Universal screening of gestational diabetes (GDM) has been advocated but is seldom practised in the United Kingdom. We report 12 months' audit of a comprehensive screening programme for women attending a teaching hospital for antenatal care. The population served is almost exclusively Caucasian. Of 4,016 women who booked in for delivery at the hospital, 3,3 16 (83%) completed the screening protocol. Sixty-seven women (1.7%) were found to have gestational diabetes after a 75 gram oral glucose tolerance test (OG n) by WHO criteria. Of these, 43 were identified by following the screening protocol (one at booking following a random venous plasma glucose result of >7.0 mmoUL, 42 after a third trimester 50 gram glucose challenge) and 24 as a result of OGTTs performed on clinical grounds. Clinical suspicion was usually raised by the identification of clinical risk factors for diabetes. Seventy OGTTs were requested despite a normal third trimester 50 gram challenge result and generated eight abnormal results, although these were at a latergestation than those generated by abnormal 50 gram results. Fifty two women (78%) were referred to the combined diabetes-antenatal clinic. Only eight women (12%) received insulin. Thirty five women (52%) returned for postnatal O G n , of which four were abnormal. We have confirmed the low prevalence of GDM in a predominantly Caucasian population and would question the value of first trimester screening in such a population. The screening protocol is only valuable if abnormal results are acted upon and, while there is disagreement among clinicians regarding the value of treating GDM, not all women with abnormal results receive dietary advice. Practical Diabetes Int 1998; 15(2): 45-48