- Split View
-
Views
-
Cite
Cite
F. Parisi, M. Rousian, A.H.J. Koning, S.P. Willemsen, I. Cetin, R.P.M. Steegers-Theunissen, Periconceptional maternal one-carbon biomarkers are associated with embryonic development according to the Carnegie stages, Human Reproduction, Volume 32, Issue 3, March 2017, Pages 523–530, https://doi.org/10.1093/humrep/dew349
- Share Icon Share
Abstract
Is periconceptional maternal one-carbon (I-C) metabolism associated with embryonic morphological development in non-malformed ongoing pregnancies?
Serum vitamin B12, red blood cell (RBC) folate and plasma total homocysteine (tHcy) are associated with embryonic development according to the Carnegie stages.
Derangements in maternal I-C metabolism affect reproductive and pregnancy outcomes, as well as future health of the offspring.
Between 2010 and 2014, women with singleton ongoing pregnancies were enrolled in a prospective periconceptional cohort study.
A total of 234 pregnancies, including 138 spontaneous or IUI pregnancies with strict pregnancy dating and 96 pregnancies derived from IVF, ICSI or cryopreserved embryo transfer (IVF/ICSI pregnancies), underwent longitudinal transvaginal three-dimensional ultrasound (3D US) scans from 6+0 up to 10+2 weeks of gestation. Carnegie stages were defined using internal and external morphologic criteria in a virtual reality system. Maternal venous blood samples were collected at enrollment for serum vitamin B12, RBC folate and plasma tHcy assessment. Associations between biomarker concentrations and longitudinal Carnegie stages were investigated using linear mixed models.
We performed a median of three 3D US scans per pregnancy (range 1–5) resulting in 600 good quality data sets for the Carnegie stage annotation (80.5%). Vitamin B12 was positively associated with embryonic development in the total study population (β = 0.001 (95% CI: 0.000; 0.002), P < 0.05) and in the subgroup of strictly dated spontaneous pregnancies (β = 0.002 (95% CI: 0.001; 0.003), P < 0.05). Low vitamin B12 concentrations (−2SD, 73.4 pmol/l) were associated with delayed embryonic development by 1.4 days (95% CI: 1.3–1.4) compared with high concentrations (+2SD, 563.1 pmol/l). RBC folate was positively associated with Carnegie stages only in IVF/ICSI pregnancies (β = 0.001 (95% CI: 0.0005; 0.0015), P < 0.05). In this group, low RBC folate concentrations (−2SD, 875.4 nmol/l) were associated with a 1.8-day delay (95% CI: 1.7–1.8) in development compared with high concentrations (+2SD, 2119.9 nmol/l). tHcy was negatively associated with embryonic development in the total study population (β = −0.08 (95% CI: −0.14; −0.02), P < 0.01), as well as in the IVF/ICSI subgroup (β = −0.08 (95% CI: −0.15; −0.01), P < 0.05). High tHcy concentrations (+2SD, 10.4 µmol/l) were associated with a delay of 1.6 days (95% CI: 1.5–1.7) in embryonic development compared with low concentrations (−2SD, 3.0 µmol/l).
The study was performed in a tertiary care center, resulting in high rates of folic acid supplement use and comorbidity that may reduce the external validity of our findings.
In periconceptional care, maternal I-C biomarkers should be taken into account as predictors of embryonic morphological development. Combining embryonic size measurements with morphological assessment could better define normal embryonic development.
The work was funded by the Department of Obstetrics and Gynaecology, Erasmus MC, University Medical Centre, Rotterdam, The Netherlands. RPMST is CSO of the startup company Slimmere Zorg and CEO of eHealth Care Solutions. The authors declare no conflicts of interest.
Not applicable.
Introduction
One-carbon (I-C) metabolism is known to play a crucial role in cellular metabolism and proliferation, as well as in the regulation of gene expression through epigenetic mechanisms. Useful biomarkers of I-C metabolism for research and clinical practice are serum vitamin B12 and folate, red blood cell (RBC) folate and plasma total homocysteine (tHcy). Several studies have linked maternal I-C biomarkers to reproductive, pregnancy and health outcomes in the offspring (Bergen et al., 2012; Steegers-Theunissen et al., 2013; van Uitert and Steegers-Theunissen, 2013a; Yajnik et al., 2014; Kalhan, 2016). Most evidence is available on the association between maternal folate deficiency, folic acid supplement use and congenital anomalies (Steegers-Theunissen et al., 2013). Nevertheless, the plasma tHcy concentration seems to be a more sensitive marker, with increased concentrations strongly associated with miscarriage, hypertensive disorders, preterm birth and birth defects (Steegers-Theunissen et al., 1991; Vollset et al., 2000; Ronnenberg et al., 2007; Hogeveen et al., 2012). Due to the increased adherence to a vegetarian diet and the frequent association with vitamin deficiency, recent research has also focused on the associations between vitamin B12, birth defects and birth weight (Finkelstein et al., 2015).
The introduction of high-resolution three-dimensional ultrasound (3D US) scans combined with visualization in immersive virtual reality (VR) systems, providing real depth perception and more sensitive embryonic size measurements and morphological evaluations, has markedly improved the opportunity to accurately study the periconceptional period (time window: 14 weeks pre-conception to 10 weeks post-conception) (Rousian et al., 2010; Steegers-Theunissen et al., 2013; Baken et al., 2014). So far these innovative techniques have been used to study embryonic crown-rump length (CRL) and volume (EV) trajectories as non-invasive measures of first trimester embryonic growth (Steegers-Theunissen et al., 2016). On the other hand, the Carnegie stages of human embryonic development were introduced as a century-old morphological classification of fixated embryos dividing the embryonic period (58 post-conceptional days) into 23 stages (O'Rahilly, 1987). The combination of 3D US and VR visualization allows us to investigate embryonic morphological development in vivo, according to the longitudinal annotation of the Carnegie stages (Blaas et al., 1998; Verwoerd-Dikkeboom et al., 2008; O'Rahilly and Müller, 2010). Despite the fact that the normal sequence of developmental events is constant and predictable in every embryo, different times and velocities can occur, making comparisons possible and worthwhile.
Here, we aimed to investigate the associations between periconceptional maternal biomarkers of I-C metabolism and first trimester embryonic development, using serial Carnegie stage annotation obtained by 3D US and VR.
Materials and Methods
This study was performed in the setting of the Rotterdam Periconception Cohort (Predict Study), a prospective periconceptional tertiary hospital-based cohort study started in 2009 at the Department of Obstetrics and Gynaecology of the Erasmus MC, University Medical Centre, Rotterdam, with the aim to assess periconceptional determinants and predictors of pregnancy outcome and offspring health (Steegers-Theunissen et al., 2016).
Study population and sample
General data
Self-administered general questionnaires reporting items on age, geographical origin, education, obstetric and medical history, and periconceptional lifestyle (smoking, alcohol consumption, folic acid and multivitamin supplement use) were collected at enrollment. Anthropometric measures were recorded by trained researchers.
Blood sample analysis
One first trimester fasting venous blood sample for serum vitamin B12, RBC folate and plasma tHcy assessment was collected at enrollment and drawn in a vacutainer EDTA tube and in a dry vacutainer tube (BD Diagnostics, Plymouth, UK). The dry vacutainer tubes were centrifuged at 2000g, and serum was collected and analyzed for vitamin B12 measurement using an immunoelectro-chemoluminescence assay (E170; Roche Diagnostics GmbH, Mannheim, Germany). Plasma was separated by centrifugation within 1 h for determination of tHcy by using a sensitive liquid chromatography tandem mass spectrum method (HPLC-Tandem MS, Waters Micromass Quattro Premier XE Mass Spectrometer with Acquity UPLC system, Milford, MA, USA). EDTA-blood was kept on ice and 0.1 ml EDTA-blood was hemolysed with 0.9 ml freshly prepared 1.0% ascorbic acid. The hematocrit was determined with the ADVIA 120 Hematology Analyzer (Bayer Diagnostics, Leverkusen, Germany). RBC folate was calculated with the following formula: (nM hemolysate folate × 10/hematocrit) − [nM serum folate × (1 − hematocrit) /hematocrit] = nM RBC folate.
Ultrasound data
From 6+0 up to 10+2 weeks of gestation, all included women underwent serial 3D US scans performed by trained researchers using the high frequency (4.5–11.9 MHz) vaginal probe of a GE Voluson E8 (GE Healthcare, Zipf, Austria). Ultrasound scans were performed on a weekly basis until 2013 and then reduced to a two weekly basis after the pilot study showed an accurate modeling of growth trajectory obtained with three scans per pregnancy (at 7, 9, 11 weeks of gestation) (van Uitert et al., 2013b). The obtained 3D data sets were stored as Cartesian volumes and transferred to the BARCO I-Space VR system at the Department of Bioinformatics, Erasmus University Medical Centre, Rotterdam. This system, running the V-Scope volume rendering application, aims to improve data set visualization by projecting a hologram in a 4-walled CAVE-like (Cave Automatic Virtual Environment) VR system, allowing full depth perception and intuitive interaction with the volume (Koning et al., 2009; Verwoerd-Dikkeboom et al., 2010). The Carnegie criteria for external and internal morphological characteristics were used by one trained researcher to stage all embryos, as previously described (Verwoerd-Dikkeboom et al., 2008; O'Rahilly and Müller, 2010). As external morphological characteristics, we used the Carnegie criteria for the development of limbs (arms and legs) and embryonic curvature. Internal morphological characteristics primarily included the criteria for the brain cavity development. The assessment of Carnegie stages required 1–2 min per embryo.
Statistical analysis
In order to evaluate selection bias, we compared maternal baseline characteristics and biomarker concentrations between excluded and included pregnancies using Chi-square or exact tests for ordinal variables and Mann–Whitney U test for continuous variables. Univariable linear regression was performed to evaluate associations between maternal baseline characteristics and biomarker concentrations.
To estimate associations between maternal biomarkers of I-C metabolism and embryonic development, we treated the Carnegie stages as a continuous variable that was censored at its maximum value of 23. This was used as the response variable in separate linear mixed models estimated for the total study population and second for the subgroups of strictly dated spontaneous and IVF/ICSI pregnancies. This analysis allows the linear modeling of longitudinal measurements, taking into account the existing correlation between serial measurements within the same pregnancy and potential confounders for adjustment (parity, alcohol use, smoking, folic acid/multivitamin supplement use, fetal gender, maternal age, BMI and comorbidity). First, we performed a crude analysis with adjustment for gestational age only (Model 1) and second we adjusted for additional confounders (Model 2). Finally, the estimates of embryonic developmental change expressed in days were determined comparing women with high (+2SD) and low (−2SD) concentrations of the biomarkers that were significantly associated with the Carnegie stages in Model 2. Due to the exclusion of pregnancies with uncertain dating and the possibility of selection bias, we additionally performed a sensitivity analysis including pregnancies with discordant CRL (n = 15). P-values ≤0.05 were considered as significant. All analyses were performed using IBM SPSS version 21.0 (IBM Corp, Armonk, NY) and R version 3.2.1 (The R Foundation for Statistical Computing).
Ethical approval
The protocol had been approved by the local medical ethics committee and all women signed a written informed consent before participation.
Results
We included 234 pregnancies with a median of three scans per pregnancy (range 1–5), counting for a total of 745 3D US scans. The Carnegie stage annotation was feasible in 600 good quality data sets (success rate 80.5%). Carnegie stage distribution in the total study population ranged from stage 13 to 23 (6+0–10+2 weeks of gestation). Table I shows maternal characteristics and biomarker concentrations at baseline with comparisons between included and excluded ongoing pregnancies. The prevalence of hyperhomocysteinemia in the total study population was 1.3% (>13 µmol/l). Vitamin B12 was significantly associated with maternal age (β = 0.15 pmol/l (95% CI: 0.13; 0.16), P < 0.05), RBC folate (β = 0.20 pmol/l (95% CI: 0.19; 0.20), P < 0.01) and tHcy concentrations (β = −0.34 pmol/l (95% CI: −0.37; −0.32), P < 0.001). RBC folate was significantly associated with maternal age (β = 0.20 nmol/l (95% CI: 0.19; 0.21), P < 0.01), smoking (β = −0.16 nmol/l (95% CI: −0.25; −0.07), P < 0.05), folic acid supplement use (β = 0.23 nmol/l (95% CI: 0.04; 0.43), P < 0.001), comorbidity (β = −0.14 nmol/l (95% CI: −0.24; −0.04), P < 0.05) and tHcy concentrations (β = −0.25 nmol/l (95% CI: −0.27; −0.24), P < 0.001).
Maternal characteristics . | Total study population (n = 234) . | M . | Excluded population (n = 118) . | M . | P-value . |
---|---|---|---|---|---|
Age (years), median (range) | 32 (22–42) | 0 | 30 (21–44) | 0 | 0.00 |
Geographical origin | 1 | 5 | 0.30 | ||
Western, n (%) | 206 (88.0) | 104 (88.1) | |||
Non-Western, n (%) | 27 (11.5) | 9 (7.6) | |||
Educational level | 1 | 5 | 0.08 | ||
High, n (%) | 135 (57.7) | 65 (55.1) | |||
Intermediate, n (%) | 93 (39.7) | 45 (38.1) | |||
Low, n (%) | 5 (2.1) | 3 (2.5) | |||
BMI (kg/m2), median (range) | 24.2 (17–42.3) | 1 | 25.8 (17.8–45.0) | 2 | 0.01 |
Nulliparous, n (%) | 74 (31.8) | 1 | 39 (33.9) | 2 | 0.69 |
Alcohol use, n (%) | 83 (35.8) | 2 | 38 (34.2) | 7 | 0.78 |
Periconception smoking, n (%) | 32 (13.7) | 1 | 21 (19.1) | 8 | 0.20 |
Periconception folic acid/multivitamin use, n (%) | 224 (97.4) | 4 | 108 (93.9) | 3 | 0.11 |
Chronic diseases, n (%) | 25 (10.7) | 0 | 22 (18.6) | 0.05 | |
Vitamin B12 (pmol/l), median (range) | 297 (95–953) | 0 | 295.5 (109–915) | 20 | 0.76 |
RBC folate (nmol/l), median (range) | 1408 (541–2811) | 12 | 1294 (634–1942) | 23 | 0.01 |
tHcy (µmol/l), median (range) | 6.4 (3.7–17.6) | 3 | 6.2 (3.4–13.6) | 23 | 0.51 |
Maternal characteristics . | Total study population (n = 234) . | M . | Excluded population (n = 118) . | M . | P-value . |
---|---|---|---|---|---|
Age (years), median (range) | 32 (22–42) | 0 | 30 (21–44) | 0 | 0.00 |
Geographical origin | 1 | 5 | 0.30 | ||
Western, n (%) | 206 (88.0) | 104 (88.1) | |||
Non-Western, n (%) | 27 (11.5) | 9 (7.6) | |||
Educational level | 1 | 5 | 0.08 | ||
High, n (%) | 135 (57.7) | 65 (55.1) | |||
Intermediate, n (%) | 93 (39.7) | 45 (38.1) | |||
Low, n (%) | 5 (2.1) | 3 (2.5) | |||
BMI (kg/m2), median (range) | 24.2 (17–42.3) | 1 | 25.8 (17.8–45.0) | 2 | 0.01 |
Nulliparous, n (%) | 74 (31.8) | 1 | 39 (33.9) | 2 | 0.69 |
Alcohol use, n (%) | 83 (35.8) | 2 | 38 (34.2) | 7 | 0.78 |
Periconception smoking, n (%) | 32 (13.7) | 1 | 21 (19.1) | 8 | 0.20 |
Periconception folic acid/multivitamin use, n (%) | 224 (97.4) | 4 | 108 (93.9) | 3 | 0.11 |
Chronic diseases, n (%) | 25 (10.7) | 0 | 22 (18.6) | 0.05 | |
Vitamin B12 (pmol/l), median (range) | 297 (95–953) | 0 | 295.5 (109–915) | 20 | 0.76 |
RBC folate (nmol/l), median (range) | 1408 (541–2811) | 12 | 1294 (634–1942) | 23 | 0.01 |
tHcy (µmol/l), median (range) | 6.4 (3.7–17.6) | 3 | 6.2 (3.4–13.6) | 23 | 0.51 |
The total study population includes strictly dated pregnancies achieved after spontaneous conception (n = 138) or IVF/ICSI (n = 96). Excluded pregnancies include oocyte(s) donation (n = 5), missing 3D US scans before 10+2 weeks of gestation (n = 7) and spontaneous pregnancies with discordant CRL measurements (≥7 days, n = 15), unknown LMP (n = 14) or self-reported irregular cycle (n = 77). Chronic diseases include cardiovascular, autoimmune, endocrine and metabolic diseases. The comparison was performed using Chi-square or exact tests for ordinal variables and Mann–Whitney U test for continuous variables. M, missing values; RBC, red blood cell; tHcy, total homocysteine; LMP, last menstrual period; CRL, crown-rump length.
Maternal characteristics . | Total study population (n = 234) . | M . | Excluded population (n = 118) . | M . | P-value . |
---|---|---|---|---|---|
Age (years), median (range) | 32 (22–42) | 0 | 30 (21–44) | 0 | 0.00 |
Geographical origin | 1 | 5 | 0.30 | ||
Western, n (%) | 206 (88.0) | 104 (88.1) | |||
Non-Western, n (%) | 27 (11.5) | 9 (7.6) | |||
Educational level | 1 | 5 | 0.08 | ||
High, n (%) | 135 (57.7) | 65 (55.1) | |||
Intermediate, n (%) | 93 (39.7) | 45 (38.1) | |||
Low, n (%) | 5 (2.1) | 3 (2.5) | |||
BMI (kg/m2), median (range) | 24.2 (17–42.3) | 1 | 25.8 (17.8–45.0) | 2 | 0.01 |
Nulliparous, n (%) | 74 (31.8) | 1 | 39 (33.9) | 2 | 0.69 |
Alcohol use, n (%) | 83 (35.8) | 2 | 38 (34.2) | 7 | 0.78 |
Periconception smoking, n (%) | 32 (13.7) | 1 | 21 (19.1) | 8 | 0.20 |
Periconception folic acid/multivitamin use, n (%) | 224 (97.4) | 4 | 108 (93.9) | 3 | 0.11 |
Chronic diseases, n (%) | 25 (10.7) | 0 | 22 (18.6) | 0.05 | |
Vitamin B12 (pmol/l), median (range) | 297 (95–953) | 0 | 295.5 (109–915) | 20 | 0.76 |
RBC folate (nmol/l), median (range) | 1408 (541–2811) | 12 | 1294 (634–1942) | 23 | 0.01 |
tHcy (µmol/l), median (range) | 6.4 (3.7–17.6) | 3 | 6.2 (3.4–13.6) | 23 | 0.51 |
Maternal characteristics . | Total study population (n = 234) . | M . | Excluded population (n = 118) . | M . | P-value . |
---|---|---|---|---|---|
Age (years), median (range) | 32 (22–42) | 0 | 30 (21–44) | 0 | 0.00 |
Geographical origin | 1 | 5 | 0.30 | ||
Western, n (%) | 206 (88.0) | 104 (88.1) | |||
Non-Western, n (%) | 27 (11.5) | 9 (7.6) | |||
Educational level | 1 | 5 | 0.08 | ||
High, n (%) | 135 (57.7) | 65 (55.1) | |||
Intermediate, n (%) | 93 (39.7) | 45 (38.1) | |||
Low, n (%) | 5 (2.1) | 3 (2.5) | |||
BMI (kg/m2), median (range) | 24.2 (17–42.3) | 1 | 25.8 (17.8–45.0) | 2 | 0.01 |
Nulliparous, n (%) | 74 (31.8) | 1 | 39 (33.9) | 2 | 0.69 |
Alcohol use, n (%) | 83 (35.8) | 2 | 38 (34.2) | 7 | 0.78 |
Periconception smoking, n (%) | 32 (13.7) | 1 | 21 (19.1) | 8 | 0.20 |
Periconception folic acid/multivitamin use, n (%) | 224 (97.4) | 4 | 108 (93.9) | 3 | 0.11 |
Chronic diseases, n (%) | 25 (10.7) | 0 | 22 (18.6) | 0.05 | |
Vitamin B12 (pmol/l), median (range) | 297 (95–953) | 0 | 295.5 (109–915) | 20 | 0.76 |
RBC folate (nmol/l), median (range) | 1408 (541–2811) | 12 | 1294 (634–1942) | 23 | 0.01 |
tHcy (µmol/l), median (range) | 6.4 (3.7–17.6) | 3 | 6.2 (3.4–13.6) | 23 | 0.51 |
The total study population includes strictly dated pregnancies achieved after spontaneous conception (n = 138) or IVF/ICSI (n = 96). Excluded pregnancies include oocyte(s) donation (n = 5), missing 3D US scans before 10+2 weeks of gestation (n = 7) and spontaneous pregnancies with discordant CRL measurements (≥7 days, n = 15), unknown LMP (n = 14) or self-reported irregular cycle (n = 77). Chronic diseases include cardiovascular, autoimmune, endocrine and metabolic diseases. The comparison was performed using Chi-square or exact tests for ordinal variables and Mann–Whitney U test for continuous variables. M, missing values; RBC, red blood cell; tHcy, total homocysteine; LMP, last menstrual period; CRL, crown-rump length.
Embryonic development
Biomarkers . | Effect estimates Carnegie stages, β (95% CI) . | |
---|---|---|
Model 1 . | Model 2 . | |
Total study population (n = 234) | ||
Vitamin B12 | 0.001 (0.000; 0.002)* | 0.001 (0.000; 0.002)* |
RBC folate | 0.0004 (0.0001; 0.0007)* | 0.000 (0.000; 0.001) |
tHcy | −0.09 (−0.15; −0.03)** | −0.08 (−0.14; −0.02)** |
Strictly dated spontaneous pregnancies (n = 138) | ||
Vitamin B12 | 0.002 (0.001; 0.003)* | 0.002 (0.001; 0.003)* |
RBC folate | 0.000 (−0.000; 0.001) | 0.003 (0.002; 0.004) |
tHcy | −0.07 (−0.17; 0.03) | −0.07 (−0.10; 0.02) |
IVF/ICSI pregnancies (n = 96) | ||
Vitamin B12 | −0.0004 (−0.002; 0.0008) | −0.000 (−0.002; 0.001) |
RBC folate | 0.000 (−0.000; 0.001) | 0.001 (0.0005; 0.0015)* |
tHcy | −0.09 (−0.16; −0.02)** | −0.08 (−0.15; -0.01)* |
Biomarkers . | Effect estimates Carnegie stages, β (95% CI) . | |
---|---|---|
Model 1 . | Model 2 . | |
Total study population (n = 234) | ||
Vitamin B12 | 0.001 (0.000; 0.002)* | 0.001 (0.000; 0.002)* |
RBC folate | 0.0004 (0.0001; 0.0007)* | 0.000 (0.000; 0.001) |
tHcy | −0.09 (−0.15; −0.03)** | −0.08 (−0.14; −0.02)** |
Strictly dated spontaneous pregnancies (n = 138) | ||
Vitamin B12 | 0.002 (0.001; 0.003)* | 0.002 (0.001; 0.003)* |
RBC folate | 0.000 (−0.000; 0.001) | 0.003 (0.002; 0.004) |
tHcy | −0.07 (−0.17; 0.03) | −0.07 (−0.10; 0.02) |
IVF/ICSI pregnancies (n = 96) | ||
Vitamin B12 | −0.0004 (−0.002; 0.0008) | −0.000 (−0.002; 0.001) |
RBC folate | 0.000 (−0.000; 0.001) | 0.001 (0.0005; 0.0015)* |
tHcy | −0.09 (−0.16; −0.02)** | −0.08 (−0.15; -0.01)* |
Effect estimates represent the change in Carnegie stage per unit of increase of biomarker concentration. Model 1 shows the crude model with adjustment for gestational age. Model 2 includes adjustment for potential confounders (parity, alcohol use, smoking habit, folic acid use, age, BMI, chronic diseases, fetal gender). Bold text shows significant results. *P < 0.05, **P ≤ 0.01.
Biomarkers . | Effect estimates Carnegie stages, β (95% CI) . | |
---|---|---|
Model 1 . | Model 2 . | |
Total study population (n = 234) | ||
Vitamin B12 | 0.001 (0.000; 0.002)* | 0.001 (0.000; 0.002)* |
RBC folate | 0.0004 (0.0001; 0.0007)* | 0.000 (0.000; 0.001) |
tHcy | −0.09 (−0.15; −0.03)** | −0.08 (−0.14; −0.02)** |
Strictly dated spontaneous pregnancies (n = 138) | ||
Vitamin B12 | 0.002 (0.001; 0.003)* | 0.002 (0.001; 0.003)* |
RBC folate | 0.000 (−0.000; 0.001) | 0.003 (0.002; 0.004) |
tHcy | −0.07 (−0.17; 0.03) | −0.07 (−0.10; 0.02) |
IVF/ICSI pregnancies (n = 96) | ||
Vitamin B12 | −0.0004 (−0.002; 0.0008) | −0.000 (−0.002; 0.001) |
RBC folate | 0.000 (−0.000; 0.001) | 0.001 (0.0005; 0.0015)* |
tHcy | −0.09 (−0.16; −0.02)** | −0.08 (−0.15; -0.01)* |
Biomarkers . | Effect estimates Carnegie stages, β (95% CI) . | |
---|---|---|
Model 1 . | Model 2 . | |
Total study population (n = 234) | ||
Vitamin B12 | 0.001 (0.000; 0.002)* | 0.001 (0.000; 0.002)* |
RBC folate | 0.0004 (0.0001; 0.0007)* | 0.000 (0.000; 0.001) |
tHcy | −0.09 (−0.15; −0.03)** | −0.08 (−0.14; −0.02)** |
Strictly dated spontaneous pregnancies (n = 138) | ||
Vitamin B12 | 0.002 (0.001; 0.003)* | 0.002 (0.001; 0.003)* |
RBC folate | 0.000 (−0.000; 0.001) | 0.003 (0.002; 0.004) |
tHcy | −0.07 (−0.17; 0.03) | −0.07 (−0.10; 0.02) |
IVF/ICSI pregnancies (n = 96) | ||
Vitamin B12 | −0.0004 (−0.002; 0.0008) | −0.000 (−0.002; 0.001) |
RBC folate | 0.000 (−0.000; 0.001) | 0.001 (0.0005; 0.0015)* |
tHcy | −0.09 (−0.16; −0.02)** | −0.08 (−0.15; -0.01)* |
Effect estimates represent the change in Carnegie stage per unit of increase of biomarker concentration. Model 1 shows the crude model with adjustment for gestational age. Model 2 includes adjustment for potential confounders (parity, alcohol use, smoking habit, folic acid use, age, BMI, chronic diseases, fetal gender). Bold text shows significant results. *P < 0.05, **P ≤ 0.01.
Discussion
This study shows significant associations between periconceptional maternal biomarkers of I-C metabolism and embryonic morphological development according to the Carnegie classification in ongoing non-malformed pregnancies. Moreover, IVF/ICSI conception did not affect embryonic morphological development compared with spontaneous conception in strictly dated pregnancies. The inclusion of pregnancies with discordant CRL revealed the same associations.
Our results are in line with previous data showing associations between maternal I-C metabolism and several reproductive, pregnancy and health outcomes (Yajnik and Deshmukh, 2012; Solé-Navais et al., 2016). Recently, maternal early pregnancy high tHcy (≥8.31 μmol/l) and low folate concentrations (≤9.10 nmol/l) have been negatively associated with fetal growth parameters, finally affecting birth weight (Bergen et al., 2016). We also showed that an optimal periconceptional RBC folate level is associated with increased first trimester longitudinal CRL measurements compared with the lowest (β = 0.24 √mm (95% CI: 0.04; 0.44), P = 0.02) and highest (β = 0.29 √mm (95% CI: 0.09; 0.49), P < 0.01) quartiles of concentrations (van Uitert et al., 2014). This result emphasizes that CRL accuracy in pregnancy dating is impacted by maternal I-C metabolism, as well as by several maternal characteristics and exposures (van Uitert et al., 2013b). Moreover, EV has been described as a more sensitive marker of first trimester growth restriction compared with CRL (Baken et al., 2013). We focused on the Carnegie stages as a century-old classification that, together with embryonic size measurements, could implement the first trimester investigation and better define a proper embryonic development. Since we excluded all pregnancies with congenital anomalies detected both in utero and after birth, our results indicate that even the developmental events of normal ongoing pregnancies are impacted by maternal I-C metabolism. This and previous findings indicate that first trimester growth and development are important embryonic outcomes affected by maternal environment. Nevertheless, CRL, EV and Carnegie stages also represent non-invasive reproducible markers with predictable associations with gestational age, leading to their potential use for pregnancy dating and raising the question of which biomarker should be the best candidate (Robinson and Fleming, 1975; O'Rahilly and Müller, 2010). Due to the lack of an optimal pregnancy dating strategy and to unavoidable systematic errors related to the recall of the LMP, imprecise ovulation/implantation dates and parental characteristics impacting embryonic ultrasound measurements, we defined gestational age based on a known LMP, regular cycle and concordant CRL. In this way, all ultrasound measurements could be read as response variables and outcome measurements. In order to reduce selection bias, we compared maternal baseline characteristics, showing that excluded women had a higher BMI, lower age and lower RBC folate concentrations. This may be mainly explained by the inclusion of a large population of subfertile women and pregnancies achieved after IVF/ICSI treatment (higher age, lower BMI, higher use of folic acid supplements). We also compared the subgroup of included and excluded spontaneous pregnancies showing indeed no significant results (data not shown). Moreover, the sensitivity analysis including pregnancies with discordant CRL confirmed the detected associations, reducing the possibility of selection bias.
The mechanisms linking maternal I-C metabolism and embryonic development are not fully understood. Animal data showed that abnormal activations of I-C metabolism were associated with hypermethylation of mitochondrial DNA, mitochondrial malfunction and decreased oocyte quality (Jia et al., 2016). Recently, a suppression of the inflammatory and upregulation of the high-density lipoprotein pathways have been demonstrated in human follicular fluid of pre-conception folic acid supplement users (Twigt et al., 2015). Cellular apoptosis and protein homocysteinylation, both dependent on tHcy concentrations, have been suggested as contributors to neural tube, orofacial and cardiac defects (Jakubowski, 2006; Taparia et al., 2007). Finally, periconceptional I-C biomarker-mediated epigenetic modifications could modify subsequent gene expression in the embryo (Steegers-Theunissen et al., 2013). All these events may finally lead to impaired first trimester development, thereby supporting our results.
Our findings also reveal that conception mode seems to modify the associations between blood biomarkers and Carnegie stages, despite the fact that no differences in embryonic development have been detected between the two subgroups. As expected, biomarker concentrations differed between spontaneous and IVF/ICSI pregnancies. Besides higher and longer pre-conceptional folic acid supplement use in the IVF/ICSI subgroup, also the ovarian stimulation treatment may affect I-C blood biomarker concentrations (Boxmeer et al., 2008). Moreover, the IVF/ICSI technique has been associated with different epigenetic patterns, gene expression and preimplantation embryo phenotypes compared with spontaneous conception, possibly affecting embryonic responses to maternal I-C biomarkers and explaining different associations detected in our results (Giritharan et al., 2007; Song et al., 2015; Zandstra et al., 2015; Kroener et al., 2016).
The major strength of our study is the longitudinal evaluation of embryonic development using a median of three scans per patient, the use of 3D US with VR visualization and the consequent high success rate of the Carnegie stage assessment. This gives an accurate and precise picture of the course of first trimester development. Confounding by gestational age is minimized by including women with strict pregnancy dating only. The high rate of folic acid supplement use, resulting in an extremely low rate of hyperhomocysteinemia and high RBC folate concentrations, strongly underlines the importance of our results, since even clinically normal values of tHcy and a non-deranged I-C metabolism could impact embryonic development of non-malformed ongoing pregnancies. The most relevant limitation of this study is related to the tertiary care setting, resulting in expected high rates of folic acid supplement use, chronic comorbidity and pregnancy complications. This may reduce the external validity of our findings. Although it is reassuring that significant associations were confirmed in IVF/ICSI pregnancies where the conception date is known by definition, the implantation date is not known and systematic errors in pregnancy dating are expected. Therefore, it is also possible that the small differences detected in embryonic development reflect an impact on the timing of implantation.
Inadequacies in dietary B vitamins and lifestyle (i.e. smoke, alcohol and coffee consumption) have led to increased dangerous plasma tHcy concentrations in the last decades (Steegers-Theunissen et al., 2013). Our results suggest that this may negatively impact first trimester embryonic development resulting in the highest effect estimates in line with previous findings (Steegers-Theunissen et al., 2013; Blanco et al., 2016). Since plasma tHcy is an overall stable marker within the same individual and in uncomplicated pregnancies, a random periconceptional tHcy measurement is reflective of an individual's status and therefore could represent a potential useful predictor of embryonic development in a clinical setting (McKinley et al., 2001; López-Alarcón et al., 2015). Conversely, the small estimates detected for vitamin B12 and RBC folate may not address their clinical use as embryonic development predictors. Nevertheless, while reduced CRL measurements have been associated with adverse pregnancy and health outcomes in the offspring (Mook-Kanamori et al., 2010; van Uitert et al., 2013c; Jaddoe et al., 2014), nothing is known about the clinical implications of first trimester developmental delay in ongoing pregnancies.
In conclusion, we have shown significant associations between periconceptional maternal biomarkers of I-C metabolism and Carnegie stages of embryonic development. Further research is needed to investigate associations between Carnegie stages and birth outcomes and to evaluate the validity of our results in the general population.
Authors’ roles
F.P. contributed to data collection, analysis and interpretation, and wrote the first draft and revised all versions of the manuscript; M.R. performed embryonic measurements; A.H.J.K. provided essential materials (V-scope software); S.P.W. analyzed data and contributed to the interpretation of results; I.C. supervised the writing of the manuscript; R.P.M.S.-T. had primary responsibility for final content, initiated the study and research questions and supervised and contributed to all aspects of the study. All authors read and approved the final manuscript.
Acknowledgements
We would thank all of the participating women and the whole Predict study group for the great effort they put into the Rotterdam Periconception cohort from 2009 and onwards.
Funding
Department of Obstetrics and Gynaecology, Erasmus MC, University Medical Centre, Rotterdam, The Netherlands.
Conflict of interest
No conflict of interest has to be declared by any of the authors regarding the material discussed in the manuscript. R.P.M.S.-T. is CSO of the startup company Slimmere Zorg and CEO of eHealth Care Solutions.
References
- pregnancy
- homocysteine
- metabolism
- erythrocytes
- fertilization in vitro
- folic acid
- academic medical centers
- biological markers
- carbon
- comorbidity
- embryo
- mothers
- obstetrics and gynecology department
- plasma
- reproductive physiological process
- sperm injections, intracytoplasmic
- vitamin b12
- embryologic development
- vitamin b12 measurement, serum
- offspring
- total study
- datasets