Identification of rhinitis phenotypes using an unsupervised approach in the Constances cohort

Unsupervised approach has been scarcely used for rhinitis in adults population-based studies. We aimed to identify phenotypes using this the French cohort Constances. Among participants answering 2014 questionnaire, we included those with current i.e. reporting sneezing, runny or blocked nose out a cold flu, last 12 months. Based on 25 variables referring characteristics (persistence, severity, triggers, seasonality, treatments) and co-morbidities, dimension reduction Factor Analysis of Mixed Data followed by K-means algorithm were clusters. 5516 (50 years old, 57% women, 20% ever-asthmatics), three clusters identified: cluster 1 (C1) n=2586, 47%, C2 n=2379, 43% C3 n=551, 10%. C1 was characterized few identified allergic triggers (dust dust mites, animals pollens reported less than 10% 53% did not know what triggered their symptoms). In C3, more 95% ever nasal allergies 100% from had severe rhinitis. Gradual increase observed asthma co-morbidity (C1=7%, C2=31%, C3=38%), combined medication (oral antihistamines intranasal corticosteroids) C2=37%, C3=56%) eosinophils count (109/L) (C1=181, C2=208, C3=231), all p-trend andlt