The effusions for which PCHOL and most of the >ther parameters had the worst misclassification rate yere the neoplastic exudates (Table 3). Similar findings lave been reported by others, who suggested that the nisclassified exudates had low cell component concentrations because the pleura had only recently been iffected by the tumor. Since only rarely is it possible 😮 determine when pleural involvement commences, t is hardly feasible to use this argument to explain >vhy a neoplastic exudate has the biochemical charac-:eristics of a transudate. Like other authors, we relieve that a more likely explanation is that the pathogenesis of neoplastic exudates involves more than >ne mechanism more frequently than that of other kinds.

Although we excluded from our study all those cases in which it was reasonable to suspect the presence of more than one cause of pleural effusion, we cannot categorically guarantee that neoplastic invasion of the pleura was the sole cause of all the neoplastic exudates included. Because of the existence jf cases with multiple etiology, it seems improbable that any biochemical parameter can perfectly separate transudates from exudates without prior screening ‘although this is, of course, no reason not to continue looking for the best parameter there may be).
We conclude that PCHOL and P/SCHOL are both lighly effective in distinguishing between pleural exudates and transudates. Both perform better than :he parameters proposed by Light et al and are extremely cost efficient. High PCHOL and P/SCHOL values appear to be related to increased permeability )f pleural capillaries. We suggest that determination >f these parameters should be routine practice in pleural effusion cases.