Biological markers in nasal secretions provide valuable information on nasal pathophysiology. However, published data on biomarker concentrations in nasal fluids are remarkably inconsistent, and the bias due to different sampling techniques, has not yet been systematically evaluated.Concentrations of various protein were repeatedly determined in nasal secretions of 16 healthy volunteers. The proteins were detected by using: 1) a 2 -macroglobulin as a marker for plasma contamination; 2) lactoferrin as a marker for glandular secretion; 3) lactate dehydrogenase as a marker for tissue injury; and 4) interleukin (IL)-1b, IL-8, tumour necrosis factor-a, and eosinophil cationic protein and tryptase as indicators for tissue inflammation. A total of four different sampling methods, including nasal lavage (NL) and a new polyurethane foam sampler technique (PFST) were employed.Analyte concentrations in NL were approximately10-times lower than in specimens obtained by PFST. Due to the unpredictable dilution during NL, various analytes were below the detection limit of the high sensitivity assays employed. With PFST, concentrations below the detection limit rarely occurred. The specimens did not significantly differ regarding plasma contamination, glandular secretion or tissue injury.The considerable variability of reported analyte concentrations in nasal secretions mainly results from different sampling techniques. To collect nasal secretions, samplers are considered superior to nasal lavage techniques. Eur Respir J 2003; 21: 600-605.