Main findings
The BEFORE THE BEGINNING study implemented a combination of time restricted eating and exercise training before conception and throughout pregnancy in people at increased risk of developing GDM. Contrary to our hypothesis, we found no significant between group differences in two hour plasma glucose level after a 75 g glucose load in gestational week 28, or in secondary outcome measures of glycaemic control at any time points before or during pregnancy. Although the intervention reduced body weight and fat mass gain at gestational week 28, there was no significant effect on GDM incidence. The participants were able to adhere to the ≤10 hour time restricted eating intervention before pregnancy, with a slight increase in the time window of energy intake during pregnancy. While participants in the intervention group earned more than 100 PAI points before pregnancy, their activity levels declined during pregnancy, and there were no significant between group differences in self-reported energy intake or physical activity throughout the study period.
Strengths and limitations
The main strengths of the BEFORE THE BEGINNING trial are the randomised controlled design and longitudinal measurements throughout the prepregnancy and pregnancy periods. The study had a high pregnancy rate (110/166 participants, 66% in both groups), and a low dropout rate during pregnancy (8/109, 7% in both groups). However, the study has several limitations. Most participants were well educated, potentially resulting in healthy volunteer bias as they were likely knowledgeable about their health and wellbeing.49 Most of our participants were also white and so it is difficult to generalise our findings to other ethnicities. However, recruitment through social media is thought to give a better representation of the general population than other recruitment methods.50 Because we had a low number of participants (n=3) with previous GDM, adjustment for the stratification variable (GDM) was not included in the analyses. Despite good adherence to time restricted eating and exercise training in the prepregnancy period, we did not observe any improvements in cardiometabolic outcomes compared with the control group.
A previous trial of lifestyle intervention in pregnancy for people at high risk of GDM showed that individual changes in physical activity and dietary intake do not need to be large to have a beneficial effect on the incidence of GDM.16 The changes induced by the intervention in our trial could, however, be too small to affect our primary outcome measure. The declining adherence during pregnancy, along with a control group that was motivated to adopt a healthy lifestyle before pregnancy, are probably important reasons for this finding. Self-reported activity levels did not differ significantly between groups, suggesting that the level of physical activity in the control group was similar to the intervention group.
Comparisons with other studies
Previous research indicated that the effectiveness of lifestyle interventions for GDM prevention is greater when they are started early in pregnancy, target high risk populations, limit gestational weight gain, and include moderate intensity exercise for 50-60 minutes twice weekly.51 Most lifestyle interventions for GDM prevention begin during pregnancy, missing the crucial window of opportunity before pregnancy to improve cardiometabolic health outcomes in those at risk.15 Systematic reviews and meta-analyses have highlighted the lack of randomised controlled trials focusing on lifestyle interventions before pregnancy in this population.5253 To date, only a handful of studies have investigated the impact of lifestyle interventions before pregnancy and continued throughout pregnancy on cardiometabolic outcomes.2526545556 We found no significant effect of the intervention on glucose tolerance at gestational week 28, or on any other glycaemic indices. In contrast to our hypothesis, we estimated a higher mean two hour plasma glucose concentration in the intervention group at gestational week 28 according to the intention-to-treat and per protocol analysis (0.48 and 0.64 mmol/L, respectively) compared with the control group. However, the estimated mean differences were smaller than what was considered a clinically relevant difference (1.0 mmol/L). A slight, non-significant increase in two hour plasma glucose concentration (0.2 mmol/L) was also observed after a five week time restricted eating intervention during pregnancy.32
In contrast to our findings, Price and colleagues55 reported that a 12 week very low energy diet intervention before pregnancy significantly reduced two hour glucose during a 75 g oral glucose tolerance test by 0.8 mmol/L compared with a standard diet intervention group among women with body mass index between 30 and 55. In that study, the participants in the intervention group consumed approximately 800 kcal/day for 12 weeks, followed by a maintenance period of energy expenditure matched energy intake throughout the prepregnancy period, while also being advised to remain physically active (>10 000 steps/day). The 12 week prepregnancy intervention induced weight loss of 9.2 kg compared with the standard diet, but gestational weight gain remained unaffected.55 Based on the findings from Price and colleagues55 and our study, it seems likely that a prepregnancy intervention must induce substantial weight loss to impact glucose tolerance in pregnancy. Our trial, along with others,26545556 did not find any significant effect of prepregnancy lifestyle interventions on GDM incidence at gestational week 28. However, GDM rates were lower in the intervention group in the Prepare trial in early pregnancy after a prepregnancy weight loss intervention.25 Despite being at higher risk of GDM, with 146/166 participants (86%) having a body mass index greater than 25 and 44/166 (27%) having a family history of diabetes, only 14/102 (14%) of our participants were diagnosed with GDM in gestational week 28. The GDM incidence in our trial was substantially lower than in the RADIEL study (60% in the control group and 54% in the intervention group)54 and in the study by Phelan and colleagues (40% in the control group and 25% in the intervention group).56 The low incidence of GDM observed in our study might be attributed to the high education level in 144/166 participants (87%) and 148/166 (89%) being white.57 All our participants included in the analyses were normoglycaemic at baseline, which could help explain the lack of effect because people with impaired glucose metabolism benefit the most from time restricted eating.58
There was a non-significant delay in the mean time to pregnancy in the intervention group (112 days) compared with the control group (84 days) in the intention-to-treat analysis. Six participants in the intervention group and three participants in the control group had a prolonged period in the trial because of spontaneous abortions. In our analysis, we estimated the time to pregnancy as the time from inclusion in the trial until the current pregnancy, and this was not always the first pregnancy after inclusion. Therefore, three participants in the intervention group were defined as pregnant more than one year after baseline (after 400, 397, and 374 days, respectively). We do not believe that the intervention affected fecundity. Evidence about the association between physical activity and fertility is inconsistent, but lifestyle interventions can increase natural conception in women with subfertility who are overweight or have obesity.59 Some studies indicate that large amounts of high intensity physical activity are negatively associated with fertility outcomes. However, an increased risk of infertility is typically only seen in those with the highest levels of intensity and frequency of physical activity. In our study, participants in the intervention group were asked to obtain at least 100 weekly PAI points, which is far less than the physical activity levels associated with reduced fertility outcomes in previous studies.
The intervention group had significantly lower body weight gain (2.0 kg) at gestational week 28 compared with the control group, but there were no significant differences between groups in the prepregnancy period or during the first trimester of pregnancy. In contrast, some of the previous prepregnancy lifestyle interventions induced weight loss before pregnancy, but did not affect weight gain during pregnancy.26545556 In the Prepare study, the intervention group lost weight before pregnancy, followed by a greater weight gain in late pregnancy.26 The addition of exercise training to the time restricted eating intervention in our study might have prevented lean mass loss, which is common in time restricted eating interventions.30 Although not statistically significant, the estimated mean visceral fat area was numerically lower in the intervention group compared with the control group before and during pregnancy. The amount of visceral fat in early pregnancy has been shown to better predict GDM than body mass index before pregnancy.60 However, the observed changes in body composition might not have been large enough to improve the other cardiometabolic outcomes in our study.
There were no significant between group differences in total energy intake or macronutrient distribution before or during pregnancy, which could partly explain the neutral effect of the intervention on most outcome measures. In contrast, Haganes and colleagues28 reported that seven weeks of combined time restricted eating and HIIT induced a reduction in energy intake of approximately 200 kcal/day compared with the control group, along with improvements in several cardiometabolic outcomes, among reproductive aged women with overweight or obesity. Correspondingly, the combination of time restricted eating and high intensity functional training for 12 weeks resulted in a 175 kcal/day reduction in total energy intake among women with obesity.61 In BEFORE THE BEGINNING, we chose a modified time restricted eating regimen, allowing for unrestricted intake on two days per week. Even if such a regimen allows for more flexibility and potentially improved long term adherence than a stricter time restricted eating intervention, our intervention was not sufficiently potent to reduce energy intake. Participants undergoing dietary interventions often underreport their total energy intake,62 including when using the electronic application that we used.63 However, there is no reason to believe that such underreporting would be different between the intervention and control groups.
There is little research on the safety and acceptability of time restricted eating during pregnancy, with one previous randomised controlled trial on the effects of time restricted eating on glycaemic control.32 In a recent online survey study, only around half of the participants agreed that a time restricted eating pattern is safe during pregnancy and 23.7% were willing to try time restricted eating during pregnancy to improve their health.64 There was a gradual decrease in the adherence to time restricted eating in our study from before pregnancy and during pregnancy. In our previous study of time restricted eating in pregnancy,32 the participants could adhere to a 10 hour time restricted eating intervention on around five days per week for five weeks in the second or third trimester of pregnancy. However, in that study, the trial period was markedly shorter, and the participants were already pregnant at inclusion, and so likely less affected by nausea and other barriers to time restricted eating than in the BEFORE THE BEGINNING study. Collectively, the experimental evidence to date on time restricted eating during pregnancy indicates that there is no positive effect on maternal glycaemic control.
HIIT is now considered safe and feasible during pregnancy.343536 However, the long term adherence to HIIT interventions during pregnancy remains to be explored. In our study, the participants in the intervention group substantially increased their physical activity levels before pregnancy with a decrease during pregnancy, suggesting suboptimal adherence to the exercise component of the intervention. Only 36/83 participants (43%) in the intervention group who became pregnant met the goal of 100 weekly PAI points before pregnancy, decreasing gradually to 8/55 (15%) during the third trimester. Although self-reported physical activity levels did not differ significantly between the groups, the reported physical activity levels were higher in the intervention group throughout the study. Adherence to exercise training is typically lower in unsupervised compared with supervised situations.65 The declining adherence to exercise during the study period was likely because of a combination of the unsupervised nature of the intervention, pregnancy related side effects (eg, nausea, pelvic pain), and decreasing motivation over a long study period. Overall, adherence to lifestyle interventions in pregnancy remains a significant challenge, especially in real life settings without close supervision.
