Lung function changes before and after pulmonary exacerbation
antimicrobial treatment in cystic fibrosis

BACKGROUND: In cystic fibrosis, observation of a lung function drop (as percent predicted forced expiratory volume in 1 s [FEV1 ]; ppFEV1 ) frequently precedes pulmonary exacerbation (PEx) diagnosis. Recovery of ppFEV1 to a previous ``baseline'' is commonly used to assess antimicrobial treatment response. However, not all diagnosed PEx are associated with a ppFEV1 drop, and it is unclear whether these are a different type of PEx from those associated with a ppFEV1 drop. METHODS: We analyzed pre- and posttreatment ppFEV1 for PEx recorded in the Epidemiologic Study of Cystic Fibrosis from 2003 through 2005. Baseline, pretreatment, and follow-up ppFEV1 were the best recorded within 12-months pre-PEx, the lowest recorded -30 to +3 days of treatment, and the best recorded during 6-month follow-up, respectively. Logistic regression models for return of ppFEV1 to baseline during follow-up were developed separately for PEx with $\geq$10%, <10%, and no ppFEV1 drop before treatment. RESULTS: Of 15 147 PEx, 10 166 (67.1%), 3479 (23.0%), and 1502 (9.9%) presented with a $\geq$10%, <10%, or no ppFEV1 drop at diagnosis, respectively. 19.5%, 35.2%, and 65.6% of PEx, respectively, had follow-up ppFEV1 equal to or exceeding baseline; overall 27.7% of all PEx treatments resulted in complete recovery of baseline ppFEV1 . Significant predictors of ppFEV1 recovery at follow-up were younger patient age, absence of Aspergillus, lower baseline ppFEV1 , fewer visits during the baseline, lower frequency of prior-year PEx, shorter elapsed time from baseline measure to treatment, smaller relative ppFEV1 drop before treatment, and non intravenous (ie, oral or inhaled antibiotic) treatment. PEx with $\geq$10%, <10%, and no ppFEV1 drop before treatment had only modest differences in covariate odds ratios associated with complete ppFEV1 recovery. CONCLUSIONS: Among the 10% of PEx presenting with no apparent ppFEV1 drop, more than one-third resulted in a decreased ppFEV1 during follow-up. Risk factors for this outcome were the same as those associated with lack of ppFEV1 recovery among PEx with pretreatment ppFEV1 drops. These results suggest that inherent FEV1 variability, baseline and follow-up sampling methodologies, ppFEV1 regression to the mean, and underlying lung disease progression complicate this approach for assessing effects of PEx and treatment response.