Secondary Logo

Journal Logo

Reviews

Non-invasive Markers of Portal Hypertension

Appraisal of Adult Experience and Potential Utilisation in Children

Sutton, Harry*; Dhawan, Anil*; Grammatikopoulos, Tassos*,†

Author Information
Journal of Pediatric Gastroenterology and Nutrition: April 2018 - Volume 66 - Issue 4 - p 559-569
doi: 10.1097/MPG.0000000000001882
  • Free

Abstract

What Is Known

  • Children with severe portal hypertension are at risk of gastrointestinal bleeding from gastroesophageal varices.
  • Many paediatric liver centres use surveillance endoscopies to monitor and some will prophylactically treat high-risk varices.

What Is New

  • Non-invasive markers can be used to identify patients at risk of developing gastrointestinal bleeding.
  • A combination of imaging techniques and serum biomarkers offer the greatest predictive ability for portal hypertension.

Portal hypertension (PHT) is the result of impaired blood flow from the portal vein to the hepatic vein, either as a result of congestion within the liver due to chronic liver disease (CLD) (at a pre- or sinusoidal level) or from extra-hepatic obstruction of the portal vein (Fig. 1) (1). The most life-threatening complication of PHT is gastrointestinal (GI) bleeding, which results from the development and rupture of varices along the GI tract (most commonly in the oesophagus). GI variceal bleeding is associated with significant mortality up to 30% in adults and between 0% and 15% in children, with extra-hepatic PHT variceal bleeding being associated with very low mortality (2–4). Varices that are more than equal to grade II or grade I with red wales, as per UK national guidelines, are clinically significant varices (CSV) and are at a greater risk of bleeding (5,6).

FIGURE 1
FIGURE 1:
Portal Hypertension subdivided on pre-, intra- and post-hepatic level.

The criterion standard of diagnosing PHT in adults is by direct measurement of hepatic venous pressure gradient (HVPG), which is invasive, only possible at specialized centres, and only provides useful data in PHT secondary to CLD (7). The normal pressure in the portal system, of adults, is classified as an HVPG between 1 and 5 mmHg (8). HVPG represents the gradient pressures between the portal vein and the intra-abdominal portion of the inferior vena cava (IVC). It is measured by placing a wedged catheter into the hepatic vein and recording the difference between the wedged hepatic venous pressure and the free hepatic venous pressure. In adults, PHT is defined as HVPG measurement ≥6 mmHg and clinically significant portal hypertension (CSPH), the pressure at which varices begin to form, as ≥10 mmHg. Due to its invasive nature, HVPG is not routinely performed on children. In paediatrics, PHT diagnosis is commonly based on clinical/biochemical findings of PHT complications such as splenomegaly, and thrombocytopenia. There are only minimal data in the paediatric population and mostly in comparison with other non-invasive tests (9). There is no clear consensus in the management of PHT in children. Many centres, including our own, manage PHT in children with surveillance endoscopies to directly visualize and further more prophylactically treat varices (10). Our centre currently uses the criteria of clinically and/or radiologically confirmed splenomegaly and persistent thrombocytopenia (platelet count below 100 × 109), for patients to be offered surveillance endoscopy. A recent review of our department's program found 57% of patients selected for endoscopy, based on our criteria, had CSV. An appropriate replacement score should have a positive predictive value >57%, for the prediction of CSV, which would be directly affected by the prevalence of CSV in the studied population. Due to the risks and expense of surveillance endoscopies in children, a screening test should be one that optimises specificity to its maximum potential, in order to stratify the risk of PHT and the procedure itself. In ideal circumstances clinicians would like a marker/score that identifies all patients without CSV, but this is not feasible in most conditions and more so in PHT.

The use of multiple endoscopists in studies using endoscopic visualisation of oesophageal varices (EV) as the criterion standard diagnosis could potentially also complicate the management of PHT as inter-observer variability in diagnosis and grading varices has been demonstrated (11,12). This issue has been widely observed across studies in adults and children reflecting local services.

The pattern of PHT tends to be different in children compared with adults. In adults PHT is predominately secondary to intra-hepatic pathology, with a small minority of cases being secondary to extra-hepatic portal vein obstruction (EHPVO) (13). The treatment of PHT is also different between intra- and extra-hepatic PHT. Liver cirrhosis and decompensated liver disease are indications for transplantation and subsequently improve portal pressure, whereas EHPVO tends to be managed conservatively for a long period, although surgical intervention, such as shunt placement is also used. The rate of PHT secondary to EHPVO in children has been shown to vary greatly, with ranges from 7% to >50% (14). Some geographical variations in these rates have been found, but EHPVO is consistently more prevalent in children than in adults. Therefore, for a prediction marker to be useful clinically in paediatrics it must be applicable for both intra and extra-hepatic causes of PHT (15).

In order to reduce the number of unnecessary endoscopies preformed on children a non-invasive method that can predict the presence of PHT and/or varices is needed. A variety of methods have been proposed as potential non-invasive markers of PHT. This is an exploratory literature review that aims to review and analyse the already existent data (quality and limitations) on those suggested non-invasive methods that could be used clinically to determine PHT in children, including serum biomarkers, radiology and various scoring algorithms. A literature search was conducted on “PubMed,” using search terms that included “portal hypertension,” “paediatric portal hypertension,” “non-invasive markers of portal hypertension,” “spleen stiffness,” “liver stiffness,” “elastography,” and “endothelial damage.” The articles included were selected based on their relevance to the purpose of our review. Adult studies have been included to supplement areas of limited paediatric data.

BIOMARKERS

Serum biomarkers have been investigated as a potential area to identify a non-invasive marker that could be used to predict the presence and grade of gastroesophageal varices. The pathophysiology of PHT involves a complex relationship between the liver, the spleen, and the vasculatury system that connects them. This means that biochemical changes associated with each aspect of the pathophysiology may provide useful information for the diagnosis and monitoring of PHT. Changes in serum biomarkers associated with liver fibrosis, immune activation, vascular epithelium, and haemostasis have been looked at as potential markers of PHT.

Markers of Immune Activation

Increased portal pressure in CLD is a result of impaired blood flow through a cirrhotic liver. The process of liver cirrhosis results from chronic tissue damage and subsequent inflammation leading to fibrosis and impaired liver architecture (16). Therefore, several inflammatory markers have been investigated as potential markers for the development of PHT secondary to CLD.

In liver cirrhosis Kupffer cells, the macrophages of the liver, mediate hepatic inflammation. Activated Kupffer cells release soluble CD163, a scavenger receptor, which is involved in clearing haemoglobin-heptoglobin complexes. The clearing of these complexes involves the enzyme heme-oxygenase-1 (HO-1), which is specific to Kupffer cells (17). Several studies have looked at the relationship between PHT and the inflammatory markers soluble CD163 and HO-1. Gørnbaek et al (18) found that soluble CD163 was independently correlated to HVPG measurements in cirrhotic adults, and a cut-off value of 3.95 mg/L being able to predict an HVPG ≥ 10 mmHg with a sensitivity of 66%, specificity 94%, and an area under the receiver operating characteristic curve (AUROC) of 0.83. The study was limited by its observational cross-sectional design. A separate study found CD163 levels were also associated with an increase in variceal bleeding (19). Yang et al (20) found that a soluble CD163 value ≥7.05 mg/L was able to predict EV with 80% sensitivity, 89% specificity, and an AUROC = 0.81. Additionally, they found elevated soluble CD163 above the mentioned cut-off in combination with platelets <80 × 109/L and the presence of high-risk haplotypes of HO-1 and vascular endothelial growth factor increased the predictive ability (sensitivity = 90%, specificity = 94%, AUROC = 0.91) (Table 1). Although this score had impressive predictive ability, the use of high-risk haplotypes may be considered impractical for clinical practice but certainly worth investigating further.

TABLE 1
TABLE 1:
Summary of biomarkers and prediction scores developed for the prediction of portal hypertension and varices

Other markers of inflammation such as IL-1β, IL-1Rα, VCAM-1, Fas-R, TNF-β, and HSP-70 were studied by Buck et al (21) looking for a possible correlation with PHT in adults with cirrhosis. They reported that all above markers were significantly correlated with increased HVPG pressures. The authors used multivariate logistic regression combining the 4 strongest predictors (TNF-β, HSP-70, at-risk alcohol use, and Child-Pugh class B) and the resulting score had a sensitivity of 87%, but a specificity of 44% and an AUROC of 0.76 for the prediction of HVPG ≤ 12 mmHg (Table 1). The study cohort comprised cirrhotic patients without varices, and so it was not reflective of the PHT population. Additionally, it did not have a comparable control group.

Markers of Endothelial Dysfunction

PHT as a result of EHPVO is more common in children than in adults; therefore, markers associated with vasculature changes may have greater application in the paediatric population (14,22). Nitric oxide (NO), a molecule with vasodilation properties, and circulating endothelial cells (CECs), a marker of vascular injury, have both been studied as markers of PHT. NO levels were found to be elevated in both intra- and extra-hepatic causes of PHT. Goel et al (23) described in extra-hepatic portal vein occlusion NO levels were raised compared to controls (43.16 vs 5.76 μmol/L, P < 0.001). Furthermore, NO levels were significantly reduced post-shunt surgery, and increasing NO levels indicated shunt blockage. Similar results were also described in studies looking at NO levels in patients with PHT secondary to biliary atresia (BA) and choledochal cysts (24,25). These studies have been able to demonstrate that NO levels are closely correlated with portal pressure, but did not assess NO's predictive power in the consequences of PHT.

CECs are sloughed off from the vascular wall under mechanical stress and are considered a marker of endothelial dysfunction. CEC's numbers have been shown to be higher in patients with cirrhosis (diagnosed on histological, radiological, laboratory, and clinical criteria) compared with controls (73.7 vs 28.7 cells/4 mL, P = 0.02) (26). Additionally, Sethi et al (27) identified a combination of CECs with aspartate aminotransferase-to-platelet ratio (termed CAPRI), which distinguished cirrhotic patients from controls with 98% sensitivity, 85% specificity, and AUROC = 0.98 (Table 1). The study also showed CECs could not differentiate non-cirrhotic liver disease patients from controls or cirrhotic patients, but was limited by its small sample of non-cirrhotic patients (n = 9). CECs use for diagnosing PHT secondary to EHPVO, as well as their association to variceal formation, still requires further study.

Markers of metabolic dysfunction

Obesity and metabolic disorders have been associated with decompensation of liver disease (28). Pathophysiological mechanisms driving this include insulin resistance that is known to develop with hepatic fibrosis, as well as insulin's ability to induce endothelial dysfunction by stimulating the synthesis of NO (29,30). Eslam et al (31) identified a significant correlation between EV and raised homeostatic model assessment for insulin resistance (HOMA-IR) index (a measure of insulin resistance) and adiponectin levels. A HOMA-IR > 4 was able to predict the presence of EV with 89% sensitivity, 71% specificity with an AUROC of 0.79, whereas adiponectin > 19.2 had a 79% sensitivity, 74% specificity and an AUROC of 0.64 (Table 1). Furthermore, HOMA-IR was associated with an increased risk of GI bleeding (P = 0.04). A second study arm evaluated HOMA-IR relationship to HVPG measurements but since the selection criteria for HVPG measurements were not the same as those for endoscopy selection bias could not be ruled out.

Fibrosis Markers

In adults, various markers of fibrosis have been proposed and studied as potential predictors of PHT. Although these would not be applicable for PHT secondary to EHPVO, they may be useful in children with known CLD to test for developing PHT. During the process of fibrosis there is an increase in the production and destruction of the extracellular membrane (ECM). Various components of the ECM such as hyaluronic acid, a glycosaminoglycan found in connective tissue produced by hepatic stellate cells, and laminin, a basement membrane glycoprotein have been studied as potential markers of PHT (32). One study in adults with CLD found that a laminin cut-off of 1.45 U/mL was able to predict an HVPG ≥ 5 mmHg with a sensitivity of 87%, specificity 74%, and diagnostic efficiency of 81% (33). Another study (34) in adults with cirrhosis secondary to alcohol consumption found a higher laminin cut-off concentration of 2.19 U/mL was able to somewhat predict patients with an HVPG ≥ 12 mmHg, a known threshold for EV bleeding (sensitivity = 73%, specificity = 60%), although the study cohort comprised of only 20 patients. Serum hyaluronic acid was also found to be correlated to PHT in a cohort of 45 adults with various aetiologies of liver fibrosis, and a linear regression model combining laminin and hyaluronic acid was able to identify adults with HVPG > 5 mmHg with a sensitivity of 83%, specificity 82%, and AUROC of 0.82 (35,36) (Table 1). This new prediction model was developed using data from an earlier study group and was tested on patients being assessed for PHT, but there was no control arm.

Other fibrosis markers that have been studied include the degradation and formation products from the ECM, such as collagen types I, II, III, IV, V, VI, elastin, and biglycan. These markers were all shown to be significantly increased in adults with HVPG measurements >16 mmHg, and were able to separate patients based on degree of HVPG. Degraded elastin, collagen IV and collagen V were all significantly increased in patients with HVPG measurements >10 mmHg (37). That same study developed a model using the 2 strongest biochemical markers, degraded elastin (a non-collagen degradation marker), and pro-collagen 3 (a collagen formation marker) combined with the model for end-stage liver disease (MELD) score: [−5.6 + (0.4 × MELD) + (2.8 × log(pro-collagen 3)) + (1.3 × log(degraded elastin))]. This score was able to predict the presence of PHT, determined by HVPG, better than any individual marker (AUROC = 0.92) (37) (Table 1). The study included only cirrhotic patients who met the centre's criteria for HVPG measurements and so is only reflective of patients with high MELD scores and moderate to severe PHT.

Coagulation Pathway

One of the sequelae of PHT is the development of splenomegaly, and thrombocytopenia as a result of platelet sequestration. Although in adult patient's platelet count has been shown to be an unreliable predictor of EV, in several paediatric studies isolated platelet count has been able to successfully predict the presence of EV (38–42). Adami et al (38) found a low platelet count was the greatest predictor of EV (AUROC = 0.82) and a cut-off value of 115 × 109/L had a sensitivity of 67% and specificity of 81%. Although this study included children with both intra- and extra-hepatic causes of PHT, the extra-hepatic group only had 5 children. Gana et al (43) similarly found an optimal platelet cut-off value to be 115 × 109/L (AUROC = 0.79). Giannini et al (44) found that platelet count/spleen diameter ratio was reliable method of predicting EV in patients with cirrhosis, and that a cut-off of 909 had a 100% negative predictive value, which was validated in a second group (sensitivity = 100%, specificity = 93%, AUROC = 0.98) (Table 1). The aetiology of PHT was secondary to hepatitis C in 75% of the patients, giving the geopolitical representation of their population. Additionally, the score was not validated against a control group. In children, since there is variation in normal spleen size based on age spleen size z score may be considered for more reliable assessment of splenomegaly (45).

Von Willebrand factor antigen (vWF-Ag) is released from dysfunctional vascular endothelial cells, as a result of mechanical stress from increased intra-portal pressure. It was initially shown to be elevated in patients with cirrhosis and was able to predict clinical outcome (46). It was also investigated as a potential marker of PHT and found to be significantly elevated in patients with EV. Furthermore, a cut-off value of >214% increase was able to predict CSPH determined by HVPG (AUROC = 0.85) (47). This study does not determine whether the raised vWF-Ag was a result of the increased PHT or secondary to the coagulation changes due to the cirrhosis, and so it is unclear if this marker would be applicable in non-cirrhotic patients. Recently vWF-Ag was combined with platelet count to develop a novel prediction score, the Von Willebrand factor antigen/thrombocyte ratio (the VITRO score). Hametner et al (48) reported that the VITRO score performed better than vWF-Ag alone at predicting CSPH diagnosed by HVPG (AUROC = 0.86 vs 0.79) (Table 1). As with the study above, this study did not assess whether this was a result of PHT or cirrhotic changes.

Scores

As seen with vWF-Ag, the combination of multiple non-invasive markers into prediction scores can improve their ability to detect EV and CSPH. A number of novel prediction scores have been developed and tested as possible predictors of PHT both in children and adults. A recent systematic meta-analysis looked at the most commonly used EV predictors in adults with cirrhosis, including aspartate aminotransferase-to-platelet ratio, aspartate aminotransferase-to-alanine aminotransferase ratio, FIB-4 (age × AST/platelet × (ALT)1/2), Lok index (−5.56 − 0.0089 × platelet + 1.26 × AST/ALT + 5.27 × INR) and Forns index [7.811 − 3.131 ln(PLT) × 0.781 ln(GGT) + 3.467 × ln(age) − 0.014 (cholesterol)] (49). The analysis found AUROC for the studied scores in predicting varices were 0.67, 0.72, 0.77, 0.78, and 0.75, respectively; and the AUROC for predicting large varices were similarly 0.72, 0.74, 0.70, 0.72, and 0.65, respectively. A study by Park et al (50) computed a novel score “Risk Score” (14.2 − 7.1 × log(10) (platelet) + 4.2 × log(10) (bilirubin)) which was reliably able to predict CSPH measured by HVPG and EV (AUROC = 0.91 and 0.82) (Table 1).

A number of EV prediction scores for children have also been developed. Through multivariate logistic regression Gana et al developed a clinical prediction rule (CPR) using platelet count, spleen length z score and albumin ((0.75 × platelets/SAZ + 5) + 2.5 × albumin)) (41). CPR was able to predict the presence of EV better than any individual marker (AUROC = 0.93, sensitivity 94%, specificity 81%) (Table 1). This score was developed using children with both intra and extra hepatic causes of PHT, although the extra-hepatic group contained only 5 of 51 (9%) children. A validation study of CPR showed an AUROC of 0.80, sensitivity of 81%, and specificity of 73% (43).

Recently another EV prediction score for children, developed at our centre, was the King's variceal prediction score (KVaPS), which utilises albumin and equivalent adult spleen size [(3 × albumin) − (2 × equivalent adult spleen size)]. KVaPS was modelled using children with CLD only, but was shown to perform better than CPR (AUROC = 0.77 vs 0.73), with a sensitivity and specificity of 72% and 73%, respectively. In the validation cohort AUROC was 0.81 with a sensitivity and specificity of 78% and 73%, respectively (51) (Table 1). Despite this study having a large sample of 124 children, only 89 were actually endoscoped for varices. The rest did not meet departmental criteria for endoscopy and were assumed to not have varices.

Another study from King's focused exclusively on infants with BA. Isted et al. developed the variceal prediction score (VPS) which utilizes albumin and platelet count (albumin × platelets/1000) (52). For the prediction of CSV VPS had an AUROC of 0.75, sensitivity of 86% and specificity of 71% (Table 1). This score was created on a specific subgroup of children and not all children included in this study underwent OGD so variceal presence and severity were assumed in the non-scoped children.

The application of CPR in children with intra and extra-hepatic PHT, as well as its high AUROC (0.93 and 0.80), make it a promising area of future research. Its use differentiating the presence of any EV and high risk EV still requires further studying as the presence of CSV has more important clinical significance.

IMAGING TECHNIQUES

In addition to laboratory biomarkers, a number of imaging techniques have been used to assess PHT both in children and adults. These methods include ultrasound (US) to measure spleen size, portal vein dilatation, and vascular resistance with Doppler; as well as FibroScan measuring liver and spleen stiffness and others. As is the case with above biomarkers, the majority of the studies have been conducted in adult patients.

Ultrasonography

As mentioned earlier, spleen size (measured by US) is commonly used in the assessment of PHT in various scoring systems. The presence and degree of splenomegaly, though associated with developing PHT, is not specific enough to be used as an individual marker as it can be caused by a variety of unrelated haematological and infectious diseases. Although not specific, US spleen size measurements during follow-up are beneficial as spleen enlargement during follow-up was shown to be predictive of the development of EV. One study, following cirrhotic adults for 5 years, found those with spleen enlargement (defined as >1 cm/y) had a higher probability of EV formation (84.6% vs 16.6%, P = 0.001) and growth (63.3% vs 20.6%, P = 0.001) (53). As in any study involving human operated devices, inter-observer variation in ultrasound measurements of spleen size may limit the reliability of the results and objective parameters need to be applied.

In addition to determining spleen size diameter US can be used to monitor vasculature and haemodynamic changes within the portal system. One of these changes is the development of porto-system shunting through collateral formation as a diversion route of a high-pressure system. The presence of porto-systemic shunts were shown to be associated with EV, and in a longitudinal study the development of port-systemic shunts during follow-up was associated with EV formation (54,55). Other haemodynamic changes can be measured with US with Doppler; these include portal blood reversal, congestion index, portal flow velocity, and others (56–59). Although some of these results have shown significant correlation with HVGP, they have not been investigated in children in order to determine applicable cut-offs that could predict PHT (60). Additionally, inter-observer and inter-equipment variability has been shown to impact the reliability of these measurements (61,62).

Magnetic Resonance Imaging

In addition to US, magnetic resonance (MR) technology has been studied in the adult population for its applicability in the prediction of PHT. Although MR is limited in its practicality by cost and time of procedure, preliminary studies have shown promising results.

Magnetic resonance elastography (MRE) has been used to determine liver stiffness measurement (LSM) and spleen stiffness measurement (SSM) in both animal and human models (63,64). A study, by Morisaka et al (65), found raised LSM and SSM, measured by MRE, to both be associated with an increased risk of any varices, whereas raised SSM was closely associated with a risk of CSV. This study also attempted to determine cut-off values for the prediction of varices, but the values were obtained from retrospective patients and were not validated prospectively. A study from 2009 of only 38 adults with CLD found 100% of those with EV had an SSM ≥ 10.5 kPa (66). Shin et al (67) found that both SSM and LSM measured by MRE were associated with EV and high risk EV. The study found for the prediction of any EV an LSM cut-off of 4.58 kPa had a sensitivity of 86%, a specificity of 72%, and an AUROC of 0.82, whereas an SSM cut-off of 7.23 kPa had a sensitivity of 85%, specificity of 79%, and AUROC of 0.83. For the prediction of high-risk varices, an LSM cut-off of 4.81 kPa had sensitivity of 89%, specificity of 56%, and AUROC of 0.75, whereas an SSM cut-off of 7.6 kPa had a sensitivity of 76%, specificity of 66%, and an AUROC of 0.75. This was a retrospective study with a large variation in the time intervals between endoscopy and MRE, limiting the usefulness of the findings.

Abdominal MRI, in addition to MRE, has also been found useful for the prediction of PHT. A study by Chouhan et al (68) looked at total liver and hepatic arterial blood flow, using caval subtraction 2D phase-contrast MRI (PCMRI). Total liver blood flow is estimated from the difference between 2D PCRMI flow measured in the suprahepatic, subcardiac IVC and infrahepatic, suprarenal IVC. The hepatic arterial flow is then calculated by subtraction from directly measured 2D PCMRI portal vein flow measurements. This study found that caval subtraction hepatic artery fraction was significantly correlated (r = 0.78) with HVPG measurements. It is important to note this study was conducted on a sample of 13 healthy volunteers; a study using patients with PHT is still needed. A different study looked at various MR parameters, such as longitudinal vascular relaxation time, perfusion of the liver and spleen and blood flow in the portal, splanchnic and collateral circulation (69). The researchers found relaxation time and splenic artery velocity to be significantly associated with HVPG (r = 0.9) as well as CSPH (r = 0.85). Like the previous study the sample size here was small, with only 24 adults included in the final analysis.

FIBRO SCAN

The use of FibroScan to measure liver and spleen stiffness (elastography) as a means of predicting PHT has become an area of research recently. In the Baveno consensus meeting in 2015 the working group focused on the stratification of adult patients developing significant PHT. Among other key points, the group recommended the combination of liver stiffness reflected by Transient Elastography measurements <20 kPa on 2 occasions and platelet count >150 × 109/L to accurately identify adults at low risk of having CSV and therefore avoid screening endoscopy (70). The majority of data where the recommendations were based at were extracted from studies on adult patients with viral induced compensated advanced liver disease, somewhat different to the paediatric population. This non-invasive method of predicting PHT has been of particular interest in paediatrics and has several studies looking exclusively at children. Although there are several types of elastography techniques, the most commonly studied in children utilises transient elastography (TE).

Chongsrisawat et al (71) found that in children with BA LSM was significantly correlated to the presence of EV or gastric varices, and a cut-off value of 12.7 kPa was able to predict the presence of varices (AUROC = 0.89). Unfortunately, this was a cross-sectional study and so was unable to comment on the subject of how changes in LSM may relate to variceal growth and possibility of bleeding. A study looking at children with cystic fibrosis found a similar cut-off (12.5 kPa) for LSM would have reduced the number of unnecessary endoscopies by two-thirds (6 vs 2) (72). Another study by Colecchia et al (39) on children with BA found a LSM cut-off of 10.6 kPa was able to predict the presence of EV (AUROC = 0.92). The same group found that by combining LSM with spleen size diameter and platelet count to make the LSM-spleen diameter-to-platelet ratio score the prediction ability could be improved (AUROC = 0.96 vs 0.92) (Table 2). A major limitation of TE FibroScan is that it is contraindicated in patients with ascites, a common complication of PHT. As a result, studies using TE FibroScan risk selection bias by excluding many with severe PHT.

TABLE 2
TABLE 2:
Imaging and elastography techniques studied for the prediction of gastroesophageal varices

Spleen stiffness measurements are of particular interest in the paediatric population as it can be used for both intra- and extra-hepatic PHT. A systematic review and meta-analysis identified that SSM was able to detect the presence of EV with 78% sensitivity and 76% specificity (73). Furthermore, it found SSM was able to detect clinically significant EV (defined as ≥grade 2) with 81% sensitivity and 66% specificity. Of note, the studies reviewed were again in the adult population and included a variety of FibroScan modalities. In exclusively EHPVO adult patients, an SSM cut-off of 42.8 kPa was able to detect variceal bleeding with a sensitivity of 88% and specificity of 94% (74) (Table 2). A study by Goldschmidt et al (75) looked at children exclusively with CLD and found SSM to be elevated in those with EV (75 vs 24 kPa) and variceal haemorrhage (75 vs 50.25 kPa), and found patients with bleeding had an SSM ≥ 60 kPa. As this study was conducted as part of clinical practice, where endoscopy was only offered to those with a high suspicion of varices, selection bias meant only those with severe PHT were included. A study including children with intra and extra-hepatic causes of PHT is still needed to determine SSM's use as a screening tool for PHT in children.

SUMMARY

The desired outcome of this literature review is to identify potential non-invasive markers that could be used clinically to select children with CSV. To be useful clinically, novel prediction markers should be an improvement in our current practice.

An ideal prediction marker/scoring system would be one that is applicable in both intra- and extra-hepatic causes of PHT. Focusing on markers associated with endothelial dysfunction would offer the best chance at identifying a widely applicable marker/scoring system, as both intra- and extra-hepatic PHT feature endothelial dysfunction as a crucial part of its pathophysiology. It is also possible that 2 different markers/scoring systems are needed, based on the aetiology of the PHT, and so research into all aspects of PHT pathophysiology are important.

Although there is significant literature on the topic of PHT prediction in adult population it is difficult to apply the conclusions from these studies to children. The aetiology both of PHT and CLD is different between adults and children with this issue reflected across the studies as they are mostly targeting specific patient groups such as hepatotropic virus induced cirrhosis. Additionally, cut-off values for biomarkers and imaging techniques that have been developed on adults are not necessary applicable in children while majority of studies have utilised HVPG as an objective measurement of PHT, rarely reproducible in paediatric population. There are no adequately powered, randomised or even prospective studies in children with CLD to guide the best screening modality. Even more there is a lack of validated cohort studies, particularly in paediatrics minimising the validity of the reported findings.

As with any attempt to extract meaningful conclusions on a specific complex medical question such as PHT in children it was paramount to include all studies that dealt with the issue of non-invasive markers of PHT. We assessed the quality and limitations of these studies by utilising the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) framework (76). The majority of the listed studies in this review were representative of the local population benefiting from the suggestive marker and all had their population clearly described along with the suggested test, with acceptable time-interval between different diagnostic modalities. The majority of the studies did not clarify the issue of specialists blinded in their interpretation and reporting of results such as endoscopic, radiological, and laboratory tests. Issues around patient withdrawal and other limitations were also not clear in most reports (Table 3). Although the practicalities of such studies can be limiting if we are to produce as meaningful as possible data in PHT we have to consider all the different aspects listed in Table 3.

TABLE 3
TABLE 3:
Utilisation of the QUADAS quality assessment tool in the review of the diagnostic accuracy of the listed studies
TABLE 3
TABLE 3:
(Continued) Utilisation of the QUADAS quality assessment tool in the review of the diagnostic accuracy of the listed studies

An ideal study would be done prospectively with similar number of children with intra- and extra-hepatic PHT selected for first endoscopy based on a high sensitivity, low specificity screening test to optimise a range of PHT severity. These children would then have selected biomarkers recorded and imaging techniques performed before endoscopy, which would be performed by a singular endoscopist blinded to all other markers.

CONCLUSIONS

Several potential non-invasive markers have been studied as possible predictors of PHT, including biomarkers and imaging techniques. In order to be clinically useful, the ideal marker would be one that could not only predict the presence of PHT but could also differentiate between grades of severity and risk of bleeding. Although it would be beneficial to be applicable in both intra and extra-hepatic PHT, since the latter makes up a significant amount of the paediatric PHT population, it is also possible that 2 separate prediction markers/scoring systems may be necessary. Additionally, it would need to be inexpensive and simple enough to be performed at regular follow-up appointments. Markers associated with endothelial dysfunction, such as vWF-Ag, NO, and CEC, may be the most applicable as a predictor for all children regardless of PHT aetiology, as endothelial dysfunction is seen in PHT secondary to intra and extra-hepatic pathology. Though currently, there is a lack of research linking NO, vWF-Ag and CECs with complications of PHT.

Based on the findings of the studies discussed it seems unlikely that a single marker would be sufficiently accurate in predicting PHT alone. The methods that had the greatest success were those that used a combination of markers to create a scoring system, such as the VITRO and Risk scores in adults and diameter-to-platelet ratio score, CPR and KVaPS in children. The recommendation by the Baveno consensus group on utilisation of TE and platelet count in adults is easily applicable and potentially useful in a paediatric setting. In conclusion, additional prospective paediatric studies are still needed to determine the best non-invasive marker of PHT in children.

REFERENCES

1. Maruyama H, Yokosuka O. Pathophysiology of portal hypertension and esophageal varices. Int J Hepatol 2012; 2012:895787.
2. Carbonell N, Pauwels A, Serfaty L, et al. Improved survival after variceal bleeding in patients with cirrhosis over the past two decades. Hepatology 2004; 40:652–659.
3. Kim SJ, Kim KM. Recent trends in the endoscopic management of variceal bleeding in children. Pediatr Gastroenterol Hepatol Nutr 2013; 16:1–9.
4. Chaudhary N, Mehrotra S, Srivastava M, et al. Management of bleeding in extrahepatic portal venous obstruction. Int J Hepatol 2013; 2013:784842.
5. Duche M, Ducot B, Tournay E, et al. Prognostic value of endoscopy in children with biliary atresia at risk for early development of varices and bleeding. Gastroenterology 2010; 139:1952–1960.
6. Wanty C, Helleputte T, Smets F, et al. Assessment of risk of bleeding from esophageal varices during management of biliary atresia in children. J Pediatr Gastroenterol Nutr 2013; 56:537–543.
7. Bosch J, Abraldes JG, Berzigotti A, et al. The clinical use of HVPG measurements in chronic liver disease. Nat Rev Gastroenterol Hepatol 2009; 6:573–582.
8. Kumar A, Sharma P, Sarin SK. Hepatic venous pressure gradient measurement: time to learn! Indian J Gastroenterol 2008; 27:74–80.
9. Yoon HM, Kim SY, Kim KM, et al. Liver stiffness measured by shear-wave elastography for evaluating intrahepatic portal hypertension in children. J Pediatr Gastroenterol Nutr 2017; 64:892–897.
10. Grammatikopoulos T, McKiernan PJ, Dhawan A. Portal hypertension and its management in children [published online August 16, 2017] Arch Dis Child doi: 10.1136/archdischild-2015-310022.
11. Conn HO, Smith HW, Brodoff M. Observer variation in the endoscopic diagnosis of esophageal varices. A prospective investigation of the diagnostic validity of esophagoscopy. N Engl J Med 1965; 272:830–834.
12. Bendtsen F, Skovgaard LT, Sorensen TI, et al. Agreement among multiple observers on endoscopic diagnosis of esophageal varices before bleeding. Hepatology 1990; 11:341–347.
13. Eckhauser FE, Appelman HD, Knol JA, et al. Noncirrhotic portal hypertension: differing patterns of disease in children and adults. Surgery 1983; 94:721–728.
14. Imanieh MH, Dehghani SM, Khoshkhui M, et al. Etiology of portal hypertension in children: a single center's experiences. Middle East J Dig Dis 2012; 4:206–210.
15. Poddar U, Thapa BR, Rao KL, et al. Etiological spectrum of esophageal varices due to portal hypertension in Indian children: is it different from the West? J Gastroenterol Hepatol 2008; 23:1354–1357.
16. Twedt DC. Cirrhosis: a consequence of chronic liver disease. Vet Clin North Am Small Anim Pract 1985; 15:151–176.
17. Philippidis P, Mason JC, Evans BJ, et al. Hemoglobin scavenger receptor CD163 mediates interleukin-10 release and heme oxygenase-1 synthesis: antiinflammatory monocyte-macrophage responses in vitro, in resolving skin blisters in vivo, and after cardiopulmonary bypass surgery. Circ Res 2004; 94:119–126.
18. Grønbaek H, Sandahl TD, Mortensen C, et al. Soluble CD163, a marker of Kupffer cell activation, is related to portal hypertension in patients with liver cirrhosis. Aliment Pharmacol Ther 2012; 36:173–180.
19. Waidmann O, Brunner F, Herrmann E, et al. Macrophage activation is a prognostic parameter for variceal bleeding and overall survival in patients with liver cirrhosis. J Hepatol 2013; 58:956–961.
20. Yang YY, Hou MC, Lin MW, et al. Combined platelet count with sCD163 and genetic variants optimizes esophageal varices prediction in cirrhotic patients. J Gastroenterol Hepatol 2013; 28:112–121.
21. Buck M, Garcia-Tsao G, Groszmann RJ, et al. Novel inflammatory biomarkers of portal pressure in compensated cirrhosis patients. Hepatology 2014; 59:1052–1059.
22. Ryckman FC, Alonso MH. Causes and management of portal hypertension in the pediatric population. Clin Liver Dis 2001; 5:789–818.
23. Goel P, Srivastava K, Das N, et al. The role of nitric oxide in portal hypertension caused by extrahepatic portal vein obstruction. J Indian Assoc Pediatr Surg 2010; 15:117–121.
24. Chand K, Bhatnagar V, Agarwala S, et al. The incidence of portal hypertension in children with choledochal cyst and the correlation of nitric oxide levels in the peripheral blood with portal pressure and liver histology. J Indian Assoc Pediatr Surg 2015; 20:133–138.
25. Khanna V, Bhatnagar V, Agarwala S, et al. Portal pressure and blood nitric oxide levels as predictors of outcome in biliary atresia. J Indian Assoc Pediatr Surg 2016; 21:49–53.
26. Abdelmoneim SS, Talwalkar J, Sethi S, et al. A prospective pilot study of circulating endothelial cells as a potential new biomarker in portal hypertension. Liver Int 2010; 30:191–197.
27. Sethi S, Simonetto DA, Abdelmoneim SS, et al. Comparison of circulating endothelial cell/platelet count ratio to aspartate transaminase/platelet ratio index for identifying patients with cirrhosis. J Clin Exp Hepatol 2012; 2:19–26.
28. Berzigotti A, Garcia-Tsao G, Bosch J, et al. Obesity is an independent risk factor for clinical decompensation in patients with cirrhosis. Hepatology 2011; 54:555–561.
29. Sud A, Hui JM, Farrell GC, et al. Improved prediction of fibrosis in chronic hepatitis C using measures of insulin resistance in a probability index. Hepatology 2004; 39:1239–1247.
30. Caballero AE. Endothelial dysfunction in obesity and insulin resistance: a road to diabetes and heart disease. Obes Res 2003; 11:1278–1289.
31. Eslam M, Ampuero J, Jover M, et al. Predicting portal hypertension and variceal bleeding using non-invasive measurements of metabolic variables. Ann Hepatol 2013; 12:588–598.
32. Liu T, Wang X, Karsdal MA, et al. Molecular serum markers of liver fibrosis. Biomark Insights 2012; 7:105–117.
33. Gressner AM, Tittor W, Kropf J. The predictive value of serum laminin for portal hypertension in chronic liver diseases. Hepatogastroenterology 1988; 35:95–100.
34. Kondo M, Miszputen SJ, Leite-mor MM, et al. The predictive value of serum laminin for the risk of variceal bleeding related to portal pressure levels. Hepatogastroenterology 1995; 42:542–545.
35. Kropf J, Gressner AM, Tittor W. Logistic-regression model for assessing portal hypertension by measuring hyaluronic acid (hyaluronan) and laminin in serum. Clin Chem 1991; 37:30–35.
36. Snowdon VK, Guha N, Fallowfield JA. Noninvasive evaluation of portal hypertension: emerging tools and techniques. Int J Hepatol 2012; 2012:691089.
37. Leeming DJ, Karsdal MA, Byrjalsen I, et al. Novel serological neo-epitope markers of extracellular matrix proteins for the detection of portal hypertension. Aliment Pharmacol Ther 2013; 38:1086–1096.
38. Adami MR, Ferreira CT, Kieling CO, et al. Noninvasive methods for prediction of esophageal varices in pediatric patients with portal hypertension. World J Gastroenterol 2013; 19:2053–2059.
39. Colecchia A, Di Biase AR, Scaioli E, et al. Non-invasive methods can predict oesophageal varices in patients with biliary atresia after a Kasai procedure. Dig Liver Dis 2011; 43:659–663.
40. Fagundes ED, Ferreira AR, Roquete ML, et al. Clinical and laboratory predictors of esophageal varices in children and adolescents with portal hypertension syndrome. J Pediatr Gastroenterol Nutr 2008; 46:178–183.
41. Gana JC, Turner D, Roberts EA, et al. Derivation of a clinical prediction rule for the noninvasive diagnosis of varices in children. J Pediatr Gastroenterol Nutr 2010; 50:188–193.
42. Qamar AA, Grace ND, Groszmann RJ, et al. Platelet count is not a predictor of the presence or development of gastroesophageal varices in cirrhosis. Hepatology 2008; 47:153–159.
43. Gana JC, Turner D, Mieli-Vergani G, et al. A clinical prediction rule and platelet count predict esophageal varices in children. Gastroenterology 2011; 141:2009–2016.
44. Giannini E, Botta F, Borro P, et al. Platelet count/spleen diameter ratio: proposal and validation of a non-invasive parameter to predict the presence of oesophageal varices in patients with liver cirrhosis. Gut 2003; 52:1200–1205.
45. Rosenberg HK, Markowitz RI, Kolberg H, et al. Normal splenic size in infants and children: sonographic measurements. AJR Am J Roentgenol 1991; 157:119–121.
46. La Mura V, Reverter JC, Flores-Arroyo A, et al. Von Willebrand factor levels predict clinical outcome in patients with cirrhosis and portal hypertension. Gut 2011; 60:1133–1138.
47. Ferlitsch M, Reiberger T, Hoke M, et al. von Willebrand factor as new noninvasive predictor of portal hypertension, decompensation and mortality in patients with liver cirrhosis. Hepatology 2012; 56:1439–1447.
48. Hametner S, Ferlitsch A, Ferlitsch M, et al. The VITRO Score (Von Willebrand Factor Antigen/Thrombocyte Ratio) as a new marker for clinically significant portal hypertension in comparison to other non-invasive parameters of fibrosis including ELF test. PLoS One 2016; 11:e0149230.
49. Deng H, Qi X, Guo X. Diagnostic accuracy of APRI, AAR, FIB-4, FI, King, Lok, Forns, and FibroIndex scores in predicting the presence of esophageal varices in liver cirrhosis: a systematic review and meta-analysis. Medicine (Baltimore) 2015; 94:e1795.
50. Park SH, Park TE, Kim YM, et al. Non-invasive model predicting clinically-significant portal hypertension in patients with advanced fibrosis. J Gastroenterol Hepatol 2009; 24:1289–1293.
51. Witters P, Hughes D, Karthikeyan P, et al. King's variceal prediction score: a novel non-invasive marker of portal hypertension in paediatric chronic liver disease. J Pediatr Gastroenterol Nutr 2017; 64:518–523.
52. Isted A, Grammatikopoulos T, Davenport M. Prediction of esophageal varices in biliary atresia: derivation of the “varices prediction rule”, a novel noninvasive predictor. J Pediatr Surg 2015; 50:1734–1738.
53. Berzigotti A, Zappoli P, Magalotti D, et al. Spleen enlargement on follow-up evaluation: a noninvasive predictor of complications of portal hypertension in cirrhosis. Clin Gastroenterol Hepatol 2008; 6:1129–1134.
54. Berzigotti A, Merkel C, Magalotti D, et al. New abdominal collaterals at ultrasound: a clue of progression of portal hypertension. Dig Liver Dis 2008; 40:62–67.
55. von Herbay A, Frieling T, Haussinger D. Color Doppler sonographic evaluation of spontaneous portosystemic shunts and inversion of portal venous flow in patients with cirrhosis. J Clin Ultrasound 2000; 28:332–339.
56. Kim G, Cho YZ, Baik SK, et al. The accuracy of ultrasonography for the evaluation of portal hypertension in patients with cirrhosis: a systematic review. Korean J Radiol 2015; 16:314–324.
57. Kondo T, Maruyama H, Sekimoto T, et al. Reversed portal flow: clinical influence on the long-term outcomes in cirrhosis. World J Gastroenterol 2015; 21:8894–8902.
58. Moriyasu F, Nishida O, Ban N, et al. “Congestion index” of the portal vein. AJR Am J Roentgenol 1986; 146:735–739.
59. Zoli M, Iervese T, Merkel C, et al. Prognostic significance of portal hemodynamics in patients with compensated cirrhosis. J Hepatol 1993; 17:56–61.
60. Berzigotti A, Ashkenazi E, Reverter E, et al. Non-invasive diagnostic and prognostic evaluation of liver cirrhosis and portal hypertension. Dis Markers 2011; 31:129–138.
61. Sabba C, Weltin GG, Cicchetti DV, et al. Observer variability in echo-Doppler measurements of portal flow in cirrhotic patients and normal volunteers. Gastroenterology 1990; 98:1603–1611.
62. de Vries PJ, van Hattum J, Hoekstra JB, et al. Duplex Doppler measurements of portal venous flow in normal subjects. Inter- and intra-observer variability. J Hepatol 1991; 13:358–363.
63. Nedredal GI, Yin M, McKenzie T, et al. Portal hypertension correlates with splenic stiffness as measured with MR elastography. J Magn Reson Imaging 2011; 34:79–87.
64. Ronot M, Lambert S, Elkrief L, et al. Assessment of portal hypertension and high-risk oesophageal varices with liver and spleen three-dimensional multifrequency MR elastography in liver cirrhosis. Eur Radiol 2014; 24:1394–1402.
65. Morisaka H, Motosugi U, Ichikawa S, et al. Association of splenic MR elastographic findings with gastroesophageal varices in patients with chronic liver disease. J Magn Reson Imaging 2015; 41:117–124.
66. Talwalkar JA, Yin M, Venkatesh S, et al. Feasibility of in vivo MR elastographic splenic stiffness measurements in the assessment of portal hypertension. AJR Am J Roentgenol 2009; 193:122–127.
67. Shin SU, Lee JM, Yu MH, et al. Prediction of esophageal varices in patients with cirrhosis: usefulness of three-dimensional MR elastography with echo-planar imaging technique. Radiology 2014; 272:143–153.
68. Chouhan MD, Mookerjee RP, Bainbridge A, et al. Caval subtraction 2D phase-contrast MRI to measure total liver and hepatic arterial blood flow: proof-of-principle, correlation with portal hypertension severity and validation in patients with chronic liver disease. Invest Radiol 2017; 52:170–176.
69. Palaniyappan N, Cox E, Bradley C, et al. Non-invasive assessment of portal hypertension using quantitative magnetic resonance imaging. J Hepatol 2016; 65:1131–1139.
70. de Franchis R, Baveno VIF. Expanding consensus in portal hypertension: report of the Baveno VI Consensus Workshop: stratifying risk and individualizing care for portal hypertension. J Hepatol 2015; 63:743–752.
71. Chongsrisawat V, Vejapipat P, Siripon N, et al. Transient elastography for predicting esophageal/gastric varices in children with biliary atresia. BMC Gastroenterol 2011; 11:41.
72. Malbrunot-Wagner AC, Bridoux L, Nousbaum JB, et al. Transient elastography and portal hypertension in pediatric patients with cystic fibrosis transient elastography and cystic fibrosis. J Cyst Fibros 2011; 10:338–342.
73. Singh S, Eaton JE, Murad MH, et al. Accuracy of spleen stiffness measurement in detection of esophageal varices in patients with chronic liver disease: systematic review and meta-analysis. Clin Gastroenterol Hepatol 2014; 12:935.e4–945.e4.
74. Sharma P, Mishra SR, Kumar M, et al. Liver and spleen stiffness in patients with extrahepatic portal vein obstruction. Radiology 2012; 263:893–899.
75. Goldschmidt I, Brauch C, Poynard T, et al. Spleen stiffness measurement by transient elastography to diagnose portal hypertension in children. J Pediatr Gastroenterol Nutr 2014; 59:197–203.
76. Whiting P, Rutjes AW, Reitsma JB, et al. The development of QUADAS: a tool for the quality assessment of studies of diagnostic accuracy included in systematic reviews. BMC Med Res Methodol 2003; 3:25.
Keywords:

biomarkers; children; portal hypertension; varices

© 2018 by European Society for Pediatric Gastroenterology, Hepatology, and Nutrition and North American Society for Pediatric Gastroenterology,