The best available pupil response predicts neurological outcome following cardiac arrest—a prospective cohort study

Abstract Background Early neuroprognostication after cardiac arrest (CA) remains challenging. Manual pupillary light reflex assessments can be inconsistent, leading to interest in automated, quantitative pupillometry. This study used the better Neurological Pupil Index (NPi) and pupillary percentage values from bilateral measurements for neuroprognostication following CA. Methods We evaluated 90 adult survivors of in- and out-of-hospital CA admitted to a medical ICU at a tertiary care university hospital in Berlin. Automated pupillometry was performed every 8 hours for 5 days post-admission using the NeurOptics NPi-100 pupillometer. The better measurement from either eye was selected for analysis. Outcomes at hospital discharge were classified as good (cerebral performance category scale [CPC] 1–2) or poor (CPC 3–5) using the Pittsburgh CPC scale. Results Patients with favorable neurological outcomes consistently showed higher NPi values ( P < 0.001) and greater percentage changes in pupillary diameter ( P < 0.001). At 72 hours, the median NPi was significantly higher in the good outcome group (4.7 [interquartile range (IQR), 4.5–4.8] vs. 4.1 [IQR, 3.7–4.5]; P < 0.05). However, NPi values overlapped between groups, and most patients with unfavorable outcomes still exhibited values within the normal range (NPi > 3.0). A receiver operating characteristic analysis of the pupillary percentage change revealed a threshold of 16% to discriminate between the prespecified outcome groups. Conclusion Higher NPi values and/or greater pupillary diameter changes of the better reading in bilateral measurements were associated with favorable neurological outcomes after CA. However, previously proposed cutoff thresholds could not be confirmed in our cohort.