Hyperosmolar environments induce histamine release from mast cells and basophils in vitro. To assess whether the same stimulus induces mediator release in vivo, 15 healthy human volunteers underwent nasal challenges with instilled solutions of differing osmolalities: lactated Ringer's solution (257 +/- 3 mOsm/kg), isosmolar mannitol (277 +/- 6 mOsm/kg), and hyperosmolar mannitol (869 +/- 8 mOsm/kg). The effect of these challenges on the volume, osmolality, and inflammatory mediator content of subsequent 5-ml isosmolar lavages was determined. The volumes of lavages returned after hyperosmolar challenges were significantly greater than those after isosmolar challenges (5.5 +/- 0.2 ml versus 4.2 +/- 0.1 ml; p less than 0.01) and these lavage solutions had higher osmolalities. Even when corrected for increased volumes, the lavages after hyperosmolar challenges contained significantly higher quantities of inflammatory mediators such as histamine (29.0 versus 10.1 ng; p less than 0.01), TAME-esterase activity (32.7 versus 11.1 cpm x 10(-3); p less than 0.01), and immunoreactive leukotrienes (9.9 versus 3.4 ng; p less than 0.01). The changes in mediators were dose dependent in that incremental increase in challenge osmolality were associated with incremental increases in histamine release. Therefore, when exposed to hyperosmolar stimuli in vivo, the nasal respiratory airway releases inflammatory mediators and fluid rapidly shifts into the airway lumen. It has been suggested that the mediator release observed on breathing cold and dry air is due to increased osmolality of airway secretions; the present data confirm that osmotic variations at the airway surface can provide an adequate stimulus for cell activation.