Background
The sense of smell provides us with crucial information about the environment and contributes to important aspects of life. Lately, interest in olfactory dysfunction has increased due to findings indicating that olfactory dysfunction is associated with increased risk of mortality [
1], degenerative disease such as Alzheimer’s disease and Parkinson’s disease [
2,
3], as well as due to the general appreciation of olfaction being important for quality of life. Olfactory dysfunction can result in altered food enjoyment, poor appetite, difficulties preparing food, inability to detect hazardous conditions (e.g., gas, smoke and spoiled food), lack of understanding of others and social isolation, depression and mood changes [
4,
5].
Although a rather large number of medical conditions are well documented regarding their impact on olfaction [
6], only a few studies have investigated olfactory function in stroke patients. Population-based studies that have estimated the risk of olfactory dysfunction in the aging segment of the population provide ambiguous results regarding stroke. Whereas both Murphy et al. [
7] and Karpa et al. [
8] showed that stroke was a risk factor for olfactory dysfunction, Schubert et al. [
9] and Landis et al. [
10] could not support these findings. The remaining available literature on stroke and olfaction includes case studies. For example, these studies show that stroke patients do not have a complete loss of olfactory function, but rather experience unpleasant sensation of odors [
11‐
16]. Cecchini et al. [
17] reported slightly reduced olfactory function in two out of the three patients they investigated.
One aim of this study was to investigate the ability to identify odors in chronic stroke patients. This was studied by comparing patients’ performance on a common clinical measure of olfactory function [
18‐
20] with an age- and sex-matched sample of referents from a population-based study. The stroke patients and referents were compared with respect to performance score on odor identification as well as distribution of olfactory diagnosis. This included normosmia (normal olfaction), hyposmia (reduced olfactory function) and functional anosmia (significantly impaired olfaction, including both total loss and minimal residual perception) [
21]. A second aim was to investigate predictors of odor identification performance. This comprised functional effects of stroke at the time of admission and during follow-up. A third aim was to compare patients being classified as normosmic versus hyposmic/anosmic with respect to self-reported olfactory functioning and odor pleasantness ratings.
Discussion
In summary, the findings demonstrate olfactory dysfunction in a significantly larger proportion of chronic stroke patients compared to a group of matched controls. Moreover, we found that age and NIHSS score (on admission) were significant predictors for olfactory dysfunction. Self-reported olfactory function did not differ between stroke patients with and without olfactory dysfunction. Yet, average pleasantness ratings differed significantly, indicating that patients with reduced olfactory function, on average, perceived olfactory items as less pleasant compared to patients with normal olfactory status.
The results indicating that olfactory dysfunction occurs as frequent as in 43 % in chronic stroke patients even one year after initial admission are notable and new. It has to be acknowledged that four patients in our study completely lacked ability to perceive odor (due to the initial screening procedure). This number is in accordance with findings from population-based studies, reporting approximately 5 % to be unable to smell [
21,
26]. Additionally, another 10.3 % of the patients showed functional anosmia, based on results from the odor identification test. It should be emphasized that the high correlation between odor detection sensitivity and identification ability validates the use of identification tasks to assess hyposmia and anosmia [
32]. Individuals with functional anosmia do indeed also experience impact on quality of life [
5]. Thus, the frequent occurrence of olfactory dysfunction in patients stresses that clinicians should be aware of this issue in clinical routines, though maybe not in the acute phase as focus lies on other medical issues.
Our findings, indicating age to be a significant predictor for olfactory performance, are in accordance with the well-established literature [
18,
28,
33]. In addition, NIHSS score remained a significant predictor even when controlling for age. The current findings extend the existing studies addressing the long-term consequences of stroke. The, NIHSS has been shown to be a predictor for stroke outcome and there is evidence showing that it may be associated with the development of vascular cognitive impairment (VCI) and eventually a diagnosis of dementia [
34,
35]. Olfactory dysfunction is a well-established marker for dementia [
3,
36] and is even known to occur in individuals at risk for dementia who demonstrate mild cognitive impairment [
37,
38]. However, findings on VCI and olfactory dysfunction are still scarce. It has been shown that patients with vascular dementia perform below normative performance in olfactory tests [
39]. Nonetheless, when comparing patients with vascular dementia and patients with AD, findings are mixed in that Gray et al. [
39] report a similar degree of olfactory impairment in both patient groups, whereas Duff et al. [
40] have shown lower performance in AD patients. In our sample, none of the patients had received a diagnosis of dementia at the time of assessment, although we cannot rule out that some patients may have been on a path towards decline. Moreover, the results indicate only a tendency of an association between global cognitive function (MMSE) and odor identification performance. Odor identification draws on memory functions [
41], but processing speed, language proficiency and reasoning have also shown to be involved [
42‐
44]. Cognitive changes in VCI are variable depending on the lesion location [
35], plausible explanations for finding a trend only may be that the sample was too heterogeneous to detect possible associations. Additional, screening instruments for dementia such as the MMSE, which was primarily developed for assessing cognitive deficits associated with AD, often do not detect those subtle deficits as seen in VCI [
35]. Yet, since up to 50 % of individuals with VCI develop dementia in a five-year period [
45], the current findings underline the urgent need for more knowledge.
The finding indicating no differences for self-reported olfactory functioning between normosmic patients and those with an olfactory dysfunction are not surprising since it has repeatedly been shown that self-reports are unreliable predictors of olfactory status [
28,
46,
47]. It has been discussed that due to the subtle progression of decline a certain threshold must be reached before loss becomes noticeable [
37]. The present finding of unawareness of olfactory dysfunction in stroke patients underlines the need for clinicians to objectively assess olfactory function in this patient population.
It is notable that those patients with an olfactory dysfunction perceive the odors as less pleasant than normosmic patients. This is in accordance with case studies which repeatedly have reported qualitative smell and taste disturbances [
11‐
14,
16]. These reports vary regarding the duration of the disturbances, varying between weeks and years. Our results showing that dysfunction may occur even a year after injury underline a prolonged course. Additional, existing findings have commonly stressed that the patients had not completely lost their ability to smell, but that change in perceptual quality of the odorous stimuli was a serious problem as it resulted in changes in diet and eventually weight loss [
11,
16]. Our results indicate a clear trend between symptoms of depression and odor identification performance. Signs of depression are a known consequence in patients with smell loss [
48]. Croy et al. [
5] suggested two potential links between depression and olfactory disorders: One arising through reduced quality of life in functions depending on olfaction (eating, social communication, environmental hazards etc.) and thus indirectly leading to depression, and another directly related to brain function in that olfactory input from the olfactory bulb via amygdala and limbic system is changed. Possible consequences for our patients have to remain speculative. Yet, two recent studies have stressed a possible link between nutritional status and stroke. Paquereau et al. [
49] found malnutrition to be associated with changed food intake and food preferences in 50 % of the stroke patients although their study did not include olfactory assessment. Aliani et al. [
50] concluded that olfactory dysfunction in the elderly and patients with dementia can lead to dietary restrictions and thus negative implications on nutrition and overall health.
The small number of patients in this study limits the generalizability of our findings. Additionally, small cell frequencies prevented us from analyzing relationships between olfactory dysfunction and stroke classifications (i.t. OCPS/TOAST) or other medical factors occurring in patients. We included all available patients, and based on MMSE score and NIHSS (baseline) it is assumed that the patients included had rather mild to moderate impact of the stroke. Thus, olfactory dysfunction may rather be under-estimated in this sample since patients not returning after one year may have even poorer functioning than those included here. Lastly, we did not register drugs patients were taking at the time of follow-up assessment, although drug-induced olfactory (and even more taste) dysfunction is well-known [
21]. Yet, although drug-induced olfactory dysfunction may be reversible, this may not always be possible in these patients.
Competing interests
The authors declare that they have no competing interests.
Authors’ contributions
EW, HN, HH, SN study design; EW, HN, HH, AB, MB: data collection; EW, DW, SN: analysis and interpretation; EW, HN, DW, HH, AB, MB, SN manuscript draft. All authors read and approved the final manuscript.