Characterizing the follicular fluid metabolome: quantifying the correlation across follicles and differences with the serum metabolome.

objectiveTo compare the variability in metabolomes between the serum and follicular fluid, as well as across 3 dominant follicles.

designProspective cohort study.

settingAn academic fertility clinic in the northeastern United States, 2005-2015. PATIENTS: One hundred thirty-five women undergoing in vitro fertilization treatment who provided a serum sample during ovarian stimulation and up to 3 follicular fluid samples during oocyte retrieval. INTERVENTION(S): None. MAIN OUTCOME MEASURE(S): Samples were analyzed using liquid chromatography with high-resolution mass spectrometry and 2 chromatography columns (C18 hydrophobic negative and hydrophilic interaction chromatography [HILIC] positive). We calculated overall, feature-specific, and subject-specific correlation coefficients to describe how strongly the intensity of overlapping metabolic features were associated between the serum and follicular fluid and between the 1st-2nd, 1st-3rd, and 2nd-3rd follicles. Feature-specific correlations were adjusted for age, body mass index, infertility diagnosis, ovarian stimulation protocol, and year. RESULT(S): From the C18-negative column and the high-resolution mass spectrometry, 7,830 serum features and 10,790 follicular fluid features were detected in ≥20% of samples. After screening retention times and checking for 1:1 matching, 1,928 features overlapped between the 2 metabolomes. From the HILIC-positive column and the high-resolution mass spectrometry, after applying the same exclusion criteria, there were 9,074 serum features, 5,542 follicular fluid features, and 1,149 features that overlapped. When comparing the feature intensity of overlapping metabolites in the serum and the follicular fluid, the overall (C18, 0.45; HILIC, 0.63), median feature-specific (C18, 0.35; HILIC, 0.37), and median subject-specific (C18, 0.42; HILIC, 0.59) correlations were low to moderate. In contrast, among the overlapping features across all 3 follicles, the overall (C18, all 0.99; HILIC, all 0.99), median feature-specific (C18, 0.74-0.81; HILIC, 0.79-0.85), and median subject-specific (C18, 0.88-0.89; HILIC, 0.90-0.91) correlations between follicular fluid metabolomics features within a woman were high. CONCLUSION(S): We observed minimal overlap and weak-to-moderate correlation between metabolomic features in the serum and follicular fluid but a large overlap and strong correlation between metabolomic features across follicles within a woman. The follicular fluid appears to represent a novel matrix, distinct from serum, which may be a rich source of biologic predictors of female fertility and reproductive outcomes.