Portal vein congestion index in patients with and without chronic liver disease

Chronic liver disease (CLD) is a progressive deterioration of liver functions (synthesis of clotting factors/proteins, detoxification of harmful products of metabolism, and excretion of bile) for more than six months. It is a continuous process of inflammation, destruction, and regeneration of liver parenchyma, which leads to fibrosis and cirrhosis. Cirrhosis is the final stage of CLD that results in disruption of liver architecture, the formation of widespread nodules, vascular reorganization, neo-angiogenesis, and deposition of the extracellular matrix⁽¹⁾. Progressive hepatic fibrosis leads to increased intrahepatic resistance and portal hypertension⁽²⁾. Liver cirrhosis is the most common cause of portal hypertension⁽³⁾. In addition, portal venous pressure correlates significantly with the degree of disease chronicity and fibrosis in liver disease^(4,5).

CLD has a wide variety of causes, including long-term alcoholism, chronic viral infection (hepatitis B and C viruses), autoimmune diseases, metabolic diseases (e.g., non-alcoholic fatty liver disease, NAFLD), and hereditary disorders^(1,2).

As of 2021, there were 1.5 billion people living with CLD (including any stage of disease severity) worldwide⁽⁶⁾. Currently, the pre-eminent causes of liver cirrhosis globally are NAFLD, hepatitis B virus (HBV), hepatitis C virus (HCV), and alcoholic liver disease, in that order⁽⁶⁾. In sub-Saharan Africa, CLD accounted for an estimated 157,558.69 deaths (equivalent to 2.11% of all deaths) in 2017⁽⁷⁾. The prevalence rates of HBV and HCV in our locality are 4.3–23.3% and 0.5–15%, respectively⁽⁸⁾.

Liver biopsy has been the traditional gold standard for diagnosing liver disease, while hepatic venous pressure gradient (HPVG) is the standard for diagnosing portal hypertension^(9–11). However, these procedures are highly invasive. Furthermore, liver biopsy is prone to significant sampling errors, as it samples only 1/50,000 part of the liver⁽¹²⁾.

Ultrasonography is a key non-invasive method for diagnosing liver diseases using B-mode/grayscale imaging, elastography, and Doppler parameters. The diameter of the portal vein is variable, and various studies carried out to determine the sonographic value for their respective populations reveal a range of about 7–15 mm^(13–15).

On Doppler interrogation, the normal flow through the portal vein is hepatopetal (towards the liver), a finding which may be absent or reversed in portal hypertension. The normal blood flow velocity in the portal vein is 13–55 cm/s. The portal waveform is monophasic with an undulating appearance due to its variations with cardiac activity and respiration^(15–17).

The portal vein congestion index (PVCI) was first introduced and analyzed by Moriyasu et al.⁽¹⁸⁾ as a duplex Doppler tool for detecting portal venous hemodynamic changes in certain cases of liver disease, and is defined as the ratio of the portal vein cross-sectional area (cm²) and the mean blood flow velocity (cm/sec)^(3,18,19). By relating two quantifiable factors affected by portal hypertension, it tends to magnify the differences in measured parameters between study groups, and is therefore said to be the most sensitive and specific triplex Doppler parameter for detecting portal hypertension⁽²⁰⁾.

The aim of this study was to compare the PVCI of adult patients with CLD to that of healthy controls, and to evaluate the differences in PVCI, if any, between the common etiologies of CLD (CVH, ALD, and NAFLD).

Portal hypertension is a common finding in decompensated chronic liver disease/liver cirrhosis, and several researchers have validated the suitability of Doppler sonography for evaluating portosystemic changes within the liver⁽¹⁴⁾. In this study, the hepatic and portal venous changes in 80 participants with CLD were compared to healthy controls.

The most common cause of CLD was alcoholic liver disease (ALD), which was seen in 57 (71.3%) subjects. The number of CLD patients with ALD was significantly higher than those diagnosed with CVH. This is at variance with studies which found chronic viral hepatitis B to be the most common cause of CLD^(29–31). The higher prevalence of ALD in this study was probably due to the high rate of alcohol consumption (locally brewed gin called ‘ogogoro’ or ‘kai-kai’) in our region⁽³²⁾.

A greater number of subjects in the ALD group showed more features of hepatic decompensation (ascites, splenomegaly, hepatic nodularity, shrunken liver with irregular outline) than the CVH group. This can probably be explained by the fact that subjects with alcoholic liver disease were more inclined to show signs of more advanced signs of liver disease due to late presentation, poor level of compliance and higher rates of relapse. Some of the hepatic and extrahepatic sonographic findings in the CLD group of this study were similar to those documented by Maàji et al.⁽³³⁾: hepatomegaly (31%), shrunken liver (25%), abnormal liver echotexture (95%), liver edge irregularities (93%), splenomegaly (9%), and ascites (74%). These values suggest a more advanced decompensated CLD in comparison with the index study.

The portal vein diameter (PVD) was significantly higher in the CLD group than the controls. This is in agreement with the findings from the study by Achim et al.⁽¹³⁾ and Ayele et al.⁽³⁴⁾ who observed a significantly higher PVD in the CLD groups than the controls. The reason for this significant difference can be adduced to the presence of portal hypertensive changes in patients with CLD, resulting in a number of hemodynamic portal venous alterations, one of which is increased venous diameter. Furthermore, subjects with ALD had a significantly higher PVD than those with CVH. The difference in PVD between these two groups is probably due to the aforementioned reason of generally more advanced disease seen in ALD than the CVH subgroup. However, the PVD should not be used in isolation to diagnose portal hypertension because of its poor sensitivity to the presence of collateral channels that decompress the portovenous system⁽¹⁴⁾.

The mean PVCSA of the CLD group was significantly higher than that of controls. This was similar to the findings by Moriyasu et al.⁽¹⁸⁾ where a similar mean portal vein cross-sectional area value for liver cirrhosis patients was obtained (1.49 ± 0.49 cm²), which was significantly higher than that of controls (p = 0.001). This similarity is most probably attributable to portal hypertensive changes which result in slightly increased portal vein caliber in both study groups. The ALD subgroup had a significantly higher mean PVCSA than the CVH subgroup. The variable disease severity between the two subgroups might be responsible for this disparity. There were no significant differences in mean PVCSA and portal venous velocity between males and females in both groups (controls and CLD patients). The mean portal vein cross-sectional area (PVCSA) for controls was 1.07 ± 0.26 cm². This is similar to the mean PVCSA obtained by Moriyasu et al.⁽¹⁸⁾ (0.99 ± 0.28 cm²) and Aiyekomogbon et al.⁽²⁰⁾ (1.097 ± 0.20 cm²), and also close to the value reported by Ibinaiye et al.⁽³⁵⁾ (1.10 ± 0.20 cm²). The similar PVCSA values from these studies suggest that there might be no significant ethnic differences in PVCSA.

The CLD group had a mean PVCI of 0.104 ± 0.03 cm/sec, which was significantly higher than the value for controls. Of the two CLD subgroups, the ALD subgroup showed a significantly higher mean PVCI than the CVH subgroup. The reason ALD patients had a higher mean PVCI than CVH patients may be due to the poor health seeking behavior of this category of patients, leading to more advanced liver disease states at presentation. This is probably worsened by the higher number of males than females in this subcategory, as men in this society have been shown to have worse health seeking habits than women⁽³⁶⁾. Furthermore, non-compliance may also play a significant role, as alcoholics tend to show poor compliance with health intervention strategies, with higher rates of relapse to the detriment of their liver health.

Other studies also found a significant difference in PVCI between CLD cases and controls^{(19,26,37–39)}. Ehtisham et al.⁽³⁸⁾ showed a more significant statistical difference between the mean PVCI for controls and CLD subjects (0.05 ± 0.00 cm/sec vs. 0.14 ± 0.02 cm/sec respectively; p = 0.0001). However, the authors did not subcategorize the CLD participants into etiological subgroups to observe for PVCI differences therein.

Chakravarthy and co-researchers⁽¹⁹⁾, on the other hand, subdivided the CLD participants into three subcategories of ALD, CVH, and non-alcoholic fatty liver disease (NAFLD), and measured the PVCI in each of these subgroups. The PVCI values recorded for each of these conditions were significantly higher than the value obtained in their controls, i.e., medians of 0.027, 0.050 and 0.060 for NAFLD, CVH, and ALD, respectively. The absolute values in their study differed significantly from those of the index research possibly due to interobserver variability, differing Doppler angles, and racial differences.

Similarly, Iliopoulos et al.⁽²⁶⁾ recorded a mean PVCI of 0.06 ± 0.03 cm/sec for subjects in the CVH group. Unlike the index study, however, this was not significantly higher than the PVCI of the control group (0.05 ± 0.02 cm/sec) (p = 0.251). This discordance with the index study may be explained by the fact that Iliopoulos et al. further subdivided subjects with CVH into five subgroups (F1–F5) on the basis of histologic activity index (necroinflammatory score). Those in subgroups F1–F4 had lower necroinflammatory scores than those in the F5 group, who showed a more advanced stage of fibrosis; this was reflected in the PVCI, which was significantly higher in F5 than the F1–F4 stages (p = 0.041). This significant difference was, however, not reflected upon averaging of the subgroups into the composite CVH group due to a higher number of participants in the F1–F4 subgroups than the F5 subgroup.

Haag et al.⁽³⁷⁾ found the PVCI to be highly sensitive and specific in diagnosing elevated portal pressure. Similarly to this study, they reported a significant difference in PVCI between control and CLD groups. However, there was no attempt to correlate the PVCI with Class-Pugh classification in this study, as was done by Haag and colleagues⁽³⁷⁾. On the other hand, the authors did not investigate possible variations in PVCI based on etiology, as was done in the index study.

Bolognesi et al.⁽³⁹⁾ in Italy investigated the relationship between the etiology of cirrhosis and splanchnic hemodynamics in patients with CLD scheduled for orthotopic liver transplant. Only patients with ALD or CVH were included. Most of the splanchnic hemodynamic parameters evaluated, including the PVCI, did not show any significant difference between the two disease categories. This is in discordance with our study in which there was a statistically significant PVCI difference between the two groups. Furthermore, the mean PVCI recorded in their research was significantly higher than that recorded in the present study, most probably because the patients recruited by the authors had very advanced stages of liver disease requiring a transplant. This same reason could also account for why the PVCI values between the ALD and CVH groups in their study did not show a significant difference.

We encountered some limitations in this study. Firstly, due to the location of the portal vein, a large body habitus or significant bowel gas usually made scanning difficult. Oblique views were employed to overcome this limitation. The subjects whose portal veins could still not be assessed were excluded from the study. Secondly, we could not correlate the PVCI with the Child-Pugh classification.

In conclusion, the patients with CLD had a significantly higher PVCI than healthy controls. Similarly, the PVCI of the ALD subgroup was significantly higher than that of the CVH subgroup. The significant difference in PVCI between the control and CLD groups suggests that PVCI can be an adjunct parameter for detecting chronic liver diseases.