ABSTRACT
Aims
This study aimed to examine the associations of maternal first‐trimester atherogenic index of plasma (AIP) with gestational diabetes mellitus (GDM) and birth weight outcomes, and the mediating role of second‐trimester blood glucose.
Methods
This prospective cohort study included 5,706 pregnant women without pre‐existing diabetes or other chronic diseases. AIP was calculated as log10[TG (mg/dL)/HDL‐C (mg/dL)] from fasting lipid profiles. Generalized linear models and mediation analysis were used to assess the relationships between AIP and outcomes, and the mediating effects of blood glucose.
Results
Participants had a mean maternal age of 29.3 years, a mean first‐trimester AIP of 0.28 ± 0.19, GDM prevalence of 20.64%, and mean neonatal birth weight of 3314.05 ± 442.11 g. A 1‐unit increase in first‐trimester AIP was associated with greater birth weight (β: 93.12 g, 95% CI: 39.38–146.87), elevated GDM risk (RR: 2.08, 95% CI: 1.52–2.85), and higher risk of large‐for‐gestational‐age infants (RR: 1.67, 95% CI: 1.09–2.58). A nonlinear association was identified between first‐trimester AIP and birth weight, with a 158.95 g (β: 158.95 g, 95% CI: 90.70–227.20) rise per 1‐unit elevation in AIP above 0.14. Fasting plasma glucose (FPG) mediated 28.35% of the association between AIP and birth weight, rising to 60.04% among women with GDM. The third‐trimester rise in AIP was significantly smaller in women with GDM.
Conclusions
First‐trimester AIP appears to correlate with GDM and birth weight. Second‐trimester FPG mediates this association, particularly in mothers with GDM. These findings support the clinical value of early lipid monitoring in prenatal care.
Keywords: Atherogenic index of plasma, Birth weight, Gestational diabetes mellitus
Abbreviations
- AIP
atherogenic index of plasma
- BMI
body mass index
- CBCS
China Birth Cohort study
- CHOD‐PAP
cholesterol oxidase‐phenol 4‐aminoantipyrine peroxidase
- CI
confidence interval
- EDC
enterprise data center
- FPG
fasting plasma glucose
- GDM
gestational diabetes mellitus
- GLM
generalized linear model
- GPO‐PAP
glycerol phosphate oxidase‐4‐aminophenazone
- HDL‐C
high‐density lipoprotein cholesterol
- IADPSG
International Association of Diabetes and Pregnancy Study Groups
- LDL‐C
low‐density lipid cholesterol
- LGA
large for gestational age
- OGTT
oral glucose tolerance test
- RCS
restricted cubic spline
- RR
relative risk
- SD
standard deviation
- SGA
small for gestational age
- TC
total cholesterol
- TG
triglyceride
- UA
uric acid
- VIFs
variance inflation factors
INTRODUCTION
Lipid metabolism disorders, encompassing elevated triglycerides (TG), abnormal cholesterol profiles characterized by increased low‐density lipoprotein cholesterol (LDL‐C) and decreased high‐density lipoprotein cholesterol (HDL‐C), and lipoprotein metabolism abnormalities, are considered to exert a crucial pathogenic function in the development of diverse conditions 1 , 2 . In recent years, reports on composite indicators integrating multiple lipid metabolism parameters have increased, among which the atherogenic index of plasma (AIP) has been established as a novel and robust biomarker among indicators related to dyslipidemia 3 , 4 . AIP serves as a comprehensive lipid indicator that incorporates both pro‐atherogenic and anti‐atherogenic lipoproteins and reflects the balance of lipid metabolism. It has been identified as a potential predictor for metabolic syndrome, cardiovascular risk, and other related health conditions in the general population 5 , 6 , 7 , 8 .
Gestational diabetes mellitus (GDM) is one of the most prevalent metabolic disorders in pregnancy, elevating the risk for multiple adverse outcomes, including fetal overgrowth, large‐for‐gestational age (LGA) infants, preeclampsia, and the long‐term risk of developing type 2 diabetes and cardiovascular diseases for both mothers and offspring 9 , 10 . Dyslipidemia is recognized as a critical factor in the development of abnormal glucose metabolism, and some studies have demonstrated that AIP correlates with higher blood glucose levels, insulin resistance, and increased diabetes risk in the general population, with this association more pronounced in women 11 , 12 , 13 . Pregnancy is characterized by physiological changes in maternal lipid metabolism, which are an adaptive response aimed at supporting fetal development 14 . Excessive maternal dyslipidemia can disproportionately exacerbate insulin resistance and β‐cell dysfunction, and several prospective cohort studies have identified that elevated first‐trimester AIP was significantly correlated with an increased risk of developing GDM 15 , 16 . Notably, this metabolic dysregulation creates a dual threat, as both elevated lipid levels and the consequent hyperglycemia contribute to altered fetal growth, ultimately influencing neonatal birth weight and increasing the chance of macrosomia 17 , 18 . Nevertheless, evidence regarding the relationship between first‐trimester AIP and neonatal birth weight outcomes remains insufficient, and it is unclear whether this association may be mediated by second‐trimester blood glucose levels.
Based on a prospective birth cohort study, this study aimed to comprehensively assess the associations of early pregnancy AIP with the risk of GDM and neonatal birth weight outcomes. Moreover, the mediating effect of second‐trimester blood glucose levels in the relationship between first‐trimester AIP and birth weight was explored.
METHODS
Study design and participants
Data utilized in the present study were derived from an ancillary research project carried out in Xi'an, Shaanxi Province, from July 2018 to December 2023, which was a part of the China Birth Cohort Study (CBCS). CBCS is a nationwide prospective birth cohort that collects information on pregnant women and infants across China 19 . Pregnant women aged 18 years or older with a gestational age of 6–13+6 weeks were enrolled in the birth cohort. Participants were permitted to withdraw voluntarily at any time throughout the study. Maternal sociodemographic characteristics, lifestyle behaviors, anthropometric measurements, and pregnancy outcomes were collected and documented, with follow‐up conducted through the enterprise data center cloud (EDC) platform over the full duration of pregnancy. Among 26,944 pregnant women, participants with spontaneous and induced abortions (n = 998), missing lipid profiles (n = 16,933), missing 75 g OGTT results (n = 1,579), missing birth outcome records or those with birth defects (n = 1,020), twin/multiple pregnancies (n = 76), a history of prior diabetes or pre‐existing metabolic/endocrine conditions (n = 523), and incomplete baseline characteristics (n = 109) were excluded. Finally, 5,706 subjects were included in the final analysis (Figure S1). Table S1 shows the baseline comparison between included and non‐eligible participants with complete baseline covariates.
Assessment of AIP
For lipid profile assessment, fasting venous blood specimens were collected from pregnant women during the first trimester (≤14 weeks), the second trimester (14–28 weeks), and the third trimester (>28 weeks). The measured lipid parameters included TG, total cholesterol (TC), HDL‐C, and LDL‐C. Standardized assay kits (Biosino Bio‐Technology and Science Inc., Beijing, China) were used for all determinations, in accordance with established experimental protocols. The AIP was calculated as log10 [TG (mg/dL)/HDL‐C (mg/dL)].
Outcome ascertainment
The primary outcomes were GDM and neonatal birth weight, and the secondary outcomes included LGA and small‐for‐gestational‐age (SGA). GDM was diagnosed via the 75 g oral glucose tolerance test (OGTT) performed at 24–28 weeks of gestation, with plasma glucose assayed by the hexokinase method. The diagnosis was confirmed when any of the following thresholds was met: fasting plasma glucose (FPG) ≥ 5.1 mmol/L, 1‐h post‐load glucose ≥ 10.0 mmol/L, or 2‐h post‐load glucose ≥ 8.5 mmol/L.
Neonatal birth weight was measured immediately after delivery with a calibrated electronic scale with an accuracy of ±5 g. LGA was characterized as a birth weight at or exceeding the 90th percentile for gestational age, as defined by the Intergrowth‐21st International Fetal and Newborn Growth Standards, whereas SGA was identified as a birth weight at or below the 10th percentile based on the same reference standards.
Covariates
Sociodemographic characteristics, reproductive history, lifestyle habits, and health status of participants were collected as covariates. Maternal age, educational level, ethnicity, occupation, household income, pre‐pregnancy body mass index (BMI), conception method, primiparity, prior pregnancy complications, adverse pregnancy history, cold/fever history, smoking/passive smoking, drinking, folic acid and multivitamin supplementation, and uric acid (UA) levels were collected before 14 gestational weeks. Prior pregnancy complications referred to a history of GDM, gestational hypertension, or thyroid disorders in previous pregnancies. Adverse pregnancy history referred to prior miscarriage, preterm delivery, or fetal malformation. Blood glucose control methods were collected after GDM diagnosis. Neonatal sex and gestational age at delivery were collected after delivery.
Statistical analysis
Normally distributed continuous variables are expressed as mean ± standard deviation (SD) and compared by one‐way analysis of variance. Categorical variables are reported as frequencies and percentages and compared using the χ 2 test. Participants were divided into three AIP tertile groups (T1: <0.19, T2: 0.19–0.35, T3: >0.35) for baseline comparison.
Generalized linear models (GLMs) were applied to assess associations between AIP and outcomes: a Poisson link function with robust variances was used for categorical outcomes to calculate relative risks (RRs) and 95% confidence intervals (CIs), and an identity link function was used for the continuous outcome to estimate β coefficients and 95% CIs. The maximum variance inflation factor (VIF) was 2.6 (Table S2), indicating no significant multicollinearity in the multivariable model. Restricted cubic spline (RCS) regression with four knots at the 5th, 35th, 65th, and 90th percentiles evaluated dose–response relationships. Piecewise regression was used for threshold effect analysis when non‐linearity was detected.
Mediation analysis was performed to examine whether second‐trimester blood glucose levels mediate the association between AIP and neonatal birth weight. Two linear regression models were constructed: the mediator model estimated the impact of AIP on blood glucose levels (path a), the outcome model assessed the direct effect of AIP on birth weight (path c′) while controlling for blood glucose levels, as well as the effect of second‐trimester blood glucose levels on birth weight (path b). Mediation effects were calculated as the product of path a and path b (a × b). Stratified mediation by the RCS‐identified inflection point (AIP = 0.14) was additionally conducted to account for nonlinearity.
Trimester‐specific AIP changes and the GDM‐trimester interaction were analyzed to complement the mediation analysis. Linear regression calculated observed AIP differences between GDM and non‐GDM groups at each trimester. Given missing AIP data in the second (19.05%) and third (13.21%) trimesters (Table S3), a linear mixed‐effects model with a random participant intercept was used for repeated measurements, including fixed effects of GDM status, trimester, their interaction, and covariates. Marginal predicted AIP values stratified by GDM status were generated for visualization.
Subgroup and sensitivity analyses tested result robustness. GLMs explored AIP associations with outcomes across subgroups and in a sensitivity model that further adjusted for gestational age at lipid sampling based on Model 3. Quartile categorization of AIP was added as a sensitivity analysis for grouping strategy robustness. FPG mediation effects were stratified by neonatal sex.
All analyses were performed in R version 4.2.3. A two‐tailed P < 0.05 was defined as statistically significant.
RESULTS
Baseline characteristics
Table 1 presents baseline characteristics of 5,706 pregnant women stratified by first‐trimester AIP tertiles. AIP was measured at 12.39 ± 1.11 gestational weeks, with a mean value of 0.28 ± 0.19. At 25.58 ± 2.02 gestational weeks, 1,178 (20.64%) women were diagnosed with GDM. The mean infant birth weight was 3,314.05 ± 442.11 g, with the rates of LGA and SGA infants being 11.04% and 4.40%, respectively. Pregnant women in higher AIP tertiles were more likely to be older, less educated, unemployed, obese, non‐primiparous, IVF conception, take pre‐pregnancy multivitamin nutrients, and had a history of pregnancy complications and adverse outcomes than those in the lowest tertile. Women in higher AIP tertiles had higher levels of TG, TC, LDL‐C, UA, and blood glucose levels but lower HDL‐C compared with those in the lowest tertile. In addition, compared with participants in the lowest AIP tertile, those in higher tertiles had greater neonatal birth weight, higher prevalence of GDM and LGA, a higher proportion receiving glycemic control interventions, and an earlier gestational age at delivery.
Table 1.
Baseline characteristics of pregnant women
| Characteristic | First‐trimester AIP quantile | P | |||
|---|---|---|---|---|---|
| Overall | T1 | T2 | T3 | ||
| No. of subjects | 5,706 | 1,902 | 1,902 | 1,902 | |
| First trimester | |||||
| Age, years | 29.32 ± 3.44 | 28.96 ± 3.30 | 29.22 ± 3.45 | 29.78 ± 3.52 | <0.001 |
| <30 | 3,184 (55.80) | 1,151 (60.52) | 1,080 (56.78) | 953 (50.11) | <0.001 |
| 30 ~ 34 | 2,082 (36.49) | 636 (33.44) | 683 (35.91) | 763 (40.12) | |
| ≥ 35 | 440 (7.71) | 115 (6.05) | 139 (7.31) | 186 (9.78) | |
| Education | <0.001 | ||||
| Senior high school or lower | 619 (10.85) | 171 (8.99) | 193 (10.15) | 255 (13.41) | |
| Junior or regular college | 4,122 (72.24) | 1,355 (71.24) | 1,380 (72.56) | 1,387 (72.92) | |
| Graduate or above | 965 (16.91) | 376 (19.77) | 329 (17.30) | 260 (13.67) | |
| Ethnicity | 0.967 | ||||
| Han nationality | 5,616 (98.42) | 1,871 (98.37) | 1,872 (98.42) | 1,873 (98.48) | |
| Ethnic minorities | 90 (1.58) | 31 (1.63) | 30 (1.58) | 29 (1.52) | |
| Occupation | <0.001 | ||||
| Working regularly | 4,510 (79.04) | 1,580 (83.07) | 1,523 (80.07) | 1,407 (73.97) | |
| Peasants/migrant workers | 73 (1.28) | 13 (0.68) | 25 (1.31) | 35 (1.84) | |
| Unemployed | 1,123 (19.68) | 309 (16.25) | 354 (18.61) | 460 (24.19) | |
| Household income, CNY | 0.086 | ||||
| Poverty (<50,000) | 638 (11.18) | 194 (10.20) | 211 (11.09) | 233 (12.25) | |
| Medium (50,000 ~ 200,000) | 3,660 (64.14) | 1,203 (63.25) | 1,237 (65.04) | 1,220 (64.14) | |
| Rich (>200,000) | 1,408 (24.68) | 505 (26.55) | 454 (23.87) | 449 (23.61) | |
| BMI, kg/m2 | 21.39 ± 3.02 | 20.41 ± 2.46 | 21.15 ± 2.89 | 22.61 ± 3.23 | <0.001 |
| Underweight (<18.5) | 801 (14.04) | 387 (20.35) | 276 (14.51) | 138 (7.26) | <0.001 |
| Normal (18.5 ~ 23.9) | 3,938 (69.02) | 1,377 (72.40) | 1,359 (71.45) | 1,202 (63.20) | |
| Overweight (24.0 ~ 27.9) | 779 (13.65) | 113 (5.94) | 217 (11.41) | 449 (23.61) | |
| Obesity (≥28.0) | 188 (3.29) | 25 (1.31) | 50 (2.63) | 113 (5.94) | |
| Conception method | <0.001 | ||||
| Natural pregnancy | 5,353 (93.81) | 1,824 (95.90) | 1,785 (93.85) | 1,744 (91.69) | |
| Artificial insemination | 47 (0.82) | 11 (0.58) | 22 (1.16) | 14 (0.74) | |
| In vitro fertilization | 306 (5.36) | 67 (3.52) | 95 (4.99) | 144 (7.57) | |
| Primiparity | <0.001 | ||||
| No | 2,414 (42.31) | 721 (37.91) | 774 (40.69) | 919 (48.32) | |
| Yes | 3,292 (57.69) | 1,181 (62.09) | 1,128 (59.31) | 983 (51.68) | |
| Prior pregnancy complications | <0.001 | ||||
| No | 5,459 (95.67) | 1,841 (96.79) | 1,830 (96.21) | 1,788 (94.01) | |
| Yes | 247 (4.33) | 61 (3.21) | 72 (3.79) | 114 (5.99) | |
| Adverse pregnancy history | <0.001 | ||||
| No | 4,104 (71.92) | 1,430 (75.18) | 1,374 (72.24) | 1,300 (68.35) | |
| Yes | 1,602 (28.08) | 472 (24.82) | 528 (27.76) | 602 (31.65) | |
| Cold/fever | 0.456 | ||||
| No | 4,482 (78.55) | 1,506 (79.18) | 1,476 (77.60) | 1,500 (78.86) | |
| Yes | 1,224 (21.45) | 396 (20.82) | 426 (22.40) | 402 (21.14) | |
| Smoking/Passive smoking | 0.574 | ||||
| No | 4,908 (86.01) | 1,623 (85.33) | 1,642 (86.33) | 1,643 (86.38) | |
| Yes | 798 (13.99) | 279 (14.67) | 260 (13.67) | 259 (13.62) | |
| Drinking | 0.453 | ||||
| No | 5,481 (96.06) | 1,819 (95.64) | 1,834 (96.42) | 1,828 (96.11) | |
| Yes | 225 (3.94) | 83 (4.36) | 68 (3.58) | 74 (3.89) | |
| Folic acid | 0.728 | ||||
| No | 162 (2.84) | 60 (3.15) | 48 (2.52) | 54 (2.84) | |
| Taking before pregnancy | 2,915 (51.09) | 962 (50.58) | 967 (50.84) | 986 (51.84) | |
| Taking after pregnancy | 2,629 (46.07) | 880 (46.27) | 887 (46.64) | 862 (45.32) | |
| Multivitamin | 0.045 | ||||
| No | 2,728 (47.81) | 930 (48.90) | 868 (45.64) | 930 (48.90) | |
| Taking before pregnancy | 1,141 (20.00) | 364 (19.14) | 379 (19.93) | 398 (20.93) | |
| Taking after pregnancy | 1,837 (32.19) | 608 (31.97) | 655 (34.44) | 574 (30.18) | |
| TG, mmol/L | 1.49 ± 0.60 | 1.02 ± 0.21 | 1.39 ± 0.23 | 2.05 ± 0.66 | <0.001 |
| TC, mmol/L | 4.37 ± 0.71 | 4.30 ± 0.68 | 4.36 ± 0.71 | 4.43 ± 0.74 | <0.001 |
| HDL‐C, mmol/L | 1.69 ± 0.33 | 1.90 ± 0.31 | 1.69 ± 0.26 | 1.49 ± 0.26 | <0.001 |
| LDL‐C, mmol/L | 2.24 ± 0.56 | 2.07 ± 0.51 | 2.26 ± 0.54 | 2.40 ± 0.58 | <0.001 |
| UA, μmol/L | 214.52 ± 43.18 | 207.14 ± 40.44 | 211.68 ± 41.76 | 224.74 ± 45.27 | <0.001 |
| Second trimester | |||||
| FPG, mmol/L | 4.69 ± 0.43 | 4.61 ± 0.37 | 4.68 ± 0.43 | 4.79 ± 0.46 | <0.001 |
| OGTT‐1 h, mmol/L | 7.53 ± 1.66 | 7.27 ± 1.56 | 7.43 ± 1.68 | 7.87 ± 1.69 | <0.001 |
| OGTT‐2 h, mmol/L | 6.61 ± 1.28 | 6.40 ± 1.16 | 6.57 ± 1.30 | 6.87 ± 1.32 | <0.001 |
| GDM | <0.001 | ||||
| No | 4,528 (79.36) | 1,624 (85.38) | 1,526 (80.23) | 1,378 (72.45) | |
| Yes | 1,178 (20.64) | 278 (14.62) | 376 (19.77) | 524 (27.55) | |
| Blood glucose control methods | <0.001 | ||||
| No | 4,616 (80.90) | 1,645 (86.49) | 1,559 (81.97) | 1,412 (74.24) | |
| Diet/exercise | 1,040 (18.23) | 245 (12.88) | 335 (17.61) | 460 (24.19) | |
| Drug therapy | 50 (0.88) | 12 (0.63) | 8 (0.42) | 30 (1.58) | |
| Birth outcomes | |||||
| Neonatal sex | 0.486 | ||||
| Male | 3,014 (52.82) | 993 (52.21) | 995 (52.31) | 1,026 (53.94) | |
| Female | 2,692 (47.18) | 909 (47.79) | 907 (47.69) | 876 (46.06) | |
| Gestational weeks at delivery | 39.31 ± 1.46 | 39.41 ± 1.36 | 39.33 ± 1.45 | 39.18 ± 1.56 | <0.001 |
| Birth weight, g | 3,314.05 ± 442.11 | 3,283.32 ± 425.23 | 3,314.22 ± 431.96 | 3,344.60 ± 466.14 | <0.001 |
| LGA | <0.001 | ||||
| No | 5,076 (88.96) | 1,741 (91.54) | 1,702 (89.48) | 1,633 (85.86) | |
| Yes | 630 (11.04) | 161 (8.46) | 200 (10.52) | 269 (14.14) | |
| SGA | 0.086 | ||||
| No | 5,455 (95.60) | 1,804 (94.85) | 1,819 (95.64) | 1,832 (96.32) | |
| Yes | 251 (4.40) | 98 (5.15) | 83 (4.36) | 70 (3.68) | |
Values were expressed as mean (standard deviation) or n (%).
AIP, atherogenic index of plasma; BMI, pre‐pregnancy body mass index; CNY, Chinese Yuan; FPG, fasting plasma glucose; GDM, gestational diabetes mellitus; HDL‐C, high‐density lipoprotein cholesterol; LDL‐C, low‐density lipid cholesterol; LGA, large for gestational age; OGTT, oral glucose tolerance test; SGA, small for gestational age; TC, total cholesterol; TG, triglyceride; UA, uric acid.
Association between first‐trimester AIP and GDM, birth weight outcomes
Table 2 shows the association between first‐trimester AIP and maternal and infant outcomes. In Model 3, for per 1‐unit increase in AIP, birth weight increased by 93.12 g (β: 93.12 g, 95% CI: 39.38, 146.87), and the risks of GDM and LGA increased by 108% (RR: 2.08, 95% CI: 1.52, 2.85) and 67% (RR = 1.67, 95% CI: 1.09, 2.58), respectively. Similar results were observed for standardized AIP. Additionally, birth weight, along with the risks of GDM and LGA, gradually rose as AIP tertiles increased (P trend <0.05). Moreover, a linear dose–response relationship between first‐trimester AIP and the risks of GDM and LGA (P overall <0.05), and a nonlinear association with birth weight (P non‐linear = 0.040) were observed in RCS analysis (Figure 1). At AIP ≥0.14, per 1‐unit increase in AIP was linked to a 158.95 g higher birth weight (β: 158.95 g, 95% CI: 90.70, 227.20), whereas no association was observed at AIP <0.14, the log likelihood ratio test (P = 0.036) confirmed nonlinearity (Table S4). Additionally, no significant associations were observed between AIP and macrosomia or low birth weight (Table S5).
Table 2.
Association between first‐trimester AIP and GDM, birth weight, LGA and SGA in different models
| Variable | Model 1 | Model 2 | Model 3 | |||
|---|---|---|---|---|---|---|
| β/RR (95% CI) | P | β/RR (95% CI) | P | β/RR (95% CI) | P | |
| GDM | ||||||
| AIP (continuous) | 3.89 (2.92, 5.17) | <0.001 | 2.09 (1.53, 2.85) | <0.001 | 2.08 (1.52, 2.85) | <0.001 |
| AIP (standardized) | 1.29 (1.22, 1.36) | <0.001 | 1.15 (1.08, 1.22) | <0.001 | 1.15 (1.08, 1.22) | <0.001 |
| AIP (tertile) | ||||||
| T1 | 1 | 1 | 1 | |||
| T2 | 1.35 (1.16, 1.58) | <0.001 | 1.24 (1.06, 1.45) | 0.007 | 1.24 (1.06, 1.45) | 0.007 |
| T3 | 1.88 (1.63, 2.18) | <0.001 | 1.46 (1.25, 1.71) | <0.001 | 1.46 (1.25, 1.70) | <0.001 |
| P trend | <0.001 | <0.001 | <0.001 | |||
| Birth weight | ||||||
| AIP (continuous) | 110.30 (49.52, 171.08) | <0.001 | 49.23 (−16.12, 114.57) | 0.140 | 93.12 (39.38, 146.87) | <0.001 |
| AIP (standardized) | 20.80 (9.34, 32.26) | <0.001 | 9.28 (−3.04, 21.60) | 0.140 | 17.56 (7.42, 27.69) | <0.001 |
| AIP (tertile) | ||||||
| T1 | 0 | 0 | 0 | |||
| T2 | 30.89 (2.83, 58.95) | 0.031 | 20.78 (−7.33, 48.90) | 0.147 | 26.58 (3.50, 49.65) | 0.024 |
| T3 | 61.27 (33.21, 89.33) | <0.001 | 35.07 (5.45, 64.69) | 0.020 | 46.02 (21.66, 70.38) | <0.001 |
| P trend | <0.001 | 0.020 | <0.001 | |||
| LGA | ||||||
| AIP (continuous) | 3.02 (2.03, 4.49) | <0.001 | 1.74 (1.13, 2.67) | 0.012 | 1.67 (1.09, 2.58) | 0.020 |
| AIP (standardized) | 1.23 (1.14, 1.33) | <0.001 | 1.11 (1.02, 1.20) | 0.012 | 1.10 (1.02, 1.20) | 0.020 |
| AIP (tertile) | ||||||
| T1 | 1 | 1 | 1 | |||
| T2 | 1.24 (1.01, 1.53) | 0.041 | 1.15 (0.93, 1.42) | 0.184 | 1.15 (0.93, 1.41) | 0.199 |
| T3 | 1.67 (1.37, 2.03) | <0.001 | 1.33 (1.08, 1.64) | 0.006 | 1.31 (1.07, 1.62) | 0.010 |
| P trend | <0.001 | 0.006 | 0.010 | |||
| SGA | ||||||
| AIP (continuous) | 0.54 (0.28, 1.05) | 0.071 | 1.03 (0.49, 2.13) | 0.945 | 1.02 (0.49, 2.13) | 0.950 |
| AIP (standardized) | 0.89 (0.78, 1.01) | 0.071 | 1.00 (0.88, 1.15) | 0.945 | 1.00 (0.87, 1.15) | 0.950 |
| AIP (tertile) | ||||||
| T1 | 1 | 1 | 1 | |||
| T2 | 0.85 (0.63, 1.13) | 0.265 | 0.92 (0.69, 1.24) | 0.592 | 0.92 (0.69, 1.24) | 0.589 |
| T3 | 0.71 (0.53, 0.97) | 0.032 | 0.93 (0.67, 1.28) | 0.646 | 0.93 (0.67, 1.28) | 0.640 |
| P trend | 0.031 | 0.625 | 0.618 | |||
Model 1: Crude model, with no covariates adjusted. Model 2: Adjusted for baseline characteristics during early pregnancy, including maternal age, educational level, ethnicity, occupation, household income, pre‐pregnancy BMI, conception method, primiparity, history of previous pregnancy complications, adverse pregnancy history, history of cold/fever, smoking/passive smoking, alcohol drinking, folic acid supplementation, multivitamin supplementation, and uric acid level. Model 3: Extended from Model 2 with additional outcome‐specific covariates: neonatal sex for GDM; neonatal sex, blood glucose control methods, and gestational age at delivery for birth weight; neonatal sex, and blood glucose control methods for LGA and SGA.
CI, confidence interval; RR, relative risk; other abbreviations as in Table 1.
Figure 1.
Restricted cubic spline (RCS) analysis of the association between first‐trimester AIP and GDM (a), birth weight (b), LGA (c), and SGA (d). All models were adjusted for baseline characteristics during early pregnancy and outcome‐specific covariates: neonatal sex for GDM; neonatal sex, blood glucose control methods, and gestational age at delivery for birth weight; neonatal sex and blood glucose control methods for LGA and SGA. The solid line and red area represent the point estimates and 95% confidence intervals, respectively.
Mediation analysis of blood glucose levels in the first‐trimester AIP‐birth weight relationship
To investigate the mediating effect of blood glucose levels between first‐trimester AIP and birth weight, several assumptions for mediation analysis were assessed in Tables S6 and S7, and the mediation results are presented in Figure 2. FPG and 1‐h glucose significantly mediated the association between AIP and birth weight, with mediation proportions estimated at 28.35% and 14.70%, respectively. Additionally, FPG significantly mediated the association between AIP and birth weight in both pregnant women with GDM and those without, with mediation proportions estimated at 60.04% and 15.33%, respectively. However, the direct effect of AIP on birth weight was not statistically significant in pregnant women with GDM (P ≥ 0.05). Given the nonlinear relationship shown by RCS, mediation was stratified by the 0.14 inflection point. No significant mediation was detected below this threshold. For AIP ≥ 0.14, FPG significantly mediated the association between AIP and birth weight, with a mediation proportion of 25.19% (Figure S2).
Figure 2.
Mediation analyses for the roles of glucose levels in the association between first‐trimester AIP and birth weight. Panels (a–c) show the mediation effects of FPG, OGTT 1‐h glucose, and OGTT 2‐h glucose in the overall cohort, respectively. Panels (d) and (e) show the mediation effect of FPG stratified by GDM status: women with GDM (d) and women without GDM (e). All mediator models were adjusted for baseline characteristics during early pregnancy and neonatal sex. All outcome models were additionally adjusted for blood glucose control methods and gestational age at delivery.
Interaction between GDM status and gestational trimester on longitudinal AIP trajectories
As shown in Figure 3a, the adjusted observed AIP was significantly higher in the GDM group compared to the non‐GDM group during the first and second trimesters (P < 0.05), whereas the difference was not statistically significant in the third trimester (P ≥ 0.05). As shown in Table S8, women who later developed GDM already had significantly higher AIP levels in the first trimester (β = 0.038, P < 0.001), prior to GDM diagnosis. AIP levels increased with advancing gestation in women without GDM (βsecond trimester = 0.154, βthird trimester = 0.308, both P < 0.001). Specifically, a significant interaction was observed between GDM status and third‐trimester gestation (β = −0.035, P for interaction <0.001), indicating a smaller AIP increase from the first to third trimester in GDM than in non‐GDM women. The marginal predicted values of AIP (Figure 3b) were consistent with these patterns.
Figure 3.
Observed values (a) and estimated marginal predicted values (b) of AIP in women with and without GDM across gestational trimesters. All analyses were adjusted for all baseline characteristics during early pregnancy and neonatal sex. Predicted values are derived from the mixed‐effects model in Table S8.
Subgroup and sensitivity analyses
Figure 4 presents subgroup analysis results examining the associations of first‐trimester AIP with GDM and birth weight. Positive associations of AIP with GDM and birth weight were consistently observed across most subgroups. After further adjustment for gestational age at lipid sampling, effect sizes and statistical significance for primary outcomes also remained essentially unchanged (Table S9). When first‐trimester AIP was categorized into quartiles, its lowest‐group cutoff (0.152) closely matched the RCS derived inflection point of 0.14, with largely consistent dose–response patterns and effect sizes relative to the primary tertile analysis (Table S10). In addition, the mediation effect of FPG in the first‐trimester‐AIP‐birth weight relationship was similar between male and female infants, with the mediation proportions estimated at 29.14% and 27.70%, respectively (Figure S3).
Figure 4.
Forest plot of the exploratory stratified analysis for the associations of first‐trimester AIP with GDM and birth weight. All subgroup analyses were adjusted for all baseline characteristics during early pregnancy and neonatal sex, excluding the corresponding stratification variable. For the birth weight outcome, models were additionally adjusted for blood glucose control methods and gestational age at delivery.
DISCUSSION
In this prospective birth cohort study, we observed a linear dose–response relationship between first‐trimester AIP and the risk of both GDM and LGA, and a nonlinear relationship with birth weight. FPG explained approximately 60.04% and 15.33% of the association between first‐trimester AIP and birth weight in pregnant women with GDM and those without GDM, respectively. Notably, we described the longitudinal trajectory of AIP across gestation and found that the physiological increase in AIP during the third trimester was significantly smaller in women with GDM.
Several epidemiological investigations have explored the association between first‐trimester AIP and the risk of GDM. In a prospective cohort study conducted in Korea, a positive linear relationship was detected between AIP and GDM, and GDM risk increased by 58% with per 0.1‐unit AIP increase 15 . However, findings from the Beijing Birth Cohort Study suggested a significant nonlinear association between AIP and GDM risk, with GDM risk rising 339% per 1‐unit increase in AIP 16 . Notably, such variations in the dose–response shape are not unique to pregnancy, and accumulating evidence from general populations also reveals both linear and nonlinear associations between AIP and diabetes risk 20 , 21 . This study suggested a linear dose–response association, with per 1‐unit increase in AIP associated with a 108% elevated GDM risk. Variations in these observations may be attributable to differences in sample size, ethnic background, geographical variations, lifestyle factors, or the distribution of AIP values. Nevertheless, all studies consistently indicate that elevated AIP is a significant risk factor for GDM, underscoring its potential value in early metabolic risk assessment during pregnancy.
Results have been inconsistent regarding the relationships between maternal lipid profiles during pregnancy and offspring birth weight. It has been reported that higher first‐trimester concentrations of TC, TG, and LDL‐C are associated with higher birth weight and an elevated risk of LGA 22 , 23 , 24 . In contrast, a Mendelian randomization analysis found no causal link between maternal gestational levels of HDL‐C, LDL‐C, or TG and birth weight 25 . Notably, it lacked precise specification of gestational timing for lipid measurement, and since the effects of lipids on fetal development are highly time‐sensitive 26 , this gap may explain its null result as it missed the causal effect of lipids in the critical window. Furthermore, prior research has predominantly focused on individual lipid parameters, and the association between comprehensive indicators such as AIP and birth weight remains largely unexplored. Our study demonstrated a nonlinear association between first‐trimester AIP and birth weight and identified a threshold effect at an AIP value of 0.14. The low threshold value, which is below the population mean, indicates that the fetal growth trajectory is susceptible to relatively small shifts in maternal lipid metabolism and that early lipid screening is necessary for pregnant women.
Our mediation analysis identified FPG and OGTT‐1 h glucose as key mediators of the association between first‐trimester AIP and birth weight, indicating that atherogenic dyslipidemia marked by AIP affects fetal growth through both glucose‐dependent and glucose‐independent pathways. Consistent with existing studies, maternal dyslipidemia commonly coincides with insulin resistance, and the hyperglycemic intrauterine environment supplies excess nutrients to the fetus, elevating birth weight via enhanced fetal insulin secretion and fat deposition 27 , 28 , 29 . In terms of the difference between fasting and post‐load glucose, the stronger mediation effect of FPG compared with post‐load glucose in this study is possibly because FPG underlies sustained basal hyperglycemia and chronic fetal glucose exposure. This is further supported by a population‐based study also showing that pregnancies with elevated FPG, either alone or combined with post‐load glucose elevation, had higher rates of LGA than those with isolated post‐load glucose elevation 30 . Stratified analyses by GDM status revealed distinct effect patterns. In women with GDM, FPG mediated 60.04% of the association, and the direct effect of first‐trimester AIP on birth weight was non‐significant, in line with previous findings that first‐trimester lipids exert no independent effect on birth weight in GDM populations 31 , 32 . A recent study also reported that AIP was associated with adverse pregnancy outcomes in women without GDM, but not in women with GDM 33 . This heterogeneity can be explained by the dominant effect of overt hyperglycemia in GDM pregnancies, which overwhelms the more subtle effect of AIP on fetal growth. In contrast, in non‐GDM pregnancies with glucose levels within the normal range, dyslipidemia related to AIP exerts a relatively more pronounced independent effect on fetal growth. Notably, stratified mediation based on the inflection point derived from RCS further revealed that the FPG‐mediated pathway was only active when first‐trimester AIP exceeded 0.14, suggesting a threshold effect of AIP on fetal growth via glucose metabolism.
We further examined longitudinal AIP changes and the interaction between GDM and trimester. AIP increased significantly across gestation, reflecting physiological adaptive lipid remodeling 14 . However, the magnitude of AIP elevation from the first to third trimester was significantly smaller in women with GDM, and although AIP was higher in this group during early pregnancy, the between‐group difference narrowed by the third trimester. One plausible explanation is post‐diagnosis clinical interventions: 93% of GDM participants received dietary, exercise, or pharmacological glucose control. These treatments may improve metabolic homeostasis, alleviate hyperglycemia, and indirectly modulate lipid metabolism, thus suppressing further AIP elevation in late pregnancy 34 , 35 . Regarding temporality, AIP was already elevated in the first trimester among women later developing GDM, supporting an upstream role of AIP in metabolic dysregulation. Since first‐trimester FPG was unavailable in our cohort, independence from baseline glycemia could not be definitively determined. Studies adjusting for FPG 36 or homeostasis model assessment of insulin resistance (HOMA–IR) 15 still found significant AIP effects beyond fasting glucose and basal insulin resistance. However, reverse causality remains possible, as very early subclinical hyperglycemia and insulin resistance may also contribute to lipid dysregulation. Notably, the pathway whereby AIP mediates the association between GDM and preeclampsia 37 emerges after GDM onset and does not contradict its upstream role before diagnosis. Collectively, elevated first‐trimester AIP precedes the clinical diagnosis of GDM, supporting its potential as an early marker, but causality remains unconfirmed. The distinct trajectory in GDM may reflect both pre‐existing intrinsic metabolic dysfunction and glycemic intervention–induced improvement in lipid profiles. Although mediation results suggest hyperglycemia largely mediates the effects of AIP on fetal growth, future studies with serial glucose and lipid measurements across pregnancy and baseline glycemic adjustment are needed to clarify the temporal sequence and the role of AIP independent of glucose.
The mechanisms underlying AIP‐related GDM and fetal overgrowth involve insulin resistance, inflammation, and β‐cell dysfunction, with maternal hyperglycemia as a core mediating pathway. Elevated free fatty acids from enhanced TG breakdown induce lipotoxicity, impairing insulin signaling in skeletal muscle and liver, reducing glucose uptake and increasing hepatic glucose production to form a pro‐diabetic early‐pregnancy milieu 38 . Adverse lipid profiles also upregulate pro‐inflammatory cytokines and oxidative stress markers, which further blunt insulin signal transduction and promote β‐cell dysfunction, accelerating GDM progression 39 . Dyslipidemia reduces peripheral insulin sensitivity, induces β‐cell apoptosis and functional impairment, and compromises compensatory hyperinsulinemia, leading to overt hyperglycemia 40 . Resultant maternal hyperglycemia and hyperinsulinemia increase transplacental glucose and lipid transport, alter placental function, stimulate fetal insulin secretion, and disrupt fetal growth 41 . Additionally, dysregulated maternal lipid metabolism may induce aberrant DNA methylation and histone modifications in the placenta and fetus, modifying genes involved in glucose metabolism and fetal growth 42 . Further research is warranted to identify the precise mechanism by which AIP contributes to GDM and changes in birth weight.
Several limitations should be acknowledged. First, as a single‐center birth cohort with clinically driven non‐random missing lipid profiles, our findings may be subject to potential selection bias and have limited generalizability to broader populations of pregnant women. Second, the observational design cannot fully exclude unmeasured or residual confounding from factors like detailed dietary intake, physical activity levels, and genetic predispositions. Third, glucose was measured only at a single second‐trimester time point, which fails to capture gestational glycemic fluctuations that may affect fetal growth and precludes adjustment for baseline glucose to disentangle the independent effect of AIP and clarify its causal relationship with GDM. Future prospective cohorts with serial glucose monitoring across all trimesters are warranted to address this gap.
CONCLUSIONS
In summary, elevated first‐trimester AIP correlated with increased birth weight and a higher risk of GDM and LGA. Second‐trimester FPG mediated the association between AIP and birth weight, particularly in pregnant women with GDM. These findings highlight the importance of early pregnancy lipid profiling for identifying pregnancies at high risk and suggest that optimizing glucose control may confer additional benefits through modulating lipid metabolism.
FUNDING
This work was supported by the Shaanxi Health Science and Technology Innovation Capacity Enhancement Program (2025YF‐09), the National Natural Science Foundation of China (82574206, 82103924), the Key Research and Development Program of Shaanxi Province (2025SF‐YBXM‐227, 2024SF2‐GJHX‐20), the Clinical Research Center for Gestational Metabolic Diseases of Shaanxi Province (2022LCZX‐08), and the Xi'an Science and Technology Plan Project (24YXYJ0025).
DISCLOSURE
The authors declare no conflict of interest.
Approval of the research protocol: This study followed the Declaration of Helsinki and was approved by the Human Research Ethics Committee of Beijing Obstetrics and Gynecology Hospital (No. 2018‐KY‐003‐02) and Northwest Women's and Children's Hospital (No. 2018018).
Informed consent: Informed consent was obtained from all subjects involved in the study.
Registry and the registration no. of the study/trial: N/A.
Animal studies: N/A.
Supporting information
Figure S1 Flowchart of study participants.
Figure S2 Association of first‐trimester AIP with birth weight mediated by FPG in participants with AIP <0.14 (A) and AIP ≥0.14 (B).
Figure S3 Association of first‐trimester AIP with birth weight mediated by FPG in male (A) and female (B) infants.
Table S1 Baseline characteristics of included and excluded participants.
Table S2 Variance inflation factors (VIF) in multifactor model.
Table S3 The pattern of missing AIP data.
Table S4 Threshold effect analysis of first‐trimester AIP levels and birth weight.
Table S5 Association between first‐trimester AIP, Macrosomia, and low birth weight.
Table S6 Association between first‐trimester AIP and second‐trimester blood glucose levels.
Table S7 Association between second‐trimester blood glucose levels and birth weight outcomes.
Table S8 Mixed‐effects model of GDM on AIP across gestational trimesters.
Table S9 Association between first‐trimester AIP and GDM, birth weight, LGA and SGA after additional adjustment for gestational age at lipid sampling.
Table S10 Association between first‐trimester AIP quartiles and GDM, birth weight, LGA and SGA.
ACKNOWLEDGMENTS
The authors wish to thank all mothers who participated in this study. They also acknowledge the personnel from Northwest Women's and Children's Hospital and Xi'an Jiaotong University for collaborating on data collection in the present study.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.
References
- 1. Pirillo A, Casula M, Olmastroni E, et al. Global epidemiology of dyslipidaemias. Nat Rev Cardiol 2021; 18: 689–700. [DOI] [PubMed] [Google Scholar]
- 2. Xu L, Yang Q, Zhou J. Mechanisms of abnormal lipid metabolism in the pathogenesis of disease. Int J Mol Sci 2024; 25: 8465. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Moussavi Javardi MS, Madani Z, Movahedi A, et al. The correlation between dietary fat quality indices and lipid profile with Atherogenic index of plasma in obese and non‐obese volunteers: A cross‐sectional descriptive‐analytic case‐control study. Lipids Health Dis 2020; 19: 213. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4. Mahdavi‐Roshan M, Mozafarihashjin M, Shoaibinobarian N, et al. Evaluating the use of novel atherogenicity indices and insulin resistance surrogate markers in predicting the risk of coronary artery disease: A case–control investigation with comparison to traditional biomarkers. Lipids Health Dis 2022; 21: 126. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Wen L, Wu Y, Fu X, et al. Atherogenic index of plasma and its 5‐year changes associated with type 2 diabetes risk: A 10‐year cohort study. Cardiovasc Diabetol 2025; 24: 349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Liu Z, Zhou D, Tang Y, et al. Association of Atherogenic Index of plasma with stroke risk in a population with cardiovascular‐kidney‐metabolic syndrome stages 0 to 3: The first evidence from the China health and retirement longitudinal study. J Am Heart Assoc 2025; 14: e039983. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Yang Q, Zhu X, Zhang L, et al. Dyslipidemia and aging: The non‐linear association between atherogenic index of plasma (AIP) and aging acceleration. Cardiovasc Diabetol 2025; 24: 181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Hajihashemi P, Mohammadifard N, Bateni M, et al. Comparing the association of novel anthropometric and atherogenicity indices with all‐cause, cardiovascular and non‐cardiovascular mortality in a general population of Iranian adults. Am J Prev Cardiol 2025; 21: 100936. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Ye W, Luo C, Huang J, et al. Gestational diabetes mellitus and adverse pregnancy outcomes: Systematic review and meta‐analysis. BMJ 2022; 377: e067946. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Xie W, Wang Y, Xiao S, et al. Association of gestational diabetes mellitus with overall and type specific cardiovascular and cerebrovascular diseases: Systematic review and meta‐analysis. BMJ 2022; 378: e070244. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Lan Y, Chen G, Wu D, et al. Temporal relationship between atherogenic dyslipidemia and inflammation and their joint cumulative effect on type 2 diabetes onset: A longitudinal cohort study. BMC Med 2023; 21: 31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Shi Y, Wen M. Sex‐specific differences in the effect of the atherogenic index of plasma on prediabetes and diabetes in the NHANES 2011–2018 population. Cardiovasc Diabetol 2023; 22: 19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Yin B, Wu Z, Xia Y, et al. Non‐linear association of atherogenic index of plasma with insulin resistance and type 2 diabetes: A cross‐sectional study. Cardiovasc Diabetol 2023; 22: 157. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Mulder J, Kusters DM, Roeters van Lennep JE, et al. Lipid metabolism during pregnancy: Consequences for mother and child. Curr Opin Lipidol 2024; 35: 133–140. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Zhang J, Suo Y, Wang L, et al. Association between atherogenic index of plasma and gestational diabetes mellitus: A prospective cohort study based on the Korean population. Cardiovasc Diabetol 2024; 23: 237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Zhang L, Yan X, Zheng W, et al. Assessment of first‐trimester insulin resistance indices for gestational diabetes mellitus: A prospective cohort study. J Endocrinol Invest 2025; 48: 2167–2176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Haj‐Husein I, Kubow S, Koski KG. Untargeted lipidomic profiling of amniotic fluid reveals dysregulated lipid metabolism in healthy normal‐weight mothers with fetal macrosomia. Nutrients 2024; 16: 3804. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Sweeting A, Wong J, Murphy HR, et al. A clinical update on gestational diabetes mellitus. Endocr Rev 2022; 43: 763–793. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. Yue W, Zhang E, Liu R, et al. The China birth cohort study (CBCS). Eur J Epidemiol 2022; 37: 295–304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Yang S, Gou X, Dong H, et al. Physical activity modifies the association between atherogenic index of plasma and prediabetes and diabetes: A cross‐sectional analysis. J Diabetes 2024; 16: e70006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Liao Y, Han Y, Cao C, et al. Association between atherogenic index of plasma and risk of type 2 diabetes mellitus and the mediating effect of BMI: A comparative analysis in Chinese and Japanese populations. Diabetol Metab Syndr 2025; 17: 349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Zhu SM, Zhang HQ, Li C, et al. Maternal lipid profile during early pregnancy and birth weight: A retrospective study. Front Endocrinol (Lausanne) 2022; 13: 951871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Isabel C, Faro Rebecca V, Vrijkotte TGM, et al. Early pregnancy triglycerides and not fructosamine are associated with birth weight (with foetal sexual dimorphism). Eur J Clin Invest 2023; 53: e13896. [DOI] [PubMed] [Google Scholar]
- 24. Adank MC, Benschop L, Kors AW, et al. Maternal lipid profile in early pregnancy is associated with foetal growth and the risk of a child born large‐for‐gestational age: A population‐based prospective cohort study: Maternal lipid profile in early pregnancy and foetal growth. BMC Med 2020; 18: 276. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Hwang LD, Lawlor DA, Freathy RM, et al. Using a two‐sample Mendelian randomization design to investigate a possible causal effect of maternal lipid concentrations on offspring birth weight. Int J Epidemiol 2019; 48: 1457–1467. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Rojas‐Rodriguez R, Price LL, Somogie J, et al. Maternal lipid metabolism is associated with neonatal adiposity: A longitudinal study. J Clin Endocrinol Metab 2022; 107: e3759–e3768. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Wang Y, Huang Y, Wu P, et al. Plasma lipidomics in early pregnancy and risk of gestational diabetes mellitus: A prospective nested case‐control study in Chinese women. Am J Clin Nutr 2021; 114: 1763–1773. [DOI] [PubMed] [Google Scholar]
- 28. Li N, Li J, Wang H, et al. Branched‐chain amino acids and their interactions with lipid metabolites for increased risk of gestational diabetes. J Clin Endocrinol Metab 2022; 107: e3058–e3065. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Albrecht M, Worthmann A, Heeren J, et al. Maternal lipids in overweight and obesity: Implications for pregnancy outcomes and offspring's body composition. Semin Immunopathol 2025; 47: 10. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Yeung RO, Retnakaran R, Savu A, et al. Gestational diabetes: One size does not fit all‐an observational study of maternal and neonatal outcomes by maternal glucose profile. Diabet Med 2024; 41: e15205. [DOI] [PubMed] [Google Scholar]
- 31. Ji J, Wu P, Li G, et al. The associations of ferritin, serum lipid and plasma glucose levels across pregnancy in women with gestational diabetes mellitus and newborn birth weight. BMC Pregnancy Childbirth 2023; 23: 478. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Herrera E, Ortega‐Senovilla H. Implications of lipids in neonatal body weight and fat mass in gestational diabetic mothers and non‐diabetic controls. Curr Diab Rep 2018; 18: 7. [DOI] [PubMed] [Google Scholar]
- 33. Zhao L, Huang Y, Wang L, et al. Trimester‐specific association of gestational Atherogenic index of plasma (AIP) with adverse pregnancy outcomes among women with and without GDM: A cohort study. BMC Endocr Disord 2026; 26: 87. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Niu Y, Cao H, Zhou H, et al. Effects of a vegetarian diet combined with exercise on lipid profiles and blood pressure: A systematic review and meta‐analysis. Crit Rev Food Sci Nutr 2024; 64: 2289–2303. [DOI] [PubMed] [Google Scholar]
- 35. Rigato M, Avogaro A, Vigili de Kreutzenberg S, et al. Effects of basal insulin on lipid profile compared to other classes of Antihyperglycemic agents in type 2 diabetic patients. J Clin Endocrinol Metab 2020; 105: dgaa178. [DOI] [PubMed] [Google Scholar]
- 36. Ren J, Lin Y, Cai W, et al. First‐trimester maternal triglycerides and atherogenic index of plasma in relation to large for gestational age and macrosomia among non‐diabetic pregnant women: A propensity score‐matched retrospective cohort study. Early Hum Dev 2026; 223: 106631. [DOI] [PubMed] [Google Scholar]
- 37. Zhang N, Wu B, Li J, et al. Atherogenic index of plasma in early pregnancy: Associations with gestational diabetes mellitus and preeclampsia and its mediating role. Lipids Health Dis 2026; 25: 147. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38. Montgomery MK, De Nardo W, Watt MJ. Impact of lipotoxicity on tissue “cross talk” and metabolic regulation. Physiology (Bethesda) 2019; 34: 134–149. [DOI] [PubMed] [Google Scholar]
- 39. Barbosa KB, Volp AC, Hermsdorff HH, et al. Relationship of oxidized low density lipoprotein with lipid profile and oxidative stress markers in healthy young adults: A translational study. Lipids Health Dis 2011; 10: 61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Tricò D, Mengozzi A, Baldi S, et al. Lipid‐induced glucose intolerance is driven by impaired glucose kinetics and insulin metabolism in healthy individuals. Metabolism 2022; 134: 155247. [DOI] [PubMed] [Google Scholar]
- 41. Anam AK, Cooke KM, Dratver MB, et al. Insulin increases placental triglyceride as a potential mechanism for fetal adiposity in maternal obesity. Mol Metab 2022; 64: 101574. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Waldrop SW, Niemiec S, Wood C, et al. Cord blood DNA methylation of immune and lipid metabolism genes is associated with maternal triglycerides and child adiposity. Obesity (Silver Spring) 2024; 32: 187–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Figure S1 Flowchart of study participants.
Figure S2 Association of first‐trimester AIP with birth weight mediated by FPG in participants with AIP <0.14 (A) and AIP ≥0.14 (B).
Figure S3 Association of first‐trimester AIP with birth weight mediated by FPG in male (A) and female (B) infants.
Table S1 Baseline characteristics of included and excluded participants.
Table S2 Variance inflation factors (VIF) in multifactor model.
Table S3 The pattern of missing AIP data.
Table S4 Threshold effect analysis of first‐trimester AIP levels and birth weight.
Table S5 Association between first‐trimester AIP, Macrosomia, and low birth weight.
Table S6 Association between first‐trimester AIP and second‐trimester blood glucose levels.
Table S7 Association between second‐trimester blood glucose levels and birth weight outcomes.
Table S8 Mixed‐effects model of GDM on AIP across gestational trimesters.
Table S9 Association between first‐trimester AIP and GDM, birth weight, LGA and SGA after additional adjustment for gestational age at lipid sampling.
Table S10 Association between first‐trimester AIP quartiles and GDM, birth weight, LGA and SGA.
Data Availability Statement
The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.