Mass spectrometry of water-soluble vitamins to establish a risk model for predicting recurrent spontaneous abortion

The adverse pregnancy outcomes, including recurrent spontaneous abortion (RSA), are strongly correlated with water-soluble vitamins, but how to predict RSA occurrence using them remains unsatisfactory. This study aims to investigate the possibility of predicting RSA based on the baseline levels of water-soluble vitamins tested by ultra-liquid chromatography-tandem mass spectrometry. A total of 918 pregnant women was consecutively enrolled in this cross-sectional study. According to the miscarriage numbers, they were divided into normal first pregnancy (NFP, n = 608), once spontaneous abortion (OSA, n = 167), and continuous spontaneous abortion (CSA, n = 143) groups. The Cox proportional-hazards regression model was employed to establish a risk model for predicting RSA. The RSA occurrence was 6.54% in overall pregnant women, with a prevalence of 12.57% in the OSA group and 27.27% in the CSA group. Significant differences were observed in baseline deficiencies of vitamin B3, B5, B6, and B9 among NFP, OSA, and CSA groups (χ2 = 12.191 ~ 37.561, all P < 0.001). Among these vitamins, B9 (HR = 0.89 and 0.88, all P < 0.001) and B6 (HR = 0.83 and 0.78, all P < 0.05) were identified as independent factors in both the OSA and CSA groups; whereas B5 was identified as an additional independent factor only in the CSA group (HR = 0.93, P = 0.005). The Cox proportional-hazards model established using these three vitamins exhibited poor or satisfactory predictive performance in the OSA (Sen = 95.2%, Spe = 39.0%) and CSA (Sen = 92.3%, Spe = 60.6%) groups, respectively. However, B5, B6, and B9 compensatory levels were not associated with RSA occurrence (all P > 0.05). Our study presents a highly sensitive model based on mass spectrometry assay of baseline levels in B vitamins to predict the RSA occurrence as possible.