Identification of phenotypic clusters of nonsteroidal anti-inflammatory drugs exacerbated respiratory disease

Abstract

BACKGROUND: Clinical presentation of nonsteroidal anti-inflammatory drugs exacerbated respiratory disease (NERD) is found to be heterogeneous. This study classified phenotypic clusters to determine NERD subtypes. METHODS: We performed two-step cluster analysis using urticaria, chronic rhinosinusitis (CRS), and atopy, in a NERD cohort comprising 302 patients. Asthma exacerbation was defined as receiving at least one burst of intravenous steroid treatment and/or at least two bursts of oral steroid use (>/= 45 mg/3 days) per year. The possession rate of anti-asthmatic medications was estimated during the follow-up period. RESULTS: There were four subtypes: subtype 1 (NERD with CRS/atopy and no urticaria), subtype 2 (NERD with CRS and no urticaria/atopy), subtype 3 (NERD without CRS/urticaria), and subtype 4 (NERD with urticaria). Significant differences were found between the four subtypes in the female proportion, baseline FEV1%, serum total IgE level, and sputum/peripheral eosinophil count. A higher frequency of asthma exacerbations was noted in subtype 1 compared to subtype 3. The possession rates of medium- to high-dose inhaled corticosteroids/long-acting beta2 -agonists showed significant differences among the four subtypes. Metabolomic analysis showed that the four subtypes of NERD had a higher serum leukotriene E4 (LTE4) level than those with aspirin-tolerant asthma. The patients with subtypes 1 and 3 had a higher urine LTE4 level than those with subtype 2. CONCLUSION: We found four distinct subtypes with different clinical/biochemical findings and asthma exacerbations in a NERD cohort. These findings suggest that stratified strategies by applying subtype classification may help achieve better outcomes in the management of NERD.