Pooling of bronchoalveolar lavage in children with cystic fibrosis does not affect the microbiological yield and may allow earlier detection of pulmonary inflammation

Abstract

Background: Bronchoalveolar lavage (BAL) is a potentially useful outcome measure for clinical trials in children with CF but its use is limited by variations in approach in different centres. First lavages (the airway component) are frequently used for culture, and in many institutions, subsequent lavages (distal/alveolar fractions) are used for measurement of inflammation. Pooling of BAL offers the opportunity to gather a homogenous sample of greater volume with a greater representation of the proximal airway component. We sought to determine if pooling of all lavage fractions adversely affected the relationship between infection and inflammation in BAL compared to a standard approach of using first lavages for culture and subsequent lavages for inflammatory markers. Results: 53 children with CF (mean age 4.5 years range 1-12) and 32 controls (mean age 4.3 years range 1-11) were recruited. Of 234 culture results in 85 samples, 192 (82%) showed concordance between the first and pooled lavages in terms of identification of organisms. Of these 192, semiquantitative culture results were similar in 186 and differed in 6 (4 greater growth in pooled lavage and 2 greater in first lavage). Of the 42 instances where culture results differed, in 8 cases (3.4%) this represented a known pathogen (4 cases where it grew in first lavage alone, and 4 in the pooled lavage alone). ROC curves demonstrated similar efficacy of pooling and first lavage as diagnostic tools for detection of known pathogens. Bland Altman analysis demonstrated strong agreement between the two methods in terms of detection of inflammation when infection was present. Neutrophil elastase (NE) was detected in 28 samples and undetected in 57. In 4 cases NE was detected in pooled lavage where it was undetected in standard lavage. Mean levels of interleukin-8 (IL-8) were higher in the pooled samples (4776pg/mL) compared to standard lavage (2820pg/mL), showed a greater spread of values and were better able to discriminate between CF and control and between those with and without infection in BAL (not statistically significant). Conclusion: Pooling of BAL allows for the collection of larger volume samples without adversely affecting the relationship between infection and inflammation, and may be better suited to detecting changes in inflammation as a result of treatment interventions.L6123590812016-10-03 <br />