Relationship Between Temporal Variability in Pupils and Neurological Prognosis in Out-of-Hospital Cardiac Arrest Patients.

Background Quantitative pupillometry has been widely used for neurological prognostication after cardiac arrest. However, the timing and frequency of pupillary assessment vary considerably across studies, and the clinical significance of temporal fluctuations in pupillary parameters under intensive care conditions remains unclear. This study aimed to evaluate temporal variability in serial quantitative pupillometry during the early post-cardiac arrest period and to explore its relationship with neurological prognostication. Methods This retrospective observational study included adult patients with out-of-hospital cardiac arrest who achieved return of spontaneous circulation (ROSC) and underwent quantitative pupillometry for at least 72 hours at Toho University Omori Medical Center, Tokyo, between October 2018 and March 2020. Pupillary measurements were initiated within three hours after resuscitation and performed every six hours using an automated pupillometer. The Neurological Pupil Index (NPi) and constriction rate (CH) were analyzed bilaterally. Temporal variability was defined using clinically established thresholds for abnormal changes. Neurological outcomes at 90 days were classified according to the Cerebral Performance Category (CPC), with CPC 1-2 considered good and CPC 3-5 poor. Results A total of 27 patients were included in the final analysis; 11 (40.7%) had good neurological outcomes and 16 (59.3%) had poor outcomes. Temporal variability in NPi and CH was observed throughout the 72-hour observation period in all patients, regardless of neurological outcome. Although right NPi at 72 hours showed a statistically significant difference between outcome groups, no consistent or reproducible differences in NPi or CH were observed across time points, between both eyes, or across pupillometric parameters. Persistently absent pupillary responses (NPi = 0) were associated with poor outcomes; however, isolated or fluctuating pupillometric values did not reliably predict neurological prognosis. Conclusions Quantitative pupillometric parameters exhibit substantial temporal variability in patients after cardiac arrest under intensive care conditions. Fixed time-point assessment of pupillometry may therefore be insufficient for reliable neurological prognostication. Serial and dynamic evaluation, particularly the identification of persistently absent pupillary responses, may provide more clinically meaningful information than isolated measurements. These findings should be considered hypothesis-generating.