Cardiopulmonary bypass-induced organ dysfunction remains a clinical problem in certain groups of patients. Although the pathogenesis is multifactorial, it is likely that a panendothelial injury consequent upon widespread humoral and cellular activation is a major contributor to this process. The biologically active products of complement activation are certainly capable of inducing many of the features of the post-perfusion syndrome. The complex interactions between complement and many of the other proposed mediators of this response also supports this contention. However, it is equally certain that many of the other proposed mediators have some role to play. Inhibition of one cell type or inflammatory cascade is therefore unlikely to abolish all the adverse effects of CPB but will, at least in experimental systems, permit a more precise determination of the pathogenesis of this problem. The temptation to simply measure elevated circulating levels of newly identified mediators must be resisted and more effort applied to examining the pathophysiological effects of specific inhibitors. This type of investigation should initially be effected in experimental models where reproducible conditions can be ensured. In conjunction with this, far more precise end-points are required in order to assess the effect of any potential therapeutic intervention in a clinical setting. In particular, new techniques of evaluating endothelial injury need to be developed. In clinical studies careful consideration must be given to the patient population studied. Whilst patients undergoing routine coronary artery surgery form a relatively homogeneous group, the magnitude of endothelial injury sustained is probably small and, especially in terms of lung function, the signal will be diluted by other non-bypass-related events. The study of high risk groups would seem more appropriate despite their heterogeneity. An important unanswered question is why certain sub-populations of patients are at increased risk of clinically relevant bypass-induced injury. The endothelium of these patients may be different: the neonatal pulmonary microcirculation is not the same as that of an adult (with increased fluid filtration pressure and a higher microvascular surface area per unit lung mass [5,6]), children with pulmonary hypertension have histological evidence of an altered/damaged endothelium (S.G. Haworth, Personal Communication) whilst pre-existing sepsis could clearly induce a degree of endothelial dysfunction. A further possibility is that the inflammatory response in these patients is already "primed". Some patients with heart failure have been shown to have elevated circulating TNF.(ABSTRACT TRUNCATED AT 400 WORDS)