The use of intranasal (IN) midazolam in adults for sedation in upper gastrointestinal endoscopy has been evaluated in a controlled clinical study. Eighty-one patients with a mean age of 37.02 +/- 12.50 years who underwent upper gastrointestinal endoscopy for various reasons were included in the study. Three groups were formed according to the sedation regimen. In the first group (n = 30), patients received IN midazolam. In the second group (n = 28) intravenous (IV) midazolam was given for sedation, and the third group of patients (n = 23) received placebo before the procedures. Patients were monitored (using a pulse oximeter with an interval of 5 minutes until the 45th minute after the procedure) for arterial oxygen saturation, heart rate, systolic and diastolic arterial blood pressure, and respiratory rate. Efficacy of sedation, amnesia, side effects, and patients' preferences were evaluated. Superior results regarding the efficacy of sedation has been documented with the use of IV midazolam (p < 0.001), and this was the preferred route for drug application according to the patients' answers (p < 0.01). However, regarding amnesia, IN midazolam was found to be almost equally effective as IV midazolam (p < 0.05); moreover, IN route of drug application caused significantly fewer side effects than did the IV form (p < 0.001 ). Intranasal application of midazolam for gastrointestinal endoscopy appeared to be an interesting alternative to the IV route, the usage of which might be limited because of its potentially serious side effects. In contrast to the IV application of midazolam, the IN route may not even necessitate the monitoring of the patient during upper gastrointestinal endoscopy.