OBJECTIVE To determine whether there is a relationship between plasma glucose level in a glucose screening test and the occurrence of pre-eclampsia in non-diabetic pregnant women. METHODS All pregnant non-diabetic women attending the antenatal clinic at the Aga Khan University Medical Center were screened with a 75 g-2 h glucose challenge test (GCT). From 1988-90, a data base of 1316 cases was compiled for use in this study. Of these, 67 had to be excluded because of one or more abnormal levels in the 75 g-3 h oral glucose tolerance test (GTT) that required treatment to maintain euglycemia. Among the remaining patients (n = 1249) who had no evidence of glucose intolerance and were included in the analysis, there were 42 cases of pre-eclampsia. The association between pre-eclampsia and plasma glucose level in the GCT was evaluated using logistic regression analysis that adjusted for effects of age and gravidity. RESULTS The odds of having pre-eclampsia were increased by 20% (95% confidence interval 0%-44%) per mmol/l rise in plasma glucose level in the GCT. The same statistic for age was 9% (95% confidence interval 2%-17%), and for primigravidity it was 210% (95% confidence interval 55%-517%). There was no significant interaction between these variables. CONCLUSIONS Minor degrees of glucose intolerance, age and primigravidity are associated with a higher occurrence of pre-eclampsia in non-diabetic pregnant women.