Maternal Physical Activity and Neonatal Cord Blood pH: Findings from the Born in Bradford Pregnancy Cohort

Objective: Evidence suggests that physical activity whilst pregnant is beneficially associated with maternal cardiometabolic health and perinatal outcomes. It is unknown if benefits extend to objective markers of the neonate condition at birth. This study investigated associations of maternal pregnancy physical activity with cord blood pH. Methods: Cord blood pH was measured when clinically indicated in a subgroup of Born in Bradford birth cohort participants ( n = 1,467). Pregnant women were grouped into one of four activity categories (inactive/somewhat active/moderately active/active) based on their self-reported physical activity at 26–28 weeks gestation. Linear regression was used to calculate adjusted mean differences in cord blood pH, and Poisson regression was used to quantify relative risks for moderate cord blood acidaemia (pH < 7.10), across physical activity groups. Results: More than half of pregnant women (52.0%) were inactive, one-fifth were somewhat active (21.7%), fewer were moderately active (14.6%) and active (11.7%), respectively. Pregnancy physical activity was favourably associated with higher cord blood pH. Compared to neonates of inactive women, there was some evidence that neonates of women who were at least somewhat active in pregnancy had lower relative risk of moderate cord blood acidaemia (for arterial blood: relative risk = 0.70 (95% confidence interval 0.48–1.03)). Conclusions: Modest volumes of mid-pregnancy maternal physical activity do not appear to adversely influence cord blood pH and may enhance the neonate condition at birth. hypertension, gestational weight gain, emergency caesarean and medical or surgical induction), measurement factors related to maternal physical activity (gestational age at the time of reporting, if women were feeling well and were able to enjoy their normal daily activities, if maternal physical activity was reported during Ramadan) and cord blood pH (the time delay from birth to blood sampling (the median (IQR) time delay to arterial and venous cord blood samples was 10 (10) and 12 (10) minutes, respectively) did not appreciably influence associations and so were not retained in models. Models were rerun excluding samples that were missing data for sampling time ( n = 61) and excluding samples that were delayed by ≥ 30 minutes following delivery ( n = 82), the results were substantively unchanged.


Introduction
Safety concerns are cited by pregnant women as a barrier to physical activity (Coll et al., 2017). This may have originated from cultural beliefs that women should 'rest' in pregnancy (Coll et al., 2017) alongside former recommendations that physical activity should be limited because of concerns for mother and offspring welfare, including risk of impaired oxygen supply to the fetus during supine exercise (American College of Obstetricians and Gynecologists (ACOG), 1985; Pivarnik and Mudd, 2009).
It is still good practice to avoid supine exercise, particularly late in pregnancy, but otherwise physical activity can enhance transportation of oxygenated blood to the developing fetus, via physiological adaptations such as increased placental surface area, and enhanced blood flow and perfusion balance between maternal and fetal circulations (Melzer et al., 2010;Ferraro, Gaudet and Adamo, 2012). Physical activity can also aid regulation of circulating blood glucose levels (Collings et al., 2020a) which could help to reduce the risk of placental dysfunction and fetal hyperinsulinemia, thereby impacting fetal metabolism, oxygen demand and lactate accumulation (Jarmuzek, Wielgos and Bomba-Opon, 2015). As well as acute neonatal stress, each of these factors can have implications for the cord blood pH, levels of which provide an objective marker of the neonate condition immediately prior to birth and closely relate to Apgar scores (Victory et al., 2004). Low cord blood pH has been shown to predict neonate morbidity and mortality and longer term neurodevelopmental impairment (Malin, Morris and Khan, 2010;Yeh, Emary and Impey, 2012;Kelly et al., 2018;Vesoulis et al., 2018). There is currently limited support for an association between pregnancy physical activity with cord blood pH, but the evidence-base is low quality, characterised by small studies that were likely underpowered to detect associations (Davenport et al., 2018).
The objective of this study was to investigate associations of maternal pregnancy physical activity with cord blood pH measured in neonates who were born in poor condition, or for whom there was a concern during labour or immediately following birth. Evidence for a positive association with cord blood pH could strengthen the case that physical activity is safe and potentially beneficial for healthy pregnant women and their neonates.

Study design
Born in Bradford (BiB) is a prospective birth cohort study of 12,453 pregnant women who were recruited between 2007-10 whilst attending routine antenatal appointments at Bradford Royal Infirmary, the only maternity unit serving the city. Bradford is the sixth largest metropolitan borough in England and is one of the most deprived and ethnically diverse cities in the country (Wright et al., 2013). In a subsample of BiB participants, cord blood samples were taken and pH was analysed on the labour ward. This occurred when there was ' concern about the baby during labour or immediately following birth' or if the baby was born in poor condition (National Institute for Health and Care Excellence, 2007). The final sample comprised 1,467 mother-neonate pairs. As cord blood pH was measured only when clinically indicated the pH subgroup was skewed toward greater representation of nulliparous women (who are higher risk for pregnancy complications (Bai et al., 2002)) and more caesarean section deliveries. The characteristics of the included subgroup were otherwise similar to all other BiB participants (n = 12,391 mother-neonate pairs; see Supplementary Table S1) who were broadly representative of the obstetric population in Bradford at the time of recruitment (Wright et al., 2013). The BiB study was approved by the Bradford Research Ethics Committee (ref 07/H1302/112) and all mothers provided written informed consent.
Maternal physical activity was assessed at 26-28 weeks gestation using the General Practice Physical Activity Questionnaire (GPPAQ) which has been validated against accelerometry and exhibits face validity in the BiB pregnancy cohort (National Health Service, 2009;Collings et al., 2020aCollings et al., , 2020b. Mothers were grouped into one of four activity levels (inactive/somewhat active/moderately active/active) based on their self-reported occupational physical activity level, physical exercise and walking. The active category is consistent with meeting the recommended minimum of 150 minutes per week of moderate intensity physical activity (Department of Health and Social Care, 2019). Full details of the GPPAQ including the scoring system used to derive activity categories are shown in Supplementary Figure S1.
Following clinical guidelines that were in operation at the time, in the event there was ' concern about the baby either in labour or immediately following birth', or if the baby was born in poor condition with a 1-minute Apgar score of 5 or less, the umbilical cord was double-clamped and venous and arterial blood samples were taken for gas analysis by clinical staff (Thorp, Dildy and Yeomans, 1996;National Institute for Health and Care Excellence, 2007). The data were retrieved from obstetric records and were used to derive a third variable that represented the lowest cord blood pH recorded from either umbilical sample (Kelly et al., 2018).
Women consented to the abstraction and use of their data from obstetric medical records and at recruitment completed an interviewer administered questionnaire. Full details of all covariables have previously been described (Collings et al., 2020a).

Statistical analysis
Linear regression models were used to calculate differences in cord blood pH levels between the four groups of maternal physical activity (reference group: inactive); p-values from trend tests across physical activity categories are also presented. Models were initially adjusted for maternal age, ethnicity, early-pregnancy body mass index (measured at ~1 2 weeks gestation), socioeconomic status, parity, season of physical activity assessment, and neonate sex. Adjustments for maternal smoking, delivery mode, gestational age, birth weight, and neonate abdominal circumference were subsequently included because they changed β-coefficients between exposures and outcomes by ≥10% (Maldonado and Greenland, 1993). All dependent variables (arterial, venous and the lowest cord blood pH) were approximately normally distributed and results from the linear regression analysis are presented as marginal means with 95% confidence intervals. Adjusted for the same covariates, Poisson regression was used to quantify the relative risk of moderate cord blood acidaemia (pH < 7.10 (Yeh, Emary and Impey, 2012;Vesoulis et al., 2018)) between neonates of inactive women compared to women who were at least somewhat active in pregnancy (somewhat active, moderately active, and active groups combined). All analyses were performed in Stata/SE version 15.0 software and p < 0.05 was deemed statistically significant.

Participant characteristics
Descriptive statistics for the study sample are presented in Table 1. More than half of women (52.0%) were inactive in pregnancy, one-fifth were somewhat active (21.7%) and fewer were moderately active (14.6%) and active (11.7%), respectively. Inactive women were more frequently of Pakistani-origin, multiparous and were from moderately or the most deprived socioeconomic strata. Nearly one-tenth (8.5%) of the lowest recorded cord blood pH samples were moderately acidaemic (Supplementary Table S2 shows the number of cases of acidaemia stratified by cord blood source and physical activity level). Cord blood pH was positively related to Apgar scores ( Supplementary Figures S2 and S3). Table 2 shows cord blood pH levels stratified by maternal pregnancy physical activity. There was no evidence for effect modification by ethnicity (p ≥ 0.42) hence results are presented for the whole sample combined, adjusted for ethnic group. Significantly higher cord blood pH was observed in neonates of women who were somewhat active and moderately active in pregnancy compared to neonates of women who were inactive. The Poisson regression analysis provided some indication that, compared to neonates of inactive women, neonates of women who were at least somewhat active in pregnancy had lower risk of moderate cord blood  , pregnancy complications (gestational diabetes, gestational hypertension, gestational weight gain, emergency caesarean and medical or surgical induction), measurement factors related to maternal physical activity (gestational age at the time of reporting, if women were feeling well and were able to enjoy their normal daily activities, if maternal physical activity was reported during Ramadan) and cord blood pH (the time delay from birth to blood sampling (the median (IQR) time delay to arterial and venous cord blood samples was 10 (10) and 12 (10) minutes, respectively) did not appreciably influence associations and so were not retained in models. Models were rerun excluding samples that were missing data for sampling time (n = 61) and excluding samples that were delayed by ≥30 minutes following delivery (n = 82), the results were substantively unchanged. acidaemia (arterial pH (relative risk (95% confidence interval): 0.70 (0.48-1.03), p = 0.071; venous pH (0.62 (0.34-1.12), p = 0.11); lowest pH (0.73 (0.51-1.05), p = 0.086).

Discussion
This study found that maternal pregnancy physical activity was favourably associated with higher cord blood pH. The association was modest in size but may be underestimated due to errors in self-reported physical activity (Celis-Morales et al., 2012). It is also conceivable that any positive influence on cord blood pH may be clinically meaningful as lower values are associated with higher risk of serious adverse neonatal outcomes in a dose-dependent manner (Malin, Morris and Khan, 2010;Yeh, Emary and Impey, 2012;Kelly et al., 2018;Vesoulis et al., 2018). A recent meta-analysis concluded that there was low quality evidence from six randomised trials which together indicated there was no evidence for an association between pregnancy physical activity with cord blood pH (Davenport et al., 2018). However, an observational study reported that arterial cord pH was significantly higher in neonates of women who performed high-intensity exercise throughout pregnancy compared to neonates of women who had discontinued exercising (Clapp et al., 1995). Furthermore, cohort analysis of a controlled trial found that objectively measured physical activity at 16 weeks gestation was associated with higher venous cord blood pH and arterial oxygen saturation (Baena-García et al., 2019). The results provided by Baena-García et al. may be of limited clinical relevance, however, as their analysis was restricted to participants with 'normal' cord blood pH values >7.20 (Baena-García et al., 2019).
This study is the first to report associations between pregnancy physical activity and the full range of cord blood pH in a large and diverse sample of mother-neonate pairs from a deprived location. Our results appear to provide evidence for a threshold-type effect. Compared to neonates of women who were inactive, only neonates of women who were somewhat and moderately active in pregnancy exhibited higher cord blood pH. This is may be explained by fewer women being categorised as active, and hence insufficient statistical power at the higher end of the activity spectrum to detect what appear to be small associations. To account for this and to help decipher the clinical relevance of associations we calculated relative risks for moderate cord blood acidaemia between neonates of inactive women versus women who were at least somewhat active in pregnancy. We found some evidence that being at least somewhat active was associated with a lower risk of moderate acidaemia. The results only approached statistical significance, but all 95% confidence intervals were largely consistent with a protective effect (confidence limits indicated considerable risk reductions of ~4 0-50% associated with being at least somewhat active in pregnancy versus comparatively marginal risk increases). It is encouraging that modest volumes of mid-pregnancy physical activity do not appear to be detrimental and seem to be associated with some benefit in terms of higher cord blood pH, and possibly lower risk of moderate acidaemia. This is contrary to the popular belief of many women that physical activity in pregnancy is harmful (Coll et al., 2017). Our results support new UK guidelines which emphasise that ' every activity counts' and that inactive pregnant women should gradually accumulate physical activity throughout the week (Department of Health and Social Care, 2019).
The results of this study add to growing evidence that pregnancy physical activity is not only safe but beneficial for the short and long-term health prospects of both mother and child (Collings et al., 2020a(Collings et al., , 2020b. Placental function likely underpins the link between maternal physical activity and higher neonatal cord blood pH. Biological adaptations to regular pregnancy physical activity include increased size and vascularity of the placenta which aids oxygen perfusion to the fetus and removal of deoxygenated lactate filled blood (Melzer et al., 2010;Ferraro, Gaudet and Adamo, 2012). Activity-mediated improvements in maternal adiposity and glucose regulation (which we have previously observed in the BiB pregnancy cohort (Collings et al., 2020a)) may also assist prevention of placental dysfunction and fetal hyperinsulinemia, which could help curb adverse fetal metabolism and cord blood acidosis (Jarmuzek, Wielgos and Bomba-Opon, 2015;Aalipour et al., 2018). As with the results of any observational study, residual confounding by imperfectly measured covariates and unknown or unmeasured confounding cannot be excluded. Additional research is needed to replicate the current findings. Future studies should be carried out in unselected population-based samples of mother-neonate pairs, rather than clinically identified samples, which hinders generalisability of results.

Conclusion
In neonates born in poor condition, or for whom there was concern during labour or immediately following birth, maternal physical activity was favourably associated with slightly higher neonatal cord blood pH. Modest volumes of mid-pregnancy physical activity do not appear to be harmful and may enhance the condition of neonates at birth.

Data Accessibility Statement
Data generated and analysed for the current study are available from the corresponding author on reasonable request.