External validation of the SICH-score
Sebina Bulic, MD
van Asch CJJ, Velthuis BK, Greving JP, van Laar PJ, Rinkel GJE, Algra A, et al. External Validation of the Secondary Intracerebral Hemorrhage Score in The Netherlands.Stroke. 2013
Despite considerable research advances, management of ICH remains devastating event with considerable morbidity and mortality. In 2010 Delgado Almandoz at al published a. groundbreaking scoring system; The Secondary Intracerebral Hemorrhage (SICH) score for assessment of patient’s risk of harboring an underlying vascular etiology for ICH. This score has been successfully externally validated and used in the USA.
In this retrospective analysis of the prospectively collected data for 204 non-traumatic ICH patients, admitted to the University Medical Center Utrecht, Netherlands, authors aimed to externally validate SICH score outside the US. All patients had non-contrast CT and angiographic imaging or pathological evaluation.
Univariable logistic regression analysis was performed for each of the separate SICH items, as well as SICH score using original regression coefficient of the multivariable model originally used. Discriminative ability in the validation cohort was modest (which improved with the application of the original regression coefficient). Calibration was found to be modest as well. Non contrast CT scan analysis was found to be strongest predictor for the underlying vascular malformation, thus Non contrast CT analysis was endorsed by the authors rather than SICH score.
Univariable logistic regression analysis was performed for each of the separate SICH items, as well as SICH score using original regression coefficient of the multivariable model originally used. Discriminative ability in the validation cohort was modest (which improved with the application of the original regression coefficient). Calibration was found to be modest as well. Non contrast CT scan analysis was found to be strongest predictor for the underlying vascular malformation, thus Non contrast CT analysis was endorsed by the authors rather than SICH score.
These two cohorts had notable baseline differences: validation cohort had underlying vascular cause in 23.5%, and 22.5% of the patients were aged 71-94 year, in comparison to derivation cohort where 14.6% had underlying vascular cause of ICH and 43.8% were aged 71-94 years. Given these differences, despite the lack of strong correlation between SICH score and existence of underlying vascular abnormality in this study, in my opinion, SICH score remains very important tool for rapid selection of the ICH patients requiring further neurovascular evaluation in the US.
