Circulating biomarkers associated with placental dysfunction and their utility for predicting fetal growth restriction

Abstract Fetal growth restriction (FGR) leading to low birth weight (LBW) is a major cause of neonatal morbidity and mortality worldwide. Normal placental development involves a series of highly regulated processes involving a multitude of hormones, transcription factors, and cell lineages. Failure to achieve this leads to placental dysfunction and related placental diseases such as pre-clampsia and FGR. Early recognition of at-risk pregnancies is important because careful maternal and fetal surveillance can potentially prevent adverse maternal and perinatal outcomes by judicious pregnancy surveillance and careful timing of birth. Given the association between a variety of circulating maternal biomarkers, adverse pregnancy, and perinatal outcomes, screening tests based on these biomarkers, incorporating maternal characteristics, fetal biophysical or circulatory variables have been developed. However, their clinical utility has yet to be proven. Of the current biomarkers, placental growth factor and soluble fms-like tyrosine kinase 1 appear to have the most promise for placental dysfunction and predictive utility for FGR.


Introduction
A small for gestational age (SGA) infant is variably defined as one with an estimated fetal weight (EFW) or birthweight (BW) less than the 10 th centile for gestation [1][2][3]. Globally, almost 21 million infants are born SGA each year, the majority in low-income and middle-income countries [4]. These infants are at higher risk of morbidity and mortality particularly if they are born preterm [5] and are also more likely to develop chronic health complications through childhood and in adulthood [6].
Although SGA and fetal growth restriction (FGR) are often used interchangeably, FGR is defined as an infant that has not achieved its genetic growth potential [7]; however, as this is inherently unknown, it is impossible to determine if any infant has indeed achieved that potential prenatally. Importantly, not all infants with FGR will be SGA and not all SGA infants will have FGR. Regardless of this caveat, many SGA/FGR liveborn infants will have low birth weight (LBW) defined as a BW <2500 g irrespective of gestational age. Worldwide, LBW is an important public health indicator, especially in settings where accurate gestational age assessment is not possible and prenatal assessment of fetal size or growth is not available [5,8].
Normal placental development [9] involves a combination of highly regulated processes requiring a plethora of angiogenic growth factors, hormones, transcription factors, cytokines and cell adhesion molecules [10]. Failure to establish a high capacitance, low pressure maternal fetal vascular interface [9] leads to placental dysfunction and is causally related to several obstetric syndromes including pre-eclampsia and FGR [11]. Although there is considerable overlap between the pathogenesis of pre-eclampsia and FGR, the relationship between the extent of failure of spiral artery conversion, gestation at onset, type of disease, maternal and infant phenotype as well as clinical outcomes remains poorly understood [12][13][14].
The challenge obstetricians face is identifying the truly growth restricted fetus regardless of size, as these are the infants that are most at risk of adverse outcomes. Because placental dysfunction leading to inadequate nutrient and oxygen transfer [15] accounts for the majority of SGA/FGR infants [16,17], many investigators have focused attention on circulating biomarkers indicative of aberrant placental function [18][19][20]. As there is currently no treatment for placental dysfunction, early recognition of at-risk pregnancies is important because careful maternal and fetal surveillance can be instituted and adverse outcomes potentially prevented by judicious timing of birth [21][22][23]. The aim of this narrative review is to provide an overview of available evidence regarding circulating biomarkers associated with placental dysfunction and to discuss their clinical utility for the prediction of FGR. A comprehensive review of PubMed, Cochrane Library, and CINAHL was performed to identify appropriate publications between 1995 and October 2022 relevant to this review.
Other studies have shown that although there is good correlation between first trimester maternal serum ADAM12 concentrations with BW centile, it performs poorly as a standalone screening test [29]. A more recent study evaluated maternal plasma ADAM12 late in the third trimester (36 weeks) and despite finding significantly lower median concentrations in women with SGA infants compared with controls [14115 pg/ml (Interquartile range (IQR): 11510-16592 pg/ml) vs. 16582 pg/ml (IQR: 13658-20322 pg/ml)], it was not suitable as a screening test [35].
A systematic review and meta-analysis (32 studies; 175240 pregnancies) assessing the predictive utility of first trimester maternal serum PAPP-A concentrations for birth of an SGA infant, revealed poor predictive value with low positive (PLR) and negative (NLR) likelihood ratios for BW < 10 th centile: PLR  [36], suggesting that PAPP-A was not suitable as a standalone biomarker to predict SGA infants [19,36]. In another case-control study at 11-13 weeks of gestation, a combination of uterine artery pulsatility index (UtA-PI), maternal mean arterial pressure (MAP), and serum concentrations of PAPP-A, βhCG, PlGF, PP13, ADAM12 and fetal nuchal translucency thickness, yielded a detection rate of 73% and 46% for birth of a preterm and term SGA infant, respectively [37].

PP13 and AFP
PP13 is a glycan-binding protein mainly expressed in syncytiotrophoblast and secreted into the maternal circulation via exosomes or microvesicles [38]. Although an earlier observational study showed an association between low first trimester PP13 concentrations in maternal serum and FGR [39], subsequent studies [33,40] found limited predictive utility with no significant differences in median PP13 MoM levels in FGR-affected pregnancies even when combined with other first trimester screening markers such as PAPP-A [33] and ADAM12 [40]. Another study also showed that median PP13 concentrations in the first trimester of women with SGA infants (BW < 3 rd , < 5 th , and < 10 th centiles) were not significantly lower than the control arms (0.978, 1.058, 1.051, and 1.083 MoM (controls) for each BW centiles, respectively) [41]. These findings were further corroborated in another study [42], which also failed to demonstrate the utility of PP13 for prediction of FGR [41]. In another study, median first trimester PP13 concentrations were significantly lower in FGR pregnancies compared with controls (86.6 vs. 132.5 pg/ml) but the overall sensitivity for the prediction of FGR was low at 33% at a specificity rate of 90% [43]. Two systematic reviews of first trimester serum PP13 in combination with maternal characteristics for the prediction of SGA infants reported low sensitivity of 32% (95% CI: 18-48%) [44] and 36% (95% CI: 33-41%) [45], respectively. The overall evidence thus far suggests that PP13 has limited clinical utility for predicting FGR.
In a study of 9715 singleton pregnancies (including 481 SGA infants with BW < 5 th percentile), higher mean log 10 MoM value of maternal serum AFP at 19-24 weeks of gestation was seen in the SGA cohort. When AFP levels were combined with maternal factors, fetal biometry, and maternal PlGF concentrations, detection rates of 100%, 76%, and 38% were achieved for SGA infants delivered at <32, 32-36, and ≥37 weeks gestation, respectively [34]. The addition of UtA-PI measurement further improved detection rates of SGA infants to 78% at 32-36 weeks and 42% at >37 weeks, respectively [46]. Another retrospective study [47] showed that while elevated serum AFP concentrations (≥2.5 MoM) in the first trimester was associated with birth of an SGA infant, its predictive utility for SGA and FGR was low with an AUROC of <0.6 [47]. Similarly, other studies have also demonstrated that although third trimester AFP concentrations are significantly lower in women with SGA infants, the overall detection rate using this biomarker is low at 26%, and even when it is combined with maternal PlGF concentrations detection rates only modestly increase to 32% [48]. A recent meta-analysis (39 cohort studies; 93968 women) reported that the relative risk (RR) for birth of an SGA infant in women with elevated AFP concentrations was increased (RR: 2.02, 95% CI: 1.75-2.33) and this risk was higher when ultrasound evidence of SGA was present (RR: 5.28, 95% CI: 3.46-8.06) [49].

Inhibin A, activin A, and follistatin
Maternal serum concentrations of activin A, inhibin A, and the activin:follistatin ratio in the third trimester have been reported to be significantly increased in FGR pregnancies compared with controls [50]. However, in another study, there was no difference in activin A or inhibin A concentrations in normotensive women with SGA infants compared with controls [51]. A study by Miranda et al. [52] showed that although mean inhibin A concentrations were significantly higher in women with SGA infants, a multivariable integrative model of maternal characteristics, fetoplacental ultrasound, and maternal biochemical markers only modestly improved the detection of SGA/FGR cases at 32-36 weeks' gestation when compared with screening based on EFW centiles alone. Other studies [53] have also shown that the clinical utility of activin A, inhibin A, and follistatin as predictors for SGA/FGR is poor.

PGH, IGF-I, IGFBP, N-CAM, and FGF
PGH is mainly expressed by syncytiotrophoblast and stimulates gluconeogenesis and anabolic pathways to support the growing fetoplacental unit [54]. Earlier observational studies reported an association between low maternal serum PGH concentrations in the second and third trimesters and birth of an SGA infant [55,56]. However, in a later study [57], first trimester median maternal serum PGH concentrations in SGA pregnancies were not different to controls (0. 95 [58], after multiple regression analyses and adjustment for maternal characteristics, these biomarkers were ultimately not useful for the prediction of SGA. Similarly, in another study, although a significant negative correlation between log IGFBP-1 and BW standard deviation score was noted, after adjusting for maternal body mass index, the relationship became nonsignificant [59]. IGFBP-4 is highly expressed by extravillous trophoblasts at the maternal-fetal interface [60] and circulating maternal IGFBP-4 concentrations in early pregnancy have been reported to be significantly higher in women with FGR infants (defined as BW < 5 th centile) compared with controls [Odds ratio (OR) 22.3, (95% CI: 2.7-181.5)] with 93% positive predictive value (PPV) [60]. Current evidence, however, does not support the use of PGH, IGF-I, or IGFBP as reliable markers for the prenatal prediction of FGR [57][58][59][60].

SPINT1
SPINT1 is a circulating protein highly expressed by villous cytotrophoblasts. It is involved in the conversion of maternal spiral arteries into low pressure, high capacitance vessels by modulating trophoblast secretion of proteolytic enzymes (serine proteinases, metalloproteinases, and collagenases) that regulate transformation of spiral arteries in normal placentation [67]. In vitro and animal studies suggest that SPINT1 is modulated by hypoxia and decreased in FGR placentae. In a study of 2003 women [68] at 36 weeks' gestation, a strong association between low plasma SPINT1 concentrations and SGA (defined as <10 th centile) was seen. Using a SPINT1 cutoff threshold of <0. 63 MoM, the risk of delivering an SGA infant with BW <3 rd , <5 th , and <10 th centile was 14.1%, 19.7%, and 28.2%, respectively [68]. A recent cohort study [69] found maternal plasma SPINT1 concentrations were significantly lower at 20 weeks gestation in women who subsequently delivered an SGA infant; however, the AUROC was modest at 0.62 for BW <3 rd centile and 0.56 for BW <19 th centile, respectively. Murphy et al. [70] reported reduced plasma SPINT1 concentration in women with pre-eclampsia who subsequently delivered an SGA infant (median [IQR]: 18857 pg/ml  in SGA vs. 40168 pg/ml  in controls). Another study by Murphy et al. [67] found an association between elevated plasma serine protease inhibitor Kunitz type 2 (SPINT2), which is functionally related to SPINT1 in pregnancies complicated by pre-eclampsia and/or SGA. However, the evidence for SPINT1 as a suitable biomarker to predict SGA/FGR is limited, and more evidence is required to validate its clinical utility. Table 1 presents angiogenic biomarkers associated with SGA or FGR. Both vascular endothelial growth factor (VEGF) and placental growth factor (PlGF) play an important role in facilitating angiogenesis in placenta and transforming spiral arteries into low resistance capacitance vessels [71][72][73]. Failure of remodeling of spiral arteries by extravillous trophoblast is seen in placentae from pregnancies complicated by pre-eclampsia or FGR and when there is an imbalance of proangiogenic (PlGF) and antiangiogenic factors [soluble fms-like tyrosine kinase 1 (sFlt-1)] [72]. Indeed the 2021 International Society for the Study of Hypertension in Pregnancy (ISSHP) [74] now includes angiogenic factors as a criterion to define uteroplacental dysfunction (placental abruption, PlGF <5 th centile for gestational age or sFlt-1/PlGF ratio >38, FGR, abnormal umbilical artery Doppler waveform analysis or intrauterine fetal death).
In a prospective cohort study [76], low plasma PlGF (<5 th percentile for gestational age) identified FGR infants and significant placental dysfunction on histopathological examination with sensitivity of 98.2% (95% CI: 90.5-99.9) and PPV of 58.5% (95% CI: 47.9-68.6), respectively. Low maternal PlGF outperformed gestational age, fetal abdominal circumference, and umbilical artery Doppler resistance indices in predicting FGR secondary to placental dysfunction. In another study, high sFlt-1 expression was present in 28% of placental tissue from pregnancies complicated by SGA/FGR without pre-eclampsia and in this group, 90% had abnormal umbilical Doppler and lower mean BW [77].
The sFlt-1/PlGF ratio is inversely correlated with BW [78,79] and a high ratio is present in pregnancies complicated by FGR [80]. Furthermore, although the sFlt-1/PlGF ratio is elevated regardless of the gestation at which FGR is diagnosed, early-onset FGR is associated with higher ratios compared with late-onset FGR, suggesting a possible lesser degree of placental dysfunction in the latter group [78,81]. A high sFlt-1/PlGF ratio also predicts a shorter time to delivery interval, which in turn is even more strongly correlated with the magnitude of daily increase of the ratio [82]. A recent study by Mitlid-Mork et al. [83] showed that compared with controls, women with pregnancies complicated by placental syndromes (pre-eclampsia and/or FGR) median maternal concentrations of PlGF (104 vs. 165 pg/ml) were significantly lower while sFlt-1 (6927 vs. 4371 pg/ml) and the sFlt-1/PlGF ratio (73.1 vs. 28.4) were significantly higher. In another study that evaluated 120 cases of early-onset FGR, 75% had an sFlt-1/PlGF ratio ≥85 with an associated probability of delivery within 1 week of diagnosis of 36%. In contrast, a ratio of <85 was associated with a >70% probability of prolongation of pregnancy for >4 weeks [84]. A more recent study of early-onset FGR demonstrated a negative predictive value (NPV) (using an sFlt-1/PlGF cutoff threshold of 38) of 100% (95% CI: 0.92-1.00) for delivery within 2 weeks of diagnosis and a NPV of 50% for delivery within 1 week if the ratio was >85 [85]. Gaccioli et al. [86] reported that using an EFW of <10 th centile for gestation and sFlt-1/PlGF ratio of >5.78 at 28 weeks resulted in a PLR of 41.1 (95% CI: 23.0-73.6) and PPV of 21.3% (95% CI: 11.6-35.8) for preterm birth of an SGA infant. Using a higher threshold sFlt-1/PlGF ratio of >38 at 36 weeks resulted in a PLR of 17.5 (95% CI: 11.8-25.9) for subsequent birth of an SGA infant associated with either maternal pre-eclampsia or perinatal morbidity or mortality [86]. Other observational studies have also reported similar associations between a high sFlt-1/PlGF ratio with SGA/FGR and a shorter duration to delivery interval [87,88]. A recent systematic review and meta-analysis (33 studies; 9426 women) showed that while PlGF, sFlt-1, and the sFlt-1/PlGF ratio showed promise for the prediction of adverse maternal and perinatal outcomes including SGA/FGR and time to delivery, PlGF was equivalent to the sFlt-1/PlGF ratio for predictive utility [89].
In a prospective study of 3953 singleton pregnancies at 35-37 weeks of gestation, Valino et al. showed that prediction of SGA (detection rate: 62.8%) was best achieved by maternal serum PlGF, ultrasound EFW, and UtA-PI [90]. However, compared with maternal serum PlGF, sFlt-1 does not provide significant independent prediction of SGA [90,91]. Another study by Valino et al. that screened 8268 singleton pregnancies at an earlier gestation of 30-34 weeks demonstrated that prediction of SGA using EFW, PlGF, sFlt-1, UtA-PI, umbilical artery pulsatility index, and middle cerebral artery pulsatility index resulted in a detection rates of 88% and 51% for birth of a preterm and term SGA infant, respectively [92].
In twin pregnancies, the sFlt-1/PlGF ratio measured in the second trimester is associated with increased odds for FGR (OR: 39.6, 95% CI: 6.31-248.17) [106]. However, as a standalone marker, PlGF does not appear to be sufficiently robust (sensitivity 27%) for the prediction of FGR for women with multiple pregnancy [107].

Soluble endoglin
Another placenta-derived antiangiogenic factor associated with placental dysfunction is soluble endoglin (sEng). sEng is a soluble transforming growth factor-β (TGF-β) coreceptor, which has been shown to be elevated in sera of women with pre-eclampsia and FGR [108][109][110]. An early small observational study (44 women) reported positive correlation between sEng and sFlt-1 concentrations (Pearson 0.653; P<0.05) with significantly higher sEng concentrations in FGR pregnancies compared with controls. However, concentrations of sEng were lower in FGR compared with pre-eclampsia pregnancies [111]. More recent study showed that sEng is strongly correlated with sFlt-1/PlGF ratio with higher concentrations observed in FGR (OR: 2.28, 95% Cl: 1.55-3.4 and 2.38, 95% CI: 1.64-3.44 for sEng and sFlt-1/PlGF, respectively) [112]. Another study, however, did not find any significant association between maternal concentrations of sEng and time of delivery in pregnancies complicated by FGR [113]. The current evidence for the utility of sEng for the prediction of SGA/FGR is limited.

Cell-free fetal DNA
Circulating cell-free fetal DNA (cffDNA) is used for aneuploidy screening, determination of fetal red cell antigen status, fetal sex, and screening for single-gene disorders [114][115][116]. cffDNA concentrations increase with gestational age and significantly higher levels are seen pregnancies complicated by placental dysfunction [117,121]. The data however from pregnancies complicated by FGR are conflicting, with some studies suggesting an increase [117][118][119] in cffDNA concentrations while others showing a decrease [120][121][122] compared with controls.
In a retrospective cohort study of 4317 singleton pregnancies [120], the fetal fraction was inversely correlated with MAP, UtA-PI, and positively associated with maternal PAPP-A and PlGF concentrations. A lower fetal fraction was associated with a higher risk of preterm FGR. Given the limited and inconsistent data regarding the relationship between maternal cffDNA concentrations and SGA/FGR, its utility as a reliable predictive marker remains unclear and further research is required [126][127][128].

MicroRNAs
MicroRNAs (miRNAs) are small nonprotein-coding, single-stranded RNA molecules of up to 19-25 nucleotides. They influence post-transcriptional gene expression and help regulate cell development, differentiation, proliferation, and apoptosis [129,130]. Because they are relatively stable and resistant to degradation by temperature and pH changes circulating miRNAs have potential as biomarkers for the prediction of adverse placenta-related outcomes [131].

Conclusions
Early prenatal identification of infants at high risk of SGA/FGR or adverse perinatal outcomes such as stillbirth, neonatal morbidity, and mortality is important because it potentially allows decisions regarding intensity of antenatal surveillance, timing of birth, model of maternity care, parental counselling, and co-ordination of neonatal resources to be made. Thus, the attraction of a simple and acceptable screening test early in pregnancy is obvious. However, there are several circulating biomarkers that are clearly associated with adverse outcomes, none have yet, either alone or in combination, been shown to be sufficiently reliable to be used in clinical practice [147]. Some, such as PlGF [148] and sFlt-1 [149] show the most promise but require further validation to determine their screening performance. More importantly, however, it is important to determine if a policy of screening for disorders related to placental dysfunction results in improvements in clinical outcomes.

Data Availability
All data are included within the main article.

Competing Interests
The authors declare that there are no competing interests associated with the manuscript.