Clinical and Imaging Evaluation of COVID-19-Related Olfactory Dysfunction

COVID-19 has been closely associated with chemosensory dysfunctions, including olfactory loss, which may be the sole presenting symptom in some patients. Prior meta-analyses report OD in approximately 45–48% of COVID-19 cases. Olfactory loss can be conductive (due to nasal airflow obstruction) or sensorineural (due to olfactory epithelium or central pathway dysfunction). Unlike common post-viral anosmia that may be related to nasal obstruction, COVID-related OD often lacks sinonasal symptoms, suggesting a different pathophysiology. A plausible mechanism is sensorineural loss related to olfactory bulb (OB) involvement and potential neural invasion via the olfactory cleft (OC). MRI of the olfactory pathway is useful to evaluate OB volume, shape, and signal. Abnormal OB volumes/signal in COVID-19 OD support central involvement. This study analyzes clinical and imaging findings in COVID-related OD to explore underlying mechanisms.

Prior studies and meta-analyses have documented a high prevalence of OD in COVID-19 (approximately 45–48%). Imaging studies reported increased OB signal intensity and volume consistent with transient inflammation/edema, supporting neural spread to the OB. Reports also describe OB hypersignal or enhancement (suggesting microbleeding/blood–brain barrier disruption) and punctate T2 hypointensities. OC edema/obstruction has been observed in early disease, and systematic reviews list OC opacification and sometimes normal OB morphology as common findings. Histopathologic reports in COVID-19 patients with OD show olfactory epithelial atrophy, leukocytic infiltration, and axonal damage. While many cases show OB swelling acutely, OB atrophy has also been reported in persistent anosmia or follow-up imaging, suggesting dynamic changes over time.

Design: Prospective study conducted at the Otorhinolaryngology and Diagnostic Radiology Departments, Mansoura University Hospitals, Egypt, from October 2020 to October 2021, with IRB approval (MFM-IRB: MS.21.01.1340) and informed consent. Participants: 110 adults with post-COVID-19 OD (anosmia/hyposmia) and 50 COVID-negative controls with normal olfaction. Inclusion: Prior SARS-CoV-2 infection confirmed by RT-PCR; recovery confirmed by at least two negative nasopharyngeal swabs; self-reported OD validated by olfactory testing scores. Controls were individuals undergoing MRI for unrelated indications. Exclusion: Preexisting olfactory loss before the pandemic, prior nasal surgeries, head trauma, chronic sinusitis, degenerative neurological disorders, or MRI contraindications. Sample size/power: Post hoc power analysis (alpha 0.05) indicated study power 77–100% depending on OB dimensions as primary outcome. Clinical assessment: Detailed history emphasizing smell dysfunction and nasal complaints (obstruction, discharge). Endoscopic nasal examination with focus on the olfactory cleft (OC). OC boundaries defined anatomically. OC mucosa graded with the olfactory cleft Lund-Kennedy (OC-LK) score (discharge, polyps, edema, scarring, crusting; each 0–2 per side; total 0–20). Olfactory testing: Smell diskettes test (Briner et al.), 8 odors (coffee, vanilla, peach, grass, pineapple, rose, chocolate, fish). Patients abstained from eating/drinking/smoking 15 min prior. Scores 0–8; 7–8 normal; ≤6 hyposmia/anosmia. Imaging: All participants underwent olfactory pathway MRI on a 1.5T Philips Ingenia with a 16-channel coil. Coronal 3D-FLAIR: TR/TE 8000/133 ms; bandwidth 120 Hz/pixel; TSE factor 80; slice thickness 2 mm; gap 0.5 mm; FOV 230×230 mm2 (from tip of nose to posterior sphenoid; skull vault to hard palate). Neuroradiologists blinded to clinical data analyzed images. Measurements: OB dimensions (length, width, height) measured; OB volume calculated using Box-frame method (length × width × height). OC status assessed for edema/opacity. Presence of sinusitis (opacity in any paranasal sinus) recorded. Statistics: SPSS v26. Normality by Kolmogorov–Smirnov. Qualitative variables as n(%); associations via Chi-square. Continuous data as mean±SD (parametric) or median (min–max) (nonparametric). Student t test for group comparisons (parametric). Significance threshold P ≤ .05.

Participants: 110 patients with COVID-related OD and 50 controls; no significant differences in age (P = .105) or gender (P = .102). Olfactory testing: Patients scored 1–3 (score 1 in 42, 2 in 55, 3 in 13); controls scored 7–8 (7 in 8, 8 in 42), consistent with OD in the patient group and normal olfaction in controls. Associated symptoms in patients: Altered taste (ageusia/hypogeusia) in 98/110 (89.1%). OD with or without taste change not associated with other nasal/general symptoms in 57/110 (51.8%). Other symptoms included sore throat 52 (47.3%), fever/fatigue 42 (38.9%), headache 40 (36.4%), nasal discharge 38 (34.5%), nasal obstruction 36 (32.7%), diarrhea 20 (18.2%), abdominal pain 17 (15.5%), dyspnea 16 (14.5%), nausea/vomiting 16 (14.5%), productive cough 22 (20%), dry cough 27 (24.5%). OC assessment: Endoscopy (OC-LK) showed significantly higher OC edema in patients vs controls (P = .002). MRI showed OC edema/opacity in 63/110 (57.3%) patients vs 1/50 (2%) controls (P ≤ .001). MRI evidence of sinusitis was present in 17/110 patients (15.5%) and in 1 control. OB measurements (mean±SD):

• Right OB height: 3.57±0.72 mm (patients) vs 2.35±0.41 mm (controls), t=11.14, P ≤ .001.
• Left OB height: 3.57±0.75 vs 2.43±0.46 mm, t=9.83, P ≤ .001.
• Right OB width: 4.51±0.60 vs 3.80±0.77 mm, t=6.03, P ≤ .001.
• Left OB width: 4.40±0.69 vs 3.45±0.59 mm, t=8.28, P ≤ .001.
• Right OB length: 10.03±2.77 vs 8.92±2.25 mm, t=2.42, P = .017.
• Left OB length: 10.16±2.43 vs 9.44±2.08 mm, t=1.78, P = .077.
• Right OB volume: 81.13±32.47 vs 40.48±15.04 mm3, t=8.35, P ≤ .001.
• Left OB volume: 81.50±31.05 vs 39.79±16.06 mm3, t=8.89, P ≤ .001.
• Mean OB volume (R+L): 80.81±29.01 vs 40.14±13.42 mm3, t=9.36, P ≤ .001. OB atrophy: Observed in 8/110 patients (7.3%); unilateral right in 2 and bilateral in 6. Overall: Patients demonstrated significantly increased OB dimensions and volume compared with controls, and more frequent OC edema, with relatively low rates of sinusitis.

The findings support a predominantly sensorineural mechanism for COVID-19-related olfactory dysfunction. Significantly increased OB dimensions and volumes indicate OB edema/swelling consistent with neural involvement, likely due to viral spread through the olfactory epithelium and OC to central olfactory pathways. The frequent absence of sinonasal symptoms (over half of patients reported OD without other nasal/general symptoms) and the low incidence of radiologic sinusitis argue against a primarily conductive mechanism. OC edema/opacity in more than half of patients further supports localized inflammatory changes at the olfactory interface. These results align with prior imaging and histopathologic studies reporting transient OB edema and olfactory epithelial injury in COVID-19. Notably, a minority exhibited OB atrophy, which may represent later-stage or persistent damage in some cases. The integration of clinical, endoscopic, and MRI findings cohesively indicates sensorineural impairment as the principal pathophysiology in COVID-related OD.

OB volumes were significantly higher in patients with COVID-related OD compared to controls, indicating OB edema/swelling. Most patients lacked sinonasal symptoms and had low rates of sinusitis on imaging. Collectively, these findings suggest that COVID-related OD is most likely due to sensorineural loss via viral spread and injury to the olfactory epithelium and pathways. Future work should include longitudinal follow-up with repeat MRI to assess the temporal evolution of OB changes and recovery, and exploration of biomarkers or interventions that may predict or enhance olfactory recovery.

Cross-sectional assessment at presentation without longitudinal follow-up; MRI was not repeated to evaluate reversibility or progression of findings. Generalizability may be limited by single-center design and lack of pre-COVID baseline imaging. Affiliation-specific factors and potential selection bias (controls undergoing MRI for other reasons) may also influence results.