Changes in diet and exercise in pregnant women after diagnosis with gestational diabetes: findings from a longitudinal prospective cohort study

Abstract

Background:

Lifestyle changes are recommended for women diagnosed with gestational diabetes (GDM), yet there is little data available documenting whether women change their diet and exercise after GDM diagnosis.

Objective:

The aim of this study was to assess if, and to what extent, pregnant women receiving usual prenatal care change their diet and exercise after GDM diagnosis.

Design:

This study was a post-hoc secondary analysis using data from the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies-Singletons (2009–2013), a prospective pregnancy cohort study.

Participants/setting:

Pregnant U.S. women without major chronic medical conditions were enrolled from 12 participating hospital centers at 8–13 weeks gestation. Diet analyses were based on 5,194 dietary recalls from 1,371 women. Exercise analyses were based on 6,440 physical activity assessemnts from 1,875 women. GDM was ascertained from medical records according to Carpenter and Coustan criteria. Women completed 24-hour dietary recalls and exercise assessemnts at weeks 16–22, 24–29, 30–33, 34–37, and 38–41 (exercise only).

Main Outcome Measures:

The main outcome was the diet and exercise change from before to after GDM diagnosis/screening.

Statistical Analyses:

Diet and exercise changes with 95% confidence intervals from before to after GDM diagnosis/screening for women with and without GDM were estimated using weighted multivariable linear mixed models.

Results:

Women with GDM (n=72) significantly reduced their total energy intake [−184 kcal/day (95% confidence interval −358, −10)] and carbohydrate intake [−47.6g/d (−71.4, −23.7)] from before to after GDM diagnosis; these changes were unique to women with GDM and not observed among women without GDM (n=1299). Women with GDM decreased intakes of juice [−0.4 cups/d (−0.7, −0.2)] and added sugar [−3.2 teaspoons/d (−5.5, −0.5)], and increased cheese [0.3 cups/d (0.1, 0.6)] and artificially sweetened beverages [0.2 cups/d (0.0, 0.3)]. Women with GDM (n=84) did not change their exercise duration after diagnosis; women without GDM (n=1791) significantly decreased moderate [−19.5 min/wk (−24.7, −14.3)] and vigorous exercise [−8.8 min/wk (−10.6, −6.9)] after GDM screening.

Conclusions:

Women with GDM made modest dietary improvements and maintained their pre-diagnosis exercise routine, yet opportunities remain to further improve dietary intake and exercise following diagnosis of GDM.

INTRODUCTION

The American College of Obstetricians and Gynecologists (ACOG)¹ and the American Diabetes Association (ADA)² recommend that women with gestational diabetes (GDM) be provided nutritional and exercise counseling after diagnosis. As such, lifestyle intervention is the first line of therapy for GDM,³ for which the primary goal is to reduce hyperglycemia and the secondary goal is to reduce GDM associated neonatal complications.¹ The Academy of Nutrition and Dietetics recommends that women with GDM are referred to a dietitian for individualized medical nutrition therapy.⁴ Women historically were counseled to limit carbohydrate intake; currently women are recommended to meet the Dietary Reference Intakes for pregnant women and there are no specific guidelines regarding how much to limit carbohydrates.^2,4 Women are also recommended to participate in a moderate exercise program to better manage their glucose levels.^1,2,4

Numerous lifestyle intervention trials have been conducted among women with GDM. Compliance to nutritional therapy within randomized trials generally ranges from 60–80%.⁵ While intensive trials can attempt to maximize compliance by providing food and other incentives to participants,⁶ it is unclear how much women receiving usual care actually modify their diet and exercise after diagnosis of GDM. Identifying areas where women with GDM were more or less successful in modifying their diet and exercise routine may help to guide clinicians in providing more targeted nutritional and behavioral therapy, particularly if sub-groups of high-risk women can be identified. The aim of this study was to use a prospective longitudinal cohort with repeated assessments of diet and exercise to assess if, and to what extent, women receiving usual care outside of a research setting changed their diet and exercise after being diagnosed with GDM and to identify differences across maternal characteristics.

MATERIALS AND METHODS

This study was a post-hoc secondary analysis of data from the prospective, longitudinal Eunice Kennedy Shriver National Institutes of Child Health and Human Development (NICHD) Fetal Growth Studies – Singletons cohort (2009–2013; CLINICALTRIALS.GOV Identifier: NCT00912132). Data were collected from 2,334 women without obesity and 468 women with obesity (total n=2,802) at 12 U.S. clinical centers.⁷ The primary aim of the NICHD Fetal Growth Studies was to develop fetal growth standards, thus enrollment was restricted to women without preexisting chronic diseases, medical conditions, or obesity (n=2,334).⁸ Secondary aims were to examine etiology of GDM and associations of obesity and fetal growth and thus a supplemental cohort of women with obesity was recruited (n=468).⁹ Briefly, eligibility among non-obese women was limited to non-smoking women with a pre-pregnancy body mass index (BMI) of 19–29.9 kg/m², women not currently drinking ≥ 1 alcoholic drink per day or using illicit drugs, women who conceived without fertility drugs or IVF, women without prior pregnancy complications, women without chronic diseases including diabetes, asthma, hypertension, cancer, psychiatric disorders, autoimmune diseases, thyroid disease, renal disease, hematologic disorders, HIV or AIDS, epilepsy, or anorxia or bulimia. Eligibility among women with obesity was limited to women with a pre-pregnancy BMI of 30.0–45.0 kg/m², women without chronic diseases including diabetes, hypertension requiring two or more medications, cancer, psychiatric disorders, autoimmune diseases, renal disease, or HIV or AIDS. Women were screened by study coordinators, and if they met the study eligibility criteria and agreed to participate, they were enrolled. All women were enrolled at 8–13 weeks gestation and followed through delivery. Prenatal and maternal delivery records were abstracted by research staff. All women provided written informed consent. Approval was received from the institutional review board at each participating site, the data coordinating center, and the NICHD.

A total of 18 women were determined to be ineligible from the NICHD Fetal Growth Studies-Singletons after enrollment and excluded from analyses. For the current analyses, women without medical chart abstraction (n=201; 7.2%) were excluded. Women were classified accoding to whether they were diagnosed with GDM. The non-GDM comparison group was limited to normoglycemic women women (i.e., known to not have GDM or any degree of impaired glucose tolerance). Therefore women not screened for GDM (n=126; 4.9%) as well as women who did not have GDM but had an abnormal result on the glucose challenge test or oral glucose tolerance test (n=306; 12.4%) were excluded from the analysis. Additionaly, women without at least one dietary and physical activity assessment before and after GDM diagnosis/screening were excluded; the final sample for the dietary analyses included 5,194 dietary recalls from 1,371 women and the physical activity analyses included 6,440 physical activity assessemnts from 1,875 women (Supplemental Figure 1). The sample size for the dietary analyses was much smaller than the physical activity analyses because the dietary assessment component was deployed only midway through the cohort enrollment due to additional funding.¹⁰ Women excluded due to missing dietary assesments were of higher pre-pregnancy BMI, more likely to be Hispanic, more likely to have less attained education, less likely to be nulliparous, and less likely to have private or managed care insurance (Supplemental Table 1). Women excluded due to missing physical activity assesments were younger, of higher pre-pregnancy BMI, more likely to be Hispanic, more likely to have less attained education, and less likely to have private or managed care insurance (Supplemental Table 2). To account for potential selection bias due to missing data all anaylses were weighted to be represenative of the original full NICHD Fetal Growth Studies Singelton Cohort using inverse probablity of selection weights.¹¹

Diet and Exercise

Past day dietary intake was collected at each study follow-up visit targeted at 16–22, 24–29, 30–33, and 34–37 weeks using the Automated Self-Administered 24-hour (ASA24) Dietary Assessment Tool (versions Beta/2011) developed by the National Cancer Institute (Bethesda, MD). The ASA24 is based on the validated USDA Automated Multiple-Pass Method (AMPM).¹² Intakes were autocoded by the ASA24 system using the USDA Food and Nutrient Database for Dietary Studies (versions 1.0/4.1).¹³ Consistent with prior studies, records with total caloric intake <600 or >6000 kcal/d were considered implausible and excluded (<1%).¹⁴ The Healthy Eating Index (HEI)-2010 total score and component scores were calculated using the simple HEI scoring algorithm.¹⁵ The HEI-2010 was used as a measure of overall dietary quality defined as conforming to the 2010 Dietary Guidelines for Americans.¹⁶ Serving sizes for food group variables was based on the USDA MyPyramid Equivalents Database.¹⁷

Physical activities since the preceding study visit were assessed using the validated Pregnancy Physical Activity Questionnaire at 16–22, 24–29, 30–33, 34–37, and 38–41 weeks.¹⁸ The current analysis focused on exercise activities classified as moderate (≥3.0 to ≤6.0 metabolic equivalents; e.g., walking slowly or quickly, prenatal exercise classes, swimming) or vigorous (>6.0 metabolic equivalents; e.g., walking up hills or jogging). Minutes per week were used to improve translation of the results, as most public health recommendations for exercise are in minutes per day or week opposed to metabolic equivalent minutes per week.

Gestational Diabetes

Women received standard clinical care for GDM screening and diagnosis, which included a glucose challenge test and/or an oral glucose tolerance test, as necessary. Diagnostic testing for GDM occurred at a median of 27.0 weeks (standard deviation 2.2). GDM was identified by medical record review of oral glucose tolerance test results according to Carpenter and Coustan criteria, as currently endorsed by the ACOG and ADA.^1,2

Additional variables

Women underwent a screening ultrasound at enrollment to confirm accurate gestational dating by last menstrual period and completed interviewer-administered medical history and socio-demographic questionnaires. Maternal height and pre-pregnancy weight were self-reported. Pre-pregnancy body mass index (BMI; kg/m²) was calculated and categorized as normal weight (19.0–24.9 kg/m²), overweight (25.0–29.9 kg/m²), or obese (30.0–45.0 kg/m²).

Statistical methods

Descriptive characteristics were compared between women with and without a GDM diagnosis using t-tests for continuous variables and chi-square tests for categorical variables.

Linear mixed effects models were used to estimate changes in diet (total energy, macronutrient composition, foods, and HEI-2010 total and component scores) and exercise levels from before to after GDM diagnostic testing among women with and without GDM. Women without GDM were included as a comparison group to assess whether any observed changes among women with GDM were specifically due to GDM and not an artifact of progression of pregnancy. The dependent variable for these models was the continuous dietary and exercise variables. Each model included independent fixed effects of GDM status (yes/no) and the measurement timing (before/after GDM screening/diagnosis) as well as an interaction term between the two to estimate changes from before to after diagnostic testing among women with and without GDM, as well as season of the recall and weekday/weekend status of the recall (diet only). The models also included a random effect for participant and measurement timing. From the unadjusted model, the average dietary intake and exercise levels were estimated and plotted in each of the following four groups of women based on the intercept, the main effects for GDM status and measurement timing, and the interaction effect between the two, assuming the reference season of winter and weekday measurement: women with GDM before diagnosis, women with GDM after diagnosis, women without GDM before screening, women without GDM after screening. P-values testing whether dietary intake and exercise differed from before and after diagnosis/screening among women with and without GDM were computed using contrasts based on the main effects of GDM status and measurement timing, and their interaction, as appropriate. The change in dietary intake and exercise levels with 95% CIs and corresponding P-values was estimated from before to after GDM diagnosis/screening among women with and without GDM in adjusted models based on the model effects for GDM status and measurement timing, and their interaction, as appropriate. The following covariates were included in the adjusted models: age, self-reported race/ethnicity, pre-pregnancy BMI, education, nulliparity, marital status, insurance status, study site, year, and gestational week at GDM diagnostic testing. All models were weighted using inverse probability of selection weights to address the potential for selection bias among the sample with observed data.

A second set of linear mixed effects models assessed the association between women’s baseline characteristics and changes after GDM diagnosis in carbohydrate intake among women with GDM to identify specific groups of women who were more or less successful at making changes. These models focused on carbohydrate intake only as no overall significant changes were observed in the other macronutrients or in exercise. The dependent variable for these models was carbohydrate intake. Each model included independent fixed effects of measurement timing (before/after GDM diagnosis), the baseline characteristic of interest, as well as an interaction term between the two. The models also included a random effect for participant and measurement timing. All models were adjusted for age, self-report race/ethnicity, pre-pregnancy BMI, education, nulliparity, marital status, insurance status, study site, year, gestational week at GDM diagnostic testing, season, and weekend/weekday status of the recall. All models were weighted using inverse probability of selection weights to address the potential for selection bias among the sample with observed data. From the model effects, the change with 95% CIs was estimated within each group of women.

There was no missing covariate data among the analytic sample. A two-sided p-value of less than 0.05 was considered significant. All analyses were performed using SAS v9.4 (Cary, NC).¹⁹

RESULTS

Among women with dietary data, 72 women (5.3%) were diagnosed with GDM. Women with GDM were older, had a higher pre-pregnancy BMI, were more likely to be Hispanic or Asian, more likely to be married, and more likely to have some college or an associate degree (Table 3). Sample characteristics were similar for the larger sample of women with physical activity data (Supplemental Table 4). Women had a median of two dietary recalls before (interquartile range 1–2; range 1–3) and two recalls after (interquartile range 2–3; range 1–4) GDM testing and a median of one physical activity assessment before (interquartile range 1–2; range 1–3) and two assessments after (interquartile range 2–2; range 1–3) GDM testing (data not shown).

Table 3.

Sample characteristics for dietary recall sample according to gestational diabetes status among pregnant women in the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies-Singletons, 2009–2013, United States.

Characteristicsa	GDMb (n=72) Mean (SD) or n (%)	Non-GDM (n=1299) Mean (SD) or n (%)	P-Valuec
Age, years	30.3 (6.0)	27.6 (5.6)	<0.001
Race/Ethnicity			0.006
Non-Hispanic White	17 (23.6)	265 (20.4)
Non-Hispanic Black	11 (15.3)	441 (34.0)
Hispanic	24 (33.3)	368 (28.3)
Asian	20 (27.8)	225 (17.3)
Pre-pregnancy body mass index, kg/m2	28.0 (6.2)	25.0 (4.8)	<0.001
<25.0	25 (34.7)	765 (58.9)
25.0–29.9	25 (34.7)	354 (27.3)
≥30.0	22 (30.6)	180 (13.9)
Education			0.02
High school or less	20 (27.8)	419 (32.3)
Some college or Associate degree	33 (45.8)	396 (30.5)
Bachelor’s degree or higher	19 (26.4)	484 (37.3)
Nulliparous	34 (47.2)	610 (47.0)	0.97
Married or living with partner	58 (80.6)	900 (69.3)	0.04
Private or managed care insurance	46 (63.9)	768 (59.1)	0.42

^aSample characteristics shown for dietary recall sample. For sample characteristics of physical activity sample see Supplemental Table 4.

^bGDM = gestational diabetes.

^cP-values computed using either t-tests for continuous variables and chi-square tests for categorical variables.

In unadjusted analyses, among women diagnosed with GDM, from before to after diagnosis, the proportion of calories from carbohydrates (52.2% to 46.1%) was significantly reduced, while the proportion of calories from protein (17.1% to 19.3%) and fat (31.8% to 35.7%) were significantly increased (Figure 2a). Although significant for carbohydrates (P=0.02) and fat (P=0.02), no meaningful changes were observed in protein (16.6% to 16.4%), fat (33.1% to 32.5%), or carbohydrates (51.7% to 52.4%) among women without GDM across the similar time frame. The HEI-2010 total score for did not differ from before to after GDM diagnostic testing for women with (51.7 to 53.9; P=0.20) or without (51.9 to 51.9; P=0.90) GDM (data not shown). Women with and without GDM reduced their duration of moderate and vigorous exercise, although the change was statistically significant only among women without GDM (Figure 2b).

Figure 2.

Unadjusted macronutrient composition (A) and exercise duration (B) among women with and without gestational diabetes from before to after gestational diabetes (GDM) diagnostic testing in the NICHD Fetal Growth Studies-Singletons, 2009–2013, United States. Panel A plots the unadjusted distribution of the percentage of calories from carbohydrate, fat, and protein estimated from a linear mixed effects model with repeated measures of dietary intake. Among women with GDM (n=72), the p-values for the difference in carbohydrate, fat, and protein from before to after diagnosis were <0.001 for all. Among women without GDM (n=1299), the p-values for the difference in carbohydrate, fat, and protein from before to after GDM screening were 0.02, 0.02, and 0.23, respectively. Panel B plots the number of minutes women participated in moderate and vigorous exercise linear mixed effects model with repeated measures of reported exercise. Among women with GDM (n=84), the p-values for the difference in moderate and vigorous exercise from before to after GDM diagnosis were 0.50 and 0.21, respectively. Among women without GDM (n=1791), the p-values for the difference in moderate and vigorous exercise from before to after GDM screening was <0.001 for both.

Adjusted diet and exercise changes from before to after GDM diagnostic testing are presented in Table 5. Among women with GDM there was a significant decrease in total energy intake [−184 kcal/day (95% confidence interval (CI) −358, −10)] and a significant decrease in carbohydrate intake [−47.6g/d (95% CI, −71.4, −23.7)] after diagnosis. No significant change in the intake of protein or fat were observed, which resulted in a significant decrease in the proportion of calories from carbohydrates [−6.1 (95% CI, −8.6, −3.6)] and a significant increase in the proportion of calories from protein [2.3 (95% CI, 1.1, 3.4)] and total fat [4.0 (95% CI, 2.0, 6.0)]. Comparatively, over the similar time frame, women without GDM significantly increased their proportion of calories from carbohydrate by 0.8 (95% CI 0.2, 1.4) suggesting that the decrease in carbohydrate intake among women with GDM post-diagnosis was related to GDM diagnosis and not an artifact of the timing of gestation (P-interaction<0.001). After diagnosis women with GDM did not significantly alter their weekly amount of moderate (P=0.45) or vigorous (P=0.23) exercise. Comparatively, women without GDM significantly decreased their moderate (P<0.001) and vigorous (P<0.001) exercise over the similar timeframe.

Table 5.

Adjusted changes in diet and exercise from before to after gestational diabetes diagnostic testing among pregnant women in the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies-Singletons, 2009–2013, United States.

	GDMa Adjusted Change (95% CI)b,c,d,e	Non-GDM Adjusted Change (95% CI)b,c,d,e
Total energy, kcal/day	−184 (−358, −10)	10 (−34, 55)
Dietary composition, % of total energy
Carbohydrates	−6.1 (−8.6, −3.6)	0.8 (0.2, 1.4)
Protein	2.3 (1.1, 3.4)	−0.2 (−0.5, 0.1)
Total fat	4.0 (2.0, 6.0)	−0.6 (−1.1, −0.1)
Saturated fat	1.3 (0.4, 2.2)	−0.3 (−0.5, −0.1)
Polyunsaturated fat	0.7 (−0.0, 1.4)	0.0 (−0.2, 0.2)
Monounsaturated fat	1.6 (0.8, 2.5)	−0.3 (−0.5, −0.1)
Macronutrient intake, grams/day
Carbohydrates	−47.6 (−71.4, −23.7)	5.5 (−0.6, 11.5)
Protein	0.4 (−8.6, 9.5)	−0.6 (−2.9, 1.7)
Total fat	0.5 (−8.2, 9.2)	−1.1 (−3.3, 1.2)
Saturated fat	−0.5 (−3.8, 2.8)	−0.7 (−1.5, 0.2)
Polyunsaturated fat	0.4 (−1.9, 2.7)	0.2 (−0.3, 0.8)
Monounsaturated fat	0.7 (−2.7, 4.0)	−0.5 (−1.4, 0.4)
Exercise, minutes/week
Any Moderate	−8.8 (−31.8, 14.2)	−19.5 (−24.7, −14.3)
Any Vigorous	−4.9 (−13.0, 3.2)	−8.8 (−10.6, −6.9)

^aGDM = gestational diabetes.

^bAnalyses adjusted for age, race/ethnicity, pre-pregnancy BMI, nulliparity, marital status, education, insurance, study site, season, year, weekday/weekend status (diet only), and gestational week at GDM diagnostic testing.

^cDiagnostic testing between women with and without GDM (i.e., a significant interaction) for total energy, carbohydrates %, protein %, total fat %, saturated fatty acids %, monounsaturated fatty acids %, and carbohydrates g/d.

^dBeta estimates and 95% confidence intervals, and P-values estimated from an adjusted weighted linear mixed effects model.

^eDietary analyses are based on 72 women with GDM and 1299 women without GDM. Exercise analyses are based on 84 women with GDM and 1791 women without GDM.

Figure 3 shows the adjusted change in foods and the HEI-2010 total and component scores from before to after GDM diagnostic testing. After diagnosis, women with GDM significantly decreased intake of juice by 0.4 cups/d (95% CI, −0.7, −0.2) and added sugar intake by 3.2 teaspoons/d (95% CI, −5.5, −0.5), and increased intake of cheese by 0.3 cups/d (95% CI, 0.1, 0.6) and artificially sweetened beverage by 0.2 cups/d (95% CI, 0.0, 0.3). Further, while there was not a significant change in the total HEI-2010 score, there was a significant increase in the HEI-2010 components scores for total vegetables, whole grains, and diary, but a decrease in the components scores for total fruit and sodium. The latter changes were not observed among women without GDM over the similar time frame who comparatively increased intake of milk by 0.1 cups/d (95% CI, 0.0, 0.2), and sugar sweetened beverages by 0.1 cups/d (95% CI, 0.0, 0.2). A significant decrease in the HEI-2010 component score for total vegetables was also observed among women without GDM.

Figure 3.

Changes in food servings and the Healthy Eating Index 2010 total score and component score from before to after gestational diabetes (GDM) screening and diagnosis among women with and without GDM in the NICHD Fetal Growth Studies-Singletons, 2009–2013, United States. Estimates are obtained from weighted linear mixed effects models adjusted for age, race/ethnicity, pre-pregnancy body mass index status, marital status, parity, education, insurance, study site, year, season, gestational weeks at GDM diagnosis, and weekend status of the recall. The units for food servings are either cups per day (whole fruit, juice, vegetables, milk, cheese, yogurt, sugar sweetened beverages, artificially sweetened beverages), ounces per day (whole grains, non-whole grains, red meat, processed meat, poultry, fish, eggs, soy, legumes, nuts and seeds), or teaspoons per day (added sugar). The maximum points for the Healthy Eating Index 2010 total score and each of the component scores is shown in parathesis.

Further, the study investigated which groups of women were likely to change their carbohydrate intake after GDM diagnosis by examining changes from before to after GDM diagnosis by the pre-pregnancy characteristics of women with GDM (Table 6). The change in carbohydrate intake from before to after GDM diagnosis was significant in all groups of women except women who were 35–41 years of age, women who were Hispanic, women with obesity, women who had a high school education or less, and women with at least one prior birth.

Table 6.

Changes in carbohydrate intake from before to after gestational diabetes diagnosis by pre-pregnancy characteristics among pregnant women diagnosed with gestational diabetes (n=72) in the Eunice Kennedy Shriver National Institute of Child Health and Human Development Fetal Growth Studies-Singletons, 2009–2013, United States.

Characteristic	Carbohydrates, grams/day, Change (95% CI)a
Age, y
18–24	−69.8 (−118.1, −21.4)
25–29	−71.8 (−115.8, −27.9)
30–34	−52.3 (−91.0, −13.6)
35–41	−18.7 (−57.8, 20.5)
Race/Ethnicity
Non-Hispanic White	−49.7 (−86.7, −12.7)
Non-Hispanic Black	−67.2 (−127.6, −6.8)
Hispanic	−38.0 (−76.7, 0.6)
Asian	−58.8 (−103.1, −14.6)
Pre-pregnancy weight status
Normal weight	−64.6 (−100.3, −28.9)
Overweight	−51.4 (−88.9, −14.0)
Obese	−34.6 (−71.2, 2.1)
Education
≤High school	−39.2 (−80.0, 1.7)
Some college or Associate degree	−48.2 (−80.1, −16.4)
≥Bachelor’s degree	−65.0 (−105.5, −24.4)
Nulliparous
Yes	−84.0 (−112.7, −55.3)
No	−15.2 (−44.7, 14.4)

^aAnalyses adjusted for age, race/ethnicity, pre-pregnancy body mass index status, marital status, parity, education, insurance, study site, year, season, gestational weeks at GDM diagnosis, and weekend status of the recall. Beta estimates and 95% confidence intervals obtained from a weighted linear mixed effects model.

DISCUSSION

For women diagnosed with GDM, a healthy diet and exercise regimen are regarded as the first-line of treatment to reduce hyperglycemia and lessen the adverse impacts from GDM.³ Yet there are limited data examining real-world diet and exercise changes made by women following GDM diagnosis and little understanding of factors associated with making desired lifestyle changes. In this diverse U.S. cohort of pregnant women receiving usual prenatal care, after being diagnosed with GDM, women made significant dietary changes which resulted in lower caloric intake by limiting intake of juice, and added sugars, and by increasing artificially sweetened beverages. These changes led to a substantial reduction in carbohydrate intake without a compensatory increase in fat or protein. Interestingly, there was also a small increase in cheese intake. Also, the degree of change in caloric intake and carbohydrates were not equitable across all groups of women. Thus, even with modest improvements in diet made by most women, greater efforts in optimizing diet may be necessary for some women with GDM. Also, while their pre-diagnosis exercise was on average less than recommended, women with GDM tended to maintain their pre-diagnosis exercise routines unlike women without GDM who significantly reduced the time spent exercising after GDM screening.

Overall women made modest dietary changes after being diagnosed with GDM, specifically including a reduction in carbohydrate intake of approximately 48 grams. Women with GDM did not replace carbohydrates with other sources of energy such as saturated fat, which is reassuring as replacement with saturated fats could further promote excessive fetal growth in this high-risk group of women.^6,20 Furthermore, women did not decrease whole fruits or whole grains, but rather reduced overall added sugars and juice, which may be useful as a diet high in complex carbohydrates may be useful in managing maternal metabolic health after GDM diagnosis.²¹ Furthermore, as indicated by the HEI-2010 total score, adherence to the 2010 Dietary Guidelines for Americans was low throughout pregnancy among both women with and without GDM. This is notable as a lower HEI-2010 score has been associated with suboptimal maternal and neonatal outcomes.^22,23 Further research is needed to identify ways to increase dietary quality among all pregnant women.

Women typically lowered their carbohydrate intake by reducing added sugars and juice. Also, an increase in daily consumption of artificially sweetened beverages was observed; thus, it is possible that some women replaced juice for artificially sweetened drinks. While some,^24,25 though not all studies,²⁶ have observed an association between artificial sweeteners in pregnancy and increases in birthweight or childhood obesity, the broad implications of consuming these beverages in pregnancy remains unknown and is an ongoing area of research.

An important finding of this study was that dietary changes after diagnosis were not equitable across all groups of women. No changes in carbohydrate intake were observed among women with obesity which is of great importance given that women with obesity already have an elevated risk for macrosomia independent of GDM.²⁷ Further, no changes in carbohydrate intake were observed among Hispanic women, women with a high school education or less, women aged 35–41 years, or multiparous women. These findings highlight importance of individualized programs to ensure that all women with GDM are successful at modifying their diet post diagnosis.

In addition to modifying diet after diagnosis of GDM, women are encouraged to integrate exercise into their regular routine to decrease the risk for adverse outcomes. The ACOG recommends 150 minutes per week of moderate intensity exercise for the majority of pregnant women, regardless of GDM,²⁸ and the ADA recommends that women with GDM undertake 20–30 min of moderate-intensity exercise most days of the week.²⁹ While some studies have found that exercising after meals is effective for maintaining glucose control,³⁰ others have been limited as to whether exercise post-GDM diagnosis improves outcomes.³¹ At an average of just over 90 minutes of moderate to vigorous exercise per week, most women in this study did not achieve adequate exercise prior to or after GDM diagnosis. While women without GDM significantly reduced their time spent exercising as they entered the third trimester, women with GDM notably maintained their exercise routine post diagnosis. Most prior interventions have been unsuccessful at increasing women’s physical activity during pregnancy.³² More innovation and research is needed to understand how to acceptably intensify pregnant women’s physical activity, particularly for women with GDM where it may be associated with improved glycemia and potentially improved neonatal outcomes.

One limitation of this study was the utilization of self-reported diet and exercise. However, multiple assessments of both diet and exercise before and after diagnosis of GDM were used. The Pregnancy Physical Activity Questionnaire provides reliable and valid estimates of pregnant women’s activities, particularly for sports/exercise activities.¹⁸ The ASA24 utilizes the well validated USDA multi-pass method,¹² but underreporting remains a concern. Also, the number of women with GDM in this study was relatively small which may have limited power to detect differences particularly by women’s characteristics. In addition, the sample size limited us from examining whether diet changes among women with GDM are associated with changes in infrequent adverse maternal and neonatal outcomes, which have been shown to be reduced in some GDM dietary intervention trials.³³ Further research is needed to understand if the observed dietary changes are effective at improving maternal and neonatal outcomes. This study did not have information on the actual prenatal care or nutritional guidance that was provided to women in this multi-site cohort study. Given that there is no standardized approach for nutritional therapy for GDM management, it is difficult to make assumptions about the specific guidance women in this study were given. The analyses adjusted for study site to account for potential differences across sites. We did not have information on women’s glucose monitoring during pregnancy and could not assess how this related to women’s change in diet and exercise.

This study has several unique strengths. First, it demonstrates that while there are modest changes in diet and exercise post-diagnosis among women receiving usual care for GDM, opportunities remain for improvements especially for some sub-groups of women. This data may inform the need to scale up novel lifestyle intervention strategies that have been found to be successful for the management of GDM.^34–36 Another strength of this study was that it examined dietary and exercise behaviors post-GDM screening among normoglycemic women to determine whether changes observed among women with GDM were unique to these women or if they were an artifact of changes in gestational timing. Notably, the decrease in carbohydrate intake was unique to women with GDM post-diagnosis. Lastly, the study utilized a large diverse, multisite, U.S. cohort generalizable to women without major chronic diseases prior to pregnancy.

CONCLUSION

After diagnosis with GDM women reduced their overall carbohydrate intake and improved the quality of consumed carbohydrates. While positive, these changes may represent ‘low hanging fruit’ and there are likely remaining opportunities to further improve education and patient behavior related to dietary intake and exercise following diagnosis of GDM. Further, these data suggest that changes were not uniform across all groups of women. These data underscore the need for more research to identify effective nutritional and behavioral therapy strategies and innovative approaches to enhance maternal diet and exercise levels which may promote the health of all women with GDM and their babies.

Supplementary Material

1

Supplemental Figure 1. Flow chart for the analysis of diet and exercise changes before and after screening and diagnosis of gestational diabetes (GDM), NICHD Fetal Growth Studies-Singletons, 2009–2013, United States.

RESEARCH SNAPSHOT:

Research Question:

Do pregnant women diagnosed with gestational diabetes modify their diet and exercise post-diagnosis?

Key Findings:

In this large diverse prospective cohort study, women with gestational diabetes significantly reduced their total energy intake and specifically their carbohydrate intake after diagnosis. Post-diagnosis, women with gestational diabetes significantly decreased their intake of juice, and added sugar, and increased their intake of cheese and artificially sweetened beverages. Women with gestational diabetes maintained their pre-diagnosis exercise routine.