Assessment of sedation by automated pupillometry in critically ill patients: a prospective observational study.

ABSTRACT

BACKGROUND: Quantitative measurement of pupil change has not been assessed against the Richmond Agitation and Sedation Scale (RASS) and spectral edge frequency (SEF) during sedation. The aim of this study was to evaluate pupillometry against these measures in sedated critically ill adult patients. METHODS: In ventilated and sedated patients, pupillary variables were measured by automated pupillometry at each RASS level from -5 to 0 after discontinuation of hypnotics, while processed electroencephalogram variables were displayed continuously and SEF was recorded at each RASS level. Correlations were made between percentage pupillary light reflex (%PLR) and RASS, and between %PLR and SEF. The ability of %PLR to differentiate light sedation (RASS ≥-2), moderate (RASS =-3), and deep sedation (RASS ≤-4) was assessed by areas under receiver operating characteristic (ROC) curves. RESULTS: A total of 163 paired measurements were recorded in 38 patients. With decreasing sedation depth, median %PLR increased progressively from 20% (interquartile range 17-25%) to 36% (interquartile range 33-40%) (P<0.001). Strong correlations were found between %PLR and RASS (Rho=0.635) and between %PLR and SEF (R=0.641). Area under the curve (AUC) of 0.87 with a %PLR threshold of 28% differentiated moderate/light sedation from deep sedation with sensitivity of 83% and specificity of 83%. An AUC of 0.82 with a threshold of 31% distinguished light sedation from moderate/deep sedation with a sensitivity of 81% and a specificity of 75%. CONCLUSIONS: Quantitative assessment of %PLR correlates with other indicators of sedation depth in critically ill patients.