The Multiple Faces of Non-Cystic Fibrosis Bronchiectasis. A Cluster Analysis Approach

    loading  Checking for direct PDF access through Ovid



The clinical presentation and prognosis of non-cystic fibrosis bronchiectasis are both very heterogeneous.


To identify different clinical phenotypes for non-cystic fibrosis bronchiectasis and their impact on prognosis.


Using a standardized protocol, we conducted a multicenter observational cohort study at six Spanish centers with patients diagnosed with non-cystic fibrosis bronchiectasis before December 31, 2005, with a 5-year follow-up from the bronchiectasis diagnosis. A cluster analysis was used to classify the patients into homogeneous groups by means of significant variables corresponding to different aspects of bronchiectasis (clinical phenotypes): age, sex, body mass index, smoking habit, dyspnea, macroscopic appearance of sputum, number of exacerbations, chronic colonization with Pseudomonas aeruginosa, FEV1, number of pulmonary lobes affected, idiopathic bronchiectasis, and associated chronic obstructive pulmonary disease. Survival analysis (Kaplan-Meier method and log-rank test) was used to evaluate the comparative survival of the different subgroups.

Measurements and Main Results:

A total of 468 patients with a mean age of 63 (15.9) years were analyzed. Of these, 58% were females, 39.7% had idiopathic bronchiectasis, and 29.3% presented with chronic Pseudomonas aeruginosa colonization. Cluster analysis showed four clinical phenotypes: (1) younger women with mild disease, (2) older women with mild disease, (3) older patients with severe disease who had frequent exacerbations, and (4) older patients with severe disease who did not have frequent exacerbations. The follow-up period was 54 months, during which there were 95 deaths. Mortality was low in the first and second groups (3.9% and 7.6%, respectively) and high for the third (37%) and fourth (40.8%) groups. The third cluster had a higher proportion of respiratory deaths than the fourth (77.8% vs. 34.4%; P < 0.001).


Using cluster analysis, it is possible to separate patients with bronchiectasis into distinct clinical phenotypes with different prognoses.

Related Topics

    loading  Loading Related Articles