In the paper by McShane et al (this issue, pp739–743), the olfactory function of 92 patients with dementia and 94 control subjects, accessed through the Oxford Project to Investigate Memory and Ageing (OPTIMA), was assessed and related to neuropathological findings at necropsy.1 Patients with Lewy body dementia were more likely to be anosmic than those patients with Alzheimer's disease, whose olfactory function was comparable with that in control subjects. The extent of the anosmia in Lewy body dementia was greater in those patients with higher counts of Lewy bodies, as detected by antiubiquitin immunohistochemistry, and was not influenced by the presence or absence, or degree, of Alzheimer-type pathological changes. Such findings are consistent with previous studies in Parkinson's disease, in which olfactory deficits are well documented.2 Lewy bodies may therefore be pathological markers of anosmia, whether these occur in the brain stem or cerebral cortex.

The data presented by McShane et al are important because they suggest a simple but sensitive, and perhaps more objective, way of discriminating in life between patients with Lewy body dementia and those with Alzheimer's disease. In so doing, olfactory testing may provide a valuable adjunct to present clinical criteria for Lewy body dementia,3 which depend heavily on the presence of visual hallucinations and fluctuating cognition, symptoms difficult to define and quantify objectively and which can be masked in the presence of severe Alzheimer's disease. Unfortunately, in the study there was only one patient with Lewy body dementia who did not have some degree of Alzheimer-type pathological changes in the brain, so the authors were unable to validate their findings in a group of patients with Lewy body dementia entirely free from such (potentially) compounding influences.

Perhaps, the main unanswered question raised by this study is why patients with Alzheimer's disease do not become anosmic, given that the same topographic areas of brain are affected by plaques and tangles in Alzheimer's disease as those affected by Lewy bodies (and variable Alzheimer pathology) in Lewy body dementia.4 Additive effects in Lewy body dementia are possible. In this disease, it is pyramidal neurons of deeper layers 5 and 6 of the cerebral cortex, particularly those of the cingulate gyrus, which receive a dense dopaminergic innervation and project to the corpus striatum and other subcortical regions, which are mostly affected by Lewy bodies. By contrast, Alzheimer-type changes tend to be more abundant in the upper cortical layers where corticocortical projections are formed, but generally in Lewy body dementia these pathological changes are much less abundant than in Alzheimer's disease itself.4Different, and additional, circuitry within the brain may therefore become disconnected by Lewy body pathology in Lewy body dementia, compared with that primarily targeted by Alzheimer-type pathology in Alzheimer's disease. It is possible that the visual hallucinations, fluctuating cognition, and olfactory impairment of Lewy body dementia are symptoms mediated by changes in function of these layer 5/6 pyramidal neurons, rather than as a result of events taking place at the level of the olfactory bulbs (which were not analysed in this study), a region well known to be severely damaged in Alzheimer's disease.5