We evaluated the results of Guglielmi detachable coil (GDC) treatment in partially thrombosed aneurysms and determined if there is high rate of recanalisation on follow-up. Among 149 treated aneurysms in 141 patients, 25 CT- or MR-confirmed partially thrombosed aneurysms were selected for evaluation. The features of thrombosed aneurysms and percentage of occlusion were analysed on initial angiograms. Follow-up angiograms, which were available in 18 cases, were evaluated for aneurysm lumen recanalisation. The recanalisation rate was compared with that of non-thrombosed aneurysms treated with GDCs. Locations of aneurysms were as follows: cavernous carotid ten; ophthalmic four; p-com. two; MCA one; A-com. one; basilar tip four; midbasilar two; PICA one. The size of the aneurysm lumen ranged from 5 to 30 mm (mean 16.8 mm) on angiograms, but on cross sectional images the size of aneurysms ranged from 13 to 70 mm (mean 24.6 mm). The extent of aneurysmal thrombosis ranged from 10 to 90 per cent (mean 46.4 per cent). On initial GDC treatment, total to subtotal occlusion was achieved in 18 cases out of 25 (72%). Of the 18 follow-up angiograms, 14 cases (77.8%) showed recanalisation ranging from 10 to 60 per cent of aneurysm size. Luminal recanalisation was due to migration (10 of 14) or compaction (4 of 14) of coil masses. In two cases, symptoms recurred in association with aneurysm recanalisation, but in no instance was haemorrhage noted. Attempts for retreatment were made in ten cases with success in six. In comparison, 14 (15.9%) out of 88 nonthrombosed cases revealed recanalisation on follow-up angiography. Midterm follow-up angiograms in partially thrombosed aneurysms treated with GDC revealed a fivefold higher rate of recanalisation than in non-thrombosed cases. Close follow-up is necessary in patients with thrombosed aneurysms treated with GDCs.