To analyse radiation related complications and secondary enucleation after irradiation of malignant uveal melanoma with ruthenium-106 plaques. A series of 100 consecutive eyes irradiated in 1981-91 was analysed using the life table method and the Cox proportional hazards model. The median apical and scleral tumour dose was 100 Gy (range 15-200 Gy) and 1000 Gy (range 200-1200 Gy), respectively. The median follow up time was 2.8 and 2.0 years (range 1 month to 10 years) for anterior and posterior segment complications, respectively. The 3 and 5 year probabilities of being without radiation cataract were 73% and 63%, without neovascular glaucoma 91% and 81%, without vitreous haemorrhage 83% and 74%, without radiation maculopathy 85% and 70%, and without radiation optic neuropathy 90% and 88%, respectively. The risk of radiation cataract was highest with large tumour size (T1 + T2 v T3, p = 0.0027; height < or = 5 v > 5 mm, p = 0.029; largest basal diameter (LBD) < or = 15 v > 15 mm, p < 0.0001) and location of anterior tumour margin anterior v posterior to the equator (p = 0.0003); the risk of neovascular glaucoma with large size (T1 + T2 v T3, p = 0.039; LBD < or = 15 mm v 15 mm, p = 0.021); and the risk of maculopathy and optic neuropathy with proximity of the posterior tumour margin to the fovea and the optic disc (< or = 1.5 v > 1.5 mm; p = 0.030 and p = 0.0004, respectively). In Cox's multivariate analysis the strongest risk indicator for radiation cataract (RR 1.5, 95% CI 1.4-1.6) and vitreous haemorrhage (RR 1.6, 95% CI 1.4-1.8) was the height of the tumour; for neovascular glaucoma the TNM class (RR 6.2, 95% CI 2.7-13.8); for radiation maculopathy location of posterior tumour margin within 2 mm from the fovea (RR 3.4, 95% CI 2.0-6.0); and for radiation optic neuropathy location of tumour margin within 1 DD of the optic disc (RR 6.1, 95% CI 3.0-12.4). The 3 and 5 year probabilities of avoiding enucleation were 92% and 85%, respectively. Ten eyes were enucleated--six because of recurrent tumour growth, three because of treatment complications, and one because of mistakenly suspected extraocular growth. The results suggest that the frequency of radiation related complications after ruthenium brachytherapy of uveal melanoma is acceptable, in particular as regard irradiation of small and medium sized tumours for which ruthenium therapy generally is recommended.