Association between maternal haemoglobin and stillbirth: a cohort study among a multi-ethnic population in England

The study objectives were to examine the association of maternal haemoglobin with stillbirth and perinatal death in a multi-ethnic population in England. We conducted a retrospective cohort analysis using anonymised maternity data from 14 001 women with singleton pregnancies ≥24 weeks' gestation giving birth between 2013 and 2015 in two hospitals - the Royal Wolverhampton NHS Trust and Guy's and St Thomas' NHS Foundation Trust. Multivariable logistic regression analyses were undertaken to analyse the associations between maternal haemoglobin at first visit and at 28 weeks with stillbirth and perinatal death, adjusting for 11 other risk factors. Results showed that 46% of the study population had anaemia (haemoglobin <110 g/l) at some point during their pregnancy. The risk of stillbirth and perinatal death decreased linearly per unit increase in haemoglobin concentration at first visit (adjusted odds ratio [aOR] stillbirth = 0·70, 95% confidence interval [CI] 0·58-0·85, aOR perinatal death = 0·71, 95% CI 0·60-0·84) and at 28 weeks (aOR stillbirth = 0·83, 95% CI 0·66-1·04; aOR perinatal death = 0·86, 95%CI 0·67-1·12). Compared with women with haemoglobin ≥110 g/l, the risk of stillbirth and perinatal death was five- and three-fold higher in women with moderate-severe anaemia (haemoglobin <100 g/l) at first visit and 28 weeks, respectively. These findings have clinical and public health importance.