Evaluating the predictive value of systemic inflammation indices in early pregnancy loss: a retrospective observational study.

This retrospective study received approval from the Ethical Committee of Wuxi Women’s Hospital, Jiangnan University, Wuxi, China (approval numbers: 2023/06/0721 − 28). The Ethical Committee waived the patient’s informed consent form due to the nature of the retrospective study. This study was performed in accordance with the Medical Association Declaration of Helsinki. A total of 663 women clinically diagnosed with unexplained spontaneous early pregnancy loss (within 13 weeks of gestation) from January 2022 to December 2022 were included in this cross-sectional retrospective study. Early pregnancy loss was diagnosed and confirmed with the absence of foetal cardiac activity by ultrasound in the first trimester. Additionally, during the same study period, 700 women with a healthy pregnancy outcome at term and without any medical condition were included as a control group. The definition of a healthy pregnancy outcome was the absence of maternal or perinatal complications throughout the gestation period. Blood samples (3 mL) were collected from the vein using EDTA-K2 Vacutainer blood collection tubes, and blood tests were performed on a Haematology Analyzer (Mindray, Shenzheng, China). Blood test data for the study group were collected at the time of diagnosis, before surgical procedures, meaning after the pregnancy loss had already occurred. Due to the nature of early pregnancy loss, there may have been a delay between the actual pregnancy loss and the time of blood collection. Additionally, the peripheral blood test data in the control group, which were performed at the first trimester (between 10 and 13 weeks), were retrospectively collected from the hospital databases. All cases involving in vitro fertilization (IVF), foetal abnormalities, molar pregnancies, or cases with a history of any chronic inflammatory diseases, such as rheumatic diseases and endometriosis, were excluded. Women with repeated pregnancy loss were excluded. Data included maternal age, parity, gravidity, gestational age, serum levels of b-hCG (human chorionic gonadotropin), and complete blood test results at the time of diagnosis. The systemic immune-inflammation index (SII) was calculated as platelet count × neutrophil /lymphocyte count (10^9/L). The systemic inflammation response index (SIRI) was calculated by neutrophil count × monocyte count/lymphocyte count (10^9/L). The neutrophil-to-lymphocyte ratio (NLR) was calculated by neutrophil /lymphocyte count (10^9/L). Pan-immune inflammation value (PIV) was calculated by neutrophil count × monocyte count × platelet count/lymphocyte count (10^9/L). The optimal cut-off value was calculated by identifying the point on the curve with a minimum distance from the left-upper corner of the unit square, maximizing Youden’s index [ 22 ]. Data were presented as mean and standard deviation (SD). T-test (Mann-Whitney test) was performed for Tables  1 and 3 using GraphPad Prism (version 10). Odds ratio (OR), or receiver operating characteristic curves (ROC), and area under the ROC curve (AUC) were performed using IBM SPSS (version 30.0). Statistical difference was considered when p  < 0.05. Table 1 The comparison of the complete blood cell counts between women with early pregnancy loss and those without complications Early pregnancy loss ( n  = 663) Controls ( n  = 700) P -value (Mann-Whitney test) Whole white cell counts (10^9/L, mean/SD) 7.36 ± 2.1 7.92 ± 2.0 < 0.0001 Neutrophil (10^9/L, mean/SD) 5.05 ± 1.81 5.46 ± 1.8 < 0.0001 Monocyte (10^9/L, mean/SD) 0.41 ± 0.13 0.42 ± 0.13 0.0010 Lymphocyte (10^9/L, mean/SD) 1.74 ± 0.52 1.92 ± 0.51 < 0.0001 Platelet (10^9/L, mean/SD) 237 ± 55 251 ± 53 < 0.0001 The comparison of the complete blood cell counts between women with early pregnancy loss and those without complications Table 2 The comparison of systemic inflammation indices between women with early pregnancy loss and those without complications Early pregnancy loss ( n =663) Controls ( n =700) P value (Mann-Whitney test) SII (mean/SD) 733 ± 387 755 ± 364 0.0629 SIRI (mean/SD) 1.28 ± 0.81 1.33 ± 0.81 0.2804 NLR (mean/SD) 3.07 ± 1.3 3.01 ± 1.33 0.4072 PIV (mean/SD) 313 ± 239 338 ± 225 0.0071 The comparison of systemic inflammation indices between women with early pregnancy loss and those without complications

The mean maternal age at the time of pregnancy loss was 31 (± 5) years, which did not significantly differ from the control group (29 ± 8 years, p  < 0.0001, Mann-Whitney test). The mean gestational age in women with early pregnancy loss was 9.75 ± 1.63 weeks. We first compared the complete blood count parameters between the two groups (Table  1 ). The mean counts of total white cells, neutrophils, monocytes, lymphocytes, and platelets in women with early pregnancy loss were significantly lower than those of controls (all p  < 0.001, Mann-Whitney test). The systemic immune-inflammation index (SII), systemic inflammation response index (SIRI), and neutrophil-to-lymphocyte ratio (NLR) did not differ significantly between the two groups. In contrast, pan-immune inflammation value (PIV) was significantly lower in women with early pregnancy loss than in controls (Table  2 , p  = 0.0071, Mann-Whitney test). However, although PIV was statistically associated with early pregnancy loss ( p  < 0.001, logistic regression analysis), its odds ratio was 1.000 (95% CI: 0.999–1.000), indicating a negligible effect size. In the multivariable logistic regression model including SII, SIRI, NLR, and PIV (Table  3 ), NLR emerged as the strongest independent predictor of early pregnancy loss, with an OR of 2.08 (95% CI: 1.274, 3.389; p  = 0.003). SII showed a small inverse association (OR = 0.998, p  = 0.021), while SIRI and PIV were not statistically significant predictors ( p  = 0.053 and 0.139, respectively). The multivariable model achieved an AUC of 0.579 (95% CI: 0.549–0.609, p  < 0.0001; Fig.  1 ). The optimal cut-off value was 0.4787, yielding a sensitivity of 57% and a specificity of 56%. Table 3 The odds ratio of systemic inflammation indices in women developing early pregnancy loss, compared to controls Odds Ratio 95% CI P -value SII  0.998 0.996, 1.000 0.021 SIRI 0.354 0.123, 1.015 0.053 NLR 2.078 1.274, 3.389 0.003 PIV  1.003 0.999, 1.007 0.139 The odds ratio of systemic inflammation indices in women developing early pregnancy loss, compared to controls Fig. 1 The area under the receiver operating characteristic for predicting early pregnancy loss using the combined systemic inflammation indices The area under the receiver operating characteristic for predicting early pregnancy loss using the combined systemic inflammation indices

In this retrospective study with a substantial sample size, we observed significantly lower counts of total white cells, neutrophils, monocytes, lymphocytes, and platelets in women with early pregnancy loss at the time of diagnosis compared to controls. The inflammation index PIV was also significantly lower in women with early pregnancy loss at the time of diagnosis. However, in the multivariable logistic regression model, the NLR was the strongest independent predictor of early pregnancy loss, although the model’s overall discriminative ability was limited (AUC = 0.579). During the implantation window, a controlled pro-inflammatory environment is necessary to promote trophoblast cell invasion and endometrial vascular remodelling [ 23 ]. Disruption of this balance, such as inappropriate local and systemic inflammation response, is associated with the pathogenesis of early pregnancy loss [ 21 ]. Recent evidence has suggested systemic inflammation response indices, including SII, SIRI, or NLR, in predicting inflammation-associated pregnancy complications, such as placenta accreta, ectopic pregnancy, and preeclampsia [ 15 – 17 ]. In addition, PIV has recently been used as a new composite circulating immune biomarker to predict inflammation status [ 18 ]. Given these associations, in this retrospective study, we investigated whether systemic inflammation indices could provide an additional clinical diagnostic or prognostic value in early pregnancy loss. Two previous studies with moderate sample sizes (up to 200 cases and up to 400 controls) [ 19 , 20 ] reported increased total white blood cell and neutrophil counts, reduced lymphocyte and platelet counts, and no differences in monocyte counts in women with early pregnancy loss. In contrast, we found significant reductions across all cell types in women with early pregnancy loss compared to controls who had no complications at term. These discrepancies could reflect differences in the timing of blood sampling. The complete blood test was performed at the sixth week of gestation in previous studies [ 19 , 20 ]. Our complete blood tests were performed at the time of diagnosis of pregnancy loss (9.7 ± 1.6 weeks of gestation). Additionally, those control groups did not confirm healthy pregnancy outcomes, potentially introducing selection bias. Physiologically, a successful pregnancy entails a state of mild inflammation in the early stage, suggesting the white cell counts, including lymphocytes and monocytes, could fluctuate with gestational age [ 24 ]. Neutrophils are good indicators of inflammation, and extracellular vesicles released from the placenta can activate maternal circulation neutrophils and stimulate neutrophil formation [ 25 , 26 ], and are involved in maternal immune adaptation [ 10 ]. Neutrophils play a pivotal role in the innate immune response, including phagocytosis and the release of various cytokines and molecular mediators [ 27 ]. Lower counts of lymphocytes indicated an increased risk of infection [ 28 ], and lower counts of monocytes also indicate medical conditions [ 29 ]. Our findings of lower peripheral blood parameter count may therefore reflect dysregulated inflammatory activity at the time of early pregnancy loss. However, ratios of these parameters are thought to provide more stable measures of inflammatory balance [ 30 ]. In our current study, we found no significant differences in SII, SIRI, and NLR between women with early pregnancy loss and controls. This result contrasts with prior studies that showed increased values of SII [ 20 , 31 ] or NLR [ 19 ] in women with early pregnancy loss, as a predictive biomarker. But our finding with no difference in NLR was similar to another study reporting no difference in NLR between healthy pregnancies and early pregnancy loss [ 32 ]. One difference in these indices between our study and the other studies may be due to the timing of blood test performance. In our study, the blood tests were performed after diagnosis, which indicated that pregnancy loss had already occurred. In previous studies, blood tests were performed before the onset of pregnancy loss. Inflammatory parameters may change dynamically and could normalize once pregnancy loss has occurred, even though in a very short window. Additionally, the controls in the two previous studies did not confirm whether they had no pregnancy complications across the gestation period. Furthermore, there is a limitation in comparing individual indices, as they share the same parameters and are therefore interrelated. A multivariable model can help adjust for these relationships. In our multivariable logistic regression analysis, including SII, SIRI, NLR, and PIV, we found that only NLR was an independent predictor of early pregnancy loss, with an OR of 2.08. This suggests that the odds of having an early pregnancy loss are 2 times higher than those of controls when using NLR as a predictive marker. It is essential that diagnostic testing be used to determine the presence or absence of a disease in clinical practice. The area under the ROC curve (AUC) is commonly used to measure the overall probability of a diagnostic test. It can be interpreted as the average value of sensitivities for all possible values, with values above 0.5 indicating diagnostic accuracy [ 33 ]. In our current study, when we combined four indices, the AUC was 0.579 (95% CI: 0.549, 0.609), exceeding the threshold of 0.5 [ 33 ]. However, the specificity in our study was relatively low, suggesting that using the combined four indices to predict early pregnancy loss remains limited. Although the interpretation of the AUC is dependent on the disease, the AUC presented in this study was not high enough to predict a potential early pregnancy loss. Future study combining systemic inflammation indices with clinical symptoms, medical history, or pregnancy history may enhance the predictive value of clinical insights. In conclusion, women with early pregnancy loss showed reduced peripheral blood cell counts, whereas systemic inflammation indices did not differ significantly from those of controls. The combined indices demonstrated limited predictive performance (AUC 0.579). These findings suggest that systemic inflammation, as measured by peripheral blood tests, is not strongly associated with early pregnancy loss.

Early pregnancy loss refers to the sudden death of a foetus or loss of a pregnancy before the 20th week, with 95% of early pregnancy loss occurring in the first trimester of pregnancy [ 1 ]. The incidence of early pregnancy loss ranges from 10% to 25%, depending on the region and ethnicity [ 2 , 3 ]. Clinical symptoms of early pregnancy loss may be dependent on the cause but often include abnormal vaginal bleeding, abdominal pain, and cramping. The underlying pathogenesis of early pregnancy loss remains unclear, although chromosomal abnormalities contribute to about half of early pregnancy loss [ 4 ]. Dysfunction in placental development is also associated with early pregnancy loss [ 5 – 7 ]. Normal implantation and pregnancy necessitate mild inflammation in the early stage of pregnancy, but excessive systemic or local inflammation, particularly in the uterus, may contribute to the pathogenesis of complications of pregnancy, including early pregnancy loss (reviewed in [ 8 ]). During normal pregnancy, extracellular vesicles are released from the placenta, and sex hormones such as progesterone [ 9 ] may stimulate decidual immune cells to produce inflammatory cytokines (reviewed in [ 8 ]) and regulate maternal immune adaptation [ 10 ]. Pro-inflammatory cytokines secreted by macrophages then activate lymphocytes. Furthermore, inflammatory cytokines also activate natural killer cells (NK cells) and the maternal adaptive immune system (reviewed in [ 8 ]). Given the pivotal role of the maternal immune system in regulating the inflammatory response [ 11 , 12 ] and its correlation with the pathogenesis or consequences of the disease, the circulating inflammatory response, including changes in parameters, may be associated with and serve as predictive or diagnostic markers for inflammation-associated conditions, such as early pregnancy loss. Neutrophils and lymphocytes play crucial roles in the inflammatory response, particularly in chronic inflammation, by migrating to destroy microorganisms and coordinating the immune response to infection. Therefore, the neutrophil-to-lymphocyte ratio (NLR) is widely used as a systemic inflammatory marker in response to various infectious and non-infectious diseases [ 13 , 14 ]. The systemic immune inflammation index (SII) has recently been suggested as a predictive marker for placenta accreta and preeclampsia [ 15 , 16 ]. Similar markers, such as platelet-to-lymphocyte ratio, have also been used as systemic inflammatory markers to predict tubal ectopic pregnancies [ 17 ]. The pan-immune inflammation value (PIV) has recently been proposed as a new composite circulating immune biomarker to predict or assess inflammation status [ 18 ]. Recent studies with moderate sample sizes reported that higher systemic immune-inflammation indices are associated with early pregnancy loss, suggesting these indices may serve as a potential predictive marker [ 19 , 20 ]. Although inflammatory processes are involved in normal pregnancy, early pregnancy loss is associated with an inappropriate inflammatory response [ 21 ]. Early diagnosis and/or prediction, along with appropriate management, are essential for women’s overall well-being. Therefore, we conducted this cross-sectional and retrospective study with a large sample size to investigate whether white blood cell counts, including their subtypes and platelet counts in peripheral blood from routine blood tests during the early pregnancy stage, as well as systemic immune-inflammation indices, are associated with early pregnancy loss.