Maternal–fetal transfer of indocyanine green: a systematic review

Abstract

Rational

In a survey of 1101 members of vitreoretinal trained physicians regarding the use of ICG angiography during pregnancy, 434 (83%) of 520 respondents had seen at least one pregnant woman requiring ICG angiography or fluorescein angiography. One hundred and five (24%) withheld ICG angiography, mostly because of fear of teratogenicity or lawsuit. Adverse reactions to fluorescein and ICG are rare and may be classified as toxic, hypersensitivity, and non-specific. This literature review aimed to review evaluate the maternal-to-fetal transfer of ICG and resume the most recent recommendations for ICG use in its obstetric applications.

Methods

The available literature was examined using PubMed-Medline, and web of science, and using the MeSH terms “fluorescein,” “Indocyanine green,” and “pregnancy” according to PRISMA-P guidelines.

Results

Studies in humans demonstrated that ICG is not detectable in fetal cord blood or umbilical vein blood collected immediately after birth. ICG maternal-to-fetal transfer is slow and is safe during pregnancy. ICG in the fetus accumulates in the liver and accumulation is enhanced by the administration of OATPs or P-gp inhibitors.

Conclusions

ICG’s transplacental transfer is minimal and is probably medicine-mediated, like rifampin. The placenta is an effective protective barrier to ICG’s distribution into the fetus.

Keywords:

• Fluorescein
• indocyanine green
• pregnancy
• safety
• placenta

Introduction

Indocyanine green (ICG) is the only United States Food and Drug Administration-approved near-infrared (NIR) fluorophore. In 1961, Caesar et al. [1] first reported the use of ICG in the measurement of hepatic blood flow and hepatic function. It is administered intravenously at doses of 25 mg in adults [2]. Although ICG has been used to evaluate and further the understanding of retinal diseases that are associated with pregnancy or that occurred in pregnant patients, and studies since the 1960s demonstrated that ICG does not cross the placenta, there is little information on the distribution of ICG within the fetus and the outcomes of concomitant use of modulators of transporter function in terms of fetal exposure to ICG. Over the past years, ICG has been mostly used for ophthalmic angiography [3–8]. In a survey of 1101 members of vitreoretinal trained physicians regarding the use of ICG angiography during pregnancy, 434 (83%) of 520 respondents had seen at least one pregnant woman requiring ICG angiography or fluorescein angiography. One hundred and five (24%) withheld ICG angiography, mostly because of fear of teratogenicity or lawsuit, and only 5% thought it was safer than fluorescein angiography [9].

There was a little review on the maternal–fetal transfer of indocyanine green. Hence, the purpose of the study was to evaluate the maternal-to-fetal transfer of ICG. The main message of the studies was reported, and a summary statement was processed for each topic.

Methods

PubMed, and web of science, were searched for original articles on the maternal–fetal transfer of indocyanine green.

Eligibility criteria

This review incorporates all articles about the maternal–fetal transfer of indocyanine green, including randomized controlled trials (RCTs), observational studies, and interventional studies published between 1974 and August 2020.

Exclusion criteria

Studies published in a language other than English, those regarding indocyanine green imaging assessing perfusion in an infant, were all excluded from this review.

Study selection

In PubMed, web of science, and Google Scholar databases, we performed the search with MeSH terms, “fluorescein,” “Indocyanine green,” and “pregnancy.” The search was performed for publications reported between 1974 and August 2020. Titles and abstracts of the identified studies were checked, and irrelevant studies were removed. Only full-text papers in English were retained. Furthermore, the reference lists of the retained articles were further analyzed for articles that corresponded to the criteria for the present review.

Assessment of study quality

The quality of evidence in the one cohort study was rated by the GRADE working group and was both judged as moderate (2b). The interventional and observational studies were of low-quality evidence (2c–3b) [10].

Data synthesis

Summary measures of association were not computed because of the heterogeneity of study designs, study populations, and interventions. As a result, a meta-analysis was not made.

Results

After the search string was run on the different databases, we initially retrieved 332 articles. 306 were removed for irrelevance and the remaining 26 were screened by abstracts. At that point, among the 26 articles screened by full-text, 12 met the inclusion criteria. Out of 12 articles, eight were qualitative researches, three were cross-sectional studies, and one article used a mixed method (Figure 1).

Figure 1. Course of literature research and study design.

Discussion

Perfusion of placental cotyledons

Previous studies in humans demonstrated that ICG is not detected in umbilical vein blood or fetal cord blood collected at once after birth [6,7,9,11]. However, it is hard to precisely determine the transplacental kinetics of a particular molecule by sampling maternal and cord blood at term.

Ameer Bishara et al. [12] have demonstrated that ICG is detectable in mice fetuses and that medications can significantly enhance ICG distribution into the fetus. They found that ICG accumulates in the fetal liver and that concomitant medications that affect maternal ICG’s kinetics significantly enhance fetal exposure. P-gp [13] and OATPs [14] are crucial transporters at the placenta. Two transporter inhibitors were used, the non-selective OATP inhibitor rifampin (an established inhibitor of OATPs 1A2, 1B1, 1B3, and 2B1) and valspodar, a P-gp inhibitor [15]. ICG being a large, charged, and highly plasma protein-bound molecule, its extravascular distribution is limited. Thus, ICG’s transfer through membranes largely depends on transporters. Both valspodar and rifampin directly inhibited to placental accumulation of ICG at late pregnancy, indicated by the lower placenta: blood ratios. In other words, the extent of the increase in placental and fetal ICG concentration was lesser than expected based on blood ICG emission. The mechanism may be inhibition of OATPs localized at the maternal-facing apical membrane of syncytiotrophoblasts. Ameer Bishara et al. [12] indicated that, as in the mother, ICG accumulates in the fetal liver, and compounds that interact with ICG may enhance hepatic accumulation. Even the very limited ICG penetration into the fetal brain was enhanced at the presence of each of the transporter inhibitors. It appears that these compounds indirectly inhibit fetal hepatic or blood-brain barrier transporters, which may be playing an important role in fetal protection because the fetal liver, leg, and brain: leg ratios remained unchanged. Notably, the impact of ICG on the pharmacokinetics of concomitant medications should also be taken into consideration. Hence, until further data are available on the safety of ICG when combined with medications that affect its hepatic handling, such combinations should be used prudently.

Rubinchik-Stern et al. [16] represented a step toward improved understanding of the mechanisms mediating ICG’s transplacental transfer. They evaluated the maternal-to-fetal transfer of ICG using the ex vivo placental perfusion model of humans and the effects of transporter inhibitors on these processes. Placentas were obtained from cesarean deliveries. Cotyledons were cannulated and dually perfused. ICG, 9.6 g/mL and antipyrine (50 g/mL) were added to the maternal circulation in the absence (n = 4) or the presence of the organic anion transporting polypeptide (OATPs) inhibitor rifampin (10 g/mL; n = 5) or the P-glycoprotein inhibitor valspodar (2 g/mL; n = 3). ICG’s maternal-to-fetal transfer was evaluated over 180 min. ICG’s transfer to the fetal compartment was slow and became linear with time after the first hour of perfusion, either in the absence or the presence of rifampin or valspodar, <1% of the perfusate flow rate, and at 180 min, only 2% of the initial concentration in the maternal reservoir were measured, which further validates placental integrity across all studies. Using the dually perfused human placenta, agree with a very low fetal: maternal ICG ratio. ICG’s low transfer across the placenta is attributed to high binding to plaznot known to be metabolized in humans [17]. Previously, valspodar increased 7.9-fold [18] the placental transfer of saquinavir, an OATPs and P-gp substrate [15], not affecting the transfer of talinolol [19], another P-gp/OATP substrate [20]. OATPs were considered to be greater determinants of ICG’s maternal-to-fetal transfer compared to that of saquinavir. Although the formation of the placenta differs between mice and humans, the trends in pregnancy regarding expression and activity of drug transporters appear to be overall similar. In contrast to the fast placental transfer of many drugs, ICG maternal-to fetal transfer is slow.

Kausik Basak et al. [21] developed an optoacoustic tomography approach for real-time imaging through entire 4 cm cross-sections of pregnant mice. Functional changes in both maternal and embryo regions were studied at different gestation days when subjected to an oxygen breathing challenge and perfusion with indocyanine green. The perfusion of ICG from the maternal region to the fetoplacental area after a tail vein injection into an E14 mouse was explored. Such observations may enable the understanding of perfusion of ICG through the placental barrier and substantiate its function in protecting fetuses from external stimuli.

ICG detected in fetal cord blood or umbilical vein blood

Studies in humans demonstrated that ICG is not detectable in fetal cord blood or umbilical vein blood collected immediately after birth.

Probst et al. [22] detected that no ICG could be tested either in fetal scalp blood or umbilical vein blood following its injection into the maternal circulation in g full-term pregnant women. ICG was injected intravenously into mothers during the second stage (seven women) or the first stage (two women) of labor. Before, 2–4 and 6–8 min after dye injection blood samples were obtained from the mother and the fetal scalp. Delivery occurred 9 min to 6 h after ICG-injection. In addition to fetal scalp blood sampling, umbilical vein blood was collected immediately after birth in four cases. ICG concentrations in maternal blood after injection of 5 mg ICG per kg were 2.59–4.80 and 1.08–3.68 mg%, 2–4 and 6–8 min after dye injection. No ICG could be found in fetal blood samples which were obtained simultaneously or in umbilical vein blood collected immediately after birth. No effects of ICG on the mother, child, or on delivery were presented.

Meir et al. [23] studied the effects of valproic acid on the placental barrier in the pregnant mouse. They demonstrated that several AEDs, including VPA, affect the expression of transporters for endogenous compounds and xenobiotics in a human placental cell line [24]. They further investigated the effects of VPA on the placental barrier in vivo, in pregnant mice. Studies were conducted on gestational days 12.5 (mid-gestation) or 17.5 (late gestation), following intraperitoneal treatment with 200 mg/kg VPA or the vehicle. Indocyanine green (0.167 mg, i.v.) served as a marker for the placental barrier permeability. VPA treatment was related to a 40% increase (p < .05) in accumulation of ICG in the maternal liver in mid‐pregnancy and a decrease by one-fifth (p < .05) in late pregnancy. Ex vivo, VPA treatment resulted in a 20% increase (p < .05) in fetal ICG emission in mid-pregnancy. Besides mechanistic studies in pregnant animals, ICG may provide a tool for studying the effects of epilepsy and AEDs on hepatic transport mechanisms in humans.

Studies were performed in 168 primi- and multigravidas with normal medical and obstetric histories (mean age of 23.5 years), 18 healthy non-pregnant women serving as controls (mean age of 24.2 years). 0.5 mg ICG/kg body mass was injected intravenously as a bolus [25]. No placental transfer of the dye could be detected by simultaneous measurement of ICG in the maternal and fetal cord blood. ICG is recommended for the estimation of excretion function of the liver during pregnancy and childbed.

In this study, the authors aimed to review and resume the last recommendations of ICG applications in pregnant women. ICG is a water-soluble compound that has been extensively used to assess cardiac output, hepatic function, and ophthalmic angiography for more than 50 years. Binding to protein in the blood or the tissue, ICG emits energy in the NIR region between 750 and 810 nm. It obtains a short plasma half-life (3–5 min) that ensures it an excellent safety profile, particularly when repeated assessments are necessary. The most frequent adverse effects of ICG (nausea and/or vomiting; rarely discomfort, hot flush, or skin rash; very rarely anaphylactic shock) were inappreciable compared to the benefits of its clinical potential [26].

ICG does not cross the placenta and it has been given to pregnant women without adverse effects on the mother or fetus [27–29].

Limitations

The power to detect ICG in fetal cord blood, in umbilical vein blood or placental cotyledons was limited by small numbers. The studies had relatively small sample sizes and may have been underpowered to detect significant differences between groups. In addition, gray literature, reports, and unpublished studies were not included in this review.

Conclusions

In conclusion, this study revealed that ICG maternal-to fetal transfer is slow, and is safe during pregnancy. ICG in the fetus accumulates in the liver and accumulation is enhanced by the administration of OATPs or P-gp inhibitors. Hence, until further data are available on the safety of ICG when combined with medications that affect its hepatic handling, such combinations should be used with caution. Thus, ICG can potentially be leveraged in periodic monitoring of maternal and fetal health and also in clinical applications to study and aid pregnancy-related abnormalities.

Acknowledgments

The authors acknowledge the support of the Scientific Research Seed Fund of Peking University First Hospital (2020SF30). We are grateful to the Peking University First Hospital.

Disclosure statement

The authors report no conflicts of interest.