Table 3—

Crude and adjusted ORs for independent variables tested in multiple logistic regression with the dependent variable overt diabetes at follow-up in 1990 for the old cohort and in 2002 for the new cohort

Crude OR (95% CI)Adjusted OR (95% CI)*
n (diabetes/total)164/556142/506
New cohort 2002 vs. old cohort 19903.1 (2.1–4.7)3.1 (1.9–5.3)
Prepregnancy BMI <25 kg/m21.01.0
Prepregnancy BMI ≥25 and <30 kg/m22.4 (1.5–3.8)2.0 (1.1–3.4)§
Prepregnancy BMI ≥30 kg/m23.2 (2.0–5.0)2.6 (1.5–4.5)§
GDM diagnosis before gestational age of 24 weeks3.7 (2.1–6.3)2.3 (1.2–4.5)
Fasting plasma glucose at diagnosis of GDM (mmol/l)2.1 (1.3–3.2)§2.5 (1.5–4.2)
IGT 2 months postpartum3.9 (2.5–6.0)5.8 (3.3–9.9)
Preterm delivery (gestational age <37 weeks)2.2 (1.3–3.8)§
Age at follow-up1.1 (1.0–1.1)
Family history of diabetes1.5 (1.0–2.1)
Other ethnic origin than Nordic1.7 (1.1–2.6)
  • Only statistically significant ORs are presented.

  • *

    * Adjusted for the other independent variables in the final model and follow-up length as a noncontinuous variable.

  • The total number includes women from the old cohort at the 1990 follow-up and from the new cohort at the 2002 follow-up. Women with overt diabetes postpartum were not included in the analyses (n = 12). As a result of the OGTT, postpartum is included in the final model, and as only 87.5% had an OGTT postpartum, the total number of women included in the final model was lower than the total number of all the participants. The final model only includes women with available data on all the variables.

  • P < 0.0005

  • §

    § P < 0.01

  • P < 0.05

  • Highest quartile of fasting glucose compared with the three lowest quartiles: 75th percentile = 5.6 mmol/l.