Perihepatic lymphadenopathy in children with chronic viral hepatitis

Abstract

Objective

To assess whether lymph node enlargement in the hepatoduodenal ligament occurs in children with chronic viral hepatitis B and C in comparison to healthy controls.

Subject and methods

In 49 patients with chronic viral hepatitis (38 with chronic hepatitis B, 11 with chronic hepatitis C, 31 male, 18 female; age range 1 to 17 years), and in 51 healthy controls (25 male, 26 female; age range 4 to 16 years), the total perihepatic lymph node volume was assessed using transabdominal ultrasonography as previously described in adult patients.

Results

Adequate visualization of the liver hilum was achieved in 46/49 (94%) pediatric patients with chronic viral hepatitis and in 46/51 (90%) pediatric healthy controls. In patients with adequate liver hilum visualization, enlarged perihepatic lymph nodes (longitudinal diameter >14 mm) were detected in 32/46 (70%) patients with chronic viral hepatitis and in 5/46 (11%) healthy controls. The total perihepatic lymph nodes volume [mean ± SD] was 1.0 ± 1.2 mL (0.1–5.4 mL) in patients with chronic viral hepatitis and 0.1 ± 0.1 mL (0.0–0.4 mL) in healthy controls (p < 0.05). A maximal lymph node diameter >14 mm identified patients with chronic viral hepatitis with 70% sensitivity and 89% specificity.

Conclusion

Transabdominal ultrasound can detect lymph nodes within the hepatoduodenal ligament not only in adults but also in children. Paediatric patients with chronic viral hepatitis have significantly enlarged perihepatic lymph nodes compared to controls. Therefore, sonographic assessment of perihepatic lymphadenopathy might be a non-invasive diagnostic tool to screen paediatric patients for chronic viral hepatitis.

Introduction

The clinical value of transabdominal ultrasound (TUS) in the detection of perihepatic lymph nodes (LNs) in adult patients is well established^(1–6). Normal or enlarged LNs in the liver hilum can be identified by TUS between the inferior vena cava and the portal vein, just next to the right renal artery⁽⁷⁾. Since inflammatory processes in organs frequently lead to hyperpla sia of regional LNs^(8–11), LN enlargement within the hepatoduodenal ligament is present in patients with acute or chronic liver disease. Enlarged perihepatic LNs have been detected in patients with chronic viral hepatitis B (CHB)^{(2, 12)}, chronic viral hepatitis C (CHC)^{(4, 13–16)}, primary biliary cirrhosis^{(6, 17)} and primary slcerosing cholangitis⁽¹⁸⁾. The total perihepatic LN volume (LNV) is associated with viremia and liver histology in adult patients with CHC^{(4, 19)}. It changes with progressive normalization according to the antiviral response and the liver histology improvement⁽¹⁶⁾. The LNV reflects progression of disease in primary biliary cirrhosis⁽⁶⁾. In patients with perihepatic lymphadenopathy of unknown origin malignant diseases have to be ruled out.

Previous studies reported that adequate sonographic visualization of the liver hilum can be achieved in more than 90% of adult patients and LNs in the hepatoduodenal ligament can noninvasively be assessed^{(4, 16)}. The method used has been validated in postmortem examinations, healthy subjects and patients with inflammatory liver disease^{(4, 5)}. Furthermore, validation for correct sonographic detection of perihepatic LNs (location and size) was obtained in patients undergoing elective abdominal surgery. The predicted size by TUS revealed a high correlation with the anatomical size which was assessed after excision of the LNs (r = 0.94 for examination during autopsy)⁽⁴⁾. The reproducibility of the method was also previously investigated by our team by repeated TUS examinations of the LNs in the hepatoduodenal ligament in 10 healthy subjects for 7 consecutive days. The mean coefficient of variation for intraindividual assessment of LNV was 12%⁽⁴⁾.

In healthy children, TUS has been reported to occasionally detect small LNs around the portal vein, with the largest diameter up to 10 mm⁽²⁰⁾. Still little is known about TUS findings in children with chronic viral hepatitis (CVH), even if hepatitis B and C viral infections are the most common causes of CVH and liver cirrhosis in older children in Europe^(21–28) . In contrast to adult patients, portal lymphadenopathy has not been systematically examined in children with CVH in comparison to healthy controls (HC).

In the present, prospective, controlled study, therefore, we investigated detection rate and size of LNs (expressed as LNV) within the hepatoduodenal ligament in children with CHB and CHC, and healthy subjects serving as controls.

Patients and methods

Patients and controls

49 consecutive pediatric patients with CVH attending our gastroenterology outpatient clinic and 51 HC were primarily enrolled into the present study. The epidemiological data are summarized in Table 1.

Tab. 1.

Demographic, Biochemical, and Serological Profile of Patients with Chronic Virus Hepatitis (CVH) and Healthy Controls (HC)

Characteristics	CVH	HC
Demography
No (M/F)	49 (31/18)	51 (25/26)
Mean age (yr)*	12 ± 3.8 (1–17)	12 ± 2.8 (4–16)
Biochemistry*
Alanine aminotransferase (U/L)	30 ± 24.3 (10–152)	-
Aspartate aminotransferase (U/L)	15.9 ± 8.8 (5–53)	-
γ-Glutamyl transpeptidase (U/L)	12.7 ± 5.2 (6–30)	-
Bilirubin (mg/dL)	0.5 ± 0.4 (0.1–2)	-
Serology
HBV-DNA positive	38/49 (77.5%)	0/51 (0%)
HCV-RNA positive	11/49 (22.5%)	0/51 (0%)

Abbreviations: HBV-DNA, hepatitis B virus DNA; HCV-RNA, hepatitis C virus RNA.

^*Mean ± SD (range).

Inclusion criteria were the diagnosis of CHC viral infection based on the consistent detection of serum hepatitis C virus RNA by reverse transcription-polymerase chain reaction assay for more than 6 months or the diagnosis of CHB viral infection based on the detection of serum hepatitis B virus DNA by reverse transcription-polymerase chain reaction assay for more than 6 months.

Other possible causes of perihepatic LN enlargement (e.g. acute hepatitis, autoimmune liver disease, including primary biliary cirrhosis, hereditary liver disease, and malignant diseases) were excluded by appropriate clinical, laboratory, and imaging investigations as recently described⁽⁴⁾.

All HC had no complaints, liver function tests were within the normal range and chronic liver diseases (infectious, metabolic, toxic, autoimmune, and hereditary) were excluded by appropriate biochemical and serological tests.

Institutional board approval for the study and informed consent from parents/legal tutors were obtained in all patients and HC.

Methods

Detection of lymph nodes in the hepatoduodenal ligament by transabdominal sonography

In all patients and HC the number and size [the long (a) and the short (b) diameters] of all detectable perihepatic LNs were determined as previously described for adult patients⁽¹⁾. The individual LN volume was calculated assuming a (rotating) ellipsoid [volume = (a/2) × (b/2)2 × (4/3) × π]^{(1, 2, 16)}. The volume of all individual LNs within the hepatoduodenal ligament were subsequently summarized and expressed as total perihepatic lymph node volume (LNV).

Perihepatic LNs of all patients and controls were assessed by the same experienced investigator (CFD), who was blinded to patients’ data at time of examination. All US examinations were performed with a 3.5 MHz probe using high resolution B-mode scanner (Siemens Elegra, Siemens, Erlangen, Germany or Acuson Sequoia, Siemens, Erlangen, Germany).

A sonographically definite diagnosis of LNs in the hepatoduodenal ligament is possible only if the size is more than 5 mm, as previously reported⁽²⁹⁾; therefore we included only LNs of 5 mm or more in size. Lymphadenopathy was defined as a longitudinal LN diameter >14 mm, value found to be the best cut-off by receiver operator curve analysis.

Serological and molecular laboratory tests

Testing for anti-hepatitis C virus antibodies (2^nd generation), hepatitis B surface antigen, hepatitis B core antibodies or anti-human immunodefiency virus antibodies 1 and 2 were performed using commercially available enzyme-linked immunosorbent assays.

Statistical analysis

Demographic, clinical and sonographic characteristics of patients were expressed as mean SD. Students t test was used to compare the LN volume between CVH/CHB/CHC and HC. The distribution of the variables was checked and the variables were normally distributed. Chi-square test was applied to compare the percentage of visible liver hilum, detectable hilar LNs and perihepatic lymphadenopathy between CVH/CHB/CHC and HC (p < 0.05 was judged to be statistically significant). LNs with more than 14 mm in length were considered pathologically enlarged.

Results

In the present study, we investigated 49 consecutive patients with CVH and 51 HC. The clinical, biochemical, and serological characteristics of patients are summarized in Table 1. Even young children <4 years tolerated the ultrasound investigation of the liver hilum well without sedation.

Adequate visualization of the liver hilum was obtained in 46/49 (94%) CVH patients, and in 46/51 (90%) HC. In subjects with adequate visualization of the liver hilum, LNs within the hepatoduodenal ligament were detectable in all 46/46 (100%) patients with CVH, and in 41/46 (89%) of the HC. Representative examples for sonographic visualization of LNs within the hepatoduodenal ligament are shown in Fig. 1 A, B. Lymphadenopathy, defined as >14 mm in the longitudinal diameter was detectable in 32/46 patients with CVH, and in 5/46 HC. Values for sensitivity, specificity, positive and negative predictive values, and accuracy for sonographically detection of enlarged LNs within the hepatoduodenal ligament in CVH patients were of 70%, 89%, 86%, 75%, and 79%, respectively.

Fig. 1.

Transabdominal image of perihepatic lymph nodes (LN or LK) in a patient with chronic viral hepatitis C [conventional B-mode (A) and Colour Doppler imaging (B)]. Liver (L), pancreatic head (PH), portal vein (PV), and renal arteries (RA, NA1, NA2) are also indicated. Arrow: hepatic artery

The total perihepatic LN volume (LNV) was calculated in each subject as sum of each detected LN volume. The mean LNV was 1.0 ± 1.2 mL (0.1–5.4 mL) in patients and 0.1 ± 0.1 mL (0.0–0.4 mL) in controls. In patients with CHB, mean LNV was 0.87 ± 1.19 mL (0.08–5.41 mL), while in patients with CHC it was 1.43 ± 1.19 mL (0.06–3.28 mL). In patients with CVH, the mean total volume values were significantly higher than in healthy subjects (p < 0.05). The LNV in CHB and CHC patients did not differ significantly (p = 0.23). Table 2 presents the sonographic findings in patients with CVH and HC.

Tab. 2.

Sonographic findings in children with chronic hepatitis B and C and in healthy controls

	Liver Hilum Visible	Hilar LNs Detectable	No. of Hilar LNs Detected*	Perihepatic Lymphadenopathy	Total LNs Volume (mL)*
Healthy controls	46/51 (90%)	41/46 (89%)	1.3 ± 0.9 [0–4]	5/46 (11%)	0.1 ± 0.1 mL (0.0–0.4 mL)
Chronic virus hepatitis	46/49 (94%)	46/46 (100%)	3.1 ± 1.8 [1–7]	32/46 (70%)¤	1.0 ± 1.2 mL (0.1–5.4 mL)¤
Cronic virus hepatitis B	36/38 (95%)	36/36 (100%)	2.9 ± 1.7 [1–7]	25/36 (69%)¤	0.87 ± 1.19 mL (0.08–5.41 mL)¤
Cronic virus hepatitis C	10/11 (91%)	10/10 (100%)	3.7 ± 2.2 [1–7]	7/10 (70%)¤	1.43 ± 1.19 mL (0.06–3.28 mL)¤

^*Mean ± SD (range)

^¤statistically different from healthy control.

Perihepatic lymph node analysis in healthy subjects

An adequate visualization of the liver hilum and adequate sonographic visualization of the hepatoduodenal ligament was achieved in 46/51 healthy subjects. In 41 healthy subjects, LNs were depictable. Mean (±SD) number of detected LNs was 1.43 ± 0.80 (0–4). The largest size of detected LNs was 17 × 5 mm (0.2 mL). No healthy person has had more than two detectable LNs in the ventral and dorsal hepatoduodenal ligament, respectively.

Lymph node analysis in patients with chronic virus hepatitis B

Of the 49 patients with CVH, 38 patients had CHB. In 36/38 (95%) of the CHB patients, adequate visualization of the liver hilum and adequate sonographic visualization of the hepatoduodenal ligament was achieved. In 36/36 (100%) of patients with CHB and adequate visualization of the liver hilum, perihepatic LNs were detectable. The largest size of the detected LNs was 35 × 14 mm (3.6 mL). The largest number of detected LNs in the ventral and dorsal hepatoduodenal ligament was n = 7 in one patient. Perihepatic lymphadenopathy was found in 25/36 (69%) CHB patients. The total perihepatic LN volume was 0.87 ± 1.19 mL (0.08–5.41 mL). Values for sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for sonographically detection of enlarged LNs within the hepatoduodenal ligament in CHB patients were of 69%, 89%, 83%, 79%, and 80%, respectively.

Lymph node analysis in patients with chronic virus hepatitis C

Of the 49 patients with CVH, 11 patients had CHC. In 10/11 (91%) of the CHC patients, adequate visualization of the liver hilum and adequate sonographic visualization of the hepatoduodenal ligament was achieved. In 10/10 (100%) of patients with CHC and adequate visualization of the liver hilum, perihepatic LNs were detectable. The largest size of the detected LN was 32 × 10 mm (1.7 mL). The largest number of detected LNs at the level of the hepatoduodenal ligament was 7 in one patient. Perihepatic lymphadenopathy was found in 7/10 CHC patients (70%). The total perihepatic LN volume was 1.43 ± 1.19 mL (0.06–3.28 mL). Values for sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for sonographically detection of enlarged LNs within the hepatoduodenal ligament in CHC patients were of 70%, 89%, 58%, 93%, and 86%, respectively.

Lymph node analysis in patients with chronic virus hepatitis B and C

An adequate visualization of the liver hilum and adequate sonographic visualization of the hepatoduodenal ligament was achieved in 46/49 (94%) patients with CVH (CHB and CHC). In all patients with adequate liver hilum visualization, perihepatic LNs were detectable. The largest size of the detected LNs was 35 × 14 mm (3.6 mL) seen in a patient with CHB. The largest number of detected LNs at the level of the hepatoduodenal ligament was n = 7 (one patient with CHB, and one patient with CHC). In 32/46 (70%) patients with CVH and adequate visualization of the liver hilum perihepatic lymphadenopathy was found, whereas only 5/46 (11%) of the HC with adequate liver hilum visualization revealed lymphadenopathy. Total lymph node volume in CVH patients was 1.0 ± 1.2 mL (0.1–5.4 mL). Values for sensitivity, specificity, positive predictive value, negative predictive value, and accuracy for sonographically detection of enlarged LNs within the hepatoduodenal ligament in CVH patients were of 70%, 89%, 86%, 75%, and 79%, respectively.

Discussion

Studies in adults have proved that TUS technology enables accurate visualization, not only of enlarged, but also of normally sized perihepatic LNs within the hepatoduodenal ligament in healthy subjects as well as in patients with infectious, autoimmune, or neoplastic liver disease. The sonographic method in assessing the total perihepatic LN volume has been validated by TUS studies in patients undergoing elective surgery, and by postmortem examinations⁽⁴⁾. However, the diagnostic importance of sonographically detectable perihepatic lymphadenopathy in pediatric patients with CVH has never been systematically evaluated. Even more, sonographically detection of perihepatic LNs in healthy children has previously been evaluated only by Toppet et al. The results of their study showed that in healthy children, TUS can occasionally detect small LNs around the portal vein, with the largest diameter of 7 to 10 mm⁽²⁰⁾. The data of the present study confirm that LNs within the hepatoduodenal ligament can readily be detected by TUS in healthy children and in children suffering from CVH.

Lymph node analysis in healthy subjects

The range of normal sized LNs in the liver hilum is often defined for adults as longitudinal LN diameter up to 19 mm^{(4, 6, 16, 30, 31)}. In this study, we calculated the sensitivity and specificity of cut-off 10–17 mm, and found by receiver operator curve analysis that 14 mm is the best cut-off. Therefore, we have considered pathologically enlarged only LNs with a longitudinal diameter of more than 14 mm. In this report, normal sized LNs could be visualized in the liver hilum of healthy children in 89% of the cases.

Lymph node analysis in patients with chronic virus hepatitis B and C in comparison

As shown for adults we found a significant difference in lymph node volume between affected children compared to HC, with no significant difference between children suffering chronic viral hepatitis B and C. These data are in accordance with the literature^{(4, 6)}. Enlargement of these perihepatic LNs could be predictive for the presence of severe inflammatory activity, with and without cirrhosis, as for adults already published⁽⁴⁾. Also, studies in adults with CHC have shown that the total perihepatic LN volume, sonographically determined, is related to liver histology and viral load^{(4, 16, 19, 32)}. Further studies should investigate whether successful antiviral therapy may result in a decline of perihepatic LN size together with histological improvement.

Mechanism

The mechanism of portal lymphadenopathy with CHC is unknown, but appears to be related to viral replication and the immune-mediated inflammatory response of the host. Recruitment of HCV-infected lymphocytes and/or macrophages into the draining LNs may be another contributing mechanism⁽³²⁾. Zheng et al. demonstrated in a recent experimental study that portal and celiac LNs are draining the mouse liver. Also, their study brought evidence that the liver-draining LNs induce an anti-HBV-specific immune response responsible for HBV clearance⁽³³⁾.

Because in the present study, no patients with severe portal hypertension were enrolled, impaired venous drainage of the perihepatic LNs appears unlikely as a relevant component of the LN enlargement. Lymphedema as a possible cause of perihepatic lymphadenopathy can be excluded, since lymphatic drainage of the liver is hepatofugal and most likely not influenced by architectural changes of the liver parenchyma⁽⁴⁾.

The advantages of ultrasound (in general)

TUS is an important tool in the diagnosis and follow-up of large populations of patients, due to its favourable cost, availability, flexibility, real-time examination, and user friendliness⁽³⁴⁾. It stands out as the imaging method with highest temporal and spatial resolution, reasons why it is regarded as the first method of choice to evaluate LNs^{(29, 34)}. Ultrasound evaluation of LNs is an important method to detect pathological changes in nearly all body areas including the abdomen^{(35, 36)}. The currently possible LN detection rate by ultrasound is limited by a minimal required LN size which is between 5–10 mm⁽²⁹⁾. Ultrasound equipment may detect LNs in the hepatoduodenal ligament in healthy subjects⁽³¹⁾. LNs are detectable within the hepatoduodenal ligament in almost all patients with CHB^{(12, 13, 32)} and CHC^{(4, 16, 32, 37–39)} as shown by adult studies. In our study, LNs within the hepatoduodenal ligament have been detected in all paediatric patients with CVH and adequate liver hilum visualization. Also, in adults, LNs volume is related to liver histology and viremia, and may be predictive for the presence of severe inflammatory activity^{(4, 39)}. TUS, being a non-invasive technique, may be freely used for repeated follow-up examinations in children with CVH⁽¹⁾. Since European Society of Paediatric Gastroenterology, Hepatology and Nutrition (ESPHAGAN) guidelines for management of children with CHB⁽²²⁾ recommend hepatocellular carcinoma surveillance with liver ultrasound every 6–12 months (depending on the stage of fibrosis)^{(40, 41)}, sonographically assessment of perihepatic LNs would be of even more great value if data would report correlations with disease progression.

Perihepatic LNs can also be observed on computed tomography (CT) scans, which offer an excellent spatial resolution for measuring LNs^{(42, 43)}. Still, there are limitations. To be considered pathologically enlarged and measurable, a LN must be at least 15 mm in short axis when assessed by CT scan. LNs that are at least 10 mm but less than 15 mm in short axis may be pathologic but are considered non-measurable^(44–46). Also, on cross-sectional imaging, measurements of long-axis and short-axis LNs diameters are distorted because of variable nodal orientation⁽⁴⁷⁾. There is also the amount of radiation exposure, which should be avoided as much as possible in children.

The advantages of ultrasound compared to MRI

Magnetic resonance imaging (MRI) can depict the locations of enlarged LNs relative to adjacent bile ducts or vascular structures, providing better contrast between LNs and the adjacent tissues than does sonography or CT^(48–51). MRI depicts perihepatic LNs in most patients with CHC. LN number, size, and hyperintensity have been demonstrated to be related to the activity of CHC, but not to the results of liver function tests⁽⁴⁹⁾. Still, MRI is not widely available and it is expensive. In addition, it needs more time for the examination reason why most of the examinations require patient sedation. This can be an impediment in pediatric age group.

Review of the literature (ultrasound)

Publications on sonographic evaluation of perihepatic lymph nodes in the paediatric population are limited. Toppet et al., sonographically assessed perihepatic LNs in 58 pediatric patients with acute hepatitis A and in 68 controls. Small LNs were occasionally seen in normal subjects, located around the portal vein and their largest diameter was 7 to 10 mm⁽²⁰⁾.

Other studies investigating perihepatic lymphadenopathy were conducted in adults: Ierna et al. examined 1222 subjects by TUS and perihepatic lymphadenopathy at the hepatoduodenal ligament was found in 184 subjects (15.1%). Of these 184 subjects, 142 were positive for anti-HCV (77.1%), while only six of 1038 subjects (0.005%) in whom perihepatic lymphadenopathy was not detected were anti-HCV-positive⁽⁵²⁾.

Braden et al. assessed the total perihepatic LN volume in 40 consecutive patients with transaminases >500 U/L without known liver disease and compared the results with the US findings in 263 patients with known liver disease and also 49 healthy patients. Perihepatic lymphadenopathy was found in none of the HC, in 94% of patients with acute viral hepatitis, in 86% of patients with chronic viral hepatitis, in 90% of autoimmune disease but in none of the patients with toxic liver damage⁽¹⁾.

Dietrich et al. examined 59 patients with chronic viral hepatitis C, and the total perihepatic LN volume (LNV) was assessed using TUS before the initiation of antiviral treatment, at the end of treatment, and at the end of a 6-month follow-up period. At the end of follow-up, they found a significantly smaller LNV in patients with a sustained virologic response than in patients who failed to respond to treatment (0.5 ± 0.3 mL versus 2.0 ± 1.2 mL; p < 0.0001). In the group of sustained virologic responders, the decline of LNs was associated with an improvement in liver histology⁽¹⁶⁾.

Nakanishi et al. investigated the relationship between the pathology of chronic liver diseases and the common hepatic arterial LNs using an ultrasonographic diagnostic system. They designated a LN index by multiplying the long and short LN diameters. The results of their study showed that LN appearance rate and LN index were significantly higher in the patients with hepatitis C, primary biliary cirrhosis and autoimmune hepatitis than in the healthy subjects. Also, LN index was significantly correlated with values for alanine aminotransferase and aspartate aminotransferase, with increase of the portal pressure, and with response to interferon therapy in patients with chronic viral hepatitis C⁽⁵³⁾.

Kuo et al. evaluated the clinical significance of sonographically enlarged LNs in the hepatoduodenal ligament in 600 outpatients categorized in 4 groups using viral markers (nonviral, CHB, CHC, and CHB and CHC). The incidence of detectable LNs in both the CHB group and the CHC group was significantly increased (56.9% and 69.4%, respectively; both p < 0.001) compared with the nonviral group, and this rate was independent of aminotransferase levels. Nodal width was the only significant parameter when viral and nonviral groups were compared (p < 0.05). If a width of more than 5 mm was used to predict HBV or HCV infection, the positive predictive rate was 88% and the specificity was 89%⁽⁵⁴⁾.

Hikita et al. evaluated by TUS 846 CHC patients during a six months period regarding presence of perihepatic LNs enlargement (defined as LN with the long axis diameter of more than 10 mm) as well as development of hepatic carcinoma. They concluded that patients with perihepatic LN enlargement had a lower risk of development of hepatocarcinoma than those without perihepatic LN enlargement. Their results may provide new insights regarding hepatocarcinogenesis. In this regard further studies are needed⁽⁵⁵⁾.

Summary

In conclusion, TUS can detect LNs within the hepatoduodenal ligament not only in adults but also in children. Patients with CVH have significantly enlarged perihepatic LNs compared to controls. Therefore, sonographic assessment of perihepatic lymphadenopathy might be a non-invasive diagnostic tool to screen paediatric patients for CVH.

Conflict of interest

The authors do not report any financial or personal connections with other persons or organizations, which might negatively affect the contents of this publication and/or claim authorship rights to this publication.