Novel inflammatory markers for predicting adverse pregnancy outcomes in pregnant women with polycystic ovary syndrome: a retrospective case‒control study
摘要
To investigate the potential predictive value of novel inflammatory markers for adverse pregnancy outcomes in pregnant women with polycystic ovary syndrome (PCOS).
MethodsA retrospective case‒control study was conducted. Clinical data were collected from pregnant PCOS patients and healthy pregnant women who received regular prenatal care and delivered at our hospital between 2021 and 2023. All included subjects had relevant clinical data collected. Basic characteristics, novel inflammatory markers, and adverse pregnancy outcomes were compared between the two groups. A random forest classification model was applied to evaluate the potential predictive value of combined novel inflammatory markers for adverse pregnancy outcomes in women with PCOS.
ResultsA total of 528 pregnant women were included. Compared with those in the control group, pregnant women in the PCOS group demonstrated statistically significant differences in height, first-trimester body mass index (BMI), gestational weight gain, BMI at admission, systolic blood pressure (SBP) at admission, and diastolic blood pressure (DBP) at admission (P < 0.05). Among participants with adverse pregnancy outcomes, compared with the control group, the PCOS group had an increased monocyte-to-lymphocyte ratio (MLR) (0.29 ± 0.02), neutrophil-to-lymphocyte ratio (NLR) (3.56 ± 0.21), systemic inflammation response index (SIRI) (1.98 ± 0.19), platelet-to-lymphocyte ratio (PLR) (133.76 ± 5.26), and systemic immune inflammation index (SII) (834.27 ± 55.72) (P < 0.05). After adjusting for confounding factors and optimizing hyperparameters, the combined prediction model using these five novel inflammatory markers (MLR, NLR, SIRI, PLR, and SII) achieved an AUC of 0.815, an F1 score of 0.776, an accuracy of 79.40%, and a precision of 79.40% for adverse pregnancy outcomes in patients with PCOS.
ConclusionEvaluation via the random forest classification model revealed that the combined use of these novel inflammatory markers has potential predictive value for adverse pregnancy outcomes in pregnant women with PCOS. These findings merit attention from clinical practitioners, and corresponding preventive measures should be implemented.