BACKGROUND : Pupillary dysfunction is recognized as a sign of acute neurological deterioration due to worsening mass effect in patients with hemispheric strokes. Recent neuroimaging studies suggest that horizontal displacement of brain structures may be more important than vertical displacement in explaining these pupillary findings. Pupillometers allow objective and standardized evaluation of the pupillary light reflex. We hypothesized that pupillary data (Neurological Pupil index [NPi] and constriction velocity [CV]) obtained with a hand-held pupilometer, correlate with horizontal intracranial midline shift in patients with ischemic and hemorrhagic strokes. METHODS : The ENDPANIC registry is a prospective database of pupillometer readings in neurological patients. There were 134 patients in the database with an acute ischemic stroke or intracerebral hemorrhage who had at least 2 neurologic imaging studies (CT or MRI) and pupillometer assessments performed within 6 hours of the imaging. Horizontal shift of the septum pellucidum (SPS) was measured in 293 images. We computed the correlation between SPS and the following pupillary variables: size, NPi, CV (left, right, and left-right difference), followed by a regression model to control for confounders. RESULTS : There were 94 patients (70.1%) with an ischemic stroke and 40 patients (29.9%) had an intracerebral hemorrhage. After controlling for age, race, and gender, there was a significant correlation between the SPS and NPi (left [P < .001], right [P < .001]), CV (left [P < .005], right [P < .001]) pupillary asymmetry (absolute difference between right and left; P < .05), but not between SPS and pupillary size (left or right). There was a significant correlation between the NPi and CV for the right pupil when there was a right-to-left SPS (P < .001 and P < .05, respectively), but none between the NPi and CV for the left pupil and left-to-right SPS. CONCLUSIONS : In patients with ischemic and hemorrhagic strokes, there is a significant correlation between SPS and the NPi, CV and pupillary asymmetry, but not with pupillary size.