Outcomes of pregnancies complicated by cirrhosis: a retrospective cohort study

A retrospective analysis was conducted on 81,554 pregnant women who gave birth at the Third Affiliated Hospital of Sun Yat-sen University between 2010 and 2022. Women with cirrhosis were initially identified according to the International Classification of Diseases, 10th Revision (ICD-10) [13]. A computer search of these women aimed to identify cases with liver cirrhosis (ICD-10 O26.604, pregnancy with cirrhosis; ICD-10 K74.100, liver cirrhosis; ICD-10 K74.607, decompensated cirrhosis; ICD-10 K74.3, primary biliary cirrhosis; ICD-10 K71.7, drug-induced cirrhosis; ICD-10 K70.3, alcoholic cirrhosis; ICD-10 K74.6, other causes of cirrhosis). The ICD-10 code for cirrhosis before or during pregnancy, the ICD-10 code for hepatic decompensation, and the ICD-10 code for causes of cirrhosis were applied to reduce the potential for missing a case of liver cirrhosis. In addition, each case that was considered to be liver cirrhosis was reviewed by two authors to reduce the number of false-positive cases. We identified 50 cases of pregnant women with cirrhosis. The inclusion criteria for this study were patients aged ≥ 18 years with a gestational age ≥ 28 weeks at the time of giving birth. Women with a previous history of liver transplantation were excluded from the study. Out of the 50 pregnant women with cirrhosis, seven chose to terminate their pregnancies before 28 weeks due to reasons such as decompensated cirrhosis, thrombocytopenia, and coagulopathy. None of these women had spontaneous abortions. The remaining 43 pregnant women met the inclusion criteria and were assigned to the cirrhosis group. All 43 pregnant women had been diagnosed with cirrhosis prior to pregnancy. Additionally, 172 pregnant women without cirrhosis were randomly selected from the non-cirrhosis group at a 1:4 ratio using the SPSS random number table. Statistical analyses were performed using SPSS software (Version 25.0; SPSS Inc, Chicago, IL).

Decompensated cirrhosis was identified when the following clinical symptoms occurred concomitantly with cirrhosis: hepatic encephalopathy, variceal hemorrhage, hepatorenal syndrome, spontaneous bacterial peritonitis, hepatopulmonary syndrome, and/or ascites [14]. For the cirrhosis group, data regarding the etiology of cirrhosis, treatment measures for cirrhosis, and examination indicators related to cirrhosis were collected. Covariates included maternal age, body mass index (BMI), twin pregnancy, breech presentation, prior cesarean section, parity, pregestational diabetes mellitus, and chronic hypertension [8, 15].

Maternal and neonatal indicators were compared between the cirrhosis group and non-cirrhosis group. Obstetric outcomes included cesarean section, induction of labor (artificially stimulating the uterus to initiate labor), postpartum hemorrhage (defined as blood loss ≥ 500 ml at vaginal delivery or blood loss ≥ 1,000 ml at cesarean section), preeclampsia, gestational diabetes, placental abruption, oligohydramnios, intrahepatic cholestasis of pregnancy (ICP; defined as itching in skin of normal appearance with a raised peak while detecting total random bile acid concentrations of 10 µM/L or more), thrombocytopenia (a platelet count below 150 × 10⁹/L), maternal death, fetal distress (fetus’s acute and chronic hypoxic symptoms under the influence of various uterine factors), transfusion, and peripartum maternal intensive care unit (ICU) admission. Neonatal outcomes included fetal or neonatal demise, preterm birth (giving birth at or after 28 0/7 weeks of gestation and before 37 0/7 weeks of gestation), small for gestational age (SGA; defined as a birthweight below the tenth centile for gestational age), and an Apgar score of less than 7 at 5 min.

To investigate their associations with preterm birth in women with cirrhosis, we assigned values to intrahepatic cholestasis of pregnancy (ICP) and decompensated cirrhosis. The assignment for ICP (Without = 0, With = 1). The assignment for decompensated cirrhosis (Without = 0, With = 1).

All data analyses were performed using the statistical software package SPSS version 25.0. Clinical characteristics and pregnancy outcomes were compared using Student t-tests and Mann–Whitney tests for continuous variables, as well as Fisher’s exact test or χ² test for categorical variables. Logistic regression was performed to calculate obstetric outcomes among women with cirrhosis compared to those without cirrhosis. The results are presented as the odds ratio (OR) with 95% confidence intervals. All tests were two-tailed, and a P-value < 0.05 was considered statistically significant when comparing the cirrhosis group and the non-cirrhosis group.

This study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee of Third Affiliated Hospital of Sun Yat-sen University (No. II2023–053-01). The requirement for informed consent was waived by the Ethics Committee of Third Affiliated Hospital of Sun Yat-sen University because of the retrospective nature of the study.

The incidence of cirrhosis in pregnancy was 0.06% (50/81, 554). Characteristics of pregnant women in the cirrhosis group are presented in Table 1. The most common etiology was viral hepatitis B (53.49%), whereas no cases of alcoholic liver disease were reported among pregnant women in the cirrhosis group. Decompensated cirrhosis was identified in 13 pregnant women in our study. Only 17 patients underwent endoscopy either before or during pregnancy, and 14 out of 17 patients were found to have varices. Of the 17 patients, 10 underwent endoscopy during pregnancy, 4 were examined in the second trimester, and 6 were examined in the third trimester. No immediate risk to the mother or fetus was observed in the patients who underwent endoscopy during pregnancy. There were no cases of variceal hemorrhage nor maternal death during the peripartum period; however, six pregnant women were admitted to the general obstetric ward during pregnancy and then were transferred to the ICU after delivery.

The gestation at time of birth occurring in pregnant women with decompensated cirrhosis was 33.38 ± 3.78 weeks. The number of preterm births in the pregnant women with decompensated cirrhosis was 9. All 13 cases of decompensated cirrhosis were cesarean section deliveries. One of these women had spontaneous onset but refused vaginal delivery. None of the pregnant women with decompensation cirrhosis delivered with labor induction.

There were no significant differences between the two groups in terms of age, BMI, twin pregnancy, breech presentation, prior cesarean section, nulliparous, multiparous, pregestational diabetes, or chronic hypertension (Table 2).

In comparison to pregnant women without cirrhosis, pregnant women with cirrhosis displayed a higher risk of various adverse outcomes, including cesarean section, preterm birth, ICP, thrombocytopenia, and postpartum hemorrhage (p < 0.05; Table 3). None of the pregnant women in the cirrhosis group underwent induction of labor.

Twenty pregnant women in the cirrhosis group experienced preterm birth, leading to an incidence rate of 46.51% (20/43), which was significantly higher than that of the non-cirrhosis group. The gestation at time of preterm birth in the cirrhosis group was significantly smaller than that in the non-cirrhosis group (32.25 ± 2.81 weeks versus 34.83 ± 1.47 weeks). Of the twenty pregnant women, 17 were emergency cesarean section; 1 was an elective caesarean section; and 2 were spontaneous onset. No preterm birth occurred as a result of iatrogenic induction of labor. Reasons for emergency cesarean section in patients with preterm birth included fetal distress (2), elevated bile acids (6), placental abruption (1), placenta previa with hemorrhage (2), spontaneous onset with breech presentation (1), spontaneous onset but refused vaginal delivery (1) and decompensated cirrhosis (4). Most of the emergency caesarean sections were due to obstetric indications.

Univariate logistic regression analysis was carried out on the dependent variables of the outcome of preterm birth and the independent variables listed in Table 4. The regression coefficient β for variables PT, TBIL, ICP, and decompensated cirrhosis were positive, and the OR values were greater than 1, indicating that prothrombin time (PT), total bilirubin (TBIL), ICP, and decompensated cirrhosis were risk factors for the prognosis of preterm birth (p < 0.05; Table 4).

As shown in Table 5, the regression equation used was as follows:

Of the 43 pregnant women with cirrhosis, the model predicted 20 women had preterm births and 23 women had term births. Of the 20 preterm births predicted by the model, 4 were not preterm births, while 4 of the 23 term births predicted by the model were preterm births, producing a model prediction accuracy rate of 81.4%.

Using the above-mentioned regression equation, we obtained probability predictions for preterm birth for the group of 43 pregnant women with cirrhosis. We then plotted the ROC curve, which demonstrated an area under the ROC curve (AUC) of 0.847 (95% CI: 0.723–0.971). The point closest to the upper left of the curve exhibited a sensitivity of 90% and specificity of 78%. The Youden Index was calculated as follows: Youden Index = Sensitivity + Specificity − 1. The cutoff point was 0.31, when the Youden Index was at its maximum (Fig. 1).