocol set of patients who took at least 65 percent of their assigned dosage showed a similar trend, but a one-step linear mixed-effects model applied as a further sensitivity analysis found no significant effect.

The baseline imbalance complicates interpretation. Despite randomisation, patients in the ketone ester group had worse slow vital capacity and higher baseline NfL levels, indicating a more advanced stage and faster progression, since baseline NfL correlated positively with the rate of functional loss during the intervention. Conversely, placebo patients were older, a known negative prognostic factor, and a higher share of ketone-ester patients were already receiving nutritional supplements. The authors emphasise that the adjusted significance emerged only after correction for these imbalances and was not confirmed in the mixed-effects model, so the biomarker finding must be regarded as preliminary.

Prespecified subgroup analyses suggested the effect was more pronounced in patients with higher baseline NfL and in those with better functional status at entry, mirroring patterns seen in the precursor LIPCAL-ALS trial, in which an additional intake of 405 kilocalories and 45 grams of fat per day reduced NfL and lengthened survival in fast-progressing patients over 18 months. In the high-NfL subgroup, median levels rose by 1.7 picograms per millilitre per month on placebo but fell by 1.3 on the ketone ester (adjusted p = 0.045); among patients with higher baseline ALSFRS-R scores the corresponding figures were plus 1.1 versus minus 1.3 (adjusted p = 0.028). These subgroup findings are explicitly hypothesis-generating, as the trial was not powered for them.

Exploratory post-hoc analyses of urinary acetone, which reflects the metabolic turnover of β-hydroxybutyrate, showed that levels in the treatment group rose during the first half of the study and then fell by the end, a pattern the authors liken to keto-adaptation, the increased renal reabsorption and more efficient ketone utilisation seen after weeks of ketogenic diet or starvation. Patients whose acetone increased more after three months had better NfL slopes than placebo (p = 0.014), while those with smaller increases resembled placebo, though the subgroups were small and the analysis post-hoc.

On clinical outcomes, no significant differences emerged for ALSFRS-R, body mass index, slow vital capacity, survival or body composition. Notably, the ketone ester group showed a faster unadjusted decline in ALSFRS-R, 1.4 versus 0.8 points per month (p = 0.020), which lost significance after adjustment and which the authors attribute plausibly to the higher baseline NfL and faster natural progression in that group, though the data cannot prove this. Quality of life on the EQ-5D-5L declined more slowly in the treatment group in the adjusted analysis (p = 0.040). Tolerability was favourable: adverse events were mainly mild gastrointestinal symptoms or a burning sensation in the throat, no intervention-related serious adverse events occurred, and nausea and vomiting were actually more frequent in the placebo group, possibly due to its high glucose content. Dropouts, at 22.2 percent overall, were driven mainly by the bad taste of the ester, and the authors conclude that a larger, longer follow-up trial with clinical endpoints is needed to determine whether the NfL signal translates into meaningful clinical benefit.

Subject of Research: A randomised placebo-controlled trial of oral D-β-hydroxybutyrate ester in amyotrophic lateral sclerosis

Article Title: Efficacy and tolerability of D-β-hydroxybutyrate ester in patients with amyotrophic lateral sclerosis (KETO-ALS)–a randomised, placebo-controlled, double-blind study

Article References: Herrmann, C., Jagodzinski, L., Elmas, Z., Herschmann, T., Fröhlich, E., Bachhuber, F., Ruschitzka, T., Willkomm, F., Parlak, Ö., Weiland, U., Becker, F., Richter, L., Wiesenfarth, M., Uzelac, Z., Michels, S., Witzel, S., Tumani, H., Weishaupt, J. H., Maier, L., … Dorst, J. (2026). Efficacy and tolerability of D-β-hydroxybutyrate ester in patients with amyotrophic lateral sclerosis (KETO-ALS)–a randomised, placebo-controlled, double-blind study. eClinicalMedicine, 101, Article 104272. https://doi.org/10.1016/j.eclinm.2026.104272

Image Credits: AI Generated

DOI: 10.1016/j.eclinm.2026.104272

Keywords: ALS, ketone ester, beta-hydroxybutyrate, neurofilament light chain, hypermetabolism, randomised controlled trial, KETO-ALS, biomarker, keto-adaptation, clinical trial, eClinicalMedicine, Ulm University