Cluster Analysis to Identify Clinical Subtypes of Mnire's Disease

Objective: To identify distinct clinical subtypes of Meniere's disease by analyzing data acquired from a UK registry of patients who have been diagnosed with Meniere's disease.Study Design: Observational study.Methods: Patients with Meniere's disease were identified at secondary/tertiary care clinics. Cluster analysis was performed by grouping participants sharing similar characteristics and risk factors into groups based on a defined measure of similarity.Results: A total of 411 participants were recruited into this study. Two main clusters were identified: participants diagnosed with ear infections (OR = 0.30, p < 0.014, 95% CI: 0.11-0.78) were more likely to be allocated in Cluster 1 (C1). Participants reporting tinnitus in both ears (OR = 11.89, p < 0.001, 95% CI: 4.08-34.64), low pitched tinnitus (OR = 21.09, p < 0.001, 95% CI: 7.47-59.54), and those reporting stress as a trigger for vertigo attacks (OR = 14.94, p < 0.001, 95% CI: 4.54-49.10) were significantly more likely to be in Cluster 2 (C2). Also, participants diagnosed with Benign Paroxysmal Positional Vertigo (OR = 13.14, <0.001, 95% CI: 4.35-39.74), autoimmune disease (OR = 5.97, p < 0.007, 95% CI: 1.62-22.03), depression (OR = 4.72, p < 0.056, 95% CI: 0.96-23.24), migraines (OR = 3.13, p < 0.008, 95% CI: 1.34-7.26), drug allergy (OR = 3.25, p < 0.029, 95% CI: 1.13-9.34), and hay fever (OR = 3.12, p < 0.009, 95% CI: 1.33-7.34) were significantly more likely to be clustered in C2.Conclusions: This study supports the hypothesis that Meniere's disease is a heterogeneous condition with subgroups that may be identifiable by clinical features. Two main clusters were identified with differing putative etiological factors.Level of Evidence: 3