Associations of body composition and physical fitness with gestational diabetes and cardiovascular health in pregnancy: Results from the HealthyMoms trial

Pregnancy induces major cardiovascular adaptations, and maternal cardiovascular health during pregnancy predicts both perinatal outcomes and long-term maternal health. Gestational diabetes mellitus (GDM) and hypertensive disorders are linked to adverse maternal-infant outcomes and future cardiometabolic risk. While obesity is a strong risk factor, body mass index (BMI) is a crude proxy that cannot distinguish fat mass (FM) from fat-free mass (FFM). Limited evidence suggests FM relates to dysglycemia and insulin resistance in pregnancy, but the role of FFM is unclear. In addition, although cardiorespiratory fitness (CRF) is a robust health marker in non-pregnant populations and can mitigate obesity-related risk, its association with GDM and cardiovascular health during pregnancy is unknown. This study aimed to examine how body composition (FMI, %FM, FFMI, BMI) and physical fitness (CRF, handgrip strength) relate to GDM and cardiovascular health markers (glycemia, insulin resistance, blood pressure, and metabolic syndrome score) in early pregnancy.

Prior studies show that pregnancy cardiovascular health influences adverse outcomes and long-term risk. Obesity increases risk for GDM, gestational hypertension, and preeclampsia, yet BMI imperfectly reflects adiposity and cannot separate FM from FFM, which may have distinct health effects. A few pregnancy studies reported positive associations of FM with glycemia and insulin resistance, while evidence on FFM is conflicting, with some suggesting FFM predicts HOMA-IR comparably to FM and others showing weaker associations for FFM. Despite criticism of BMI as an adiposity measure, it often predicts cardiovascular risk reasonably well. CRF is consistently associated with better cardiometabolic profiles outside pregnancy and may buffer obesity-related risk, but no prior study had evaluated CRF or muscular strength in relation to GDM and cardiovascular health during pregnancy. These gaps motivated evaluating whether precise body composition measures outperform BMI and whether fitness independently relates to cardiometabolic health in pregnant women.

Design: Cross-sectional analysis of baseline data from the HealthyMoms randomized controlled trial (NCT03298555), conducted prior to randomization and intervention exposure. Participants: 303 pregnant women (of 305 measured; two excluded from CRF due to inability to complete test) recruited during routine first trimester visits (October 2017–March 2020) at maternity clinics in Linköping, Norrköping, and Motala, Sweden. Inclusion criteria: singleton pregnancy, age ≥18 years, Swedish language proficiency, informed consent. Baseline visit at mean 13.9 ± 0.7 gestational weeks after overnight fast. Data collection: fasting venous blood; body composition; physical fitness tests; questionnaires (age, pre-pregnancy weight/activity, parity, occupation, education). Measurements: Body composition and anthropometry: Height via stadiometer; weight and body composition by Bod Pod air-displacement plethysmography. Body density derived from body volume and weight. FM and FFM calculated using a two-compartment model with pregnancy-appropriate densities at gestational week 14 (FM 0.900 g/cm³; FFM 1.098 g/cm³; FFM density estimated per Most et al.). Derived indices: BMI (kg/m²), FMI (FM kg/height²), FFMI (FFM kg/height²), %FM. Physical fitness: Cardiorespiratory fitness assessed by 6-minute walk test (6MWT) distance in a 30 m corridor; average heart rate during test recorded as exertion (Polar M400). Muscular strength assessed by handgrip dynamometry (TKK 5001), with hand span adjusted; two trials per hand, best of each averaged. Cardiovascular health outcomes: Fasting plasma glucose (hexokinase method) and serum insulin (Electrochemiluminescence immunoassay, Cobas 602); lipids (total cholesterol, HDL-C, triglycerides by enzymatic colorimetry; LDL-C calculated by Friedewald equation). Blood pressure measured after 5-min seated rest with automated sphygmomanometer; two measurements (third if >10 mmHg discrepancy); averages used. Derived metrics: GDM defined as fasting plasma glucose ≥5.1 mmol/L (IADPSG). HOMA-IR = (fasting insulin [µU/L] × fasting glucose [mmol/L]) / 22.5; ln-transformed for analyses. MetS score computed as standardized sum of z-scores of triglycerides, inverted HDL-C, glucose, and mean of systolic/diastolic BP (waist omitted). High HOMA-IR, high BP (mean of SBP/DBP), and high MetS defined as ≥1 SD above sample mean. Statistical analysis: Linear regression to assess associations of body composition and fitness with continuous outcomes; logistic regression to estimate odds ratios (ORs) per 1 SD increase in exposures for GDM, high HOMA-IR, high BP, and high MetS. Three models: unadjusted; partially adjusted (age, education [university vs not], parity [0 vs ≥1]); fully adjusted (plus mutual adjustments for CRF, handgrip, FMI, FFMI; models with BMI and %FM excluded FMI/FFMI to avoid collinearity). Assumptions checked; two-sided P<0.05 considered significant; no multiple comparison adjustment. Sensitivity analyses: Additional adjustment for self-reported pre-pregnancy physical activity; models assessing the mediating/attenuating role of FMI; fitness analyses restricted to participants with average 6MWT HR >60% or >70% of age-predicted maximum and models including 6MWT HR% as covariate. Analyses performed in SPSS v26. Ethics: Regional Ethical Review Board in Linköping approvals (2017/112-31; 2018/262-32); written informed consent.

Sample: n=303; mean age 31.3±4.1 y; pre-pregnancy BMI 23.7±3.9 kg/m²; GDM prevalence 12.3% (n=37). Body composition: In linear models, BMI, FMI, and %FM showed strong positive associations with fasting glucose, HOMA-IR, SBP, DBP, and MetS score (all β ≥ 0.30; all P<0.001). FFMI showed weaker positive associations that became non-significant after adjustment for FMI and fitness. Logistic regression (per 1 SD higher exposure): - GDM: BMI, FMI, %FM associated with higher odds (ORs 1.72–2.14; P≤0.003) in unadjusted and adjusted models. FFMI associated with GDM in unadjusted/partially adjusted models (OR up to 1.82; P=0.001) but attenuated in full model (OR 1.54; P=0.039). - High HOMA-IR: BMI, FMI, %FM ORs 3.01–3.80; P<0.001; FFMI OR 1.54–1.63; P<0.006 unadjusted/partially adjusted, attenuated in full model (ns). - High blood pressure (mean SBP/DBP): BMI, FMI, %FM ORs 1.81–2.05; P<0.001. - High MetS score: BMI, FMI, %FM ORs 3.29–3.71; P<0.001; FFMI OR 1.63; P=0.004 unadjusted/partially adjusted, attenuated to ns in full model. Physical fitness: Higher CRF associated with lower HOMA-IR and MetS score in unadjusted models (e.g., OR for high HOMA-IR 0.60, P=0.001; high MetS 0.57, P=0.001) and with small favorable linear associations for HOMA-IR, SBP, MetS (β about -0.12 to -0.20), but all associations were fully attenuated after adjusting for body composition (FMI/FFMI) and handgrip strength. Handgrip strength showed no significant associations with cardiometabolic outcomes. Sensitivity analyses indicated FMI largely mediated/attenuated associations of FFMI and CRF with outcomes; results were robust to adjustment for pre-pregnancy physical activity and to restricting analyses to higher 6MWT exertion or adjusting for 6MWT HR%.

The findings demonstrate that adiposity, whether assessed by FMI or %FM, is strongly and consistently associated with GDM risk, insulin resistance, blood pressure, and a composite MetS score in early pregnancy. In contrast, FFMI shows only weak associations that disappear after accounting for fat mass, indicating that FFM does not confer independent cardiometabolic benefit in this context. BMI performed similarly to FMI and %FM in predicting risk, aligning with prior evidence that BMI, despite being an imperfect measure of adiposity, tracks cardiometabolic risk effectively when FFM lacks an independent protective association. CRF exhibited favorable associations with insulin resistance and overall metabolic risk but these were not independent of adiposity, suggesting that higher fat mass mediates or explains much of the apparent fitness benefit in early pregnancy. Clinically, results emphasize focusing on excess fat mass rather than FFM during prenatal care for cardiometabolic risk assessment and indicate that BMI remains a practical, effective screening measure. The novel examination of fitness suggests potential benefits but underscores the need for further research to clarify causal pathways and the independent role of fitness in pregnancy health.

State-of-the-art measures of fat mass (FMI, %FM) in early pregnancy are strongly associated with GDM and adverse cardiometabolic markers, but not more strongly than BMI. FFMI has weak, non-independent associations once fat mass is considered. Greater CRF relates to lower insulin resistance and MetS risk, yet these associations are attenuated after accounting for fat mass, indicating limited independent predictive value. Future research should use more precise fitness assessments, longitudinal designs to assess causality and temporal dynamics across gestation, and more diverse populations to enhance generalizability and to explore whether interventions targeting adiposity and fitness can improve pregnancy and long-term maternal cardiometabolic outcomes.

Cross-sectional design limits causal inference. Cardiorespiratory fitness assessed via a submaximal 6-minute walk test; potential variability in exertion could bias estimates despite sensitivity analyses. Sample had a high proportion of women with university education, potentially limiting generalizability; overweight/obesity prevalence was somewhat lower than national figures. Classification of high HOMA-IR, blood pressure, and MetS as ≥1 SD above the mean lacks pregnancy-specific clinical thresholds. Inability to distinguish fetal from maternal body composition, though fetal contribution at ~14 weeks is minimal; nevertheless, maternal body composition estimates could include minor fetal contributions. No adjustment for multiple comparisons.