Association of first-trimester complete blood count–derived systemic immune-inflammatory indices with miscarriage in women with recurrent pregnancy loss: a retrospective cohort study

Objective To investigate whether first-trimester complete blood count (CBC)-derived systemic immune-inflammatory indices are associated with miscarriage in women with recurrent pregnancy loss (RPL). Methods This retrospective cohort study included 688 women with RPL managed at the First Affiliated Hospital of Henan University of Science and Technology. Participants were classified by the outcome of the index pregnancy into a miscarriage group ( n = 308) and a non-miscarriage group ( n = 380). Complete blood counts obtained in early pregnancy were used to calculate the neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), monocyte-to-lymphocyte ratio (MLR), and platelet-to-lymphocyte ratio (PLR). Each index was analyzed both as a continuous variable (per 1-SD increase) and across quartiles in three progressively adjusted logistic regression models. Non-linearity was assessed using generalized additive models and threshold analysis, discrimination using receiver operating characteristic (ROC) analysis, and robustness using sensitivity analyses that excluded extreme values. Results Compared with the non-miscarriage group, women who miscarried had higher neutrophil counts, lower lymphocyte and monocyte counts, and higher values of all five indices, namely NLR, SII, SIRI, MLR, and PLR (all P < 0.001). In the fully adjusted model, each 1-SD increase was associated with higher odds of miscarriage for NLR (OR = 5.37, 95% CI 3.94–7.31), SII (OR = 3.73, 2.82–4.95), SIRI (OR = 2.04, 1.67–2.48), MLR (OR = 1.70, 1.42–2.02), and PLR (OR = 2.48, 2.00–3.08). Quartile analyses showed consistent dose-response gradients for all indices (all P for trend < 0.0001). SII, SIRI, MLR, and PLR showed significant non-linear associations, whereas NLR was monotonic. NLR provided the highest discrimination (AUC = 0.796), and the remaining indices were modest (AUC 0.63–0.67). The associations were unchanged after excluding extreme values. Conclusion Higher first-trimester CBC-derived immune-inflammatory indices, particularly NLR, were associated with miscarriage in women with RPL. NLR showed the strongest discriminative performance among the five indices. The data-derived threshold patterns and inflection points identified in this cohort are exploratory and require validation in independent multicenter populations before clinical application.