These results indicate that, despite having grossly impaired olfaction, as is typical of many neurodegenerative conditions [14–19], some PD patients were able to discriminate the scent of lavender from another floral scent at levels that exceeded chance. Furthermore, at least in women with PD, that ability was related somewhat to discrimination of a floral scent on a standardized test of olfaction (UPSIT). These results do not directly address whether a scent such as lavender, when presented at night to promote sleep in patients with advanced neurologic disease [4–10], might induce or sustain sleep in some patients, but they are certainly compatible with such an effect, to the extent that this selected aspect of olfaction evidently remains intact in a subset of these patients. Given the potential danger of many medications used to induce sleep in dementia patients (including older and newer generation anti-psychotic medications), many of which carry so-called “black box” warnings from the United States Food and Drug Administration [26], alternative medicine treatments for nocturnal agitation and disturbed sleep in dementia (often subsumed under the rubric of “sundowning,” [27]), such as aromatherapy, become a potentially attractive intervention for caregivers, despite the adequacy of the empirical data base confirming their efficacy [3].
Our findings do not address directly the mechanisms as to how lavender might promote sleep in patients with neurodegenerative disease. It has been long-established that some of the earliest stages of neurodegeneration may be manifested by protein aggregates in the olfactory bulb and/or olfactory nucleus [28–31]. However, even patients with complete anosmia can detect a wide range of scents [32], indicating that maxillary trigeminal afferents can play a role in identification of certain odors. In PD patients, such function may compensate to some extent in otherwise compromised odor identification via the first cranial nerve [33]. Lavender is a complex molecule containing various components such as linalool, camphor, eucalyptol, terpinene-4-ol, and linalyl acetate [34, 35], and various formulations of lavender oils may contain 2- to 3-fold the concentrations of these various components relative to other formulations [36]. We did not ascertain the varying amounts of these substances present in our samples, though specific odor discrimination studies involving linalool, often considered a key constituent for lavender’s anti-microbial properties, have been shown to be detectable in non-PD anosmia patients and be rated as having modest levels of intensity [32]. More relevant for the current discussion are basic science receptor binding and infusion studies that have shown that linalool has anti-glutamatergic properties throughout the rat cerebral cortex and may dampen seizure susceptibility [37, 38], physiological effects which are compatible with the presumed sedative/hypnotic effects of lavender. Additionally, murine models using oral administration have demonstrated immediate effects on the sleep electroencephalogram (EEG) [39]. Further evidence for intactness of an alternative olfaction pathway for lavender and its composite volatile oils in PD patients derives from studies of EEG-derived event related potentials, in which olfactory stimuli (non-lavender) with presumed trigeminal influence showed similar responses in PD and controls [40, 41], and a study of single nostril odor identification in PD, which was interpreted to suggest greater likelihood of trigeminal mediation of scent identification [42].
Arguing against the interpretation that lavender identification relied solely on trigeminal involvement, however, was the suggestion that, at least in women, identification of the scent of rose on the UPSIT, a floral scent usually considered less likely to engage the trigeminal system [32], may have been more likely among those able to identify lavender in the forced-choice task. Like lavender, rose is a monoterpene, but more recent molecular characterization has shown that this scent consists of hundreds of volatile oils that demonstrate considerable molecular variability across different rose types [43]. When one such floral scent, geraniol, was placed head-to-head with linalool, younger anosmia patients without neurodegenerative disease showed far greater accuracy identifying the latter and also rated linalool as more intense, suggesting primacy of the trigeminal system in its detection [32].
It should be noted that not all studies have reported that lavender aromatherapy has beneficial effects on agitation in dementia. Using a within-subjects repeated ABCBA design, Snow et al [12] reported no such effects but left open the possibility that, consistent with post-mortem receptor binding studies suggesting that linalool may have direct anti-convulsant properties and indirect sedative properties [44], dermal contact with lavender oil might still have medicinal effects. For example, linalool and linalyl acetate have rapid skin absorption [34], which would be compatible with such effects. However, a cross-over trial using dermal application of lavender oil also failed to demonstrate immediate beneficial effects on agitation in a nursing home population [13].
There are some weaknesses associated with the approach that we have taken here. Our forced-choice discrimination approach, essentially constructed with available household materials (i.e., commercially available candle-shavings), demonstrates proof-of-concept for lavender discrimination but does not allow precise standardization of stimuli presentation, nor does it control for sniffing effort, which has been suggested to play a role in at least some of the olfactory performance deficits demonstrated by PD patients but which may otherwise be unelated to overall motor impairment [45]. We also did not find an association between PD patients’ ability to discriminate lavender and the severity of their motor impairment on the UPDRS. The ability of at least some of the PD patients to discriminate lavender is somewhat surprising, since lack of standardization might have even increased the probability of a negative result. Though our results could be interpreted as compatible with the potential beneficial effects of lavender on sleep in some studies of patients with advanced neurodegenerative diseases [4–10], the presentation of odors during sleep has complex effects, since it is not always apparent that such stimuli, whether pleasant or noxious, impact objective sleep measures. For example, when presented intermittently during EEG monitored sleep in tightly controlled temporally sequences, even stimuli thought to activate olfaction primarily through trigeminal afferents may not always result in altered sleep architecture [46, 47], though acute changes in inspiration and expiration during stimuli presentation have been noted [48]. On the other hand, when presented prior to sleep, certain scents, like lavender, have been reported to enhance sleep quality and increase polysomnographically defined measures such as slow wave sleep, at least in young adults with presumably intact olfactory function [49, 50]. Interestingly, stimuli with both higher (e.g., lavender) and lower (e.g., vanilla) likelihood of trigeminal mediation were shown to enhance delta activity when presented during sleep in healthy younger persons as well [51]. The latter results cannot necessarily be extrapolated to those with neurodegenerative diseases, but they do allow for the possibility that aromatherapy might indeed impact sleep in such populations.
In summary, our results are compatible, at least in principle, with certain types of aromatherapy (e.g., lavender) being candidates for improving sleep in neurodegenerative disease. Clinical trials of lavender have varied widely in quality, and most have been poorly controlled to date [3]. Clearly additional work would be required to determine both the efficacy and mechanisms underlying such an intervention.