Background: Diabetes mellitus (DM) is a metabolic disorder and the 5th cause of death in developed countries. Gestational diabetes mellitus (GDM) is an impaired glucose tolerance in pregnancy period. Fourteen-eighteen percent of pregnant women are at risk of GDM, 3-4 percent of whom suffer from it. The present study was conducted to determine the most important causes of DM in a regression model.

Materials and Methods : In a case control study 70 GDM patients were compared with 350 non-GDM pregnant women referring to Shiraz hospital for delivery. The data were gathered and analyzed using Fisher exact test, in univariate t-test method and binary logistic regression at multivariate model.

Results : The mean age in cases and controls was 31.2 ± 6.03 and 26.32 ± 5.03, respectively. All of variables showed significant difference between two groups, except history of infertility malformation of twine pregnancy, toxemia and previous preterm delivery having no relation with GDM. In final regression model, GD history (OR=21.93), glocosuria in three trimester (OR=10.34), age older 35 years (OR=9.57), family history of GD (OR=8.13) and preeclampsia (OR=3.81) were the most important predictors of GDM, respectively.

Conclusion : To prevent diabetes and its unpleasant outcomes in pregnant women it is suggested that they should avoid creating factors of glocosuria, especially in older mothers with family history of GD and GDM history.