La. Kearse et al., BISPECTRAL ANALYSIS OF THE ELECTROENCEPHALOGRAM PREDICTS CONSCIOUS PROCESSING OF INFORMATION DURING PROPOFOL SEDATION AND HYPNOSIS, Anesthesiology, 88(1), 1998, pp. 25-34
Background: The bispectral index (BLS) measures changes in the interfr
equency coupling of the electroencephalogram (EEG). The purposes of th
is study mere (1) to determine whether BIS correlates with responses t
o command during sedation and hypnosis induced by propofol or propofol
and nitrous oxide, and (2) to compare BIS to targeted and measured co
ncentrations of propofol in predicting participants' responses to comm
ands. Methods: Twenty volunteers (15 men and 5 women, aged 22-50 yr) w
ere given propofol by computer-controlled infusion, and EEG was record
ed for off-line analysis of BIS. Responses to randomly ordered verbal
commands or voice plus touch were measured with two categorical scales
(CS1 and CS2, respectively). All subjects received a propofol infusio
n targeted to achieve effect site concentrations of 1, 2, 4, 2, 1, and
0 mu g/ml. Ten participants had repeated infusion, whereas 10 others
breathed 30% nitrous oxide and oxygen and received a propofol infusion
targeted for 0.5, 1, 2, 4, 2, 1, 0.5, and 0 mu g/ml. Five minutes aft
er each targeted concentration had been reached, CS1, CS2, and arteria
l propofol concentration were determined. The area under the receiver
operating characteristic curve was used to compare the accuracy of (1)
BIS, (2) targeted propofol concentration, (3) measured concentration,
and (4) treatment history as predictors of response. Results: Bispect
ral index was a strong predictor of CS1 and CS2 (P < 0.0001) and signi
ficantly more accurate than targeted or measured propofol concentratio
ns (P < 0.0003 and P < 0.003, respectively), It also provided addition
al predictive power when combined with treatment history (P < 0.02). N
itrous oxide slightly decreased the probability of response at a given
value of BIS (P < 0.05), but accuracy was unaffected. Conclusions: Bi
spectral index accurately predicts response to verbal commands during
sedation and hypnosis with propofol or propofol plus nitrous oxide. Ac
curacy is maintained in situations likely to be encountered during cli
nical use: when propofol concentrations are increasing or decreasing a
nd when repeated measurements are made over time.