Contrast Agent Induced Obliteration of a Spinal Arteriovenous Malformation

Discussion

Spontaneous obliteration of central nervous system AVMs has occasionally been described.^6–8 Most of these cases were cerebral AVMs that thrombosed after symptomatic intracerebral or subarachnoid hemorrhage. This spontaneous thrombosis of a cerebral AVM may be explained by reduction of blood flow as a result of compression on the proximal draining veins by the hematoma or edema, feeding artery occlusion through vasospasm, thromboembolism, or systemic coagulation disorders.⁹

Spontaneous obliteration of spinal AVMs has been described in only 3 cases. One spinal AVM at the T11–12 level thrombosed after hemorrhage. However, in this case, a previous angiogram was not available and surgical exploration was not performed.² The second case involved complete thrombosis of a perimedullary spinal cord AVM at the T11-L1 level. The thrombosis probably occurred soon after diagnostic angiography and was associated with neurologic deterioration. MR imaging and angiography performed 1 week later revealed complete occlusion of the lesion. Thrombosis of the dominant draining vein triggered by contrast injection was believed to be the most probable cause of the observed phenomenon.⁹ The third case was an angiographically confirmed spinal AVM at the T12-L1 level that was left untreated. The patient returned 3 years later with sudden paraparesis and bladder retention. Repeat angiography demonstrated complete thrombosis of the AVM, also at follow-up. The patient did not recover completely. It was thought that thrombosis of the draining veins, triggered by the interaction of genetic and hemodynamic factors, was the most probable cause.¹

Obliteration during angiography with confirmation 3 months later, as in our case, has not been described previously, to our knowledge. The mechanism that led to thrombosis of the AVM remains unclear. Possible mechanical explanations are thromboembolism, dissection of the vessel wall, and vessel spasm. However, in our case, dissection of the vessel wall or vasospasm did not occur, and there were no thrombi visualized on the angiogram. Nonionic contrast agents, as used in our patient, can induce coagulation via activation of platelets, aggregation of red blood cells, changes in the endothelium, and changes of the osmolality and ionic strength.^3,10 We believe this toxic effect of the contrast agent is the most likely explanation for the acute thrombosis. The patient's symptoms resolved soon after obliteration of the AVM, unlike in the 2 previously documented cases. Apparently, the thrombosis of the AVM remained restricted to the nidus itself with unaffected patency of the anterior spinal axis.

It is suggested that spontaneously thrombosed AVMs may recanalize and that follow-up angiographic evaluation is recommended, preferably after 2 months or later.⁷ We performed follow-up angiography at 3 and 6 months, which showed persistent complete occlusion of the AVM by both MR and conventional angiography.