Number of Affected Relatives, Age, Smoking, and Hypertension Prediction Score for Intracranial Aneurysms in Persons with a Family History for Subarachnoid Hemorrhage

Abstract

Background: Persons with a positive family history of aneurysmal subarachnoid hemorrhage are at increased risk of aneurysmal subarachnoid hemorrhage. Preventive screening for intracranial aneurysms (IAs) in these persons is cost-effective but not very efficient. We aimed to develop and externally validate a model for predicting the probability of an IA at first screening in persons with a positive family history of aneurysmal subarachnoid hemorrhage. Methods: For model development, we studied results from initial screening for IA in 660 prospectively collected persons with ≥2 affected first-degree relatives screened at the University Medical Center Utrecht. For validation, we studied results from 258 prospectively collected persons screened in the University Hospital of Nantes. We assessed potential predictors of IA presence in multivariable logistic regression analysis. Predictive performance was assessed with the C statistic and a calibration plot and corrected for overfitting. Results: IA were present in 79 (12%) persons in the development cohort. Predictors were number of affected relatives, age, smoking, and hypertension (NASH). The NASH score had a C statistic of 0.68 (95% CI, 0.62-0.74) and showed good calibration in the development data. Predicted probabilities of an IA at first screening varied from 5% in persons aged 20 to 30 years with two affected relatives, without hypertension who never smoked, up to 36% in persons aged 60 to 70 years with ≥3 affected relatives, who have hypertension and smoke(d). In the external validation data IA were present in 67 (26%) persons, the model had a C statistic of 0.64 (95% CI, 0.57-0.71) and slightly underestimated IAs risk. Conclusions: For persons with ≥2 affected first-degree relatives, the NASH score improves current predictions and provides risk estimates for an IA at first screening between 5% and 36% based on 4 easily retrievable predictors. With the information such persons can now make a better informed decision about whether or not to undergo preventive screening.