O-234 Using serum metabolomics to identify biomarkers of viable early, intrauterine pregnancy: an untargeted 1H NMR-based approach

Study question Which metabolites are associated with a viable intrauterine pregnancy (VIUP) when compared to other early pregnancy outcomes (failed intrauterine and ectopic pregnancies)? Summary answer Serum levels of four metabolites (phenylalanine, alanine, glutamate and glutamine) were significantly altered in VIUPs compared to other early pregnancy outcomes. What is known already Around 10% of all intrauterine pregnancies are lost in the first trimester. A further 1-2% of pregnancies are located outside the endometrial cavity; these ectopic pregnancies are the leading cause of maternal mortality in the first trimester of gestation. Early miscarriages may also cause significant morbidity when bleeding or infection occurs. The symptoms of miscarriages and ectopic pregnancy are often similar (pain and bleeding), however, such symptoms are also common in VIUPs. To date, no biomarkers have been identified to differentiate VIUPs from non-viable and ectopic pregnancies. Study design, size, duration This is a prospective cohort study that included 332 pregnant women at less than ten weeks of gestation, who attended the early pregnancy assessment unit (EPAU) at Liverpool Women’s Hospital with pain and/or bleeding. Participants/materials, setting, methods Blood samples were collected from the 332 pregnant women prior to final clinical diagnosis of pregnancy outcome. Serum samples were subjected to NMR metabolomics profiling (14 spectra that did not meet the recommended minimum reporting standards were removed from subsequent analysis). 1D 1H-NMR spectra were acquired at 37 °C on a 700 MHz spectrometer. Relative metabolite abundances underwent statistical analysis using MetaboAnalyst 5.0 (p-value FDR adjusted). Main results and the role of chance Final pregnancy outcomes were as follows: one hydatidiform mole (0.3%), 48 ectopic pregnancies (14.4%), three pregnancies of unknown location (PULs, 0.9%), 78 failed pregnancies of unknown location (FPULs, 23.4%), 47 miscarriages (14.1%), two vanishing twin pregnancies (0.6%) and 153 VIUPs (45.8%). Due to small sample numbers, the hydatidiform mole, PULs and vanishing twin pregnancies were excluded from further analysis. To compare VIUPs to other pregnancy outcomes, ectopic pregnancies, FPULs and miscarriages were grouped together. Univariate analysis of serum metabolite concentrations identified four metabolites (phenylalanine, alanine, glutamate and glutamine) as significantly different in VIUPs compared to other pregnancy outcomes. Multivariate partial least squared discriminant analysis provided only weak correlation between the serum metabolome and pregnancy outcome. In summary, we have identified differences in the metabolome of women with VIUPs compared to other common pregnancy outcomes, which may provide diagnostic utility. Limitations, reasons for caution In this study, women with VIUPs presented with pain and/or bleeding. The presence of symptoms may influence the metabolome of this group versus VIUPs without symptoms, thus limiting the translation of our findings. Furthermore, environmental factors were not controlled (e.g. fasting status), making it likely that cohort heterogeneity was enhanced. Wider implications of the findings This study identifies a metabolite profile associated with VIUPs. These findings may be useful in the development of a diagnostic test to confirm VIUPs and thus exclude potentially life-threatening pregnancy outcomes. Such a test would be invaluable in clinical emergencies. Trial registration number NA