Feasibility and Variability of Automated Pupillometry Among Stroke Patients and Healthy Participants: Potential Implications for Clinical Practice

Abstract

Background: Early neurological deterioration (END) is common after stroke and represents a poor prognostic marker. Manual pupillary assessment to detect END is subjective and has poor interrater reliability. Novel methods of automated pupillometry may be more reliable and accurate. This study aims to evaluate the acceptability and feasibility of automated pupillometry in patients with acute stroke and healthy volunteers and compare its interrater reliability with that of the traditional manual method. Methods: Automated and manual pupillary assessments were recorded between 2 independent observers alongside routine neurological observations from 12 acute stroke patients at a high risk of END. The proportion of completed measurements, adverse events, and qualitative feedback from patients and staff nurses was used to assess acceptability and feasibility of automated pupillometry. Paired automated and manual assessments were supplemented with measures from healthy volunteers to analyze measures of variability and agreement. Results: Automated pupillometry was acceptable and safe among 12 acute stroke patients, but feasibility criteria were not attained. Interrater agreement for automated pupillometry was superior to manual assessment for measurements of pupil size, anisocoria, and pupillary light reactivity, for both patients and healthy volunteers. Substantial disparity existed in agreement between automated and manual assessments of these parameters. Conclusions: Automated pupillometry represents an alternative to manual pupillary assessment that may have greater interrater agreement and reliability. As an optimized method of neurological assessment, it has the potential to improve detection and treatment of conditions leading to END after stroke.