Abstract
Aims
This parallel two-arm, randomized, controlled trial (RCT) tested the efficacy of wearing ankle weights during daily activities (e.g., childcare and housework) on body weight and composition changes, and glycemic, lipid, and inflammatory biomarkers in postpartum women diagnosed with gestational diabetes mellitus (GDM) during pregnancy.
Methods
The intervention group (1.1 kgs each, 2 hours/day, from 25-35 to 190-210 days’ postpartum) was compared to a standard-of-care control group in women with GDM during pregnancy. The primary outcome was body weight. Secondary outcomes included body composition and glycemic, lipid, and inflammatory biomarkers. Linear mixed models with time*treatment arm interaction compared changes in outcomes among the intervention group (ANK, n = 18) and control group (CTRL, n = 21), adjusted for clinic site and pre-pregnancy body mass index (BMI) category.
Results
The intervention had no significant effect on weight change (mean (SD): 3.8 (1.8) kg for ANK vs 2.2 (1.6) kg for CTRL, P-for-interaction = 0.61) or percent body fat (−0.6 (0.9) for ANK vs −1.5 (0.8) for CTRL, P-for-interaction = 0.52). Both groups reduced their total cholesterol, but there was no intervention effect on this change (−25.3 (7.3) mg/dL for ANK vs −18.5 (6.7) mg/dL for CTRL, P-for-interaction = 0.40) or on other secondary outcomes. Registered under Physical Activity Intervention for Gestational Diabetes (GDM) on Clinicaltrials.gov NCT03664089 (https://clinicaltrials.gov/study/NCT03664089?term=NCT03664089&rank=1).
Keywords: gestational diabetes, postpartum, physical activity, weight loss, pregnancy, type II diabetes
“One meta-analysis found that diet and/or physical activity interventions did not have significant effects on fasting plasma glucose, 2-hr glucose tolerance test, or HbA1C.”
Introduction
Gestational Diabetes Mellitus (GDM) affects 63.5 per 1000 live births in the US, which is an increase of about 33% from the previous decade. 1 It is estimated that one third of women with a history of GDM have a subsequent diagnosis of type 2 diabetes mellitus (T2DM) within 15 years of their pregnancy, 2 and the relative risk for T2DM is nearly 10 times greater in women with GDM compared with healthy controls. 3 Women with history of GDM are also more likely to experience complications associated with the T2DM diagnosis including increased risk for myocardial infarction and coronary artery disease. 4 Postpartum weight loss has been shown to improve glucose metabolism in women with GDM 5 ; whereas, excessive postpartum weight retention is associated with long-term obesity and can exacerbate the risk of T2DM. 6
Many previous studies have examined the effects of lifestyle interventions in both reducing the risk of GDM during pregnancy and in reducing the risk of T2DM in women with GDM postpartum. Many of these lifestyle interventions focus on both physical activity and nutrition.7-18 The multiple health behavior change theory supports that targeting multiple behaviors may be more efficacious for weight loss than targeting a single behavior 19 ; however, it may be challenging for new mothers to implement multiple health behavior changes in their daily life due to large lifestyle changes associated with caring for an infant. Three web and phone-based interventions targeting both diet and exercise found a positive effect on diet behaviors, but had no effect on physical activity and mixed effects on weight loss.7-9 Similarly, another study found that a diet and physical activity intervention did not affect glycemic status or development of type 2 diabetes in postpartum women with GDM compared to standard of care. 10
Having moderate or high levels of physical activity has been independently associated with postpartum weight loss in women with obesity 20 ; however, many women report barriers to physical activity participation in the postpartum period, including: fatigue, lack of motivation and confidence, time constraints, and lack of access to affordable and appropriate activities. 21 Previous interventions targeting these barriers have been successful in addressing lack of energy and motivation, but were unable to address time constraints. 11 One previous study found that women with GDM in a resistance exercise program during pregnancy had improved blood glucose levels and less gestational weight gain compared to the control group 22 ; however, no similar studies exist for women postpartum. All previous interventions relied on participants devoting additional time outside of their normal responsibilities.
It is evident there is a need to identify an efficacious intervention that postpartum women can balance with other responsibilities and that has ease of use and a low barrier to entry, making it more scalable to the many women affected by GDM. Our intervention attempted to address this by increasing the physical intensity of routine daily activities. New mothers spend much of their time engaged in light-intensity physical activity including cooking, cleaning, and childcare, which have been shown to lower blood glucose. 23 We hypothesize that by increasing the intensity of daily activities, our intervention will increase the energy expenditure of participants and promote weight loss (primary aim) and improve body composition and biomarkers (secondary aim).
Subjects, Materials, and Methods
Study Design and Setting
The Moms in Motion was a randomized controlled trial to evaluate the efficacy of a simple, novel, physical activity-boosting intervention on postpartum weight loss among women with GDM. The study protocol has been published. 24 Participants were recruited from Ohio State University Wexner Medical Center Maternal-Fetal Medicine Diabetes in Pregnancy Clinics in Columbus, Ohio, USA between September 2018 and April 2021. These clinics served a large and diverse population of privately and publicly insured patients referred from a large catchment area. The study was reviewed and approved by the Ohio State University Institutional Review Board.
Participants, Sample Size and Power
Participants were women diagnosed with GDM during their current pregnancy. GDM was diagnosed with an oral glucose tolerance test using the Carpenter-Coustan criteria, 25 typically between 24- and 28-weeks’ gestation, or earlier if a patient was high-risk. Eligible women included those who were 18 years of age or older, English-speaking, planned to remain in the area for the duration of the study and were physically capable of engaging in moderate physical activity. Women with a prior type 1 or type 2 diabetes diagnosis, surrogate gestational carriers, pregnant with multiples, delivered prior to 35 weeks’ gestation, had a pre-pregnancy BMI of less than 18.5 (underweight), or lived more than 35 miles away were ineligible. All participants provided written informed consent.
Clinic staff distributed informational flyers to potentially eligible participants who attended the Diabetes in Pregnancy education class (first clinical point of contact after a GDM diagnosis). Study staff approached patients during a prenatal care appointment to discuss the study and invite participation. Women could enroll and provide written informed consent in clinic at 30-40 weeks’ gestation or at a home visit at 7-14 days’ postpartum.
A sample size of 160 participants was planned to provide 80% power to detect a Cohen’s d effect size of 0.5 for the effect of the intervention on the primary outcome (weight loss), assuming 20% attrition, and 2 groups of 80. This sample size was selected based on preliminary data from our pilot study which observed an effect size of 0.63 (mean weight loss of 2.9 kg in the intervention group, 1.9 kg in control group).
Run-in Period, Randomization, Masking, and Intervention
To improve adherence to the intervention, a run-in period started at a visit in the participant’s home at 7-14 days’ postpartum. Participants received an ActiGraph wGT3X-BT accelerometer to wear for all waking hours during the subsequent 7 days. Those who wore the accelerometer for more than 49 hours over the 7-day period were invited to continue in the study; others were dismissed.
Randomization occurred at approximately 30 days’ postpartum during a study (baseline) visit at Nationwide Children’s Hospital Clinical Research Services. A pseudorandom number generator in the statistical software package SAS was used to implement randomization, stratified by study clinic (clinic A or B) and BMI group (18.5 – 24.9 or ≥25) with randomly varying block sizes. Co-author MAK, who had no involvement or contact with the participants, maintained the concealed randomization sequence unavailable to other study staff and investigators. Participants were randomized 1:1 to either the intervention or the control condition. Blinding of study investigators, staff, and participants was not possible with this intervention.
Women assigned to the intervention arm received the standard recommendation to engage in 150 minutes of physical activity per week from their obstetric care provider. They received a pair of ankle weights (2.5 pounds [1.1 kg] per ankle) at baseline after randomization to wear 2 hours daily during routine daily activities (i.e., childcare, household chores) until the final in-person study visit at approximately 220 days’ postpartum. Women assigned to the control arm received the standard recommendation from their obstetric care provider to engage in 150 minutes of physical activity per week (standard of care).
Data Collection
At the time of consent, participants answered surveys about sociodemographic information (income, race, ethnicity, age, insurance, employment). Parity was abstracted from the obstetric record. At baseline, study staff measured the participant’s weight and body composition (percent body fat) using the Tanita TBF-310GS Body Composition Analyzer. Waist circumference (cm) was assessed using metric tape at the middle point between the ribs and the iliac crest, with the participant in a standing position. Hip circumference (cm) was measured at the widest circumference of the buttocks. Before conducting measurements and periodically throughout the study, staff completed anthropometric measurement training and reliability checks to ensure consistency and accuracy. Height was abstracted from the obstetric record, which may have been a measured or self-reported height. Body mass index was calculated as weight (kg) divided by height (m2). A registered nurse performed venipuncture for biomarker measurements of cardiometabolic health. Then, a fasting 2-hour, 75g oral glucose tolerance test (OGTT) was conducted. Phone calls at approximately 90, 120, 150, and 180 days’ postpartum allowed staff to answer questions and encourage compliance with study-related activities. A final study visit at approximately 220 days’ postpartum at Nationwide Children’s Hospital repeated all baseline measures. These served as the final measures.
Primary Outcome
Weight Loss
The change in weight (kg) from baseline to the final study visit (approximately 190 days’ duration) served as the primary outcome measure.
Secondary Outcomes
Body Composition
Change in body composition including body fat percentage, BMI, and waist-hip ratio from baseline to the final study visit served as secondary outcome measures.
Glycemia and Associated Biomarkers
Other secondary outcome measures included blood biomarkers of glucose (2 hour), insulin, hemoglobin A1c, and lipids (total cholesterol, high density lipoprotein, low density lipoprotein, and triglycerides) analyzed by the clinical laboratory at Nationwide Children’s Hospital. Adiponectin and high sensitivity C-reactive protein assays were performed by ARUP Laboratories (Utah), and leptin assays were performed by Esoterix Endocrinology (California). Insulin sensitivity and β-cell function were measured using HOMA-IR, which was calculated using fasting serum insulin (μU/ml) × fasting plasma glucose (mmol l-1)/22.5. Laboratory values that were reported as being below the lower limit of detection for the assay were assigned the value of the lower of limit of detection.
Other Variables
Energy Expenditure and Physical Activity
All women received an ActiGraph wGT3X-BT accelerometer to monitor their daily activity and energy expenditure. Using a belt clip, women wore the ActiGraph on their waist for 21 days after the baseline visit and for 21 days before the final visit. All devices were calibrated based on the participant’s anthropometric data and configured for recording. Participants received reminders asking them to wear their ActiGraph upon request or if study staff observe that the participant is not wearing the device as often as desired. Participants could not view the information collected by the ActiGraph. Physical activity was measured by the percentage of time in moderate or vigorous physical activity (MVPA) as calculated by ActiGraph ActiLife software. We also reported percentage of time in sedentary and light activity.
Diet
Participants completed the Diet History Questionnaire III (DHQ III) at baseline and the final study visit, a web-based food frequency questionnaire based on a collection of national 24-hour dietary recall data from the National Health and Nutrition Examination Surveys (NHANES) conducted from 2007 to 2014. 26 Diet was characterized by the average energy intake per day and total healthy eating index (HEI) 2015 score, a measure of adherence to the US dietary guidelines.
Compliance
Study staff instructed participants in the intervention arm to document the date, time, and activities undertaken while wearing ankle weights in a study journal for the duration of the intervention. Compliance was measured by calculating the mean number of minutes per days the ankle weights were worn. If a participant reported that they did not wear their ankle weights on a particular day, that value was treated as zero. If a participant did not report whether they wore their ankle weights on a particular day, that value was treated as missing and was not included in the calculation. Staff continued to collect outcome data from participants who discontinued the intervention as long as the participant did not formally withdraw and was not lost to follow-up.
Statistical Analysis
The study followed CONSORT statement guidelines using intent-to-treat as the primary approach for analysis. All individuals were kept in the group to which they were randomized, regardless of protocol violations or dropouts. Balance between the treatment groups was tested by Fisher’s exact test for baseline variables between study arms. Outliers were inspected and retained. Outcome variables were examined over time at the group and the individual level graphically as well as via descriptive statistics stratified on time and group. No interim analyses or stopping guidelines were planned.
Bivariate relationships between outcome variables and potential covariates were examined to determine if substantial deviations from linearity existed. Analyses for the primary and secondary outcomes involved mixed effects model for repeated measures, similar to ANCOVA but based on maximum likelihood. 27 Each body composition and biomarker outcome were evaluated in relation to treatment group and time in separate models. Because some very high values were observed for hsCRP, we conducted a sensitivity analysis with log10 hsCRP as the outcome. Pre-specified variables including race, pre-pregnancy BMI category, GDM severity, parity, clinic location, and education were planned to be included as covariates in the primary and secondary analyses, but the smaller than expected sample size precluded inclusion of so many variables in the models. Thus, only the variables used to stratify the randomization were included: clinic location and pre-pregnancy BMI category. Analyses were planned to evaluate whether diet and physical activity were mediators of observed treatment effects, but the sample was too small to conduct robust mediation analyses. Instead, indicators of average diet quality and daily moderate or vigorous physical activity were calculated by treatment group at baseline and the differences in change between by treatment group were investigated using a mixed effect model using a time*treatment interaction. Analysis for diet and physical activity was conducted post-hoc. Analyses used SAS 9.4.
Impacts of the COVID-19 Pandemic
At the onset of the COVID-19 pandemic, participant recruitment slowed because clinical research facilities were closed, and clinical care was disrupted. The study home visits were offered as a virtual telehealth option. Our smaller-than-anticipated sample size required the statistical analyses to be streamlined and reduced the statistical power to detect small and medium effects. Although scholars have highlighted the pitfalls of post-hoc power analysis, 28 we estimated that our achieved sample size enabled the detection of an effect size of Cohen’s d = 0.18 with 20% power rather than the projected effect size of d = 0.63 with 80% power. To detect an effect size of d = 0.63, the study had 33% power, in the end.
Results
Participant Characteristics
Of 1265 patients screened for eligibility, 722 were excluded because of ineligibility or logistical barriers that prevented recruitment, leaving 543 who were approached to invite enrollment (Figure 1). Of those, 82 were ineligible, 232 declined participation, and 154 were nonresponsive. A total of 75 participants consented to participate, but before they could be randomized 15 withdrew, 10 were lost to follow-up, 6 failed the run-in period, 3 were withdrawn by the investigators because COVID-19 related restrictions prevented continuation in the trial, 1 became ineligible, and 1 enrolled just before all study activities ceased. Thus, 39 were randomized and received their assigned intervention, 18 to the treatment arm and 21 to the control arm. All of these were included in the analysis. Thirty (77%) completed the final study visit. All data collection was completed by July 2022. There was no statistically significant difference in any baseline variables between the treatment and control groups (Table 1).
Figure 1.
CONSORT Diagram.
Table 1.
Participant Characteristics by Treatment Group, Moms in Motion Trial (n = 39, 2018-22, Ohio, USA).
| Characteristic | All n (%) | Ankle Weight (Intervention) Arm n (%) | Control Arm n (%) | P-value |
|---|---|---|---|---|
| Maternal age | .89 | |||
| 22-27 | 9 (23) | 5 (28) | 4 (19) | |
| 28-32 | 12 (31) | 6 (33) | 6 (29) | |
| 33-35 | 9 (23) | 4 (22) | 5 (24) | |
| 36+ | 9 (23) | 3 (17) | 6 (29) | |
| Ethnicity | 1.00 | |||
| Hispanic/Latin | 3 (8) | 1 (6) | 2 (10) | |
| Not Hispanic/Latin | 36 (92) | 17 (94) | 19 (90) | |
| Race | .41 | |||
| Asian | 7 (18) | 4 (22) | 3 (14) | |
| Black or African-American | 5 (13) | 3 (17) | 2 (10) | |
| White | 19 (49) | 6 (33) | 13 (62) | |
| Multiple | 2 (5) | 1 (6) | 1 (5) | |
| Other | 4 (10) | 3 (17) | 1 (5) | |
| Missing | 2 (5) | 1 (6) | 1 (5) | |
| US born | 30 (77) | 13 (72) | 17 (81) | .70 |
| Currently employed | 28 (72) | 12 (67) | 16 (76) | .70 |
| Annual household income | .73 | |||
| <$20,000 | 8 (21) | 5 (28) | 3 (14) | |
| $20,000-<$50,000 | 9 (23) | 3 (17) | 6 (29) | |
| $50,000-<$100,000 | 8 (21) | 4 (22) | 4 (19) | |
| >$100,000 | 10 (26) | 5 (28) | 5 (24) | |
| Missing | 4 (10) | 1 (6) | 3 (14) | |
| Health insurance | 1.00 | |||
| Private insurance | 31 (79) | 14 (78) | 17 (81) | |
| Medicaid (Caresource) | 7 (18) | 3 (17) | 4 (19) | |
| Missing | 1 (3) | 1 (6) | 0 (0) | |
| Number of children (not counting current pregnancy) | .35 | |||
| 0 | 13 (33) | 8 (44) | 5 (24) | |
| 1 | 13 (33) | 4 (22) | 9 (43) | |
| 2+ | 13 (33) | 6 (33) | 7 (33) | |
| Clinic | .42 | |||
| Upper Arlington | 32 (82) | 16 (89) | 16 (76) | |
| McCampbell | 7 (18) | 2 (11) | 5 (24) |
Compliance
On average, those randomized to the intervention arm reporting wearing the ankle weights for 94.9 minutes/day (SD = 36.7), which is 79% of the recommended 120 minutes of ankle weight wear/day. Two participants reported at least 120 minutes/day of ankle weight wear. Five participants did not report any compliance data and were excluded from this calculation due to missing data.
Primary Outcomes
Overall, participants had an average weight change of 2.8 kg from baseline to the final study visit (SD = 6.6, range = [-9.9, 17.2]). That is, they tended to gain, rather than lose, weight. This did not differ by treatment group (mean for the intervention arm = 3.7 (SD = 6.9), mean for the control arm = 2.2 (SD = 6.6), P = 0.61) (Table 2).
Table 2.
Change in Anthropometric Measurements and Biomarkers From Baseline to the Final Study Visit, Comparing Ankle-Weight Intervention Arm to the Control Arm, Moms in Motion Trial (Ohio, 2018-2022, USA).
| Outcome | Baseline (Mean (SD)) | End of Trial (mean(SD)) | Change (mean(SD)) | Difference in Change (95%CI) | P-value** | |||
|---|---|---|---|---|---|---|---|---|
| Ankle Weight (n = 18) | Control (n = 21) | Ankle Weight (n = 13) | Control (n = 17) | Ankle Weight | Control | |||
| Weight (kg) | 84.1 (17.1) | 83.4 (17.5) | 90.5 (23.0) | 84.7 (19.5) | 3.7 (6.9) | 2.2 (6.6) | 1.2 (−3.6-6.1) | P = .61 |
| BMI (kg/m2) | 31.5 (6.8) | 30.9 (5.4) | 33.8 (9.4) | 31.4 (6.6) | 1.5 (2.6) | 0.8 (2.4) | 0.5 (−1.3-2.4) | P = .54 |
| Fat percent | 42.0 (5.3) | 41.5 (5.2) | 42.2 (8.4) | 39.4 (6.6) | −0.6 (3.2) | −1.3 (3.2) | 0.6 (−1.4-2.7) | P = .52 |
| Fat free mass (kg) | 48.0 (5.7) | 48.0 (6.6) | 51.7 (6.1) | 49.9 (8.2) | −2.4 (1.4) | −1.8 (3.5) | 0.27 (−1.7-2.2) | P = .77 |
| Waist to hip ratio | 0.79 (0.07) | 0.83 (0.06) | 0.81 (0.07) | 0.83 (0.08) | 0.02 (0.03) | 0.00 (0.04) | 0.00 (−0.02-0.02) | P = .89 |
| 2-hr glucose (mg/dL) | 118.2 (22.8) | 129.0 (30.3) | 130.8 (26.3) | 138.8 (55.0) | 19.8 (26.2) | 7.4 (34.1) | −3.6 (−18.9-11.7) | P = .63 |
| Total cholesterol (mg/dL) | 208.9 (28.9) | 212.2 (41.8) | 181.1 (23.2) | 185.2 (34.7) | −24.5 (25.4) | −15.9 (23.9) | −5.2 (−16.8-6.5) | P = .40 |
| LDL (mg/dL) | 134.2 (25.8) | 133.2 (33.7) | 112.4 (16.2) | 114. 9 (32.4) | −18.6 (20.5) | −10.0 (20.6) | −3.2 (−13.4-7.1) | P = .53 |
| HDL (mg/dL) | 51.1 (11.3) | 57.1 (19.1) | 48.9 (15.2) | 48.3 (14.0) | −2.1 (12.0) | −5.2 (8.1) | −1.0 (−6.5-4.5) | P = .71 |
| A1C* (%) | 5.3 (0.3) | 5.3 (0.4) | 5.4 (0.3) | 5.7 (1.3) | 0.2 (0.2) | 0.3 (1.2) | −0.1(-0.3-0.2) | P = .59 |
| Triglycerides (mg/dL) | 118.3 (58.8) | 108.7 (66.3) | 99.1 (60.4) | 109.7 (64.9) | −19.3 (40.1) | −4.1 (63.3) | −3.9 (−30.4-22.6) | P = .76 |
| Insulin (uU/mL) | 8.2 (7.8) | 9.2 (9.0) | 15.4 (12.7) | 17.1 (18.0) | 5.1 (8.2) | 4.7 (12.6) | −0.3 (−5.2-4.6) | P = .90 |
| hsCRP (mg/dL) | 3.6 (1.8-5.7)* | 5.7 (4.2-9.1)* | 3.4 (2.3-8.4)* | 3.7 (3.1-5.6)* | 1.0 (2.8) | −4.5 (7.3) | 0.3 (−1.2-1.9) | P = .65 |
| Adiponectin (ug/mL) | 9.1 (5.0) | 8.7 (3.6) | 11.1 (5.7) | 9.1 (4.8) | 1.1 (4.5) | 0.2 (3.0) | 0.4 (−1.5-2.4) | P = .65 |
| Leptin (ng/mL) | 24.0 (12.8) | 20.3 (10.0) | 30.7 (21.7) | 20.9 (9.3) | 5.0 (18.0) | 0.0 (6.2) | 4.4 (−1.0-9.8) | P = .11 |
| HOMA-IR | 1.9 (1.8) | 2.2 (2.4) | 4.4 (3.4) | 5.4 (6.5) | 2.1 (2.2) | 1.8 (4.7) | −0.2 (−1.6-1.1) | P = .74 |
*Median (IQR).
**P-value for mixed effect model time*treatment interaction.
Secondary Outcomes
Among the secondary outcomes, biomarkers like total cholesterol tended to decrease across the time of the trial, as might be expected during the postpartum period. 29 However, other outcomes including 2-hour glucose values and BMI appeared to be steady or trending in an unfavorable direction overall. Formal comparisons by treatment group revealed no statistically significant differences by treatment group in changes over time among the secondary outcomes (Table 2). The sensitivity analysis for the log10 of hsCRP did not differ from hsCRP.
Physical Activity and Diet
According to the Actigraph data, our sample spent around two thirds of their time in sedentary behavior, regardless of timepoint or treatment group. Of the time not spent in sedentary behavior, most of the time was spent in light physical activity, which is expected of new mothers. The physical activity distribution was very similar for the control group from baseline to follow-up, and there was no statistically significant difference in change in light physical activity or MVPA between the treatment groups. There was also no statistically significant difference between the change in energy intake between the two treatment groups. On average, the HEI for our participants was greater than 50 showing that they met at least half the dietary guidelines. There was a significant difference in the change in diet quality between the 2 treatment groups (P = 0.01); diet quality improved from baseline to follow-up in the ankle weight group and worsened in the control group (Table 3).
Table 3.
Measures of Diet and Physical Activity at Baseline and End of Trial by Study Arm.
| Baseline (Mean (SD)) | End of Trial (mean(SD)) | P-value | |||
|---|---|---|---|---|---|
| Ankle Weight (n = 17*) | Control (n = 20) | Ankle Weights (n = 13) | Control (n = 18**) | ||
| Energy take/day (kcal) | 1795 (655) | 2007 (545) | 2070 (884) | 1912 (583) | .30 |
| HEI score | 68.7 (11.0) | 64.1 (11.0) | 71.3 (12.1) | 62.2 (9.8) | .01 |
| Percent time in sedentary | 69.3 (9.1) | 72.0 (8.5) | 64.0 (8.2) | 71.5 (10.0) | .03 |
| Percent time in light | 29.6 (8.9) | 26.7 (8.1) | 33.0 (7.4) | 27.0 (9.7) | .06 |
| Percent time in MVPA | 1.1 (0.9) | 1.2 (1.3) | 3.0 (2.6) | 1.5 (1.3) | .21 |
*n = 16 for physical activity.
**n = 17 for energy intake and HEI.
Discussion
This study examined the efficacy of wearing ankle weights during daily activities on postpartum weight loss and other biological indicators in women with recent GDM. We hypothesized that those randomized to the ankle weight group would experience greater weight loss and improved biomarkers compared to those receiving the standard of care, thus reducing the risk of developing T2DM. However, no difference in weight loss was observed between the two groups. There was no difference in any of the biomarkers of interest. We also observed low adherence to the intervention, which is common in behavioral interventions in this population.7,18
Lifestyle interventions involving diet and/or physical activity are commonly studied among postpartum women with recent GDM due to their evidence in promoting weight loss and reducing T2DM risk in other populations. In contrast to our findings, a recent systematic review and meta-analysis of randomized controlled trials found that behavioral interventions showed significant effects on weight loss, BMI, and waist circumference among women with recent GDM. 30 However, most of the interventions used in that analysis targeted both diet and physical activity. And although pooled effects were statistically significant, there were individual studies that showed null results, similar to the present study. One study that was efficacious in increasing weight loss involved 10 phone-based sessions provided by credentialled diabetes educators, 8 which differs from our intervention’s hands-off approach. Another intervention that did not achieve improved weight loss was an app-based intervention where participants were given activity and energy intake targets to help them achieve weight loss, which was also much more intensive than our ankle-weight intervention. 7
Most studies examining biomarkers focus on measures of glycemia. One meta-analysis found that diet and/or physical activity interventions did not have significant effects on fasting plasma glucose, 2-hr glucose tolerance test, or HbA1C. 31 Another intervention utilizing facilitated group sessions, automated voice and text messages, and phone calls did not see improvement glycemic status in their intervention group, 10 which is consistent with our findings. In our population, measures of glycemia, including 2-hr glucose, insulin, and A1C, did not improve, regardless of randomization group.
There remains a need to identify an effective intervention that addresses new mothers’ barriers to participating in lifestyle interventions. One significant shortcoming of previous physical activity interventions is that they require mothers to take time out of their day to complete additional tasks, and these studies were unable to increase the amount of time mother’s spent completing physical activity. Our intervention attempted to address this barrier by increasing the intensity of everyday activities without requiring additional time. Our hypothesis relied on the assumption that mothers are already engaging in physical activity while completing everyday activities; however, our population spent much of their time in sedentary behavior both at the start and end of the study. Since the physical activity levels were so low, the addition of ankle weights alone may have been insufficient for increased weight loss in the intervention group.
Additionally, compliance to the intervention was low, consistent with previous studies,7,18 which may explain why we did not observe more weight loss in the intervention group. Although we addressed the key barrier of time constraints, we speculate that postpartum women may have hesitated to engage in the intervention because they found the ankle weights to be uncomfortable, bulky, or unfashionable. Postpartum women may simply forget to wear the ankle weights because they are juggling the many other responsibilities of new motherhood. The main shortcomings of the study were the small sample size compared to original plans which reduced statistical power, loss to follow-up and withdrawal of participants, low adherence to the intervention, and recruiting from a limited geographical area. Due to small sample size, we were also unable to investigate if energy expenditure or dietary intake mediated the effects of the intervention. We used weight loss as a proxy for T2DM risk but did not measure incidence of T2DM to evaluate direct effects of the intervention on development of T2DM. Ongoing monitoring of the sample for T2DM is warranted, notably, the mean A1C in the control group at the end of the trial was at the threshold for pre-diabetes.
This trial was unique in that it investigated a physical activity intervention that integrated into the life of postpartum mothers rather than requiring additional time, thereby addressing one of the key barriers that postpartum women have reported to physical activity participation. We were able to measure a variety of outcome measures including both body measurements and biomarkers that are key indicators of T2DM risk. The study population was indicative of the geographical area and consisted of a variety of racial and ethnic groups (Table 1). Those in the intervention group improved their diet quality throughout the during of the trial, although diet was not the target of the intervention.
Conclusion
In conclusion, we did not observe increased weight loss or improved biomarkers from the ankle-weight intervention in postpartum GDM women. The much-reduced sample size compared to original plans played an important role. Further research in larger samples examining physical activity and energy intake as mediators is warranted to understand why the intervention did not have an effect. Overall, there remains a need to identify effective interventions for postpartum weight loss in women with previous GDM that are simple and address key barriers to participation.
Acknowledgments
We thank the families who participated in the study. Research staff and clinicians: Amanda James, Brenda Widmayer, Lisa Buccilla, Julie Somppi, Patricia Guittar, Kajal Gandhi, Andrea Bonny, Jessica Schiering, Bonny Bowen, Lydia Ward, Andria Parrott, Taniqua Ingol, Holly Blei, Abigail Dean, Anna Wiese, Katie Smith, Melissa Kravets, Samantha Buls, Shelli Farley, Emily Viall, Rachel Mason, Anna Bartholomew, Alyson Johnson, Alexia Neri, Jessica Cohen, Dawn Cline, Stephanie Brindle, and Sounali Perez. Nationwide Children’s Hospital and Ohio State University Wexner Medical Center provided data collection and administrative support.
Footnotes
Author Contributions: Steven Gabbe: Investigation, Resources, Writing—Review and Editing Sarah Keim: Conceptualization, Data Curation, Formal Analysis, Investigation, Project Administration, Resources, Supervision, Writing—Review and Editing Mark Klebanoff: Investigation, Writing—Review and Editing Mark Landon: Investigation, Resources, Writing—Review and Editing Reena Oza-Frank: Conceptualization, Funding Acquisition, Investigation, Project administration, Resources, Supervision, Writing—Review and Editing Caroline Racke: Data Curation, Formal Analysis, Investigation, Writing—Original, Writing—Review and Editing Briana Stith: Investigation, Project Administration, Writing—Review and Editing Stephen Thung: Investigation, Project Administration, Resources, Writing—Review and Editing Samrawit Yisahak: Data Curation, Formal Analysis, Investigation, Writing—Review and Editing
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the American Diabetes Association (grant #1-18-JDF-062) and, in part, by the National Center for Advancing Translational Sciences of the National Institutes of Health under Grant Number UM1TR004548. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. These sources were not involved in the design of the study and collection, analysis, and interpretation of data nor in writing the manuscript.
Ethical Statements
Ethical Approval
The study was reviewed and approved by the Ohio State University Institutional Review Board on 4/23/2018 (approval # IRB18-00178). All participants provided written informed consent prior to enrollment in the study. This research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki.
Consent to Participate
All participants provided written informed consent.
Consent for Publication
Written informed consent for publication was provided to the participants or a legally authorized representative.
ORCID iD
Sarah A. Keim https://orcid.org/0000-0003-3490-3649
Data Availability Statement
Data may be available upon request from the Corresponding author.*
References
- 1.Shah NS, Wang MC, Freaney PM, et al. Trends in gestational diabetes at first live birth by race and ethnicity in the US, 2011-2019. JAMA. 2021;326(7):660-669. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Dennison RA, Chen ES, Green ME, et al. The absolute and relative risk of type 2 diabetes after gestational diabetes: a systematic review and meta-analysis of 129 studies. Diabetes Res Clin Pract. 2021;171:108625. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Vounzoulaki E, Khunti K, Abner SC, Tan BK, Davies MJ, Gillies CL. Progression to type 2 diabetes in women with a known history of gestational diabetes: systematic review and meta-analysis. BMJ. 2020;369:m1361. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Cui Y, Marshall A, Tsaih SW, Palatnik A. Impact of prior gestational diabetes on long-term type 2 diabetes complications. J Diabetes Complications. 2022;36(9):108282. [DOI] [PubMed] [Google Scholar]
- 5.Ehrlich S, Hedderson M, Quesenberry JC, et al. Post-partum weight loss and glucose metabolism in women with gestational diabetes: the DEBI Study. Diabet Med. 2014;31(7):862-867. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Gore SA, Brown DM, West DS. The role of postpartum weight retention in obesity among women: a review of the evidence. Ann Behav Med. 2003;26(2):149-159. [DOI] [PubMed] [Google Scholar]
- 7.Lim K, Chan S-Y, Lim SL, et al. A smartphone app to restore optimal weight (SPAROW) in women with recent gestational diabetes mellitus: randomized controlled trial. JMIR Mhealth Uhealth. 2021;9(3):e22147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Reinhardt JA, Van Der Ploeg HP, Grzegrzulka R, Timperley JG. Implementing lifestyle change through phone-based motivational interviewing in rural-based women with previous gestational diabetes mellitus. Health Promot J Austr. 2012;23(1):5-9. [DOI] [PubMed] [Google Scholar]
- 9.Bijlholt M, Ameye L, Van Uytsel H, Devlieger R, Bogaerts A. The INTER-ACT E-health supported lifestyle intervention improves postpartum food intake and eating behavior, but not physical activity and sedentary behavior—a randomized controlled trial. Nutrients. 2021;13(4):1287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tandon N, Gupta Y, Kapoor D, et al. Effects of a lifestyle intervention to prevent deterioration in glycemic status among south Asian women with recent gestational diabetes: a randomized clinical trial. JAMA Netw Open. 2022;5(3):e220773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jelsma JG, van Poppel MN, Smith BJ, et al. Changing psychosocial determinants of physical activity and diet in women with a history of gestational diabetes mellitus. Diabetes Metab Res Rev. 2018;34(1):e2942. [DOI] [PubMed] [Google Scholar]
- 12.Dennison RA, Griffin SJ, Usher-Smith JA, Fox RA, Aiken CE, Meek CL. Post-GDM support would be really good for mothers: a qualitative interview study exploring how to support a healthy diet and physical activity after gestational diabetes. PLoS One. 2022;17(1):e0262852. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Simmons D, Devlieger R, Van Assche A, et al. Effect of physical activity and/or healthy eating on GDM risk: the DALI lifestyle study. J Clin Endocrinol Metab. 2017;102(3):903-913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Kunath J, Günther J, Rauh K, et al. Effects of a lifestyle intervention during pregnancy to prevent excessive gestational weight gain in routine care–the cluster-randomised GeliS trial. BMC Med. 2019;17(1):1-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Koivusalo SB, Rönö K, Klemetti MM, et al. Gestational diabetes mellitus can be prevented by lifestyle intervention: the Finnish Gestational Diabetes Prevention Study (RADIEL): a randomized controlled trial. Diabetes Care. 2016;39(1):24-30. [DOI] [PubMed] [Google Scholar]
- 16.Guo XY, Shu J, Fu XH, et al. Improving the effectiveness of lifestyle interventions for gestational diabetes prevention: a meta-analysis and meta-regression. BJOG. 2019;126(3):311-320. [DOI] [PubMed] [Google Scholar]
- 17.Lin X, Yang T, Zhang X, Wei W. Lifestyle intervention to prevent gestational diabetes mellitus and adverse maternal outcomes among pregnant women at high risk for gestational diabetes mellitus. J Int Med Res. 2020;48(12):0300060520979130. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Horn CE, Seely EW, Levkoff SE, Isley BC, Nicklas JM. Postpartum women's experiences in a randomized controlled trial of a web-based lifestyle intervention following Gestational Diabetes: a qualitative study. J Matern Fetal Neonatal Med. 2023;36(1):2194012. [DOI] [PubMed] [Google Scholar]
- 19.Maisano MS, Shonkoff ET, Folta SC. Multiple health behavior change for weight loss: a scoping review. Am J Health Behav. 2020;44(5):559-571. [DOI] [PubMed] [Google Scholar]
- 20.Dalrymple KV, Uwhubetine O, Flynn AC, et al. Modifiable determinants of postpartum weight loss in women with obesity: a secondary analysis of the UPBEAT trial. Nutrients. 2021;13(6):1979. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Saligheh M, McNamara B, Rooney R. Perceived barriers and enablers of physical activity in postpartum women: a qualitative approach. BMC Pregnancy Childbirth. 2016;16(1):131. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Huifen Z, Yaping X, Meijing Z, et al. Effects of moderate-intensity resistance exercise on blood glucose and pregnancy outcome in patients with gestational diabetes mellitus: a randomized controlled trial. J Diabetes Complications. 2022;36(5):108186. [DOI] [PubMed] [Google Scholar]
- 23.Bassuk SS, Manson JE. Epidemiological evidence for the role of physical activity in reducing risk of type 2 diabetes and cardiovascular disease. J Appl Physiol. 2005;99(3):1193-1204. [DOI] [PubMed] [Google Scholar]
- 24.Stith BJ, Buls SM, Keim SA, et al. Moms in motion: weight loss intervention for postpartum mothers after gestational diabetes: a randomized controlled trial. BMC Pregnancy Childbirth. 2021;21(1):1-10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Carpenter MW, Coustan DR. Criteria for screening tests for gestational diabetes. Am J Obstet Gynecol. 1982;144(7):768-773. [DOI] [PubMed] [Google Scholar]
- 26.Diet History Questionnaire . National Institutes of Health, Applied Research Program. Jhajjar: National Cancer Institute. [Google Scholar]
- 27.Winkens B, van Breukelen GJ, Schouten HJ, Berger MP. Randomized clinical trials with a pre- and a post-treatment measurement: repeated measures versus ANCOVA models. Contemp Clin Trials. 2007;28(6):713-719. [DOI] [PubMed] [Google Scholar]
- 28.Heinsberg LW, Weeks DE. Post hoc power is not informative. Genet Epidemiol. 2022;46(7):390-394. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Wild R, Feingold KR. Effect of pregnancy on lipid metabolism and lipoprotein levels. In: Feingold KR, Anawalt B, Blackman MR, eds. Endotext. MDText.Com, Inc. Copyright © 2000-2024. Portland: MDText.com, Inc.; 2000. [PubMed] [Google Scholar]
- 30.He J, Hu K, Wang B, Wang H. Effect of dietary and physical activity behavioral interventions on reducing postpartum weight retention among women with recent gestational diabetes: a systematic review and meta-analysis of randomized controlled trials. Obes Rev. 2024;25(4):1-771. [DOI] [PubMed] [Google Scholar]
- 31.Dennison RA, Oliver-Williams C, Qi HLJ, et al. The effectiveness of pharmacological and lifestyle interventions to reduce the risk of diabetes and hyperglycaemia following gestational diabetes: a systematic review and meta-analysis. Diabet Med. 2024;41(6):e15316. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
Data may be available upon request from the Corresponding author.*