When less is more: Non-contrast head CT alone to work-up hypertensive intracerebral hemorrhage

Abstract

Hypertension is a common cause of intracerebral hemorrhage (ICH). The work up typically involves neuroimaging of the brain and blood vessels to determine etiology. However, extensive testing may be unnecessary for presumed hypertensive hemorrhages, and instead prolong hospital stay and increase costs. This study evaluates the predictive utility of hemorrhage location on the non-contrast head CT in determining hypertensive ICH. Patients presenting with non-traumatic ICH between March 2014 and June 2019 were prospectively enrolled. Hemorrhage etiology was determined based on previously defined criteria. Chi square and Student’s t tests were used to determine the association between patient demographics, ICH severity, neuroimaging characteristics, and medical variables, with hypertensive etiology. Multivariable regression models and an ROC analysis determined utility of CT to accurately diagnose hypertensive ICH. Data on 380 patients with ICH were collected; 42% were determined to be hypertensive. Along with deep location on CT, black race, history of hypertension, renal disease, left ventricular hypertrophy, and higher admission blood pressure were significantly associated with hypertensive etiology, while atrial fibrillation and anticoagulation were associated with non-hypertensive etiologies. Deep location alone resulted in an area under the curve of 0.726. When history of hypertension was added, this improved to 0.771. Additional variables did not further improve the model’s predictability. Hypertensive ICH is associated with several predictive factors. Using deep location and history of hypertension alone correctly identifies the majority of hypertensive ICH without additional work-up. This model may result in more efficient diagnostic testing without sacrificing patient care.