Elevated plasma serotonin is associated with primary pulmonary hypertension (PPH). To test whether this elevation could be related to platelet activation, the 2 pools of blood serotonin (platelets and plasma) and plasma 5-hydroxyindoleacetic acid (5-HIAA) as well as markers of platelet activation (alpha(IIb)beta(3), CD36, P-selectin, and CD63 membrane epitopes) were measured in 16 patients with severe PPH (group 1) before and at days 10 and 40 of treatment with a continuous infusion of epoprostenol (prostacyclin). The same biological parameters were also measured in 19 healthy subjects (group 2) and in 10 patients after cardiovascular surgery with extracorporeal circulation (group 3), a condition known to profoundly activate the platelets. Twelve PPH patients showed hemodynamic and clinical improvement, 3 remained stable, and 1 had the treatment stopped because of clinical aggravation. At day 0, mean plasma serotonin (5-hydroxytryptamine [5-HT]) concentration was much higher in PPH patients than in normal subjects (34.4+/-21.2 versus 9.1+/-6.0 nmol/L, respectively; P:<0.001) and positively correlated with total pulmonary resistance. The mean platelet 5-HT content was not significantly different in PPH compared with normal individuals. Mean plasma 5-HIAA concentrations were much higher in PPH than in normal patients (162+/-57 versus 61+/-7 nmol/L, respectively; P<0.001). These parameters did not significantly change during epoprostenol treatment. There was no correlation between the changes in plasma 5-HT during treatment and clinical or hemodynamic improvement. In PPH patients, the mean platelet volume significantly decreased (ANOVA, P<0.01) during treatment. Positive correlations were evidenced between the size of platelets and the number of alpha(IIb)beta(3) and CD36 epitopes. When compared with control platelets, the number of alpha(IIb)beta(3) epitopes detected on PPH platelets at day 0 tended to be higher, but this difference did not reach a statistical significance (41 300+/-7140 for PPH patients versus 36 010+/-3930 for control subjects, P=0.069). The number of CD36 epitopes, in the range of controls at day 0 (11 590+/-5080 for PPH patients versus 11 900+/-1790 for control subjects), decreased during treatment (ANOVA, P=0.038) and became significantly low at day 40 (8660+/-3520, P<0.001). The number of CD63 epitopes was not elevated, and P-selectin was never detected at any time point on PPH platelets. This glycoprotein profile indicates that the platelets of PPH patients were not highly activated but had an accelerated turnover and returned to normal under epoprostenol treatment without change of the elevated plasma serotonin, characteristic of PPH. In conclusion, neither platelet activation nor a significant alteration of the 5-HT endothelial metabolism explains the high level of plasma 5-HT in PPH patients. The 5-HT plasma concentration is not a predictive marker of the severity of PPH, and its evolution is independent of the clinical and hemodynamic status. Treatment by a potent antiaggregating agent, epoprostenol, does not affect the increase of plasma 5-HT, despite a therapeutic benefit.