Table 4.
Mediation analysis of the estimated effect of maternal glucose concentrations and neonatal anthropometric measures mediated by DNA methylation of genes related to the PPAR signaling pathway in the human placenta#
| Analyses | Total effect | Indirect effect | Direct effect | Proportion mediated (%) |
|---|---|---|---|---|
| β (95% CI) | β (95% CI) | β (95% CI) | ||
| FPG concentrations | ||||
| Abdominal skinfold thickness | ||||
| ACAA1b | 0.23 (0.00, 0.47)* | 0.07 (0.02, 0.14)** | 0.16 (−0.07, 0.40) | 29.00 |
| ACADMb | 0.22 (0.01, 0.43)** | 0.03 (0.00, 0.08)** | 0.19 (−0.03, 0.40) | 15.39 |
| CPT2a | 0.23 (0.01, 0.46)** | 0.03 (−0.00, 0.07)* | 0.21 (−0.02, 0.43)* | 11.38 |
| Back skinfold thickness | ||||
| ACAA1b | 0.32 (0.05, 0.58)** | 0.07 (0.00, 0.18)** | 0.25 (−0.04, 0.52)* | 21.75 |
| ACADMb | 0.33 (0.03, 0.61)** | 0.06 (0.00, 0.13)** | 0.27 (−0.04, 0.55)* | 17.69 |
PPAR, peroxisome proliferator-activated receptors; FPG, fasting plasma glucose; CI, confidence interval
#The models were adjusted for maternal age, pre-pregnancy BMI, maternal education, household income per capita, maternal passive smoking before gestation, parity, gestational age, infant sex, gestational week at the time of maternal glucose measurement, and gestational weight gain
aThe DNA methylation levels were not ln-transformed
bThe DNA methylation levels were ln-transformed
* 0.05 ≤ P < 0.1
** P < 0.05