Depressive symptoms and olfactory function in older adults


*Anna Scinska, MD, PhD, Department of Otolaryngology, Warsaw Medical Academy, Czerniakowski Hospital, Stepinska 19/25 Street, Warsaw 00-739, Poland. Email:


Aims:  Neuroimaging studies suggest a significant overlap between brain regions involved in the regulation of olfaction and mood. The aim of the present study was to search for correlations between depressive symptomatology measured by the 15-item Geriatric Depression Scale (GDS) and olfactory function assessed with Sniffin' Sticks in non-demented older adults (aged 53–79 years).

Methods:  Taste detection thresholds were also measured by means of electrogustometry on the anterior tongue.

Results:  No correlation was found between the GDS scores (range: 0–12) and olfactory thresholds or olfactory identification scores. Similarly, there was no relationship between depressive symptoms and electrogustometric thresholds. Subjects (n = 25) scoring ≥5 on the GDS were classified as ‘depressed’ and all other individuals (n = 60) were classified as ‘non-depressed’. The two groups did not differ in terms of the olfactory measures and electrogustometric threshold.

Conclusion:  Depressive symptoms are not associated with any major olfactory deficit in non-clinical older adults.

THE SENSE OF smell plays important and potentially even life-saving roles for subjects from all age groups. Preserved olfactory function can be particularly beneficial for food hedonics, safety, personal hygiene and social life in older adults.1,2 Olfaction is also known to be closely associated with brain regions regulating emotions and memory processes.3–5 Thus, it is not surprising that olfactory tests have been considered as simple tools for early diagnosis of various neuropsychiatric disorders.3,4,6

Early and specific diagnosis is of vital importance for successful treatment of neurodegenerative disorders, including Alzheimer-type dementia, in older adults. A growing body of evidence indicates that patients with Alzheimer-type dementia suffer from marked olfactory dysfunction, including deficits in olfactory acuity (e.g. increased thresholds), discrimination, recognition, and identification. In line with the above, it has been reported that olfactory tests facilitate screening for dementia in the elderly.3,6

It is widely recognized that depressive symptoms can mimic cognitive impairment and dementia in older patients.7 Therefore, it has been suggested that olfactory tests could also improve differential diagnosis between dementia and depression in older subjects.8–10 Clinically diagnosed depression is not consistently associated with olfactory abnormalities, and olfactory deficits observed by some authors in severely depressed subjects were relatively mild. Olfactory thresholds in depressed psychiatric patients were either increased,11,12 reduced13,14 or unchanged13,15–17 as compared to healthy controls. Olfactory identification abilities in depressed psychiatric patients were intact in most studies on this topic.10,11,15,18–20 Similarly, intensity ratings of various odors did not differ between psychiatric inpatients with depressive episode and control subjects.11,12,17 In some of these studies,12 positive correlations were found between olfactory deficits and the self-rated intensity of depressive symptoms assessed by the Beck Depression Inventory (BDI).21 In contrast, olfactory measures were not correlated with the clinician-rated severity of depression measured with the Hamilton Depression Rating Scale (HAM-D).13,22 Taken together, the conflicting results suggest that possible alterations in olfactory function can be a state, rather than a trait, marker of depressive episode and depend on the self-rated severity of depression.12

The value of olfactory tests in screening for dementia and for differential diagnosis between dementia and depression in primary care cannot be fully ascertained without studies on olfactory function in non-clinical older adults with depressive symptoms. Previous reports on this topic are sparse. Satoh et al. have found a negative relationship between the Zung Self-rating Depression Scale (SDS) scores and perceived intensities of three of eight odors in elderly Japanese men.23 There was no correlation between the SDS scores and olfactory intensity ratings in elderly Japanese women. A single study addressed possible relationship between olfactory measures and depressive symptomatology in older, non-clinical population of European origin. No association between olfactory identification abilities and the BDI scores was observed in Greek adults (aged 49–88 years).24 The two studies did not address the problem of reliability and validity of the instruments used for the assessment of depressive symptoms.23,24 This issue is important because somatic complaints in older subjects may be misinterpreted as representing depressive symptomatology. The BDI used in the Greek study contains several somatic items24 and thus has been criticized as a tool for the measurement of late-life depression.25,26 Moreover, although some authors reported alterations in olfactory threshold in depressed psychiatric patients,11–14 olfactory threshold was not assessed in the two studies on non-clinical older adults.23,24

The primary purpose of the present study was to search for correlations between depressive symptoms measured by the Geriatric Depression Scale (GDS-15) and olfactory function (threshold and identification abilities) in non-clinical older adults. The GDS was selected because it is a simple instrument with yes/no answers specifically developed to measure depression in older populations.26

Recently our group has shown no relationship between electrogustometric thresholds and depressive symptomatology assessed with the BDI in older adults.27 In order to repeat and extend the latter observation, taste detection thresholds assessed by means of electrogustometry were also correlated with the GDS scores.

Although high reliability and validity of the GDS-15 has been confirmed,28,29 we decided to examine test–retest reliability and concurrent validity of its Polish version before the start of the major study.


Reliability and validity of the Geriatric Depression Scale

An original English version of the 15-item GDS26,30 was translated into Polish by a group of experienced clinical researchers. Back-translation was used to assess the accuracy of the Polish version. Test–retest reliability of the instrument was evaluated in a sample of non-psychiatric patients with non-malignant disorders (chronic sinusitis, nasal septum deviation, sleep apnea, tinnitus). The sample consisted of 51 non-demented individuals, between 51 and 78 years old, admitted for routine medical care to the departments of otolaryngology and neurology (mean age ±SE, 63.0 ± 1.1, 45.1% women, 37.3% smokers). The subjects assessed their depressive symptoms in the previous week with the GDS twice with an interval of 1–6 days (mean interval, 2.0 ± 0.15 days). The GDS scores ranged from 0 to 14. The mean value was 4.4 ± 0.47 (median value, 4) in the first test and 4.1 ± 0.47 (median value, 3) in the retest. The reliability of the GDS was very good (Spearman's R = 0.88, P < 0.00001 for the whole sample, R = 0.85, P < 0.00001 for the women participants).

Concurrent validity of the GDS was measured against the 21-item HAM-D22 in a subgroup of 25 subjects (57.7% women). The subjects were first asked to complete the GDS. Their scores ranged from 0 to 14. On the same day, a board-certified psychiatrist (L.S.) who was blind to the GDS scores assessed their depressive symptomatology by means of the HAM-D. The HAM-D scores ranged from 0 to 21, with a mean value of 5.0 ± 1.1. The validity of the GDS was satisfactory (R = 0.75, P < 0.0001).

Depressive symptoms and chemosensory function in older adults


Study participants were recruited through word of mouth from families of staff members and local community. All potential participants completed physical, neurological and laryngological examination before inclusion to the study. Subjects were excluded if they had a history of post-traumatic or upper respiratory infection-induced olfactory loss, severe facial trauma, a head injury with significant loss of consciousness or other medical conditions (e.g. stroke, epilepsy, nasal polyps, diabetes) that might alter their sense of smell.31 Only non-demented subjects whose Mini Mental State Examination scores (MMSE)32 were >24 were included to the study. The recruited subjects, 29 men and 56 post-menopausal women, were white, aged 53–79 years and free of any psychotropic medication (Table 1). No study participant reported receiving a formal diagnosis of drug dependence (except of nicotine dependence) or other psychiatric disorder.

Table 1.  Subject characteristics
  • *

    P < 0.0001; all other P were >0.05.

  • Scoring ≥5 on the 15-item GDS.

  • Mean ± SE.

  • §

    Sniffin' Sticks, range 1–16; the higher the number, the lower (i.e. better) threshold and better the identification abilities.

  • AUDIT, Alcohol Use Disorders Identification Test; GDS, Geriatric Depression Scale; MMSE, Mini Mental State Examination.

Sociodemographic parameters
Age (years)67.2 ± 1.266.7 ± 0.9
Weight (kg)71.8 ± 2.273.0 ± 1.1
Height (cm)164.0 ± 1.7164.4 ± 1.6
University degree11/2534/60
Presently employed2/259/60
GDS score6.5 ± 0.42.6 ± 0.2*
MMSE score28.8 ± 0.228.7 ± 0.2
AUDIT score1.6 ± 0.32.2 ± 0.3
Current smokers3/258/60
Olfactory and gustatory measures
Olfactory threshold§5.7 ± 0.55.7 ± 0.4
Olfactory identification§12.8 ± 0.712.5 ± 0.4
Electrogustometric threshold (µA)
Mean88.7 ± 21.3133.4 ± 16.0
Right side95.6 ± 26.3143.0 ± 18.3
Left side81.9 ± 20.7123.8 ± 15.7

The whole study was carried out in accordance with the Declaration of Helsinki. All participants read and signed an informed consent form after study procedures had been fully explained. The protocol for the study was reviewed and approved by the Ethics Committee on Human Studies of the Warsaw Medical Academy. The subjects were not compensated for their participation.

General design

A single test session was conducted between 10.00 hours and 12:30 hours in an air-conditioned, quiet room. The participants were asked to refrain from eating, drinking and smoking for at least 1 h prior to the test session. Consumption of water was allowed.

The subjects were questioned regarding basic sociodemographic variables, smoking cigarettes, and drinking alcohol. The Alcohol Use Disorders Identification Test (AUDIT) was used to assess alcohol consumption.33 The 15-item GDS was then completed by each participant. The GDS scores ranged from 0 to 12. The subjects (n = 25) scoring ≥5 were classified as ‘depressed’.28,30 All other individuals (n = 60) were classified as ‘non-depressed’.

Olfactory tests started 15 min after the completion of the GDS during which the participants were familiarized with experimental procedures. Electrogustometry started 10 min after the completion of olfactory tests. The whole session lasted 60–75 min. The olfactory tests and electrogustometry were performed by an experimenter blind to the subject's medical history and questionnaire scores.

Assessment of olfactory detection threshold and olfactory identification using sniffin' sticks

The Sniffin’ Sticks (Burghardt, Wedel, Germany) is a test of olfactory function based on pen-like odor dispensing devices. Test–retest reliability and validity of the test was confirmed in several studies.34–36 Recently, normative data from 3282 subjects have been published.37

Each subject wore a sleeping mask to prevent visual identification of odor-dispensing pens. Birhinal detection threshold for n-butanol was assessed using a single-staircase, three-alternative forced choice procedure. Three odor-dispensing pens were presented, in a randomized order, 2 cm below both nostrils. Two pens contained deionized water and the third pen contained one of sixteen concentrations of n-butanol. The procedure began with the presentation of triplet no. 16 including the pen with the lowest concentration of n-butanol. The participant had to identify the pen with the odorant. Reversal of the staircase was triggered when the odor-containing pen was correctly identified in two successive trials. The triplets were presented at intervals of approximately 20 s. No feedback was given to the participants as to the correctness of their responses. Olfactory threshold was defined as the mean of the last four of eight reversals. The test scores could range from 1 to 16. The higher the score, the lower (better) was the subject's threshold.34,37

Birhinal olfactory identification was assessed 5 min after the completion of threshold measurement. Sixteen pens with common odors were presented in front of both nostrils. The subject had to identify each odor by choosing one of four descriptors. The test scores could range from 1 to 16. The higher the score, the better were the subject's identification abilities.34,37


An electrogustometer (TR-06, Rion, Tokyo, Japan) used for electrogustometry is a standardized device for human taste studies. The apparatus allows delivery of anodal currents of low intensity (4–400 µA) at known duration. In the present study the stimulus duration was kept at 0.5 s.27,38,39 The electrogustometer was equipped with a stainless-steel, flat, circular stimulus rod (5 mm in diameter) and a larger indifferent electrode. During the test, the stimulus probe was placed on the tongue margin, 1.5 cm to the right or to the left from the midline. The indifferent electrode was attached to the subject's neck. The sequence of threshold measurement (right–left or left–right) was randomized across the subjects.

A modified version of initially ascending, single-staircase detection threshold procedure was used to assess electrogustometric taste thresholds.27 The subject was asked to signal any new taste sensation on the tongue by pressing the response button connected to a small buzzer. The current intensity was increased if no response occurred within 3 s. The current intensity was decreased (reversal) if the subject signaled detection of the stimulus. No feedback was given to the participants as to the correctness of their responses. The reported threshold (µA) was an average of the last four of six reversals.


The Spearman rank correlation test was used to search for correlations between the GDS scores and other parameters in the whole group. The Mann–Whitney U-test or Fisher exact test was used to compare the depressed and non-depressed subjects. P < 0.05 was considered significant. No correction for multiple comparisons was applied. All statistical analyses were performed with the aid of the Statistica 5.0 software package (StatSoft, Tulsa, OK, USA).


The Spearman rank correlation test indicated a moderate negative correlation between age and olfactory threshold (R = −0.28, P < 0.05). Because the lower values indicated the higher (worse) threshold, the latter correlation most likely reflected the physiological decline in olfactory function with increasing age.37 In line with this, a negative correlation was observed between age and olfactory identification (R = −0.32, P < 0.05). Electrogustometric thresholds did not correlate with age and the olfactory parameters (P > 0.05). No effect of gender was observed in the present study.

The GDS scores did not predict olfactory thresholds (R = 0.02, P > 0.05) or olfactory identification abilities (R = 0.10, P > 0.05). No relationship was observed between the GDS scores and the mean electrogustometric threshold (R = −0.15, P > 0.05) or threshold measured on the right (R = −0.14, P > 0.05) and left side of the anterior tongue (R = −0.16, P > 0.05).

The depressed and non-depressed individuals did not differ in terms of basic sociodemographic characteristics. The two groups did not differ in the olfactory and gustatory parameters (P > 0.05; Table 1).


High reliability and concurrent validity of the GDS-15 against other measures of depression have been reported for various populations.28–30,40 In the present study the initial validation of the GDS in non-psychiatric patients confirmed its high test–retest reliability and good concurrent validity against the HAM-D.22

The mean olfactory identification scores observed in the present study were similar to that reported for healthy people aged ≥55 years.37 The mean olfactory threshold for n-butanol in the previous report (7.15 and 7.44 for older men and women, respectively37) tended to be lower than the threshold observed in the present study. But Hummel et al. recruited subjects ‘in excellent health’,37 while most individuals recruited for the present study suffered from mild cardiovascular and/or metabolic disorders and took 1–2 prescription medications.

In general, we found no relationship between olfactory acuity and identification and depressive symptoms assessed by the GDS in the older non-clinical adults. The present data replicate and extend the results reported for the older Greek adults.24 In the latter study the BDI scores were not correlated with performance on the University of Pennsylvania Smell Identification Test or Cross-Cultural Smell Identification Test. Olfactory thresholds were not examined. In addition, the present findings partially replicate the report by Satoh et al.23 Those authors found negative correlations between depressive symptomatology and rated intensities of three out of eight olfactory stimuli in the elderly Japanese men, but no correlation between olfactory intensity ratings and depressive symptomatology was observed in the elderly Japanese women.23 Olfactory thresholds and identification abilities were not examined in the latter study.23 Comorbid depressive symptoms were not associated with olfactory identification scores in young adults with attention deficit-hyperactivity disorder,41 olfactory threshold, discrimination, identification and judgments in alcohol-dependent patients,35,42 or olfactory thresholds in patients with multiple chemical sensitivities.43

Two studies addressed possible relationship between depressive symptoms and olfactory function in non-clinical young adults.44,45 In both studies, no association was found between olfactory discrimination and depressive symptoms assessed with the BDI. In one of these studies some correlation was observed between the BDI scores and olfactory thresholds measured with the Sniffin Sticks.45 But the reported BDI scores (≤9) suggest that all subjects had minimal, if any, depressive symptoms.21,45

In the previous study we observed no correlation between the BDI scores (range: 3–33) and electrogustometric thresholds in 33 older adults without a history of psychiatric disorders.27 In line with this, no relationship between electrogustometric thresholds and depressive symptomatology measured with the GDS was found in the present study. Detection thresholds for sweet substances did not vary for severity of depression in psychiatric46 and non-psychiatric patients.47 More recently, no relationship between depressive symptoms and taste thresholds has been reported for non-psychiatric Japanese subjects.48 Thus, it seems that depressive symptomatology is not associated with any central or peripheral deficit precluding detection of threshold taste stimuli. This notion is further supported by findings that patients with major depressive disorder and non-psychiatric subjects with depressive symptomatology gave similar intensity ratings to weak taste stimuli as compared to non-depressed individuals.18,27 Interestingly, depressive symptoms may lead to abnormal judgment of relatively strong taste stimuli.27,49

In conclusion, the present study casts into doubt any relationship between depressive symptoms and olfactory threshold or identification in non-psychiatric older adults. In addition, the present results confirm that depressive symptomatology is not associated with any major alterations in electrogustometric thresholds.


This work was supported by the Institute of Psychiatry and Neurology (grant no. 67/07) and by the Ministry of Science and Graduate Education (grant no. 0864/P05/2005/29).