Aims. In a multicentre, randomised study of adolescents undergoing posterior spinal fusion for idiopathic scoliosis, we investigated the effect of adding gelatine matrix with human thrombin to the standard surgical methods of controlling blood loss. Patients and Methods. Patients in the intervention group (n = 30) were randomised to receive a minimum of two and a maximum of four units of gelatine matrix with thrombin in addition to conventional surgical methods of achieving haemostasis. Only conventional surgical methods were used in the control group (n = 30). We measured the intra-operative and total blood loss (intra-operative blood loss plus post-operative drain output). Results. Each additional hour of operating time increased the intra-operative blood loss by 356.9 ml (p < 0.001) and the total blood loss by 430.5 ml (p < 0.001). Multiple linear regression analysis showed that the intervention significantly decreased the intra-operative (-171 ml, p = 0.025) and total blood loss (-177 ml, p = 0.027). The decrease in haemoglobin concentration from the day before the operation to the second post-operative day was significantly smaller in the intervention group (-6 g/l, p = 0.013) than in the control group. . Conclusion. The addition of gelatine matrix with human thrombin to conventional methods of achieving haemostasis reduces both the intra-operative blood loss and the decrease in haemoglobin concentration post-operatively in adolescents undergoing posterior spinal fusion for idiopathic scoliosis. Take home message: A randomised clinical trial showed that gelatine matrix with human thrombin decreases intra-operative blood loss by 30% when added to traditional surgical haemostatic methods in adolescents undergoing posterior spinal fusion for idiopathic scoliosis. Cite this article: Bone Joint J 2016;98-B:395–401