To investigate clinical features of childhood vasovagal syncope (VVS) and the possible relationship between changes of plasma and platelet 5-hydroxytryptamine (5-HT) and childhood VVS. Forty-one children who were diagnosed as VVS because of positive head-up tilt test (HUTT) in Capital Institute of Pediatrics were enrolled as HUT-positive group, while 36 healthy children as control group. Clinical features of all children were analyzed, and blood samples of all children were obtained. Plasma and platelet 5-HT was measured by enzyme-linked immunosorbent assay (ELISA). (1) The mean age of 41 VVS children was (10.5 +/- 1.8) years, and there were more girls than boys with the boys to girls ratio of 1:1.4. (2) Presyncopal symptoms occurred in 33 patients (80.4%), among whom dizziness had a high rate: 78.8%. (3) Commonly, there were some provocation factors before syncope, among which long-time standing was the most common one with the rate of 91.7%. (4) The mean time of positive response in BHUT and SNHUT were (20.6 +/- 8.6) minutes and (5.0 +/- 2.2) minutes, respectively. Duration of syncope was shorter than 5 minutes. (5) HUTT positive response included vasodepressor type with the rate of 61.0%, cardioinhibitory type with 14.6%, and mixed type with 24.4%. (6) There were no significant differences in baseline heart rate, systolic blood pressure and diastolic blood pressure between VVS children and healthy children. And it was the same among different types of VVS children. (7) There were no significant differences in plasma 5-HT between VVS group of baseline or HUTT-positive and control group [(27.51 +/- 1.32) microg/L vs.(27.28 +/- 2.48)microg/L, t = 0.518, P = 0.606; (27.51 +/- 1.32) microg/L vs.(28.05 +/- 1.40) microg/L, t = 2.044, P = 0.167]. There were no significant differences in platelet 5-HT concentration between VVS group of baseline and control group [(82.30 +/- 6.06) 10(9) ng/L vs. (79.88 +/- 5.79) 10(9) ng/L, t = 1.788, P = 0.780].(8) HUTT-positive platelet 5-HT concentration of VVS children was significantly higher than baseline value [(97.90 +/- 6.59) 10(9) ng/L vs. (82.30 +/- 6.06) 10(9) ng/L, t = 11.26, P = 0.00]. There were no significant changes in plasma 5-HT in children with VVS during baseline, syncope or pre-syncope, which suggests that plasma 5-HT might not be valuable for the prediction of syncope trigger. However, platelet 5-HT of VVS children was obviously higher during syncope and presyncope, which suggests that central serotonergic system might be involved in the pathogenesis of VVS.