To assess whether corrosion of tungsten coils is related to residual shunting and to evaluate whether elevated tungsten serum levels are associated with local or systemic toxicity. Tungsten coils (SPI, Balt, France) were implanted into the subclavian artery of New Zealand white rabbits leading to a residual high-flow shunt in 5/10 rabbits. Serial serum tungsten levels, complete blood count and clinical chemistry were analysed prior to the implantation as well as 15 min, 2 and 4 months thereafter. After 4 months the rabbits underwent repeat angiography before they were sacrificed and the internal organs were evaluated histopathologically. Mean tungsten levels rose from 0.48 microg/l prior to the implantation to 12.4 microg/l 4 months post-implantation. The rise in serum tungsten levels was neither associated with residual shunting present at the time of implantation nor with residual shunting at the time of explantation. One animal had to be sacrificed because of non-resolving palsy of the upper extremity. The remaining animals had an uneventful clinical course with no signs of toxicity of the elevated tungsten levels. Histological examination revealed no evidence of local or systemic toxicity of the tungsten coils. Tungsten coils corrode and lead to a steady increase in serum tungsten levels starting as early as 15 min after implantation. Residual shunting does not seem to influence the kinetics of corrosion of tungsten coils. Despite markedly elevated serum tungsten levels 4 months after implantation degradation of tungsten coils is not associated with local or systemic toxicity.