A prospective, randomized, controlled trial was conducted in 200 consecutive patients undergoing endoscopy of the upper alimentary tract. One hundred patients received supplementary oxygen at 4 liters/minute through nasal cannulae, while 100 patients received no additional oxygen. Within each of these two groups, 50 patients were sedated with midazolam and 50 patients with diazepam suspension ("Diazemuls"). The patients' weights were recorded and correlated with their height to assess whether they were over or under their ideal weight. Oxygen saturation was recorded at baseline and throughout the endoscopic procedure. The principal finding of this study was that hypoxia (oxygen saturation less than 93%) was prevented in all cases by the administration of 4 liters/minute of oxygen, whereas 48 of the 100 patients who did not receive oxygen exhibited falls in oxygen saturation to less than 93% (p < 0.0001). Those with the highest risk were the obese patients (p < 0.01). There was no significant difference between the two sedative drug groups in either frequency or severity of associated hypoxia (p = 0.77, patients not given oxygen; p = 0.31, patients receiving oxygen). The cost of administering oxygen during upper gastrointestinal endoscopy would be an average of 95 pence ($1.60) per patient. In conclusion, the administration of oxygen during endoscopy is a worthwhile prophylactic measure to prevent hypoxia and its potential adverse effects.