Results:
The Flow diagram is shown in Fig. 2 . Three patients refused to intervention management after entering the operating room, four patients were excluded for surgery method changed, and two were lost to follow-up. Finally, enrolled 78 patients in the study: 25 patients in high concentration group, 24 patients in middle concentration group, 28 patients in low concentration group.
No differences in age, BMI, HR and MAP among the three groups (Table 1 ).
3.1 Primary endpoint
Concentration-change data for each patient received by the up-and-down treatment were performed in Fig. 3 (A, B, C). According to up-and-down method and probit analysis, REM EC50 1.82 ng/mL (95% CI 0.94 to 2.27 ng/mL) in high group, 2.96 ng/mL (95% CI 2.75 to 3.14 ng/mL0) in middle group, 5.14 ng/mL (95% CI 4.86 to 5.47 ng/mL) in low group (Fig. 3 D).
There was also no significant difference in MAP, HR, PSI and BIS related to surgery among the three groups (Fig. 4 ), but compared with middle and low group, the change of MAP and HR of high group were more obvious at T4.
Operation duration, hospital day, VAS score in the first day after sugery did not make imparity among three groups, however, the dose of REM made statistical significance (P<0.05). Compared with other two group,the incidence of adverse events, the extubation and lengths of PACU stay in middle group was shortest (Table. 2 ). All three groups denied intraoperative awareness and hyperpathia, no postoperative delirium occurred.
3.2 Secondary endpoints
The analgesic response of REM can be obtained by DSA (Figure. 5 ), EEG changes caused by different concentrations of REM were analyzed. It can be seen that the slow delta wave (0.5-4HZ) and alpha wave (8-12HZ) in a is more obvious than b in every part of the figure.
The area under the ROC curve (Fig. 6 ) for the PSI and BIS after incision were 0.953 ± 0.02 and 0.888 ± 0.04, respectively, it was confirmed that the two indexes could predict the consciousness state of patients, and PSI was more sensitive than BIS.