Ozone at ambient concentrations affects lung function and initiates an inflammatory response of the airways. However, the underlying mechanisms are poorly understood. In vitro studies have shown that ozone reacts with water to give reactive hydroxyl radicals capable of oxidizing a wide range of biomolecules. We conducted a study to determine if in vivo hydroxyl radical attack on human airways occurs under natural exposure to ozone. The relation of orthotyrosine to para-tyrosine as a measure of hydroxyl radical attack was analyzed in nasal lavage samples of 44 primary school children in an epidemiologic study. Repeated nasal lavages were performed between May and October 1991 both following "low" (daily half-hour maximum < 140 micrograms/m3, approximately 70 ppb) and "high" (daily half-hour maximum > 180 micrograms/m3, approximately 90 ppb) ozone exposure. Concomitantly, lung function tests were performed. On average, 11.6 (6-16) nasal lavages were performed for each of 24 study days (10 days following "low" ozone exposure and 14 days following "high" ozone exposure). Average ortho-tyrosine (median; 5-95% percentile) for each child was 0.037 mumol/L (0.016-0.064 mumol/L) and average para-tyrosine was 15.7 mumol/L (9.8-24.1 mumol/L). Ortho-tyrosine (as percentage of tyrosine) was significantly higher following days with "high" ozone exposure (0.18%) vs. days following "low" ozone exposure (0.02%; p = .0001). Ortho-tyrosine showed an inverse relationship with forced vital capacity (p = .01) but was not related to inflammation of the upper airways as assessed by cell counts of polymorphonuclear neutrophils. Hydroxyl radical attack subsequent to ambient ozone occurs in the upper airways of healthy children and is related to lung function decrements.