Materials and Methods

This study was embedded in the Generation R study, a prospective cohort study from early pregnancy onwards in Rotterdam, The Netherlands. The study was approved by the medical ethics committee of the Erasmus University Medical Center, Rotterdam, The Netherlands. Written informed consent was obtained from all mothers for maternal and child data. From 2001 through 2005, 9778 women were included.

All prenatally enrolled Dutch women with a liveborn singleton (n = 4195) and with a CRL measurement were selected from the total study sample ( Figure 1 ). Women with no CRL measurement within the recommended gestational age (GA) range (10+0 weeks to 13+6 weeks) were excluded. Only mothers with a known first day of the last menstrual period (LMP) and a regular menstrual cycle of 28 ± 4 days were included. LMP was obtained of the referring letter from the community midwife or hospital and was verified at the inclusion moment. Additional information on regularity and cycle duration was obtained. If mothers participated in Generation R with ≥2 pregnancies, 1 sibling was randomly selected to avoid bias due to paired data (n = 130). After exclusion of women who underwent a fertility treatment (n = 10) or experienced a stillbirth or miscarriage (n = 21) 1491 women were eligible for the present study. Sensitivity analyses were carried out with exclusion of: (1) diabetics; and (2) women who later develop preeclampsia (PE). As this did not substantially change our results, these women were included in the analyses.

Flow chart

PAI , plasminogen activator inhibitor; PlGF , placental growth factor; sFlt , soluble fms-like tyrosine kinase.

Bouwland-Both. Maternal biomarkers in pregnancy and early fetal size. Am J Obstet Gynecol 2013 .

In early pregnancy, venous blood samples were drawn at the same moment as the CRL was measured (mean 12.4 weeks’ GA, SD 0.8). The samples were stored at room temperature before being transported to the regional laboratory for processing and storage for future studies. Processing was planned to finish within 3 hours after venous puncture. The samples were centrifuged and thereafter stored at –80°C. In plasma, sFlt-1, PlGF, and folate were analyzed using a microparticle-enhanced immunoassay on the Architect System (Abbott Diagnostics BV, Hoofddorp, The Netherlands) as described previously. Between-run coefficients of variation and analytical ranges are listed in Supplemental Table 1 ( Appendix ). Plasma PAI-2 was analyzed using enzyme-linked immunosorbent assay with the same experimental setup as described previously. The analytical and functional sensitivity were 11 pg/mL and 32 pg/mL, respectively. The mean PAI-2 was 88.4 pg/mL and the intraassay variation, between-plates variation, and interassay variation were 3.4%, 2.9%, and 8.4%, respectively.

Standard ultrasound planes for CRL measurement were used as described previously. Intraclass correlation coefficients for intraobserver (0.998) and interobserver (0.995) reproducibility of CRL measurements have been described before. GA-adjusted SD scores (SDS) for CRL were constructed and were based on reference growth curves from the entire study population.

From self-administered questionnaires, data were extracted on maternal age, education, marital status, parity, smoking, folic acid use, nausea/vomiting/fever during the first trimester, and repeated (≥2) miscarriages. Ethnicity was defined according to the classification of Statistics Netherlands. Education was assessed by the highest completed level of the mother and classified as: (1) low (none/primary); (2) medium (secondary); or (3) high (college/university). Maternal smoking habits were assessed for each trimester. Women, who reported any or no smoking during pregnancy were respectively classified as smokers and nonsmokers. Folic acid supplement use was assessed by asking women about the use and moment of initiation of a folic acid supplement. This information was used to categorize women into 3 groups of folic acid use: (1) adequate, defined as preconceptional initiation; (2) inadequate, defined as postconceptional initiation; and (3) no use. At enrollment maternal weight and height were measured to calculate body mass index (BMI) (kg/m ² ). Information on fertility treatment, pregnancy outcome, date of birth, birth anthropometrics, and the sex of the child were obtained from community midwives, obstetricians, and hospital registries.

Analyses were performed to correlate sFlt-1, PlGF, PAI-2, and folate with GA. Multivariable linear regression analyses were performed with the biomarker concentrations continuous and in quintiles as regressors and CRL (SDS) as main outcome measure. This approach was chosen to explore linearity and nonlinearity of the association. The lowest quintile of each biomarker was used as reference group. In the final analysis, we accounted for potential confounding variables, which were first selected on previous literature and determined a priori. Thereafter, covariates were selected as confounding variables if the effect estimates changed ≥10% in the exploratory analyses. By using this approach, maternal age, parity, smoking, and the use of a folic acid supplement (not for the analyses of folate) were included in the final analyses. Maternal ethnicity, maternal education, height, BMI, and fetal sex were included by default. Furthermore, as sFlt-1 and PlGF belong to the same pathway, analyses of sFlt-1 and PlGF were additionally adjusted for PlGF and sFlt-1, respectively. To assess the possibility of multicolinearity, we have calculated the variance inflation factor (VIF) and tolerance. If the VIF is <10 and/or the tolerance is >0.2, then there is no cause for concern.

Missing data of the covariables ethnicity (3.8%), education (4.9%), parity (0.7%), smoking (10.0%), height (0.3%), BMI (0.3%), and use of a folic acid supplement (19.2%) were completed using the Markov chain Monte Carlo multiple imputation technique for the linear regression models. Five datasets were created. Subsequently multivariable regression analyses were performed separately on each completed data set and combined to 1 pooled estimate. For all analyses, results including imputed missing data are presented.

Nonresponse analyses were performed by comparing characteristics of mothers (age, education, parity, smoking, folic acid supplement use, BMI) and birth characteristics of the children of mothers who were included in the analyses to those who were excluded using t , Mann-Whitney U , and χ ² tests. Furthermore, analyses were performed to examine correlations between the biomarkers and GA. All analyses were performed using software (SPSS, version 17.0; IBM Corp, Armonk, NY).

Results

The maternal and fetal characteristics are presented in Table 1 . Nonresponse analyses showed that women who had been excluded were less often Dutch (62.8%), were younger (mean = 29.5 years, SD 5.0), had lower education levels (9.4%), were more often multiparous (43.8%), and less frequently used a folic acid supplement (no use = 24.9%) (all P < .05).

Within the GA range (from 10+0 to 13+6 weeks’ GA), we observed positive correlations between GA and the concentrations of PlGF (Spearman ρ = 0.29, P < .001) and PAI-2 (Spearman ρ = 0.30, P < .001), and a small negative correlation with folate (Spearman ρ = –0.09, P = .002). Furthermore, positive correlations were observed between sFlt-1 and PlGF (Spearman ρ = 0.12, P < .001), sFlt-1 and PAI-2 (Spearman ρ = 0.23, P < .001), and PlGF and PAI-2 (Spearman ρ = 0.36, P < .001). No correlation was observed between folate and the other biomarkers.

The associations between maternal sFlt-1, PlGF, PAI-2, and folate levels and CRL (millimeters, SDS) are shown in Table 2 and Figure 2 . In all analysis the first quintile was taken as a reference category.

Maternal sFlt-1, PlGF, PAI-2, and folate concentrations and early fetal size

Values are regression coefficients (error bars indicate 95% confidence interval) and represent difference in crown-to-rump length (CRL) SD score (SDS), compared to quintile (Q)1, in A , sFlt-1, B , PlGF, C , PAI-2, and D , folate. Linear trend analyses reflect change in CRL (SDS) as continuous measure per unit (biomarker) change. All values are adjusted for gestational age at venous puncture (based on last menstrual period [LMP]), duration of LMP, maternal age, maternal height, maternal body mass index, maternal ethnicity, fetal sex, parity, educational level, smoking. Analyses of sFlt-1, PlGF, and PAI-2 were additionally adjusted for folic acid supplement use.

PAI-2, plasminogen activator inhibitor 2; PlGF, placental growth factor; sFlt-1, soluble fms-like tyrosine kinase 1.

Bouwland-Both. Maternal biomarkers in pregnancy and early fetal size. Am J Obstet Gynecol 2013 .

In the crude analyses, a larger CRL was observed in women with sFlt-1 levels in quintile (Q)5 (difference SDS = 0.16; 95% confidence interval [CI], 0.00–0.32). A significant linear trend was observed (change 0.15 mm; 95% CI, 0.03–0.26; SDS = 0.03; 95% CI, 0.01–0.04 per ng/mL change in sFlt-1). In the multivariable analyses, the association remained significant in Q5. After additional adjustment for PlGF levels, the associations between sFlt-1 and CRL remained in Q5, but slightly weakened (difference 1.33 mm; 95% CI, 0.09–2.57; SDS = 0.21; 95% CI, 0.04–0.39), with a significant linear trend (change 0.18 mm; 95% CI, 0.05–0.32; SDS = 0.03; 95% CI, 0.01–0.05 per ng/mL change in sFlt-1).

In the crude analyses, positive associations were observed between CRL and PlGF in Q3 (difference 1.58 mm; 95% CI, 0.48–2.69; SDS = 0.22; 95% CI, 0.06–0.38), Q4 (difference 2.73 mm; 95% CI, 1.60–3.86; SDS = 0.39; 95% CI, 0.23–0.56), and Q5 (difference 2.65 mm; 95% CI, 1.44–3.87; SDS = 0.34; 95% CI, 0.17–0.52). After adjustment for confounders in the multivariable analysis, all significant associations remained. After additional adjustment for sFlt-1, all associations between PlGF and CRL remained in Q3 (difference 1.96 mm; 95% CI, 0.75–3.17; SDS = 0.29; 95% CI, 0.12–0.46), Q4 (difference 2.86 mm; 95% CI, 1.63–4.09; SDS = 0.42; 95% CI, 0.24–0.59), and Q5 (difference 3.18 mm; 95% CI, 1.80–4.56; SDS = 0.42; 95% CI, 0.22–0.61).

In the crude analyses, women with PAI-2 levels in Q2 (difference 2.52 mm; 95% CI, 1.41–3.62; SDS = 0.38; 95% CI, 0.22–0.54), Q3 (difference 1.66 mm; 95% CI, 0.54–2.77; SDS = 0.25; 95% CI, 0.09–0.41), Q4 (difference 3.70 mm; 95% CI, 2.56–4.85; SDS = 0.52; 95% CI, 0.35–0.68), and Q5 (difference 3.23 mm; 95% CI, 2.02–4.43; SDS = 0.45; 95% CI, 0.27–0.62) showed associations with CRL. The linear trend tests were significant (change 0.06 mm; 95% CI, 0.03–0.09; SDS = 0.01; 95% CI, 0.00–0.01 per ng/mL change in PAI-2). In the multivariable analyses, the associations remained significant.

The crude analyses showed significant associations between folate and CRL in Q3 (difference 1.26 mm; 95% CI, 0.16–2.36) and Q5 (difference 1.86 mm; 95% CI, 0.75–2.96; SDS = 0.24; 95% CI, 0.08–0.40). Significant trends were observed between folate and CRL (change 0.06 mm; 95% CI, 0.02–0.09; SDS = 0.01; 95% CI, 0.00–0.01 per nmol/mL change in folate). In the multivariable analyses the associations remained significant in Q5.

Last, to assess multicolinearity, we calculated the VIF and tolerance. In all the models, the VIF was <10 and the tolerance was >0.2.

Results

The maternal and fetal characteristics are presented in Table 1 . Nonresponse analyses showed that women who had been excluded were less often Dutch (62.8%), were younger (mean = 29.5 years, SD 5.0), had lower education levels (9.4%), were more often multiparous (43.8%), and less frequently used a folic acid supplement (no use = 24.9%) (all P < .05).

Table 1

Baseline characteristics

Characteristic	n = 1491
Maternal
Age, a y	30.7 (4.7)
GA at intake, a wk	12.4 (0.8)
Height, a cm	168.6 (7.1)
Body mass index at intake, b kg/m²	23.5 (21.5-26.0)
Ethnicity, %
Dutch/Caucasian	67.9
Surinamese	5.5
Turkish	6.1
Moroccan	3.8
Indonesian	3.2
Other	9.7
Missing	3.8
Education, %
Low	5.9
Medium	38.5
High	50.7
Missing	4.9
Gravidity, %
Primigravida	48.8
Multigravida	50.6
Missing	0.6
Parity, %
Nullipara	61.0
Multipara	38.3
Missing	0.7
Folic acid supplement use, %
Adequate (preconception initiation)	42.0
Inadequate (postconception initiation)	26.3
No	12.5
Missing	19.2
Smoking, %
Yes	13.5
Until pregnancy recognition	9.0
No	67.5
Missing	10.0
Daily nausea, %	30.0
Daily vomiting, %	7.0
Fetal
Crown-to-rump length, a mm	60.9 (11.4)
Male, %	49.0
GA at birth, b wk c	40.1 (39.0-41.0)
Birthweight, a g	3455 (563)

Only gold members can continue reading. Log In or Register to continue

Share this:

• Click to share on Twitter (Opens in new window)
• Click to share on Facebook (Opens in new window)

Related

Related posts:

1. Does the dawn phenomenon have clinical relevance in normal pregnancy?
2. Outcomes of infants exposed to oseltamivir or zanamivir in utero during pandemic (H1N1) 2009
3. Reply
4. Reply

Stay updated, free articles. Join our Telegram channel

Join