Cluster Analysis of World Trade Center Related Lower Airway Diseases

Rafael E. de la Hoz, Yunho Jeon, John T. Doucette, Anthony P. Reeves, Raúl San José Estépar, Juan C. Celedón

Research output: Contribution to journalArticlepeer-review


Introduction: Cluster analysis can classify without a priori assumptions the heterogeneous chronic lower airway diseases found in former workers at the World Trade Center (WTC) disaster site. Methods: We selected the first available chest computed tomography scan with quantitative computed tomography measurements on 311 former WTC workers with complete clinical, and spirometric data from their closest surveillance visit. We performed a nonhierarchical iterative algorithm K-prototype cluster analysis, using gap measure. Results: A five-cluster solution was most satisfactory. Cluster 5 had the healthiest individuals. In cluster 4, smoking was most prevalent and intense but there was scant evidence of respiratory disease. Cluster 3 had symptomatic subjects with reduced forced vital capacity impairment (low FVC). Clusters 1 and 2 had less dyspneic subjects, but more functional and quantitative computed tomography evidence of chronic obstructive pulmonary disease (COPD) in cluster 1, or low FVC in cluster 2. Clusters 1 and 4 had the highest proportion of rapid first-second forced expiratory volume decliners. Conclusions: Cluster analysis confirms low FVC and COPD/pre-COPD as distinctive chronic lower airway disease phenotypes on long-term surveillance of the WTC workers.

Original languageEnglish
Pages (from-to)179-184
Number of pages6
JournalJournal of Occupational and Environmental Medicine
Issue number2
StatePublished - 1 Feb 2024


  • 2001
  • World Trade Center Attack
  • chronic obstructive pulmonary disease
  • longitudinal changes in lung function
  • occupational lung disease
  • smoke inhalation injury
  • spirometry


Dive into the research topics of 'Cluster Analysis of World Trade Center Related Lower Airway Diseases'. Together they form a unique fingerprint.

Cite this