The aim of the study was to measure the peripheral blood levels of soluble E-selectin in patients with systemic inflammation and compare them with in vivo granulocyte activation, pulmonary intravascular granulocyte pooling, pulmonary extravascular granulocyte migration and 99mTc-diethylenetriaminepenta-acetic acid (DTPA) aerosol clearance, an index of lung injury. The level of soluble E-selectin was measured by capture ELISA. Granulocytes were labelled with 111In and 99mTc for quantification of pulmonary granulocyte kinetics. The pulmonary vascular granulocyte pool (PGP) was expressed as a fraction of the total blood granulocyte pool. Pulmonary granulocyte migration was quantified on 24-h images using the 111In signal. Granulocyte activation was quantified as the percentage of circulating cells showing shape change (‘primed’). Lung injury was assessed from the clearance rate of inhaled 99mTc-DTPA aerosol. Eighteen patients with systemic inflammation were studied: five with inflammatory bowel disease, eight with systemic vasculitis, four with graft versus host disease and one with a recent renal transplant. The peripheral blood levels of soluble E-selectin were significantly elevated in patients with systemic inflammation. The level of soluble E-selectin showed a significant association with granulocyte migration (Spearman rank correlation coefficient, Rs = 0.53; P< 0.05) but not with PGP or with the percentage of cells showing shape change (P> 0.05 for both). Granulocyte migration was bimodal: patients were therefore subdivided into ‘migrators’ and ‘non-migrators’. Soluble E-selectin level, 99mTc-DTPA clearance and PGP, but not the percentage of cells showing shape change, were significantly higher in migrators than in non-migrators. We conclude that pulmonary intravascular granulocyte pooling is increased in the presence of increased numbers of circulating primed granulocytes but increased pooling does not by itself promote granulocyte migration into the lung interstitium. Insofar as an elevated level of E-selectin in peripheral blood reflects vascular endothelial activation, the data are consistent with the notion that pulmonary endothelial activation is required, in addition to granulocyte activation and an expanded PGP, for granulocyte migration into lung parenchyma and, therefore, for lung injury to occur.