Key Points
The emotional burden of olfactory dysfunction bears distinct psychological profiles.
Objective olfactory dysfunction findings do not reliably reflect patients' emotional burden.
Olfactory‐related quality of life assessment may identify patient risk of emotional distress.
1. Introduction
Olfactory dysfunction (OD) is associated with neuropsychiatric comorbidity and reduced quality of life (QOL) [1]. OD may also co‐occur with depressive disorders [2, 3], although findings from large population‐based studies have been inconsistent [4]. Emotional responses to olfactory loss vary widely, ranging from profound distress to minimal impact. Notably, subjective olfactory‐related QOL was found to be a stronger predictor of well‐being than objective performance in psychophysical tests [5]. Machine‐learning approaches applied to large olfactory datasets have recently identified distinct patient characteristics not captured by conventional analyses [6]. Here, we applied a data‐driven clustering approach to subjective and affective outcomes with the aim of identifying distinct psychological responses to OD and examined their relationship with psychophysical olfactory performance [5].
2. Methods
The study patients were recruited from the Smell and Taste Clinic at a tertiary urban medical center between 2024 and 2025. The healthy controls were volunteers without known conditions affecting olfactory function. The exclusion criteria were pregnancy, recent upper respiratory infection, and corticosteroid use within 4 weeks. Psychophysical olfactory function was assessed by the validated Hebrew Sniffin’ Sticks Threshold‐Discrimination‐Identification (TDI) test [7, 8]. Participants completed the He‐WHO‐5 Well‐Being Index, Positive and Negative Affect Schedule (He‐PANAS), Questionnaire of Olfactory Disorders (He‐QOD), Nasal Obstruction Symptom Evaluation (He‐NOSE) scale, and Sino‐Nasal Outcome Test‐22 (He‐SNOT‐22), and subjective taste impairment was assessed by a dichotomous scale.
Group comparisons were performed by independent‐samples t‐tests, χ 2 tests, or Mann–Whitney U tests, as appropriate. K‐means clustering was performed on standardized PANAS, WHO‐5, and QOD scores, with the SNOT‐22 emotional and sleep subdomains additionally included based on preliminary regression analyses examining predictors of well‐being. The optimal number of clusters was determined by the Gap Statistic (k = 3). TDI scores were compared across clusters by one‐way analysis of variance with Tukey post hoc testing. Statistical significance was set at p < 0.05, and size effects are reported. Analyses were conducted using Jamovi (v2.3) and Python (v3.11).
3. Results
3.1. OD Patients Versus Controls
In total, 80 OD patients (mean age 44.7 ± 16.6 years; 57.5% male) and 48 controls (mean age 40.8 ± 14.3 years; 45.8% male) were enrolled. The OD patients had significantly lower He‐TDI scores (t(117) = 10.1, p < 0.001, d = 1.89) and a higher sinonasal symptom burden on both He‐SNOT‐22 (U = 747.5, p < 0.001, r = 0.56) and He‐NOSE (U = 1073.5, p < 0.001, r = 0.39), as well as a higher burden of olfactory‐related QOL (He‐QOD) (U = 116.0, p < 0.001, r = 0.92). In contrast, general well‐being and affective measures showed mixed results: the patients reported significantly reduced positive affect (U = 1255, p = 0.013, r = 0.27), but there were no significant group differences for negative affect (U = 1705, p = 0.957, r = 0.01) or well‐being (U = 1560, p = 0.258, r = −.012) (Table S1).
3.2. Clustering Based Upon Subjective and Affective Measures
K‐means clustering (BSS/TSS = 49.5%; silhouette = 0.25) identified three profiles. Cluster 1 (high distress; n = 26) had elevated He‐SNOT‐22 sleep/emotional scores, high He‐QOD impairment, low He‐WHO‐5 scores, and high negative affect. Cluster 2 (intermediate distress; n = 29) showed moderate impairment and a mild effect on well‐being. Cluster 3 (low distress; n = 44) showed minimal impairment and high well‐being scores, comprising mostly controls (91%) (Table S3). The three clusters were well‐separated in principal component analysis space (PC1: 54.0%; PC2: 18.4%; combined: 72.4%), supporting the distinctiveness of the derived profiles (Figure 1). Cluster membership was strongly associated with group assignment (χ 2(2) = 73.4, p < 0.001, Cramer's V = 0.86) (Figure S1).
FIGURE 1.
Three distinct psychological response profiles emerging from cluster analysis.
Two‐dimensional visualization of the three psychological response profiles derived from K‐means clustering of standardized affective (PANAS), well‐being (WHO‐5), and OD‐related quality‐of‐life (SNOT‐22, QOD) variables. Each symbol represents an individual participant. The shapes and colors are coded by cluster membership.
3.3. Intermediate‐Distress Versus High‐Distress Profiles
Group comparisons between the high‐distress (Cluster 1) and intermediate‐distress (Cluster 2) profiles revealed no significant difference in age or etiology, although there was a significantly higher proportion of females in the high‐distress profile (χ 2(1) = 5.26, p = 0.022, Φ = 0.31). Interestingly, subjective taste impairment was more prevalent in the high‐distress profile (χ 2(1) = 5.59, p = 0.018, Φ = 0.33) (Table 1). Analysis of the psychophysical olfactory performance revealed that the profiles differed significantly on the TDI total score (t(47) = 2.54, p = 0.015) with a higher odor threshold (U = 158, p = 0.006, r = 0.45) and a higher odor discrimination (t(47) = 2.82, p = 0.007, d = 0.81) in the high‐distress cluster (Figure S2).
TABLE 1.
Groups differences between high and intermediate distress.
| Measure | Cluster 1: high‐distress profile | Cluster 2: intermediate distress profile | p value; effect size |
|---|---|---|---|
| Patients, n | 26 | 29 | |
| Age | 38.7 ± 14.7/38.5 | 43.6 ± 16.8/47.0 | 0.254; d' = 0.31 |
| Male (%) | 34.6 | 65.5 | 0.022; Φ = 0.31 |
| Conductive etiology (%) | 43.5 | 41.4 | 0.879; Φ = 0.02 |
| Phantosmia and parosmia (%) | 30.4 | 34.5 | 0.757; Φ = 0.04 |
| Subjective taste impairment (%) | 52.2 | 20.7 | 0.018; Φ = 0.33 |
| OD duration (months) | 49.0 ± 91.8/24.0 | 80.7 ± 97.4/48.0 | 0.192; r = 0.21 |
| TDI total score | 24.9 ± 10.03/26.6 | 18.2 ± 8.54/15.5 | 0.015; d' = 0.72 |
| Threshold | 5.89 ± 3.92/5.63 | 3.48 ± 3.30/1.50 | 0.006; r = 0.45 |
| Discrimination | 10.00 ± 3.22/11.00 | 7.24 ± 3.62/8.00 | 0.007; d' = 0.81 |
| Identification | 9.04 ± 4.16/10.00 | 7.60 ± 3.63/7.00 | 0.202; d' = 0.37 |
Note: Continuous variables are presented as mean ± standard deviation (SD) and /median. Conductive etiology was determined based on physical examination, radiological findings, and other relevant clinical assessments. Phantosmia and parosmia were identified based on patient self‐report.
Abbreviations: OD, olfactory dysfunction; TDI, Olfactory Threshold, Discrimination and Identification.
4. Discussion
Unlike prior studies that have focused primarily on psychophysical testing, the current study clusters participants based on their subjective responses to validated olfactory‐related QOL assessments. We identified three distinct psychological response profiles, demonstrating that emotional burden is not determined by the severity of sensory loss according to TDI, but rather by the perceived deprivation of olfactory capability. Notably, the intermediate‐distress cluster exhibited the poorest psychophysical olfactory performance yet better psychological functioning than the high‐distress cluster. This dissociation supports prior evidence that subjective olfactory‐related QOL, rather than psychophysical olfactory performance, more strongly influences emotional burden outcomes [5, 6, 9].
The higher prevalence of subjective taste impairment in the high‐distress cluster suggests that perceived multisensory disruption may amplify the psychosocial burden [4]. Furthermore, differences in the TDI threshold and discrimination, rather than decreased identification scores, point toward a potential role for early perceptual processing in shaping emotional experience in this setting. These findings support the routine assessment of olfactory‐related QOL and affective state alongside psychophysical testing to better identify patients at risk for emotional repercussions of sensory loss. Targeted interventions incorporating olfactory training, hedonic re‐learning, and psychoeducation may be particularly beneficial for high‐distress patients [10].
This study has several limitations, including its cross‐sectional design, recruitment from a single tertiary referral center, and reliance upon self‐report measures. Longitudinal studies incorporating cognitive and neurobiological measures are needed to further elucidate causal mechanisms. Overall, the identification of distinct psychological response profiles conceptualizes OD as a multidimensional condition requiring integrated sensory and affective assessment.
Ethics Statement
The study received Institutional Review Board approval (#TLV‐0017‐24).
Conflicts of Interest
The authors declare no conflicts of interest.
Supporting information
Supplementary Figure 1. Distribution of psychological response profiles. Distribution of participants with olfactory dysfunction and healthy controls across the three psychological response profiles identified by k‐means clustering.
Supplementary Figure 2. Validation of psychological response profiles. Distribution of TDI total scores by cluster, demonstrating that objective olfactory performance does not linearly correlate with emotional response.
Supplementary Table 1. Group differences between patients with olfactory dysfunction and healthy controls
Supplementary Table 2. Correlations between QOL‐related PROMs Measurements.
Supplementary Table 3. Affective state, well‐being, and QOL scores across the clusters detected in our cohort.
References
- 1. Zald D. H. and Pardo J. V., “Functional Neuroimaging of the Olfactory System in Humans,” International Journal of Psychophysiology 36, no. 2 (2000): 165–181, 10.1016/S0167-8760(99)00110-5. [DOI] [PubMed] [Google Scholar]
- 2. Kohli P., Soler Z. M., Nguyen S. A., Muus J. S., and Schlosser R. J., “The Association Between Olfaction and Depression: A Systematic Review,” Chemical Senses 41, no. 6 (2016): 479–486, 10.1093/chemse/bjw061. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Croy I., Nordin S., and Hummel T., “Olfactory Disorders and Quality of Life—An Updated Review,” Chemical Senses 39, no. 3 (2014): 185–194, 10.1093/chemse/bjt072. [DOI] [PubMed] [Google Scholar]
- 4. Lassale C., Liljas A. E. M., Jones A., Cadar D., and Steptoe A., “Association of Multisensory Impairment With Quality of Life and Depression in English Older Adults,” JAMA Otolaryngology–Head & Neck Surgery 146, no. 3 (2020): 278–285, 10.1001/jamaoto.2019.4470. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Yoo J. H., Kim T. S., Kim J. S., Lee S. H., and Seo M. Y., “Subjective Distress Mediates the Association between Olfactory Dysfunction Duration and Depression in Post COVID 19 Patients,” Scientific Reports 15, no. 1 (2025): 22216, 10.1038/s41598-025-07953-z. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Lötsch J., Kringel D., and Hummel T., “Machine Learning in Human Olfactory Research,” Chemical Senses 44, no. 1 (2019): 11–22, 10.1093/chemse/bjy067. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Peled I., Fischer T., Wasserman G., et al., “Validation of the Hebrew Version of the “Sniffin′ Sticks” Olfactory Test: A Cross‐Sectional Study in an Israeli Cohort,” European Archives of Oto‐Rhino‐Laryngology 2025 282 (2025): 5655–5664, 10.1007/S00405-025-09627-Y. [DOI] [PubMed] [Google Scholar]
- 8. Hummel T., Sekinger B., Wolf S. R., Pauli E., and Kobal G., “‘Sniffin’ Sticks′. Olfactory Performance Assessed by the Combined Testing of Odor Identification, Odor Discrimination and Olfactory Threshold,” Chemical Senses 22, no. 1 (1997): 39–52, 10.1093/chemse/22.1.39. [DOI] [PubMed] [Google Scholar]
- 9. Liu D. T., Prem B., Sharma G., Kaiser J., Besser G., and Mueller C. A., “Depression Symptoms and Olfactory‐Related Quality of Life,” Laryngoscope 132, no. 9 (2022): 1829–1834, 10.1002/lary.30122. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Oleszkiewicz A., Croy I., and Hummel T., “The Impact of Olfactory Loss on Quality of Life: A 2025 Review,” Chemical Senses 50 (2025): bjaf023, 10.1093/chemse/bjaf023. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Supplementary Figure 1. Distribution of psychological response profiles. Distribution of participants with olfactory dysfunction and healthy controls across the three psychological response profiles identified by k‐means clustering.
Supplementary Figure 2. Validation of psychological response profiles. Distribution of TDI total scores by cluster, demonstrating that objective olfactory performance does not linearly correlate with emotional response.
Supplementary Table 1. Group differences between patients with olfactory dysfunction and healthy controls
Supplementary Table 2. Correlations between QOL‐related PROMs Measurements.
Supplementary Table 3. Affective state, well‐being, and QOL scores across the clusters detected in our cohort.