A predictive score for identification of rupture site in subarachnoid haemorrhage with multiple intracranial aneurysm

Abstract

Purpose The rigid method of identifying the rupture site in cases of subarachnoid haemorrhage (SAH) with multiple intracranial aneurysms (MIAs) is still unclear. Here we present a reliable method by using a combination of six predictors. Materials and methods Concerning the 48 SAH patients with MIAs who visited the Showa General Hospital during the period from January 2005 to March 2016, several predictors of rupture site such as the aneurysm-related morphologic features (size, aspect ratio, shape, bleb), aneurysm location, and the distribution of SAH were investigated. Compared with other coexisting aneurysms in each predictor, each aneurysm was categorized into ‘suspicion’ or ‘non-suspicion’, and we analyzed the association between ‘suspicion’ and rupture. Results In the first analysis, all variables were associated with rupture and included in the multivariate logistic regression analysis. The presence of bleb (OR, 20.7; CI, 2.3–186; p = .007) and the aneurysm location (OR, 23.5; CI, 5.2–106; p < .001) were significantly associated with rupture in multivariate logistic regression analysis. Based on the results, a predictive score for rupture was created and calculated for each aneurysm, and the aneurysm with highest predictive score in each patient was categorized into ‘suspicion’. ‘Suspicion’ in the predictive score was significantly associated with rupture (OR, 412.5; CI, 52.2–16384; p < .001). The sensitivity (0.90), specificity (0.98) and the accuracy (0.94) of identifying the rupture site by the predictive score were quite satisfactory. Conclusion Our results suggest that the predictive score may be an excellent parameter to identify the rupture site in cases of SAH with MIAs.