Gestational diabetes mellitus (GDM), a prevalent pregnancy complication affecting approximately 6.1% (1.8–31.0%) of European pregnancies [1], is escalating due to rising obesity rates among pregnant women [2]. GDM significantly increases adverse outcomes for mothers and offspring, including fetal overgrowth, higher cesarean section rates, and long-term risks like offspring obesity and type 2 diabetes in mothers [3, 4]. Pre-conceptional obesity is a major modifiable risk factor for GDM [5], but it's often addressed too late in prenatal care. Therefore, focusing on modifiable risk factors during pregnancy, such as physical activity, is crucial. Research indicates that physical activity during pregnancy improves glucose uptake and reduces circulating insulin [6, 7], with meta-analyses showing lower GDM risk with physical activity before and during pregnancy [8]. However, many studies focus on early pregnancy, overlooking the typical decrease in activity levels as gestation progresses [9, 10]. This longitudinal change in activity could significantly influence the association with glucose and insulin. Excessive sedentary behavior is an independent risk factor for type 2 diabetes, cardiovascular disease, and premature death outside pregnancy [11–14]. However, studies on its association with the glucose-insulin axis in pregnancy yield inconsistent results [15], likely due to variations in definitions and measurements of sedentary behavior. The DALI (vitamin D And Lifestyle Intervention for gestational diabetes mellitus prevention) study compared a lifestyle intervention promoting physical activity to usual care for GDM prevention [16]. While the intervention reduced sedentary time and increased moderate-to-vigorous physical activity (MVPA), it didn't improve glucose, insulin, or insulin resistance [17, 18]. The small differences in MVPA and sedentary time between intervention groups prompted this secondary analysis of the entire cohort to examine a wider range of activity levels. Using objectively measured accelerometer data, this study aims to investigate the longitudinal relationship between physical activity, sedentary time, and the glucose-insulin axis in obese pregnant women, hypothesizing that reduced sedentary time and increased MVPA will improve this axis.

Existing literature highlights the strong association between obesity during pregnancy and the increased risk of gestational diabetes mellitus (GDM). Numerous studies have demonstrated the benefits of physical activity in improving glucose metabolism and reducing insulin resistance during pregnancy. However, the existing research shows inconsistencies in the measurement and definition of sedentary behavior and its impact on the glucose-insulin axis. A meta-analysis has shown a correlation between pre-pregnancy and pregnancy physical activity and a reduction in the risk of GDM; however, these studies often lack the longitudinal perspective needed to fully understand the changing relationship between activity levels and metabolic parameters throughout gestation. This study seeks to address the gap in understanding by using objective accelerometer data to assess both sedentary time and physical activity at multiple points during pregnancy in a population of obese pregnant women.

This secondary analysis utilized data from the DALI (vitamin D And Lifestyle Intervention for gestational diabetes mellitus prevention) study, a multicenter randomized controlled trial conducted across nine European countries between 2012 and 2015. The study included pregnant women under 20 weeks gestation with a pre-pregnancy BMI ≥ 29 kg/m², who were screened for and excluded if they had pre-existing GDM. Data was collected at three time points: <20 weeks, 24-28 weeks, and 35-37 weeks of gestation. At each time point, participants underwent a 75 g oral glucose tolerance test (OGTT) after an overnight fast, with blood samples analyzed for glucose and insulin levels. Insulin resistance was calculated using HOMA-IR, and first and second-phase insulin release were calculated using the Stumvoll method. Anthropometric measurements and questionnaire data were collected, capturing information on demographics, lifestyle factors, and pregnancy details. Physical activity and sedentary time were objectively measured using accelerometers (ActiGraph GTIM, GT3X+, or Actitrainer) worn for three valid days at each time point, with data processed to determine minutes per day spent in sedentary, light, and moderate-to-vigorous physical activity (MVPA). Longitudinal linear mixed regression models were used to analyze the data, accounting for the nested structure of the data (observations within individuals) with random intercepts and slopes. Models were adjusted for gestational age, randomization group, pre-pregnancy BMI, education, and country. Sensitivity analyses included complete case analysis, exclusion of participants who developed GDM, analysis with time as a categorical variable, and multiple imputation for missing data. Statistical significance was set at p < 0.05. Log transformations were used to address skewness in the data.

The analysis included 232 women with sufficient accelerometer data. Baseline characteristics were comparable between included and excluded participants except for ethnicity and employment status. Accelerometer wear time and MVPA decreased significantly over pregnancy, while sedentary time (in minutes) remained relatively constant. However, as a percentage of wear time, sedentary time increased from <20 weeks to mid-pregnancy. Fasting glucose decreased slightly between early and mid-pregnancy. Other metabolic parameters (postprandial glucose, fasting and postprandial insulin, HOMA-IR, Stumvoll first and second phase) increased significantly throughout pregnancy. The main analysis revealed that higher average sedentary time was significantly associated with higher fasting glucose, fasting insulin, HOMA-IR, and Stumvoll first and second phase. Conversely, higher average MVPA was associated with lower Stumvoll first and second phase. An increase in sedentary time during gestation was associated with elevated Stumvoll first and second phase. Sensitivity analyses generally supported these findings.

This study demonstrates a stronger association between sedentary time and the glucose-insulin axis than between MVPA and the axis in obese pregnant women. This finding highlights the importance of targeting sedentary time reduction, in addition to promoting MVPA, in interventions aimed at improving metabolic health during pregnancy. The longitudinal aspect of this research is significant, as it shows that increased sedentary time during pregnancy itself is a risk factor for adverse metabolic changes. The results underscore the need for early intervention, even before or in early pregnancy, to modify lifestyle behaviors and mitigate GDM risk. The use of objective accelerometer data strengthens the reliability of the findings compared to self-reported measures. While the DALI study did not show an overall impact of the intervention, this secondary analysis, using more precise measurements, reveals important lifestyle factors associated with metabolic health.

This study provides robust evidence highlighting the importance of reducing sedentary time, in addition to increasing MVPA, for improving glucose-insulin homeostasis in obese pregnant women. The significant association between increasing sedentary time during gestation and elevated insulin secretion suggests that interventions should focus on reducing sedentary behavior as early as possible, ideally before or at the start of pregnancy. Future research could explore the effectiveness of targeted interventions specifically focused on reducing sedentary time and the ideal timing and intensity of such interventions to optimize metabolic outcomes during pregnancy.

The study's limitations include the relatively small sample size after applying inclusion criteria based on accelerometer data availability, potential for residual confounding despite adjustments made, and the potential for variations in accelerometer wear time to affect the accuracy of activity measurements. Generalizability might also be affected by the specific population studied and variations in cultural and lifestyle factors among the study's participants in multiple countries.