Women with impaired glucose tolerance are at high risk of developing noninsulin dependent diabetes mellitus (NIDDM). The Mercy Hospital for Women has a long-term follow-up programme for women with gestational diabetes, which identifies many women with impaired glucose tolerance. Two hundred of these women were entered into a randomized controlled trial of intensive versus routine dietary advice. Seven women were lost to follow-up. The annual incidence rates of diabetes mellitus for the 2 groups were 6.1% (intervention) and 7.3% (control), an incident rate ratio of 0.83, 95% confidence interval 0.47-1.48, p = 0.50. Overall, there was a return to normal glucose tolerance in 44% of patients. Multivariate analysis showed that body mass index, fasting and 2-hour plasma glucose levels at trial entry were significantly associated with an increased risk of diabetes mellitus. Impaired glucose tolerance is an important condition that should be treated with advice about lifestyle modification (diet and/or exercise). We consider that future trials in the management of women with previous gestational diabetes who have impaired glucose tolerance should investigate the effect of pharmacological intervention in addition to diet and/or exercise, the latter providing a therapy that it would be unethical to exclude on the evidence presently available.