Abstract Background: Propofol may result in hypotension, bradycardia, and loss of protective reflexes, especially in elderly patients, while esketamine, a N-methyl-D-aspartate receptor antagonists, has analgesic, anaesthetic and sympathomimetic properties and is known to cause less cardiorespiratory depression. We hypothesized that esketamine may reduce the median effective concentration (EC50) of propofol and cause more stable haemodynamic responses during gastrointestinal endoscopy in elderly patients. Methods: Ninety elderly patients, aged 65-89 years, undergoing gastrointestinal endoscopy were randomly assigned into three groups: SK0.25 group (0.25 mg/kg esketamine), SK0.5 group (0.5 mg/kg esketamine) and saline control group. Anaesthesia was achieved by target-controlled infusion of propofol with an initial plasma concentration of 2.5 μg/ml with different bolus doses of esketamine during gastrointestinal endoscopy. The EC50 of propofol for gastrointestinal endoscopy was determined by using an up-and-down method of Dixon with an adjacent concentration gradient at 0.5μg/mL to prohibit purposeful movements. Cardiovascular parameters were also measured and recorded. Results: Propofol EC50 and its 95% confidence interval for gastrointestinal endoscopy in elderly patients were 1.71 (1.15-2.27) μg/mL in SK0.5 group, 2.45 (1.85-3.05) μg/mL in SK0.25 group and 3.69 (2.59-4.78) μg/mL in control group respectively (P < 0.05). The average percent change to baseline mean arterial pressure (MBP) was -19.7 (7.55), -15.2 (7.14) and -10.1 (6.73) with P＜0.001, in the control group, the SK0.25 group and the SK0.5 group, respectively. Conclusions: Combination medication of propofol with esketamine reduced the propofol EC50 during gastrointestinal endoscopy in elderly patients and caused more stable haemodynamic responses compared with single administration of propofol.