The bispectral index as a predictor of anterograde amnesia caused by premedication with intranasal midazolam

To determine the ability of the bispectral index (BIS) to predict anterograde amnesia for events occurring between the arrival of patients in the operating theater and anesthetic induction, following premedication with 10 mg of intranasal midazolam. We enrolled patients scheduled for any type of surgery under general anesthesia. Patients were first monitored using the BIS and administered 10 mg of intranasal midazolam. The BIS was monitored until it began to fall steadily and remained below a value of 90; the elapsed time was recorded and the patient was taken to the theater. All patients underwent 6 routine maneuvers before induction. After surgery, patients were asked about their memory of the maneuvers performed. Sensitivity, specificity, and predictive values were studied to obtain the BIS value that provided the best combination of sensitivity and specificity, and a receiver operating characteristic (ROC) curve was drawn. We enrolled 55 patients. The mean (SD) time taken for the BIS value to fall was 5.93 (2.93) minutes. The sensitivity of the test was 0.96 and specificity was 0.60. The positive predictive value was 0.91 and the negative predictive value was 0.75. The test classified 89.1% of the patients correctly. The ROC curve showed the best cutoff to be 83. A fall in the BIS to below 90 can be used as a predictor for anterograde amnesia following administration of 10 mg of intranasal midazolam.