BACKGROUND We evaluated whether preoperative carbohydrate drink would be able to decrease postoperative nausea and vomiting (PONV) and improve the quality of recovery (QoR) in diabetic patients undergoing total knee arthroplasty (TKA). METHODS Eighty-two patients were randomized to either the intravenous (IV) Dextrose group (n = 41) or Oral carbohydrate (CHO) group (n = 41). The IV Dextrose group received dextrose solution mixed with insulin while fasting, and the Oral CHO group received carbohydrate drinks preoperatively. PONV was assessed up to postoperative 36 hours, and QoR was assessed before surgery and on postoperative day (POD) 1. Blood glucose was measured from the morning of surgery until POD 1. RESULTS PONV scores were not different between the groups. Postoperative QoR scores were significantly higher in the Oral CHO group (median [interquartile range]; 160 [153 to 167]) than the IV Dextrose group (155 [147 to 159]) (P = 0.009), but the difference did not meet the minimal clinically important difference. Blood glucose was comparable between the groups. CONCLUSIONS Preoperative CHO drink did not reduce PONV in diabetic patients after total knee arthroplasty. A statistically significant but clinically questionable improvement in the QoR was seen in the Oral CHO group. However, preoperative CHO drink did not increase hyperglycemia, which suggests that it may be a safe component of perioperative care in diabetic patients.