Interpretation
Lipid profile
We observed differences in lipid profiles between ethnicities. Pregnancy is a state of physiologic hyperlipidaemia to accommodate the growing fetus,43 but reference values for serum cholesterol during pregnancy have not been established. Values for both ethnicities fell within a ‘normal’ range reported in a single cross-sectional pregnancy study.44 The overall lipid profile was less favourable among Indian women with increased triglycerides and lower HDL levels, which are considered important components of the metabolic syndrome.26,45 In line with our findings and hypothesis, a systematic review and meta-analysis showed that triglyceride levels were significantly increased in women with GDM, compared with those without insulin resistance (across all three trimesters). In that study, HDL was also reduced among those with GDM (in the 2nd and 3rd trimesters), while there were no differences in LDL and total cholesterol levels.46 In addition to a less favourable lipid profile, our study showed that adiponectin (i.e. an adipokine protein with insulin-sensitising and ani-inflammatory effects) was lower among mothers of Indian ethnicity. Reduced adiponectin has been associated with obesity, diabetes and components of the metabolic syndrome,47,48 and the findings in this study may be of clinical importance to Indian mothers.
Placental biomarkers
As infants of Indian women are generally born with a lower mean birthweight compared with other ethnic groups,16,49-51and as birthweight and placental weight are highly correlated,52 one may expect lower levels of PlGF among Indian mothers. Nonetheless we observed the contrary, in line with an Asian study that found significantly higher PlGF levels among Indian, compared with Chinese and Malay women ≥18 weeks gestation (although not between 11 and 14 weeks).53 PlGF belongs to the vascular endothelial growth factor family, and is highly expressed in the placenta throughout pregnancy.54 PlGF promotes vascular placental growth, but is also found in other organs and stimulates angiogenesis during inflammatory response.55 Limited clinical data shows increased PlGF in obese non-pregnant adults and children with the metabolic syndrome and diabetes.56-58 While studies are small and data sometimes conflict, a positive association between elevated PlGF and GDM has been established.59-61 A recent literature review concludes that although large-scale research is missing, most studies agree with this observation.62 Similarly, increased levels of PlGF have been detected in the placentas of anaemic women, compared with non-anaemic controls.63 In both diabetic disease and anaemia the increase in PlGF is thought to be caused by an adaptive angiogenic response to a relatively hypoxic placental environment, seen with hyperglycaemia and low haemoglobin.63,64 While placental hypoxia is considered physiological during first trimester pregnancies,65 biomarkers were collected at 15±1 weeks gestation in this study. As women of Indian ethnicity have higher rates of both disorders, these pathways may explain some of the differences seen in PlGF in our study.
Metabolic/obesity biomarkers
Certain adipokines such as NGAL have been positively associated with low-level systemic inflammation in the metabolic syndrome, obesity, hyperglycaemia and insulin resistance in human studies.66,67 We found significantly lower levels of NGAL among Indian women compared with European. Animal studies, however, show conflicting results, with some reporting diet-induced obesity and insulin resistance in NGAL knockout mice, accompanied by an increase in pro-inflammatory mediators.68,69 The latter finding is consistent with our hypothesis, that lower NGAL in Indian women is associated with a less favourable metabolic health profile. Further research needs to be done into the relationship between NGAL and metabolic health, to fully understand our findings by ethnicity.
Other significant metabolic biomarkers in multivariable analysis were periostin and MMP-9. Periostin is an extracellular matrix protein involved with tissue remodelling and repair,70 and has been positively associated with obesity, type II diabetes, insulin resistance, cardiovascular disease, lipids, chronic inflammation, and polycystic ovarian syndrome.71-74 Periostin levels were significantly increased among Indian women in this study, compared with European. MMP-9 is also responsible for tissue remodelling, and is particularly overexpressed with diabetes and the metabolic syndrome, as a consequence of oxidative stress in endothelial cells.75,76 In our analysis MMP-9 levels were significantly reduced among Indian, compared with European mothers, which is in contrast to our hypothesis. The reason for this is unknown.
Inflammatory biomarkers
In a recent study we reported that Indian mothers have higher rates of perinatal death at extremely preterm gestations.8 We hypothesised that a pro-inflammatory environment during these early weeks may lead to preterm birth, as an important antecedent factor for perinatal death.77 In the current study, several important pro-inflammatory biomarkers were increased among women of Indian ethnicity, compared with European. For example, levels of ICAM-1 were significantly higher, which has been associated with an increased risk of preterm birth when analysed in serum,78-80cervicovaginal fluid,81 and amnion.82 Of these studies, only Chen et al. included serum samples at early gestation (mean=16 weeks), granted with a standard deviation of ±4.5 weeks.78 ICAM-1 is a surface glycoprotein expressed on endothelium and immune cells which regulates several inflammatory pathways.83 It is elevated in activated endothelium,83 in the metabolic syndrome,84,85 and seems to be inversely correlated to HDL.85 Another biomarker with pro-inflammatory properties, CXCL10, was significantly increased among Indian mothers compared with European. CXCL10 has been associated with a range of chronic inflammatory disorders, including type II diabetes and the metabolic syndrome.86-88 Some studies even suggest that CXCL10 is a potential biomarker for the onset of adipose tissue inflammation in obese people.89,90
Cardiovascular biomarkers
ANP is mostly known for its properties in cardiovascular, endocrine and renal functions, and is stored as the biologically inactive variant proANP in atrial myocytes.91 In addition, ANP indirectly influences glucose metabolism.92,93 In this study we observed significantly lower levels of proANP among women of Indian ethnicity. This is in line with our hypothesis, as other studies report an inverse relationship between different fragments of proANP (NT-ANP, MR-ANP) and the risk of developing diabetes.94,95 Furthermore, a Turkish study found progressively lower ANP levels in women with GDM,96although non-significant, while another study was able to confirm this.97
Furthermore, our findings are consistent with a previous study observing overall higher angiogenin levels among South Asian subjects, compared with European ethnicities in the United States.98Angiogenin levels rise during pregnancy with increasing gestation,99 and seem to promote placental vascular development during the prenatal period.99,100 Although not widely investigated, a meta-analysis found decreased levels of angiogenin among patients with type II diabetes, but this was not statistically significant.101 It is unclear whether this also occurs with GDM. We hypothesise that relative placental hypoxia (e.g. due to maternal hyperglycaemia or low haemoglobin) may play a role in the observed increased angiogenin level in pregnant Indian women in our study. Finally, ST2, a cardiovascular biomarker positively associated with diabetes and the metabolic syndrome (including related markers such as triglycerides, liver function and glucose),102-104 was found in lower levels among Indian women compared with European in the current study. The reason for this is unclear.
Routine testing
Although it is widely recognised that women of Indian ethnicity have higher rates of pre-existing diabetes,8,17,18 and increased odds of developing GDM,9-12,16 random blood glucose levels at 15±1 and 20±1 weeks did not differ between groups. This is not surprising, as women with pre-existing diabetes were excluded from enrolment in the SCOPE study. Additionally, there was a statistically significant difference in mean haemoglobin between ethnic groups. It is, however, unclear whether this finding is of clinical relevance as both mean values fell within normal range for pregnant women (>110 g/L). It may indicate that women of Indian ethnicity have a lower threshold for developing anaemia and associated disorders, as pregnancy progresses further into a physiologic anaemic state. In this study there was no difference in haemoglobin levels between non-vegetarian and vegetarian women of Indian ethnicity. While we observed a large difference in rates of anaemia between Indian and European (6.5% vs. 1.8% respectively), when anaemic women were excluded in sensitivity analyses biomarker profiles between Indian and European women did not change. This may be due to the low number of women with anaemia in each ethnic group, and further investigation is warranted.