Intrahepatic cholestasis of pregnancy: risk factors for severe disease

Abstract

Introduction

Intrahepatic cholestasis of pregnancy (ICP) is a pregnancy-associated hepatic disorder characterized by pruritus in the setting of elevated serum bile acids (BA). Risk factors for the disease include preexisting hepatobiliary disease, personal or family history of ICP, and advanced maternal age. Recent data suggests that patients with severe ICP (BA ≥100 μmol/L) have a higher risk of adverse pregnancy outcomes including stillbirth.

Material and methods

This was a retrospective cohort study of patients diagnosed with ICP between 2012 and 2019 at a tertiary referral center. ICP was defined as symptomatic pruritus combined with serum BA >10 μmol/L. Maternal characteristics and outcomes were abstracted from electronic medical records. Baseline characteristics were compared between patients with mild (BA <40 μmol/L), moderate (BA 40-99 μmol/L) and severe (BA ≥100 μmol/L) ICP. Obstetrics and neonatal outcomes for patients in each category were then analyzed. Shapiro-Wilk test was used to test for normality for continuous variables, and ANOVA, Kruskal-Wallis, Chi-squared or Fisher’s exact tests were used as appropriate. A p-value <.05 was considered statistically significant.

Results

438 patients were included in the analysis. Individuals with pregestational diabetes (p < .01), history of ICP (p < .01), prior cholecystectomy (p < .01), and tobacco use (p < .05) were more likely to have severe disease. When compared to individuals with moderate and mild disease, individuals with severe disease were more likely to be diagnosed earlier (29w1d vs 34w1d vs 34w1d, p < .05), have gestational diabetes (50% vs 6% vs 13%, p < .01), hypertensive disorders of pregnancy (42% vs 10% vs 15%, p = .02), and abnormal aspartate aminotransferase (91% vs 65% vs 27%, p < .01) and alanine aminotransferase levels (91% vs 60% vs 26%, p < .01). There were no differences in preterm labor, meconium-stained amniotic fluid, or neonatal respiratory distress syndrome and no stillbirths in this cohort.

Conclusions

In patients with ICP, those with pregestational diabetes, history of ICP, prior cholecystectomy, and tobacco use are more likely to develop severe disease. Given the adverse outcomes associated with severe disease, serial BA measurements to monitor for development of severe disease may be warranted in this population.

Keywords:

• Cholestasis
• intrahepatic
• pregnancy
• liver
• risk

Introduction

Intrahepatic cholestasis of pregnancy (ICP) is a pregnancy-associated hepatic disorder characterized by pruritus in the setting of elevated serum bile acids (BA) [1]. ICP typically presents in the third trimester and resolves spontaneously after delivery. It is the most common liver disease unique to pregnancy and affects 0.32% to 5.6% of pregnancies in the United States [1,2].

While maternal outcomes are overall favorable, ICP has been associated with adverse neonatal outcomes such as preterm birth, meconium-stained amniotic fluid, neonatal respiratory distress syndrome, and stillbirth [3,4]. BA levels > 40 μmol/L have been associated with an increased risk of adverse outcomes [4,5]. However, a recent systematic review suggests that the risk of stillbirth only significantly increases in the setting of severe cholestasis, with BA levels ≥ 100 μmol/L [3]. Both the Society for Maternal Fetal Medicine and the American College of Obstetricians and Gynecologists therefore recommend delivery at 36 weeks’ gestation for patients with BA ≥ 100 μmol/L, and between 36 and 39 weeks’ gestation for patients with BA <100 μmol/L [6].

The primary objective of this study was to identify risk factors for severe disease in patients with a diagnosis of ICP. The secondary objective was to compare obstetric and neonatal outcomes based on severity of ICP.

Material and methods

This study was a retrospective cohort study that included all patients diagnosed with ICP during pregnancy between 2012 and 2019 who delivered at a tertiary referral center. This study was approved by the Ohio State University Institutional Review Board (Study ID 2019H0416).

All patients with a diagnosis of ICP during a pregnancy were identified through the Ohio State University Information Warehouse using ICD-9 and ICD-10 codes for liver and biliary tract disorders in pregnancy (646.7* and O26.61*, respectively) and had BA level testing during pregnancy encounter. Patient charts were then individually reviewed to confirm the diagnosis of ICP with BA >10 μmol/L and symptoms of pruritus. In patients with more than one pregnancy complicated by ICP during timeframe, each pregnancy was treated as a separate event.

Maternal and neonatal baseline characteristics, associated disease outcomes, obstetric and neonatal outcomes were abstracted from the medical record. ICP was defined as serum BA >10 μmol/L with symptomatic pruritus. The cohort was stratified into mild ICP (BA 10–39 μmol/L), moderate ICP (BA 40–99 μmol/L), and severe ICP (BA ≥100 μmol/L). In women with repeated BA measurements, the highest value was used for categorization. Abnormal liver function tests were set by institutional laboratory reference range of >40 U/L for abnormal alanine aminotransferase (AST) and > 48 U/L for aspartate aminotransferase (ALT).

Statistical analysis was performed using IBM SPSS v26 (2019, Aramonk, New York). Categorical variables were compared using Chi-Squared or Fisher’s exact tests as applicable. Continuous variables were tested for normality using the Shapiro-Wilk test and compared using ANOVA or Kruskal-Wallis test as applicable. A p-value < .05 was considered statistically significant.

Assuming that among patients with ICP, 20% of those with mild, 40% with moderate and 60% with severe disease would have a known risk factor for severe disease, a sample size of 206 women in a 6:3:1 ratio would identify differences with 80% power and an alpha of <0.05 [4].

Results

1,283 patients were identified with a possible diagnosis of ICP during a pregnancy episode through the electronic medical record query. Of those, 438 met criteria for diagnosis of ICP on chart review and were included in the data analysis: 377 (86%) had mild, 49 (11%) moderate, and 12 (3%) severe disease. There were a total of 39,654 deliveries during the study period, and the incidence of ICP was therefore 1.1% in our population. There were 21 patients who did not initially meet the diagnosis of ICP by bile acid levels, but had elevated levels on serial collections. There were 14 patients with a progression in severity of ICP with serial bile acid levels: 12 progressed from mild to moderate ICP, and 2 progressed from mild to severe ICP. 3 of those who progressed from mild to moderate and 1 who progressed from mild to severe were not treated with ursodeoxycholic acid.

There were no differences in age, race/ethnicity, Hepatitis C infection, body mass index at delivery, or multifetal gestations between the groups (Table 1). No patients in this cohort had hepatitis B or other liver diseases. Patients with severe ICP were more likely to have pregestational diabetes (p < .01), history of cholecystectomy (p < .01), history of ICP in prior pregnancy (p < .01), and tobacco use (p < .04). During the pregnancy, those who developed severe disease were more likely to have gestational diabetes mellitus (p < .01), hypertensive disorders of pregnancy (p < .01), and elevated ALT or AST (p < .01) (Table 2). They were also more likely to use ursodeoxycholic acid after diagnosis (92% vs 80% vs 67%, p = .05). Pregnancies complicated by severe ICP were more likely to be diagnosed with ICP at an earlier gestational age (29w1d vs 34w1d vs 34w1d, p < .05) and have a longer latency from diagnosis to delivery (5w3d vs 2w0d vs 2w3d, p < .01).

Table 1. Patient baseline characteristics.

Bile acid level (μmol/L)	<40	40–99	≥100	p-value
n	377	49	12
Age (years)	30 ± 5	29 ± 5	29 ± 5	.47
Race/ethnicity
African American	61 (16)	10 (20)	0	.45
Caucasian	228 (61)	26 (53)	9 (75)
Hispanic	37 (10)	6 (12)	3 (25)
Asian	41 (11)	6 (12)	0
Other / Unknown	10 (3)	1 (2)	0
BMI at delivery	31.4 ± 6.3	31.4 ± 8	31.2 ± 6.9	.99
Tobacco use	53 (14)	3 (6)	4 (33)	.04
Pregestational diabetes	29 (8)	3 (6)	4 (33)	<.01
Hepatitis C	29 (8)	8 (16)	1 (8)	.13
History of cholecystectomy	40 (11)	5 (10)	6 (50)	<.01
History of ICP	39 (10)	9 (18)	6 (50)	<.01
Prior preterm birth	71 (19)	11 (22)	2 (17)	.81
Multiple pregnancy				.84
Twins	21 (6)	4 (8)	0
Triplets	1 (0.3)	0	0

Data presented as n (%) or mean ± SD.

ICP: intrahepatic cholestasis of pregnancy; BMI: body mass index.

Table 2. Obstetric outcomes of patients with intrahepatic cholestasis of pregnancy stratified by bile acid level.

Bile acid level (μmol/L)	<40	40–99	≥100	p-value
N	377	49	12
Gestational age at diagnosis	34w1d [31w1d–36w3d]	34w1d [31w3d–36w1d]	29w1d [26w6d–33w2d]	.04
Time from diagnosis to delivery	2w3d [1w0d–4w0d]	2w0d [0w5d–4w0d]	5w3d [2w3d–8w6d]	.09
Mode of Delivery				.45
Vaginal	260 (69)	38 (78)	8 (67)
Cesarean	117 (31)	11 (22)	4 (33)
Cesarean section indication				.95
Dystocia	19 (15)	2 (17)	0
Non-reassuring fetal status	26 (20)	4 (33)	2 (50)
Abnormal presentation	18 (14)	0	0
Abruption	1 (1)	0	0
Fetal anomaly	1 (1)	0	0
Elective	3 (2)	0	0
Previous cesarean	43 (33)	6 (50)	2 (50)
Prior shoulder dystocia	2 (2)	0	0
Multiple pregnancy	15 (12)	0	0
Maternal medical condition	2 (2)	0	0
Estimated blood loss	350 [250 − 600]	350 [250 − 500]	300 [200 − 500]	.82
Blood transfusion	13 (3)	1 (2)	1 (8)	.56
Gestational diabetes mellitus	49 (13)	3 (6)	6 (50)	<.01
Hypertensive disorders of pregnancy	55 (15)	5 (10)	5 (42)	.02
Fetal growth restriction	12 (3)	1 (2)	1 (8)	.54
PPROM	12 (3)	2 (4)	1 (8)	.61
Preterm labor	15 (4)	1 (2)	1 (8)	.58
Meconium-stained amniotic fluid	16 (4)	5 (10)	1 (8)	.17
Abnormal AST (>40 U/L)	95 (27)	31 (65)	10 (91)	<.01
Abnormal ALT (>48 U/L)	93 (26)	29 (60)	10 (91)	<.01
Gestational age at delivery	37w1d [36w5d–37w4d]	36w3d [36w1d–37w0d]	36w3d [36w0d–37w2d]	<.01
Ursodeoxycholic acid use	252 (67)	39 (80)	11 (92)	.05

Data presented as n (%) or median [interquartile range].

ALT: alanine aminotransferase; AST: aspartate transaminase; PPROM: preterm premature rupture of membranes.

There was no difference in mode of delivery between the groups (p = .45). There was no difference in the incidence of meconium-stained amniotic fluid, preterm delivery (spontaneous or iatrogenic), transient tachypnea of the newborn or NICU admission. Though not statistically significant (p = .38), the incidence of respiratory distress syndrome in the severe ICP group (25%) was twice that in the mild (12%) and moderate (13%) ICP groups. Patients with severe ICP were more likely to have an infant be intubated (17% vs 2.3%, p < .01).

Although there were no stillbirths in the cohort, there were two neonatal deaths. The first was in a patient who had mild disease, whose fetus had known major congenital anomalies that contributed to the neonatal death. The second was in a patient who had severe disease whose neonate was born at 24 weeks gestation due to preterm premature rupture of membranes and preterm delivery and expired due to complications of prematurity.

Discussion

This was a retrospective analysis of patients with ICP to identify risk factors for severe disease. Our main finding demonstrated that history of cholecystectomy, pregestational diabetes mellitus, history of ICP in prior pregnancy, and tobacco use are associated with an increased risk of severe ICP.

It is widely accepted that cholecystectomy may have adverse effects on bile acid metabolism. One suggested mechanism is removal of the gallbladder results in continuous enterohepatic cycling of the BA pool, thereby providing negative feedback to the hepatic synthesis of BA and prolonging primary BA exposure in the small intestine leading to conversion of cholic acid to deoxycholic acid [7]. Another hypothesis is that gallbladder cholangiocytes secrete human fibroblast growth factor 19 (FGF19), which is a protein involved in bile acid synthesis, and both expression of FGF19 and secretion is regulated by BA [8]. Cholecystectomy results in a reduction in FGF19 and doubling of C4, which is a bile acid synthesis marker [8]. This alteration in bile acid metabolism may result in women who have undergone prior cholecystectomy being more likely to development severe ICP when combined with other risk factors.

The increased rates of severe ICP in patients with gestational and pregestational diabetes align with current literature regarding BA in patients with diabetes outside of pregnancy. Studies have shown an increased in fasting and post prandial BA in patients with insulin resistance [9]. Bile acids appear to play a crucial role in gluconeogenesis, insulin secretion, insulin sensitivity, and glycogen synthesis, and dysregulation in bile acid levels may be a contributing factor in the pathogenesis of diabetes [10–13]. Additionally, ICP has been associated with increase in maternal glucose concentrations both in the ambulatory setting via continuous glucose monitoring and after glucose tolerance tests [14]. This association has downstream fetal effects, as infant birthweights were found to be significantly higher in ICP pregnancies [14]. The incidence of gestational diabetes in patients with ICP in our study was 13.2%, which is lower than the 17.1% incidence in the multicenter study by Liu, et. al. however the incidence of gestational diabetes in the severe ICP population was much higher at 50%, suggesting an association between gestational diabetes and severe ICP [15].

Minimal and conflicting data exist regarding the effect of tobacco use on gallbladder disease [16]. Tobacco smoking has been shown to increase post-prandial bile salt excretion offering one physiologic mechanism for the association of severe ICP and concurrent tobacco use (33%) in our study [17]. Additionally, patients with severe disease had a significantly increased risk of hypertensive disorders of pregnancy. This association has previously been described in the literature [18,19]. Current research suggests that BAs can cause an imbalance in soluble fms-like tyrosine kinase 1 (sFlt-1) and soluble endoglin (sEng) through a combination of placental vasoconstriction and oxidative damage [20,21]. These molecules have consistently been found to be associated with preeclampsia [22,23]. Additionally, recent data suggests a common genetic pathway that contributes to the development of both preeclampsia and ICP [24].

Although a small number of patients in our cohort had progressive disease, many did not have serial bile acid measurements as this is not currently recommended by professional societies [6], Of note, there was a subset of patients who initially did not meet criteria for ICP and were diagnosed on serial bile acid testing. In high risk patients with persistent symptoms and normal bile acids, serial bile acid collection may be warranted. Patients with progressive disease were managed according to provider preference with regard to serial bile acids, antenatal testing, ursodeoxycholic acid dosing, and timing of delivery. Patients with mild/moderate disease were less likely to be on ursodeoxycholic and had a shorter latency from diagnosis to delivery, which could be in part due to delivery recommendation at the time of diagnosis and therefore deference of medical management, though this was not directly studied. Amongst patients with mild disease who did not receive ursodeoxycholic acid, only 0.024% progressed to moderate disease and 0.008% progressed to severe disease suggesting other factors are more likely to influence disease progression. This study suggests that certain subgroups may be at higher risk for severe disease, and serial BA measurements may help inform obstetrical management.

Previous studies have shown a higher risk of ICP in patients with hepatitis C. However, there were no differences noted with regard to severity of ICP in patients with Hepatitis C [24]. This study replicates those findings, where Hepatitis C was not noted to be a risk factor for severe disease.

This study is important as many previous studies have looked at factors associated with ICP, but to our knowledge no studies have broken down these associations by severity of ICP. Strengths of this study include its relatively large size and representative population. The study is limited by the inherent bias of retrospective study design. The study may also be limited by selection bias, as some women who carried the diagnosis of ICP during pregnancy at our institution may have delivered at a different institution. Another limitation was the inconsistency in obtaining serial bile acids measurements, as there could be patients who progressed with regard to BA severity that were not recognized. In patients with abnormal liver function tests, the charts were reviewed to verify that these elevations were attributed to ICP rather than alternative diagnoses such as preeclampsia, however etiology for transaminitis was unable to be determined for two patients. The study was also limited by the small number of patients in the severe ICP subgroup, increasing the possibility of a type II error.

Conclusion

In summary, our findings suggest an association between prior cholecystectomy, pregestational diabetes, tobacco use, ICP in prior pregnancy, and risk of severe ICP. Patients with these risk factors may be at higher risk for severe disease, and serial BA measurements may be warranted to help inform obstetrical management.

Abbreviations
BA	=	bile acids
ICP	=	intrahepatic cholestasis of pregnancy

Acknowledgments

SM: study conception and design, acquisition of data, drafting of manuscript, critical revisions.

ES: study conception and design, acquisition of data, drafting of manuscript, critical revisions.

PC: acquisition of data, critical revisions.

CS: analysis and interpretation of data, critical revisions.

KR: study conception and design, analysis and interpretation of data, critical revisions.

MM: study conception and design, acquisition of data, analysis and interpretation of data, drafting of manuscript, critical revisions.

Disclosure statement

No potential conflict of interest was reported by the author(s).