Intrahepatic cholestasis of pregnancy: Introduction and overview 2024

Abstract

Considerable progress has been made to explain the aetiology of intrahepatic cholestasis of pregnancy (ICP) and of the adverse pregnancy outcomes associated with high maternal total serum bile acids (TSBAs). The reported thresholds for non-fasting TSBA associated with the risk of stillbirth and spontaneous preterm birth can be used to identify pregnancies at risk of these adverse outcomes to decide on appropriate interventions and to give reassurance to women with lower concentrations of TSBA. Data also support the use of ursodeoxycholic acid to protect against the risk of spontaneous preterm birth. A previous history of ICP may be associated with higher rates of subsequent hepatobiliary disease: if there is a suspicion of underlying susceptibility, clinicians caring for women with ICP should screen for associated disorders or for genetic susceptibility and, where appropriate, refer for ongoing hepatology review.

Intrahepatic cholestasis of pregnancy (ICP) is the most common pregnancy-specific hepatic disorder. It affects 0.6–0.7% pregnant women of European ancestry in Australia and the UK, with relatively higher rates in Indigenous Australian women and in UK women of Indian sub-continent descent, somewhat higher in Scandinavia (1.5%), and markedly increased in the Andean Indian populations of South America (∼4%). ¹ A 1.75% incidence has also recently been reported in mainland China, ² whilst there are anecdotal reports of a high incidence in First Nations people in western Canada (Personal communication – Xavier Thompson (xavier.thompson@medportal.ca)). There is also an increased risk of ICP in the sisters of women with ICP. ³ The reasons for the varying incidence are not clear. Amongst other factors, a potential impact of dietary selenium has been discussed. ⁴ Genetic variants in bile acid and phospholipid transport may also contribute, as discussed by Peter Dixon later in this publication: up to 25% of women with ICP have been shown to be heterozygous for ABCB4 mutations, encoding a phospholipid floppase essential for biliary phospholipid secretion, and/or ABCB11 mutations, encoding the bile salt export pump (BSEP). ⁵

ICP has had various synonyms over the years since its possible first description by Ahlfield in 1883 as “recurrent jaundice in pregnancy”. ⁶ However, the reports from Scandinavia of, first, Svanborg, then Thorling in the 1950s, were perhaps the first to characterize recurrent jaundice of pregnancy with pruritus.^7,8 Alternative names for the condition have included: Jaundice in late pregnancy, Endogenous hepatotoxaemia of pregnancy, Benign jaundice of pregnancy, Icterus gravidarum, Hepatosis of pregnancy and Obstetric cholestasis, but the recent consensus is for Intrahepatic cholestasis of pregnancy, with the support from the main national groups: RCOG and ACOG, the specialist groups MOMS, BMFMS and SMFM, and the supranational groups SOMANZ and EASL.

The main maternal symptom of ICP is pruritus without rash or other skin alterations (apart from self-inflicted excoriations), which can range from mild, often just affecting soles and palms, to devastating and generalised, with sleep deprivation and major mental stress. Most common in the third trimester, pruritus can start also in the second trimester or – very rarely – even in the first trimester. ⁹ The potential mechanisms of pruritus are discussed below.

The most studied biomarker for ICP is the increased total serum bile acid (TSBA) concentration, although this may not be present initially and, in its absence, should be rechecked if symptoms are continuing, in particular in combination with increased serum ALT. ¹⁰ The importance of increased TSBA is its concentration-dependent association with complications of ICP, including the very rare, but most feared, stillbirth, as well as spontaneous preterm birth and neonatal asphyxia: the link relates to the peak values of TSBA that have been identified, ¹¹ as discussed by Caroline Ovadia later in this publication. Serum should be drawn in the non-fasting state for the measurement of TSBA in women with ICP, as TSBA rise post-prandially and in ICP it is important to know the peak concentration of bile acids (BAs) in the serum; many high-risk cases will be missed if only fasting samples are used to identify the risk. ¹² The upper limit of normal for enzymatic determination of TSBA differs between fasting (∼10 μmol/L) and postprandial (∼20 μmol/L) serum concentrations. ¹² It is recommended that measures are repeated in ICP as the woman gets closer to term (at least weekly after week 32 ¹³ ) so that appropriate decisions can be made in the light of risks associated with postprandial TSBA ≥100 μmol/L. ¹² It is not yet clear, however, whether a fall in TSBA from an identified peak value can indicate a reduction in risk. There has not yet been a consensus on the methodology and platforms for the laboratory analysis of TSBA, but harmonisation efforts are currently in progress, ¹⁴ as discussed later by Corey Markus and Bill Hague. Affordable routine enzymatic tests for determination of TSBA are available and commonly used in clinical practice. Mass-spectrometry techniques, such as liquid chromatography/tandem mass-spectrometry (LC-MS/MS), for the determination of individual serum BA appear unnecessary for routine clinical practice, take considerably longer time to provide a result and are much more costly. In addition, good data to validate the associated risks of increased individual serum BA concentrations are lacking for ICP. Still, individual serum BA determinations remain of interest for detailed academic studies.

From an obstetric perspective, other hepatic disorders of pregnancy, such as acute fatty liver, have been described alongside ICP so that clinicians need to be alert to such features. Multifetal pregnancy also increases the risk of ICP, perhaps related to the associated increase in pregnancy hormone concentrations. However, it is not established how the risks of ICP are affected by the common comorbidities, such as gestational diabetes (GDM) and pre-eclampsia, and advice should be sought from experienced clinicians in their management.

Other hepatic morbidities, such as metabolic dysfunction-associated fatty liver disease (MAFLD) [formerly: non-alcoholic fatty liver disease (NAFLD)], can be associated with an increase in serum BA, but pruritus is only seen in about 20%. ¹⁵ In a large prospective US study, 14% of pregnant women screened in early pregnancy were diagnosed with NAFLD. ¹⁶ Gilbert's syndrome, on the other hand, may be associated with lower concentrations of serum BA, with the possibility of false negative testing ¹⁷ : the increase in unconjugated bilirubin in Gilbert's may, therefore, be a useful distinguishing feature.

In the event of persisting pruritus in the absence of increased BA and any dermatosis apart from excoriations, and whilst awaiting any later change in BA, ¹⁰ empirical topical treatment can be considered despite the absence of clinical trials to support this. Our practice is to offer a menthol-based or a coal tar-based aqueous cream, sorbolene cream/lotion, pine tar solution, liquid paraffin or bicarbonate of soda baths, coupled with a regular non-sedating antihistamine agent (e.g. cetirizine) by day and/or a sedating antihistamine agent (e.g. promethazine) at night. Ultraviolet phototherapy has been used in pregnancy for other dermatoses, but there are no data available for its use in ICP. ¹⁸ If used, care should be taken to provide folic acid supplements as there is a risk of depletion. ¹⁹

Other laboratory surveillance, including ALT, bilirubin, INR, is not regularly required to monitor progress in ICP unless there are concerns about coincident pathology, such as pre-eclampsia, HELLP syndrome or acute fatty liver of pregnancy, or, if steatorrhoea develops, suggesting poor absorption of Vitamin K (and other fat-soluble vitamins), which warrants consideration of treatment with parenteral Vitamin K.

The pruritus in ICP can be devastating, but its molecular pathogenesis remains an enigma. In particular, the well-recognised predilection for palmar and solar pruritus in ICP is so far unexplained. Clinical observations and experimental therapeutic interventions have helped to improve understanding of pruritus and the potential pruritogens involved ²⁰ : (i) pruritogens in cholestasis are accumulating in the systemic circulation, as supported by the transient marked antipruritic effect of experimental interventions, such as plasmapheresis or molecular adsorbents recirculating system (MARS) dialysis ²¹ ; (ii) pruritogens are secreted into bile as suggested by impressive transient antipruritic effects of nasobiliary drainage in genetic cholestatic syndromes ²² or primary biliary cholangitis ²³ although with considerable risk of post-ERCP pancreatitis^22,24; (iii) pruritogens are (biotrans)-formed in the liver and/or small intestine as suggested by the antipruritic effect of rifampicin, which, in addition to its action as an antibiotic in tuberculosis, is a potent inducer of the nuclear pregnane-X receptor (PXR) and, thereby, biotransformation enzymes in the liver and intestine ²⁵ ; (iv) pruritogen signal through the endogenous neural opioidergic and serotonergic systems, which can be modulated by opioid antagonists, such as naltrexone, and serotonergic agents, such as sertraline. ²⁰

The extent to which TSBA contribute to itch in ICP and other cholestatic disorders is, in our view, at best minor, ²⁰ as already claimed more than 60 years ago. ²⁶ Also, endogenous µ-opioids appear unlikely to play a relevant role. ²⁷ However, in identifying the pruritogens of ICP (and other cholestatic disorders), links have been established between cholestatic pruritus and the volatile phospholipid lysophosphatidic acid (LPA) and the enzyme autotaxin which forms the LPA.^27,28 Autotaxin has been shown to be increased in ICP compared with HELLP syndrome and gestational pruritus. ²⁹ Serum autotaxin activity is decreased by treatment of severe itch with rifampicin, MARS dialysis or nasobiliary drainage, but not colesevelam, a potent bile acid-binding resin. ²⁸ Persistent serum autotaxin activity has been identified up to a year postpartum of ICP pregnancies, suggesting an endogenous predisposition to cholestatic pruritus. ²⁹ Autotaxin, however, is only available as a research assay at present. Urinary sulphated metabolites of progesterone ³⁰ have also been correlated with ICP and its treatment, ³¹ and, when used in combination with autotaxin, provide increased sensitivity and specificity for the diagnosis of ICP compared with the individual compounds. ³¹ Thus, LPA and sulphated progesterone metabolites are regarded as potential pruritogens in ICP. There is also evidence that serum 5β-pregnane-3α,20α-diol-3-sulphate (PM3S) acts as a Takeda G-protein receptor-5 (TGR5) agonist in humans at concentrations that are reached at the end of a pregnancy with ICP. ³¹ TGR5 has also been suggested to play a part in BA-induced activation of itch neurons. ³² However, the study was performed with non-physiological types and concentrations of BA, whilst intradermal injection of TGR5-ligand in humans was reported not to induce itch perception. ³³ Hence, the potential role of TGR5 in cholestasis-associated itch remains controversial.

Various antipruritic strategies have been proposed for ICP. Ursodeoxycholic acid (UDCA) is the first-line treatment of ICP-associated pruritus. ¹³ It is regarded as safe during pregnancy.^13,34 The physiological taurine conjugate of UDCA, tauroursodeoxycholic acid, has been shown to stimulate impaired vesicular exocytosis and membrane insertion of key transporters into the apical membrane of cholestatic hepatocytes in a Ca⁺⁺/PKCα/PKA-dependent way, thereby enhancing the impaired secretory capacity of hepatocytes in ICP.^35,36 This mechanism might explain the anticholestatic and antipruritic action of UDCA in ICP. Notably, UDCA is reported to be more effective to reduce pruritus in ICP than the anion-exchange resin cholestyramine. ³⁷

Rifampicin is regarded as a safe second-line treatment of pruritus in ICP during pregnancy. As mentioned, rifampicin impairs expression of the LPA-forming enzyme autotaxin by PXR-dependent mechanisms. ²⁸ Rifampicin is an efficient first-line antipruritic agent in non-pregnancy cholestatic disorders, such as primary and secondary sclerosing cholangitis.^38,39 There is a low incidence of adverse effects, and hepatotoxicity is usually short-lived and reversible on drug cessation. ⁴⁰ Acute kidney injury is rare (incidence 0.1%) and usually antibody-mediated. ⁴¹ Rifampicin may also rarely be associated with disseminated intravascular coagulation (DIC) due to an immuno-allergic reaction, typically in individuals with previous exposure to rifampicin, and with onset within days of re-exposure. ⁴² This becomes important when considering the treatment for ICP recurring in subsequent pregnancies. In addition, rifampicin may rarely cause anaphylaxis, serum sickness, autoimmune haemolytic anaemia, thrombocytopenia and hypotension due to hypersensitivity. ⁴³

For bezafibrate, the pan-peroxisome proliferator-activated receptor (PPAR) agonist and also a first-line antipruritic agent in non-pregnancy cholestatic disorders, such as primary and secondary sclerosing cholangitis,^38,39 there are no efficacy and safety data as yet to support its use in pregnancy. ¹³

Similarly, ileal bile acid transporter (IBAT)/apical sodium-dependent bile acid transporter (ASBT) inhibitors have been shown to be effective in the inhibition of bile acid uptake in children with cholestatic pruritus (e.g. progressive familial intrahepatic cholestasis [PFIC]). ⁴⁴ One randomised trial of their use in pregnancy was abandoned because of poor recruitment. However, the emerging data from trials of hereditary and acquired cholestasis, and the fact that IBAT/ASBT inhibitors are not systemically absorbed, suggest that future studies may demonstrate that they have a role in ICP treatment for a subgroup of affected women.

Dysregulated BA synthesis might be a key pathogenic finding in ICP: BAs are synthesized from cholesterol via cholesterol 7α-hydroxylase (CYP7A1) and excreted in the bile into the duodenum. In the distal ileum, ∼99% of BAs are actively taken up by the IBAT/ASBT and transported back to the liver via the portal vein (enterohepatic circulation). In the terminal ileal enterocytes, BA binds to the nuclear farnesoid X receptor (FXR), which induces the excretion of fibroblast growth factor 19 (FGF19) that, in a negative feedback loop, reduces hepatic BA synthesis by downregulating CYP7A1. In normal murine and human pregnancy there is reduced enterocyte FXR signalling, likely related to microbial alterations of the luminal BA, resulting in lower FGF19 secretion and enhanced BA synthesis. ⁴⁵ Furthermore, ingestion of UDCA enhances FGF19 secretion in women with ICP, likely by improving hepatocellular cholestasis and biliary BA secretion, thereby contributing to normalisation of BA homeostasis. ⁴⁶ In severe cholestasis, FXR may also regulate BA synthesis in the liver via a small heterodimer partner (SHP). FXR further regulates hepatic BA uptake via sodium (Na⁺)-taurocholate co-transporting polypeptide (NTCP) and hepatic BA excretion via the bile salt export pump (BSEP). BA synthesis can reliably be estimated by measuring the BA precursor 7α-hydroxy-cholestene-4-one (C4). It appears that (1) women with ICP differ from unaffected pregnant women in the FXR-dependent regulation of BA synthesis and that (2) some women with ICP may have only a limited response to UDCA. Both conditions would be reflected by normal, or even increased, C4 concentrations, perhaps due to improved secretion of hydrophobic BA despite the increased circulating BA. Work is in progress to assess these possibilities. In particular, a new approach of assessing serial changes in the complex intrahepatic systems is to use a “liquid biopsy” approach to examine hepato-specific extracellular vesicles and their contents. ⁴⁷ Zivile Useckaite discusses the possibilities in her paper below.

The maternal gut microbiome changes significantly throughout pregnancy. BA has complex interactions with the gut microbiome, and there is a dynamic equilibrium between diet–gut microbiome–BA pool size/composition. ⁴⁸ In addition, recent evidence has shown that the maternal gut microbiome plays an important, but ill-defined, role in the aetiology of ICP. The pathogenic role of alterations in the intestinal microbiome in the development of cholestatic liver disease was recently demonstrated in two elegant studies using different mouse models of cholangitis. Dysbiosis was found to cause intestinal barrier dysfunction, resulting in T helper 17 cell/interleukin (IL)17-mediated hepatobiliary injury. ⁴⁹ IL17 may be a possible mediator of severe ICP. ⁵⁰

The gut microbiota are a major regulator of BA through the conjugation of primary BA, thereby affecting their re-uptake. They undergo large changes during pregnancy, becoming less diverse and likely contributing to maternal weight gain and changes in metabolism. ⁵¹ The gut microbiome in ICP women differs from that in unaffected pregnancy and UDCA treatment alters its composition in a UK population, as mentioned above. ⁵² In addition, the gut microbiome provides an important contribution to the maintenance of normal carbohydrate tolerance, and, in the largest population study to date examining the relationship between ICP and gestational diabetes mellitus (GDM), there was an odds ratio (OR) of 2.8 for GDM in Swedish women with ICP, ⁵³ whilst in an Australian population, an OR of 3.5 for GDM was seen in women with severe ICP. ⁵⁴ A US study demonstrated that the association between GDM and ICP is only seen after TSBAs are increased, indicating that hypercholanaemia may contribute. ⁵⁵ Marloes Dekker and Caroline Ovadia discuss the importance of these observations in their paper below.

TSBA, as well as the incidence of pruritus, may also increase in women with other hepatic disorders, including viral and autoimmune liver disease, as well as with biliary disease, so that a careful history, examination and perhaps imaging are necessary to consider the possibility of such in symptomatic pregnant women with an atypical history for ICP. ICP has been reported to be more common and may occur earlier in women with previous hepatitis C, ⁵⁶ although this may have been misclassified as ICP, and has been rather a gestational cholestasis secondary to chronic hepatitis C and its liver damage. The presence of active hepatitis, whether viral or autoimmune, excludes the diagnosis of ICP, and there are insufficient data to enable the application of similar risk profiles to ICP from the associated increased TSBA.

The small but significant increase in risk of morbidities associated with ICP has probably led to an increase in obstetric interventions to prevent such, especially in litigation-conscious societies, without good evidence to underwrite them. Research is urgently required to analyse the associated risks and costs of interventions despite the relatively infrequent occurrence of ICP, which makes the designing and funding of such trials difficult.

A number of treatment trials in ICP have explored various treatment options. ⁵⁷ The largest randomised trial to date has been PITCHES, a double-blind, multicentre, randomised placebo-controlled trial at 33 hospital maternity units in England and Wales of UDCA in all-comers with ICP, with a gestational age (GA) between 20 and 40⁶ weeks, with a singleton or twin pregnancy and no known lethal fetal anomaly. The primary outcome was a composite of perinatal death (in utero fetal death after randomisation or known neonatal death up to 7 days after birth), preterm delivery (<37 weeks), or neonatal unit admission for at least 4 h (from birth until hospital discharge). A total of 605 women were randomly allocated to receive UDCA (n = 305) or placebo (n = 300). The primary outcome analysis included 304 women and 322 infants in the UDCA group and 300 women and 318 infants in the placebo group, and the primary composite outcome occurred in 74 (23%) of 322 infants in the UDCA group and 85 (27%) of 318 infants in the placebo group (adjusted risk ratio 0·85 (95% CI 0·62–1·15)). None of the eight serious adverse events were regarded as being related to treatment, and there were no perinatal deaths. The authors concluded that treatment with UDCA does not reduce adverse perinatal outcomes in women with ICP, and therefore, its routine use for this condition should be reconsidered. ⁵⁸ It should be noted, however, that approximately 75% of participants had TSBA <40 µmol/L at randomisation. In the recent Cochrane review, when compared with placebo, UDCA resulted in a small improvement in pruritus score measured on a 100 mm visual analogue scale (VAS) (mean difference (MD) −7.64 points, 95% CI −9.69 to −5.60 points; 2 trials, 715 women; GRADE moderate certainty), where a score of 0 indicates no itch and a score of 100 indicates severe itching. ⁵⁷ There was insufficient evidence to indicate if S-adenosylmethionine (SAMe), guar gum, activated charcoal, dexamethasone, cholestyramine and various herbal treatments, either alone or in combination, are effective in treating women with ICP.

The available data have recently been subjected to a meta-analysis, as discussed by Nadejda Capatina and Caroline Ovadia later in this publication. In summary, in women with ICP and serum BA concentrations ≥40 µmol/L, UDCA should be offered as a treatment, as an individual participant meta-analysis demonstrated that it reduced the risk of spontaneous preterm birth ⁵⁹ and it may also be protective against stillbirth.

A current major research focus in ICP is the TURRIFIC trial in women with severe ICP in which participants are randomised to UDCA or rifampicin with the primary outcome of pruritus. ⁶⁰ Rifampicin is a commonly used antibiotic in the treatment of tuberculosis (including in pregnant women), and, as discussed above, it is also used for the treatment of cholestatic pruritus outside of pregnancy, ⁶¹ where it acts through its function as a pregnane-X receptor agonist, increasing BA metabolism and possible suppression of BA synthesis. ²⁵ In addition to the potential side effects discussed above, rifampicin may also cause impairment of vitamin K metabolism, increasing INR: neonates of women being treated with rifampicin should be administered parenteral vitamin K at birth because of potential neonatal haemorrhage. ⁶²

Interestingly in the light of the increased incidence of GDM in ICP, there are laboratory studies identifying an impact of metformin, the primary treatment for type 2 diabetes and a first-line treatment for GDM in some countries on BA metabolism in laboratory animals. ⁶³ Cases have also been reported ⁶⁴ and a clinical trial is in progress.

The risk of recurrence of ICP in a subsequent pregnancy is reported to range from 40% ⁵⁴ to 70%. ⁶⁵ The longer-term risks have been explored in large Swedish population registry studies. Whilst pruritus usually resolves quickly postpartum, affected women have an increased risk of hepatobiliary disease in later life, including liver and biliary tree cancer (Hazard ratio [HR] 3.61, 95% CI 1.68–7.77 and 2.62, 95% CI 1.26–5.46, respectively), ⁶⁶ although again, this may perhaps have been overlooked hepatobiliary disease during pregnancy, misclassified as ICP. In the same study, ICP was also associated with later immune-mediated diseases (HR 1.28, 95% CI 1.19–1.38), specifically diabetes mellitus (HR 1.47, 95% CI 1.26–1.72), thyroid disease (HR 1.30, 95% CI 1.14–1.47), psoriasis (HR 1.27, 95% CI 1.07–1.51), inflammatory polyarthropathies (HR 1.32, 95% CI 1.11–1.58) and Crohn's disease (HR 1.55, 95% CI 1.14–2.10). Women who have had severe ICP are at risk of chronic liver disease, which may in part be due to their being heterozygous carriers of transporter mutations, and should have long-term follow-up.

For the offspring, whilst the hazard of preterm birth and of stillbirth is dependent on the peak titre of TSBA, once delivered, the baby is removed from the toxic maternal environment, and good outcomes can be expected. Few studies have investigated the long-term outcome for children of pregnancies complicated by ICP.

A recent study of 6-month-old infants (n = 540) from a birth cohort (2018–2020) in Southwest China (n = 27,058) identified an increased incidence of infantile food allergy in 27/188 (14.4%) children of mothers identified with severe ICP compared with 39/430 (9.1%) children of mothers with no or only mild ICP (χ²: p = 0.04). ⁶⁷ Over 90% of the women were delivered by caesarean section (mostly electively), and no data were provided as to peripartum antibiotic administration.

In an earlier study (n = 45) of 16-year-olds (27 male and 18 female) born to mothers who experienced ICP, who had no other known metabolic disorder (including diabetes and pre-eclampsia) and with a similar maternal body mass index (BMI) in pregnancy, birthweights, placental weights and GA at birth, males had increased BMI and fasting insulin compared with children from uncomplicated pregnancies. Females had increased waist and hip girth, as well as decreased fasting high-density lipoprotein (HDL) cholesterol, relative to females from uncomplicated pregnancies. ⁶⁸ Further studies are needed to explore the potential impact of treatment on such children following maternal treatment during pregnancy.

A very recent study from Sweden has examined the long-term neurodevelopmental outcomes for a cohort of over 10,000 children born to women diagnosed with ICP from a birth cohort of over 2.3 million singleton pregnancies in 1.2 million mothers between 1987 and 2010 and followed up to the end of 2016. ⁶⁹ A total of 143,746 (6.05%) of children were diagnosed with a neurodevelopmental condition: attention deficit/hyperactivity disorder (ADHD), autism or non-specific. After adjusting for child's sex, birth year, birth month, maternal age, highest parental education level, maternal birth country, birth order and maternal psychiatric history, ICP was associated with increased OR (95% CI) of the offspring having a neurodevelopmental condition (1.22: 1.13–1.31), particularly amongst those exposed to early-onset ICP (2.38: 1.71–3.30) as compared with ICP diagnosed after reaching the term (37 weeks’) (1.08: 0.97–1.20). Importantly, when adjustment was performed for GA at birth in this study (shown in their Table 3) as a secondary analysis requested by the reviewers, inclusion of GA at birth attenuated this association – although non-specific neurodevelopmental disorders and ADHD were more common with maternal ICP, the associations with autism and intellectual disability were not significant. The value of the adjustment GA at birth supports the prolongation of mild–moderate ICP rather than early delivery (as practised in many centres). A strength of the study was that within-family comparisons of full maternal cousins (where 12.5% might, on average, have a shared genotype) were possible, yielding an OR of 2.99 (1.48–6.04), and comparisons of full siblings (where, on average, 50% might be expected to share a genotype) showed an OR of 1.92 (0.92–4.02) though the latter was less precise. Full siblings were also considered more likely, on average, to share environmental factors than cousins. However, there were no data available relating to treatment (UDCA was not in common use at that period) nor to the severity of the BA disturbance. Further studies are required to follow up these important data.

In summary, considerable progress has been made to explain the aetiology of ICP and of the adverse pregnancy outcomes associated with high maternal TSBA. The reported thresholds for non-fasting TSBA associated with risk of stillbirth and spontaneous preterm birth are now cited in most international guidelines. They can be used to identify pregnancies at risk of these adverse outcomes to decide on appropriate interventions and to give reassurance to women with lower concentrations of TSBA. There are also data to support the use of UDCA to protect against the risk of spontaneous preterm birth. Ongoing research will establish the role of additional drugs, including rifampicin, metformin and IBAT/ASBT inhibitors, for management of maternal symptoms, hepatic impairment and reduction of fetal risk. For affected women, there is evidence that a previous history of ICP may be associated with higher rates of subsequent hepatobiliary disease, although this “ICP” may possibly reflect misdiagnosis of a previously unrecognised disorder: if there is a suspicion of underlying susceptibility, clinicians caring for women with ICP should screen for associated disorders or for genetic susceptibility and, where appropriate, refer for ongoing hepatology review. Thus, emerging research findings will enable improved outcomes for affected women and their children.