The Role of Exercise in Reducing the Risks of Gestational Diabetes Mellitus

Sarah A Hopkins; Raul Artal


Women's Health. 2013;9(6):569-581. 

In This Article

Can Exercise Prevent GDM?

A number of observational studies have demonstrated that women who report regular physical activity both before and during early pregnancy have a lower risk of developing GDM.[25–28] These studies have used a variety of validated physical activity questionnaires to examine self-reported exercise behaviors among a range of population-based study cohorts. After reviewing data from eight large studies, a recent meta-analysis reported a 55% reduction in the risk of GDM for women in the highest quantiles of prepregnancy physical activity compared with the lowest (odds ratio [OR]: 0.45; 95% CI: 0.28–0.75; p = 0.002).[29] Regular physical activity in early pregnancy was also associated with a reduction in GDM risk, however, the impact appears to be somewhat reduced compared with pregravid exercise (OR: 0.76; 95% CI: 0.70–0.83; p < 0.001). These findings may reflect a dose–response relationship between intensity or volume of physical activity and GDM risk, as has previously been described with similar health outcomes in the nonpregnant population.[30] While the optimal volume of physical activity for GDM prevention is unclear, current observational data support the public health recommendation of accumulating 150 min of moderate-intensity activity each week to confer significant reduction in the risk of developing diabetes.[19]

The beneficial association between physical activity and lower GDM risk may, at least, partially reflect the underlying relationship between an active lifestyle and leaner body composition.[25,26] Alterations in GDM risk in observational studies have been adjusted for prepregnancy BMI and, therefore, reflect an independent effect of regular physical activity on GDM risk. However, few studies have examined whether the effect of physical activity on GDM risk varies by prepregnancy body size. There are some suggestions that the protective effect of exercise may be stronger in women with a higher BMI, who have a higher risk for developing GDM. In 1997, a study published by Dye et al. utilized a population-based birth registry in central New York State between October 1995 and July 1996, with 12,800 women included in the analysis.[31] When stratified by prepregnancy BMI, low levels of physical activity was associated with increased rates of GDM among women with a BMI greater than 33 kg/m2 (OR: 1.9; 95% CI: 1.2–3.1). By contrast, in the lower BMI categories, women who reported regular exercise had a slightly higher incidence of GDM than nonexercisers. By contrast, a study by Oken et al. found that the protective effect of prepregnancy physical activity for the development of abnormal glucose tolerance was only significant in women with a BMI less than 25 kg/m2,[26] while other studies have found no interaction between prepregnancy BMI and the association between physical activity and GDM risk.[25,27] These conflicting findings may be related to the type and timing of physical activity questionnaires. In the study by Dye et al., physical activity was assessed after delivery. It is possible that women in the lower BMI categories (<33 kg/m2) who developed GDM reported physical activity that was initiated after their GDM diagnosis, and were therefore misclassified. These findings may also reflect an interaction between the timing of exercise and underlying differences in the pathophysiology of GDM between lean and obese women. Regular exercise during the prenatal period may have more impact on the prevention of GDM in obese women, in whom the development of GDM may rely more upon the contribution of obesity-associated insulin resistance, in addition to an underlying β-cell defect. By contrast, the development of GDM in lean women may reflect a greater reliance on a β-cell defect that may limit the ability of exercise to prevent the onset of GDM. Additional studies to examine these interactions are warranted before targeted recommendations can be made on the basis of prepregnancy size.

Investigators have recently turned their attention towards examining whether a prescribed exercise program initiated in pregnancy can reduce the risk of developing GDM. Seven randomized controlled trials, published between 2009 and 2012, are summarized in Table 1.[32–38] In contrast to observational data, a recent summary of five of these studies concluded that they could find no clear effect on preventing GDM.[39] There are a number of potential reasons to explain the lack of significant findings from the randomized controlled trials published to date. First, there have only been a limited number of trials conducted in a small number of women. With the exception of two recently published trials,[34,38] all of the other trials have included less than 100 patients. Therefore, it is likely that a lack of statistical power and the failure to control for significant confounding factors, such as diet and gestational weight gain, may have contributed to nonsignificant findings in both individual trials and the published meta-analysis. Power calculations to determine adequate sample size were reported in only one of the five included studies.[34] In this study, the authors estimated that 381 patients were required in each group. This calculation was based upon having statistical power to detect a predicted reduction in GDM incidence from 9% in the control group to 4% in the exercise intervention group. Studies in high-risk populations may not require as many participants, particularly if the incidence of GDM in the control group is higher and the effect of exercise training may be more protective.

Noncompliance to an exercise program and low-intensity exercise volumes have also been limitations of some of the trials. Stafne et al. randomized 855 Norwegian women (mean BMI: 25 kg/m2) to a standard exercise program or routine care control for a 12 week duration, beginning at 18–22 weeks gestation.[34] The intervention program included a weekly supervised 60-min exercise session, containing a combination of aerobic, strength and balance exercises. Subjects were also encouraged to complete at least two 45-min unsupervised sessions each week. Exercise adherence, defined as completion of at least three exercise sessions per week, was achieved by only 55% of women in the intervention group and 10% of the control group. On average, women in the exercise group exercised on 2.0 days per week compared with the 0.7 days per week in the control group. No differences in GDM risk were observed in this study; GDM was diagnosed in 7% of women in the intervention group and 6% in the control. Similar findings were reported in the study by Oostdam et al., who randomized 85 Dutch women (mean BMI: 33.5 kg/m2) who were classified as high risk for GDM to a supervised exercise program or control from 15 to 32 weeks gestation.[37] Their exercise program included two 60-min supervised exercise sessions each week. However, only eight of 53 (16%) intervention individuals who started the exercise program completed at least 50% of the required sessions. Furthermore, when physical activity was assessed using accelerometry at three follow-ups during pregnancy, the exercise group spent fewer minutes performing physical activity per week than controls. GDM was reported in 14.6% of intervention individuals and 21.6% of controls (p = 0.65). In both of these studies, the lack of an exercise effect on GDM risk may be, at least, partially due to the low level of exercise reported in the intervention group. Commonly reported barriers to exercise have included a lack of motivation, physical limitations, and a lack of resources and time. Future studies should continue to address perceived barriers to maintaining regular physical activity in obese pregnant women, and to evaluate optimal exercise prescription and program design to maximize compliance in order to achieve desired health benefits.

Exercise & Glycemic Control in Pregnancy

Maintaining an active lifestyle before pregnancy has been associated with reduced insulin resistance and better glucose tolerance in late pregnancy and may, therefore, protect against the development of GDM. Retnakaran et al. assessed glucose tolerance, insulin sensitivity and β-cell function in 851 (predominantly normal weight) women who underwent a glucose challenge test and 3-h oral glucose tolerance test in late pregnancy.[40] Glucose tolerance status, ranging from normal glycemic control to GDM, improved across increasing quartiles of self-reported physical activity in the 12 months prior to pregnancy (p = 0.02). Furthermore, in multiple linear regression analysis, vigorous exercise prior to pregnancy was a significant independent predictor of insulin sensitivity in late pregnancy. However, it is unclear whether these late pregnancy enhancements in glycemic control were due to the effects of increased pregravid exercise per se, or whether women who were the most physically active in the pregravid period maintained their exercise regimens into their pregnancy.

Few studies have examined the impact of regular exercise on changes in insulin sensitivity during pregnancy, particularly in overweight or obese women. We have recently reported that 15 weeks of moderate-intensity cycling training in nondiabetic women had no impact upon the pregnancy-induced changes in insulin sensitivity in late pregnancy.[32] We randomized a cohort of healthy nulliparous pregnant women, with a range of mid-pregnancy BMIs, primarily within the normal weight range (BMI: 18–32 kg/m2), to a home-based stationary cycling program or nonexercise control for the remainder of their pregnancy. After a ramping period to establish a regular exercise routine, women in the exercise group were asked to maintain five 40-min sessions of moderate-intensity exercise per week at approximately 65% of their predicted aerobic capacity or maximal aerobic capacity. Our observations may indicate that the chronic changes in insulin sensitivity seen in nonpregnant individuals in response to exercise training may be overshadowed in normal weight pregnant women by the persistent regulation of insulin sensitivity required to achieve optimal fetal growth (Figure 2). Future studies should assess the frequency and intensity of exercise training required to impact insulin sensitivity in overweight and obese mothers in light of their increased insulin resistance and elevated risks for GDM. It is likely that chronic adaptations to regular exercise in pregnancy, particularly if accompanied by restricted gestational weight gain, may be different to normal weight women.

During normal pregnancy, it appears that the repeated short-term responses to each exercise session may play the greatest role in influencing glycemic control in late pregnancy. Previous studies have described a significant reduction in glucose and insulin concentrations for a period of time following a single exercise session in late pregnancy.[41,42] In addition, elevated insulin sensitivity has been demonstrated at least 30 min following moderate-intensity cycling exercise in late gestation.[43] With persistent regular exercise, these acute responses may help to regulate appropriate glycemic control in late gestation for nondiabetic pregnant women and, therefore, prevent the development of GDM. However, the majority of these studies have so far been carried out in normal weight women. Future studies should focus on the acute and repeated alterations in glucose and insulin concentrations following exercise in overweight and obese women.