Vaginal microbiome and birth weight in a Zimbabwean cohort
 
Muchaneta Gudza-Mugabe^1,2,6*, Enock Havyarimana¹, Shameem Jaumdally¹, Christina Balle¹, Katie Lennard¹, Andrew Tarupiwa², Fortunate Mugabe⁴, Rooyen Mavenyengwa³ Lindi Masson¹, Heather Jaspan^1,5
1University of Cape Town, South Africa; ²National Microbiology reference laboratory, Zimbabwe; ³University of Zimbabwe; ⁴Harare Central Hospital, Zimbabwe; ⁵Seattle Children's Research Institute, USA; ⁶Letten Foundation Research Centre
 
Introduction
Poor birth outcomes are a major cause of child morbidity and mortality, particularly in developing countries.  Birth weight is an objective measure of birth outcomes, encompassing both preterm and growth restricted pregnancies. This study investigated the relationship between the vaginal microbiome during pregnancy and low birth weight (LBW).
 
Methodology
Vaginal swabs were collected from 420 pregnant Zimbabwean women between 13-35 weeks of gestation who were then followed until delivery.  Nugent score was used to determine bacterial vaginosis (BV) status. DNA was extracted and sequencing of the 16S rRNA gene V4 hypervariable region was performed using Illumina MiSeq. Cytokine analyses were performed using Luminex.
 
Results
There were 233 singleton pregnancies with complete sequencing and birth outcome data, of which 41 (17.6%) resulted in the delivery of LBW infants (<2500g). Women who delivered LBW infants had significantly lower vaginal alpha diversity than women who experienced normal births and were less likely to have BV compared to term births (Median Shannon indices 0.67 vs 1.05; p=0.02). However, no significant differences in beta diversity of Bray Curtis distances were evident between women who delivered LBW versus normal infants (ADONIS p= 0.06). The vaginal microbiota of the cohort clustered into three community types using Fuzzy clustering (C1 and C2 were Lactobacillus iners and crispatus dominant, respectively, whereas C3 was Gardnerella vaginalis dominated). Women who delivered LBW infants clustered mainly with C1.  Vaginal concentrations of multiple inflammatory cytokines and growth factors were significantly lower in women with LBW infants. In a multivariate model, preterm delivery, community type, pregnancy-induced hypertension and growth factors were independently associated with risk for LBW, after adjusting for important confounders such as maternal age.
 
Conclusion
In this cohort of Zimbabwean women, multiple factors measured during pregnancy were found to be predictive of birth weight, including preterm delivery, pregnancy-induced hypertension, a vaginal microbiome dominated by L. iners and a deficit of cellular growth factors.