Objective To investigate the biomechanical characteristics of self-developed distraction reduction fixation system (DRFS) and to evaluate its cl inical effect on thoracolumbar vertebrate fracture. Methods Twelve fresh porcine spines were prepared for the biomechanical test. The stiffness of each intact specimen were recorded on forward bending (20°),backward bending (10°), lateral bending (30°) and axial rotation (20°), and then specimens of fracture dislocation were made by cuneiform osteotomy of L3 vertebral body and excision of posterior facet joints. Finally, DRFS internal fixation was performed on each specimen. The stiffness of specimens in fracture dislocation and after DRFS fixation were measured during the same movements mentioned above, respectively. The cl inical effect of DRFS on thoracolumbar vertebrate fracture in 31 patients (aged 17-46 years with an average of 32.1) from April 1998 to October 2002 was summarized. Fracture types were classified according to Denis classification: 2 patients suffered simple compressed fracture, 16 burst fracture, and 13 fracture dislocation, including 2 cases of T11, 11 cases of T12, 14 cases of L1 and 4 cases of L2. Frankel and X-ray examination were adopted to assess the results. Results The stiffness during forward bending, backward bending, lateral bending and axial rotation in the fracturedislocation state was markedly lower than that of the corresponding movements of the intact porcine spines (P lt; 0.05). After DRFS, the stiffness during various movements increased to the level that significantly higher than that in the fracture-dislocation state (P lt; 0.05), and the stiffness during backward bending was of significant difference (P lt; 0.05), but without significant difference during the rest three movements (P gt; 0.05) when compared to that in intact spines. Regarding cl inical observation, the operation time was 2.2-4.1 hours (2.7 hours on average) and blood loss was 250-600 mL (450 mL on average). The patients were followed up for 5-20 months (10.2 months on average). The heal ing time for fracture was 5-11months (8 months onaverage). The mean anterior and posterior heights of the injured vertebrate recovered from 46.2% ± 7.5% and 76.4% ± 2.4% preoperatively to 89.8% ± 4.6% and 94.1% ± 1.5% postoperatively (P lt; 0.05).The median point height also raised from 60.8% ± 6.4% to 90.7% ± 2.9% (P lt; 0.05). The Cobb’s angle decreased from (26.3 ± 5.9)° to (5.2 ± 1.8)° (P lt; 0.05), and all the sl ipped vertebrates were well repositioned. Spine function was assessed by Frankel classification as follows: 2 of 5 Class A preoperatively improved to Class B postoperatively, and the other 3 remained unchanged ; 4 of 6 from B to C, and the other 2 to D; 6 of 13 from C to D, and the other 7 to E; and 7 from D to E. Conclusion DRFS is capable of providing sufficient stabil ity, meeting the requirement of the spine physical activity without any obvious side effect. DRFS, as a handy, safe and effective technique in cl inical practice, is featured by satisfying functions of distraction, reposition and fixation.