RESULTS
Baseline characteristics of the study population were showed inTable 1 . A total of 5,986 mother-newborn pairs were included in the analyses and 5.7% (n=344) of them were preterm birth. The mean maternal age at delivery was 32.2±4.7 years and the mean pre-pregnancy BMI was 20.7 ± 2.9 kg/m2. The median values (IQR) of ApoB, ApoA-1 and ApoB/ApoA-1 ratio were 2.07 (0.30) g/L, 0.96 (0.28) g/L and 0.47 (0.14), respectively. The results showed that preterm birth occurred more frequently in those mothers with more adverse risk profiles, including being older at delivery, multiparous, had higher pre-pregnancy BMI, with higher ApoB and ApoB/ApoA-1 levels during second trimester.
The relationships between maternal apolipoprotein levels and PTB were presented in Table 2. In the crude model, participants in the highest tertile of ApoB (OR, 1.54; 95% CI, 1.17-2.02; P for trend= 0.002) and ApoB/ApoA-1 (OR, 1.42; 95% CI, 1.09-1.85; Pfor trend= 0.007) had a significantly higher risk of PTB, compared to those in the lowest tertile. Adjustment for potential confounders did not materially alter the positive associations between ApoB, ApoB/ApoA-1 and PTB. Compared with the lowest tertiles, individuals in the highest tertile of ApoB had a 41% higher risk of PTB (OR, 1.41; 95% CI, 1.07-1.87; P for trend= 0.017) and those in the highest tertile of ApoB/ApoA-1 had a 35% higher risk of PTB (OR, 1.35; 95% CI, 1.03-1.78; P for trend= 0.024). There are no differences in the risk of PTB across the ApoA-1 tertiles. When continuous variables were used, each 1-g/L increase in ApoB and each 1 unit increase in ApoB/ApoA-1 were positively related to the risk of PTB (OR, 2.33; 95% CI 1.40-3.86 for ApoB and OR, 4.10; 95% CI, 1.47-11.36 for ApoB/ApoA-1). We further conducted restricted cubic spline to estimate the dose-response relationships between apolipoprotein levels and PTB (Fig. 2 ). The results indicated that ApoB (P for non-linearity= 0.943) and ApoB/ApoA-1 (P for non-linearity= 0.284) were both linearly associated with the risk of PTB, whereas ApoA-1 was not associated with PTB (P for all = 0.092).
The results of subgroup analyses were shown in Table 3 . Even though there was no interaction between all of the stratification variables and preterm birth (all P for interaction > 0.05), the positive relationship between maternal ApoB/ApoA-1 and preterm birth was stronger in women with pre-pregnancy BMI ≤ 24 kg/m2, age at delivery ≥ 35 years or those who were multiparous. In the sensitivity analyses, the positive relationships between ApoB, ApoB/ApoA-1 ratio during pregnancy and preterm birth were not materially changed after excluding pregnant women with GDM (Table 4 ).