OBJECTIVE Women who develop diabetes during pregnancy are at significant risk of developing Type 2 diabetes later in life, and "should be offered blood glucose testing ... to exclude persisting hyperglycaemia" after delivery. New UK guidelines now recommend that this is done by measurement of fasting blood glucose, rather than an oral glucose tolerance test (OGTT). Our study aims to establish whether assessment by fasting blood glucose gives equivalent results to an OGTT. METHODS Retrospective data were obtained from four centres in the West Midlands, UK, who performed a postnatal OGTT on 470 women during a 12-month period. All centres used a 75 g, 2-h protocol. RESULTS 28 of 470 (6.0%) women had diabetes mellitus after testing of fasting plasma glucose. After OGTT, 70 women (14.9%) were found to have impaired glucose tolerance (IGT), and 10 further cases of diabetes were detected (overall prevalence of DM 8.1%). Of the 400 women with normal fasting plasma glucose. After OGTT, 70 women (14.9%) were found to have impaired glucose tolerance (IGT), and 10 further cases of diabetes were detected (overall prevalence of DM 8.1%). Of the 400 women with normal fasting plasma glucose (< or = 6.0 mmol/l), 55 were found to have impaired glucose tolerance and 4 had overt diabetes. In all, 26% of women with diabetes were missed by FPG. CONCLUSIONS Fasting plasma glucose alone does not allow detection of IGT, and misses a substantial proportion of women with Type 2 diabetes. Recognition of ongoing IGT or DM is important in identifying women who may benefit from lifestyle modification, and who may need early intervention in subsequent pregnancies. Our results suggest that OGTT should be the investigation of choice in the postnatal assessment of women who have had GDM.