Background. Intensive care unit (ICU) delirium is a frequent secondary neurological complication in critically ill patients undergoing prolonged mechanical ventilation. Quantitative pupillometry is an emerging modality for the neuromonitoring of primary acute brain injury, but its potential utility in patients at risk of ICU delirium is unknown.
Methods. This was an observational cohort study of medical-surgical ICU patients, without acute or known primary brain injury, who underwent sedation and mechanical ventilation for at least 48 hours. Starting at day 3, automated infrared pupillometry – blinded to ICU caregivers – was used for repeated measurement of the pupillary function, including quantitative pupillary light reflex (q-PLR, expressed as % pupil constriction to a standardized light stimulus) and constriction velocity (CV, mm/sec). The relationship between delirium, using the CAM-ICU score, and quantitative pupillary variables was examined.
Results. A total of 59/100 patients had ICU delirium, diagnosed at a median 8 days from admission. Compared to non-delirious patients, subjects with ICU delirium had lower values of q-PLR (25 [19-31] vs. 20 [15-28] %) and CV (2.5 [1.7-2.8] vs. 1.7 [1.4-2.4] mm/sec) at day 3, and at all additional time-points tested (p<0.05). After adjusting for the SOFA score and the cumulative dose of analgesia and sedation, lower q-PLR was associated with an increased risk of ICU delirium (OR 1.057 [1.007-1.113] at day 3; p=0.03).
Conclusions. Sustained abnormalities of quantitative pupillary variables at the early ICU phase correlate with delirium and precede clinical diagnosis by a median 5 days. These findings suggest a potential utility of quantitative pupillometry in sedated mechanically ventilated ICU patients at high risk of delirium.