Unlike expired sevoflurane concentration, propofol lacks a biomarker for its brain effect site concentration (Ce), leading to dosing imprecision particularly in infants. Electroencephalography (EEG) monitoring can serve as a biomarker for propofol Ce, yet proprietary EEG indices are not validated in infants. We evaluated spectral edge frequency (SEF95) as a propofol anesthesia biomarker in infants. We hypothesized that the SEF95 targets will vary for different clinical stimuli and an inverse relationship existed between SEF95 and propofol plasma concentration.


This prospective study enrolled infants (3-12 months) to determine the SEF95 ranges for three clinical endpoints of anesthesia (consciousness-pacifier placement, pain-electrical nerve stimulation, and intubation-laryngoscopy) and correlation between SEF95 and propofol plasma concentration at steady state. Dixon’s Up-Down method was used to determine target SEF95 for each clinical endpoint. Centered isotonic regression determined the dose-response function of SEF95 where 50% and 90% of infants (ED50 and ED90) did not respond to the clinical endpoint. Linear mixed-effect model determined the association of propofol plasma concentration and SEF95.


Of 49 enrolled infants, 44 evaluable (90%) showed distinct SEF95 for endpoints: pacifier (ED50 21.4Hz, ED90 19.3Hz), electrical stimulation (ED50 12.6Hz, ED90 10.4Hz), and laryngoscopy (ED50 8.5Hz, ED90 5.2Hz). From propofol 0.5-6 μg/ml, a 1 Hz SEF95 increase was linearly correlated to a 0.24 (95% CI: 0.19 – 0.29, p<0.001) μg/mL decrease in plasma propofol concentration (marginal R 2 = 0.55).


SEF95 can be a biomarker for propofol anesthesia depth in infants, potentially improving dosing accuracy and utilization of propofol anesthesia in this population.