Pulmonary exacerbations in cystic fibrosis.

The clinical characteristics most relevant to the decision to treat for a pulmonary exacerbation with antibiotics in cystic fibrosis patients were determined. Variables including age, increased cough frequency and sputum production, new crackles and wheezing, asthma, symptomatic sinusitis, hemoptysis, decreased lung function, weight loss, and new acquisition of Pseudomonas aeruginosa were collected in a large prospective multicenter database (Epidemiologic Study of Cystic Fibrosis). During a 12-month baseline period, data from 11692 patients were compared with data collected during the subsequent 6-month study period. Because pulmonary function assessments were unavailable for patients <6 years of age, separate analyses were done for those <6 and >or=6 years of age. The outcome of interest was any antibiotic treatment in the 6-month study period reported as indicated for an exacerbation. Characteristics with the most discriminatory power were determined using stepwise multiple logistic regression. For patients <6 years of age, the strongest independent associations with treatment for a pulmonary exacerbation were new crackles, increased cough frequency, decline in weight, and increased sputum production. For those patients >or=6 years of age, the strongest independent associations were a relative decrease in percent predicted forced expired volume in 1 sec, increased cough frequency, new crackles, and hemoptysis. The presence of three or more of these key characteristics was strongly associated with the occurrence of a treated exacerbation. The reproducibility of the model over time was confirmed by application to a subsequent set of data. This model has potential for use as an outcome measure in clinical trials, and to assist in treatment decisions for individual patients.