Introduction
The risk of olfactory disorders increases with age and is higher in elderly people with pre-dementia and dementia.Reference Doty, Bayona, Leon-Ariza, Cuadros, Chung and Vazquez 1 , Reference Murphy, Gilmore, Seery, Salmon and Lasker 2 The most common causes of olfactory disorders are chronic sinonasal diseases, acute rhinitis and post-traumatic conditions, as well as toxic chemicals, cancer and degenerative diseases. Although degenerative diseases are not the main cause of olfactory disorders, such disorders are often found in patients in the early stages of both Alzheimer's disease (before the appearance of other cognitive and behavioural symptoms) and Parkinson's disease (prior to motor symptoms). However, Alzheimer's disease and Parkinson's disease patients rarely undergo specific assessment for olfactory disorders in daily clinical practice.
At the histopathological level, olfactory disorders in Alzheimer's disease are caused by the presence of amyloid plaques in the olfactory epithelium, the olfactory bulb, the anterior olfactory nucleus and limbic regions associated with olfactory functions such as the uncus and amygdala. At the biochemical level, cholinergic deficits could contribute to the olfactory disorders found in Alzheimer's disease patients because acetylcholine plays a major role in the olfactory learning process.
Currently, Alzheimer's disease is diagnosed based on cognitive and imaging tests, even though olfactory disorder may be an early clinical marker of dementia due to Alzheimer's disease and could thus improve early clinical diagnosis.
A problem faced by clinicians evaluating olfactory disorders in elderly people with cognitive impairment is that self-reported olfactory complaints may be inaccurate and can reflect a number of different smell and taste disturbances. To assess olfactory disorders objectively, clinicians should have access to sensitive, easy-to-use olfactory tests in their daily practice. Several tests have been developed for investigating different aspects of olfaction, ranging from odour sensitivity to odour identification. However, no ‘gold standard’ has been established and published studies have used assessment tools targeting different aspects of olfactory disorders and employing different odours (because odours are often culture and country specific), resulting in incomparable findings across studies and cultures. This review describes olfactory changes that occur during ageing and in patients with cognitive impairment, reports the tools currently available for assessing olfactory disorders, and offers new perspectives on how to improve current assessment methods for diagnosing olfactory disorders in elderly populations.
Method
To identify relevant articles published between January 2000 and October 2015, the following electronic databases were searched: PubMed (Medline), Cochrane Library, PsycINFO and Web of Science. The following keywords were used: ‘olfaction’ or ‘olfactory disorders’ or ‘smell’ combined with ‘aging’ or ‘elderly’ or ‘dementia’ or ‘Alzheimer's disease’.
The titles and abstracts of retrieved articles were independently screened by two authors (AG and RD), and rated to assess their relevance to the research question. For the studies presented below, the age of participants is reported as the mean ± standard deviation.
Biology of olfactory disorders in ageing
Olfactory disorders in ageing
Olfactory disorders are frequently observed in ageing populations,Reference Doty, Bayona, Leon-Ariza, Cuadros, Chung and Vazquez 1 with prevalence rates of around 5 per cent for people aged 45–65 years and of more than 10 per cent for people aged over 65 years.Reference Murphy, Gilmore, Seery, Salmon and Lasker 2 Olfactory disorders are usually first observed at the age of 60 years, with an earlier decline in men than in women,Reference Doty, Shaman, Applebaum, Giberson, Siksorski and Rosenberg 3 , Reference Choudhury, Moberg and Doty 4 and are estimated to affect more than 50 per cent of the population aged over 80 years old.Reference Lafreniere and Mann 5 The definition of olfactory disorder includes both hyposmia (partial loss of olfactory function) and anosmia (complete loss of olfactory function).
Several age-related factors contribute to olfactory disturbance, for example structural changes in the olfactory epithelium and olfactory sensory neurons (including the olfactory bulb that mediates the neural response to olfactory stimuliReference Enwere 6 ), pathways and processing regions.Reference Attems, Walker and Jellinger 7
Olfactory disorders can affect odour signal analysis at different levels of the nervous system. Indeed, olfactory disorders at the peripheral level can result from alterations to the detection threshold (i.e. the molecular concentration of odorant that an individual can detect) due to impairments at the peripheral nervous system level.Reference Frasnelli, Lundström, Schöpf, Negoias, Hummel and Lepore 8
At the central level, olfactory disorders can result from alterations in discrimination ability (i.e. the ability to distinguish a specific odour from other odours) and identification ability (i.e. the ability to associate an odorant molecule with related words or images), both of which result from impairment at the central nervous system level.
Olfactory disorders in Alzheimer's disease
Olfactory disorders in Alzheimer's disease have been a focus of research since 2000. The first investigations studied alterations in the detection threshold, while more recent studies have focused on alterations in identification abilities. In addition, several biological and genetic markers related to Alzheimer's disease risk and pathogenesis have been associated with olfactory changes in ageing. For instance, in cognitively healthy elderly people, olfactory disorders have been associated with increased levels of cerebral amyloid lesions and apolipoprotein E ε4 status,Reference Wilson, Arnold, Schneider, Tang and Bennett 9 – Reference Olofsson, Nordin, Wiens, Hedner, Nilsson and Larsson 11 two well-known biomarkers of Alzheimer's disease.
Different types of olfactory disorders can be found in Alzheimer's disease, including quantitative disorders, which affect odour detection thresholds; and qualitative disorders, which affect odour identification.
Quantitative disorders
The olfactory epithelium of Alzheimer's disease patients undergoes several changes.Reference Getchell, Shah, Buch, Davis and Getchell 12 Psychophysical studies indicated that Alzheimer's disease patients have higher detection thresholds, and thus lower olfactory sensitivity, compared with cognitively healthy participants,Reference Nordin and Murphy 13 , Reference Bacon, Bondi, Salmon and Murphy 14 and that the degree of impairment correlates with disease severity.Reference Nordin and Murphy 13 , Reference Murphy, Gilmore, Seery, Salmon and Lasker 15 A recent study (n = 94; age = 72.45 ± 9.4 years) showed asymmetry in olfactory thresholds depending on whether the odour was presented to the right or left nostril.Reference Stamps, Bartoshuk and Heilman 16 In this study, Alzheimer's disease patients detected odours from a closer distance when presented to the right nostril compared with the left nostril, suggesting that an important alteration had occurred in the left nostril. However, another recent study (n = 35, age = 71.05 ± 6.7 years) of Alzheimer's disease patients failed to replicate these results, finding no evidence of detection threshold asymmetry between the right and left nostrils.Reference Doty, Bayona, Leon-Ariza, Cuadros, Chung and Vazquez 1
Qualitative disorders
Qualitative alterations or distortions in smell perception (known as dysosmias) are less well studied in dementia-related diseases. Dysosmia includes parosmia, which refers to a distorted perception of an odorant (odorants are described as smelling differently, often foul-smelling, from how the patient remembers), and phantosmia (smell perception in the complete absence of a physical odour). Some studies suggest that in Parkinson's disease, qualitative abnormalities of olfaction should be more carefully examined in the prodromal phase (i.e. the early stages) of Parkinson's disease, and have proposed phantosmia as a new premotor manifestation of Parkinson's disease.Reference Landis and Burkhard 17 , Reference Hirsch 18 However, a more recent study with a larger cohort concluded that idiopathic phantosmia is more likely to be a symptom than a reliable predictor of early Parkinson's disease or other neurodegenerative diseases.Reference Landis, Reden and Haehner 19
Qualitative disorders are mainly discovered when an accurate history is taken of a patient's ability to discriminate and identify odours.
Olfactory discrimination
Olfactory discrimination is the ability to recognise a smell that has been presented, and requires the original smell to be stored in the patient's memory. Studies into the odour discrimination abilities of Alzheimer's disease patients are scarce and have received more criticism than studies on odour sensitivity and identification. The main reason is that more cognitive components are required for olfactory discrimination than for odour identification. Indeed, the mnemonic component of these tests is intertwined with the semantic component.Reference Moller, Wulff and Koster 20 – Reference Djordjevic, Jones-Gotman, De Sousa and Chertkow 22 However, despite these methodological constraints, altered discrimination ability is reported to be more predictive of cognitive decline compared with altered odour sensitivity or identification abilities.Reference Sohrabi, Bates, Weinborn, Johnston, Bahramian and Taddei 23 , Reference Naudin, Mondon, El-Hage, Desmidt, Jaafari and Belzung 24 A recent study investigated whether an olfactory discrimination test could discriminate between Alzheimer's disease and depression (participants with Alzheimer's disease, n = 20, age = 75.9 ± 9 years; participants with depression, n = 20, age = 73.4 ± 5.6 years).Reference Pentzek, Grass-Kapanke and Ihl 25 The results showed that individuals with depression had impaired olfactory discrimination ability for both familiar and unknown odours, while individuals with Alzheimer's disease made mistakes in recognising only unknown odours, suggesting that emotional olfactory memory is somewhat preserved in individuals diagnosed with Alzheimer's disease.
Olfactory identification
Olfactory identification impairments have been reported in healthy elderly participants,Reference Growdon, Schultz, Dagley, Amariglio, Hedden and Rentz 26 as well as in patients with mild cognitive impairment,Reference Petersen 27 including both amnestic mild cognitive impairment (a cognitive state more commonly associated with conversion to Alzheimer's disease) and non-amnestic mild cognitive impairment.Reference Vyhnalek, Magerova, Andel, Nikolai, Kadlecova and Laczo 28 Growdon et al. reported that diminished olfactory identification was associated with markers of neurodegeneration such as entorhinal cortex thickness and increased cortical amyloid burden.Reference Growdon, Schultz, Dagley, Amariglio, Hedden and Rentz 26 Impaired odour identification is also a better predictor of cognitive decline than some memory disorders (e.g. deficits in episodic memory) among cognitively healthy participants.Reference Devanand, Lee, Manly, Andrews, Schupf and Doty 29 Therefore, many studies have focused on the ability of patients to identify odours to aid the early diagnosis of Alzheimer's disease, and have suggested that olfactory identification may be more relevant than olfactory sensitivity for predicting conversion from mild cognitive impairment to Alzheimer's disease. In 2000, Devanand et al. assessed the predictive utility of an odour identification test for determining conversion from mild cognitive impairment to Alzheimer's disease.Reference Devanand, Michaels-Marston, Liu, Pelton, Padilla and Marder 30 This longitudinal study monitored 90 patients with mild cognitive impairment for over 3 years and found that patients with lower olfactory identification scores were more likely to develop Alzheimer's disease. These promising results suggest that inclusion of an olfactory identification test in the routine assessment might help predict conversion from mild cognitive impairment to Alzheimer's disease.Reference Makowska, Kloszewska, Grabowska, Szatkowska and Rymarczyk 31 The University of Pennsylvania smell identification test is currently considered one of the top five predictors for assessing the conversion risk to Alzheimer's disease.Reference Devanand, Liu, Tabert, Pradhaban, Cuasay and Bell 32 A recent study (n = 148, age = 67.9 ± 8.7 years) combined the University of Pennsylvania smell identification test with four other predictors: informant report of functioning, the selective reminding test – immediate recall (verbal memory), magnetic resonance imaging (MRI) hippocampal volume and MRI entorhinal cortex volume.Reference Devanand, Liu, Tabert, Pradhaban, Cuasay and Bell 32 This combination of tests was strongly predictive of conversion to Alzheimer's disease and markedly superior to combining age and mini-mental state examination. As taste is heavily dependent on olfactory abilities, the study of taste disturbances may offer similar opportunities.Reference Steinbach, Hundt, Vaitl, Heinrich, Förster and Bürger 33 Impairments in olfactory identification could also be used to measure cognitive decline in patients with amnestic mild cognitive impairment.Reference Kjelvik, Saltvedt, White, Stenumgård, Sletvold and Engedal 34 Olfactory identification has been extensively documented in patients with amnestic mild cognitive impairment, but far less so in individuals with non-amnestic mild cognitive impairment. A recent study highlighted that olfactory identification was also impaired in individuals with non-amnestic mild cognitive impairment, although the degree of olfactory impairment did not correlate with cognitive performance.Reference Vyhnalek, Magerova, Andel, Nikolai, Kadlecova and Laczo 35 An olfactory identification test was also a useful clinical marker for monitoring the effectiveness of symptomatic medications such as cholinesterase inhibitors in Alzheimer's disease patients,Reference Li, Wang, Wu, Shi, Zhou and Lin 36 and may contribute to the differential diagnosis with depression.Reference Solomon, Petrie, Hart and Brackin 37
Despite clinical interest in developing olfactory measures for identifying patients with Alzheimer's disease and related disorders, a recent review highlighted contradictory results in this area. This finding may be explained by the use of different tests and different methodologies among studies; thus, the lack of generally applicable instruments prevents olfactory testing being integrated into the routine clinical evaluation of Alzheimer's disease.Reference Sun, Raji, Maceachern and Burke 38 For acceptance by the scientific and medical communities, new olfactory tests should be both reliable for research use and suitable for assessing Alzheimer's disease patients in daily clinical practice.
Indeed, since the discovery that olfactory regions are affected in Alzheimer's disease, numerous studies have aimed to identify an olfactory test that can predict disease development or help with diagnosis. Nevertheless, at more than 10 years after the first published study, no gold standard method of measurement has yet been developed for general clinic application.
Currently available psychophysical tests
Olfactory tests differ depending on whether the aim is to explore the odour detection threshold or to identify or characterise the odour.
Most sniffing tests use odorant stimulations comprising a mixture of familiar compounds such as essential oils, raw materials or flavours. Their familiarity is designed to enable rapid completion of the description task. Most are related to odours generated by various foods, such as orange, clove, fish and vanilla.Reference Choudhury, Moberg and Doty 4 Alternatively, non-food smells generally include woody or flower smells; odours are chosen to be culture specific, although efforts have been made to set up internationally applicable tests.Reference Doty, Marcus and William Lee 39
These tests are insufficient for a standalone diagnosis and were not initially designed for use by ENT specialists or neurologists. However, the identification of olfactory disorders can help to establish a diagnosis for, and can even represent an early marker of, neurodegenerative diseases. It is necessary to use olfactory tests to assess olfactory disorders because patients rarely report these. Table I shows the tests currently used and their levels (peripheral and/or central nervous system) of assessment, the pathological conditions for which they were developed and for which they are currently used, their sensitivity and specificity in diagnosing Parkinson's and Alzheimer's disease patients, and their strengths and weaknesses.
Table I Characteristics of existing psychophysical olfactory disorder tests
UPSIT = University of Pennsylvania smell identification test; B-SIT = brief smell identification test; ETOC = European test of olfactory capabilities; AD = Alzheimer's disease; PD = Parkinson's disease; MCI = mild cognitive impairment; US = United States (of America); Q-SIT = quick smell identification test
Towards a new diagnostic tool?
This review highlighted the important points that olfactory disorders (1) are qualitatively different in cognitively healthy and Alzheimer's disease individuals and (2) may predict conversion from mild cognitive impairment to Alzheimer's disease. However, no gold standard olfactory assessment instrument is currently available for diagnosing or monitoring Alzheimer's disease in daily clinical practice. This is primarily due to a lack of consensus on the validity of existing olfactory tests for clinical practice and research purposes. At least one study tried to compare different olfactory tests to identify the most reliable one, but it failed to identify a reference tool independent of the population and pathology of interest.Reference Gu and Li 75
Pathology-specific tests
No test has been developed and used specifically for neurodegenerative diseases: tests developed for ENT diseases are usually extended to neurodegenerative diseases, and vice versa. Although the European test of olfactory capabilities, the ‘Sniffin’ Sticks’ test and the quick smell identification test were designed specifically for Alzheimer's disease, they are preferentially used for other types of ENT diseases.Reference Alt, Mace, Buniel, Soler and Smith 53 – Reference Soler, Hyer, Ramakrishnan, Smith, Mace and Rudmik 55 , Reference Mahlknecht, Pechlaner, Boesveldt, Volc, Pinter and Reiter 70 , Reference Joussain, Bessy, Faure, Bellil, Landis and Hugentobler 74 This may be because lack of a single, reliable test makes it difficult to compare results among different studies. Therefore, clinicians working on Alzheimer's disease have no reliable evidence to support the use of olfactory testing of Alzheimer's disease patients. Consistent with this hypothesis, a recent review highlighted the importance of developing a single, reliable test for routine clinical use in Alzheimer's disease patients.Reference Sun, Raji, Maceachern and Burke 38 It is unlikely that a single all-purpose test will ever be developed because of the large number of pathologies in which olfaction is affected, such as ENT, psychiatric and neurodegenerative diseases, which all have different aetiologies and effects on olfaction. Therefore, efforts should instead be made to develop tests specific to a single pathology and culture or country.
Culture-specific tests
Olfaction is strongly affected by culture, with familiar and unfamiliar odours varying among countries and regions, making a single test unlikely to have general utility. This explains why researchers in different countries have modified existing tests or developed their own tests. However, the human odorant receptor gene repertoire is also highly variable, suggesting that most people will not have the same response to odorant stimulation.Reference Secundo, Snitz, Weissler, Pinchover, Shoenfeld and Loewenthal 76 Genetic effects on the olfactory perception of single compounds are only beginning to be understood. For example, the perception of methanthiol is affected by genetic variation only in Caucasian people: no correlation has been demonstrated for African people.Reference Pelchat, Bykowski, Duke and Reed 77 As these variations are likely to be highly complex for mixtures of odorants, olfactory stimulation by single odorant compounds appears preferable.
Test composition
Testing stimulation using pure compounds might be simpler for several reasons. Quality control is more straightforward for a pure compound than for a mixture that may contain tens of chemicals. In addition, the chemical composition of the blend used in these tools may be unknown to the user and also depends on the commercial constraints of the supplier, which are likely to change over time (for example, because of economical or safety concerns). Even if the olfactory response triggered by a single compound is not simpler than that of a complex mixture, it seems intuitively better to make olfactory tests using pure odorants. As a comparison, visual or auditory tests use simple stimuli (rather than complex) to identify dysfunction.Reference Štenc Bradvica, Bradvica, Matić and Reisz-Majić 78 , Reference Näätänen, Kujala, Escera, Baldeweg, Kreegipuu and Carlson 79 Monitoring olfaction with a complex blend is similar to monitoring the auditory response using a symphony rather than using reproducible sounds at defined frequencies. Even single odorants can trigger responses corresponding to familiar stimulants with a single descriptor, at least for people of the same cultural background.
Above all, the use of pure compounds would enable odour perception to be linked to the pharmacology of the olfactory system, a task that is difficult to perform with complex mixtures. Human beings perceive odours through the stimulation of odorant receptors expressed by olfactory sensory neurons located in the nasal epithelium.Reference Buck and Axel 80 Humans possess close to 400 different functional odorant receptor genes,Reference Niimura 81 and the differential activation of these receptors encodes the olfactory signal within our brain. The current consensus is that a given odour is associated with a ‘combinatorial code’ of odorant receptor activation. Thus, the pharmacology of odorant receptors and their role in the perception of pure compounds are beginning to be uncovered.Reference De March, Ryu, Sicard, Moon and Golebiowski 82
Conclusion
Olfactory disorders may predict the conversion from mild cognitive impairment to Alzheimer's disease. Currently, no gold standard olfactory test is available for diagnosing or monitoring Alzheimer's disease in clinical practice. The development of a single, reliable assessment tool for Alzheimer's disease populations is thus critical. This tool should be specific to the pathology and culture of interest, and should use pure odorants (to simplify the analysis and to determine genetic factors and psycho-physiological effects). Future efforts should aim to understand why olfactory tests developed specifically for memory centres are not used. For example, clinicians may not be accustomed to the olfactory system, be unable to store odorant correctly, have insufficient time and may not be convinced by the available evidence. The objective is to develop a test that will account for all clinical, cultural and molecular factors mentioned in this review.
Acknowledgements
The authors thank K W Zhu and H Matsunami for their comments and revisions. This work was supported by grants from the Gen Foundation (to CAdM) and from the Centre National de la Recherche Scientifique (défi AUTON 2016) to JG. Thanks also go to Giract for providing a PhD bursary.
