Objective: There is little information on the advantages of nonglucose carbohydrates in total parenteral nutrition (TPN) for diabetic patients. The aim of this study is to evaluate glycemic control and insulin requirements in diabetic patients who received TPN with different sources of carbohydrates, and to determine whether insulin requirements are different when septic and non-septic diabetic patients are studied. Materials and Methods: One-hundred and thirty-eight patients were randomly divided into two groups receiving either glucose (G), n ¼ 71, or glucose-fructose-xylitol 2 : 1 : 1 (GFX), n ¼ 67. There were no differences between the demographic or anthropometric characteristics of the groups, nor between the patients with diabetes mellitus type 1 and type 2, nor the initial TPN composition. Acceptable glycemic control was considered when glycemia reached < 200 mg=dl. Results: Glycemic control was attained in 79.7% of patients (74.6 vs 85.1%), in the same period of treatment. At the end of treatment, insulin requirements were not different (45 AE 19 vs 45 AE 26 UI=day) in both groups, while similar amounts of carbohydrates (191 AE 36 vs 187 AE 45 g=day) were infused. The ratio insulin=body weight and insulin=carbohydrates were equal in both groups. In the GFX group nonseptic and septic patients needed less and more insulin, respectively, than their counterparts in the G group. No major adverse events related to carbohydrate infusions were observed. Conclusions: Either G or GFX could be used in TPN for diabetic patients, providing glycemic control in most cases with similar insulin requirements. GFX mixtures were slightly more beneficial to attain glycemic control in nonseptic patients, but septic diabetic patients had higher insulin needs in this group.