Stafne, Signe N. PT, MSc; Salvesen, Kjell Å. MD, PhD; Romundstad, Pål R. PhD; Eggebø, Torbjørn M. MD, PhD; Carlsen, Sven M. MD, PhD; Mørkved, Siv PT, PhD
Gestational diabetes mellitus, defined as carbohydrate intolerance with the onset or first recognition during pregnancy,1 is a risk factor for pregnancy complications.2,3 Nevertheless, it is unclear whether the latter is attributable to elevated glucose levels per se. Even minor degrees of hyperglycemia are associated with adverse outcomes.4,5 Gestational diabetes is usually diagnosed by an oral glucose tolerance test (OGTT); however, the OGTT procedure and the diagnostic criteria used vary.6 The prevalence of gestational diabetes has been reported to be as high as 14%,7 is increasing worldwide, and varies with diagnostic criteria, ethnicity, and the population studied.7,8 Women with gestational diabetes diagnosed are at increased risk for development of type 2 diabetes later in life.9
In the nonpregnant state, physical activity improves glucose homeostasis; however, exercise must be undertaken regularly to have continued benefits.10,11 Both in prevention and treatment of type 2 diabetes, exercise plays an important role.10 The effect of exercise on the development of gestational diabetes has been studied little, and the study results are conflicting.12–14 In a systematic search on PubMed, no previous randomized controlled trials (RCTs) addressing the effect of physical activity in prevention of gestational diabetes in a general population of pregnant women were found. One smaller trial (n=84) assessed offspring size and maternal insulin sensitivity after exercise training in pregnancy and found no effect on maternal insulin sensitivity.13 Observational studies suggest that higher levels of physical activity before pregnancy or in early pregnancy are associated with lower risk of development of gestational diabetes.14Results from large RCTs of high methodological quality are lacking.15 Hence, we have conducted a randomized trial in two Norwegian university hospitals to assess the efficacy of offering pregnant women a regular exercise program to prevent gestational diabetes and improve insulin resistance.
MATERIALS AND METHODS
We conducted a two-armed, two-center, RCT of a 12-week regular exercise program compared with standard antenatal care. Pretests were performed at 18–22 weeks of gestation and posttests at 32–36 weeks of gestation.
Pregnant women booking appointments for routine ultrasound scans at St. Olavs Hospital, Trondheim University Hospital, and Stavanger University Hospital were invited to participate in the trial. Women in Trondheim were recruited from April 2007 to June 2009, and women in Stavanger were recruited from October 2007 to January 2009. More than 97% of Norwegian pregnant women attend a routine scan at approximately 18 weeks of gestation, and the examinations are free of charge. During the inclusion period, approximately 12,000 pregnant women had routine scans at the two hospitals. Inclusion criteria were white women aged 18 years or older with a singleton live fetus. Exclusion criteria were high-risk pregnancies or diseases that could interfere with participation (or both). For practical reasons, we also excluded women who lived too far from the hospitals to attend weekly training groups (more than 30-minute drive).
Concealed randomization in blocks of 30 was performed at the Unit for Applied Clinical Research, Norwegian University of Technology and Science, by a Web-based computerized procedure. The staff involved with training or outcome assessments had no influence on the randomization procedure. Because of the nature of the study it was not blinded. However, the analyses of glucose and insulin levels were performed blinded for group allocation.
Women in the intervention group received a standardized exercise program including aerobic activity, strength training, and balance exercises. The training protocol followed recommendations from the American College of Obstetricians and Gynecologists and the Norwegian National Report on Physical Activity and Health.16,17 Training sessions of 60 minutes in groups of 8–15 women instructed by a physiotherapist were offered once per week over a period of 12 weeks (between 20 and 36 gestation weeks). Each group session consisted of three parts. The first included 30–35 minutes of low-impact aerobics (no running or jumping). Step length and body rotations were reduced to a minimum, and crossing of legs and sharp and sudden changes of position were avoided. The aerobic dance program was performed at moderate intensity, defined as 13 and 14 on the Borg rating scale of perceived exertion.18 The second included 20–25 minutes of strength exercises using body weight as resistance, including exercises for the upper and lower limbs, back extensors, deep abdominal muscles, and pelvic floor muscles. Three sets of 10 repetitions of each exercise were performed. The third included 5–10 minutes of light stretching, body awareness, breathing, and relaxation exercises.
In addition, women were encouraged to follow a written 45-minute home exercise program at least twice per week (30 minutes of endurance training and 15 minutes of strength and balance exercises). Adherence to the protocol was defined as exercising 3 days per week or more at moderate to high intensity. Performing the exercise program was strongly emphasized and recorded in the women's personal training diaries and through reports from the physiotherapists leading the training groups. In addition, physical activity was recorded in a questionnaire for both groups.
Women in the control group received standard antenatal care and the customary information given by their midwife or general practitioner. They were not discouraged from exercising on their own. Women in both groups received written recommendations on pelvic floor muscle exercises, diet, and pregnancy-related lumbo-pelvic pain.
The procedures followed were in accordance with ethical standards of research and the Helsinki declaration. The women received written information, and they signed informed consent forms. The participants were not compensated financially. The study was approved by the Regional Committee for Medical and Health Research Ethics (REK 4.2007.81) and was registered with clinicaltrial.gov (NCT 00476567).
The primary outcomes were prevalence of gestational diabetes and insulin resistance estimated by the homeostasis model assessment method.19 All participants underwent a 75-g OGTT at inclusion (18–22 weeks of gestation) and at the end of the training period (32–36 weeks of gestation). Fasting and 2-hour glucose levels were measured in serum by the routine methods used by the hospital laboratory. Gestational diabetes was diagnosed according to the World Health Organization criteria as fasting glucose level in fasting whole blood 6.1 mmol/L or more, or plasma glucose 7.0 mmol/L or more, or 2-hour value 7.8 mmol/L or more.20 In addition, secondary outcomes, such as maternal weight, body mass index (BMI, calculated as weight (kg)/[height (m)]2), and pregnancy complications and outcomes (eg, newborn weight, gestational age, Apgar scores) were registered.
For the power calculation, we assumed a gestational diabetes prevalence of 9% in the control group and a prevalence of 4% in the exercise group (risk difference of 5%). Under these assumptions, a two-sample comparison (χ2) with a 5% level of significance and power of 0.80 gave a study population of 381 patients in each group. This sample size was able to detect a 0.2 standard deviation difference in continuous variables given the same power and significance level. Assuming a 10% dropout rate, we needed to include approximately 880 pregnant women.
The primary data analysis was according to the intention-to-treat principle using Stata software 10.0 for Windows. To assess potential differences in the prevalence of gestational diabetes, we used a two-sample test of proportions for univariable analyses and we used a logistic regression model to assess potential differences when adjusting for baseline values. For complete case analysis of continuous variables, we used simple linear regression for univariable analysis and analysis of covariance when adjusting for baseline values. The estimated risk differences were accompanied with 95% confidence intervals (CI) and corresponding P values. To assess the potential effect from missing data, we used a mixed model with random slopes. In this model we used the interaction between treatment group and time to estimate the baseline adjusted differences between exercise and control group. We also performed exploratory analyses using the same statistical models comparing women in the intervention group who adhered to the protocol with the total control group.
In all, 875 women consented to participate in the trial. Twenty women were excluded or withdrew before the first examination: 13 did not meet the inclusion criteria, five miscarried, and two had twin pregnancies. A total of 855 women were randomly allocated to an intervention group or a control group (Fig. 1). However, 33 women in the intervention group and 61 in the control group were lost to follow-up, and 21 intervention group women and 38 control group women did not complete the OGTT. Data from 375 intervention group women and 327 control group women were included in a complete case analysis.
The groups had similar baseline characteristics, except insulin resistance, which was lower in the intervention group (Table 1). Women lost to follow-up reported performing less regular exercise before pregnancy than women completing the study. Among those who completed the study, women in the intervention group had lower fasting insulin and insulin resistance than women in the control group.
We found no differences in the prevalence of gestational diabetes between groups; 25 of 375 (7%, 95% CI 4.3–9.7) intervention group women compared with 18 of 327 (6%, 95% CI 3.3–8.6) control group women (P=.52). Complete case analysis demonstrated lower fasting insulin and insulin resistance in intervention group compared with control group at follow-up, but when adjusting for baseline values the estimated differences were reduced (Table 2). To evaluate the potential effects of the results by attrition and missing data, linear mixed-effects model was performed on all included participants (n=855). The results were similar, giving lower fasting insulin and insulin resistance in the intervention group, but the estimated differences were halved when adjusting for baseline values (Table 3).
In an additional multivariable analysis we adjusted for baseline levels of fasting insulin, fasting and 2-hour OGTT, BMI, and center, but this did not change the results (data not shown). We also checked whether the results differed between the centers by including an interaction term between centers and intervention, but we found no evidence of effect modification (data not shown).
There was no difference between groups in weight gain, weight, BMI, and blood pressure at follow-up (data not shown). At inclusion, women with gestational diabetes (n=43) were significantly older and had higher BMIs and higher levels of insulin resistance, fasting glucose, and 2-hour glucose after an OGTT. There was no difference in level of reported exercise at inclusion or follow-up between women with gestational diabetes diagnosed and those without gestational diabetes diagnosed.
Adherence to protocol (exercising 3 days per week or more at moderate to high intensity) in the intervention group was 55%. For comparison, 10% in the control group exercised 3 days per week or more at moderate to high intensity at follow-up (P<.001). Among the intervention group women exercising less than three times per week during the intervention period, 34% reported no regular exercise at the time of inclusion. In an exploratory analysis, 217 women in the intervention group, who adhered to the protocol, were compared with the total control group. Women adhering to protocol reported performing more regular exercise at inclusion (66% compared with 51%) and had lower fasting insulin (9.3±4.28 compared with 10.7±5.47) and insulin resistance (1.8±0.87 compared with 2.1±1.12) than control group women. At follow-up there was no difference in the prevalence of diabetes (15 of 212, 7%, 95% CI 4.0–11.4) among the protocol adherents compared with the control group (18 of 327, 6%, 95% CI 3.3–8.6; P=.46). In complete case analysis, fasting insulin and insulin resistance were significantly lower in women who adhered to protocol; estimated mean differences (95% CI) were −2.3 (−3.48, −1.02; P<.001) for fasting insulin and −0.43 (−0.17, −0.69; P=.001) for insulin resistance. When adjusting for baseline values the estimated differences between groups were halved, although fasting insulin was still lower compared with that in control group women with estimated mean difference of −1.2 (95% CI −2.3, −0.1; P=.03).
In the intervention period, women in the intervention group exercised 2.0 (95% CI 1.9–2.2) days per week at moderate to high intensity compared with 0.7 (95% CI 0.6–0.8) days in the control group (P<.001). No serious adverse events related to physical exercise were seen, and the outcomes of pregnancy were similar in the two groups (Table 4).
In the present trial there were no significant differences in the prevalence of gestational diabetes or levels of insulin resistance between the intervention group and control group. Fifty-five percent of the intervention group women followed the recommended exercise protocol of training at moderate intensity of 45–60 minutes for 3 days per week. An explorative analysis of protocol-adherent women compared with control participants demonstrated no differences in the prevalence of gestational diabetes but significantly lower fasting insulin at follow-up. This explorative analysis did not have the benefits of an RCT, and results therefore should be interpreted with caution.
The strengths of the present study were the large number of participants and the high methodological qualities such as computerized randomization, the use of validated measurements, and following the general recommendations for the intervention.16,17
The power calculation of the trial was based on previous studies reporting prevalence up to 14%.7 When the present trial was planned, no prevalence studies from Norway were published. Thus, we chose prevalence of 9% and assumed a reduction to 4% in the intervention group. The overall gestational diabetes prevalence in the trial was 6.1% at 32–36 gestational weeks, indicating that the study was slightly underpowered. However, even with a larger trial, there would probably be no statistically significant difference in gestational diabetes prevalence between groups (7% compared with 6% in the present trial). This is supported by a continuous variable with more power, the 120-minute OGTT, which was not different between groups.
Fewer than 10% (n=875) of the 12,000 women attending routine ultrasound scans at the two hospitals during the study period were recruited into this trial. The two hospitals serve large geographical areas, and the number of women who met all the inclusion criteria and were eligible for this trial is not known. Overall, 32% of women exercised regularly before pregnancy at moderate to high intensity three times per week or more, and 13% exercised at the time of inclusion. In another simultaneous Norwegian cohort study (the MoBa study, N=34,508), 46% reported regular exercise at moderate to high intensity three times per week or more before pregnancy, and 28% exercised at gestational week 17.21 Body mass index at enrollment was 24.8±3.2 in our trial and 25.2±4.2 in the MoBa study. Thus, both studies included women who had BMIs within the normal range and who were exercising regularly. Because the MoBa study included 42% of all eligible women, and because our results were comparable, we find the external validity of the trial to be acceptable for Norway. However, the generalizability of the results should be interpreted with caution in other pregnant populations with higher BMIs, in less physically active women, and in ethnically diverse populations.
During pregnancy, an increase in insulin resistance occurs secondary to the diabetogenic effect of one or more of the gestational hormones secreted by the placenta.22 The marked reduction in maternal insulin sensitivity in late pregnancy increases the glucose supply to the fetus.23 In nonpregnant women, the acute effect of exercise includes increased glucose uptake in skeletal muscle via an insulin-independent mechanism that bypasses the insulin signaling defects associated with insulin resistance.11 The effects of one single exercise session on insulin sensitivity persist up to 48 hours. Individuals who undertake regular training may change the expression or activity of a variety of signaling proteins involved in the regulation of skeletal muscle glucose uptake,24 and prepregnancy physical activity has been consistently associated with a reduced risk of gestational diabetes.14 Oken et al25 found that vigorous physical activity before pregnancy and continuation of activity into early pregnancy may reduce a woman's risk for development of abnormal glucose tolerance and gestational diabetes. Thus, there is a possibility that regular exercise before pregnancy and in early pregnancy is more important than exercise during the second half of pregnancy because chronic changes in the regulation of skeletal muscle glucose uptake are adapted, and women may be better able to handle the metabolic stress of a pregnancy.
Another interesting hypothesis is that the effect of exercise on glucose metabolism is different in pregnant and nonpregnant women. Because only 55% of the intervention group women adhered strictly to the exercise protocol, we cannot conclude what possible effects exercising at least three times per week may have directly on glucose metabolism. However, more women in the intervention group exercised regularly between 20 and 36 weeks of gestation compared with the control group. Although some women in the intervention group did not follow the exercise protocol, the difference in the levels of exercise between groups was significant. However, this difference may have been too low to affect insulin sensitivity and development of gestational diabetes. From a public health perspective, the adherence to the intervention must be considered. Reported reasons for not following the protocol were mostly attributable to pregnancy symptoms, child care, and work commitments. Thus, it may be difficult to implement a training program with moderate to high intensity three times per week for pregnant women as suggested in the general recommendations. Nevertheless, adherence to the exercise protocol should be of vital importance for future studies.
Our primary results on insulin resistance are in accordance with a recent smaller trial reporting that insulin resistance was unaffected by exercise in pregnancy.13 In the exploratory analysis of protocol-adherent women, we observed a difference in fasting insulin but no differences in gestational diabetes. The differences in fasting insulin were small and might have been caused by unknown covariates (eg, diet). All women in the trial received the health authorities' standard written dietary advice at inclusion. Dietary advice may have influenced the effect of exercise
Women in the present trial were white, had BMIs within the normal range, and a healthy endocrine function at the time of inclusion. In our population, 10% had a prepregnancy BMIs higher than 27, which is one criteria for screening for gestational diabetes according to current guidelines in Norway. We see two important topics for future research: to study gestational endocrine changes and glucose homeostasis in general and insulin-stimulated glucose uptake in particular, and to study exercise during pregnancy in women with high prepregnancy risk of development of gestational diabetes, such as increased BMI, previous gestational diabetes, and a family history of type 2 diabetes mellitus.
1. Gestational diabetes. ACOG Practice Bulletin No. 30. American College of Obstetricians and Gynecologists. Obstet Gynecol 2001;98:525–38.
2. Scholl TO, Sowers M, Chen X, Lenders C. Maternal glucose concentration influences fetal growth, gestation, and pregnancy complications. Am J Epidemiol 2001;154:514–20.
3. Kim C. Gestational diabetes: risks, management, and treatment options. Int J Womens Health 2010;2:339–51.
4. HAPO Study Cooperative Research Group, Metzger BE, Lowe LP, Dyer AR, Trimble ER, Chaovarindr U, et al.. Hyperglycemia and adverse pregnancy outcomes. N Engl J Med 2008;358:1991–2002.
5. Landon MB, Spong CY, Thom E, Carpenter MW, Ramin SM, Casey B, et al.. A multicenter, randomized trial of treatment for mild gestational diabetes. N Engl J Med 2009;361:1339–48.
6. Cheung NW. The management of gestational diabetes. Vasc Health Risk Manag 2009;5:153–64.
7. Jovanovic L, Pettitt DJ. Gestational diabetes mellitus. JAMA 2001;286:2516–8.
8. Yajnik CS, Fall CH, Coyaji KJ, Hirve SS, Rao S, Barker DJ, et al.. Neonatal anthropometry: the thin-fat Indian baby. The Pune Maternal Nutrition Study. Int J Obes Relat Metab Disord 2003;27:173–80.
9. Bellamy L, Casas JP, Hingorani AD, Williams D. Type 2 diabetes mellitus after gestational diabetes: a systematic review and meta-analysis. Lancet 2009;373:1773–9.
10. Colberg SR, Albright AL, Blissmer BJ, Braun B, Chasan-Taber L, Fernhall B, et al.. Exercise and type 2 diabetes: American College of Sports Medicine and the American Diabetes Association: joint position statement. Exercise and type 2 diabetes. Med Sci Sports Exerc 2010;42:2282–303.
11. Hawley JA, Lessard SJ. Exercise training-induced improvements in insulin action. Acta Physiol 2008;192:127–35.
12. Callaway LK, Colditz PB, Byrne NM, Lingwood BE, Rowlands IJ, Foxcroft K, et al.. Prevention of gestational diabetes: feasibility issues for an exercise intervention in obese pregnant women. Diabetes Care 2010;33:1457–9.
13. Hopkins SA, Baldi JC, Cutfield WS, McCowan L, Hofman PL. Exercise training in pregnancy reduces offspring size without changes in maternal insulin sensitivity. J Clin Endocrinol Metab 2010;95:2080–8.
14. Tobias DK, Zhang C, van Dam RM, Bowers K, Hu FB. Physical activity before and during pregnancy and risk of gestational diabetes mellitus: a meta-analysis. Diabetes Care 2011;34:223–9.
15. Mottola MF. The role of exercise in the prevention and treatment of gestational diabetes mellitus. Curr Diab Rep 2008;8:299–304.
16. American College of Obstetricians and Gynecologists. Exercise during pregnancy and the postpartum period. Clin Obstet Gynecol 2003;46:496–9.
17. National Council on Nutrition and Physical Activity in Norway. Extract from the Norwegian National Report on Physical Activity and Health. Scand J Med Sci Sports 2001;11:255–7.
18. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc 1982;14:377–81.
19. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia 1985;28:412–9.
20. Alberti KG, Zimmet PZ. Definition, diagnosis and classification of diabetes mellitus and its complications. Part 1: diagnosis and classification of diabetes mellitus provisional report of a WHO consultation. Diabet Med 1998;15:539–53.
21. Owe KM, Nystad W, B ø K. Correlates of regular exercise during pregnancy: the Norwegian Mother and Child Cohort Study. Scand J Med Sci Sports 2009;19:637–45.
22. Artal R. Exercise: the alternative therapeutic intervention for gestational diabetes. Clin Obstet Gynecol 2003;46:479–87.
23. Caruso A, Paradisi G, Ferrazzani S, Lucchese A, Moretti S, Fulghesu AM. Effect of maternal carbohydrate metabolism on fetal growth. Obstet Gynecol 1998;92:8–12.
24. Zierath JR. Invited review: Exercise training-induced changes in insulin signaling in skeletal muscle. J Appl Physiol 2002;93:773–81.
25. Oken E, Taveras EM, Popoola FA, Rich-Edwards JW, Gillman MW. Television, walking, and diet: associations with postpartum weight retention. Am J Prev Med 2007;32:305–11.