Abstract
Background
Gestational diabetes mellitus (GDM) presents clinically as either diet-controlled (diet-controlled gestational diabetes mellitus, A1GDM) or medication-controlled (insulin-requiring gestational diabetes mellitus, A2GDM) subtypes. Pre-pregnancy physical activity represents a significant modifiable factor in GDM prevention and management. Therefore, this study aimed to investigate the differential effects of physical activity on these two clinical GDM subtypes.
Methods
This secondary analysis utilized data from the Korean Pregnancy Outcome Study, a longitudinal multicenter cohort. From 3,457 singleton pregnant participants, physical activity was assessed at five time points from pre-pregnancy to postpartum using the International Physical Activity Questionnaire Short Form. Logistic regression models were used to estimate adjusted odds ratios (aORs) for GDM subtypes according to physical activity. Firth’s penalized logistic regression was used for A2GDM analysis.
Results
Of the participants, 231 developed GDM (overall prevalence 6.7%). Among these cases, 198 (85.7%) were diagnosed with A1GDM, and 33 (14.3%) with A2GDM. No significant differences in moderate-to-vigorous physical activity or muscle-strengthening activity were observed across three groups in baseline characteristics. However, Women who performed muscle-strengthening activity before pregnancy had a significantly lower risk of A1GDM (aOR, 0.492; 95% confidence interval, 0.262–0.923), with risk decreasing as activity frequency increased (P for trend = 0.014).
Conclusion
Pre-pregnancy muscle-strengthening activity is associated with significantly reduced risk of A1GDM. These findings highlight the importance of muscle-strengthening activity before conception as a potential strategy for GDM prevention.
Keywords: Diabetes, Gestational; Exercise; Longitudinal Studies; Pregnancy
Graphical Abstract
INTRODUCTION
Gestational diabetes mellitus (GDM) is characterized by impaired glucose metabolism during pregnancy, resulting in elevated blood glucose levels and reduced insulin sensitivity. Typically diagnosed between 24 and 28 weeks of gestation, GDM refers to glucose intolerance first identified during pregnancy.1 It affects up to 14.7% of pregnancies worldwide, with a prevalence exceeding 10% in Asia, a rate that continues to rise alongside increasing obesity and the aging maternal population.2,3 GDM poses significant short- and long-term health risks for both mothers and their offspring. In mothers, it increases the likelihood of complications such as cardiovascular disease, type 2 diabetes mellitus, macrosomia, and delivery-related challenges. In babies, GDM is associated with a heightened risk of preterm birth, neonatal hypoglycemia, and shoulder dystocia.4,5,6,7,8,9,10
Current prevention and management strategies for GDM primarily rely on lifestyle modifications, including dietary changes and physical activity, as no universally accepted pharmacological treatment exists. Although insulin therapy is sometimes used, its effectiveness can be limited by insulin resistance.11 Clinically, GDM is classified into two subtypes based on treatment requirements: diet-controlled gestational diabetes mellitus (A1GDM), which can be managed through nutritional therapy alone, and insulin-requiring gestational diabetes mellitus (A2GDM), which necessitates pharmacological intervention for glycemic control.12 Poorly managed A1GDM can progress to A2GDM, where insulin is used to control blood sugar. Furthermore, women diagnosed with A2GDM during pregnancy face a substantially increased risk of developing T2DM postpartum.11 Despite advances in understanding GDM, the underlying mechanisms and differential responses to treatment between these two clinical subtypes remain poorly understood.
International guidelines recommend that pregnant individuals engage in at least 150 minutes of moderate-intensity physical activity per week - activities such as walking, swimming, and low-impact aerobics are commonly suggested.13 Numerous studies have consistently reported the protective effects of adequate physical activity against GDM and its role in enhancing glycemic control.14,15,16,17,18 Our previous study further strengthened the evidence supporting the effectiveness of exercise in GDM prevention, demonstrating a significantly lower risk of GDM among women who participated in pre-pregnancy strength training.19 However, despite the distinct clinical management and treatment strategies for GDM subtypes, there remains a notable lack of research investigating the effects of exercise specifically for each GDM subtype. Therefore, the current study aims to extend our prior findings by stratifying GDM into its two major subtypes including A1GDM and A2GDM and longitudinally analyzing the associations between physical activity and the risk of each GDM subtype. This subtype-specific approach may contribute to more targeted preventive strategies and a deeper understanding of GDM pathophysiology.
METHODS
Study population and design
Data for this study were obtained from the Korean Pregnancy Outcome Study (KPOS), a longitudinal prospective cohort study focusing on adverse pregnancy outcomes, which was conducted at the two secondary hospitals in Seoul, Korea between March 2013 and January 2017.19 Of the 4,537 women enrolled in the cohort, 3,457 singleton pregnancies were included in the present analysis, after excluding women with pre-existing diabetes mellitus. GDM diagnosis followed a two-step screening protocol. Initial screening was performed using a 50 g glucose challenge test (GCT) between 24 and 28 weeks of gestation. Women with GCT values ≥ 140 mg/dL proceeded to a diagnostic oral glucose tolerance test. Detailed diagnostic criteria and thresholds are provided in our previous report.19 Women diagnosed with GDM were classified as A2GDM if insulin prescription was documented in their medical records during pregnancy, and as A1GDM otherwise. Based on pregnancy outcomes, participants were categorized into three groups: normal (n = 3,226), A1GDM (n = 198), and A2GDM (n = 33) (Fig. 1).
Fig. 1. Flow chart of study participant selection process.
A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-treated gestational diabetes mellitus, GDM = gestational diabetes mellitus.
Data were collected at four scheduled visits: visit 1 (approximately 12 gestational weeks [GW]), visit 2 (24 GW), visit 3 (36 GW), and visit 4 (4–6 weeks postpartum). A structured survey regarding pre-pregnancy lifestyle factors was administered during visit 1.
Demographic data
Demographic and clinical characteristics were obtained from the KPOS dataset. Socio-demographic information including maternal age, educational attainment, and household income and health-related behaviors such as smoking status, alcohol consumption, and dietary behavior were collected through structured interviews using questionnaires. Additional clinical information, including obstetric history (e.g., parity, history of GDM, history of previous macrosomia), medical history (e.g., hypertension and polycystic ovary syndrome [PCOS]), and first-degree family history of diabetes, was obtained via self-report and medical records. Dietary behavior was assessed using the Mini Dietary Assessment index which includes 10 items scored on a Likert scale (maximum 50 points), with higher scores indicating healthier dietary behaviors.20
Physical activity measurements
Physical activity was measured using the short-form International Physical Activity Questionnaire (IPAQ).21 Each domain assessed walking, moderate, and vigorous physical activities performed for at least 10 consecutive minutes daily over 7 days. The IPAQ was used to assess participants' physical activity at five time points: pre-pregnancy, visit 1 (first trimester), visit 2 (second trimester), visit 3 (third trimester), and postpartum. Analysis was restricted to subjects with complete IPAQ data at each time points, with missing responses excluded. The number of participants at each time point after accounting for missingness is presented in Fig. 1.
To analyze physical activity data as a continuous variable, metabolic equivalent task (MET) values were applied. These were calculated by multiplying the weekly duration (minutes/week) of walking (3.3 MET), moderate (4 MET), and vigorous (8 MET) activities by their respective MET values. The sum of walking, moderate, and vigorous METs was used to estimate moderate-to-vigorous physical activity (MVPA) levels. In addition to the standard International Physical Activity Questionnaire Short Form (IPAQ-SF) questionnaire items, we included supplementary questions to assess muscle-strengthening activities performed in the preceding 7 days (e.g., push-ups, sit-ups, dumbbell exercises, weightlifting, and barbell exercises). As in the previous study, this was counted as 5 METs.18,22
Next, the physical activity performance indicators were established based on the criteria outlined in the Korea Health Statistics 2020, and the IPAQ data collected in this study were used to generate performance attainment variables for statistical analysis.23
1. Walking performance: Walking at least 30 minutes per day for at least 10 minutes per session, at least 5 days per week.
2. MVPA performance: Engaging in 150 minutes of moderate-intensity physical activity, 75 minutes of vigorous-intensity physical activity, or a combination of moderate- and vigorous-intensity activities (1 minute of vigorous-intensity equals 2 minutes of moderate-intensity).
3. Muscle-strengthening activity performance: Performing muscle-strengthening activities at least 2 days in the past week.
Statistical analysis
All statistical analyses were performed using SPSS version 28.0 (IBM Corp., Armonk, NY, USA). Continuous variables are presented as means ± standard deviations or 95% confidence intervals (CI), and categorical variables as frequencies and percentages. One-way analysis of variance (ANOVA) followed by Bonferroni post hoc analysis was used to assess between-group differences in continuous variables. Pearson’s χ2 tests were used for categorical variables.
To estimate adjusted odds ratios (aORs) for the associations between physical activity and GDM subtypes, separate regression models were fitted for each outcome. For A1GDM, standard logistic regression was employed. For A2GDM, Firth’s penalized logistic regression was used to reduce small-sample bias and improve estimation stability. Both models were adjusted for the same covariates: maternal age, pre-pregnancy body mass index (BMI), education level, smoking status, hypertension, family history of diabetes, parity, and prior GDM. Dose-response relationships were examined in supplementary analyses by categorizing muscle-strengthening activity frequency (0, 1, ≥ 2 days/week) and testing for linear trends. Statistical significance was defined as a P value < 0.05 or a 95% CI that did not include 1.0.
Ethics statement
This study was approved by the Institutional Review Boards (IRBs) of Cheil General Hospital (IRB No. CGH-IRB-2013–10) and CHA Hospital (IRB No. 2013-14-KNC13–018), and informed consent was obtained from all participants at enrollment. The current analysis using data from KPOS was additionally approved by CHA Bundang Medical Center (IRB No. CHAMC 2021-04-052).
RESULTS
Study population characteristics
Among the 3,457 participants included in the analysis, 198 were diagnosed with A1GDM, and 33 with A2GDM, resulting in an overall GDM prevalence of 6.7% (Fig. 1). In the demographics of the study population, statistically significant differences were observed across the three groups. Variables showing differences included maternal age, pre-pregnancy BMI, educational attainment, household income, smoking status, history of hypertension, first-degree family history of diabetes, parity, history of GDM, and history of previous macrosomia (Table 1). Maternal age and pre-pregnancy BMI were higher in both GDM groups, with the highest averages observed in the A2GDM group. In contrast, employment status, alcohol consumption, dietary behavior index, and history of PCOS did not differ significantly between groups (Table 1).
Table 1. Baseline characteristics of the study population.
| Variables | Normal (n = 3,226) | A1GDM (n = 198) | A2GDM (n = 33) | P value | ||
|---|---|---|---|---|---|---|
| Sociodemographic characteristics | ||||||
| Maternal age, yr | 33.2 ± 3.7a | 34.2 ± 3.5b | 37.2 ± 3.5c | < 0.001 | ||
| Pre-pregnancy BMI, kg/m2 | 21.0 ± 2.8a | 23.1 ± 3.8b | 24.7 ± 3.9c | < 0.001 | ||
| Educational attainment | 0.007 | |||||
| High school or lower | 255 (7.9) | 22 (11.1) | 7 (21.2) | |||
| More than high school education | 2,971 (92.1) | 176 (88.9) | 26 (78.8) | |||
| Household income, KRW | 0.002 | |||||
| ≤ 4 million | 957 (29.7) | 58 (29.3) | 19 (57.6) | |||
| > 4 million | 2,269 (70.3) | 140 (70.7) | 14 (42.4) | |||
| Employment status | 0.060 | |||||
| Currently employed | 1,180 (36.6) | 89 (44.9) | 12 (36.4) | |||
| Currently unemployed | 2,046 (63.4) | 109 (55.1) | 21 (63.6) | |||
| Smoking status | < 0.001 | |||||
| Never smoking | 2,896 (89.8) | 175 (88.4) | 20 (60.6) | |||
| Smoking before pregnancy | 256 (7.9) | 17 (8.6) | 8 (24.2) | |||
| Smoking until early pregnancy | 74 (2.3) | 6 (3.0) | 5 (15.2) | |||
| Alcohol consumption | 0.523 | |||||
| Never drank | 624 (19.3) | 43 (21.7) | 3 (9.1) | |||
| Drinker | 2,602 (80.6) | 155 (78.3) | 30 (90.9) | |||
| Dietary behavior indexd | 33.3 ± 4.3 | 32.6 ± 4.0 | 32.8 ± 5.8 | 0.108 | ||
| Medical and family history | ||||||
| History of hypertension | < 0.001 | |||||
| No | 3,191 (98.9) | 190 (96.0) | 31 (93.9) | |||
| Yes | 35 (1.1) | 8 (4.0) | 2 (6.1) | |||
| History of PCOS | 0.213 | |||||
| No | 3,177 (98.5) | 192 (97.0) | 32 (97.0) | |||
| Yes | 49 (1.5) | 6 (3.0) | 1 (3.0) | |||
| First-degree family history of diabetes | < 0.001 | |||||
| No | 2,589 (80.3) | 136 (68.7) | 11 (33.3) | |||
| Yes | 637 (19.7) | 62 (31.3) | 22 (66.7) | |||
| Obstetric history | ||||||
| Parity | 0.008 | |||||
| 0 | 1,940 (60.1) | 97 (49.0) | 19 (57.6) | |||
| ≥ 1 | 1,286 (39.9) | 101 (51.0) | 14 (42.4) | |||
| History of GDMe | < 0.001 | |||||
| No | 1,246 (96.9) | 87 (86.1) | 12 (85.7) | |||
| Yes | 40 (3.1) | 14 (13.9) | 2 (14.3) | |||
| History of previous macrosomiae | < 0.001 | |||||
| No | 1,244 (96.7) | 88 (87.1) | 12 (85.7) | |||
| Yes | 42 (3.3) | 13 (12.9) | 2 (14.3) | |||
Continuous variables are presented as mean ± standard deviation; categorical variables are presented as number (%).
A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-requiring gestational diabetes mellitus, BMI = body mass index, PCOS = polycystic ovary syndrome, GDM = gestational diabetes mellitus.
a,b,cSame superscripts indicate that those groups are not significantly different from each other based on the Bonferroni post hoc test.
dThere is a maximum score of 50, with higher scores indicating healthier eating habits.24
eAnalyzed participants with parity 1 or higher.
Physical activity assessment
Physical activity intensity data, presented as MET-minutes per week, are summarized in Table 2. No statistically significant differences in MVPA or muscle-strengthening activity levels were identified across the three groups throughout the pre-pregnancy, pregnancy, and postpartum periods. However, pre-pregnancy muscle-strengthening activity was lower in both A1GDM and A2GDM groups compared to the normal group, although the difference did not reach statistical significance (P = 0.050) (Table 2).
Table 2. Physical activity intensity (METs-min/week) among normal, A1GDM, and A2GDM groups from pre-pregnancy to postpartum.
| Variables | Normal | A1GDM | A2GDM | P value | |
|---|---|---|---|---|---|
| Pre-pregnancy | n = 3,114 | n = 194 | n = 32 | ||
| MVPAa | 1,111.0 (1,056.5–1,165.5) | 1,009.4 (803.5–1,215.4) | 1,100.2 (673.4–1,527.0) | 0.673 | |
| Muscle-strengthening activityb | 73.4 (63.1–83.7) | 26.0 (8.1–43.8) | 21.5 (−9.9–52.9) | 0.050 | |
| Visit 1 (1st trimester) | n = 3,114 | n = 194 | n = 32 | ||
| MVPAa | 489.8 (460.4–519.2) | 434.7 (334.5–535.0) | 761.0 (226.8–1,295.2) | 0.123 | |
| Muscle-strengthening activityb | 5.9 (1.3–10.6) | 3.0 (−2.0–8.0) | 19.9 (−20.7–60.4) | 0.788 | |
| Visit 2 (2nd trimester) | n = 2,990 | n = 188 | n = 33 | ||
| MVPAa | 784.3 (740.6–827.9) | 745.0 (588.4–901.5) | 829.5 (740.7–824.2) | 0.888 | |
| Muscle-strengthening activityb | 6.49 (1.48–11.5) | 0 | 0 | 0.788 | |
| Visit 3 (3rd trimester) | n = 2,761 | n = 172 | n = 31 | ||
| MVPAa | 718.7 (678.4–759.0) | 736.6 (605.2–868.1) | 808.9 (337.9–1,280.0) | 0.879 | |
| Muscle-strengthening activityb | 8.2 (5.6–10.7) | 12.6 (−4.2–29.4) | 24.8 (−17.7–67.3) | 0.343 | |
| Postpartum | n = 2,288 | n = 141 | n = 25 | ||
| MVPAa | 340.0 (297.8–382.2) | 265.9 (102.7–429.2) | 265.6 (20.0–511.3) | 0.666 | |
| Muscle-strengthening activityb | 10.0 (6.3–13.7) | 5.8 (−1.0–12.7) | 6.4 (−6.8–19.5) | 0.842 | |
Values are presented as the average of MET-min/week (95% confidence interval). The P values were obtained using a one-way ANOVA test, Bonferroni post-hoc analysis.
MET = metabolic equivalent task, A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-requiring gestational diabetes mellitus, MVPA = moderate to vigorous physical activity.
aWalking, moderate and vigorous intensity are multiplied by 3.3 MET, 4 MET, and 8 MET respectively and then summed for all using International Physical Activity Questionnaire Short Form.
bDerived by multiplying the value of 5.0 MET for the Compendium of Physical Activities.22
Table 3 presents the number of participants who met the key performance indicators for physical activity. No significant differences were observed in walking or MVPA performance across the five time points (pre-pregnancy, first trimester, second trimester, third trimester, and postpartum). However, the proportion of participants who met muscle-strengthening activity during the pre-pregnancy period was significantly lower in the A1GDM (5.7%) and A2GDM (3.1%) groups than in the normal group (11.7%) (P = 0.013) (Table 3).
Table 3. The number of participants achieving the key performance indicators of physical activity among normal, A1GDM, and A2GDM groups from pre-pregnancy to postpartum.
| Variables | Normal | A1GDM | A2GDM | P value | |
|---|---|---|---|---|---|
| Pre-pregnancy | n = 3,116 | n = 194 | n = 32 | ||
| Walking performancea | 1,387 (44.5) | 70 (36.1) | 12 (37.5) | 0.054 | |
| MVPA performanceb | 1,891 (60.7) | 112 (57.7) | 23 (71.9) | 0.303 | |
| Muscle-strengthening activity performancec | 363 (11.7) | 11 (5.7) | 1 (3.1) | 0.013 | |
| Visit 1 (1st trimester) | n = 3,132 | n = 196 | n = 32 | ||
| Walking performancea | 789 (25.2) | 41 (20.9) | 7 (21.9) | 0.375 | |
| MVPA performanceb | 1,060 (33.8) | 56 (28.6) | 14 (43.8) | 0.151 | |
| Muscle-strengthening activity performancec | 27 (0.9) | 2 (1.0) | 1 (3.1) | 0.392 | |
| Visit 2 (2nd trimester) | n = 2,990 | n = 188 | n = 33 | ||
| Walking performancea | 949 (31.7) | 49 (26.1) | 11 (33.3) | 0.259 | |
| MVPA performanceb | 1,494 (50.0) | 80 (42.6) | 17 (51.5) | 0.139 | |
| Muscle-strengthening activity performancec | 36 (1.2) | 0 | 0 | 0.260 | |
| Visit 3 (3rd trimester) | n = 2,761 | n = 172 | n = 31 | ||
| Walking performancea | 777 (28.1) | 59 (34.3) | 11 (35.5) | 0.154 | |
| MVPA performanceb | 1,275 (46.2) | 81 (47.1) | 14 (45.2) | 0.966 | |
| Muscle-strengthening activity performancec | 47 (1.7) | 2 (1.2) | 2 (6.5) | 0.109 | |
| Postpartum | n = 2,288 | n = 141 | n = 25 | ||
| Walking performancea | 178 (7.8) | 6 (4.3) | 4 (16.0) | 0.090 | |
| MVPA performanceb | 340 (14.9) | 18 (12.8) | 5 (20.0) | 0.605 | |
| Muscle-strengthening activity performancec | 52 (2.3) | 3 (2.1) | 1 (4.0) | 0.841 | |
Values are presented as the number (percentage) of subjects who met each performance indicators of physical activity. P value was determined using Pearson χ2 test.
A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-requiring gestational diabetes mellitus, MVPA = moderate to vigorous physical activity.
aWalking performance: Walking at least 30 minutes per day for at least 10 minutes per session, at least 5 days per week.
bMVPA performance: Participating in 150 minutes of moderate-intensity physical activity or 75 minutes of vigorous-intensity physical activity, or a combination of moderate- and vigorous-intensity activities (1 minute of high-intensity equals 2 minutes of moderate-intensity) per week.
cMuscle-strengthening activity performance: at least 2 days in the past week.
Association between GDM subtype and physical activity
Separate logistic regression models were performed to evaluate whether pre-pregnancy physical activity was independently associated with each GDM subtype, adjusting for maternal age, pre-pregnancy BMI, educational attainment, smoking status, history of hypertension, first-degree family history of diabetes, parity, and history of GDM (Table 4). For A1GDM, meeting the walking performance indicator showed a protective association in crude analysis (odds ratio [OR], 0.703; 95% CI, 0.520–0.951; P = 0.022), but this association was attenuated after adjustment for covariates (aOR, 0.773; 95% CI, 0.567–1.054; P = 0.104). Meeting MVPA performance criteria was not significantly associated with A1GDM risk in either crude or adjusted models. In contrast, pre-pregnancy engagement in muscle-strengthening activities was associated with a significantly lower risk of A1GDM in both crude (OR, 0.456; 95% CI, 0.246–0.846; P = 0.013) and adjusted models (aOR, 0.492; 95% CI. 0.262–0.923; P = 0.027).
Table 4. ORs and 95% CIs for the risk of A1GDM and A2GDM by meeting key performance indicators of physical activity in pre-pregnancy.
| Variables | Pre-pregnancy period | GDM, No. (%) | Crude OR (95% CI) | P value | Adjusted ORf (95% CI) | P value |
|---|---|---|---|---|---|---|
| A1GDMa (n = 3,310) | Walkingc | |||||
| Met (n = 1,457) | 70 (4.8) | 0.703 (0.520–0.951) | 0.022 | 0.773 (0.567–1.054) | 0.104 | |
| Non-met (n = 1,853) | 124 (6.7) | 1 (ref) | 1 (ref) | |||
| MVPAd | ||||||
| Met (n = 2,003) | 112 (5.6) | 0.885 (0.660–1.187) | 0.414 | 0.976 (0.720–1.323) | 0.876 | |
| Non-met (1,307) | 82 (6.3) | 1 (ref) | 1 (ref) | |||
| Muscle-strengthening activitye | ||||||
| Met (n = 374) | 11 (2.9) | 0.456 (0.246–0.846) | 0.013 | 0.492 (0.262–0.923) | 0.027 | |
| Non-met (n = 2,936) | 183 (6.2) | 1 (ref) | 1 (ref) | |||
| A2GDMb (n = 3,148) | Walkingc | |||||
| Met (n = 1,399) | 12 (0.9) | 0.760 (0.365–1.521) | 0.442 | 0.944 (0.434–1.979) | 0.881 | |
| Non-met (n = 1,749) | 20 (1.1) | 1 (ref) | 1 (ref) | |||
| MVPAd | ||||||
| Met (n = 1,914) | 23 (1.2) | 1.601 (0.775–3.588) | 0.209 | 1.656 (0.769–3.844) | 0.202 | |
| Non-met (n = 1,234) | 9 (0.7) | 1 (ref) | 1 (ref) | |||
| Muscle-strengthening activitye | ||||||
| Met (n = 364) | 1 (0.3) | 0.361 (0.040–1.366) | 0.153 | 0.246 (0.026–1.036) | 0.057 | |
| Non-met (n = 2,784) | 31 (1.1) | 1 (ref) | 1 (ref) |
OR = odds ratio, CI = confidence interval, A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-requiring gestational diabetes mellitus, GDM = gestational diabetes mellitus, ref = reference category, MVPA = moderate-to-vigorous physical activity.
aStandard logistic regression was used for normal vs. A1GDM analysis.
bFirth’s penalized logistic regression was used for normal vs A2GDM analysis to address small sample bias and quasi-complete separation.
cWalking performance: Walking at least 30 minutes per day for at least 10 minutes per session, at least 5 days per week.
dMVPA performance: Participating in 150 minutes of moderate-intensity physical activity or 75 minutes of vigorous-intensity physical activity, or a combination of moderate- and vigorous-intensity activities (1 minute of high-intensity equals 2 minutes of moderate-intensity) per week.
eMuscle-strengthening activity performance: at least 2 days in the past week.
fAdjusted for maternal age, pre-pregnancy body mass index, educational attainment, household income, smoking status, history of hypertension, first-degree family history of diabetes, parity and history of GDM.
For A2GDM, Firth’s penalized logistic regression was employed. None of the three physical activity performance indicators showed significant associations with A2GDM risk: walking (aOR, 0.944; 95% CI, 0.434–1.979; P = 0.881), MVPA (aOR, 1.656; 95% CI, 0.769–3.844; P = 0.202), and muscle-strengthening activities (aOR, 0.246; 95% CI. 0.026–1.036; P = 0.057). Supplementary dose-response analysis examining the frequency of pre-pregnancy muscle-strengthening activities (0, 1, ≥ 2 days/week) revealed a significant linear trend for A1GDM (P for trend = 0.014), supporting a dose-dependent protective effect. No such trend was observed for A2GDM (P for trend = 0.183) (Table 5).
Table 5. Dose-response association between pre-pregnancy muscle-strengthening activity and GDM subtypes.
| Pre-pregnancy muscle-strengthening activity frequency | No. | GDM, No. (%) | Adjusted OR (95% CI)a | P value | |
|---|---|---|---|---|---|
| A1GDM (n = 3,310) | |||||
| 0 days/wk | 2,857 | 181 (6.3) | 1.000 (ref) | ||
| 1 day/wk | 79 | 2 (2.5) | 0.502 (0.121–2.083) | 0.342 | |
| ≥ 2 days/wk | 374 | 11 (2.9) | 0.473 (0.253–0.886) | 0.019 | |
| P for linear trendb | 0.678 (0.497–0.925) | 0.014 | |||
| A2GDM (n = 3,148) | |||||
| 0 days/wk | 2,705 | 29 (1.1) | 1.000 (ref) | ||
| 1 day/wk | 79 | 2 (2.5) | 4.939 (0.904–17.976) | 0.063 | |
| ≥ 2 days/wk | 364 | 1 (0.3) | 0.271 (0.029–1.133) | 0.078 | |
| P for linear trendb | 0.649 (0.279–1.196) | 0.183 | |||
GDM = gestational diabetes mellitus, OR = odds ratio, CI = confidence interval, A1GDM = diet-controlled gestational diabetes mellitus, A2GDM = insulin-requiring gestational diabetes mellitus.
aAdjusted for age, pre-pregnancy body mass index, education level, household income, smoking status, chronic hypertension, family history of diabetes, and parity. Standard logistic regression was used for A1GDM analysis; Firth’s penalized logistic regression was used for A2GDM analysis due to small sample size (32 events).
bLinear trend test treating frequency as an ordinal variable; OR represents per one-category increase in activity frequency.
DISCUSSION
This secondary analysis of prospective cohort data evaluated the association between physical activity and GDM subtypes. Our findings indicate that pre-pregnancy muscle-strengthening activity was significantly associated with a reduced risk of A1GDM, with a clear dose-response relationship. In contrast, no significant associations were observed between any physical activity indicators and A2GDM risk. Walking performance showed a protective trend for A1GDM in crude analysis, but this association was attenuated after adjustment for covariates.
Previous research has consistently demonstrated the protective effects of physical activity in preventing and managing GDM. International guidelines recommend at least 150 minutes of moderate-to-vigorous physical activity per week during pregnancy, primarily emphasizing aerobic activities such as walking, swimming, and low-impact aerobics.13,24,25 In our cohort, the proportion of women who met the recommended threshold varied substantially from pre-pregnancy to postpartum (12.8–71.9%), yet this adherence was not protective against either GDM subtype. A recent meta-analysis examining the role of lifestyle in the prevention of gestational diabetes also reported mixed results regarding the intensity of physical activity by trimester; specifically, light to moderate activity was effective in the first trimester, whereas moderate-to-vigorous exercise in the second trimester showed limited effectiveness.26 This suggests the need for subtype-specific, individualized guidance to prevent and manage GDM.
In contrast, Women who engaged in pre-pregnancy muscle-strengthening activities experienced approximately half the risk of A1GDM compared to those who did not. This finding aligns with emerging evidence suggesting that muscle-strengthening activities may be particularly effective in reducing GDM risk,16,27 potentially through distinct mechanisms compared to aerobic exercise. Given that GDM shares several pathophysiological characteristics with type 2 diabetes mellitus, muscle-strengthening activity may help reduce risk through improvements in glycemic control, insulin sensitivity, adiposity, muscular strength, and body composition.28 Supporting this mechanism, a recent Korean prospective study demonstrated that women with GDM who had higher muscle mass showed significantly lower insulin resistance after delivery and reduced progression to postpartum type 2 diabetes mellitus, independent of body fat mass.29 This finding reinforces the critical role of skeletal muscle as a metabolically active organ in glucose metabolism and diabetes prevention.
However, no significant association was observed for A2GDM. This null finding should be interpreted with caution as it may primarily reflect limited statistical power due to the small number of A2GDM cases rather than a true absence of effect. The wide CIs observed for A2GDM analyses reflect the statistical limitations inherent to small sample sizes. While it is possible that A1GDM and A2GDM differ in their responsiveness to physical activity interventions, our study cannot definitively address this question. These findings should be considered hypothesis-generating observations that require validation in larger studies with sufficient power to detect associations in rarer outcomes such as A2GDM.
These findings have important clinical implications for GDM prevention. Pre-pregnancy muscle-strengthening activity appears to reduce A1GDM risk and may inform the development of prevention strategies. The protective mechanism may partly operate through improved glucose regulation before conception. Recent evidence shows that even modest elevations in pre-pregnancy fasting glucose within the normal range are associated with increased GDM risk.30 This suggests that metabolic benefits of pre-pregnancy muscle-strengthening activity may begin before pregnancy and contribute to GDM prevention through enhanced glycemic control. Current guidelines provide limited specific recommendations regarding muscle-strengthening exercise for women planning pregnancy. Our study suggests that initiating muscle-strengthening activity before conception may be particularly beneficial for preventing A1GDM. Further research is needed to establish optimal exercise protocols and to clarify whether similar benefits extend to A2GDM prevention. Pre-pregnancy counseling should consider incorporating muscle-strengthening activity as part of comprehensive GDM prevention efforts.
This study has several strengths, including the use of a large prospective cohort and comprehensive physical activity data collected longitudinally from pre-pregnancy through postpartum. By assessing physical activity across multiple time points, we were able to examine both the timing and type of activity in relation to GDM risk. Furthermore, stratifying outcomes by GDM subtype offers novel insights into potential differential effects. The observed dose-response relationship for muscle-strengthening activity and A1GDM strengthens the evidence for a potential protective effect and demonstrates consistency across different levels of exposure.
However, several limitations should be noted. First, as an observational study, our findings demonstrate associations but cannot establish causality, even though the prospective design supports temporality. Second, pre-pregnancy physical activity was assessed retrospectively during early pregnancy using the self-reported IPAQ-SF, which may be subject to recall bias and over-reporting. Self-reported measures may have led to misclassification of physical activity physical activity performance status. Third, the small number of A2GDM cases resulted in limited statistical power and wide CIs that reflect substantial uncertainty in the effect estimates. The null findings for A2GDM should be interpreted with caution and do not necessarily indicate the absence of an association. Fourth, missing data and participant attrition across time points may have affected the precision of our estimates. Finally, the study did not explore the combined effects of physical activity with other modifiable lifestyle factors. Future studies employing objective measures of physical activity and comprehensive lifestyle assessments are needed to validate and expand on these findings.
In conclusion, pre-pregnancy engagement in muscle-strengthening activity was associated with a reduced risk of A1GDM in a dose-dependent manner, supporting its potential role as a target for GDM prevention. The lack of significant association for A2GDM likely reflects limited statistical power rather than a true null effect, though this requires confirmation in larger studies. These findings emphasize the importance of pre-pregnancy physical activity, particularly Muscle-strengthening activity, in GDM prevention strategies. Future research should employ larger sample sizes, objective activity measures, and prospective intervention designs to further clarify the causal relationships and subtype-specific effects of physical activity on GDM risk.
ACKNOWLEDGMENTS
We express our sincere gratitude to the Division of Maternal-Fetal Medicine at Cheil General Hospital and CHA Hospital for their involvement in the KPOS research (Korea National Institute of Health research project: 2015-E6302-00).
Footnotes
Disclosure of Artificial Intelligence (AI)-Assisted Technology: During the preparation of this manuscript, the authors used Claude 4.5 (Anthropic) for English language editing and grammar checking. The authors reviewed and edited the output and take full responsibility for the content of this publication.
Funding: This work was supported by grants from the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health and Welfare, Republic of Korea (RS-2022-KH129953), and the National Institute of Health (NIH) research project (project No. 2024-ER1104-01).
Disclosure: The authors have no potential conflicts of interest to disclose.
- Conceptualization: Ryu HM, Lim JH.
- Data curation: Lim JH, Ryu HM, Kim JW, Yang SJ, Kim JY, Choi YJ, Lee HJ, Kim NR, Jang JH, Kwak M.
- Formal analysis: Han HJ, Han YJ, Kwak DW, Chung JH.
- Funding acquisition: Ryu HM.
- Methodology: Ryu HM, Lim JH, Han HJ, Kim JW.
- Software: Han HJ.
- Validation: Ryu HM, Lim JH, Lee HJ.
- Visualization: Han HJ, Kim TH.
- Writing - original draft: Kim TH, Han HJ.
- Writing - review & editing: Ryu HM, Lim JH, Han HJ, Lee HJ, Kim TH.
References
- 1.Quintanilla Rodriguez BS, Vadakekut ES, Mahdy H. StatPearls. Treasure Island, FL, USA: StatPearls Publishing; 2024. Gestational diabetes. [Google Scholar]
- 2.Saeedi M, Cao Y, Fadl H, Gustafson H, Simmons D. Increasing prevalence of gestational diabetes mellitus when implementing the IADPSG criteria: a systematic review and meta-analysis. Diabetes Res Clin Pract. 2021;172:108642. doi: 10.1016/j.diabres.2020.108642. [DOI] [PubMed] [Google Scholar]
- 3.Zhou T, Du S, Sun D, Li X, Heianza Y, Hu G, et al. Prevalence and trends in gestational diabetes mellitus among women in the United States, 2006–2017: a population-based study. Front Endocrinol (Lausanne) 2022;13:868094. doi: 10.3389/fendo.2022.868094. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Teh WT, Teede HJ, Paul E, Harrison CL, Wallace EM, Allan C The WT. Risk factors for gestational diabetes mellitus: implications for the application of screening guidelines. Aust N Z J Obstet Gynaecol. 2011;51(1):26–30. doi: 10.1111/j.1479-828X.2011.01292.x. [DOI] [PubMed] [Google Scholar]
- 5.Petry CJ. Gestational diabetes: risk factors and recent advances in its genetics and treatment. Br J Nutr. 2010;104(6):775–787. doi: 10.1017/S0007114510001741. [DOI] [PubMed] [Google Scholar]
- 6.Caughey AB, Cheng YW, Stotland NE, Washington AE, Escobar GJ. Maternal and paternal race/ethnicity are both associated with gestational diabetes. Am J Obstet Gynecol. 2010;202(6):616.e1–616.e5. doi: 10.1016/j.ajog.2010.01.082. [DOI] [PubMed] [Google Scholar]
- 7.Solomon CG, Willett WC, Carey VJ, Rich-Edwards J, Hunter DJ, Colditz GA, et al. A prospective study of pregravid determinants of gestational diabetes mellitus. JAMA. 1997;278(13):1078–1083. [PubMed] [Google Scholar]
- 8.Bouthoorn SH, Silva LM, Murray SE, Steegers EA, Jaddoe VW, Moll H, et al. Low-educated women have an increased risk of gestational diabetes mellitus: the Generation R Study. Acta Diabetol. 2015;52(3):445–452. doi: 10.1007/s00592-014-0668-x. [DOI] [PubMed] [Google Scholar]
- 9.Sohn K. Parents are rapidly getting older in South Korea. Hum Fertil (Camb) 2017;20(3):212–216. doi: 10.1080/14647273.2017.1279352. [DOI] [PubMed] [Google Scholar]
- 10.Amiri FN, Faramarzi M, Bakhtiari A, Omidvar S. Risk factors for gestational diabetes mellitus: a case-control study. Am J Lifestyle Med. 2018;15(2):184–190. doi: 10.1177/1559827618791980. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Plows JF, Stanley JL, Baker PN, Reynolds CM, Vickers MH. The pathophysiology of gestational diabetes mellitus. Int J Mol Sci. 2018;19(11):3342. doi: 10.3390/ijms19113342. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Ehrenberg HM, Mercer BM, Catalano PM. The influence of obesity and diabetes on the prevalence of macrosomia. Am J Obstet Gynecol. 2004;191(3):964–968. doi: 10.1016/j.ajog.2004.05.052. [DOI] [PubMed] [Google Scholar]
- 13.Physical activity and exercise during pregnancy and the postpartum period: ACOG committee opinion, number 804. Obstet Gynecol. 2020;135(4):e178–e188. doi: 10.1097/AOG.0000000000003772. [DOI] [PubMed] [Google Scholar]
- 14.Russo LM, Nobles C, Ertel KA, Chasan-Taber L, Whitcomb BW. Physical activity interventions in pregnancy and risk of gestational diabetes mellitus: a systematic review and meta-analysis. Obstet Gynecol. 2015;125(3):576–582. doi: 10.1097/AOG.0000000000000691. [DOI] [PubMed] [Google Scholar]
- 15.Mijatovic-Vukas J, Capling L, Cheng S, Stamatakis E, Louie J, Cheung NW, et al. Associations of diet and physical activity with risk for gestational diabetes mellitus: a systematic review and meta-analysis. Nutrients. 2018;10(6):698. doi: 10.3390/nu10060698. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Davenport MH, Ruchat SM, Poitras VJ, Jaramillo Garcia A, Gray CE, Barrowman N, et al. Prenatal exercise for the prevention of gestational diabetes mellitus and hypertensive disorders of pregnancy: a systematic review and meta-analysis. Br J Sports Med. 2018;52(21):1367–1375. doi: 10.1136/bjsports-2018-099355. [DOI] [PubMed] [Google Scholar]
- 17.Martínez-Vizcaíno V, Sanabria-Martínez G, Fernández-Rodríguez R, Cavero-Redondo I, Pascual-Morena C, Álvarez-Bueno C, et al. Exercise during pregnancy for preventing gestational diabetes mellitus and hypertensive disorders: an umbrella review of randomised controlled trials and an updated meta-analysis. BJOG. 2023;130(3):264–275. doi: 10.1111/1471-0528.17304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Lim JH, Kim MH, Han HJ, Yang SJ, Kim MG, Lee HJ, et al. The preventive effect of physical activity on gestational diabetes mellitus: a korean longitudinal prospective cohort study. Diabetes Metab J. 2025;49(5):1116–1125. doi: 10.4093/dmj.2024.0280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Choi H, Kwak DW, Kim MH, Lee SY, Chung JH, Han YJ, et al. The Korean Pregnancy Outcome Study (KPOS): study design and participants. J Epidemiol. 2021;31(6):392–400. doi: 10.2188/jea.JE20200055. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kim WY, Cho MS, Lee HS. Development and validation of mini dietary assessment index for Koreans. Korean J Nutr. 2003;36(1):83–92. [Google Scholar]
- 21.IPAQ. Guidelines for Data Processing and Analysis of the International Physical Activity Questionnaire (IPAQ) – Short and Long Forms. Revised November 2005. [Updated 2025]. [Accessed April 22, 2026]. https://sites.google.com/view/ipaq/score .
- 22.Herrmann SD, Willis EA, Ainsworth BE, Barreira TV, Hastert M, Kracht CL, et al. 2024 Adult Compendium of Physical Activities: a third update of the energy costs of human activities. J Sport Health Sci. 2024;13(1):6–12. doi: 10.1016/j.jshs.2023.10.010. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Korea Health Statistics. Korea National Health and Nutrition Examination Survey. [Updated 2020]. [Accessed April 22, 2026]. http://knhanes.kdca.go.kr/
- 24.Savvaki D, Taousani E, Goulis DG, Tsirou E, Voziki E, Douda H, et al. Guidelines for exercise during normal pregnancy and gestational diabetes: a review of international recommendations. Hormones (Athens) 2018;17(4):521–529. doi: 10.1007/s42000-018-0085-6. [DOI] [PubMed] [Google Scholar]
- 25.Evenson KR, Barakat R, Brown WJ, Dargent-Molina P, Haruna M, Mikkelsen EM, et al. Guidelines for physical activity during pregnancy: comparisons from around the world. Am J Lifestyle Med. 2014;8(2):102–121. doi: 10.1177/1559827613498204. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Altemani AH, Alzaheb RA. The prevention of gestational diabetes mellitus (The role of lifestyle): a meta-analysis. Diabetol Metab Syndr. 2022;14(1):83. doi: 10.1186/s13098-022-00854-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Xie Y, Zhao H, Zhao M, Huang H, Liu C, Huang F, et al. Effects of resistance exercise on blood glucose level and pregnancy outcome in patients with gestational diabetes mellitus: a randomized controlled trial. BMJ Open Diabetes Res Care. 2022;10(2):e002622. doi: 10.1136/bmjdrc-2021-002622. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Colberg SR, Sigal RJ, Yardley JE, Riddell MC, Dunstan DW, Dempsey PC, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care. 2016;39(11):2065–2079. doi: 10.2337/dc16-1728. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Shin Y, Moon JH, Oh TJ, Ahn CH, Moon JH, Choi SH, et al. Higher muscle mass protects women with gestational diabetes mellitus from progression to type 2 diabetes mellitus. Diabetes Metab J. 2022;46(6):890–900. doi: 10.4093/dmj.2021.0334. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Jung YJ, Kim T, Kim YH. Elevated pre-pregnancy blood pressure and the risk of adverse pregnancy outcomes: evidence from a nationwide population-based study. J Korean Med Sci. 2025;40(46):e302. doi: 10.3346/jkms.2025.40.e302. [DOI] [PMC free article] [PubMed] [Google Scholar]