Detection of delayed cerebral ischemia using objective pupillometry in patients with aneurysmal subarachnoid hemorrhage.

Abstract

Cerebral vasospasm causing delayed cerebral ischemia (DCI) is a source of significant morbidity after subarachnoid hemorrhage (SAH). Transcranial Doppler is used at most institutions to detect sonographic vasospasm but has poor positive predictive value for DCI. Automated assessment of the pupillary light reflex has been increasingly used as a reliable way of assessing pupillary reactivity, and the Neurological Pupil Index (NPi) has been shown to decrease hours prior to the clinical manifestation of ischemic injury or herniation syndromes. The aim of this study was to investigate the role of automated pupillometry in the setting of SAH, as a potential adjunct to TCD. Our analysis included patients that had been diagnosed with aneurysmal SAH and admitted to the neuro-intensive care unit of the University of Texas Southwestern Medical Center between November 2015 and June 2017. A dynamic infrared pupillometer was used for all pupillary measurements. An NPi value ranging from 3 to 5 was considered normal, and from 0 to 2.9 abnormal. Sonographic vasospasm was defined as middle cerebral artery velocities greater than 100 cm/sec with a Lindegaard ratio greater than 3 on either side on transcranial Doppler. Most patients had multiple NPi readings daily and we retained the lowest value for our analysis. We aimed to study the association between DCI and sonographic vasospasm, and DCI and NPi readings. A total of 56 patients were included in the final analysis with 635 paired observations of daily TCD and NPi data. There was no statistically significant association between the NPi value and the presence of sonographic vasospasm. There was a significant association between DCI and sonographic vasospasm, χ2(1) = 6.4112, p = 0.0113, OR 1.6419 (95% CI 1.1163-2.4150), and between DCI and an abnormal decrease in NPi, χ2(1) = 38.4456, p < 0.001, OR 3.3930 (95% CI 2.2789-5.0517). Twelve patients experienced DCI, with 7 showing a decrease of their NPi to an abnormal range. This change occurred > 8 hours prior to the clinical decline 71.4% of the time. The NPi normalized in all patients after treatment of their vasospasm. Isolated sonographic vasospasm does not seem to correlate with NPi changes, as the latter likely reflects an ischemic neurological injury. NPi changes are strongly associated with the advent of DCI and could be an early herald of clinical deterioration.