Inflammatory markers in sputum and serum have been used with variable success as outcome measures in interventional studies. Limited data are available on reproducibility of such assays in cystic fibrosis (CF) particularly over a long time period. This study was designed to address this; stable patients (FEV1>40, >10 years age) were recruited into the study which ran over an 18-month period with up to four hospital visits. Patients provided a 24 h sputum collection, for weighing at each visit. Spontaneous sputum was collected at the beginning of each visit; if insufficient sample was obtained, sputum was induced with hypertonic saline. Inflammatory markers were measured in dithiothreitol-processed sputum (total and differential cell counts, IL8 (and other cytokines), calprotectin, neutrophil elastase, myeloperoxidase and extracellular DNA). Blood was collected at each visit for cytokines (IL1β, IL6, IL8, IL10, IL12 (p40) and TNFα) and calprotectin. Data are available from 189 patients at 655 visits. Adequate sputum for analysis was obtained at only 60% of visits. Sputum induction accounted for only 16% of adequate samples. Median and range for each detectable assay are shown below. Serum cytokines IL1β, IL10, IL12 (p40) and TNFα were undetectable at each visit, and IL6 was only detectable in 17% of samples. To assess intra-individual variability the coefficient of variation of results across each visit for each patient is presented. Both sputum and serum assays showed a large range of results at each visit, but the variation for each individual was much higher than the ideal 10%. Serum assays were not able to discriminate between CF and non-CF, apart from calprotectin (CF 10.1(0.8–72.5) vs non-CF 0.40 (0.2–1.12) p<0.001). Due to the difficulty in obtaining sputum samples reliably and the large variability of results between visits in these stable patients, we consider it unlikely that a change due to a new therapy would be detectable. As such, we are not considering sputum inflammatory markers as primary or secondary efficacy endpoints in our multidose gene therapy trial. Serum markers also appear to be of limited use is assessing efficacy but will still be useful for toxicology and safety studies.