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 internationally. We sought to determine if pooling adversely affected the diagnostic properties of BAL. METHODS: Children undergoing bronchoscopy for clinical reasons were included. A multi-step study protocol ensured BAL was collected and analysed both separately and as a pooled fluid. RESULTS: Eighty-five children (53 CF, 32 control) were recruited. There was a high level of concordance between pooled and non-pooled samples in terms of organism identification (76%). There was good agreement (Bland Altman) between the two methods in terms of detection of inflammation independent of centre, microbiological concordance or disease status. Bi-directional variability in IL-8 levels between pooled and non-pooled samples was seen. Free neutrophil elastase (NE) was detected in 4 cases in pooled lavage when absent in non-pooled lavage. Levels of interleukin-8 (IL-8) were similar between the two groups with pooled samples showing a greater spread of values. CONCLUSIONS: Pooling of BAL in children does not negatively impact on either the detection of pulmonary infection or inflammation or the observed relationship between infection and inflammation. Intra-patient variability in BAL IL-8 levels suggests regional differences in inflammation.