European Journal of Clinical Nutrition (2014) 68, 215–222 & 2014 Macmillan Publishers Limited All rights reserved 0954-3007/14 www.nature.com/ejcn

ORIGINAL ARTICLE

Oxidative stress markers and micronutrients in maternal and cord blood in relation to neonatal outcome D Weber1, W Stuetz1, W Bernhard2, A Franz2, M Raith2, T Grune1 and N Breusing3 BACKGROUND/OBJECTIVES: Oxidative stress and micronutrient deficiencies have been related to lower birth weight (BW), small for gestational age (SGA) offspring and preterm delivery. SUBJECTS/METHODS: The relation between neonatal outcome (BW, head circumference, SGA, preterm delivery) with markers of oxidative stress and micronutrients in maternal and cord blood was to be examined. Oxidative stress markers (protein carbonyls (PrCarb), 3-nitrotyrosine (3NT), malondialdehyde (MDA)), total protein concentration and lipid-soluble micronutrients (carotenoids, retinol, tocopherols) were measured in 200 newborns (11% preterms, 13% SGA) and 151 mothers. Associations between target parameters in cord plasma and maternal serum with BW, head circumference and risk of being SGA or preterm were explored. RESULTS: Maternal protein concentration, PrCarb, MDA and all lipid-soluble micronutrients were significantly higher compared with newborns, except for 3NT, which was significantly elevated in newborns. Newborn parameters correlated positively with those of mothers. Preterms had lower proteins and retinol but higher PrCarb than terms. Maternal PrCarb and retinol were inversely associated with BW and head circumference. Mothers with PrCarb, MDA and retinol in the highest quintile had a 3.3-fold (0.9; 12.1), 2.1-fold (0.7; 6.4) and 3.3-fold (1.2; 9.4) risk, respectively, for delivering an SGA newborn, whereas the lowest quintile of retinol in cord blood was associated with an increased risk for preterm delivery. CONCLUSIONS: Oxidative stress (elevated PrCarb) was associated with lower BW/head circumference and SGA. Inadequate hemodilution may explain the inverse relation of maternal retinol with BW and head circumference, and the association between highest maternal retinol and risk for SGA. European Journal of Clinical Nutrition (2014) 68, 215–222; doi:10.1038/ejcn.2013.263; published online 11 December 2013 Keywords: newborn; preterm delivery; lipid-soluble micronutrients; oxidative stress; cord plasma; mothers

INTRODUCTION The risks for various acute and chronic health conditions during the course of life are influenced by low birth weight (LBW) and prematurity, both in developing and developed countries.1–4 In developed countries, except for Sweden, preterm deliveries rose during the last two decades from 7.6% to 9.2%, and those below 1500 g BW rose from 0.7% to 1.3%.5,6 Depending on the definition, B10% of all newborns are classified as small for gestational age (SGA),7 which was shown to be associated with the development of metabolic syndrome, type 2 diabetes mellitus and cardiovascular diseases.8 Oxidative stress is defined as an imbalance between oxidants and antioxidants in favor of the oxidants9 and has been linked to adverse pregnancy course and outcome, such as eclampsia, fetal growth retardation resulting in SGA, and preterm delivery.10–13 The human body continuously produces reactive oxygen and nitrogen species (ROS and RNS) during normal cellular metabolism. Compensation of oxidative stress can be achieved by the endogenous antioxidant system, consisting of enzymes such as catalase, superoxide dismutase and glutathione. Furthermore various micronutrients can counteract oxidative stress. These include ascorbic acid, a-tocopherol and carotenoids, preventing that biomolecules get damaged by ROS and RNS. These damaged molecules include products of protein oxidation, lipid

peroxidation and DNA damage. The most prominent biomarkers that reflect oxidative stress are protein carbonyls (PrCarb), that is, plasma proteins containing oxidized residues such as aldehydes and ketones, 3-nitrotyrosine (3NT), a marker for nitrated proteins and malondialdehyde (MDA), a lipid peroxidation product derived from the oxidative breakdown of polyunsaturated fatty acids. Newborns are particularly susceptible to oxidative stress because of high metabolic turnover rates, low antioxidant concentrations and increased production of free radicals due to the increased postnatal oxygen pressure in plasma.11,14 Furthermore, pregnancy is a physiological state that goes along with elevated energy demands, an increased requirement of oxygen and higher rates of cellular proliferation and one-carbon metabolism.15,16 Early gestational age (GA) and LBW are associated with oxidative stress;13,17 oxidative stress may be due to an insufficient maternal antioxidant supply, because the fetus is totally dependent on maternal nutrient delivery to the placenta. In addition, well-known factors for birth outcomes such as maternal age, smoking habits and other social factors18,19 may also be linked to oxidative stress. Oxidative stress is involved in common diseases of the preterm, such as bronchopulmonary dysplasia (BPD), retinopathy of prematurity and necrotizing enterocolitis, among others. Preterm

1 Department of Nutritional Toxicology, Institute of Nutrition, Friedrich-Schiller-University of Jena, Jena, Germany; 2Department of Neonatology, Faculty of Medicine, EberhardKarls-University of Tuebingen, Tuebingen, Germany and 3Department of Applied Nutritional Science/Dietetics, Institute of Nutritional Medicine, University of Hohenheim, Stuttgart, Germany. Correspondence: Dr N Breusing, Department of Applied Nutritional Science/Dietetics (180c), Institute of Nutritional Medicine, University of Hohenheim, Fruwirthstrasse 12, Stuttgart 70593, Germany. E-mail: [email protected] Received 8 August 2013; revised 21 October 2013; accepted 22 October 2013; published online 11 December 2013

Oxidative stress and micronutrients at birth D Weber et al

216 newborns suffer from additional oxidative stress as a result of lung diseases and oxygen therapy.20 Furthermore, the activation of a series of complex cellular mechanisms in inflammation, coagulation, fibrinolysis, cell cycle and signal transduction as well as free iron are results of oxidative stress.21 Previous work has demonstrated the relation of lipid-soluble micronutrients and oxidative stress to adverse pregnancy outcomes, spontaneous abortion, LBW and preterm birth.13,21–23 However, to our knowledge, there have been no recent studies measuring this combination of oxidative stress markers and micronutrients in mother–newborn pairs of a representative cohort with B10% preterms. To investigate the oxidative stress and micronutrients at birth, we examined markers of protein and lipid oxidation (PrCarb, 3NT, MDA) and lipid-soluble micronutrients (retinol, carotenoids and tocopherols) in pair-matched maternal serum and cord plasma. We describe possible associations of these maternal and neonatal parameters with clinically defined and relevant birth outcomes, that is, BW, head circumference and GA as well as SGA and preterm birth.

MATERIALS AND METHODS The present study was part of a larger study aimed to evaluate the supply of term and preterm newborns with choline and methyl group donors using umbilical cord plasma and maternal serum.24,25 The study was approved by the institutional review board (291/2008BO1), written consent obtained and subjects’ data and samples collected in pseudonymized form.

Subjects/population Pregnant women giving birth at the University Women’s Hospital of the University of Tuebingen, Germany, were asked for consent to participate in this study at admittance to the delivery unit. The study population consisted of convenience samples of unselected consecutively admitted healthy mothers and their term and preterm newborns. None of the mothers included in the present study had diabetes (gestational diabetes, type I or type II diabetes). Inclusion criteria were written parental informed consent and availability of umbilical cord blood and maternal routine data. There were no exclusion criteria. Two cord blood samples including data sets were not assessed because they were extreme preterms (o32 weeks GA) and multiple births. Two maternal samples including data sets were not assessed because the plasma sample of the corresponding child was missing.

Data and blood collection The study group consisted of 200 newborns. Umbilical cord blood was collected into a tube containing EDTA, placed at 4 1C as soon as possible, centrifuged at 1000g for 10 min at 4 1C. A total of 151 maternal serum samples were available (151 matched pairs). Maternal serum was obtained from samples taken within 24 h before or after delivery for other clinical purposes. Cord plasma and maternal serum were transferred into separate tubes and temporarily stored at  20 1C until shipment (on dry ice) to the University of Jena where samples were stored (at  80 1C) and measured. Maternal weight and height were measured at entry into the delivery unit by trained personnel, mostly midwives, and recorded in the mother’s chart, from where this information was retrieved by the study personnel. Further maternal routine data (age, parity, smoking status) were obtained in a pseudonymized form from patient files.

Birth outcomes GA at delivery, BW and head circumference were obtained from patient files. GA was determined by best obstetric estimate based on first day of last menstrual period and first trimester ultrasound or assisted conception date. Preterm birth was defined as newborns born alive before 37 weeks of pregnancy were completed. BW was taken immediately after delivery and determined to the nearest 5 g. Newborns were classified as SGA if their BW was below the 10th percentile of BW for GA by gender according to United States national reference table for fetal growth.26 European Journal of Clinical Nutrition (2014) 215 – 222

Analytical measurements Because deliveries occur irrespective of day or night time and immediate sample processing cannot be guaranteed, we carried out experiments to assess the stability of our measured parameters. Serum and plasma samples of the same individuals (n ¼ 4) were left at 4 1C or room temperature over a period of 24 h. There was neither a significant decrease nor increase in any of our measured parameters during storage at either temperature over the given period. There was, however, a difference between serum and plasma samples of each individual, which can be attributed in part to fibrinolysis factors, and thus to the more or less diluted fractions (including/excluding fibrinogen). Another cause that might account for higher levels of oxidative stress markers in serum samples is the fact that serum vacutainers are routinely left at room temperature to clot for a longer period. Furthermore, EDTA in plasma complexes metal ions such as iron and thus inhibits lipid peroxidation. In detail, serum concentrations were higher for PrCarb ( þ 9.6%), 3NT ( þ 34.6%), MDA ( þ 62.9%) and a-tocopherol ( þ 10.3%). Lycopene and b-cryptoxanthin were also higher (18.5% and 18.1%, respectively), but these parameters were not detectable in cord plasma thus there was no need to adjust these concentrations. Serum concentrations of 3NT, MDA and a-tocopherol differed by more than 10% in comparison with plasma and were therefore adjusted to plasma concentrations for better comparison of values. The calculations were as follows: serum 3NT  0.743 ¼ plasma 3NT; serum MDA  0.614 ¼ plasma MDA; serum a-tocopherol  0.904 ¼ plasma a-tocopherol. Differences of less than 10% were assumed to be due to analytical variations and thus not of relevance for the here examined parameters.

ELISA for the measurement of protein carbonyls and protein-bound 3NT Protein carbonyls were measured as a parameter of oxidative stress after derivatization with 2,4-dinitrophenylhydrazin (DNPH) according to Buss et al.27 with modifications carried out by Sitte et al.28 Protein-bound 3NT was determined as a marker of nitrosative stress as described previously by Weber et al.29 The following devices were used for both ELISA: Nunc Immuno 96 Microwell plate MaxiSorp (Sarstedt, Nuembrecht, Germany), Microplate Reader BioTek Synergy 2 (BioTek Instruments, Friedrichshall, Germany).

HPLC measurements of lipid-soluble micronutrients and MDA Retinol, carotenoids (lutein, zeaxanthin, b-cryptoxanthin, lycopene, a-carotene and b-carotene) and tocopherols (a-tocopherol, g-tocopherol) were measured in maternal serum and cord plasma by reversed-phase HPLC with UV and fluorescence detection according to a method described by Stuetz et al.30 with modifications. In brief, 40 ml of plasma or serum were extracted with 200 ml of an ethanol/n-butanol mixture (1:1) containing b-apo-80 -carotenal-methyloxime as the internal standard. After mixing vigorously, samples were centrifuged and concentrations in the supernatant were analyzed using column and mobile phase as previously described. Retinol (325 nm), carotenoids and internal standard (450 nm) were measured by UV–vis, whereas tocopherols were analyzed by fluorescence detection (excitation/emission set at 298 nm/328 nm). Analyzed pooled samples revealed inter-batch CV for carotenoids between 3.9% (b-carotene) and 7.2% (lycopene), 3.7% for retinol and for tocopherols less than 10%. MDA was determined as a marker of lipid peroxidation after derivatization with thiobarbituric acid (TBA) as describe by Wong et al.31 A volume of 50 ml of serum or plasma was added to 750 ml phosphoric acid (0.44 M), 250 ml TBA solutions (6 g/l) and kept in a water bath at 95 1C for 1 h. After cooling samples in a water-ice bath, they were neutralized with an equal volume of methanolic sodium hydroxide (0.1 M) and centrifuged. Clear supernatants were analyzed for TBA-reactive species (TBARS) after separation on a Reprosil-Pur 120 C18 AQ analytical column (5 mm, 250  4.6 mm, Trentec, Gerlingen, Germany) with fluorescence detection (excitation/emission set at 525/550 nm). The standard Malondialdehydbis(diethylacetal) (Merck KGaA, Darmstadt, Germany) was used for calibration, diluted in ethanol to physiological blood concentrations (0–2.4 mM) and treated as sample. All reagents were of analytical or HPLC grade and purchased from Carl Roth (Karlsruhe, Germany) and Sigma-Aldrich (Steinheim, Germany). All HPLC devices were from Shimadzu and consisted of a pump, autosampler, column heater and system controller (all from the LC-20A & 2014 Macmillan Publishers Limited

Oxidative stress and micronutrients at birth D Weber et al

217 series), UV detector (SPD 20AV), a fluorescence detector (RF 10A XL) and the chromatography workstation (LC Lab Solution).

Statistics Normally distributed continuous variables (maternal age, weight and height, BW, head circumference) were described using means±s.d. Skewed data were cubed power transformed (GA), squared (3NT, MDA, retinol), log-transformed (lutein, a-carotene, b-carotene) or reciprocal transformed (PrCarb), to achieve normal distribution and are expressed as geometric mean (95% confidence interval, CI). For categorical variables (newborn’s gender, preterm birth, LBW, SGA) frequencies are reported. Pearson correlation was used to describe associations between pair-matched maternal and cord blood parameters. Concentrations of measured parameters in maternal and cord blood were compared by

Table 1.

Maternal characteristics and birth outcomes (n ¼ 200)

Maternal characteristics Maternal age (years) (Min–max) Maternal weight (kg)a (Min–max) Maternal height (cm)b (Min–max) Parity (median, IQR)c 0 (% (n)) 1 (% (n)) X2 (% (n)) Birth outcomes Birth weight (g) (Min–max) Gestational age (weeks, days) (Min–max) Preterm births (% (n)) SGA (% (n)) LBW (% (n)) Head circumference (cm) (Min–max) Newborn’s gender (male, % (n))

32.2±5.0 (19.3–44.6) 65±12 (43–99) 167±7 (150–182) 1 (0; 1) 49 (42) 42 (36) 9 (8) 3340±578 (1640–5180) 39.1±1.5 (32.6–42.1) 10.5 (21) 12.5 (25) 8.5 (17) 34.9±1.5 (30.2–39.0) 53 (105)

Abbreviations: IQR, interquartile range; LBW, low birth weight; MDA, malondialdehyde. an ¼ 70. bn ¼ 69. cn ¼ 86.

Table 2.

paired t-test. Maternal serum 3NT and MDA concentrations were adjusted to plasma concentrations for better comparison between maternal and cord means of 3NT and MDA as described above. Means of maternal and cord blood parameters between preterms and terms were compared using independent samples t-test. To assess associations of maternal and cord parameters with BW and/or head circumference, crude univariate regression was applied. Parameters that were associated with a birth outcome (P-valueo0.1) were simultaneously assessed in a multivariate regression models adjusted for GA, maternal age and newborn’s gender. Parameters that were associated with birth outcomes in univariate linear regression (P-valueo0.1) or showed significant differences between preterms and terms were categorized into quintiles. Highest or lowest quintile of maternal or cord plasma parameters showing associations (P-valueo0.1) with SGA and preterm birth were simultaneously assessed in multivariate logistic regression models adjusted for maternal age and newborn’s gender. All tests were two-sided and P-valueo0.05 was considered statistically significant. All statistical analyses were carried out using the Statistical Package of Social Sciences software (SPSS Inc., Chicago, IL, USA; Version 19.0).

RESULTS Table 1 shows the maternal and infants’ characteristics at delivery. Preterm infants accounted for 10.5%, SGA for 12.5% and LBW for 8.5% of newborns in our group. All means of measured parameters between mothers and their newborns differed significantly (Table 2). Maternal serum concentrations of total proteins, PrCarb and MDA were higher than cord plasma concentrations, even after adjusting for the difference between serum and plasma. In contrast, 3NT was elevated in newborns (Figure 1). Concentrations of lipid-soluble micronutrients (retinol, a-carotene, b-carotene, lutein, a- and g-tocopherol) were also higher in mothers compared with newborns, whereas zeaxanthin, b-cryptoxanthin, lycopene and g-tocopherol were even below the limit of detection (LOD) in newborns plasma. Maternal proteins, 3NT, MDA, lutein, a-carotene, b-carotene and a-tocopherol were positively correlated with matched cord blood parameters (Figure 2). In contrary, PrCarb and retinol between cord plasma and maternal serum showed no correlation. Preterm newborns had 842 g lower BW and 2 cm less in head circumference than term newborns (2586±610 g vs 3 428±506 g,

Oxidative stress markers and vitamins in maternal serum and cord plasmaa

Proteins (mg/ml) Protein carbonyls (pmol/mg) 3-Nitrotyrosine (pmol/mg) 3-Nitrotyrosine corrected to plasma (pmol/mg)c MDA (mM) MDA corrected to plasma (mM)d Retinol (mM) Lutein (mM) Zeaxanthin (mM) b-Cryptoxanthin (mM) Lycopene (mM) a-Carotene (mM) b-Carotene (mM) a-Tocopherol (mM) a-Tocopherol corrected to plasma (mM)e g-Tocopherol (mM)

Maternal serum (n ¼ 151)

Cord plasma (n ¼ 200)

Pearson R

P-value

72.3±10.0b 235 (220; 252)b 2.93 (2.58; 3.31)b 2.17 (1.91; 2.44)b 2.04 (1.95; 2.14)b 1.26 (1.20; 1.31)b 1.15 (1.09; 1.21)b 0.528 (0.494; 0.563)b 0.111±0.056b 0.399±0.259b 1.08±0.429b 0.323 (0.277; 0.376)b 0.755 (0.675; 0.845)b 45.7±10.1b 41.3±9.1b 1.58±0.857b

59.1±8.42 83.9 (79.4; 89.0) 6.64 (6.14; 7.16) 6.64 (6.14; 7.16) 0.86 (0.81; 0.90) 0.86 (0.80; 0.90) 0.90 (0.85; 0.95) 0.021 (0.014; 0.030) oLOD oLOD oLOD 0.004 (0.003; 0.005) 0.013 (0.010; 0.017) 6.53±1.87 6.53±1.87 oLOD

0.473 0.115 0.685 0.685 0.298 0.298 0.099 0.368    0.668 0.832 0.262 0.262 

o0.001 0.161 o0.001 o0.001 o0.001 o0.001 0.228 o0.001    o0.001 o0.001 0.001 0.001 

Abbreviations: MDA, malondialdehyde; LOD, limit of detection. aAll values are mean (s.d.) or geometric mean (95% CI). bDifference between maternal serum and cord plasma (Po0.001). cCorrected/adjusted to plasma concentrations as described in the methods section: serum 3NT  0.743. dCorrected/adjusted to plasma concentrations as described in the methods section: serum MDA  0.614. eCorrected/adjusted to plasma concentrations as described in the methods section: serum a-tocopherol  0.904.

& 2014 Macmillan Publishers Limited

European Journal of Clinical Nutrition (2014) 215 – 222

Oxidative stress and micronutrients at birth D Weber et al

218

Figure 1. PrCarb, 3NT and MDA in cord plasma and maternal serum preterms and terms. Box-plots of PrCarb (a), 3NT (b) and MDA (c) concentrations in cord plasma (n ¼ 200) and maternal serum (n ¼ 151) in preterms and terms. Maternal serum concentrations were corrected for better comparison with plasma concentrations as described in the Materials and Methods section. Cord blood samples are illustrated by the light gray boxes, maternal serum by the dark gray boxes. Horizontal lines represent the mean, with the box representing the 25th and 75th percentiles, the whiskers the 5th and 95th percentiles, outliers are shown by circles, and extremes are shown by dots. Significant differences are marked with asterisks (*Po0.05; ***Po0.001).

Figure 2. Correlation between maternal serum and newborn cord plasma. Exemplary correlation between maternal serum and newborns plasma shown for the protein concentration (a) and 3NT (b); n ¼ 151, Po0.001. Maternal proteins, 3NT, MDA, lutein, a-carotene, b-carotene and a-tocopherol were positively correlated, whereas PrCarb and retinol did not correlate with matched cord blood parameters. Maternal serum concentrations were corrected for better comparison with plasma concentrations as described in the Materials and Methods section.

33±2 cm vs 35±1 cm). Age, height or parity did not differ between mothers of term and preterm newborns. Mothers of preterms had marginal lower a-tocopherol concentrations (Table 3). All other parameters measured in maternal serum did not differ between mothers of preterms and mothers of terms. Total protein concentration in cord plasma was lower in preterm compared with term infants, whereas PrCarb were increased. However, there was no effect of GA on MDA or 3NT. In preterms, retinol concentrations were lower than in their term counterparts (0.7 vs 0.9 mM) (Table 3). Multivariate linear regression analysis on birth outcomes adjusted for covariates (GA, maternal age and newborn’s gender) European Journal of Clinical Nutrition (2014) 215 – 222

revealed that maternal PrCarb and retinol were negatively associated with BW and head circumference (Table 4). None of the parameters in cord blood was associated with newborn anthropometrics. Multivariate logistic regression analysis adjusted for maternal age and newborn’s gender revealed that the highest maternal PrCarb, MDA and retinol concentrations were associated with an increased risk for SGA. Mothers with serum PrCarb, MDA and retinol in the highest quintiles had a 3.3-fold (0.9; 12.1), 2.1-fold (0.7; 6.4) and 3.3-fold (1.2; 9.4) higher risk, respectively, for delivering a SGA newborn. In contrary, lowest quintile of retinol in cord blood was associated with an increased risk for preterm birth (Table 5). & 2014 Macmillan Publishers Limited

Oxidative stress and micronutrients at birth D Weber et al

219 Table 3.

Differences between preterm vs term in mothers and newborns

a

Maternal serum (n ¼ 151)

Proteins (mg/ml) Protein carbonyls (pmol/mg) 3-Nitrotyrosine (pmol/mg) MDA (mM) Retinol (mM) Lutein (mM) Zeaxanthin (mM) b-Cryptoxanthin (mM) Lycopene (mM) a-Carotene (mM) b-Carotene (mM) a-Tocopherol (mM) g-Tocopherol (mM)

Cord plasma (n ¼ 200)

Preterms (n ¼ 16)

Terms (n ¼ 135)

P-value

Preterms (n ¼ 21)

Terms (n ¼ 179)

P-value

70.0±11.3 239 (199; 299) 3.21 (1.98; 4.73) 2.06 (1.64; 2.52) 1.08 (0.91; 1.26) 0.50 (0.41; 0.60) 0.09±0.05 0.42±0.28 1.03±0.36 0.34 (0.25; 0.45) 0.76 (0.54; 1.07) 41.5±9.59 1.37±0.74

72.6±9.8 235 (219; 252) 2.88 (2.52; 3.26) 2.04 (1.95; 2.14) 1.16 (1.10; 1.22) 0.53 (0.50; 0.57) 0.11±0.06 0.40±0.26 1.09±0.44 0.32 (0.27; 0.38) 0.75 (0.67; 0.85) 46.2±10.1 1.60±0.87

0.338 0.862 0.588 0.950 0.420 0.511 0.175 0.685 0.593 0.764 0.976 0.084 0.306

52.0±6.3 99.8 (85.7; 120) 6.28 (5.18; 7.49) 0.80 (0.70; 0.91) 0.70 (0.63; 0.78) 0.01 (0.00; 0.04) oLOD oLOD oLOD 0.00 (0.00; 0.01) 0.02 (0.01; 0.03) 6.85±2.05 oLOD

60.2±8.3 82.9 (78.8; 87.4) 6.87 (6.39; 7.38) 0.87 (0.82; 0.91) 0.90 (0.86; 0.95) 0.03 (0.02; 0.04) oLOD oLOD oLOD 0.00 (0.00; 0.01) 0.01 (0.01; 0.02) 6.64±1.8 oLOD

o0.001 0.035 0.424 0.312 o0.001 0.268 — — — 0.963 0.531 0.616 —

Abbreviations: MDA, malondialdehyde; LOD, limit of detection. aAll values are mean±s.d. or geometric mean (95% confidence interval).

Table 4.

Multivariate regression model on birth weight and head circumferencea Birth weight (n ¼ 150)a

Maternal serum protein carbonyls (pmol/mg) Maternal serum retinol (mM)

Head circumference (n ¼ 149)a

b coefficient

95% CI

P-value

b coefficient

95% CI

P-value

 0.569  295

 1.10;  0.031  490;  100

0.038 0.003

 0.001  0.972

 0.003; 0.000  1.52;  0.425

0.058 0.001

Abbreviation: CI, confidence interval. aModels adjusted for gestational age, maternal age, newborn’s gender.

Table 5.

Associations with risks for SGA and for preterm birtha

Small for gestational age: 13% (20/151) Maternal serum protein carbonyls Maternal serum MDA Maternal serum retinol

b

(X357 pmol/mg) (X2.56 mM)b (X1.45 mM)b

Preterm birth: 11% (21/200) Cord plasma retinol

c

(o0.67 mM)

a

b coefficient

95% CI

P-value

3.31 2.13 3.29

(0.90; 12.1) (0.71; 6.4) (1.15; 9.4)

0.071 0.178 0.026

b coefficient

95% CI

P-value

2.91

(1.09; 7.8)

0.034 b

Abbreviations: MDA, malondialdehyde; SGA, small for gestational age. Multivariate model adjusted for maternal age and newborn’s gender. Highest quintile versus lower quintiles. cLowest quintile versus higher quintiles.

DISCUSSION In the present study on 200 newborns and 151 mothers, we observed that oxidative stress, by means of elevated PrCarb, was associated with lower BW and head circumference and with an elevated risk for SGA. PrCarb were elevated in preterms compared with term newborns, whereas mothers of terms and preterms had similar values. Furthermore, highest maternal PrCarb concentrations were related to SGA. Preterm newborns, in comparison with terms, have only marginal adipose tissue and consequently insufficient stores of lipid-soluble antioxidants to compensate for oxidative stress. Further problems increasing oxidative stress particularly in preterm newborns, are derived from ventilation with increased oxygen concentrations (in contrast to in utero concentrations) and inflammatory processes. High maternal and cord PrCarb were previously associated with SGA and the respiratory severity score,11,32 whereas 3NT was only elevated in preterms that developed BPD.33 We found no difference in MDA and 3NT between preterm and term & 2014 Macmillan Publishers Limited

newborns or between their mothers, which may be attributed to the absence of severe diseases such as BPD. However, we observed that MDA and 3NT concentrations in cord blood correlated with those of maternal blood. This indicates that these markers in maternal blood are predictive for the newborn’s oxidative stress. However, in contrast to Arguelles et al.,34 we did not find a correlation for PrCarb between maternal and cord blood. This group also found that oxidative stress measured as lipid hydroperoxides and PrCarb were higher in newborns than in mothers. This was true only for 3NT in our study. Arguelles et al.34 measured maternal PrCarb at beginning of delivery with a spectrophotometric DNPH assay; in contrast, the samples in our study were taken before or after delivery and measured with a sensitive ELISA, which may explain the differences observed. In the present study, no significant correlations between protein oxidation and nitration in cord blood (r ¼  0.004, P ¼ 0.843) and maternal blood (r ¼  0.107, P ¼ 0.194) were observed. 3NT is accepted as a reliable marker for peroxynitrite production because only peroxynitrite has been shown to yield detectable European Journal of Clinical Nutrition (2014) 215 – 222

Oxidative stress and micronutrients at birth D Weber et al

220 amounts of nitrated plasma proteins.35,36 Nitric oxide, responsible for placental blood flow, leads to increased formation of peroxynitrite in the presence of superoxide. Superoxide can be elevated in the fetus and newborn due to a reduced antioxidant capacity and a higher basal level of peroxynitrite formation in newborns influenced by elevated ROS from hyperoxia, as hypothesized by Banks et al.33 During inflammation, activated granulocytes and macrophages produce peroxynitrite, which may further contribute to elevated 3NT concentrations. Posttranslational protein modification is considered a pathological as well as a selective reversible physiological process.37 Nitration can influence inhibition of platelet activation,38 T-cell proliferation and activation,39 vasodilatation40 and muscle relaxation,41 and thus might represent a supportive signal rather than being detrimental in newborns. Furthermore, many reactive species, among them peroxynitrite,40 take part in redox signaling. It was reported that phosphorylation signals may even be mimicked by nitration.37 To our knowledge, this is the first study assessing 3NT simultaneously in mother–newborn pairs. Our findings on the higher level of nitration in newborns should be confirmed in future studies. High concentrations of lipid peroxidation products such as MDA, in maternal as well as in cord blood, were shown to be related to SGA, reduced BW and prematurity.11,23,42 Maternal plasma MDA levels were reported to be twice as high as in cord plasma.43 This is in accordance to our study where maternal MDA concentrations were B1.5-fold (±0.7) higher than cord plasma concentrations (after adjusting for serum-plasma differences). Maternal oxidative stress is known to consume antioxidants, which in turn leads to reduced antioxidant concentrations in the fetus. It is hypothesized that oxidative and nitrative stress, occurring in the placenta due to ischemia/reperfusion and other reasons, covalently modify protein structures in the placenta with consequent unfavorable changes in placental function.23,37 Oxidative stress may suppress blood flow, nutrient supply and hence fetal growth.22,37 However, no mechanistic link has yet been established between oxidative stress (ROS, RNS) and the reduction of BW. In the present study, the maternal serum protein concentrations were higher than those in cord plasma. Furthermore, the protein concentration was lower in the plasma of preterms than in the plasma of terms. The protein concentration in the fetus increases toward the end of gestation.44 This increase in protein concentration is accompanied by water retention to protect the fetus from dehydration in the first period, until the mother is able to provide breast milk.45 Albumin, the major protein in plasma, is known as an important extracellular antioxidant and was shown to be a good target for oxidation in newborn plasma.46 Because of lower albumin, the antioxidant capacity may be reduced, which can lead to higher PrCarb observed in preterms. The cord plasma protein concentration was lower in newborns that developed respiratory distress syndrome and correlated with the severity of lung disease.32,47 Nevertheless, the development of respiratory distress syndrome was most likely due to immaturity and other underlying diseases (inflammation leading to edema), and not necessarily due to low protein concentrations. In our study, cord blood levels of lipid-soluble micronutrients represented 86% (retinol), 15% (a-tocopherol), 13% (lutein), 2.6% (b-carotene) and 2.7% (a-carotene) of maternal levels after adjusting for the serum-plasma difference. Lower circulating levels of carotenoids, retinol and tocopherols in cord blood versus maternal blood have previously been described.11,13,22,48 This can be attributed to a decreased transport capacity for lipophilic substances, including carotenoids and tocopherols in serum/ plasma due to low levels of circulating lipoproteins in fetal blood.49,50 In contrast to other studies,13,51 we found no positive association between a-tocopherol in maternal or cord plasma with European Journal of Clinical Nutrition (2014) 215 – 222

birth outcomes, although mothers of preterms had lower atocopherol concentrations than mothers of terms (P ¼ 0.084). Mothers in the lowest a-tocopherol quintile had higher PrCarb (Po0.05), MDA (Po0.001), and marginal higher cord MDA (P ¼ 0.074), which indicates a role of a-tocopherol in the role of protecting proteins and lipids from oxidation. It was previously reported that an oxidation product of a-tocopherol, a-tocopherol quinone, was significantly higher in plasma and red blood cells of term newborns than in their mothers.52 We found that a cord blood retinol concentration in the lowest quintile (o0.67 mM), close to the conventional cut-off value of 0.7 mM for retinol deficiency, was associated with a 2.9-fold higher risk for preterm birth after adjusting for maternal age and newborn’s gender. On the contrary, we found that mothers with the highest retinol concentrations (X1.45 mM) had an elevated risk for having an SGA newborn. The inverse relation of maternal retinol with BW and head circumference, as well as the association between highest maternal retinol and an increased risk for SGA can be explained by inadequate hemodilution or defective transport. High maternal retinol and hemoglobin status in late pregnancy was associated with a lower BW, head circumference and placental size.51,53 Furthermore, maternal retinol and hemoglobin were independently predictive of BW and placental weight.53 The relatively high maternal retinol concentrations in our study may be caused by a failure to increase plasma volume, which has been related to poor fetal growth.54 Physiologically, there is an increase in plasma volume of B15% during the course of pregnancy, which results in a hemodilution with a relative decrease in hemoglobin;55 during the third trimester, mothers with elevated retinol concentrations are most likely those with inadequate plasma volume expansion.53 Mothers with the highest retinol concentrations gave birth to children with the same GA (39 weeks), but with 321 g lower BWs (3415 vs 3094 g) and 1.34 cm smaller head circumference (35.45 vs 34.11 cm) compared with those mothers with the lowest retinol concentrations; these women also had a 3.9-fold higher risk for having an SGA newborn (0.930–16.16, P ¼ 0.061). We supposed that an inadequate hemodilution, marked by highest maternal retinol concentrations would lead to a retinol deficiency of the newborn. However, there seemed to be a non-linear effect between maternal and cord retinol; mothers with highest retinol did not give birth to newborns with lowest retinol and there was no correlation between maternal and cord retinol. Other authors found only a weak correlation or no correlation at all.49,51,56 Low maternal retinol in early pregnancy represents a risk for retinol deficiency of the fetus. In contrary, highest retinol may represent a marker for inadequate hemodilution, which is known to adversely affect birth outcomes.51,53,54 Maternal retinol did not correlate with total proteins, indicating that the degree of inadequate hemodilution may not have been severe enough to detect significant increase in plasma proteins. However, the effect on BW and head circumference still persisted after adjustment for protein concentration, which supports the hypothesis of retinol as a marker for inadequate hemodilution. In summary, a high maternal retinol was associated with an increased risk for SGA; in contrary, a low cord blood retinol was associated with an increased risk for preterm birth. In maternal blood, there seems to be a non-linear relationship as a result of inadequate hemodilution. The association between cord blood retinol and preterm birth can be explained as a linear relationship between increasing retinol concentration and GA. Besides retinol, which was lower in preterm newborns, we found no differences in any of the measured lipid-soluble micronutrients in cord plasma between term and preterm newborns. Low plasma retinol in preterm infants is a possible risk factor for BPD.57 Low retinol is likely to represent a consequence of the short gestation period. Fetal retinol liver & 2014 Macmillan Publishers Limited

Oxidative stress and micronutrients at birth D Weber et al

221 stores (given in microgram per gram of liver) are doubled between 25th and 37th week of gestation.58 Plasma retinol is transported by retinol-binding protein (RBP) and homeostatically controlled. In the fetus, RBP is released from the liver only at a sufficiently high retinol liver concentration. The present study had several limitations that deserve acknowledgement. There was no sufficient documentation of maternal smoking status, nutrition (including the use of micronutrient supplements) and hemoglobin measurement. Because smoking status was not assessed, the true effect of smoking as a confounder in our study cannot be assessed.59 Future studies should ideally carry out cotinine measurement as a gold standard, as studies have shown that smoking is underestimated by 25% in pregnant women who do not admit to smoke.60 The maternal nutritional status is linked to neonatal health and can easily be modified. Thus, the maternal nutritional intake represents an important determinant of birth outcomes.61 Further studies assessing maternal or newborn blood micronutrients and oxidative stress markers should include the information on maternal food habits and/or intake of micronutrients and antioxidants during the course of pregnancy; this would allow evaluating the effect of micronutrient-rich food intake on oxidative stress status and its implication on birth outcomes. Finally, the determination of the hematocrit or hemoglobin together with the micronutrient measurements would help to identify women with an inadequate hemodilution being most likely responsible for the association between highest maternal retinol and adverse pregnancy outcomes as found in the present study. In line with this, further studies are needed to understand the effects of oxidative stress in relation to micronutrients such as carotenoids, retinol and tocopherols in the mother as well as in newborns on birth outcomes, which should include data on maternal nutrition and smoking status, as well as parity and medical or obstetrical complications, route of delivery and further perinatal factors known to influence the antioxidant status in cord blood. CONCLUSION This is the first study, to our knowledge, to investigate the association of three different markers of oxidative stress, and lipidsoluble micronutrients in cord plasma and maternal serum with SGA and prematurity. These adverse pregnancy outcomes represent important risks for various acute and chronic health conditions during the course of life. Our data also demonstrate that not all antioxidants pass the placental barrier. Although the placental transfer of oxidative stress markers to fetus or mother, respectively, cannot be ruled out by our measurements, maternal parameters can be used to assess the fetal situation. Our work corroborates previous studies showing that maternal oxidative stress and retinol levels are associated with newborn anthropometrics. Oxidative stress may more likely represent a direct effect of inflammation, whereas high maternal retinol seems to reflect inadequate hemodilution resulting in an insufficient fetal-placental blood flow and thus an indirect marker for birth outcomes. High plasma retinol concentration seems to reflect inadequate plasma volume expansion rather than a high vitamin A status adversely effecting birth outcomes. CONFLICT OF INTEREST The authors declare no conflict of interest.

ACKNOWLEDGEMENTS This study was supported by an institutional grant (project no. E1100008) from the joint Center for Nutritional Medicine (Zentrum fuer Ernaehrungsmedizin, ZEM) of the University of Hohenheim and University of Tuebingen, Germany. The project with

& 2014 Macmillan Publishers Limited

the working title ‘Deficiency of choline and methylene donors in preterm newborns’ (N. Breusing/W. Bernhard) was approved for the 2009/2010 research period under the subject ‘Malnutrition/Clinical nutrition’.

REFERENCES 1 Crump C, Sundquist K, Sundquist J, Winkleby MA. Gestational age at birth and mortality in young adulthood. JAMA 2011; 306: 1233–1240. 2 Crump C, Winkleby MA, Sundquist K, Sundquist J. Risk of diabetes among young adults born preterm in Sweden. Diabetes Care 2011; 34: 1109–1113. 3 Crump C, Winkleby MA, Sundquist K, Sundquist J. Risk of hypertension among young adults who were born preterm: a Swedish national study of 636 000 births. Am J Epidemiol 2011; 173: 797–803. 4 Boyle EM, Poulsen G, Field DJ, Kurinczuk JJ, Wolke D, Alfirevic Z et al. Effects of gestational age at birth on health outcomes at 3 and 5 years of age: population based cohort study. BMJ 2012; 344: e896. 5 Blencowe H, Cousens S, Oestergaard MZ, Chou D, Moller AB, Narwal R et al. National, regional, and worldwide estimates of preterm birth rates in the year 2010 with time trends since 1990 for selected countries: a systematic analysis and implications. Lancet 2012; 379: 2162–2172. 6 Zimmer KP. Neonatology departments under economic pressure. Dtsch Arztebl Int 2012; 109: 517–518. 7 Saenger P, Czernichow P, Hughes I, Reiter EO. Small for gestational age: short stature and beyond. Endocr Rev 2007; 28: 219–251. 8 Barker DJ. The fetal and infant origins of adult disease. BMJ 1990; 301: 1111. 9 Sies H. Oxidative stress: introductory remarks. In: Sies H (ed) Oxidative Stress. Academic Press: London, UK, 1985, pp 1–8. 10 Wang Y, Walsh SW. Increased superoxide generation is associated with decreased superoxide dismutase activity and mRNA expression in placental trophoblast cells in pre-eclampsia. Placenta 2001; 22: 206–212. 11 Saker M, Soulimane MN, Merzouk SA, Merzouk H, Belarbi B, Narce M. Oxidant and antioxidant status in mothers and their newborns according to birthweight. Eur J Obstet Gynecol Reprod Biol 2008; 141: 95–99. 12 Scholl TO, Chen X, Sims M, Stein TP. Vitamin E: maternal concentrations are associated with fetal growth. Am J Clin Nutr 2006; 84: 1442–1448. 13 Baydas G, Karatas F, Gursu MF, Bozkurt HA, Ilhan N, Yasar A et al. Antioxidant vitamin levels in term and preterm infants and their relation to maternal vitamin status. Arch Med Res 2002; 33: 276–280. 14 Wisdom SJ, Wilson R, McKillop JH, Walker JJ. Antioxidant systems in normal pregnancy and in pregnancy-induced hypertension. Am J Obstet Gynecol 1991; 165(6 Pt 1): 1701–1704. 15 Myatt L, Cui X. Oxidative stress in the placenta. Histochem Cell Biol 2004; 122: 369–382. 16 Jauniaux E, Watson AL, Hempstock J, Bao YP, Skepper JN, Burton GJ. Onset of maternal arterial blood flow and placental oxidative stress. A possible factor in human early pregnancy failure. Am J Pathol 2000; 157: 2111–2122. 17 Davis JM, Auten RL. Maturation of the antioxidant system and the effects on preterm birth. Semin Fetal Neonatal Med 2010; 15: 191–195. 18 Brooke OG, Anderson HR, Bland JM, Peacock JL, Stewart CM. Effects on birth weight of smoking, alcohol, caffeine, socioeconomic factors, and psychosocial stress. BMJ 1989; 298: 795–801. 19 Haste FM, Brooke OG, Anderson HR, Bland JM. The effect of nutritional intake on outcome of pregnancy in smokers and non-smokers. Br J Nutr 1991; 65: 347–354. 20 Saugstad OD. Bronchopulmonary dysplasia-oxidative stress and antioxidants. Semin Neonatol 2003; 8: 39–49. 21 Buonocore G, Perrone S, Longini M, Terzuoli L, Bracci R. Total hydroperoxide and advanced oxidation protein products in preterm hypoxic babies. Pediatr Res 2000; 47: 221–224. 22 Jain SK, Wise R, Yanamandra K, Dhanireddy R, Bocchini Jr. JA. The effect of maternal and cord-blood vitamin C, vitamin E and lipid peroxide levels on newborn birth weight. Mol Cell Biochem 2008; 309: 217–221. 23 Kim YJ, Hong YC, Lee KH, Park HJ, Park EA, Moon HS et al. Oxidative stress in pregnant women and birth weight reduction. Reprod Toxicol 2005; 19: 487–492. 24 Bernhard W, Full A, Arand J, Maas C, Poets CF, Franz AR. Choline supply of preterm infants: assessment of dietary intake and pathophysiological considerations. Eur J Nutr 2013; 52: 1269–1278. 25 Weber D, Stuetz W, Bernhard W, Franz AR, Raith M, Grune T et al. 5-Methyltetrahydrofolate and thiamine diphosphate in cord blood erythrocytes of preterm versus term newborns. Eur J Clin Nutr 2013; 67: 1029–1035. 26 Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996; 87: 163–168. 27 Buss H, Chan TP, Sluis KB, Domigan NM, Winterbourn CC. Protein carbonyl measurement by a sensitive ELISA method. Free Radic Biol Med 1997; 23: 361–366. 28 Sitte N, Merker K, Grune T. Proteasome-dependent degradation of oxidized proteins in MRC-5 fibroblasts. FEBS Lett 1998; 440: 399–402.

European Journal of Clinical Nutrition (2014) 215 – 222

Oxidative stress and micronutrients at birth D Weber et al

222 29 Weber D, Kneschke N, Grimm S, Bergheim I, Breusing N, Grune T. Rapid and sensitive determination of protein-nitrotyrosine by ELISA: application to human plasma. Free Radic Res 2012; 46: 276–285. 30 Stuetz W, McGready R, Cho T, Prapamontol T, Biesalski HK, Stepniewska K et al. Relation of DDT residues to plasma retinol, alpha-tocopherol, and beta-carotene during pregnancy and malaria infection: a case-control study in Karen women in northern Thailand. Sci Total Environ 2006; 363: 78–86. 31 Wong SH, Knight JA, Hopfer SM, Zaharia O, Leach Jr CN, Sunderman Jr. FW. Lipoperoxides in plasma as measured by liquid-chromatographic separation of malondialdehyde-thiobarbituric acid adduct. Clin Chem 1987; 33(2 Pt 1): 214–220. 32 Ballard PL, Truog WE, Merrill JD, Gow A, Posencheg M, Golombek SG et al. Plasma biomarkers of oxidative stress: relationship to lung disease and inhaled nitric oxide therapy in premature infants. Pediatrics 2008; 121: 555–561. 33 Banks BA, Ischiropoulos H, McClelland M, Ballard PL, Ballard RA. Plasma 3-nitrotyrosine is elevated in premature infants who develop bronchopulmonary dysplasia. Pediatrics 1998; 101: 870–874. 34 Arguelles S, Machado MJ, Ayala A, Machado A, Hervias B. Correlation between circulating biomarkers of oxidative stress of maternal and umbilical cord blood at birth. Free Radic Res 2006; 40: 565–570. 35 Gow A, Duran D, Thom SR, Ischiropoulos H. Carbon dioxide enhancement of peroxynitrite-mediated protein tyrosine nitration. Arch Biochem Biophys 1996; 333: 42–48. 36 Beckman JS, Ye YZ, Anderson PG, Chen J, Accavitti MA, Tarpey MM et al. Extensive nitration of protein tyrosines in human atherosclerosis detected by immunohistochemistry. Biol Chem Hoppe Seyler 1994; 375: 81–88. 37 Myatt L. Review: Reactive oxygen and nitrogen species and functional adaptation of the placenta. Placenta 2010; 31(Suppl): S66–S69. 38 Low SY, Sabetkar M, Bruckdorfer KR, Naseem KM. The role of protein nitration in the inhibition of platelet activation by peroxynitrite. FEBS Lett 2002; 511: 59–64. 39 Brito C, Naviliat M, Tiscornia AC, Vuillier F, Gualco G, Dighiero G et al. Peroxynitrite inhibits T lymphocyte activation and proliferation by promoting impairment of tyrosine phosphorylation and peroxynitrite-driven apoptotic death. J Immunol 1999; 162: 3356–3366. 40 Ferdinandy P. Peroxynitrite: just an oxidative/nitrosative stressor or a physiological regulator as well? Br J Pharmacol 2006; 148: 1–3. 41 Viner RI, Williams TD, Schoneich C. Peroxynitrite modification of protein thiols: oxidation, nitrosylation, and S-glutathiolation of functionally important cysteine residue(s) in the sarcoplasmic reticulum Ca-ATPase. Biochemistry 1999; 38: 12408–12415. 42 Joshi SR, Mehendale SS, Dangat KD, Kilari AS, Yadav HR, Taralekar VS. High maternal plasma antioxidant concentrations associated with preterm delivery. Ann Nutr Metab 2008; 53: 276–282. 43 Rogers MS, Mongelli M, Tsang KH, Wang CC. Fetal and maternal levels of lipid peroxides in term pregnancies. Acta Obstet Gynecol Scand 1999; 78: 120–124. 44 Linderkamp O, Stadler AA, Zilow EP. Blood viscosity and optimal hematocrit in preterm and full-term neonates in 50- to 500-micrometer tubes. Pediatr Res 1992; 32: 97–102.

European Journal of Clinical Nutrition (2014) 215 – 222

45 Cooper WO, Atherton HD, Kahana M, Kotagal UR. Increased incidence of severe breastfeeding malnutrition and hypernatremia in a metropolitan area. Pediatrics 1995; 96(5 Pt 1): 957–960. 46 Marzocchi B, Perrone S, Paffetti P, Magi B, Bini L, Tani C et al. Nonprotein-bound iron and plasma protein oxidative stress at birth. Pediatr Res 2005; 58: 1295–1299. 47 Bland RD. Cord-blood total protein level as a screening aid for the idiopathic respiratory-distress syndrome. N Engl J Med 1972; 287: 9–13. 48 Navarro J, Causse MB, Desquilbet N, Herve F, Lallemand D. The vitamin status of low birth weight infants and their mothers. J Pediatr Gastroenterol Nutr 1984; 3: 744–748. 49 Kiely M, Cogan PF, Kearney PJ, Morrissey PA. Concentrations of tocopherols and carotenoids in maternal and cord blood plasma. Eur J Clin Nutr 1999; 53: 711–715. 50 Jauniaux E, Cindrova-Davies T, Johns J, Dunster C, Hempstock J, Kelly FJ et al. Distribution and transfer pathways of antioxidant molecules inside the first trimester human gestational sac. J Clin Endocrinol Metab 2004; 89: 1452–1458. 51 Masters ET, Jedrychowski W, Schleicher RL, Tsai WY, Tu YH, Camann D et al. Relation between prenatal lipid-soluble micronutrient status, environmental pollutant exposure, and birth outcomes. Am J Clin Nutr 2007; 86: 1139–1145. 52 Jain SK, Wise R, Bocchini Jr. JJ. Vitamin E and vitamin E-quinone levels in red blood cells and plasma of newborn infants and their mothers. J Am Coll Nutr 1996; 15: 44–48. 53 Mathews F, Youngman L, Neil A. Maternal circulating nutrient concentrations in pregnancy: implications for birth and placental weights of term infants. Am J Clin Nutr 2004; 79: 103–110. 54 Salas SP, Rosso P, Espinoza R, Robert JA, Valdes G, Donoso E. Maternal plasma volume expansion and hormonal changes in women with idiopathic fetal growth retardation. Obstet Gynecol 1993; 81: 1029–1033. 55 Lund CJ, Donovan JC. Blood volume during pregnancy. Significance of plasma and red cell volumes. Am J Obstet Gynecol 1967; 98: 394–403. 56 Uotila J, Tuimala R, Pyykko K, Ahotupa M. Pregnancy-induced hypertension is associated with changes in maternal and umbilical blood antioxidants. Gynecol Obstet Invest 1993; 36: 153–157. 57 Hustead VA, Gutcher GR, Anderson SA, Zachman RD. Relationship of vitamin A (retinol) status to lung disease in the preterm infant. J Pediatr 1984; 105: 610–615. 58 Shah RS, Rajalakshmi R, Bhatt RV, Hazra MN, Patel BC, Swamy NB et al. Liver stores of vitamin A in human fetuses in relation to gestational age, fetal size and maternal nutritional status. Br J Nutr 1987; 58: 181–189. 59 Campbell MK, Cartier S, Xie B, Kouniakis G, Huang W, Han V. Determinants of small for gestational age birth at term. Paediatr Perinat Epidemiol 2012; 26: 525–533. 60 Shipton D, Tappin DM, Vadiveloo T, Crossley JA, Aitken DA, Chalmers J. Reliability of self reported smoking status by pregnant women for estimating smoking prevalence: a retrospective, cross sectional study. BMJ 2009; 339: b4347. 61 Hininger I, Favier M, Arnaud J, Faure H, Thoulon JM, Hariveau E et al. Effects of a combined micronutrient supplementation on maternal biological status and newborn anthropometrics measurements: a randomized double-blind, placebocontrolled trial in apparently healthy pregnant women. Eur J Clin Nutr 2004; 58: 52–59.

& 2014 Macmillan Publishers Limited