Correlation of bile acids and aspartate-aminotransferase with outcomes in cholestasis of pregnancy

1 Introduction

Intrahepatic cholestasis of pregnancy (ICP), the most common pregnancy-associated hepatic condition, is diagnosed by hypercholanemia (an elevation in fasting serum bile acid (BA) level ≥10μmol/L), nocturnal pruritus affecting the palms and soles, and the absence of a rash. The condition usually presents in the second or third trimester, and tends to be more prevalent among Hispanic and Scandinavian women. The diagnosis requires ruling out other dermatologic pathologies in pregnancy such as atopic eczema, pruritic urticarial papules and plaques of pregnancy (PUPPP), pemphigoid gestationis (PG), and hepatic diseases such as viral hepatitis, biliary obstruction, cholelithiasis, preeclampsia, and acute fatty liver of pregnancy. The disease process is limited to the gestation interval and typically resolves within 4 weeks postpartum [1]. The primary and most damaging complication associated with ICP is intrauterine fetal demise (IUFD). It has been shown that the lowest risks of expectant management with respect to IUFD occur at 35 weeks and begin to rise beyond 36 weeks [2] with an ICP-associated stillbirth rate of 1.2% commonly cited after 37 weeks of gestation [3]. The original studies, that prompted widespread recommendations for active management of pregnancies complicated by ICP and that advocated induction of labor at 36 weeks to reduce the risk for IUFD, reported a total of four cases of IUFD among 331 cases of ICP [3]. Surprisingly, subsequent meta-analyses have suggested that the rate of IUFD among pregnancies complicated by ICP may not be significantly different from the rate of IUFD among the general population [3]. The timing of delivery must balance potential neonatal complications of prematurity with IUFD and other perinatal complications. A review of the literature demonstrates a correlation between elevated BA and adverse perinatal outcomes including, IUFD, spontaneous preterm labor, non-reassuring fetal heart tracing, and meconium-stained amniotic fluid (MSAF) [1 –13]. A recent retrospective study found that induction of labor prior to 37 weeks gestation was justified in patients with BA levels ≥40μmol/L [14], which is the current BA level that differentiates low risk from high risk cholestasis in pregnant patients.

Traditionally, ICP has not been associated with acute adverse maternal outcomes, and while it is a hepatic disease, ICP generally does not affect coagulation or increase the risk of postpartum hemorrhage [15]. However, new studies have demonstrated a higher rate of gestational diabetes among women with ICP [16]. While research is currently underway to define the etiopathogenesis of ICP, it appears that a combination of genetic, hormonal, and environmental factors is involved in the disease process [1 , 17]. Studies have demonstrated a correlation between the disease state and maternal blood BA and liver transaminase levels. Furthermore, studies have also shown a direct correlation between the severity of the elevation in BA level and MSAF and fetal lung injury, with the proposed mechanism of action being BA– induced alteration of secretory phospholipase A2 [9, 10]. The current gold standard treatment for ICP is medical treatment with ursodeoxycholic acid (UDCA). According to a recent meta-analysis, patients who were treated with UDCA for ICP had complete resolution of pruritis in 42% of cases and an improvement in pruritis in 61% [18]. Generally, treatment response begins within one to two weeks of onset, with improvement within three to four weeks; the predominant side effects being mild nausea and dizziness. UDCA treatment is currently recommended for all patients with ICP. Current treatment recommendation is UDCA 300 mg orally every eight hours, and for treatment of refractory cases, either cholestyramine or S-adenosyl-methionine can be added to the treatment regimen; there is insufficient evidence to support the usage of these medications as single agents [19]. The aim of this study was to identify factors correlated with poor obstetrical and neonatal outcomes.

2 Methods

After receiving the Institutional Review Board (IRB) approval, current procedural terminology (CPT) codes were used to identify cases of ICP. A retrospective chart review of women admitted for ICP between January 1, 2013, and December 31, 2017 was performed. The inclusion criteria were: Women with singleton gestations who delivered at Newark Beth Israel Medical Center (NBIMC) during the aforementioned time period with the diagnosis of ICP, with BA ≥10μmol/L, and in whom diagnostic work-up ruled out other causes of liver disease. The exclusion criteria were: Women admitted to NBIMC during the aforementioned dates with bile acids <10μmol/L; those with other hepatic diseases and diseases affecting liver function tests; those who have had multiple gestations; and those who had no follow-up, who delivered at another institution, or whose medical records were incomplete. Patient characteristics such as maternal age, gravidity, parity, medical and surgical history, estimated gestational age at delivery, delivery method, neonatal data, and complications were collected. The biochemical data extracted were the maximum BA level during pregnancy, aspartate aminotransferase (AST) level, and alanine aminotransferase (ALT) levels. A total of 61 charts met inclusion criteria. Patients who delivered immediately after diagnosis were included because these patients often presented late in the third trimester and/or had markedly elevated BA levels. The delivery notes and neonatal intensive care unit (NICU) notes were analyzed for neonatal Apgar scores, weight in grams, complications, length of the NICU stay, and indication for the NICU admission. Adverse perinatal outcomes were defined as IUFD, non-reassuring fetal heart tracing, MSAF, preterm labor before 37 weeks of gestation, placental abruption, NICU admission, neonatal hyperbilirubinemia, respiratory distress syndrome (RDS), and neonatal demise.

3 Results

Examination of the pregnancy outcomes and clinical features of the 61 patients diagnosed with ICP showed no occurrence of IUFD, stillbirth, abruption, or neonatal demise. Of the 61 cases, the gestational age at delivery ranged from 27 weeks and two days to 39 weeks and six days, with a mean of 36 weeks and six days, and a standard deviation of σ: 1.53 weeks; 47.5% of deliveries were prior to 37 weeks (Table 1. Characteristics of singleton ICP pregnancies). BA levels ranged from 10 to 131, with a mean of μ: 41.8, and a standard deviation of σ: 31.1. The AST levels ranged from 11 to 1653, with a mean of μ: 115.7, and a standard deviation of σ: 223.1. There was no correlation between BA level and birth weight or age of gestation at the time of delivery.

The receiver operative characteristic curve analysis revealed a statistically significant correlation between BA and AST levels and perinatal outcomes (Fig. 1. Receiver-operator characteristic curves). Consistent with previously published studies, a BA level equal to or greater than 37μmol/L is strongly correlated with spontaneous preterm labor in women affected by ICP with a sensitivity of 100% and specificity of 60.70% (p = 0.002) (Table 2. ROC curve results table). A BA level equal to or greater than 42μmol/L is strongly correlated with MSAF with a sensitivity of 85.70% and specificity of 66.70% (p = 0.006). The AST level equal to or greater than 62 IU/L is strongly correlated with the NICU admission with a sensitivity of 81.30% and specificity of 62.20% (p = 0.002). The AST level equal to or greater than 75 IU/L is strongly correlated with hyperbilirubinemia with a sensitivity of 87.50% and specificity of 69.80% (p = 0.001).

OPEN IN VIEWER Fig.1

Although the exact mechanism of IUFD in ICP is unknown, it is hypothesized to be a sudden event. The current void in understanding leaves practitioners in favor of frequent antenatal testing and delivery prior to term, especially in women with a history of previous stillbirth. More studies that link ICP to IUFD are needed. It appears that there is an association between elevated BA, AST levels, and adverse perinatal outcomes. Our findings of adverse perinatal outcomes with BA ≥37μmol/L correlates with the current cut-off values of <40μmol/L for mild ICP and ≥40μmol/L for severe ICP. Additionally, we demonstrated the statistical significance of the specificity and sensitivity for the elevated AST levels with the NICU admission and hyperbilirubinemia. The literature has demonstrated significantly increased poor perinatal outcomes with BA levels ≥100μmol/L, however our cohort was limited to five cases meeting that criteria and no statistically significant correlation was found.

Our investigation found a significant correlation between adverse perinatal outcomes and elevated levels of AST, but not ALT. These enzymes are elevated in both acute and chronic hepatic injury. Both enzymes catalyze the transfer of α-amino groups from aspartate and alanine to the α-keto group of ketoglutaric acid to generate oxalacetic and pyruvic acids respectively [20]. These reactions require pyridoxal-5’-phosphate (vitamin B₆), which has clinical significance in patients with chronic alcoholic liver disease. In these patients, vitamin B6 deficiency can decrease the serum ALT activity and increase the AST/ALT ratio [21]. The potential hepatic biochemical process between ICP and transaminases needs to be further investigated. A meta-analysis published in 2019 by Ovadia et al., which was the largest combined cohort to date of women with ICP included a total of 4936 ICP cases. They found an association between women with severe ICP (defined as BA≥100μmol/L) and an increased stillbirth risk compared with those with milder disease and the background population [22]. The meta-analysis provided strong evidence that ICP is associated with adverse perinatal outcomes, especially when serum total BA are ≥100μmol/L. One limitation to the meta-analysis was that it was a multi-institute aggregation of data from both case-control and case-only studies, without standardized protocols or diagnostic criteria. Therefore the potential for inter-institute discrepancies in diagnoses such as neonatal asphyxia exist.

In analyzing our sub-groups of BA levels (0 to 39μmol/L, 40 to 99μmol/L, and ≥100μmol/L) there was no statistically significant difference between BA levels, LFTs, and sub-groups, however, the limitations of our study are that it was a retrospective study with a small number of patients (n = 61), and only 5 cases met the severe range criteria of bile acids ≥100μmol/L. Therefore we were only able to demonstrate an association between the elevated biochemical levels and adverse outcomes. The meta-analysis of Ovadia et al. highlighted the low incidence of ICP and the need for aggregating data [22]. In order to enhance our understanding of ICP, we recommend that a multicenter registry of ICP patients should be established as the incidence of ICP is low and the risk of stillbirth appears to approach the general population stillbirth rate in the United States of 6.05 per 1,000 births [23].

5 Conclusions

In singleton gestations affected by ICP, a BA level ≥37μmol/L is strongly correlated with spontaneous preterm labor and BA level ≥42μmol/L is strongly correlated with MSAF.

In singleton gestations affected by ICP, an AST level ≥62 IU/L is strongly correlated with the neonatal NICU admission and an AST level ≥75 IU/L is strongly correlated with the neonatal hyperbilirubinemia.

Our cohort did not have any stillbirths, and although it is possible the risk of stillbirth may be closer to the baseline rate of 6.05 stillbirths per 1,000 deliveries, further investigation is required [23]. Although we had positive correlations with bile acid level and aspartate-aminotransferase levels, we investigated short-term outcomes, and whether long-term sequellae exist has yet to be elucidated. As per our data analysis, a subcategory of patients with ICP and abnormal LFTs needs to be separately investigated.

Disclosures

A poster presentation of this research (Abstract 817) was presented at SMFM’s 39th Annual Pregnancy Meeting.