Sex hormones and bile acids in intrahepatic cholestasis of pregnancy


  • Humberto Reyes M.D.

    Corresponding author
    1. Department of Medicine (Eastern Campus) and Institute of Biomedical Sciences, University of Chile School of Medicine, Santiago de Chile
    • Departamento de Medicina Oriente, Facultad de Medicina, Universidad de Chile. Casilla 16038, Santiago de Chile
    Search for more papers by this author
    • fax: (56) (2) 274 1628.

  • See Article on Page 544

  • Potential conflict of interest: Nothing to report.

Fifty years ago, 2 clinical studies from Alvar Svanborg and Leif Thorling, in Sweden, identified the first series of patients with the basic features of a peculiar disease of pregnancy that had yet to be identified under a standard name.1, 2 The first papers were published in journals that were difficult to access, in an era when the Internet and PubMed were not even dreams. Index Medicus gave them an official recognition and then started a chain of perseverant (sometimes inadvertent) research efforts that spread worldwide. Decades later, the topic has gained recognition in the fields of hepatology and obstetrics. It is difficult to trace this “new” disease before 1950 or to define whether it corresponded with a case of recurrent jaundice in late pregnancy (without a mention of pruritus, that currently is considered a sine qua non symptom of the disease) that had been locally reported in 1882 by F. Ahlfeld in Germany. Data collected in Europe, North America, Australia, New Zealand, and Chile delineate a disease with clinical and laboratory characteristics that can, at present, be considered diagnostic requisites (Table 1) (reviewed in Bacq et al.,3 Reyes et al.,4 Lammert et al.,5 and Arrese et al.6). Although some patients may differ slightly (for example, early onset in the second or first trimester), most characteristics should be fulfilled if a proper differential diagnosis is to be made, particularly with an unsuspected underlying chronic liver disease that will become evident sometime after delivery.7–9 For clinical needs, such as for the selection of treatment, the surveillance of the fetal status, and eventually its active management, the diagnosis of intrahepatic cholestasis of pregnancy can certainly be made during the pruritic episode before delivery, but for research purposes it is indispensable to have a follow-up to document a complete clinical and laboratory improvement after delivery.

Table 1. Cardinal Features for the Diagnosis of Intrahepatic Cholestasis of Pregnancy
1. Pruritus “sine materia” appeared in late pregnancy, starting in the palms and soles and then extended to other zones of the body surface.
2. Abnormal serum liver tests:
 • Alanine and/or aspartate aminotransferases 2-fold to 10-fold over normal (in all the patients).
 • Fasting total bile salts >10 μmol/L (in all the patients).
 • Alkaline phosphatases (total) with variable increase.
 • Gamma-glutamyl transferase mildly increased (in 15% of patients).
 • Conjugated hyperbilirubinemia (in 10% of patients, with mild jaundice in a few of them).
3. Pruritus and the biochemical abnormalities persist during the remaining time of pregnancy, with spontaneous fluctuations in severity.
4. Pruritus disappears after delivery and the biochemical abnormalities improve to normal values within 3 weeks.
5. No past history, physical signs, or biochemical, virologic, or autoimmune abnormalities that could reveal other acute or chronic liver diseases.

CA, cholic acid; CDCA, chenodeoxycholic acid; DCA, deoxycholic acid; ICP, intrahepatic cholestasis of pregnancy; UDCA, ursodeoxycholic acid.

The disease has been named “intrahepatic cholestasis of pregnancy” (ICP), or “obstetric cholestasis” in the British Commonwealth. It has been diagnosed in different ethnic groups and geographic locations. The prevalence ranges from 1 case in 1000 to 1 case in 10,000 deliveries in North America, Asia, and Australia. Between 1950 and 1980, a 10-fold to 100-fold higher prevalence had been reported in Chile (particularly in the native southern Araucanian-admixed population) and in Sweden and other Baltic countries. After 1980, the prevalence of ICP has markedly decreased in Chile and currently ranges from 1.5%–4% of all pregnancies. Changes in environmental factors have been proposed for this epidemiological phenomenon, supported also by the higher prevalence of this disease in winter as has been observed in Chile, Finland, and Sweden.6 ICP may recur in 40%–70% of subsequent pregnancies, but no maternal mortality has been attributed to it during pregnancy or after delivery. The only maternal consequences so far recognized are a higher risk of developing cholestatic hepatitis if the woman receives high doses of estrogens or progesterone in nonpregnant periods, a higher prevalence of cholesterol gallstones, and the fact that episodes of ICP tend to coexist with urinary tract infections. In contrast, the babies face a higher risk of fetal distress, with premature deliveries and even stillbirths. This is the main reason to search for therapeutic agents to improve the disease in the mothers: to improve the babies' prognosis. In recent years, oral administration of ursodeoxycholic acid (UDCA) has become the best alternative for both outcomes, mainly in patients in whom the disease starts before week 35 of pregnancy, those with hyperbilirubinemia, or those with high fasting serum total bile acids (>40 μmol/L).10–12

The cause of the disease remains unknown. Its pathogenesis has been connected with female sex hormones, either through abnormalities in synthesis and metabolism generating cholestatic compounds, or through an abnormal response to the physiologic overload of these hormones in pregnancy. A search for an abnormal metabolism of female sex hormones has been a major approach, and the maternal liver has been the focus of most research efforts. But it should be kept in mind that during pregnancy the synthesis and metabolism of female sex hormones is not limited to the maternal liver, ovaries, and adrenals: there is a bidirectional flux to the fetal placental unit and the fetus also synthesizes hormonal compounds. In fact, about half of the progesterone metabolism occurs in extrahepatic sites. This gives rise to a large number of complex molecules whose identification requires sophisticated, expensive, and time-consuming techniques that are currently out of the reach of clinical diagnostic laboratories.

Serum levels and urinary excretion of estrogens and their metabolites do not differ during ICP than in normal pregnancies at the same gestational age.4, 5, 13 However, multiple pregnancies (twins, triplets) that have a higher load of female sex hormones than singletons also have a higher incidence of ICP. Changes in the metabolism of estrogens could lead to retention of conjugated metabolites, mainly precursors of D-ring glucuronides that can cause cholestasis in rodents,14 but these metabolites have not yet been related to cholestasis in humans.

In 1970, Jan and Karin Sjövall reported differences in progesterone metabolism: patients with ICP have higher serum levels of sulfated metabolites of progesterone than patients with normal pregnancies.15 Other investigators found similar changes in new studies. Twenty years later, L.J. Meng in Sjövall's laboratory, by employing gas chromatography and mass spectrometry to separate and measure different bile acids and steroidal compounds in serum and urine, clarified that the synthesis of bile acids appeared to be reduced in patients with ICP, probably due to bile secretory failure and retention of bile acids in hepatocytes, resulting in their increase in serum and urine.16–18 In patients with ICP, fasting serum total bile acid concentration is typically higher than in healthy pregnant women (>10 μmol/L). Primary bile acids are 90% of total bile acids, and cholic acid (CA) becomes the predominant primary bile acid in serum (80%) in contrast with normal pregnant and nonpregnant women, in whom its proportion is almost similar to chenodeoxycholic acid (CDCA). A low serum level of deoxycholic acid (DCA, a secondary bile acid) reflects an impaired enterohepatic circulation with a low dehydroxylation of CA by bacterial enzymes in the intestinal lumen. Treatment with UDCA in those patients decreased the total concentration of bile acids and also changed the proportion of conjugated bile acids (CA:CDCA:DCA) to values similar to those in normal pregnancies.

These studies also revealed that in healthy pregnant women, monosulfated and disulfated progesterone metabolites are increased in serum, and most of them disappear about a week after delivery. But in ICP patients, these metabolites increase 4-fold to 10-fold more whereas glucuronidated metabolites do not change, either by pregnancy or ICP. The most significant and probably specific change was the increased ratio of 3α- to 3β-hydroxysteroid sulfates in serum. Giusti et al. reported that these differences were not observed in pregnant patients with viral hepatitis, thus supporting a specific connection with ICP.19

In ICP patients treated with UDCA (1 g/day for 3 weeks), pruritus was ameliorated, standard serum liver tests indicative of cholestasis improved, and the concentration of progesterone metabolites in blood and their urinary excretion also tended to normalize. Sjövall et al. proposed that “patients with ICP have a selective defect in the excretion of steroid metabolites into bile, affecting only the excretion of sulphated steroids. These changes may be corrected by the administration of UDCA”.4

Several abnormalities in the metabolism of female sex hormones may coexist and interact; therefore, a role of estrogen metabolites is not excluded by these results. Furthermore, during enterohepatic circulation both progesterone and estrogen metabolites may be transformed into cholestatic compounds. This could be exaggerated in patients with ICP in whom an increased intestinal permeability (“leaky gut”) has been detected during and after pregnancy.20

An experimental study by Vallejo et al. showed that a glucuronidate estrogen metabolite (estradiol 17β-D-glucuronide) and a sulfated progesterone metabolite (5α-pregnan-3α-ol-20-one sulfate) are both secreted into bile, and they reduce bile flow and bile acid output in isolated perfused rat liver. These effects were not obtained by infusing the parent sulfated progesterone. Both metabolites also inhibited the bile salt export pump and impaired mitochondrial activity and bile acid efflux in frog oocytes, thus favoring intracellular toxicity of accumulated bile acids.21 A cholestatic effect of these metabolites in humans has not yet been reported.

The recent study by Glantz et al. adds valuable information on the effects of UDCA in steroid hormones and their metabolites in patients with ICP.22 In a study of 40 patients, they showed at baseline a typical increase in serum bile acids with >50% CA. The pattern of steroid hormone metabolites in urine was dominated by disulfates of pregnanediols. In 13 patients randomized to receive UDCA (1 g/day for 3 weeks), they observed a significant fall in fasting total serum bile acids, starting at 1 week, with an enrichment with isoUDCA (a main isomer of UDCA) and a significant reduction of CA and CDCA; urinary excretion of steroid hormone metabolites also diminished significantly. Pruritus was ameliorated simultaneous to these metabolic changes. An important contribution of this study is that the improvement in pruritus correlated with the reduction in urinary excretion of pregnanediol disulfate, whereas pruritus did not correlate with changes in serum total bile acids or individual CA or CDCA levels. The beneficial effects obtained by UDCA were significantly better than the minimal changes obtained in 14 patients that received oral dexamethasone (12 mg/day for 1 week). The authors also interpreted the pattern of progesterone metabolites in ICP as “most likely due to impaired hepatobiliary secretion of specific isomeric metabolites, which is improved by oral administration of UDCA”. Minor differences with the results reported by Meng et al. could be due to different methods used to identify isomers of progesterone metabolites.16–18 Besides, Glantz et al. measured progesterone metabolites only in urine whereas Meng and coworkers did it in serum and urine samples obtained simultaneously. The studies by Glantz and coworkers further stimulate research in membrane receptors and transporters of steroid hormones and their metabolites.

Hormonal and other metabolic factors that may be involved in the pathogenesis of ICP are becoming more precisely known, but their intimate action and the mechanisms of the beneficial effect of UDCA are still hypothetical. Stimulated by findings in other different familial cholestatic disorders (mainly seen in childhood) some mutations have been proposed as causally related to ICP.23–27 Nevertheless, the number of patients in whom they have been found is quite small. It is also disturbing that some patients in whom putative mutations have been described do not fulfill the diagnostic requisites of a classical case of ICP, and therefore, they may have different diseases unveiled by pregnancy. If a mutation were causally related to ICP, it should be found in all or in a large proportion of patients with an unequivocal phenotype. Homozygous patients would be expected to have a more severe and recurrent disease, with familial clustering, than the heterozygous individuals; the mutation should be absent or exceptional in proper control patients (multiparous women with several pregnancies not affected by ICP) and it should be present in different populations but with a higher prevalence in ethnic groups where ICP is also more prevalent.

Clinical and epidemiological characteristics of ICP support the proposal that its pathogenesis is multifactorial: a genetic abnormality determines a susceptibility to develop the disease while, simultaneously, environmental factors may influence the onset of the disease and modulate its severity and its prevalence worldwide.


The author is grateful to Professor Emeritus Jan Sjövall, who pioneered research on steroid metabolism in pregnant women and in patients with ICP, leading his group and stimulating others to follow a sound line of research. This editorial is dedicated, on his 81st birthday, to Irwin Arias, Founding Editor of HEPATOLOGY, a mentor and friend whose endless scientific and intellectual curiosity is a dreamt example to numerous former Fellows in the international community.