The FUDGE study: Future health and gestational diabetes

Gestational diabetes mellitus (GDM) is one of the most common pregnancy complications and is diagnosed in 5.9% of the pregnant women in Denmark. Globally, the number of women diagnosed with GDM is increasing; hence, more and more women will be impacted by the health implications of GDM, both short- and long-term.

GDM is associated with an increased risk of developing complications during pregnancy and labour affecting the woman as well as the neonate. Additionally, women with GDM face a significantly higher risk of developing diabetes later in life. Data also suggest that the future health after a pregnancy with GDM may be impaired with regard to the risk of other morbidities such as cardiovascular morbidity, psychiatric morbidity and kidney disease. However, the evidence is scarce and the conclusions are conflicting. Furthermore, it is unknown to what extent subsequent development of diabetes might be mediating the risk of future morbidity in women with previous GDM.

GDM is a condition characterised by insulin resistance and beta-cell insufficiency causing hyperglycemia, hence, GDM resembles a metabolic dysfunction. The treatment of GDM comprises counselling on diet and exercise and lifestyle changes. In the more severe cases, insulin treatment is initiated. A range of morbidities is associated with metabolic dysfunction and therefore, women with severe metabolic dysfunction during pregnancy may be at a higher risk of future morbidity.

The overall aim of this PhD project was to investigate the future health after GDM. Specifically, we investigated the long-term risk of incident cardiovascular and metabolic morbidity, psychiatric morbidity, and kidney disease. Additionally, the mediating role of subsequent diabetes in the association between GDM and the future morbidity outcomes was explored. Finally, the impact of the severity of the metabolic dysfunction on the future morbidity risk was investigated.

The PhD project was a nationwide cohort study based on prospectively collected register data on the complete population of women giving birth in Denmark from 1997 to 2018. The project comprised three studies investigating the risk of cardiovascular and metabolic morbidity (Study 1), psychiatric morbidity (Study 2), and kidney disease (Study 3). GDM was the exposure of interest and was based on GDM diagnosis code. The outcomes were based on hospital diagnosis codes and/or redemptions of prescribed medication. The severity of the metabolic dysfunction was defined by generating a proxy variable based on GDM exposure and insulin treatment during GDM pregnancy.

Hereby the study population was categorised into three groups perceived as having increasing severity of the metabolic dysfunction: (1) no GDM, (2) non-insulin-treated GDM, and (3) insulin-treated GDM.

Cox regression models were used to investigate the associations between GDM and future morbidity and to explore the risk according to the severity of the metabolic dysfunction. Mediation analyses were performed to investigate and quantify the proportion of the effect mediated by subsequent diabetes in the associations.

Based on a study population of ~700.000 women, a GDM prevalence of 3.1%–3.4% was found (numbers differed slightly between the three studies). The women were followed for up to 21.9 years with a median follow-up time ranging from 9.3 to 12.0 years depending on the specific outcome. Women with GDM had a higher risk of developing future morbidity compared to women without GDM. In Study 1, we found the following adjusted hazard ratios (aHRs) and 95% confidence intervals (CI): aHR 2.13 (95% CI 2.07–2.20) for any cardiovascular and metabolic morbidity, aHR 1.69 (1.55–1.84) for major cardiovascular disease, aHR 1.89 (95% CI 1.82–1.96) for hypertension, aHR 4.48 (95% CI 4.28–4.69) for dyslipidemia, and aHR 1.32 (95% CI 1.16–1.50) for venous thrombosis. In Study 2, we found an aHR of 1.22 (95% CI 1.18– 1.27) for depression, aHR of 1.20 (95% CI 1.13–1.27) for any psychiatric diagnosis, and aHR of 1.21 (1.17–1.25) for any psychopharmacological medication use. In Study 3, we found an aHR of 1.92 (95% CI 1.67–2.21) for chronic kidney disease. Subsequent diabetes was found to be a significant mediator of the identified associations with proportions of mediated effect ranging from 35% to 76%. Additionally, the studies showed that the risk of future morbidity increased with increasing severity of the metabolic dysfunction during pregnancy.

In conclusion, this PhD project contributed with novel knowledge regarding several aspects of future health after GDM. Women with GDM had a significantly higher risk of developing cardiovascular and metabolic morbidity, psychiatric morbidity and chronic kidney disease later in life. The identified associations were partially mediated by diabetes, however, the associations persisted irrespective of subsequent development of diabetes. Finally, the women with the most severe metabolic dysfunction during pregnancy seemed to constitute a specific high-risk group regarding future morbidity. Overall, this PhD project supported a hypothesis of GDM as a risk factor for future morbidity. Attention to the long-term health in women with previous GDM is warranted. Further studies are needed to elucidate the optimal postpartum and long-term management targeting prevention and early detection of future morbidity.