Background: Recent international recommendations for the diagnosis of gestational diabetes (GD) were implemented in a university hospital. The aim was to audit the appropriateness of use of the new diagnostic approach. Methods: The same 5-month period, one before [2009, traditional two-step oral glucose tolerance test (OGTT) approach, S1] and one after the implementation of new criteria (2010, S2) were compared. Results: In the two periods, 256 (S1) and 245 (S2) pregnant women were examined and 298 (50 g, n=195; 100 g, n=103) and 252 (75 g) OGTTs were, respectively, executed. In S1, 54 (27.7%) 50 g OGTTs resulted positive and 36 (66.7%) of those performed the 100 g OGTT. In addition, three (1.5% of total) 50 g OGTT negative women were submitted to 100 g OGTT. Sixty-three women did 100 g OGTT only. In total, 14 (13.6%) 100 g OGTTs were positive. In S2, 38 (15.1%) 75 g OGTTs were positive. In women who did the complete protocol in the hospital, 98.3% in S1 and 77.0% in S2 performed the correct protocol (p<0.0001). Conclusions: In this hospital new recommendations for GD diagnosis are not correctly applied in 23% of cases. The main issue seems to be the lack of consideration of the new threshold for fasting glycemia (5.1 mmol/L) as a main decisional driver for performing OGTT.