PURPOSE: To search for risk factors of developing irreversible cranial ischemic complications (ICIC) in patients with giant cell arteritis (GCA) and to explore whether two subsets of patients (high risk and low risk of developing ICIC) can be defined. METHODS: One-hundred seventy- eight consecutive patients with temporal arteritis (149 biopsy-proven) were diagnosed and followed up in a department of Internal Medicine between 1976 and 1999. The patients were separated into two groups, according to the presence or absence of ICIC, with comparison of 17 clinical and biological parameters prospectively recorded for each patient using a pre-established comprehensive questionnaire. RESULTS: ICIC occurred in 25 patients (14%), with amaurosis in 22 cases. Suggestive symptoms and/or signs of temporal arteritis were present in 92% of the patients, lasting 50 days (median) before the onset of ICIC. Forty-three patients (24%) complained of transient visual ischemic symptoms (TVIS), which preceded acute blindness in 11 cases. A multivariate logistic regression, from which 28 cases with upper limb artery involvement were excluded for technical reasons (no CCII in any case, thus predicting perfectly the lack of ischemic risk, P = 0.02), indicated that the only independent variables associated with the ischemic risk were: a history of TVIS (P = 0.05), the lack of signs of polymyalgia rheumatica (PMR; P = 0.02), lower blood levels of fibrinogen (P = 0.024) and higher mean blood platelets levels (P = 0.006). However, these five variables predicted only 30% of the variability of the model. Sensitivity, specificity, positive and negative predictive values of the model reached respectively 36, 96, 64 and 88%. Overall, 86% of the cases were correctly classified with respect to the ischemic risk. CONCLUSION: The rate of ICIC should be reduced by an earlier recognition of the usual signs of temporal arteritis. Several independent risk factors of ICIC have been identified. However, the logistic model failed to predict accurately the ischemic risk in 14% of the cases, indicating that as yet unrecognised factors probably exist that play a role in the occurrence of ICIC. Nevertheless, regarding the strong association between platelet levels and ICIC, patients with thrombocytosis should receive initially both corticosteroids and antiplatelet agents.