Introduction

Preeclampsia (PE) is a multisystem disorder that occurs during pregnancy and is one of the main causes of maternal, fetal, and neonatal mortality, especially in low and middle-income countries1. PE is commonly associated with a sudden increase in blood pressure with concurrent development of a high level of protein in the urine (proteinuria), and can progress to eclampsia, in which seizures or coma may occur. From 1980–2010, the rates of PE increased from 3.4% to 3.8% for all pregnancies in the U.S., and these rates continue to rise, especially among certain subgroups2,3. According to a recently published epidemiological meta-analysis, the nation-specific incidence of PE varies from 4.6%, up to 8.2%4. Hypertensive disorders are thought to be the leading causing of maternal death in Latin America and the Caribbean5. Low maternal socioeconomic status, including limited educational attainment, physically demanding work environments, low household income, and underinsurance, is a strong risk factor for PE6. Furthermore, studies have shown an increase in in-hospital mortality for Black and Asian/Pacific Islander individuals with PE across all income groups, which is exacerbated due to their higher odds of developing strokes, kidney injuries, and heart failure7.

Limited interventions exist to prevent preeclampsia other than low-dose aspirin, which has been observed to result in a moderate risk reduction8. However modifiable risk factors have been shown to mitigate disease severity and adverse sequelae of PE4,9. Thus, early identification of those individuals most likely to develop PE could enable more surveillance, patient education, behavioral modification, and other opportunities to intervene and prevent PE development. Given the significant clinical and public health implications of PE, because of both the increasing incidence and the disproportionate impact on specific socioeconomic and racial/ethnic groups, it is important to identify efficient, accurate, early, and widely usable detection strategies.

In this review, we start by summarizing the current approaches used to detect and treat preeclampsia. We describe the rationale behind the need for these approaches to be improved. We discuss in detail the biomarkers that have emerged as candidates for PE screening across preeclampsia gestational stage. Lastly, we discuss the improved use of these existing and emerging preeclampsia biomarkers for point-of-care (PoC) testing, defined as testing in proximity to the patient.

Current methods for detecting and treating preeclampsia

Based on the guidelines from various global organizations, preeclampsia is generally defined as a condition that occurs in pregnancy or the postpartum period, characterized primarily by new-onset or worsening of pre-existing hypertension accompanied by evidence of end-organ injury. Specifically, the International Society for the Study of Hypertension in Pregnancy (ISSHP)10 defines preeclampsia as gestational hypertension accompanied by proteinuria or other maternal end-organ dysfunction after 20 weeks of gestation. This includes neurological complications, pulmonary edema, hematological complications, acute kidney injury (AKI), liver involvement, and uteroplacental dysfunction. Similarly, in the United Kingdom, the National Institute for Health and Care Excellence (NICE)11 identifies preeclampsia as new-onset hypertension after 20 weeks of pregnancy, with either proteinuria or other maternal organ dysfunction. This includes renal insufficiency, liver involvement, neurological complications, hematological complications, and uteroplacental dysfunction. In contrast, the American College of Obstetricians and Gynecologists (ACOG)12 no longer considers proteinuria a necessary sign for diagnosing preeclampsia. ACOG’s guidelines focus on elevated blood pressure with either proteinuria or other complications such as thrombocytopenia, renal insufficiency, impaired liver function, pulmonary edema, or cerebral/visual symptoms.

While definitions of preeclampsia historically focused primarily on the combination of high blood pressure and proteinuria, modern guidelines recognize a broader spectrum of symptoms and complications13. These include, but are not limited to, organ dysfunction (such as renal, liver, or neurological complications), hematological issues such as HELLP (Hemolysis, Elevated Liver enzymes and Low Platelets) syndrome14, and uteroplacental dysfunction. This evolution in the definition allows for a more comprehensive and sensitive approach to diagnosing and managing preeclampsia, with the ultimate aim of improving outcomes for both pregnant people and their neonates.

Most interventions for PE primarily address advanced presentations, incorporating the use of antihypertensive drugs to control severe hypertension and prevent maternal morbidity, corticosteroids to prevent neonatal morbidity, and preterm delivery if evidence of HELLP syndrome or other end-organ dysfunction arises15,16,17,18. Several preventive strategies with mixed efficacy, among them the administration of aspirin prior to 16-weeks’ gestation, have been suggested to potentially mitigate PE risk4,19. However, the effectiveness of these strategies hinges on the availability of screening technologies capable of detecting PE biomarkers at the earliest gestational stages, thereby enabling timely intervention20,21.

The ACOG defines a blood pressure of 120/80 mm Hg as healthy for a pregnant person, while a reading of between 120–129/ ≤ 80 mm Hg is classed as elevated, 130–139/80–89 mm Hg is representative of stage 1 hypertension, and new onset elevation ≥140/ ≥ 90 mm Hg on two occasions at least 4 h apart is associated with gestational hypertension or PE22. However, BP monitoring alone lacks sensitivity and specificity23. In combination with BP criteria, biomarker tests measure protein in urine, components of the blood24, and liver-related biomolecules (e.g., lactate dehydrogenase, aspartate transaminase, or alanine transaminase) to identify PE and/or the onset of HELLP syndrome25,26. For diagnosis of PE, the proteinuria measurement uses a semi-quantitative dipstick test ( ≥ 2 + ) or a 24 h urine collection ( ≥ 300 mg in 24 h) to assess the protein content in urine27,28,29. Since the concentration of creatinine in urine can remain consistent even in pregnant people with PE, the urine protein-to-creatinine (UPC) ratio is used to assess endothelial damage in the renal glomeruli vessels, with a spot UPC ratio ≥0.3 mg/dL indicative of PE in individuals without preexisting nephropathy27.

Challenges for point-of-care screening and management of preeclampsia

PoC testing enables clinical assessments to be conducted either at or in close proximity to the patient’s location. It can be used for diagnosis, screening, or monitoring, depending on the clinical scenario at hand. The primary objective of PoC is to shift away from lengthy multi-step laboratory processes and, in doing so, deliver faster results. This accelerated feedback allows for timely and decisive clinical interventions, leading to improved patient outcomes. PoC tests have found successful applications in diverse clinical areas such as glucose monitoring30,31, infectious disease detection32,33,34, and drug testing35,36.

The task of developing a PoC test specifically designed for the early detection of PE is undoubtedly complex. Traditional diagnostic methods are usually used when clear symptoms manifest, or when there is an existing or suspected medical condition that demands verification or exclusion. However, the fact that significant symptoms or other clinical evidence of PE might be absent during the early stages, particularly the first trimester, can render the traditional diagnostic approach ineffective until much later in the pregnancy. In this context, PoC tests for PE are better suited for screening or monitoring purposes. Here, the primary goal becomes the early detection and timely intervention, which can potentially mitigate the disease’s severity, prevent its onset, or provide insights into its progression and the efficacy of the interventions utilized. Moreover, it is to be expected that biomarker expression profiles or thresholds will differ based on the disease’s stage. This implies that any PoC testing designed for PE screening or monitoring must undergo thorough validation to target the appropriate biomarkers corresponding to distinct pathogenic stages. Unlike the more straightforward decision-making often seen in infectious disease detection, a PoC tool designed for PE management needs to continuously monitor a range of biomarkers. Each of these biomarkers might have a stage-specific threshold or profile, necessitating a more detailed and precise assessment.

Furthermore, the unique challenges posed by low-resource communities that are disproportionately affected by preeclampsia must be considered. In these settings, the lack of advanced analytical facilities and specialized technical personnel presents substantial hurdles. To be truly effective, PoC platforms must be designed with an inherent adaptability to such resource constraints. This means creating devices that are not only cost-effective but also user-friendly, requiring minimal training to operate. Additionally, these tests should be robust, capable of functioning in diverse environmental conditions, and not reliant on complex supply chains for reagents or parts. Emphasizing the integration of these considerations into the design and deployment of PoC platforms will ensure their viability and effectiveness in the communities that need them the most.

Approaches to improve detection

In light of the complexities surrounding PE detection often at later stages of pregnancy and the clinical challenges posed by the presence of severe features (i.e. development of BP ≥ 160/ ≥ 110 mm Hg on two occasions at least 4 h apart and/or serum biomarker abnormalities or clinical symptoms that signal end-organ dysfunction)37, early identification is key to improve clinical care. Recognizing PE at its onset not only optimizes opportunities for surveillance and the application of therapeutic interventions, but also significantly improves maternal and neonatal outcomes19. Two scenarios where PE screening tests may be useful38 are first trimester screening (mass screening) and as an diagnostic adjunct in suspected PE cases. First trimester screening involves screening an asymptomatic population to identify those at high risk of developing PE. This approach is primarily preventative and currently utilizes a combination of clinical risk factors, biomarkers, and imaging techniques (such as uterine artery Doppler flow velocity waveform analysis) to predict the likelihood of PE. The goal is to stratify antenatal care and determine which individuals might benefit from early interventions, such as prophylactic treatment with aspirin. The effectiveness of this screening is improved by integrating multiple variables into predictive algorithms. However, the cost-effectiveness and practicality of such comprehensive screening methods can vary between healthcare systems. Diagnostic adjunct screening pertains to individuals who present later in pregnancy with ambiguous symptoms suggestive of PE. In these cases, biomarkers are used to help estimate the risk of imminent PE. The PRAECIS Study39,40, whose findings have since undergone FDA clearance, is one example of this approach. It utilized biomarkers such as PlGF and the sFlt-1 to PlGF ratio to predict the risk of severe preeclampsia in the later stages of pregnancy. While both scenarios share the objective of improving preeclampsia screening tests, first trimester screening needs integration and cost-effective application of tests on a large scale, whilst diagnostic adjunct screening requires diagnostic tools with accurate prediction risk models that are accessible and practical for use in a variety of healthcare environments. Both avenues provide engineers with opportunities for innovation, and present specific scenarios for which next-generation tests could be beneficial.

Benefits of early preeclampsia detection

The emergence of medical tools capable of early PE detection before the third trimester offers a multitude of benefits. It sets the stage for enhanced screening, monitoring, and management, enabling timely detection and mitigation of potential complications for both pregnant people and their neonates. This early insight facilitates personalized education and counseling, empowering individuals with crucial knowledge on lifestyle modifications, dietary recommendations, and the early signs of PE for timely medical attention. It also lays the groundwork for meticulous planning regarding need for hospitalization and timing of delivery, significantly contributing to the safety and well-being of both the pregnant person and child. The ripple effect of early detection extends to better maternal and fetal outcomes by potentially curbing the progression of preeclampsia to its most severe forms and minimizing the risk of preterm birth and other adverse health outcomes. Moreover, it provides early-stage data that could be used to inform research into future treatment and preventive strategies. Lastly, the economic benefits of early diagnosis and intervention cannot be overstated, as it could markedly reduce healthcare costs by averting the high expenses required to manage severe preeclampsia and related complications.

In essence, medical tools adept at early preeclampsia screening are not just technological advancements, but invaluable assets in prenatal care, enabling a proactive approach to managing this condition, improving prognosis, and elevating the standard of care provided to both the pregnant person and child. The development of portable medical devices that can be applied at the PoC has the potential to significantly enhance all of the positive features associated with the ability to detect individuals who will clinically develop PE from the earliest stages of pregnancy. PoC screening further democratizes care especially in low-resource settings or remote areas where access to centralized laboratory facilities is limited, thus expanding the reach of early screening and management of PE.

Identification and Classification of preeclampsia biomarkers

PE’s etiology remains elusive, with its diverse origin pathways and changing stage-dependent traits complicating the identification and classification of biomarkers critical to its progression41,42. The 2017 guidelines issued by the American College of Cardiology (ACC), American Heart Association (AHA), and the Heart Failure Society of America (HFSA) apply a level-based classification system that considers the number of randomized clinical trials, quality of evidence, and meta-analyses of the trials to assign scores to different diagnostic recommendations for patient care43. We applied a similar approach to summarize understanding of PE biomarkers published in primary research papers (Web of Science), systematic reviews and meta-analyses (MEDLINE), and clinical trials (CENTRAL). Full details of the methods we used are detailed in the Supplementary Note. Supplementary Fig. 1 shows the workflow we used for the database queries. Based on a weighted classification relying on the research and clinical evidence, Table 1 provides a hierarchy matrix for the biomarkers and their level of evidence for PE pathogenesis.

Table 1 The table showcases our proposed hierarchy of biomarkers for preeclampsia, based on the robustness of evidence
Full size table

Analysis of the classification results indicated that biomarkers capable of detecting PE in the 1st, or 1st to 2nd trimesters are linked to six distinct pathogenic categories: epigenetic, genetic, angiogenic, renin-angiotensin-aldosterone system (RAAS), endocrine system, and inflammatory. This further highlights the intricate and multifactorial pathophysiology of the condition.

Established biomarkers

In this section we discuss the biomarkers we identified as established and review their potential for aiding the detection of PE. In Fig. 1, we have also classified the various pathways and how the biomarkers in these pathways can affect the pathophysiology of PE.

Fig. 1: Preeclampsia Biomarker Temporal Detection.
figure 1

Derived from Sub-Level I of categories A, B, and C from Table 1, this chart delineates the detection periods of various preeclampsia biomarkers throughout gestation. The x-axis represents gestational weeks (0–40, trimester 1:weeks 0–12; trimester 2: weeks 13–26, and trimester 3: weeks 27–40), while the y-axis lists specific biomarkers according to their Level A (established, blue text), Level B (emerging, green text) and Level C (promising, red text). Biomarkers are color-coded as follows: green for the 1st trimester, orange spanning the 1st to 2nd trimester, blue from the 2nd to 3rd trimester, and pink for the 3rd trimester. Notably, the expansive yellow bar for KIR-AA(maternal)/HLA-C(fetal) and progesterone receptor polymorphisms indicates genetic/epigenetic testing that could be applied across all pregnancy stages. At the side of the chart, biomarkers detectable in the 1st and 1st to 2nd trimesters are categorized by their respective pathogenic pathways. Abbreviations: TNF: tumor necrosis factor; IL: interleukin; HIF: Hypoxia-inducible factor; HLA: human leukocyte antigen; PAPP-A: pregnancy associated plasma protein A; KIR-AA: killer immunoglobulin-like receptor AA genotype;; miR: microRNA; sFlt: soluble Fms like tyrosine; PlGF: placental growth factor; sEng: Soluble endoglin; PRPs: progesterone receptor (PR) polymorphisms; NO: nitric oxide; PP13: Placental protein 13.

Full size image

sFlt-1 and PlGF (Angiogenic, 1st Trimester)

Studies have shown that the cellular shear stress of the syncytiotrophoblast placental barrier, which occurs during the last 8–10 weeks of pregnancy, leads to biochemical changes in levels of sFlt-1 and PlGF in healthy pregnancies44,45. sFlt-1 and PlGF are pivotal in the early stages of PE and are identifiable in both the circulatory system and placental tissues. PlGF is a pro-angiogenic factor that enhances placental vascularization and trophoblast growth46,47, and has a high affinity for fms-like tyrosine kinase-1 (Flt-1) sites on cell membranes48,49. However, sFlt-1 competes with Flt-1 and acts as an anti-angiogenic factor, by binding with vascular endothelial growth factor (VEGF) thus reducing its free circulation50. Notably, low VEGF concentrations have also been associated with endothelial and glomeruli damage in PE51. As proteins, they are suitable for immunoassay tests with PoC devices that utilize biorecognition elements, such as antibodies or nucleic acid aptamers. The sFlt-1 to PlGF ratio is specific (84.2–97%) and sensitive (85–95%) at ruling out PE development, so it is a good candidate for PoC PE screenings52,53. In a two-stage cohort study of 1050 patients, a cut-off ratio of sFlt-1/PlGF of 38 was used for ruling out the development of early onset preeclampsia52. While a ratio higher than 38 had a positive predictive value of 36.7%, a ratio lower than 38 had a negative predictive value of 99.3%, indicating its value for a rule-out diagnosis.

Insulin (Endocrine System, 1st Trimester)

The relationship between insulin resistance and PE has been explored in several studies, and the connection often centers on the broader metabolic changes that can occur during pregnancy54,55. Pregnancy itself can lead to increased insulin resistance, which occurs when cells in muscles, fat, and the liver do not respond well to insulin and cannot easily take up glucose from the blood, leading the pancreas to make more insulin. Some studies have found that pregnant people with PE may have a higher degree of insulin resistance compared to those without the condition56,57. As previously mentioned, one of the hallmark features of PE is endothelial dysfunction, in which the inner lining of the small blood vessels fails to function normally. Insulin resistance can exacerbate endothelial dysfunction, and this can potentially worsen the features of PE57.

Preliminary small-scale clinical studies indicate a possible link between insulin resistance and preeclampsia in pregnancy. For example, in a sample of 675 pregnant people across all trimesters higher insulin resistance, measured by Homeostatic Model Assessment (HOMA-IR), was found in individuals at risk of preeclampsia compared to those without (P < 0.001, P = 0.021), (Table S5)58. A study of 829 pregnant people also observed a correlation between preeclampsia and insulin resistance by using fasting insulin homeostasis model assessment, quantitative insulin sensitivity check index, and measuring oral glucose insulin sensitivity59. They reported potential screening accuracy with 79% sensitivity in early pregnancy (16–20 weeks) and 83% in late pregnancy (26–30 weeks), along with 97% specificity. However, the limited case numbers in these studies suggest that the findings might not be statistically robust enough to develop screening or diagnostic tools without further large-scale validation studies.

While changes in insulin resistance can be observed in PE, it is not used as a primary diagnostic tool for the condition. Thus, even though there is a connection between insulin concentrations, insulin resistance, and metabolic dysregulation in the setting of PE, relying on insulin and associated markers of insulin resistance solely for detection seems impractical. However, understanding insulin resistance in the context of PE might help predict or manage the condition, especially in pregnant people who already have risk factors for insulin resistance, including those with polycystic ovary syndrome (PCOS) or obesity60. Also, a better understanding might reveal underlying mechanisms and suggest possible pathways that can be used for treatment and intervention.

TNF-α (Inflammatory, 1st Trimester)

Some theories suggest that PE arises from a maladaptive immune response to the placenta61,62. Tumor necrosis factor-α (TNF-α), one component of the immune response, might be involved in this maladaptation, contributing to the improper invasion of placental trophoblasts into the maternal uterus. Consistent with the abnormality in immune cell population, inflammatory cytokines exhibit dominance over anti-inflammatory cytokines in PE. Increased serum concentrations of pro-inflammatory cytokines, including interleukin (IL)-6, IL-8,TNF-α, and IL-18, can be observed while serum concentrations of anti-inflammatory cytokines, such as IL-10 and TGF-β, reduce in the 2nd and 3rd trimester63,64,65. Although systemic inflammation is commonly found in pregnancy, this inflammatory response is exacerbated in patients with PE66,67,68. Studies suggest that this inflammation may be related to the overexpression and secretion of exosomes from syncytiotrophoblasts69,70. Exosomes stimulate the production of inflammatory cytokines, including IL-1871,72. Increases in IL-18 promote expression of Interferon-γ (IFN-γ), resulting in a decrease in immunotolerance in individuals with preeclampsia73. Due to the evident role of TNF- α in PE, some researchers have explored targeting this cytokine as a therapeutic strategy. However, further studies are needed to assess the safety and efficacy of such interventions.

Several studies have investigated the relationship between serum TNF-α levels and preeclampsia. For instance, a cross-sectional study of 90 pregnant people in Lagos showed higher TNF-α levels in those with preeclampsia and suggested a screening range of 15.6 to 26.4 ng/mL in the third trimester74. Additionally, a study found in their cohort of 90 pregnant peoples, which included 10 individuals with preeclampsia, that a TNF-α threshold of 10.13 pg/mL could potentially screen for preeclampsia in the third trimester with 90% sensitivity and 78% specificity75. However, TNF-α polymorphism, which can influence expression levels76, was not accounted for in these studies. Without considering ethnic variability, results may be skewed. Future validation studies on TNF-α as a preeclampsia marker should address these genetic factors to minimize bias as well as consider larger patient sample sizes like for greater confidence and higher predictive values.

Despite studies indicating elevated TNF-α concentrations in the serum of individuals with PE, several considerations pose a challenge to using it as a biomarker. Since TNF-α concentrations can be elevated in a variety of conditions in pregnancy, not just PE, TNF-α does not offer enough specificity to be used as a standalone marker. Additionally, TNF-α concentrations change dynamically throughout pregnancy and may be influenced by external factors further complicating its use as a reliable detection tool. Yet, as with numerous identified PE biomarkers, when combined with other markers and clinical evaluations, it could provide significant insights into the disease’s presence and intensity. Nonetheless, its definitive utility for detecting PE necessitates more in-depth research and validation.

Renin (RAAS System, 2nd-3rd Trimester)

The RAAS system is closely related to the balancing of angiogenic factors and plays an important role in placental development and functioning. It exists at the uteroplacental interface, and its downregulation has been found to limit uteroplacental blood flow77,78. Angiotensin (Ang) II is a vasoconstrictor that enhances vascular flow within the fetoplacental region by binding to angiotensin II receptors present on the surface of smooth muscle cells79,80. Ang (1–7) causes vasodilation and inhibits localized inflammation in the placental region81. Aldosterone has been shown to enhance PlGF expression and trophoblast cell proliferation, therefore its downregulation restricts placental growth and trophoblast invasion82,83. During a healthy pregnancy, the Ang II levels in serum are significantly increased during the second and third trimesters (29-34 weeks’ gestation)84. In PE, however, patients have been shown to have reduced levels of renin, Ang I, Ang II, and Ang(1-7) in their serum85,86. Additional studies have shown that similar to renin and Ang I/II, the serum level of aldosterone also decreases in individuals with preeclampsia87. However, while reduced Ang II levels in maternal serum are commonly associated with proteinuria and HELLP, other studies suggest that reduced RAAS factors are not a cause but rather a manifestation of late-onset PE, and urinary aldosterone is not correlated with PE but its reduced levels are correlated with low placental weight and body weight of the fetus, both of which are common clinical findings associated with PE88. Since decreased adrenal aldosterone in serum is associated with a decreased renin production, it can be used as an indirect biomarker for preeclampsia89, but its use is controversial, as another study reports increased levels in the 1st trimester90.

A longitudinal study explored the relationship between the renin-angiotensin-aldosterone system (RAAS) and preeclampsia, including 25 pregnant people with chronic hypertension91. They observed that patients with pre-existing chronic hypertension with superimposed preeclampsia had a significant decrease in plasma renin activity (PRA) from the first to the third trimester, which was not initially evident in normal hypertension cases. A more comprehensive study examined RAAS profiles in 108 patients with chronic hypertension, considering age, gestational weeks, and race among other factors92. Their findings indicated a consistently lower PRA in patients with superimposed preeclampsia at weeks 28 and 36. These studies suggest a potential diagnostic role for renin in preeclampsia.

Whilst the RAAS system is connected to the development of PE, solely relying on renin levels for diagnosis might be misleading. Renin levels can fluctuate due to various factors including salt consumption, body position, daily cycles, and the functioning of the kidneys. This variability might make it challenging to establish a standardized threshold for diagnosing PE based solely on renin levels. Whilst development of a PoC test for renin to aid in diagnosing PE is theoretically possible, the practicality and utility of such a test would require careful evaluation. It might be more practical to develop a PoC test that incorporates renin based on its added detection value, however that value still needs to be demonstrated.

Emerging biomarkers

We classify emerging biomarkers as those identified as connected to PE after the year 2000. Details of the classification categories can be found in Suppl. Figures 1–2, Suppl. Data 3, Table 1, and Figs. 1–3. We have highlighted two such biomarkers, sEng and PP13 that hold potential for PoC platforms. Biologically, both biomarkers play a key role in placental growth. sEng suppresses placental blood flow by antagonizing TGF-β193,94,95 while PP13 is involved in the trophoblast invasion in placenta96. Moreover, various studies show that including sEng or PP13 alongside assessment of the sFlt-1/PlGF ratio strengthens the screening accuracy for PE97.

Fig. 2: Pathogenic pathways and the role of biomarkers in preeclampsia development.
figure 2

Biomarkers in the starred categories have commercial devices, with violet indicating biomarkers cleared by the FDA and CE marked, and blue indicating CE marked biomarkers. Overall, the biomarkers in square categories fall under Level A, with green indicating Sub-Level II, and yellow indicating Sub-Level III classifications. The biomarkers in circle categories fall under Levels B (orange) and C (red) respectively. Details for each biomarker classification can be found in Supplementary Data 3. Created with BioRender.com

Full size image
Fig. 3: OFET for detection of sFlt-1. Bottom Left.
figure 3

A schematic illustrating the structure of a Polyisoindigo-based organic semiconductor (pII2T-Si, chemical structure provided). pII2T-Si encapsulated the two drain and source Au electrode by spin coating with silicon oxide as the dielectric layer and silicon as the Gate. Top Left: Compared to typical FET devices, OFET combines the idea of ELISA. OEFT sensor functionalized with specific antibodies will capture sFlt-1. Sequentially, the biotinylated detector antibodies will bind on the top to form a sandwich complex. With the addition of streptavidin-conjugated glucose oxidase, glucose will be converted into gluconic acid. The more the sFlt-1 are captured, more gluconic acid formed to decrease the pH value as more glucose oxidase are available. The decrease in pH value of the solution will further increase the current flow of the device which can be an indicator corresponding to the amount of sFlt-1 captured. Au electrode will be the ‘source’ and ‘drain’ for signal transduction. Bottom Right: a graph showing normalized intensity versus time along with various addition of different concentrations of sFlt-1 (0–40 ng mL−1), demonstrating a successful real-time quantification of sFlt-1 levels in the plasma. Top Right: Organic Field-effect transistors developed for the detection of sFlt-1 utilize an antibody based catalytic sandwich-assay to quantify sFlt-1 across a range of 1.25–10 ng/mL121.

Full size image

sEng (Angiogenic, 1st to 2nd Trimester)

Soluble endoglin (sEng) is another anti-angiogenic factor that has been implicated in the pathogenesis of PE. Endoglin is a co-receptor for transforming growth factor-beta (TGF-β) and is involved in vascular development and remodeling93. In PE, an elevated level of the soluble form of endoglin, sEng, has been observed, and it is believed to

play several roles in the disease’s onset and progression. sEng contributes to the dysfunction of the vascular endothelium, which is a key feature of PE95. The endothelial dysfunction results in increased vascular permeability, vasoconstriction, and pro-inflammatory states, all of which contribute to the clinical features of PE, such as hypertension. sEng, in combination with other factors like sFlt-1, acts as an antagonist to pro-angiogenic factors, primarily vascular endothelial growth factor (VEGF) and placental growth factor (PlGF)95. By doing so, it disrupts the balance between pro-angiogenic and anti-angiogenic factors, leading to impaired placental angiogenesis. Proper vascular development in the placenta is crucial for fetal nutrient and oxygen exchange, and its impairment can lead to placental insufficiency. Furthermore, elevated levels of sEng are associated with systemic inflammation, which is a characteristic of PE. The pro-inflammatory environment can further exacerbate endothelial dysfunction. Elevated levels of sEng have been observed before the clinical onset of PE, suggesting its potential as a predictive biomarker for the disease.

A study on 1002 pregnant people found significant correlations between sEng levels and key clinical parameters of preeclampsia, such as systolic blood pressure (rs = 0.37, P < 0.001), diastolic blood pressure (rs = 0.8, P < 0.001), and the urinary protein: creatinine ratio (rs = 0.35, P < 0.001)98. These findings suggest sEng as a positively correlated biomarker with preeclampsia’s common features, indicating its potential for diagnostic use. Supporting this finding, a longitudinal study was conducted on 471 pregnant people, recording plasma sEng levels99. They reported a notable diagnostic performance of sEng, with an area under the receiver operating characteristic (ROC) curve of 0.81 (95% CI, 0.62–0.99) at 19–22 weeks. Nonetheless, both studies noted an inverse relationship between sEng levels and gestational age, implying that reliance on sEng alone as a biomarker could be problematic. However, in combination with established biomarkers such as sFlt-1 and PlGF, there is the possibility that sEng can help in risk stratification and early detection of pregnant people likely to develop PE.

PP13 (Inflammatory, 1st to 2nd Trimester)

Placental protein 13 (PP13), also known as Galectin-13, is a protein expressed by the placental syncytiotrophoblasts and is released into the maternal circulation96. Galectin-13 serum concentrations vary throughout pregnancy, with a characteristic pattern of being relatively high in the first trimester, decreasing in the second trimester, and potentially rising again in the third trimester in cases of PE, while remaining low in uncomplicated pregnancies. PP13 is proposed to play a role in regulating the invasion of trophoblasts into the maternal decidua during early pregnancy96. Adequate trophoblast invasion is critical for proper placental development and function. Abnormal trophoblast invasion is a hallmark of PE and is associated with compromised uteroplacental blood flow. Some studies suggest that lack of PP13 might contribute to endothelial dysfunction96,100, a central feature of PE. This dysfunction could be due to the direct effects of PP13 or because of placental insufficiency related to improper trophoblast invasion. PP13 might also be involved in modulating the maternal immune response to the developing placenta96. An imbalance in maternal immune tolerance can contribute to PE’s pathogenesis.

In a 2021 meta-analysis, data was synthesized from 17 longitudinal studies encompassing 40,474 cases to assess the diagnostic capability of PP13 for PE detection101. They found that the overall sensitivity and specificity of PP13 as a predictive marker for PE were 62% and 84%, respectively, both with a 95% confidence interval (CI). Importantly, the diagnostic odds ratio was determined to be nine, with a 95% CI. The findings suggest that low PP13 levels in early pregnancy could predict an increased PE risk. Nonetheless, due to individual variations in PP13 levels and potential alterations in other placental conditions, its effectiveness as a standalone predictor for PE is limited. Consequently, additional studies and validations are warranted before PP13 can be recommended for routine PE screening.

KIR-AA and HLA-C (Genetic, all trimesters)

Genotyping maternal immune cell receptors and fetal ligands will elucidate the immunotolerance of the maternal body towards fetal cells during pregnancy102. Killer cell immunoglobulin-like receptors (KIRs) on maternal immune cells interact with the human leukocyte antigen (HLA) receptors on fetal trophoblasts. During a healthy pregnancy, the maternal immune system undergoes specific adaptations to tolerate and support the semi-allogeneic fetus (which carries half of the father’s genetic material, making it foreign to the pregnant person). Trophoblasts, which are specialized cells of the placenta, invade the maternal uterus lining, facilitating nutrient and gas exchange between the pregnant person and fetus. Maternal natural killer (NK) cells, which are part of the innate immune system, express KIRs on their surface103. These KIRs can either activate or inhibit NK cell function, depending on which specific KIR genes are present and expressed. The invading trophoblasts express specific HLA molecules on their surface, notably HLA-C, which can be recognized by the KIRs on maternal NK cells. The interaction between maternal KIRs and fetal HLA-C molecules is thought to modulate the depth of trophoblast invasion into the maternal tissue. Proper trophoblast invasion is crucial for the remodeling of maternal spiral arteries, ensuring adequate blood flow to the placenta and, consequently, the fetus. PE is associated with inadequate trophoblast invasion and poor spiral artery remodeling. Studies have suggested that certain combinations of genotypes of maternal KIRs and fetal HLA-C molecules increase the risk of PE104,105. For instance, when maternal NK cells have activating KIRs and the fetal trophoblasts express specific HLA-C molecules that bind strongly to these KIRs, there might be an imbalance leading to shallow trophoblast invasion, poor arterial remodeling, and increased risk of PE. There are several dysfunctional variants of both receptors which eventually reduce the immune tolerance and prohibit trophoblast invasion. Two studies have reported that impairment of both receptors may result in PE104,106.

A longitudinal study of 401 pregnant people, of whom 201 developed preeclampsia104, found a 35% increase in the frequency of the maternal KIR-AA allele among patients with preeclampsia (p = 0.038), a figure that rose in the presence of the fetal HLA-C2 allele (p = 0.001). However, the study also highlighted possible discrepancies due to an inverse correlation between KIR-AA and HLA-C allele frequencies across ethnic groups, as well as differences in KIR phenotypes among individuals with the same genotype. To affirm the diagnostic value of these genetic markers, future investigations should address these potential sources of error by ensuring control of confounding factors, recruiting ethnically diverse participants, and seeking to reproduce results in varied clinical contexts. This approach would help in building a comprehensive and universally applicable profile of these biomarkers’ accuracy and utility in PE detection107.

If further validation through more comprehensive studies is achieved the development of a PoC test to monitor the KIR-HLA-C relationship for PE detection would be a step forward in rapid and accessible diagnostics. A blood sample from the pregnant person would be needed to analyze the presence and type of KIRs on maternal NK cells. For fetal HLA-C determination, cell-free fetal DNA from maternal plasma or other non-invasive sources could be used. Genotyping would be required to identify the presence of activating or inhibitory KIR genes in the maternal genome and to determine the fetal HLA-C allele type. Rapid PCR techniques could potentially be employed to quickly amplify and detect specific KIR and HLA-C genes. Alternatively, immunoassays could be developed to detect specific interactions between KIRs on maternal NK cells and HLA-C on trophoblast-derived vesicles or particles in maternal blood. If a certain pattern of interaction (or lack thereof) is indicative of PE, it could be used as a direct detection method. Once the types of KIRs and HLA-C are identified, the results would be analyzed based on established risk profiles. For example, if a pregnant person possesses a set of KIR genes and the fetus expresses HLA-C molecules known to increase PE risk when combined, the test will flag that preeclampsia is at a higher risk of occurring during this pregnancy.

Promising biomarkers

Promising biomarkers are still in their infancy as they have only undergone investigation for the role in the pathophysiology of PE after 2010. We discuss some genetic and epigenetic biomarkers in this category, which show potential for early-stage development of the disease.

Progesterone Receptor Polymorphisms (PRPs) (Genetic, all trimesters)

Polymorphisms are variations in a particular DNA sequence in a population. When these variations occur in genes encoding specific proteins or receptors, they can influence how those proteins or receptors function. Progesterone (PR) is a crucial hormone during pregnancy. It prepares the uterus lining for a fertilized egg to implant and supports the growth of the placenta and the fetus. The progesterone receptor is a molecule to which progesterone binds to exert its effects. Proper placental development and function are crucial for a healthy pregnancy. Given that PR and its receptor play a vital role in supporting placental development, any dysfunction or variation in PR or its receptor, such as those caused by polymorphisms, could potentially influence the risk or development of PE. Several studies have investigated the association between PRPs and the risk of PE. Some have suggested that certain PR gene variants might be linked with an increased risk of the disease108,109, while others have found no such association110. However, because PE has a multifactorial origin, with multiple genes and environmental factors contributing to its onset, PRPs might be just one of many genetic factors influencing PE development. In summary, while PRPs have been studied in relation to PE, the exact nature and significance of their relationship remain an area of ongoing research.

Before using such genetic markers in clinical practice to detect or predict PE, robust evidence from large-scale longitudinal or human cohort studies would be required. Detecting PRPs typically involves molecular genetic techniques. The first step in detecting polymorphisms is to obtain a DNA sample, usually from a blood sample or cheek swab. PCR is then used to amplify a specific region of DNA, in this case, the region surrounding the PR gene or the specific site of the known polymorphism. After amplifying the DNA region of interest, DNA sequencing is required to identify previously unknown mutations or polymorphisms. Alternatively, genotyping can be used to identify the presence of specific known mutations or polymorphisms. Common genotyping methods include restriction fragment length polymorphism analysis, allele-specific PCR, and high-resolution melting analysis. While there are rapid and portable PCR machines and some genetic PoC tests for specific conditions or pathogens, a PoC test for PRPs is not standard so developing such a test would involve a considerable amount of effort.

miR-17/20a/20b and miR-125b (Epigenetic, 1st Trimester)

Non-coding, short-sequence, microRNAs (miRNAs) provide epigenetic control in many systems111,112,113. Before miRNAs begin to circulate, they are first cleaved in the cytoplasm rendering them functional to operate as gene enhancers or silencers114. miRNAs are capable of binding with messenger RNA (mRNA) and inhibiting or enhancing the expression level of the mRNA after binding with an RNA inference silencing complex (RISC)114. They are multifunctional and can potentially act as early biomarkers of PE (e.g., miR-17 has been discovered to be upregulated in individuals with PE compared to healthy pregnant people)115. An upregulation of miRNAs-17/20a/20b is found in patients with PE116 The serum levels of miR-31, miR-155, and miR-214 have also been shown to be elevated in individuals with PE117. Upregulated levels of these three miRNAs promotes the expression of cytokine-inducible nitric oxide synthase (iNOS) and suppresses the expression of eNOS117. When the iNOS is higher than eNOS, the overexpression of inflammatory cytokines can increase NO production (a vasodilator), which increases blood pressure and can cause hypertension, as is commonly observed in individuals with PE118.

While non-coding, cell-free RNA may be vulnerable towards RNases in the blood, the specific secondary structure of miRNAs increases their resistance against RNases. Due to the low concentrations of miRNAs ( ~ picomolar, pM) present in maternal serum, their detection is often reliant on testing methods capable of amplifying the targets to detectable levels. However, the variable degradation rates of different RNA sequences in serum make it challenging to stratify the sample appropriately. Compared to genetic materials, miRNAs and non-coding cell-free RNAs are more accessible as targets in PoC assays. Nevertheless, additional extraction steps from cells or exosomes might be needed, along with the incorporation of amplification techniques, which could further complicate the testing process. Moreover, the upregulation or downregulation of small groups of these cell-free RNAs might not serve as reliable biomarkers. In a PoC assay, capturing at least partial transcriptomes or expression profiles is more desirable when using them as biomarkers for detecting PE. Moreover, further human cohort studies are required to validate and characterize the performance of this biomarker for preeclampsia, which necessitates further research.

Availability and development of point-of-care devices for PE detection

Commercial PoC devices for PE detection

As mentioned previously, sFlt-1 and PlGF can be used to help detect PE early in pregnancy. Given that both are detectable in maternal serum, companies including Roche, Abbott Rapid Diagnostics (formally Alere), and Perkin Elmer have developed assays for quantifying these biomarkers from a blood sample.

The Elecsys assay from Roche captures sFlt-1 and PlGF from patients’plasma using magnetic microparticles which are then further concentrated onto an electrode. The magnetic microparticles are coated with streptavidin and further functionalized with biotinylated, monoclonal antibodies and ruthenium-complex-labeled antibodies. After the removal of the plasma, an external voltage is applied to the electrode to induce a chemiluminescence signal. A photomultiplier is used to convert the light signal from the assay back to an electric signal for final data analysis. The detection ranges for sFlt-1 and PlGF are 10–85,000 pg/mL and 3–10,000 pg/mL, respectively. This Elecsys system is capable of detecting PE between weeks 20–33 of pregnancy and can also provide a PE prediction (serving as a reference and not a diagnosis) from week 24 to week 36.

Triage assays from Abbott Rapid Diagnostics capture PlGF from plasma using murine monoclonal antibodies labelled with fluorescent dyes. The plasma sample is then deposited onto a lateral flow platform containing the labelled antibodies. PlGF is recognized by the antibodies immobilized on the test line, which are assessed using fluorescence spectroscopy. The test takes a total of 15 min to complete and has a PlGF detection range of 12–3000 pg/mL. PlGF levels <100 pg/mL are associated with a high likelihood of the patient having PE.

Perkin Elmer has developed a system called DELFIA Xpress (Dissociation Enhanced Lanthanide Fluorescent Immuno Assay) for multiplexed detection of pregnancy-associated plasma protein A (PAPP-A), free beta-human chorionic gonadotropin (ß-hCG), PlGF, and alpha-fetoprotein (AFP). To quantify PlGF, DELFIA Xpress uses a sandwich assay that incorporates europium-labeled antibodies. The serum sample, buffer, and labeled antibodies are added to a vial containing secondary capture antibodies. This test allows the detection of PlGF across a concentration ranging from 1.9–4000 pg/mL and can be administered after the 20-week gestation period119. For PAPP-A detection, lanthanide labelled monoclonal antibodies are initially bound to the target and are then used as part of an indirect assay which utilizes the two sites of the PAPP-A/proMBP complex. The detection of PAPP-A ranges from 1.2–10,000 mIU/L119. A summary table of the commercially available devices for PE detection is given in Supplementary Data 2.

Emerging development of PoC device for PE in research

To further reduce the lower limit of detection for sFlt-1 and PlGF, the use of sensors that incorporate field-effect transistors (FET) is being investigated. In 2021, a new diagnostic platform consisting of a plasma separator, a nanoribbon, FET sensor, and a customized Ag/AgCl micro pellet electrode was developed to measure PlGF levels120. The separator and nanoribbon remove the impurities in the plasma, the FET is functionalized with antibodies specific to the recognition of PlGF, and the Ag/AgCl electrode acts as the source and drain for signal transduction. The test can detect 0.06 pg/mL of PlGF in 40 mins using only 30 μL of sample120. A similar approach was developed to measure sFlt-1, in which an organic field-effect transistor (OFET) was configured to capture and quantify the target121. This method has a similar targeting mechanism to an Enzyme-Linked Immunosorbent Assay (ELISA) whereby a biotin functionalized detection antibody binds directly to an antibody-immobilized sFlt-1 target, as summarized in Fig. 3. Streptavidin-conjugated glucose oxidase is then added to catalyze free glucose within the system into gluconic acid, causing a decrease in pH. SFlt-1 can be quantified across a range of 1.25–10 ng/mL by measuring changes in current flow induced by the increased acidity.

To detect misfolding proteins in the urine of individuals with PE, a paper-based lateral flow test, as shown in Fig. 4, has been developed by leveraging an azo-based dye, congo red122, to specifically bind proteinaceous deposits that are similar to amyloid. Despite its non-quantitative output, this was a promising platform as the mere qualitative detection of amyloid presence was able to achieve high sensitivity (80.2%), specificity (89.2%), and accuracy (86.7%)123. Specific details regarding PoC devices currently under development are listed in Table S7122.

Fig. 4: Congo red detection of amyloid-like proteins.
figure 4

The congo red paper test can detect amyloid-like proteins in the urine of pregnant women, via congophilia123,141. Specifically, excess amyloid-like proteins promote the intercalation of congo red, reducing the availability and subsequent binding of the dye on the cellulose fibers. The size and distribution of the red spot is used to estimate the concentration of misfolded proteins. The lower left image shows the actual assay. The lower right image shows an example of actual test results and their interpretation.

Full size image

A proof-of-concept study used surface-enhanced Raman scattering (SERS) in a sandwich assay consisting of silver nanoparticles (AgNPs) modified with 2 sets of nucleic acid sequences probes to target miR-17 present in bovine serum, as shown in Fig. 5124. When miR-17 is captured by the nanoparticles, it initiates the formation of nanoassemblies that are responsible for the generation of high intensity SERS signals. This method can detect miR-17 across a linear range from 1 pM to 1 nM.

Fig. 5: Surface-enhanced Raman Scattering for targeting miR-17.
figure 5

The SERS-based assay for the quantification of miR-17 uses silver nanoparticle probes modified with split DNA strands complementary to miR-17 sequence. A typical SERS detection can be achieved by functionalization of plasmonic nanoparticles with Raman dyes (Malachite green in this study) and DNA sequences or antibodies that conjugate with the targets. The mechanism is based on the ‘hotspots’ formation between the plasmonic nanoparticles when aggregation occurs after the bind of targets. The Raman signal at the ‘hotspot’ region can be amplified because of the crosstalk between the electromagnetic field and charge transfer between two nanoparticles142. DNA sequences functionalized on AgNPs can hybridize with the target hsa-miR-17-5p from 3’ to 5’ end to form a triplex structure without overhung base pair. In the presence of hsa-miR-17-5p, the SERS-active, DNA-functionalized AgNPs will aggregate. Therefore, the SERS output from the malachite green reporter significantly increases compared to the absence of hsa-miR-17-5p due to plasmonic coupling between the nanoparticles caused by their hybridization to the miRNA target. Using this method miR-17 can be detected across a range of 1 pM to 1 nM (n = 5). The error bars represent the standard deviation calculated from five spectra, with each spectrum being the average of 10 consecutive 1 second scans124. Created with BioRender.com

Full size image

Commercial tests such as the Lumella125, which is detailed in Fig. 6, leverages a similar binding of the nanoparticles to the target biomarker GlyFn, which is captured by biorecognition elements immobilized on the test line of a lateral flow strip.

Fig. 6: Commercial lateral flow approaches for detecting glycosylated fibronectin.
figure 6

Lumella is a commercial POC test developed by Diabetomics Inc for targeting glycosylated fibronectin (GlyFn)127,143. The top images show the actual device (left) and machine that reads the cartridge (right). The lower diagram explains how the device works: Diluted fingerstick whole blood is applied to lateral flow testing strips containing gold nanoparticles coated with an anti-GlyFn monoclonal antibody. Capture of colloidal gold-fibronectin complexes at the test line by another anti-GlyFn monoclonal antibody results in the formation of a reddish-purple line. Based on the depth of color at the test line the reader is able to interpret whether the GlyFn levels are in the normal (50–350 µg/mL), positive (351–600 µg/mL, increased risk of preeclampsia) or high positive ( > 600 µg/mL, increased risk of severe preeclampsia) ranges. Created with BioRender.com

Full size image

Regulatory considerations

The Roche Elecsys measurement of sFlt-1, PlGF, and their ratios, was granted the Conformite Europeenne (CE) Mark in Europe before 201453, and became widely adapted for ruling out the short-term absence of preeclampsia after the PROGNOSIS trial (Prediction of Short-Term Outcome in Pregnant Women with Suspected Preeclampsia Study)52. In contrast, Thermo Fisher’s B·R·A·H·M·S sFlt-1 KRYPTOR and B·R·A·H·M·S PlGF plus KRYPTOR devices were granted clearance by the FDA in May 2023 for the US126. The Thermo Fisher devices were validated for risk stratification for short-term development of severe PE (sPE) in the PRAECIS study (PE Risk Assessment: Evaluation of Cut-offs to Improve Stratification)39. The equivalent Roche Elecsys devices however, have still not received FDA clearance. This indicates that the statistical efficacy of the claims, after appropriate data stratification of samples, and accounting for confounding factors like gestational age, patient age, ethnic backgrounds, and other potential blood results, can impact the regulatory clearance process in the US.

Securing a CE mark requires manufacturers to submit pre-submission reviews of predicate devices and a post-market clinical follow-up, instead of a full clinical trial. This could explain why PerkinElmer’s DELFIA® Xpress system for sFlt-1 and PlGF 1-2-3 have received the CE mark for early detection and short-term prediction of PE in Europe.

In addition, detection methodologies can also affect the FDA clearance process. For instance, Lumella’s test for GlyFn uses a semi-quantitative output across all trimesters of pregnancy127, and PerkinElmer’s DELFIA® Xpress system for PP13, AFP, PAPP-A, and free ß-hCG use multiples of median (MoM) to estimate risk for early-onset of sPE128,129. These platforms have also received the CE mark, but have not been cleared by the FDA.

Thus, in the US, not only is foundational research needed for rigorous quantification of biomarkers related to PE, but the statistical efficacy of claims also needs to be stratified to accurately identify independent variables, which can then be used as bases for further clinical trials. Globally though, most commercial devices for PE utilize protein-based biomarkers that are powerful representatives of downstream PE progression. However, it would be worthwhile to investigate the add-on value of nucleic acid biomarkers such as exosomal material, cell-free DNA and RNA, and genetic and epigenetic factors, by including them in clinical trials and post-market follow-ups. Such additive biomarkers can identify risk profiles, predict upstream PE development, and prove valuable for robust prediction algorithms in personalized medicine.

Clinical considerations for development of PoC screening tests for PE

When considering development, and future implementation, of clinically useful and widely adopted PoC screening tests aimed at detection of preclinical PE in early pregnancy prior to the development of clinical manifestations of the disease later in gestation, several criteria must be met. First, the disease in question, in this case PE, must constitute a significant public health problem that carries risk of morbidity and mortality among a significant subset of the population, and have readily available treatment and management strategies that could improve clinical outcomes with early detection on a positive screening test130. Earlier in this review, we highlighted the significant public health implications of PE and the benefits of early detection, thus establishing the basis for needing more effective PoC screening tests for this disease.

Secondly, the screening test itself must be cost-effective, simple, safe and acceptable, precise, and accurate130,131. To be cost-effectiveness, tests must be inexpensive, widely available, and able to detect a sufficiently high number of cases of preclinical disease in the screened population to justify their cost130,132. To ensure simplicity, safety, and acceptability, screening tests must be easy to administer, be quick and easy to interpret, and impart minimal discomfort and morbidity to the individual undergoing screening132. Tests must be precise to ensure test results are reliably consistent when repeated and are thus reproducible133.

Finally, the screening test must be accurate and valid in effectively distinguishing those who will develop the disease, before serious disease is reached, from those without the disease. In other words, screening tests should have high sensitivity (i.e. few false-negative results) and specificity (i.e. few false-positive results), with consideration also for their positive (PPV) and negative (NPV) predictive values132. High sensitivity is especially desired for screening tests to avoid missing individuals with the disease, to ensure more individuals are afforded the opportunity for intervention134. To achieve the desired test characteristics, the appropriate cutoff values must be applied for the screening test, although it is important to note that decreasing a screening cutoff value to improve sensitivity usually occurs at the expense of specificity135.

In addition to the economic, practical, and logistical considerations that could influence the adoption and utility of PoC tests in real-world clinical settings, including in low-resource settings that we discussed earlier it is important to consider the importance of accuracy and diagnostic performance. Both are required of a PoC screening test to significantly improve clinical care and promote widespread adoption into clinical practice, as illustrated by the PRAECIS study39. In the PRAECIS study, clinical performance of sFlt-1/PlGF using a discriminatory cutoff ratio of 40 in maternal serum in pregnant people hospitalized for management of hypertension in pregnancy demonstrated 94% sensitivity, 75% specificity, 65% PPV, and 96% NPV for identifying pregnant people at risk of developing sPE within 2 weeks of the test. In receiver-operating characteristic curve analysis evaluating the average value of sensitivity for the sFlt-1/PlGF ratio across all possible values of specificity, the ratio performed better (area under the curve [AUC] = 0.92) than standard clinical measures, such as BP and routine laboratory tests (AUC < 0.75), for predicting sPE39,136. In effect, the high sensitivity of the test, in conjunction with its relatively high specificity, makes the sFlt-1/PlGF ratio an effective screening test to identify individuals at greatest risk of developing sPE who benefit from heightened maternal and fetal surveillance.

In a real-world study of clinical use of sFlt-1/PlGF ratios137, using different discriminatory cutoff ratios (high-risk: >85, intermediate-risk: 38–85, and low-risk: <38) the authors found that the biomarkers performed similarly well for sensitivity and specificity in the high- and intermediate-risk groups. Ultimately, clinical use of this biomarker ratio allowed for adjustment of maternal and fetal monitoring based on biomarker results, and promoted prolongation of pregnancy without negative outcomes for the pregnant person or neonate in the setting of these surveillance modifications137,138. Aside from PoC screening tests evaluating sFlt-1/PlGF ratios, additional PoC screening tests evaluating alternative biomarkers would need to meet, or exceed, these acceptable test characteristics that have been achieved with sFlt-1/PlGF PoC tests in order to be adopted into clinical practice.

Outlook

As discussed, the complexities in screening and diagnosis of PE largely emerge from its multifactorial pathogenic origins. Early in pregnancy, factors such as genetics, environment, and immunology can precipitate abnormal placentation and ischemia. This, in turn, might result in infants being small-for-gestational age. The consequent ischemic state disrupts the equilibrium of angiogenic markers like sFlt-1 and PlGF, leading to varied clinical manifestations that include proteinuria and hypertension. Although sFlt-1 and PlGF are well-established for PE detection, our review highlights many other biomarkers across six key pathogenic categories. However, as highlighted thoughout this article many of the markers warrant further validation in clinical cohorts to fully establish their diagnostic power.

Given the intricate nature of PE, a comprehensive, multiplexed screening approach is essential. This will enable identification of the pathogenic pathways involved and refine therapeutic efficacy and detection precision. A multifaceted approach, perhaps a multi-omics evaluation, could be required, especially in the later stages of pregnancy. However, several challenges exist. The varying degrees of clinical validation for different biomarkers, originating from the different in which they were discovered, and their differing specificity need to be considered when designing PoC devices. The development of PoC devices democratizes access to testing by decentralizing test locations and thereby improving access, including to those most susceptible to PE. But these devices must cater to their deployment scenarios. For example, PoC devices in humid, rural locales need resilience against environmental factors, which may reduce sensitivity. A multiplexed panel incorporating markers such as sFlt-1, PlGF, and other relevant biomarkers might offer close monitoring and timely interventions.

While a detailed discussion of treatments available for PE is beyond the scope of this review, it is evident that PE’s progression exhibits a nuanced shift in biomarker profiles across gestation, mandating tailored therapeutic strategies. Existing treatments often focus on BP control, using drugs such as IV labetalol or hydralazine for acute hypertensive urgency (160/110) and oral nifedipine or labetalol to maintain blood pressure below this severe threshold. Corticosteroids for fetal lung maturation if concern for preterm delivery arises and intrapartum magnesium for seizure prophylaxis if severe features are present, while beneficial, come with their own set of challenges. Nevertheless, early detection and heightened surveillance, potentially achievable with PoC devices, could revolutionize PE management. For instance, early detection of abnormal biomarker profiles could trigger preventive measures, such as administering aspirin in the first trimester or using metformin to counteract angiogenic and mitochondrial pathway damages19,139. We recommend that more studies aim to adopt the approach of the multicenter ASPRE (Combined Multimarker Screening and Randomized Patient Treatment with Aspirin for Evidence-Based Preeclampsia Prevention)140 trial carried out at 13 maternity hospitals in UK, Spain, Italy, Belgium, Greece and Israel, which integrated maternal characteristics with technical measurements including mean arterial pressure and the uterine artery pulsatility index, alongside advanced biochemical assays for serum markers. Specifically, the trial utilized the DELFIA® Xpress platform to quantitatively assess PAPP-A and PlGF, employing multiples of the median for accurate normalization. This comprehensive screening, executed at 11–13 weeks’ gestation, utilized Bayes’ theorem for enhanced risk stratification, giving improved outcomes compared to the approaches currently recommended by NICE and ACOG. There was a high detection rate for preeclampsia seen in trial participants, which further underscored the clinical utility of aspirin prophylaxis in high-risk pregnancies.

In conclusion, while significant strides have been made in understanding PE, an undeniable need remains for effective screening methods that afford swift detection and intervention. With the fusion of innovative biomarker discoveries and the development of sophisticated, multiplexed PoC devices, we anticipate a transformative shift in PE screening and diagnostics, interventions, and outcomes for both pregnant people and their neonates.