The study by Vanmierlo et al. (1) published in this issue of Acta Psychiatrica Scandinavica suggests that brassicasterol might be a relevant additional biomarker in Alzheimer’s disease (AD). Brassicasterol, sitosterol, campesterol and stigmasterol are all plant, exogenous, sterols that enter the body only via food intake. They are similar to cholesterol, but contrary to this one are able to cross the blood-brain barrier (BBB) and accumulate within the brain. Since the BBB and its choroid plexus have an impaired function in the early stages of AD, Vanmierlo et al. hypothesized that plant sterol concentrations in CSF are altered in AD and therefore might be relevant biomarkers for AD. The authors of this paper found that compared with controls, concentrations of plant sterols are in fact reduced in the CSF of AD patients, after controlling for relevant covariates. In both sitosterol and brassicasterol the differences were statistically significant, but remained significant only for brassicasterol when controlling for cholesterol. For a reasonable sensitivity of 85% in predicting AD, the specificity of brassicasterol was 70%, which is a remarkable performance. Furthermore, this plant steroid improved the predictive value when added to the best accepted biomarkers until now, pTau and Aß42. As the authors of this paper conclude, brassicasterol might indeed be a relevant, additional biomarker in AD. This interesting article emerges in a context of general concern about dementia, the most widespread neurodegenerative disease, and specifically about AD, the most frequent subtype (2). The clinical symptoms of AD are commonly diagnosed in older people, but there is evidence that the degenerative process starts years before the clinical onset. It has been suggested that a delay of few years before the clinical onset of AD might decrease dramatically the prevalence of the disease. Therefore, the presymptomatic detection of AD might be crucial to facilitate an early treatment before the destructive neurodegeneration occurs. The search of biomarkers to adequately identify AD in the early stages has become a focus of research interest (3). It would facilitate an early diagnosis and treatment, as well as the monitoring of disease progression and treatment response. Ideally, biomarkers should facilitate the detection of the fundamental, neuropathological features in a feasible process, and should be reliable, reproducible, non-invasive, and cheap. Furthermore, biomarkers should have adequate sensitivity, specificity and predictive value, and ideally should be validated against the gold standard of confirmed neuropathology, if possible in different, independent and sound studies (4). CSF biomarkers have special interest, since they have been reported to show the best validity coefficients to date. Aß42, total tau and phosphorilated tau have been the most promising biomarkers in this respect. While none of them has been shown to generate the diagnosis, a combination of biomarkers has the potential to improve the diagnostic specificity (5). Vanmierlo′s proposal is attractive, but new studies seem to be indispensable to confirm the clinical value of the test (1). The consensus in the field is that the findings reported in this type of research should be replicated in new, independent studies. It would also be important to test the validity of the procedure in AD cases with confirmed pathological diagnosis. Discrepant findings might be expected before laboratory procedures are adequately standardized in different centres. The clinical applicability of the new biomarker might also be hampered by the lack of equipment and experience in some laboratories, and by the invasive nature of the lumbar puncture according to criteria common in some clinics, so that ethical questions might arise before standard implementation (5). Nevertheless, an important support for the new biomarker could come from data suggesting that it shows a relationship with disease stage and/or progression, and also from new, stringent studies testing to what extent brassicasterol is able to discriminate different neurological diseases, in particular neurological diseases impairing the BBB.