Intrahepatic cholestasis of pregnancy and offspring neurodevelopment: bile acid-mediated mechanisms and long-term neurodevelopmental outcomes

Abstract

Purpose

Intrahepatic cholestasis of pregnancy (ICP) is a pregnancy-specific liver disorder characterized by pruritus and elevated maternal serum total bile acids. While clinical management has focused on bile acid-based risk stratification and delivery planning to reduce perinatal risks, emerging evidence suggests possible longer-term effects on offspring neurodevelopment. This review synthesizes current epidemiologic and mechanistic evidence linking ICP with offspring neurodevelopmental outcomes.

Methods

PubMed, Embase, and Web of Science were searched using controlled vocabulary and free-text terms. Relevant references were evaluated for inclusion in a stepwise approach.

Results

Large Nordic register-based studies report associations between in utero ICP exposure and increased risks of childhood neurodevelopmental diagnoses, particularly in earlier-onset disease. Follow-up cohorts suggest altered early-life growth trajectories related to biochemical severity. Mechanistically, elevated maternal bile acids are a plausible proximal exposure. Maternal bile acid overload can disrupt placental bile acid transport and handling, alter the feto-maternal bile acid gradient, and increase fetal bile acid burden. The placenta may further convert bile acid perturbations into inflammatory, oxidative stress, endoplasmic reticulum stress, and vasoactive signals that reshape the intrauterine milieu during sensitive windows of brain development. By contrast, downstream pathways involving microglial priming, neuroinflammation, and blood–brain barrier vulnerability are supported mainly by indirect evidence and require validation in pregnancy-relevant models.

Conclusion

Current evidence supports a life-course perspective on ICP beyond established perinatal risks. Maternal bile acids and placental dysfunction provide a coherent mechanistic framework for possible neurodevelopmental effects, but pregnancy-relevant mechanistic studies and longitudinal cohorts with refined exposure assessment are still needed.

Background

Intrahepatic cholestasis of pregnancy (ICP) is a liver disorder unique to pregnancy [1]. It typically presents with pruritus and elevated maternal serum total bile acids (TBA) [1], most often in the second or third trimester, and usually resolves after delivery [2]. The pathophysiology of ICP is multifactorial and remains incompletely understood, with contributions from hormonal, environmental, and genetic factors [2–6]. Many affected women report a personal and/or family history of ICP, and approximately 15–25% of cases have been attributed to pathogenic variants in genes involved in bile acids and phospholipid transport or synthesis [3, 7, 8]. Clinical management of ICP has primarily focused on reducing short-term fetal risks in the perinatal period [9, 10], including spontaneous and medically indicated preterm birth, meconium-stained amniotic fluid, fetal compromise, and stillbirth [11–13]. An individual patient data meta-analysis showed that these risks rise with worsening biochemical disturbance, particularly higher TBA, providing a key rationale for TBA-guided counseling and delivery planning [11]. Prospective population-based and registry studies have similarly reported a severity-dependent pattern of fetal risk, reinforcing TBA measurement as central to diagnosis and risk stratification [12, 13]. Major international guidelines advise confirming the diagnosis biochemically and using TBA to stratify disease severity. This stratification helps guide management decisions that balance gestational age against fetal risk [9, 10, 14]. At the same time, these guidelines acknowledge ongoing uncertainty about the optimal surveillance strategy and the degree to which treatment alters major perinatal outcomes. However, growing evidence indicates that maternal cholestasis may have effects that extend beyond the perinatal period [11, 15]. In this context, TBA is a measurable marker of cholestasis severity during pregnancy and serves as a practical exposure metric for mechanistic and epidemiologic studies of potential longer-term outcomes in offspring.

In addition to TBA-guided monitoring and delivery planning, pharmacologic treatment has focused on maternal symptom control and biochemical improvement. However, its impact on major perinatal outcomes remains uncertain. Ursodeoxycholic acid (UDCA) is the most commonly used therapy and is recommended by the Society for Maternal–Fetal Medicine (SMFM) as first-line treatment for pruritus, although guidance documents note that evidence for its impact on key perinatal outcomes is inconclusive [10, 16]. In the PITCHES randomized trial, UDCA did not reduce a composite adverse perinatal outcome, despite improving biochemical measures [16]. By contrast, an individual participant data meta-analysis reported that UDCA use was associated with lower odds of selected outcomes, including spontaneous preterm birth, while highlighting that effect estimates can vary with study design, patient population characteristics, and outcome definitions [17].

More recently, ICP has been considered within a developmental “programming” framework, in which the timing, duration, and magnitude of TBA elevation may represent a measurable prenatal exposure with potential relevance for longer-term child health [15, 18]. Within the developmental origins of health and disease framework, metabolic and inflammatory disruptions during sensitive windows of fetal development are thought to contribute to later disease susceptibility [18, 19]. This is biologically plausible for ICP because bile acids are not only involved in digestion but also act as signaling molecules that can influence placental function and fetal tissues [20, 21]. Consistent with this framework, a nationwide Swedish register-based cohort of approximately two million children reported that in utero exposure to maternal ICP was associated with higher risks of neurodevelopmental conditions in offspring, with stronger associations in early-onset disease, highlighting the need for further mechanistic and translational research [22]. This review synthesizes epidemiologic evidence linking ICP to longer-term offspring outcomes, with a focus on neurodevelopment, and maps a bile acid-centered maternal–placental–fetal framework to highlight mechanistic pathways and priorities for future translational research.

Methods

Three electronic databases (PubMed, Embase, and Web of Science) were searched using a combination of controlled vocabulary terms and free-text terms related to “intrahepatic cholestasis of pregnancy”, “bile acids”, “placenta”, “offspring”, and “neurodevelopment”. Additional records were identified by manually screening the reference lists of relevant articles and key reviews. Articles were evaluated for relevance in a stepwise manner. Studies were included if they addressed clinical features or management of ICP relevant to fetal exposure, ICP-related offspring outcomes, placental bile acid transport and handling, or mechanistic pathways linking maternal cholestasis with fetal neurodevelopment. Seminal earlier studies were retained where necessary to provide clinical or mechanistic context, particularly for placental bile acid transport, developmental programming, and blood–brain barrier or neuroimmune biology. Articles not directly relevant to the scope of this review were excluded.

ICP in brief: bile acids as a quantifiable maternal–placental–fetal exposure

ICP is typically defined by otherwise unexplained pruritus, often affecting the palms and soles, accompanied by elevated TBA, with or without abnormal liver function tests, during pregnancy, with postpartum resolution [1, 2]. Pruritus may precede biochemical abnormalities, which is clinically important because a single normal TBA measurement does not exclude evolving disease, and repeat testing may be warranted when symptoms persist [6, 10]. ICP also remains a diagnosis of exclusion [9, 10]. TBA is the cornerstone biomarker for diagnosis and risk stratification [2, 9–11]. Importantly, TBA is not only a diagnostic biomarker but also a practical metric for characterizing prenatal bile acid exposure relevant to long-term outcomes [22, 23]. Exposure to maternal hypercholanemia can be described by level (magnitude of elevation), timing (gestational window), and duration (persistence over time). Because TBA concentrations may change after UDCA is started, these exposure dimensions should be interpreted in the context of the sampling window and, where possible, longitudinal measurements. Beyond their classical digestive functions, bile acids act as signaling ligands, most notably through FXR and TGR5/GPBAR1, thereby shaping transcriptional programs and inflammatory pathways relevant to pregnancy physiology [24, 25]. Wu et al. reported that FXR is detectable in placental tissues and can influence placental bile acid handling by regulating transporter expression [26]. In this study, an FXR agonist increased placental protein of the bile salt export pump (BSEP) in a maternal cholestasis mouse model [26]. Yang et al. also found that in BeWo cells, taurocholic acid (TCA) upregulated OATP1A2, a member of the organic anion transporting polypeptide (OATP) family, and FXR overexpression further enhanced the TCA effect [27]. TGR5 is expressed predominantly in placental macrophages and is significantly reduced in ICP placenta [28]. Nevertheless, even the low TGR5 levels were sufficient to increase placental inflammatory gene expression through the PI3K/NF-κB pathway [29]. Maternal cholestasis can increase bile acid levels in the fetal circulation via the placenta [30, 31]. Fetal bile acids rely on placental export to the maternal compartment for metabolism and excretion because fetal bile acid metabolic capacity is developmentally immature and fetal enterohepatic circulation is absent [32]. Viewed as a maternal–placental–fetal condition, elevated maternal bile acids can alter placental transport and stress responses, increasing the plausibility of fetal exposure. This is especially relevant for neurodevelopment because ICP typically presents in mid-to-late gestation, coinciding with sensitive periods of fetal brain maturation [33]. This timing, in turn, provides a biological rationale for linking registry-based neurodevelopmental signals to testable placenta-to-brain pathways.

Epidemiologic signals for long-term offspring outcomes after ICP exposure

Epidemiologic interest in the potential long-term consequences of ICP has grown with the increasing availability of nationwide registers and longitudinal mother–child cohorts. Although ICP has traditionally been viewed primarily in terms of acute perinatal morbidity, accumulating evidence supports a life-course perspective in which the timing, duration, and magnitude of maternal bile acid exposure may influence offspring outcomes beyond birth.

Neurodevelopmental outcomes: register-based evidence and robustness checks

The most robust population-level evidence to date linking ICP to clinically diagnosed neurodevelopmental outcomes comes from large register-based cohorts. In a nationwide Swedish cohort of approximately two million singleton births, prenatal exposure to maternal ICP was associated with higher risks of attention deficit/hyperactivity disorder (ADHD), autism spectrum disorder (ASD), and intellectual disability [22]. Associations were stronger for earlier-onset ICP, consistent with a longer cumulative exposure period overlapping sensitive stages of neurodevelopment [22, 33]. Importantly, this study went beyond conventional registry analyses by incorporating extensive covariate adjustment and family-based comparisons, which strengthens inference against shared familial and environmental confounding [22].

In recent studies, exposure timing and biochemical severity repeatedly emerge as plausible effect modifiers. In follow-up cohorts, early childhood growth trajectories appear to vary with maternal biochemical severity, including bile acid levels [23]. This study and Chen et al.’s observation [22] are compatible with a dose–duration–timing framework and support biological plausibility. At the same time, causal interpretation requires attention to downstream pregnancy pathways. ICP is associated with higher risks of preterm birth and other perinatal complications [11–13, 34], which themselves are linked to later neurodevelopmental risk and may therefore mediate part of the observed association [11–13, 22]. Reflecting this complexity, the Swedish study complemented its prespecified models with additional analyses designed to probe robustness to pregnancy-pathway influences and shared familial factors [22]. Although bile acids are the central exposure metric used for perinatal risk stratification [11–13], single-time-point measurements may not capture longitudinal trajectories, bile acid composition, or cumulative exposure burden, which could be more informative for longer-term outcomes [35]. Residual confounding from shared familial and cardiometabolic factors, such as maternal adiposity, diabetes liability, and genetic risk, cannot be fully excluded, even though the Swedish cohort used family-based and other robustness analyses to reduce these biases [15, 22].

Beyond diagnoses: supporting postnatal phenotypes and biological plausibility

Because replication of diagnosis-based neurodevelopmental outcomes remains limited, evidence of measurable postnatal phenotypes in other domains is useful for supporting the plausibility of longer-term consequences following ICP exposure. A population-based nested case–control cohort study followed children to 36 months and reported less favorable early growth indices in infancy in relation to maternal biochemical severity, including bile acid levels, with partial catch-up in later toddlerhood [23]. Although growth measures are not surrogates for neurodevelopmental diagnoses, such follow-up shows that postnatal phenotypes after ICP exposure can be quantified in routine child-health records and may reveal differences in early-life growth trajectories [23]. Complementary support for a developmental programming interpretation comes from integrative human and animal studies suggesting altered metabolic profiles and increased susceptibility to diet-induced dysfunction in exposed offspring [15]. These findings are in keeping with the developmental origins of health and disease framework, which suggests that metabolic and inflammatory disturbances in utero during sensitive developmental windows can influence later disease susceptibility [18, 19]. While not substitutes for diagnosis-based neurodevelopmental outcomes, these measurable postnatal phenotypes strengthen biological plausibility and support broader investigation of neurodevelopment across diverse populations and study designs.

Bile acid-mediated mechanisms (as shown in Fig. 1)

Fig. 1.

Bile acid-mediated pathways linking intrahepatic cholestasis of pregnancy to offspring neurodevelopment

Elevated maternal TBA are the defining biochemical feature of ICP and are widely regarded as the most plausible proximal mediator linking maternal cholestasis to fetal outcomes. Growing evidence indicates that an increased maternal bile acid load can alter placental transporter activity and bile acid-handling capacity, shift the fetal–maternal bile acid gradient, and thereby increase the likelihood of fetal bile acid exposure. We summarize in the following sections what is currently known about how bile acid signaling may plausibly influence fetal neuroimmune pathways, blood–brain barrier vulnerability, and subsequent maturation of neural circuits, and synthesize these findings into an integrated model that highlights key evidence boundaries and translational research priorities.

Fetal bile acid exposure as the initiating event

A prerequisite for bile acid-mediated neurodevelopmental programming is that maternal cholestasis increases fetal bile acid exposure to a biologically meaningful extent [36, 37]. Human trophoblast membrane transport studies indicate that transplacental bile acid gradients are shaped by transporter activity at the fetal-facing trophoblast membrane, providing a physiological basis for regulated bile acid exchange between maternal and fetal compartments [36]. This exchange is critical because fetal bile acid handling is developmentally immature and fetal enterohepatic circulation is absent, making placental export to the maternal compartment the primary route for fetal bile acid clearance [32]. Accordingly, when maternal cholestasis increases the maternal bile acid load and/or compromises placental clearance capacity, bile acids can accumulate on the fetal side and raise bile acid levels in the fetal circulation [30, 31]. In experimental models of pregnancy cholestasis, placental transfer of bile acids from fetus to mother can be impaired, supporting the notion that disrupted “fetal-to-maternal clearance” may increase the fetal bile acid burden [37]. In human ICP, the fetal–maternal bile acid gradient can become reversed, and UDCA treatment has been associated with partial restoration of this gradient, suggesting that fetal exposure may be, at least in part, modifiable [38]. Human placental studies further provide direct evidence of transplacental bile acid passage and highlight placental handling as a determinant of fetal exposure [38–40]. At the tissue level, ICP has been linked to a reproducible “placental phenotype”, and bile acid exposure in placental models can reproduce key features consistent with placental dysfunction [41, 42]. Recent profiling studies also report altered expression of placental bile acid transporters in ICP [38], alongside transcriptomic signatures suggestive of vascular dysregulation and inflammatory activation [43, 44]. Reviews synthesizing placental bile acid homeostasis emphasize that the placenta actively regulates bile acid transport and signaling, and that fetal exposure reflects the balance between maternal bile acid load and placental handling capacity [29]. Taken together, experimental and human data support a coherent exposure pathway whereby maternal bile acid elevation disrupts placental bile acid transport and handling and, in turn, increases fetal bile acid exposure.

Placental responses to bile acid overload

Beyond serving as a transport interface, the placenta can convert bile acid perturbations into inflammatory, oxidative, and vasoactive signaling with potential downstream consequences for the fetus [29]. In trophoblast and placental models, bile acids have been shown to activate inflammatory pathways, including GPBAR1/TGR5-linked NF-κB signaling [29, 45–47], supporting a plausible route by which bile acid-driven placental inflammation could contribute to intrauterine programming. In parallel, the ICP placenta has been reported to show activation of mTOR signaling, reflected by increased pS6 and pAkt, alongside upregulation of endoplasmic reticulum (ER) stress markers such as BiP. In vitro, lithocholic acid, rather than taurocholic acid or UDCA, induces an mTOR-dependent ER stress response involving BiP and IRE1α and reduces trophoblast viability, effects that are partially attenuated by pharmacologic pathway inhibition [48]. In models of pregnancy cholestasis, maternal cholestasis has also been associated with placental oxidative stress and apoptosis, and UDCA has shown protective effects in this context, consistent with a bile acid–placental stress axis [44, 49, 50]. Altered placental vascular tone represents an additional pregnancy-specific pathway through which bile acid elevation could influence fetal oxygen and nutrient supply. Seminal studies in isolated human placental chorionic veins demonstrated vasoconstrictive responses to bile acids, providing a physiologic basis for placental perfusion disturbances and fetal stress signaling in ICP [39]. Mechanistic data further indicate that UDCA can inhibit placental uptake of taurocholate and attenuate taurocholate-associated vasoconstrictor effects in human placental tissue and vascular preparations, offering a translational link between bile acid exposure and placental functional responses [39].

More recently, placenta-based injury programs have been proposed as additional links between bile acid overload and placental dysfunction. Experimental data implicate ferroptosis, with taurocholic acid inducing ferroptotic injury via reduced FTH1, providing a bile acid-species-linked pathway to trophoblast cell death [51]. Multi-omics profiling further supports dysregulated autophagy as a mechanistic axis relevant to trophoblast function and placental stress responses [52, 53]. In parallel, bile acids have been linked to trophoblast mitochondrial dysfunction through NRF1/PGC-1α-associated mechanisms [42], and emerging evidence also implicates impaired angiogenesis and disrupted mitochondrial NAD⁺ transport via a BACH1/SLC25A51-centered pathway [54]. At the tissue level, FTIR microspectroscopy reveals broad biomolecular remodeling in ICP placenta consistent with oxidative stress and apoptosis [55]. Collectively, these emerging placenta-based pathways expand the mechanistic landscape of ICP beyond transporter dysfunction and provide testable routes by which bile acid-associated placental injury may contribute to downstream fetal effects.

Bile acid signaling and fetal brain relevance

Bile acids are increasingly recognized as multifunctional signaling mediators that act through nuclear and membrane receptors, including FXR and GPBAR1/TGR5, and can engage pathways involved in metabolism, cellular stress, and inflammation [26, 28, 40, 41, 56–58]. Once fetal hypercholanemia is established, these receptor-based effects provide a plausible route by which elevated circulating bile acids could influence the developing brain, either directly if bile acids reach the fetal central nervous system (CNS), or indirectly via bile acid-sensitive inflammatory and metabolic cues [26, 28, 40, 41, 56]. Importantly, human data from pregnancy provide emerging signals for bile acid-linked pathways in relation to later neurodevelopmental outcomes. In an untargeted metabolomics study of mid-pregnancy maternal serum, bile acid pathways were among the metabolic networks differentiating pregnancies whose offspring were later diagnosed with autism spectrum disorder [59], supporting the potential relevance of bile acid-related maternal metabolic states during gestation. Complementary experimental evidence further supports biological feasibility for bile acid-associated pathways to intersect with neurodevelopment. In a rodent study of parental microbiome-driven fetal programming, offspring behavioral phenotypes were associated with bile acid-linked features of the maternal microbiome, such as bile salt hydrolase expression, together with differences in the abundance of specific bile acids [60]. The same study reported accompanying alterations in hippocampal DNA methylation and gene expression in offspring, providing molecular correlates in the offspring brain alongside the behavioral phenotype [60].

Additionally, some indirect evidence regarding the potential impact of cerebral bile acid exposure on fetal neurodevelopment is derived from non-pregnancy models, most of which are based on adult systems. Mertens et al. summarized both direct and indirect routes by which bile acid signaling may influence the CNS, offering a useful framework once fetal exposure has been established [61]. More recent CNS-oriented syntheses focusing on bile acids and GPBAR1/TGR5 have consolidated receptor-mediated pathways relevant to neuroinflammation and neural homeostasis [62–65]. However, evidence that is both pregnancy-specific and directly informative for the fetal brain remains limited. These constraints frame the sections below, where microglial maturation and blood–brain barrier vulnerability are presented as testable downstream hypotheses that become most compelling once fetal exposure and placenta-derived signaling are in place.

Candidate fetal neurodevelopmental targets

Microglial maturation and neuroinflammatory pathways

Within the developing fetal brain, the microglial lineage serves as a plausible interface integrating signals from maternal hypercholanemia and placental inflammation. Microglia are essential for early neurodevelopment, coordinating synaptic pruning, circuit refinement, axonal myelination support, and innate immune surveillance [66, 67]. Their maturation follows tightly timed transcriptional programs and is highly sensitive to prenatal immune-metabolic perturbations, which can “prime” microglial set points and bias later neurobehavioral trajectories [67]. Direct ICP-specific evidence for microglial dysmaturation remains limited, but existing animal work supports fetal brain vulnerability under cholestatic pregnancy, including fetal brain injury-related changes in rat ICP models [68]. Importantly, bile acids can be pro-inflammatory within the developing CNS. In neonatal rodent studies, increased brain cholic acid was associated with microglial activation, reflected by an increased number of Iba-1⁺ microglia and a shift toward activated morphologies, along with higher levels of pro-inflammatory cytokines such as IL-6 and TNF-α, supporting biological plausibility that elevated bile acids can bias microglia toward a pro-inflammatory phenotype in the immature brain [69]. Taken together, a hypothesis-driven model for ICP is that placenta-driven inflammatory cues and fetal bile acid exposure shift microglial developmental trajectories, lowering the threshold for exaggerated neuroinflammatory responses during critical windows of circuit formation and thereby increasing susceptibility to later neurodevelopmental morbidity.

Blood–brain barrier development and vulnerability

The blood–brain barrier (BBB) is present and functional early in fetal life, but its tight-junction organization and transporter systems continue to mature across late gestation and the early postnatal window; these developmental dynamics can create stage-specific vulnerabilities to circulating inflammatory cues and metabolites [70]. Although direct ICP-specific BBB studies are scarce, the cholestasis literature provides a mechanistic framework that becomes highly relevant to ICP once fetal exposure is established. In a rat model of obstructive cholestasis induced by bile duct ligation, elevated circulating bile acids increased BBB permeability and disrupted tight junction organization, including altered distribution of occludin and ZO proteins, with evidence supporting Rac1-dependent mechanisms [71]. These findings are consistent with the broader bile acid–brain literature indicating that bile acids may cross the BBB through physicochemical diffusion and via transporter-mediated pathways, and that increases in systemic bile acids can be accompanied by disturbances in CNS homeostasis [72].

In ICP, placental transporter adaptations may be insufficient to prevent persistent accumulation of bile acids on the fetal side, particularly in severe disease, raising the plausibility that bile acids and placenta-linked cytokine signals could challenge BBB integrity at a developmentally sensitive time [38]. A BBB that is transiently compromised would be expected to increase fetal brain exposure not only to bile acids but also to circulating cytokines and other inflammatory mediators, thereby strengthening a mechanistic bridge to microglial activation and altered circuit maturation described in the preceding section [73]. Collectively, existing evidence is consistent with a testable ICP hypothesis in which maternal–placental adverse signaling and fetal hypercholanemia may lower the threshold for BBB dysfunction, facilitating neuroimmune and neuroinflammatory activation and increasing susceptibility to adverse neurodevelopmental trajectories.

Integrative mechanistic model and translational priorities

Overall, pregnancy- and placenta-focused studies support the view that increased fetal bile acid exposure and placental dysfunction represent central upstream events in ICP [36–38, 58]. Placental inflammation, ER stress, oxidative stress, and altered vascular tone offer biologically plausible indirect pathways through which maternal cholestasis may influence the fetal brain environment [39, 45, 48, 74]. Recent placenta-based mechanistic work, including ferroptosis, autophagy, mitochondrial dysfunction, angiogenic perturbations, and multi-omics profiling [42, 51–54, 74], highlights pregnancy-relevant nodes that can be incorporated into gestational models to test downstream effects on fetal microglial maturation, BBB integrity, and later neurobehavioral outcomes [61, 72, 73, 75–78]. By contrast, evidence for neuroimmune perturbations affecting microglial maturation and for BBB vulnerability currently comes largely from mechanistic studies in cholestasis models and from the broader CNS literature; in ICP, these should be presented as biologically plausible, testable downstream hypotheses that become most relevant once fetal bile acid exposure and placenta-derived inflammatory signaling are established, rather than as definitive pregnancy-specific mechanisms [13, 40, 41, 66, 71–73, 79]. This graded approach to the evidence strengthens biological plausibility while maintaining clear boundaries between direct and indirect support.

Clinical implications

Accumulating epidemiologic and mechanistic evidence linking ICP to longer-term offspring outcomes has implications for obstetric counseling and postnatal follow-up. Current management focuses on reducing well-established perinatal risks, particularly stillbirth, preterm birth, and fetal compromise, through bile acid-guided monitoring and gestational age-informed delivery planning [9–11, 13]. Guidelines emphasize serum bile acids for diagnosis and risk stratification, with heightened concern at higher concentrations, but longer-term child-health outcomes are not routinely considered in standard care pathways [9, 10]. From a clinical perspective, registry-based evidence suggests that ICP may represent a prenatal exposure with potential implications beyond the perinatal period. In the nationwide Swedish cohort of approximately two million singleton births, maternal ICP was associated with increased risks of offspring neurodevelopmental conditions, including ADHD, ASD, and intellectual disability, with stronger associations in early-onset disease, which may reflect differences in exposure timing and cumulative duration [22]. These observational findings are insufficient to justify changes in diagnostic criteria or recommendations on delivery timing based solely on long-term neurodevelopmental risk. However, they do support more nuanced counseling that clearly distinguishes well-established perinatal risks from evolving evidence on longer-term outcomes [9, 10, 22].

Serum TBA already underpin routine risk stratification. Signals related to biochemical severity and disease onset timing in epidemiologic analyses raise the hypothesis that longitudinal bile acid trajectories could help identify subgroups at higher risk who may benefit from targeted developmental surveillance [9, 10, 13, 22]. Although current guidelines do not recommend routine neurodevelopmental screening for children born after ICP pregnancies, it may be pragmatic to ensure that clinically severe or early-onset maternal disease is documented and communicated to primary care and pediatric teams, to support heightened vigilance during routine developmental surveillance [22]. Therapeutically, UDCA is widely used to improve maternal symptoms and liver biochemistry, whereas its effects on major perinatal outcomes remain variable across studies. The PITCHES randomized trial did not demonstrate a reduction in a composite adverse perinatal outcome with UDCA [16], and an individual participant data meta-analysis provides a more granular assessment of potential benefits across heterogeneous populations and outcome definitions [17]. At present, whether UDCA or other bile acid-modifying strategies influence longer-term offspring neurodevelopment remains unknown, representing a clinically important knowledge gap given the central role of bile acids in proposed mechanistic pathways [17, 39, 57]. ICP is a clinically identifiable condition with a quantifiable exposure metric and well-defined timing, making it a useful human model for developmental programming research at the interface of obstetrics, placental biology, pediatrics, and neurodevelopment [22, 57, 61, 79]. Integrating obstetric exposure data with longer-term pediatric follow-up may be particularly informative for refining counseling and postnatal care pathways in the future.

Evidence gaps and future research priorities

Despite rapid growth in epidemiologic and mechanistic work, several evidence gaps still limit the translation of potential long-term neurodevelopmental implications into routine care pathways. Neurodevelopmental outcomes are currently informed largely by Nordic registry-based cohorts, which offer scale and statistical power but remain susceptible to residual confounding and dependence on diagnostic coding and service access patterns [22]. In addition, registry definitions of ICP typically do not provide individual bile acid concentrations or longitudinal bile acid trajectories, and often lack information on bile acid composition and treatment exposure, which limits evaluation of biochemical severity, dose response patterns, and potentially modifiable pathways. Broader validation across healthcare systems, together with standardized developmental assessments beyond diagnostic codes, would improve phenotype resolution and strengthen causal inference [22, 76, 79, 80]. These registry-level constraints highlight the need for more granular exposure characterization. Maternal bile acid concentrations vary across gestation and may change with treatment. A frequent limitation in many datasets is incomplete information on UDCA timing, dose, and adherence, which complicates interpretation of TBA trajectories. Future cohorts should record treatment initiation and pair these data with longitudinal bile acid measurements to better separate disease severity from treatment-related changes. Single-time-point measurements may therefore fail to capture cumulative exposure, critical windows of vulnerability, or bile acid composition. Longitudinal bile acid trajectories, bile acid species profiling, and linkage to placental transporter expression would help clarify dose–response relationships and timing-sensitive vulnerability [38, 39, 44]. Human data showing altered fetal–maternal gradients and transplacental passage further support the feasibility of exposure assessment across maternal and fetal compartments in cohorts with detailed phenotyping [40].

Mechanistic inference would benefit from models that better reflect pregnancy biology. The most coherent ICP pathway begins with altered fetal exposure and placental dysfunction [36–38, 44], but downstream fetal brain endpoints are still rarely tested directly in gestational models. Integrating injury programs centered on the placenta into animal models that reflect pregnancy biology, or into advanced placental–fetal experimental systems could enable direct testing of fetal brain endpoints [81–83]. Recent molecular studies implicating ferroptosis and epithelial–mesenchymal transition-related processes provide a mechanistic basis for this approach, allowing evaluation of microglial maturation, blood–brain barrier integrity, and neurobehavioral phenotypes across development [51, 71, 74, 75, 79].

Treatment effects and long-term follow-up remain another major gap. While UDCA improves maternal symptoms and biochemistry, its effects on perinatal endpoints vary across study designs [16, 17], and experimental data support actions on placental bile acid handling and vascular responses [39]. Long-term follow-up of existing trial cohorts, alongside carefully designed observational studies capturing treatment timing and bile acid species, would help determine whether modifying maternal bile acid profiles translates into differences in later neurodevelopment [17, 39]. A more integrated maternal–placental–offspring framework is likely to be most informative. Placental biology is central to developmental programming, and conceptual models that link placental dysfunction to offspring health later in life provide a useful scaffold for studies in ICP [44, 76, 79]. Multidisciplinary studies that integrate obstetric exposure metrics with placental profiling, umbilical cord blood biomarkers, and standardized neurodevelopmental assessments in children will be critical for identifying subgroups most likely to benefit from targeted surveillance or early intervention.

Conclusion

In summary, ICP may not be confined to perinatal risks alone. An expanding body of population-based research suggests that intrauterine exposure to ICP may be associated with offspring outcomes beyond the neonatal period, including a higher risk of childhood neurodevelopmental diagnoses and measurable differences in early growth trajectories. Although the neurodevelopmental evidence is derived largely from Nordic registry studies, and observational designs cannot establish causality, these consistent signals support consideration of a life-course perspective when considering ICP. From a mechanistic standpoint, a pregnancy-specific sequence involving increased maternal bile acid burden, disrupted placental bile acid transport and handling, and consequent elevation of fetal exposure offers a coherent biological rationale for long-term effects. Placental inflammation, ER stress, oxidative stress, and vascular dysfunction may further shape the intrauterine environment during sensitive windows of brain development. Key priorities for future work include external validation in diverse populations, ideally through prospective cohorts. This should be coupled with refined exposure assessment that integrates longitudinal bile acid dynamics and profiles, placental and umbilical cord blood biomarkers, and mechanistic models that reflect pregnancy biology, to identify critical windows of exposure and potentially modifiable pathways. Such efforts will be essential to support evidence-based counseling and to guide long-term follow-up strategies for offspring exposed to ICP.

Author contributions

Conceptualization: YY H and X T. Writing—original draft: SY D. Data curation: SY D, H Z, YP X, Q C, QH C and YH L. Formal analysis: SY D. Investigation: SY D and H Z. Visualization: SY D. Writing—review and editing: SY D, H Z, YP X, Q C, QH C, YH L, D S, X T and YY H. Funding acquisition: D S and YY H. Supervision: X T and YY H. All the authors have read and approved the final manuscript.

Funding

This study was supported by the National Natural Science Foundation of China Youth Fund (No. 82301924) and the Sichuan Provincial Department of Science and Technology (No. 23LCYJ013).

Data availability

No datasets were generated or analyzed during the current study.

Declarations

Conflict of interest

The authors declare that they have no competing interests.

Ethics approval and consent to participate

Not applicable. This article is a review and does not involve human participants, human data, or animals.

Consent for publication

Not applicable.