High intra-operative oxygen concentration reportedly reduces postoperative nausea and vomiting (PONV), but recent data are conflicting. Therefore, we tested whether the effectiveness of supplemental oxygen depends on the endpoint (nausea vs. vomiting), observation interval (early vs. late) or surgical field (abdominal vs. non-abdominal). We randomly assigned 560 adult patients undergoing various elective procedures with a PONV risk of at least 40% to intra-operative 80% (supplemental) or 30% oxygen (control). Potential confounding factors were similar between groups. Incidences of nausea were similar in the groups during early (12% (supplemental) vs. 10% (control), p = 0.43) and late intervals, 26%vs. 20%, p = 0.09, as were the incidences of vomiting (early: 2%vs. 3%, p = 0.40; late: 8%vs. 9%, p = 0.75). Supplemental oxygen was no more effective at reducing PONV in abdominal (40%vs. 31%, p = 0.37) than in non-abdominal surgery (25%vs. 21%, p = 0.368). Thus, supplemental oxygen was unable to reduce PONV independent of the endpoint, observational period or site of surgery.