Association of hyperosmolar therapy with cerebral oxygen extraction after cardiac arrest.

Abstract

Background: Elevated jugular bulb venous oxygen saturation (SjvO2) after cardiac arrest may be due to diffusion-limited oxygen extraction secondary to perivascular edema. Treatment with hyperosmolar solution (HTS) may decrease this edema and thus the barrier to oxygen diffusion. Alternatively, SjvO2 may rise when cerebral metabolic rate declines due to irreversible cellular injury, which would not be affected by HTS. Electroencephalography (EEG) may differentiate between these etiologies of elevated SjvO2. We hypothesized SjvO2 would be lower after treatment with HTS and EEG could identify treatment responders.

Methods: We conducted a retrospective observational cohort study including comatose survivors of cardiac arrest who had at least one elevated SjvO2 (>75%) and were EEG-monitored. We quantified the change in consecutive SjvO2 values within a sample pair using a multivariable mixed-effects regression, treating HTS as a fixed effect, adjusting for mean arterial pressure, partial pressure of arterial oxygen, and partial pressure of carbon dioxide. We classified pretreatment EEG patterns as benign or indicative of potential metabolic failure and tested for an interaction of EEG pattern with HTS.

Results: Our primary adjusted analysis showed an independent association of HTS treatment with change in SjvO2 (β -2.2; 95% confidence interval [CI], -4.0 to -0.3%). In our interaction model, the effect of treatment differed by EEG pattern (β for interaction term -10.9%, 95% CI -17.9 to -3.9%). HTS was associated with a significant change in SjvO2 among those with benign pre-treatment EEG patterns (-12.4%, 95% CI -18.4 to -6.4%) but was not associated with a change in SjvO2 in those with ominous pre-treatment EEG patterns (-1.6%, 95% CI -3.6 to 0.4%).

Conclusions: HTS was independently associated with decreased SjvO2 in patients resuscitated from cardiac arrest, and this effect was limited to patients with benign pretreatment EEG patterns. Our results suggest diffusion-limited oxygen extraction secondary to modifiable perivascular edema as the etiology of elevated SjvO2, and EEG pattern may be useful to identify treatment responders.