Risk prediction model construction method and analysis method based on serum bile acid spectrum and cytokines

Abstract

According to the method, the serum bile acid spectrum and the cell factors are integrated to construct the multi-index prediction model, and the multi-index prediction model is used for evaluating the adverse pregnancy outcome risk of the ICP child patient. The previous research focuses on a single index, such as TBA or clinical characteristics, but neglects the synergistic effect of bile acid subfractions and immune inflammation pathways. The research finds that the OR value of the TLCA in the prediction of fetal intrauterine distress is up to 65.6 (95% CI: 14.0-306.8), and the TLCA shows a strong risk prediction capability.

Description

Technical Field

[0001] The present invention belongs to the field of public medicine and relates to a method for constructing a risk prediction model based on serum bile acid profile and cytokine and an analysis method. Background Art

[0002] Intrahepatic cholestasis of pregnancy (ICP) is a common pregnancy-specific liver dysfunction that often occurs in the second and third trimesters and is characterized by pruritus and elevated serum total bile acid (TBA). ICP not only affects maternal quality of life but also significantly increases the risk of adverse fetal outcomes, leading to fetal distress, meconium-stained amniotic fluid, spontaneous preterm birth, and a significant increase in perinatal mortality. Fetal distress is reported to occur in 10% to 30% of patients with ICP and is a major cause of unplanned cesarean sections and neonatal intensive care unit admissions. Currently, the clinical diagnosis of ICP relies primarily on serum TBA levels and clinical symptoms, but their ability to predict fetal outcomes is limited, making individualized risk stratification difficult. Therefore, the design and construction of a risk prediction model for ICP is of great significance. Summary of the Invention

[0003] In response to the deficiencies of the existing technology, the present invention provides a risk prediction model construction method and analysis method based on serum bile acid profile and cytokine.

[0004] To achieve the above object, the present invention provides the following technical solutions:

[0005] The method for constructing a risk prediction model based on serum bile acid profile and cytokine profile includes the following steps:

[0006] S1: Determination of bile acid subfractions in serum of pregnant women with ICP by high performance liquid chromatography-tandem mass spectrometry;

[0007] S2: Simultaneous detection of serum cytokine levels;

[0008] S3: Construct a binary logistic regression model using the glm function.

[0009] Disturbances in bile acid metabolism are not only a key pathological feature of ICP but may also contribute to fetal injury by activating inflammatory pathways, inducing oxidative stress, and cellular apoptosis. Bile acids are steroidal compounds synthesized from cholesterol and can be divided into primary bile acids (cholic acid (CA) and chenodeoxycholic acid (CDCA)), secondary bile acids (deoxycholic acid (DCA), lithocholic acid (LCA), and ursodeoxycholic acid (UDCA), as well as free and conjugated forms conjugated to glycine or taurine. Studies have shown that serum TBA levels are significantly elevated in ICP patients, with the predominant increase in conjugated bile acids, particularly taurocholic acid (TCA) and taurochenodeoxycholic acid (TCDCA). Different bile acid subfractions significantly impact pregnancy outcomes. The hydrophobic bile acids LCA and DCA are cytotoxic and can cause hepatocyte apoptosis and liver damage, increasing the risk of adverse pregnancy outcomes such as preterm birth, fetal distress, and intrauterine fetal death. The hydrophilic bile acid UDCA has a protective effect, mitigating liver damage caused by cholestasis. Furthermore, the role of immune inflammatory responses in ICP is gaining increasing attention. Th17 cell-related cytokines (such as IL-17), macrophage markers (such as CD-68), and anti-inflammatory factors such as progranulin (PGRN) and UDCA have all been shown to correlate with ICP severity and pregnancy outcomes. Therefore, monitoring changes in bile acid subfractions and cytokine concentrations can help assess the risk of intrauterine distress, meconium-stained amniotic fluid, and spontaneous preterm birth in pregnant women and fetuses with ICP.

[0010] This study used high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS / MS) to determine 15 bile acid subfractions in the serum of pregnant women with ICP and simultaneously detected serum-related cytokine levels. The aim was to construct a comprehensive risk prediction model integrating bile acid profiles and cytokines, assess the severity of ICP in pregnant women, and predict adverse outcomes in children, in order to improve the clinical management of ICP and reduce perinatal mortality.

[0011] Furthermore, the bile acid subcomponents in the serum of pregnant women with ICP in S1 include TUDCA, TLCA, GCDCA, PGRN, IL-17, CD-68, UDCA, IL-4, and GUDCA.

[0012] Furthermore, a binary logistic regression model was constructed in S3 using the glm function, including univariate logistic regression analysis of serum indicators;

[0013] Furthermore, a binary logistic regression model was constructed in S3 using the glm function, including multivariate logistic regression analysis of serum indicators;

[0014] Another object of the present invention is to provide an application of a risk prediction model in ICP fetal intrauterine distress.

[0015] Furthermore, an ICP fetal distress prediction model was established based on TLCA, TCA, and UDCA for its application in ICP fetal distress.

[0016] Another object of the present invention is to provide an application of a risk prediction model for the occurrence of meconium-stained amniotic fluid in an ICP fetus.

[0017] Furthermore, an ICP fetal distress prediction model was established based on TUDCA, GCDCA, PGRN, TCDCA, IL-17, GLCA, CD-68, UDCA and IL-4 for the application of meconium-stained amniotic fluid in ICP fetuses.

[0018] Another object of the present invention is to provide an application of a risk prediction model for premature birth in ICP fetuses.

[0019] Furthermore, a model for predicting premature birth in ICP fetuses was established based on TUDCA, TCDCA, IL-17, GLCA, CD-68 and UDCA for its application in ICP fetal intrauterine distress.

[0020] This study integrated serum bile acid profiles with cytokines to construct a multi-indicator prediction model to assess the risk of adverse pregnancy outcomes in infants with ICP. Previous studies have focused on single indicators, such as TBA or clinical features, while ignoring the synergistic effects of bile acid subfractions and immune-inflammatory pathways. This study found that TLCA had an OR of 65.6 (95% CI: 14.0–306.8) in predicting fetal intrauterine distress, demonstrating strong risk prediction capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 This is an analysis of the correlation between bile acid subfractions and cytokines in mild, moderate and severe ICP.

[0023] Figure 2 A risk prediction model for ICP fetal distress.

[0024] Figure 3 A risk prediction model for meconium-stained amniotic fluid in fetuses with ICP.

[0025] Figure 4 A risk prediction model for premature birth in ICP fetuses was developed. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] Serum samples were collected from 148 patients with ICP who presented to the Affiliated Hospital of Southwest Medical University and Yibin Hospital of Traditional Chinese Medicine between February 11, 2021, and January 19, 2024. The samples were stored at -80°C. Clinical information and pregnancy outcomes were also collected. According to the "Guidelines for the Diagnosis and Treatment of Intrahepatic Cholestasis of Pregnancy (2015)", the patients were divided into a mild-to-moderate ICP group (88 patients) and a severe ICP group (60 patients).

[0028] Patients with ICP met the diagnostic criteria of the "Guidelines for the Diagnosis and Treatment of Intrahepatic Cholestasis of Pregnancy (2015)", including symptoms such as pruritus and jaundice during pregnancy and elevated serum trisodium benzoate (TBA) levels. All participants provided informed consent and signed a written informed consent form. This study was approved by the Ethics Review Committee of the Affiliated Hospital of Southwest Medical University (No. KY2020073) and the Ethics Review Committee of the Yibin Hospital of Traditional Chinese Medicine (No. 2022071802).

[0029] Patients with concurrent liver diseases such as viral hepatitis, drug-induced liver injury, autoimmune liver disease, etc. Patients with concurrent medical diseases such as diabetes and hypertension. Patients who have recently taken medications that affect bile acid metabolism.

[0030] An AB SCIEX Triple Quad 4500MD HPLC-MS / MS detection system and 15 bile acid detection kits from Nanjing Pinsheng Medical Technology Co., Ltd. were used to detect cholic acid (CA), deoxycholic acid (DCA), chenodeoxycholic acid (CDCA), ursodeoxycholic acid (UDCA), lithocholic acid (LCA), as well as taurocholic acid (TCA), taurodeoxycholic acid (TDCA), taurochenodeoxycholic acid (TCDCA), tauroursodeoxycholic acid (TUDCA), taurolithocholic acid (TLCA), glycocholic acid (GCA), glycodeoxycholic acid (GDCA), glycochenodeoxycholic acid (GCDCA), glycoursodeoxycholic acid (GUDCA), and glycolithocholic acid (GLCA) in serum. Alanine aminotransferase (ALT), aspartate aminotransferase (AST), total bilirubin (TBI), triglycerides (TG), and total bile acid (TBA) were measured using a Mindray BS-2000 series fully automated biochemical analyzer. Progranulin (PGRN), CD-68, IL-4, IL-12, IL-17, and TNF-α were measured using enzyme-linked immunosorbent assays (ELISAs). PGRN and CD-68 kits were purchased from Shanghai Lianmai Biotechnology Co., Ltd., while IL-4, IL-12, IL-17, and TNF-α kits were purchased from Shenzhen Kerunda Biotechnology Co., Ltd.

[0031] Intergroup comparisons were performed using IBM SPSS 20.0 software. Data were normally distributed and homogeneous, and were expressed as mean ± standard deviation (mean ± SD). Two-group comparisons were performed using the independent sample t-test, whereas differences were expressed as median and interquartile range (P<0.05, P<0.05). Intergroup comparisons were performed using the Mann-Whitney U test. A binary logistic regression model was constructed using the glm function in R (4.2.1), and a nomogram model was constructed using the RSM (6.4.0) function.

[0032] Example 1

[0033] Comparison of serum indicators between patients with mild to moderate ICP and severe ICP

[0034] Serum TLCA, TCA, TCDCA, GLCA, and GUDCA levels were significantly higher in ICP patients than in mild-to-moderate ICP patients, whereas serum TUDCA, GCDCA, and UDCA levels were significantly lower in severe ICP patients than in mild-to-moderate ICP patients (all P < 0.05). Serum PGRN levels were significantly lower in severe ICP patients than in mild-to-moderate ICP patients (P < 0.001). Serum IL-4, IL-17, and CD-68 levels were significantly higher in severe ICP patients than in mild-to-moderate ICP patients (all P < 0.05).

[0035] Table 1. Comparison of serum parameters in patients with mild to moderate ICP and severe ICP

[0036]

[0037]

[0038] Univariate logistic regression analysis of serum indicators in patients with mild to moderate ICP and severe ICP

[0039] Univariate logistic regression analysis showed that TUDCA, GCDCA, PGRN, and UDCA played a protective role in the severity of ICP, while TLCA, IL-17, CD-68, and IL-4 played a promoting role in the severity of ICP.

[0040] Table 2. Univariate logistic regression analysis of serum indicators in patients with mild to moderate ICP and severe ICP

[0041]

[0042]

[0043] 2.3 Multivariate logistic regression analysis of serum indicators in patients with mild and severe ICP

[0044] Multivariate logistic regression analysis showed that TUDCA and UDCA were independent protective factors for the progression of ICP from mild to moderate to severe, and were positively correlated with the cytokine PGRN and negatively correlated with IL-17 and IL-4 ( Figure 1 ); TLCA, IL-17, and GUDCA were independent risk factors for the progression of mild to moderate ICP to severe ICP in patients, but TLCAT and GUDCA had no correlation with IL-7 (Table 3).

[0045]

[0046] Example 2 Univariate and multivariate logistic regression analysis of ICP fetal distress

[0047] The TLCA, TCA, UDCA, and LCA levels were significantly different between the ICP fetuses in the distress group (distress group) and the ICP fetuses in the non-distress group (non-distress group) (all P < 0.01). Univariate and multivariate analyses showed that TLCA (OR = 65.6, 95% CI: 14.0-306.8, P < 0.001) was an independent risk factor for fetal distress, as was UDCA (OR = 0.787, 95% CI: 0.643-0.964, P = 0.021), as shown in Table 4.

[0048] Table 4. Univariate and multivariate logistic regression analysis of ICP fetal distress

[0049]

[0050] Risk prediction model for ICP fetal distress

[0051] A prediction model for ICP fetal distress was established based on TLCA, TCA, and UDCA. The C index of the model was 0.954 (95% CI: 0.920-0.989, P < 0.001), which can well predict the risk of ICP fetal distress ( Figure 2 ).

[0052] Example 3

[0053] Univariate and multivariate logistic regression analysis of meconium-stained amniotic fluid in ICP fetuses

[0054] As shown in Table 4 , univariate analysis showed that IL-17, GLCA, CD-68, and IL-4 were risk factors for meconium-stained amniotic fluid in fetuses with ICP (all P < 0.001), and TUDCA, GCDCA, PGRN, and UDCA were protective factors for meconium-stained amniotic fluid in fetuses with CP (all P < 0.001). However, multivariate analysis did not identify independent risk factors for meconium-stained amniotic fluid (see Table 5 ).

[0055] Table 5. Univariate and multivariate logistic regression analysis of meconium-stained amniotic fluid in ICP fetuses

[0056]

[0057]

[0058] Risk prediction model for meconium-stained amniotic fluid in fetuses with ICP

[0059] A prediction model for meconium-stained amniotic fluid in ICP fetuses was established based on TUDCA, GCDCA, PGRN, TCDCA, IL-17, GLCA, CD-68, UDCA, and IL-4. The C index of this model was 0.973 (95% CI: 0.951-0.994, P < 0.001), which can well predict the risk of meconium-stained amniotic fluid in ICP fetuses. Figure 3 .

[0060] Example 4 Univariate and multivariate logistic regression analysis of ICP fetuses with premature birth

[0061] TUDCA, TCDCA, IL-17, GLCA, CD-68, and UDCA levels were significantly different between the ICP fetuses (preterm group) and the ICP amniotic fluid non-preterm group (non-preterm group) (all P < 0.05). Univariate and multivariate logistic regression analysis showed that TCDCA (OR = 1.269, 95% CI: 1.063-1.514) was an independent risk factor for preterm birth in ICP fetuses (Table 6).

[0062] Table 6. Univariate and multivariate logistic regression analysis of ICP fetuses with premature birth

[0063]

[0064]

[0065] Risk prediction model for premature birth in ICP fetuses

[0066] The C index of the ICP fetal premature birth prediction model established based on TUDCA, TCDCA, IL-17, GLCA, CD-68 and UDCA was 0.791 (95% CI: 0.698-0.884, P < 0.001), which can better predict the risk of premature birth in ICP fetuses ( Figure 4 ).

[0067] In practice, it has been found that as ICP worsens, serum levels of hydrophobic bile acids (e.g., TLCA, TCA, TCDCA, GLCA, and GUDCA) increase significantly, while hydrophilic bile acids (e.g., TUDCA, GCDCA, and UDCA) decrease significantly, consistent with reports by Glantz and Williamson et al. [1,2]. Studies have shown that hydrophobic bile acids have stronger cytotoxicity, can induce hepatocyte apoptosis, activate inflammatory pathways, and cross the placental barrier to affect fetal health. This study integrated the serum bile acid profile with cytokines to construct a multi-indicator prediction model to assess the risk of adverse pregnancy outcomes in children with ICP. Previous studies have focused on single indicators, such as TBA or clinical features, while ignoring the synergistic effects of bile acid subfractions and immune-inflammatory pathways. This study found that TLCA had an OR of 65.6 (95% CI: 14.0-306.8) in predicting fetal distress, demonstrating strong risk prediction capabilities. A meta-analysis by Ovadia et al. only confirmed an association between TBA and fetal distress but failed to distinguish the contributions of different bile acid isoforms. This study analyzed 15 bile acid isoforms using an LC-MS / MS system and found that TLCA was a strong predictor of fetal distress (OR = 65.6), significantly outperforming traditional indicators such as TBA, ALT, and AST. Furthermore, we revealed an independent association between IL-17 and meconium-stained amniotic fluid. This finding is consistent with the animal studies of Zhang et al., but this study is the first to validate its human applicability in a clinical cohort. In contrast, UDCA, a well-established therapeutic agent, also exhibits a clear protective effect. In this study, elevated UDCA levels were significantly associated with a reduced risk of fetal distress, meconium-stained amniotic fluid, and preterm birth, further supporting its clinical value in improving cholestatic symptoms and fetal outcomes.

[0068] The results of this study have guiding significance for clinical practice. A nomogram model based on bile acid profiles and cytokines demonstrated excellent predictive performance (C-index for intrauterine distress = 0.954, C-index for meconium-stained amniotic fluid = 0.973), providing clinicians with a personalized risk assessment tool. For example, fetal monitoring interventions should be prioritized for pregnant women with ICP and a TLCA level >3.0 μmol / L, while UDCA supplementation may be more targeted for patients with high TLCA levels. Furthermore, IL-17, a characteristic cytokine of Th17 cells, can participate in tissue damage by recruiting neutrophils and inducing the release of inflammatory factors (such as IL-6 and TNF-α), and has been shown to have proinflammatory effects in various autoimmune and infectious diseases. In this study, IL-17 was not only closely associated with the severity of ICP but was also an independent risk factor for meconium-stained amniotic fluid (OR = 6.80) and preterm birth (OR = 2.01), suggesting that Th17 immune responses may play a key role in fetal microenvironmental disturbances and placental dysfunction. CD-68, a classic marker of macrophage activation, is significantly elevated in patients with severe ICP and demonstrates good predictive efficacy in both meconium-stained amniotic fluid and preterm birth prediction models. This finding supports the hypothesis of "inflammation-driven injury," whereby abnormal activation of the maternal immune system may affect the fetal environment through the placental barrier, inducing premature birth or meconium excretion. This provides a theoretical basis for the development of new anti-inflammatory therapies (such as IL-17 inhibitors). Policymakers can refer to this model to optimize the tiered management guidelines for ICP and refer high-risk pregnant women to tertiary medical centers, thereby reducing perinatal mortality.

[0069] Although some studies have attempted to establish predictive models based on biochemical markers or single cytokines, most have been limited to small sample sizes or analyses of single complications, lacking multi-marker modeling and systematic validation. Therefore, constructing a comprehensive predictive model integrating bile acid subtypes and immune inflammatory factors and evaluating its predictive efficacy for different types of adverse pregnancy outcomes is of great significance for improving the clinical management of ICP and reducing perinatal risks.

[0070] This study has certain limitations. As a retrospective study, the sample size was relatively limited and the population source was not extensive enough. External validation is needed through a multicenter prospective cohort. Secondly, although 15 bile acid subtypes were detected, they did not fully cover all metabolic derivatives (such as sulfated bile acids), and there was a lack of in-depth analysis of the dynamic metabolic pathways of bile acids. In addition, this study did not consider the impact of genetic factors (such as ABCB11 / ABCB4 gene mutations) on the bile acid profile, which may explain the model's prediction bias in some cases. The model did not integrate fetal monitoring parameters (such as fetal heart rate monitoring and umbilical cord blood flow indicators), which may further improve the predictive power. These limitations point to the direction of future research, including expanding the sample size, integrating multi-omics data, and conducting interventional studies to verify the clinical practicality of the model.

[0071] This study demonstrated that a predictive model based on serum bile acid profiles (TLCA, UDCA) and cytokines (IL-17) can effectively identify individuals at high risk for complications in ICP patients. Its superior discriminatory performance (C-index: 0.791-0.973) provides an objective basis for clinical decision-making. In particular, a strong correlation between TLCA and fetal distress (OR = 65.6) was found, revealing that the pathological significance of specific bile acid subtypes far exceeds that of traditional TBA indicators. This model innovatively combines metabolic disorders with immune-inflammatory mechanisms, not only contributing to understanding the molecular mechanisms of ICP-induced fetal injury but also enabling precise stratified management through quantitative risk assessment. In the future, by optimizing assay standardization and integrating dynamic monitoring data, this model is expected to become an important tool for guiding personalized interventions (such as the timing of UDCA treatment and the timing of delivery), ultimately improving maternal and fetal outcomes.

[0072] Those skilled in the art will appreciate that various combinations and / or combinations of features described in the various embodiments and / or claims of the present invention may be made, even if such combinations and / or combinations are not explicitly described in the present invention. In particular, various combinations and / or combinations of features described in the various embodiments and / or claims of the present invention may be made, without departing from the spirit and teachings of the present invention. All such combinations and / or combinations fall within the scope of the present invention.

Claims

1. A method for constructing a risk prediction model based on serum bile acid profiles and cytokines, characterized by: The steps include: S1: Determination of bile acid subfractions in serum of pregnant women with ICP by high performance liquid chromatography-tandem mass spectrometry; S2: Simultaneous detection of serum cytokine levels; S3: Construct a binary logistic regression model using the glm function.

2. The risk prediction model construction method according to claim 1, characterized in that: The bile acid subcomponents in the serum of pregnant women with ICP in S1 include TUDCA, TLCA, GCDCA, PGRN, IL-17, CD-68, UDCA, IL-4, and GUDCA.

3. The risk prediction model construction method according to claim 1, characterized in that: In S3, a binary logistic regression model was constructed using the glm function, including univariate logistic regression analysis of serum indicators.

4. The risk prediction model construction method according to claim 1, characterized in that: In S3, a binary logistic regression model was constructed using the glm function, including multivariate logistic regression analysis of serum indicators.

5. Application of the risk prediction model construction method according to claim 1 in ICP fetal intrauterine distress.

6. The risk prediction model construction method according to claim 5 is to establish an ICP fetal intrauterine distress prediction model based on TLCA, TCA, and UDCA.

7. Application of the risk prediction model construction method according to claim 1 in the development of meconium-stained amniotic fluid in ICP fetuses.

8. The risk prediction model construction method according to claim 7 is to establish an ICP fetal intrauterine distress prediction model based on TUDCA, GCDCA, PGRN, TCDCA, IL-17, GLCA, CD-68, UDCA and IL-4.

9. Application of the risk prediction model construction method according to claim 1 in the occurrence of premature birth in ICP fetuses.

10. The method for constructing a risk prediction model according to claim 9 is to establish a model for predicting premature birth in an ICP fetus based on TUDCA, TCDCA, IL-17, GLCA, CD-68 and UDCA.

Images