Dr N Potdar, Reproductive Sciences Section, Department of Cancer Studies and Molecular Medicine, Robert Kilpatrick Clinical Sciences Building, University of Leicester, Leicester LE2 7LX, UK. Email email@example.com
Objective Investigation of increased oxidative stress in early pregnancy and association with an increased risk of small-for-gestational-age (SGA) fetus.
Design Longitudinal case–control study.
Setting University Hospitals of Leicester NHS Trust, Leicester, UK.
Population Low-risk pregnant women with no current or pre-existing medical illness were recruited at a large teaching hospital from 2004 to 2006.
Methods Recruitment performed at the time of the dating ultrasound scan (12 ± 2 weeks of gestation). Spot urine samples collected at 12 ± 2 and 28 ± 2 weeks of gestation were analysed for 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG) by liquid chromatography with tandem mass spectrometry). SGA was defined as birthweight <10th centile based on customised centile calculator (www.gestation.net). This identified the cases (n= 55), whereas controls (n= 55) were mothers whose babies were appropriate for gestational age (AGA, birthweight 10th–90th centile). Statistical analysis was performed using GraphPad Prism v.5. The relationship between maternal urinary 8-oxodG at different gestations and customised SGA was investigated by nonparametric tests.
Main outcome measures Customised SGA and AGA pregnancies.
Results Urinary 8-oxodG concentrations were significantly increased in pregnancies with subsequent SGA compared with concentrations in normal pregnancies; 12 weeks: 2.8 (interquartile range [IQR] 1.96–3.67) versus 2.2 (IQR 1.26–3.28) pmol 8-oxodG/μmol creatinine (P= 0.0007); 28 weeks: 2.21 (IQR 1.67–3.14) versus 1.68 (IQR 1.16–2.82) pmol 8-oxodG/μmol creatinine (P < 0.0002). Concentrations decreased significantly between week 12 and 28 (P= 0.04 and P= 0.02 for controls and cases).
Conclusions In this study, urinary 8-oxodG at 12 and 28 weeks were elevated in SGA compared with AGA pregnancies. This may reflect early placental changes predating clinical features of SGA.
Small-for-gestational-age (SGA) babies, who are not diagnosed before birth, have a significantly increased risk of perinatal mortality. Specifically, it has been shown that up to 50% of unexplained stillbirths are growth restricted.1–3 However, SGA is a heterogeneous population, which includes not only fetal growth restricted (FGR) babies but also small healthy babies. The use of customised centile calculator is a means by which the SGA group may be refined to minimise the inclusion of false positives (small healthy babies) and to maximise the identification of FGR.4 Therefore, while not synonymous with FGR, SGA, as defined by the customised centile calculator (customised SGA), represents a group with a high proportion of FGR. In the general population, FGR complicates ∼10% of pregnancies and, in addition to increased perinatal morbidity/mortality, is associated with an increased risk of developing the metabolic syndrome (dyslipidemia, insulin resistance and hypertension), type II diabetes and cardiovascular diseases in adulthood.5–7 However, despite the severe consequences of FGR, predictive tests are lacking.
The aetiopathogenesis of FGR is poorly understood, although various factors such as maternal systemic disease, smoking and recreational drugs have been implicated.8 Various markers of oxidative stress have been studied in FGR and other pregnancy complications such as diabetes and pre-eclampsia, suggesting a role of reactive oxygen species (ROS) in their aetiopathogenesis. Oxidative stress is defined as an imbalance between prooxidant–antioxidant system, in favour of the former.9 This imbalance can be the result of either (i) increased production of oxidants, such as superoxide from mitochondria and other metabolic processes or (ii) decreased antioxidant capacity, which includes enzymes, such as glutathione peroxidase, and dietary nutrients, such as vitamins C and E or (iii) combination of both. Under conditions of oxidative stress, oxidatively generated modification of cellular biomolecules increases over and above the basal level of damage. All molecules within the cell are potential targets for ROS, including DNA and unsaturated lipids. Oxidatively generated damage is followed by multiple repair pathways; the damage and repair to DNA can lead to microsatellite instability, inhibit methylation, accelerate telomere shortening, induce mutations, alter gene expression and eventually cause cytostasis.10 Oxidatively modified DNA components, such as 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxodG), have been widely used as a biomarker of oxidative stress.11,12 Furthermore, some of these products can be examined in extracellular matrices, such as urine, offering a valuable means by which oxidative stress may be assessed noninvasively, circumventing issues of DNA extraction and artefact.
It is well established that pregnancy itself is a state of oxidative stress13 secondary to increased metabolic activity in the placental mitochondria. There is a first-trimester physiological increase in the production of ROS primarily due to the hypoxia–reoxygenation injury following the establishment of intervillous space blood flow.14,15 Further rise in ROS production leads to increased oxidative stress, with attendant pathological effects at a cellular and organ level. The exact basis for this ‘pathological’ oxidative stress in complicated pregnancies, such as FGR, over and above that which is physiological is not known. Chronic hypoxia is associated with increased production of ROS and free radical species and may contribute to the pathogenesis of this condition.
We report a longitudinal case–control study, investigating the possibility of using urinary 8-oxodG, as a noninvasive marker of oxidative stress, in late first and second trimesters of pregnancy. Our purpose was to investigate whether increased oxidative stress in early pregnancy, as measured by using urinary 8-oxodG, is associated with customised SGA.
Study design and recruitment
This was a case–control cohort study of pregnant women recruited to a now completed prospective study of caffeine in pregnancy at the University Hospitals of Leicester NHS Trust and the University of Leicester, UK.16 Only low-risk women, defined as having no current or pre-existing medical illness, were recruited into the study. Subjects with recreational drug or alcohol abuse problems and psychiatric illness were also excluded. Recruitment was performed at the time of the dating ultrasound scan in the late first trimester (12 ± 2 weeks of gestation). Each subject gave a signed informed consent before entering the study, which also had been approved by the local ethics committee.
A questionnaire that included demographic details and smoking status was completed by the subjects. Spot urine samples were collected at recruitment and again in the late second trimester (28 ± 2 weeks of gestation). Cases (n= 55; age 18–40 years) included women giving birth to a SGA baby, defined as birthweight <10th centile on the customised centile calculator of the Perinatal Institute, Birmingham, UK; www.gestation.net.17 The centile calculator takes into account maternal body mass index (BMI), ethnicity, parity, gestational age at delivery, fetal gender and birthweight. The control group (n= 55; age 18–40 years) was derived from women who delivered appropriate for gestational age (AGA) babies (birthweight 10th–90th centile). Selection was on the basis of their subject identification number being adjacent to that of the corresponding case in the recruitment database. Pregnancies complicated by delivery at <37 completed weeks of gestation and pre-eclampsia (blood pressure of ≥140/90 mmHg, with proteinuria) were excluded from the study.
Urine samples (∼15–20 ml) collected at booking and at 28 weeks were stored, without preservatives, at −80°C prior to analysis (these have been shown to be stable, once frozen, for more than 10 years).18 Urinary creatinine (Department of Chemical Pathology, University Hospitals of Leicester NHS Trust) was used to normalise for variations in urine concentration, and 8-oxodG measurements adjusted as pmol 8-oxodG/μmol creatinine.19
Analysis of urinary 8-oxodG
Urinary 8-oxodG was quantified as previously reported by Cooke et al.19 Briefly, to each of the urine samples was added 24 pmol [15N5]-labelled 8-oxodG prior to processing to account for possible losses during sample work up and analysis, together with ion suppression during mass spectrometric analysis. Each urine sample (1600 μl) was acidified with 160 μl of 10% (v/v) formic acid and incubated at 4°C for 1 hour, prior to centrifugation at 10 000 ×g for 10 minutes at 4°C. The supernatants were collected and then underwent solid phase extraction using Waters Oasis® (Waters Ltd, Elstree, UK) HLB Vac cartridges (60 mg packing material) to increase the accuracy and sensitivity.19–21 We have previously shown recovery of 8-oxodG from these columns to be 81%.19 The samples were then analysed by mass spectrometry.
The liquid chromatography with tandem mass spectrometry (LC-MS/MS) system consisted of a Waters Alliance 2695 (Waters Ltd, Manchester, UK) separations module connected to a Micromass Quattro Ultima Platinum (Waters-Micromass Ltd, Manchester, UK) tandem quadrupole mass spectrometer with an electrospray interface. Selected reaction monitoring (SRM) analysis was performed for the [M+H]+ ion to oxidised base [B+H2]+ transitions of 8-oxodG (mass-to-charge ratio 284–168) and the stable isotope internal standard [15N5]8-oxodG (mass-to-charge ratio 289–173). The level of 8-oxodG in each urine sample was determined from the ratio of the peak area of 8-oxodG to that of the internal standard. The limit of detection for the optimised analysis of urinary 8-oxodG by LC-MS/MS SRM was 5 fmol on the column (signal-to-noise ratio [S/N] = 4) for pure 8-oxodG standard. Each sample was assayed in duplicate. The inter- and intra-assay coefficient of variation was less than 10% for 8-oxodG.19
Salivary cotinine analysis
Smoking has a well-known association with both FGR8,22 and oxidative stress.23,24 Salivary cotinine was therefore measured to identify smokers, confirm self-reported smoking status and explore the effect of smoking, as a potential confounder, on 8-oxodG concentrations in our population. Saliva samples were collected, from all subjects, in sterile salivettes (Sarstedt, Loughborough, UK) at recruitment (12 ± 2 weeks of gestation) and stored at −80°C. Cotinine analysis was then performed by the standard enzyme-linked immunosorbent assay (Cozart plc, Abingdon, UK) technique.
Data analysis and graphs were plotted on GraphPad Prism v.5 (GraphPad Software, San Diego, CA, USA). The normality of data distribution was determined by D’Agostino and Pearson omnibus normality test. The relationship between maternal urinary 8-oxodG at different gestations and customised SGA and possible effects of cotinine were investigated by Mann–Whitney test. As the observations reported in this study were not the primary objectives of the main study,16 a power calculation was not performed.
The demographic characteristics of the 55 subjects and 55 controls are shown in Table 1; 44% were nulliparous and 56% were multiparous. Eighty-nine percent were white Europeans, and there were no differences in any of the demographic variables between the cases and the controls (Table 1). The smoking status of the cases and controls, assessed by measuring salivary cotinine, at 12 weeks of pregnancy is shown in Table 1. They were further divided into three subgroups based on the standard reported cotinine cutoffs: nonsmoker <1 ng/ml, passive smoker = 1–5 ng/ml and smoker >5 ng/ml.25 The cotinine data were not normally distributed (determined by D’Agostino and Pearson omnibus normality test); therefore, nonparametric tests were performed. Although the median values of cotinine in the cases appeared to be greater than those in the controls, 0.71 (interquartile range [IQR] 0.11–8.30) ng/ml and 0.37 (IQR 0.14–1.99) ng/ml, these were not statistically significant (Mann–Whitney, P > 0.05).
Table 1. Participant characteristics in the two study groups
Cases (n = 55) (pregnancy with SGA)
Controls (n = 55) (pregnancy with AGA)
Mean maternal age, years (range)
BMI (mean ± SD)
25 ± 4
24 ± 3.5
94% White European (53)
89% White European (49)
64% Nulliparous (35)
38% Nulliparous (21)
Gestation at delivery (days) (median; IQR)
Neonatal gender, M/F
Neonatal birthweight (g) (median; IQR)
Birthweight centile (median; IQR)
Salivary cotinine (ng/ml) at 12 weeks (median; IQR)
The Figure 1 shows the scatter plot with the median values of urinary 8-oxodG in both cases and controls at 12 and 28 weeks of gestation. The concentrations were significantly higher in the cases compared with the controls, at both 12 weeks: 2.8 (IQR 1.96–3.67) versus 2.2 (IQR 1.26–3.28) pmol 8-oxodG/μmol creatinine (P = 0.0007); and 28 weeks: 2.21 (IQR 1.67–3.14) versus 1.68 (IQR 1.16–2.82) pmol 8-oxodG/μmol creatinine (P < 0.0002). The median urinary 8-oxodG concentrations decreased significantly from 12 to 28 weeks (P= 0.04 and P= 0.02 for controls and cases, respectively). The median urinary creatinine concentration did not differ significantly between the cases and the controls at 12 and 28 weeks (12 weeks: controls, median = 7.6 [IQR 1.97–14.38] versus cases 6.4 [IQR 3.1–12.8] mmol/l [P= 0.8]; 28 weeks: controls, median = 7.65 [IQR 4.07–10.10] versus cases 7.00 [IQR 3.55–11.60] mmol/l [P= 0.9]), indicating that the 8-oxodG concentrations were not affected by creatinine concentrations. A further correlation was examined between salivary cotinine and urinary 8-oxodG concentrations, at 12 weeks, which showed no relationship (Spearman r= 0.25, P= 0.84).
A multivariate analysis (using a multiple regression model) was performed to investigate the possible of effect of maternal parity, BMI, gender of the baby and cotinine concentrations on urinary 8-oxodG at both 12 and 28 weeks of gestation. This showed that none of these factors had effect on the trends of urinary 8-oxodG concentrations at both 12 and 28 weeks of gestation (all P > 0.05 and r2= 0.013 and 0.066 for 12 and 28 weeks, respectively).
To the best of our knowledge, this is the first report examining urinary 8-oxodG, an established marker of oxidative stress, in early pregnancy as a potential biomarker of FGR. In a study of 52 women in the early third trimester, Scholl and Stein26 demonstrated that increased urinary 8-oxodG concentrations were associated with low birthweight and preterm delivery. Our study shows that an increase in oxidative stress, over and above that which is physiological, may occur as early as 12 weeks into the pregnancy and is associated with customised SGA pregnancies.
The increase in urinary 8-oxodG concentrations in late first trimester compared with the second trimester in both cases and controls demonstrated in this study can be explained by the physiological changes occurring in the late first trimester. After implantation, the fetoplacental unit exists in a hypoxic environment due to a low intrauterine oxygen tension (pO2 < 20 mmHg).14 At the end of the first trimester, following trophoblastic invasion of the spiral arteries, there is a rise in the oxygen tension in the intervillous space (pO2∼50 mmHg). This causes the hypoxia–reoxygenation (ischaemia–reperfusion) injury, observed in normal placenta, leading to an increased level of oxidative stress in the syncytiotrophoblast.14,27 In fact, although not directly comparable, when our first-trimester data (controls) were compared with those reported for nonpregnant women in the literature,28 the median concentrations of 8-oxodG appeared higher in our pregnant controls (2.2 versus 1.83 pmol/μmol creatinine).
An association between oxidative stress and complicated pregnancies such as FGR and pre-eclampsia is well established,29–34 but its evaluation in different trimesters of pregnancy, and customised SGA, has not been considered prior to our study. Our data show that urinary 8-oxodG concentrations are much higher in the late first trimester and second trimester of pregnancies complicated by SGA compared with controls. Maternal urinary 8-oxodG and malondialdehyde (another marker of ROS) in the samples collected at the end of first stage of labour have also been shown to have an inverse association with birthweight at term.30 In another study, Takagi et al.33 showed increased immunohistochemical staining for 8-oxodG in the placenta of women who had FGR babies compared with placenta from mothers with appropriately grown babies. Furthermore, the concentrations of 8-oxodG were significantly higher in placentas of pregnancies complicated with FGR alone or pre-eclampsia with FGR compared with pre-eclampsia only and normal controls.33 Arguelles et al.29 have shown that maternal biomarkers of oxidative stress correlate closely with neonatal oxidative stress. Maternal oxidative stress therefore appears to be a reliable proxy assessment of fetal oxidative stress.
We chose to evaluate a marker of DNA oxidation, rather than lipid peroxidation, because of the prevailing evidence that oxidative stress affects the placenta differently in the pathogenesis of various complications, such as FGR and pre-eclampsia. Lipid peroxidation occurs primarily in the superficial cells of the syncytiotrophoblasts, hence higher concentrations are observed in women with pre-eclampsia, whereas, increased DNA damage is observed mainly in the rapidly growing cells of the trophoblast columns, affecting fetal growth.33
Smoking has a well-established association with FGR8,22 and has been associated with oxidative stress;23,24 however, we noted no significant correlation between cotinine and urinary 8-oxodG, implying that smoking possibly affects fetal growth by mechanisms other than oxidative damage to DNA components. In contrast, Arguelles et al.29 reported a higher concentration of other biomarkers of oxidative stress (protein carbonyls and lipid peroxidation products) in the serum of mothers who smoke, and their newborns, compared with nonsmoking pregnant women. The differences reported between this and our study may derive from a number of factors, which include assessment of different biomarkers of oxidative stress, investigation of biomarkers in different matrices (urine versus serum), determination of smoking status (self-reported [Arguelles et al.] versus salivary cotinine [present study]).
Our study has demonstrated, using a sensitive, noninvasive method, an increased level of oxidative damage to DNA in the first trimester of pregnancies that become complicated by customised SGA. We have provided some evidence to suggest that maternal oxidative stress status might reflect the oxidative stress within the fetoplacental unit. Our data also confirm physiological oxidative stress in normal pregnancy as well as the presence of increased oxidative stress in association with customised SGA pregnancies. The precise source of this remains to be clarified.
We acknowledge that our study involves a relatively small cohort, with no direct indication of fetal oxidative stress status, making it difficult to speculate on whether increased oxidative stress causes FGR or the fetus prone to growth restriction is susceptible to oxidative stress, and hence the effect on its growth potential.
Currently, there are no accurate tests to predict which pregnancies will be complicated by FGR. In this pilot study, urinary 8-oxodG concentrations (a marker of oxidative stress) were elevated in customised SGA compared with AGA pregnancies at both 12 and 28 weeks of gestation. Further studies are recommended to confirm these observations and investigate the potential of urinary 8-oxodG as a biomarker for prediction of FGR.
Disclosure of interest
Contribution to authorship
N.P.: conception, design of the study, acquisition of data, analysis, drafting of manuscript and final version. R.S.: laboratory analysis and contribution to the manuscript. V.M.: laboratory analysis and contribution to the manuscript. M.D.E.: contribution to conception and design and critical revision of manuscript. P.B.F.: laboratory analysis and critical revision of manuscript. J.C.K.: substantial contribution to design, data collection, supervision, drafting and revising manuscript. M.S.C.: substantial contribution to conception, design, supervision and revising manuscript.
Details of ethics approval
From local ethics committee (LNRREC, reference no. 7260) for the prospective study of caffeine in pregnancy, further amendment obtained for the case–control study of the same cohort.
No external funding was received for the case–control study.
The authors gratefully acknowledge support from the UK Food Standards Agency (contract T01032).