Ten non-diabetic surgical patients, submitted to parenteral hyperalimentation, were prospectively studied, in order to determine the incidence of glucose intolerance during this treatment. Additionally three methods for the control of glucose metabolism, namely glucosuria and gtycemia, given by reagent strips, and standard blood sugar, determined at the auto-analyser, were compared, with the purpose of establishing which of these tests is best suited for routine utilization. It was concluded that: 1) The incidence of glucose intolerance, defined as one or more episodes of positive glucosuria, or of blood sugar higher than 180mg/100 ml, was 60% in the first day of intravenous nutrition, and 80% if the whole period of treatment; 2) Glucose intolerance was more often detected by urine testing than by blood determinations, either with Dextrositx or at the a auto-analyser as the glucosuria test: was performed much more frequently; 3) It is concluded that although the findings of tests for glucosuria may not adequately reflect glucose metabolism in all circumstances, they represent the choice examination for basic monitoring of patients receiving intravenous hyperalimentation, as soon as renal function is satisfactory, because of its low expense and its easy application.