OBJECTIVE To explore the alterations of serum of myelin basic protein (MBP) concentration in the plasma and ultrastructure in the spinal cord after continuous intrathecal injection of different ropivacaine concentrations in rats. METHODS Ninety-six male Sprague-Dawley rats weighing 220 to approximately 280 g were randomly divided into a control group (Group N), Group R(1), R(2) and R(3) (24 rats in each group). Each group was subdivided into 4 subgroups (6 rats in each subgroup). According to the method of Yaksh's, a polyurethane microspinal catheter was inserted into the lumbar subarachnoid space in which 8 cm segment was left. Rats in each group were continuously received 40 microL of intrathecal injection of normal saline(Group N), 0.5%, 0.75%, and 1.0% ropivacaine (Group R(1),R(2),R(3)), 3 times every 1.5 hours. Blood (0.5 mL) was drawn from the femoral artery to determine serum concentrations of MBP at the detecting time T(0)(before inserted pipe)and T(1)(before the first intrathecal injection); for the subgroups, the examining time was at T(2), T(3), T(4) and T(5)(6, 12, 24 and 48 h respectively after the last time intrathecal administration). After blood was drawn, the rats in each subgroups were decapitated and the spinal cord of L(1-2) intumescentia lumbalis were immediately removed for electronic microscopic examination. RESULTS MBP levels were comparatively steady in Group N, R(1) and R(2), while there was statistical difference between Group R(3) and Group N, R(1),R(2),and R(3) (P<0.05). MBP level of Group R(3) was significantly higher at T(2),T(3),T(4) and T(5) than that at T(0)(P<0.01). The ultrastructural changes of the spinal cord in Group R(3) were pycnosis of most neurons, dilation of most rough endoplasmic reticulum, and vague structure of mitochondria and endocytoplasmic reticulum. A few neurons were completely de-generated losing the normal structure, with vacuole degeneration of crista mitochondriales or even partial loss. CONCLUSIONS The spinal cord ultrastructure is selectively vulnerable after intrathecal 1.0% ropivacaine injection, which may be one of the important pathophysiological bases for local anesthetic neurotoxicity. MBP may serve as a sensitive and specific indicator of spinal cord damage after intra-thecal administration of ropivacaine.