Anesthetics can either promote or inhibit the development of neurogenic pulmonary edema (NPE) after central nervous system (CNS) injury. The influence of isoflurane was examined in male Wistar rats using 1.5%, 2%, 2.5%, 3%, 4%, or 5% isoflurane in air. Epidural balloon compression of the thoracic spinal cord was performed. The development of NPE was examined in vivo and on histologic sections of lung tissue. Animals anesthetized with 1.5% or 3% isoflurane were behaviorally monitored using the BBB and plantar tests for 7 weeks post-injury. The spinal cord was examined using MRI and morphometry of the spared white and gray matter. All animals from the 1.5% and 2% groups developed NPE. Almost 42% of the animals in the 1.5% group died of severe pulmonary hemorrhage and suffocation; x-rays, the pulmonary index, and the histological picture revealed a massive NPE. More than 71% of the animals from the 2.5% and 3% groups did not develop any signs of NPE. Blood pressure after spinal cord compression rose more in the 1.5% group than in the 3% one. In the 1.5% group, the sympathetic ganglionic blockade prevented the neurogenic pulmonary edema development. Animals from the 3% group recovered behaviorally more rapidly than did the animals from the 1.5% group; morphometry and MRI of the lesions showed no differences. Thus, low levels of isoflurane anesthesia promote NPE in rats with a compressed spinal cord and significantly complicates their recovery. The optimal concentration of anesthesia for performing a spinal cord compression lesion is 2.5-3% isoflurane in air.