As a pivotal cause of liver cirrhosis and hepatocarcinoma, chronic hepatitis B virus (HBV) infection remains a major health problem with global impact and accounts for approximately one million deaths annually, which surpasses the deaths caused by human immunodeficiency virus, tuberculosis, and malaria infection (1). Antiviral therapy against HBV has displayed significant effectiveness in improving survival and quality of life by preventing disease progression, and even reversing liver fibrosis and cirrhosis (2). In May 2016, the World Health Organization proposed the goal for elimination of viral hepatitis as a public health threat by 2030. In this instance, aggressive antiviral treatment is suggested. According to some guidelines, an increasing number of female patients initiate antiviral treatment in their childbearing age (2, 3).

Since be cloned and purified, interferon (IFN) is the first approved biotherapeutic agent. With accumulating studies, IFNs present not only antiviral activity with a spectrum of clinical effectiveness, but are also the prototypic biological response modifiers for cancers. Currently, genetically engineered IFNs have been applied primarily in the treatment of patients with HBV and hepatitis C virus infection, multiple sclerosis (MS), malignant melanoma, neuroendocrine tumors, and certain lymphoproliferative and hematological diseases (4). It is now well established that type I IFN can induce innate immune function and inhibit HBV replication (5, 6). Since type I IFN can induce long-term immunological control of HBV replication within a finite duration treatment, sets of guidelines have recommended type I IFN as preferred initial therapy (2, 7, 8). Moreover, with accumulated data about antiviral effectiveness of IFN, an expert consensus has recommended combined therapy of IFN and nucleos(t)ide analogs to peruse functional cure for chronic HBV patients (9). Therefore, IFN has been extensively applied to treat chronic HBV patients for clinical cure, especially in the young population.

With the latest estimate, there are 65 million women of reproductive age with HBV infection (1). Globally, 44% of pregnancies are unintended. Although effective contraception is required during IFN therapy, unintended pregnancy is still inevitable (10). IFN was contraindicated during pregnancy considering the potential influence on fetuses. Whereas, among the majority of guidelines on the management of HBV infection, there are few recommendations about how to deal with unintended pregnancy during IFN treatment (2, 7, 11, 12) or even termination of pregnancy is suggested in Chinese guidelines (13). Nevertheless, it is the concern of obstetricians that the significant high prevalence of secondary infertility would lead to depression and discrimination (14). Given these potential challenges, counseling about the possible risks to fetuses after IFN exposure is an important part of disease management for chronic HBV patients. At present, there is no consensus on the management of unintended pregnancy in female patients with HBV infection during IFN therapy.

This highlights the need for up-to-date information on IFN safety during pregnancy. We found low-quality evidence about the safety of IFN during pregnancy; randomized controlled study of IFN administration during pregnancy is unlikely to be conducted given the ethical concerns of such a trial. Compared with any single study, meta-analysis has more power for statistical performance, and increased precision for confidence intervals, because the conclusions often reflect a broad spectrum of objectives and study characteristics and are more generalizable. Consequently, meta-analysis may be an appropriate alternative to help guide decision making in this special situation. Furthermore, since type I IFN has also been approved for MS (15), essential thrombocythemia (ET) (16), and other myeloproliferative diseases, systematic evaluation of all information about IFN exposure during pregnancy would provide a comprehensive safety profile for hepatologists.

Therefore, we conducted this meta-analysis of all available studies to evaluate the effect of IFN on pregnancy outcomes and provided references for clinical decisions.

The systematic review and meta-analysis were registered in advance in PROSPERO (CRD42020183239) and were reported in line with the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) (17) and Meta-analysis of Observational Studies in Epidemiology (MOOSE).

Since be cloned and large-scale produced, type I IFNs have been successfully applied as therapies in patients suffering from various diseases (44–46). Although there are scant published data discussing the safety in pregnant women with chronic HBV infection, IFN has been recommended widely in MS, ET, chronic myeloid leukemia, multiple myeloma, and others (46–48). Furthermore, with growing research data in pregnant women with MS exposed to IFN-β-1a, the Food and Drug Administration has removed the category C warning from the package insert for IFN-β-1a in 2020, which gives more confidence to women with MS to continuing IFN therapy while starting to grow their families. In this meta-analysis including the most comprehensive and up-to-date research on the safety of INF exposure in pregnant women, we found that exposure to type I IFNs during pregnancy did not adversely affect pregnancy or infant outcomes.

In these meta-analysis enrolled 3,543 pregnancies, the pooled estimate of live birth was 85.2% in IFN-exposed group, which is in accordance with that reported in the general population (49). When further compared with pregnant women with the same diseases unexposed to IFN, there was no significant difference, suggesting no evidence of the adverse effect on live birth. This similarity has also been presented in other systematic literature reviews and observational studies (50, 51). Additionally, a recent meta-analysis about pregnant women with myeloproliferative neoplasms suggests that the use of IFN-β during pregnancy was associated with higher successful pregnancy than observation alone, which may be attributed to the potential antiproliferative properties of IFN in myeloproliferative neoplasms (52). In view of these available research results, IFN exposure displays no adverse impact on successful pregnancy.

In line with the general pregnancy population, spontaneous abortion is the primary adverse outcome in the current investigation. A pooled rate of 9.4% was presented, which is not higher than the estimates based on the U. S. National Survey of Family Growth data (15–16%) (53) and that from the general population (16%) (54). Furthermore, when compared with patients unexposed to IFN, the risk of spontaneous abortion (OR 1.09; 95% CI 0.73–1.63; P = 0.672) was not increased. In addition, as presented in Figure 3, the prevalence of stillbirth and preterm delivery remained within the ranges reported in the general population (49). Moreover, the risks of these adverse pregnancy outcomes were not increased after exposure to IFNs in pregnant women with the same diseases. Several clinical trials and reviews have previously evaluated the safety profile of IFN in pregnant women, the majority of the exposure was during the first trimester (22, 51, 55). A bit higher rates of spontaneous abortions, lower birthweight (20, 21), and shorter mean birth length were presented in observational cohorts with a small sample size (20, 21, 55). Whereas, these negative pregnancy outcomes were not demonstrated with IFN exposure in systematic reviews and clinical cohorts with large samples (22, 34, 51). Based on these results, there is no evidence of a signal, suggesting that exposure to IFNs before conception and/or during pregnancy adversely influences spontaneous abortion, stillbirth, preterm delivery, or other adverse pregnancy outcomes.

In the enrolled studies, there were 67 maternal adverse events reporting from 7 cohorts, and pooled analysis did not show an increased risk of maternal complications in pregnant women exposed to IFN. These findings align with the previous meta-analysis concerning the IFN exposure in pregnant women with myeloproliferative neoplasms, and are consistent with the worldwide prevalence (56). Whereas, some IFN-associated adverse events, including general fatigue, hair loss, and IFN-associated autoimmune diseases, as well as other maternal complications including pre-eclampsia, gestational diabetes, were not sufficiently evaluated in the current analysis. It is thought that the main concern of the included studies contributes to whether or not it is reported. Furthermore, the reporting bias may also be due to the missing data associated with retrospective studies.

As the main concern among hepatologists about IFN exposure during pregnancy, the pooled estimate of birth defects was 0.51% in this meta-analysis, which is lower than that reported in the United States (3%) (57) and in the European Union (2%) (58). With the largest prospective cohort providing a safety profile of women with MS exposed to IFN-β, Hellwig and his colleagues reported that the prevalence of birth defects was within an estimated range of 2.1–2.8% (51). The lower estimates of birth defects may be attributed to reporting bias associated with retrospective cohorts, and selective termination due to concern about the “potential effect” of IFNs. Whereas, there have been reports that expression of type I IFNs is constitutive in placental, regulating the immune balance between mother and fetus during gestation (59, 60). IFN-α is reportedly present in fetal blood, amniotic fluid, and placental amniotic membrane to prevent intrauterine viral infection (61). Finally, among the 18 cases with birth defects, no consistent pattern of developmental abnormalities was identified, which suggests that there is no evidence of signal suggesting exposure to type I IFN resulting in birth defects.

As a major choice of antiviral therapy, especially in young patients with HBV, unintended pregnancies during IFN-α treatment is inevitable. Although only case reports were retrieved in pregnant women with viral hepatitis, there were cohorts presenting outcomes from patients with other myeloproliferative diseases. Similar to a previous investigation, comparing with those with same diseases without IFN-α treatment, the risk of adverse outcomes, including spontaneous abortion, stillbirth, preterm delivery, birth defects, and maternal complications, were not increased in pregnancies exposed to IFN-α (52). Although the included sample was small, the results were further confirmed with a meta-analysis concerning pregnancy outcomes after exposure to IFN-α or IFN-β, which shares the same receptor with IFN-α and is also approved for the treatment of viral hepatitis (35). Overall, the current meta-analysis presents a comprehensive assessment of the safety in pregnancies with exposure to type I IFNs and provides references for clinical decisions in unintended pregnancies during IFN treatment.

There are inevitable limitations in the present meta-analysis. First, the included studies in this research were retrospective cohorts and case–control observations, and none were randomized controlled trials. However, from an ethical point of view, it is ethically unacceptable to perform ideal randomized controlled trials in such a special population. In addition, during the process of literature inclusion, we conducted a rigorous quality assessment of the included cohort studies according to NOS, and only those with moderate quality were included in this meta-analysis. Due to the particularity of the study population and intervention conditions, it is difficult to conduct higher-quality studies in the future. Second, there were only case reports about pregnant women with viral hepatitis, and only a few cohorts were eligible for meta-analysis. Because of the conventional concern of the “potential teratogenic effects of IFN” among hepatologists, leading to selective/induced termination, only a few studies are available reporting pregnancy outcomes from women with viral hepatitis. Another limitation of the current meta-analysis is the potential for under-reporting due to retrospective studies. Whereas, prospective observational studies or registries require many years to reach a sample size that afford the statistical power to detect a difference in the relevant risk compared with the general population or untreated patients. Furthermore, due to dose and duration of IFN exposure in every single case was unavailable, we could not determine the association of adverse pregnancy outcomes with higher doses or longer duration of IFN exposure. Nevertheless, this systematic review with comprehensive evaluation of type I IFNs could provide references for hepatologists to make a strategic clinical decision.

Taken together, the current analysis presented a comprehensive assessment of the safety of IFN exposure during pregnancy. The estimated prevalence and risks of adverse pregnancy outcomes and birth defects had not been worsened with IFN exposure before conception and/or during pregnancy. Despite the low-to-moderate quality of the enrolled studies in this analysis, there is no evidence of signal suggesting termination of pregnancy after exposure to IFN prior to or during pregnancy. However, as with all drugs used in pregnancy, it should be encouraged and supported to collect data on birth outcomes to further assess any potential risk of adverse outcomes.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

MZ, SF, and DR: Study design, study identification, data collection and extraction, quality assessment, data analysis and interpretation, and manuscript drafting. NY, YF, and TN: Data collection, quality assessment, manuscript revision. YC: Study design, data analysis, interpretation, and critical review of the manuscript. TC, YZ, and JL: Study concept, study design, data analysis, interpretation of data, manuscript drafting and revision, quality control of algorithms, and study supervision. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.