Risk assessment does not explain high prevalence of gestational diabetes mellitus in a large group of Sardinian women

Background: A very high prevalence (22.3%) of gestational diabetes mellitus (GDM) was recently reported following our study on a large group of Sardinian women. In order to explain such a high prevalence we sought to characterise our obstetric population through the analysis of risk factors and their association with the development of GDM. Methods: The prevalence of risk factors and their association with the development of GDM were evaluated in 1103 pregnancies (247 GDM and 856 control women). The association of risk factors with GDM was calculated according to logistic regression. Sensitivity and specificity of risk assessment strategy were also calculated. Results: None of the risk factors evaluated showed an elevated frequency in our population. The high risk patients were 231 (20.9%). Factors with a stronger association with GDM development were obesity (OR 3.7, 95% CI 2.08-6.8), prior GDM (OR 3.1, 95% CI 1.69-5.69), and family history of Type 2 diabetes (OR 2.6, 95% CI 1.81-3.86). Only patients over 35 years of age were more represented in the GDM group (38.2% vs 22.6% in the non-GDM cases, P < 0.001). Type 2 diabetes in second-degree relatives was equally represented in GDM and non-GDM subjects, while prior poor obstetrical outcomes mostly characterized non-GDM women (17.5% vs 10.6%, P < 0.001). The "average risk" assessment better characterized non-GDM patients (76.8% vs 57.8%, P < 0.001). The logistic regression analysis confirmed that Type 2 diabetes in second-degree relatives, prior poor obstetrical outcomes and the "average risk" definition did not predict the development of GDM. Conclusion: Such a high prevalence of GDM in our population does not seem to be related to the abnormal presence of some known risk factors, and appears in contrast with the prevalence of Type 2 diabetes in Sardinia. Further studies are needed to explain the cause such a high prevalence of GDM in Sardinia. The "average risk" definition is not adequate to predict GDM in our population.