Hepatic artery duplex Doppler ultrasound in severe alcoholic hepatitis and correlation with Maddrey’s discriminant function

Patients and methods

The study subjects were 20 consecutive SAAH patients and 20 alcoholic cirrhosis patients without alcoholic hepatitis who served as controls. All patients were treated at the Gastroenterology Department, Government Medical College, Thiruvananthapuram, India, from September 2011 to September 2012. Diagnostic criteria for alcoholic hepatitis included chronic alcohol consumption of at least 20 g per day with a recent history of heavy alcohol ingestion, elevation of serum aspartate aminotransferase (AST) at least twice above the alanine aminotransferase (ALT) level, jaundice, coagulopathy, and at least two of the following features: tender hepatomegaly, presence of a hepatic bruit, fever (temperature >100°F), or leukocytosis (white blood cell count >10,000/mm3). The study included SAAH patients as defined by MDF >32 on admission. The control group included patients with clinically, radiologically, and biochemically proven alcoholic cirrhosis without alcoholic hepatitis and alcohol abstinent for at least 1 year. Exclusion criteria included patients with portal vein thrombosis, active gastrointestinal bleeding, active infection including spontaneous bacterial peritonitis, patients already on steroids or pentoxifylline, MDF <32 and patients in whom hepatic artery Doppler was not feasible due to anatomical variations.

Duplex Doppler ultrasound (Toshiba Nemio, Japan) examination was performed using transducers with a frequency of 3.5 MHz. All subjects had fasted for at least 8 h before scanning. Scans were performed with the patients in the supine position using both Gray-scale and color Doppler imaging to identify vascular landmarks. Doppler tracings were acquired during suspended respirations. The right hepatic artery was visualized at the porta hepatis as it crosses the main portal vein in all patients. At least three similar, sequential waveforms were obtained at measured angles of insonation of <60°. The best tracing was selected for analysis. RI and PI were then calculated as follows: RI = (PSV − end diastolic velocity)/PSV, and PI = (PSV − end diastolic velocity)/mean velocity. An average HAD was obtained using Gray scale imaging in the transverse plane. Patients in SAAH group had severity assessed by MDF on admission before initiating pentoxifylline or prednisolone.

Results

The clinical and laboratory characteristics of both study groups are presented in Table 1. All patients were male. The mean age of patients and duration of alcohol consumption did not differ significantly between the two groups. Fever and hepatic bruit were present in 12 SAAH patients, and in none of the cirrhotic patients. Ascites though more frequent in the alcoholic cirrhosis group was not significantly different between groups (P value >0.05), and hepatomegaly was significantly greater in SAAH group (P value <0.05) (Table 1).

Table 1 Baseline characteristics

Liver function test showed a statistically significant difference in serum bilirubin (P value<0.001), AST (P value <0.001) and ALT (P value <0.001) levels. Their mean values were higher in SAAH group compared with alcoholic cirrhosis group. The prothrombin time and albumin were comparable in two groups (P value >0.05). Leukocyte count (P value <0.001) and serum creatinine (P value<0.001) were also significantly higher in severe alcoholic hepatitis. group (Table 1).

Duplex Doppler parameters of HAD, RI and PI of both groups are summarized in Table 1. The mean hepatic artery diameter (HAD) in SAAH was 3.96±0.51 mm compared with cirrhotic patients with 2.86±0.41 mm (P<0.001) (Fig. 1). The mean RI in SAAH was lower with 0.49±0.08 compared with cirrhotic patients with 0.81±0.09 (P<0.001). The mean PI in SAAH was lower 1.67±0.13 compared with cirrhotic patients with 1.80±0.13 (P<0.003). The coefficient of variation (CV) for intraobserver variability was 7.4±4.2% for RI and 8.8±5.2% for PI; the corresponding figures for interobserver variability were 6.2±3.9% and 10.2±5.6%, respectively.

Figure 1 The mean HAD in SAAH was 3.96±0.51 mm compared to cirrhosis group with 2.86±0.41 mm which was statistically significant (P<0.001)
HAD, hepatic artery diameter; SAAH, severe acute alcoholic hepatitis

The mean MDF was 58. HAD showed a significant correlation with MDF (Table 2 and Fig. 2). Regression analysis showed that 36% of the variability in MDF could be independently predicted by HAD. Coefficients of regression analysis are provided in Table 3.

Table 2 Correlation of duplex Doppler ultrasound parameters to Maddrey’s discriminant function (MDF) in severe alcoholic hepatitis

Figure 2 In severe alcoholic hepatitis there was statistically significant co-relation between DF and HAD
MDF, Maddrey’s discriminant function; HAD, hepatic artery diameter

Table 3 Regression analysis with variables inserted and coefficients

Discussion

Duplex Doppler ultrasonography can noninvasively assess hepatic hemodynamics [6,7]. In our study, ultrasound Doppler revealed a significantly greater dilatation of the hepatic artery in SAAH compared to alcoholic cirrhosis patients (3.96±0.51 mm vs. 2.86±0.4 mm). Wing et al described a unique “parallel tubular” appearance of intrahepatic vessels in a cohort of alcoholic patients, which subsequently was demonstrated to represent dilated intrahepatic arteries [8]. In alcoholic hepatitis, intrahepatic artery dilation has been described by Sumino et al and termed as pseudoparallel channel sign characterized by dilated intrahepatic arterial branch with an adjacent portal venous tract. They observed PPCS in 90% of patients of AAH and in 23% of patients other than alcoholic liver disease. PPCS gave a sensitivity of 82%, specificity of 87% and diagnostic accuracy of 84% in diagnosing AAH [2,3]. Han et al showed quantitatively that there was a significant increase in HAD and peak systolic velocity in patients with AAH compared with cirrhotic patients and healthy subjects [9].

In our study, all patients were male. Experimental porcine model showed a significant difference in biomechanical properties of the hepatic artery with respect to gender [10]. There is a scarcity of clinical studies on hepatic artery hemodynamics with respect to gender.

A prospective multicenter study shows that severely altered portal flow is common in patients with severe AH and associated with poor outcome [11]. The hepatic arterial buffer response (HABR) is proposed to occur in conditions of decreased portal flow. HABR is an intrinsic regulatory mechanism of the liver to maintain total hepatic blood flow when portal perfusion decreases. Increased hepatic arterial blood flow mediated by adenosine washout in the portal triad compensates for the reduced portal blood flow [12,13]. Histological studies demonstrate the presence of dense sinusoidal collagen deposits in patients with AAH. Increased sinusoidal resistance impedes blood flow through the hepatic sinusoids. Ultimately, sinusoidal pressure builds until portal blood flow is compromised in a retrograde manner, with marked reduction in portal venous perfusion of the sinusoids. The reciprocal response of hepatic arterial blood flow to reduction in portal venous perfusion is well-established [12]. To preserve the perfusion to the liver, compensatory dilation of the hepatic artery occurs with subsequent increase in hepatic artery flow. The mean HAD was greater in our study as we studied only SAAH patients with MDF >32 compared to MDF range of 13.0-64.6 in the study by Han et al [9].

Hepatic bruit was present in 12 of 20 SAAH patients but none in the alcoholic cirrhosis group. In AAH, hepatic artery angiography demonstrates stretched and elongated hepatic arteries, whereas others describe long and tortuous arteries [13,14]. Hepatic bruit results from the high-frequency vibration of the perivascular tissue induced by turbulent flow through the tortuous hepatic artery, but is not specific for alcoholic hepatitis, however as it can also occur in hepatoma, arterio-venous fistula and hemangioma [15].

In our study, we found significantly lower RI and PI in SAAH compared to cirrhotic patients. Han et al in their study found RI and PI to be significantly lower in patients with AAH compared with healthy controls, although not significantly different from patients with decompensated alcoholic cirrhosis [9]. Their study did confirm an increase in hepatic arterial blood flow, evidenced by significantly higher PSV and HAD in patients with AAH when compared with cirrhotic patients and healthy controls. In the study by Colli et al a significant difference in RI and PI between patients with alcoholic hepatitis and alcoholic cirrhosis was found [16]. The variables RI and PI are indicators of vascular impedance. Ethanol itself may be responsible for hepatic artery dilatation as it has been seen in chronic alcoholic patients without liver damage [16]. However, progression of alcoholic hepatitis to cirrhosis profoundly impairs the hepatic arterial responsiveness as a result of fibrosis with a vascular distortion. In our study, the CV for intraobserver variability was 7.4±4.2% for RI and 8.8±5.2% for PI; the corresponding figures for interobserver variability were 6.2±3.9% and 10.2±5.6% respectively. The relevance of the hepatic artery RI in AAH has been challenged by Han et al, who reported a high variability of this sign in patients with liver disease due to factors such as vessel compliance and different states of fasting [9].

The most interesting finding of our study was the significant correlation of HAD with MDF of SAAH and that 36% of the variability in MDF could be independently predicted by HAD. Other Doppler parameters RI and PI showed no correlation with MDF. In the study by Han et al there was no correlation between the duplex Doppler ultrasound parameters and the severity of liver disease as assessed by MDF. We included only SAAH patients with mean MDF of 58 compared with mean MDF 34 (range of 13.0-64.6) in the study by Han et al. The SAAH group in our study had high mean MDF and mean serum creatinine indicating severe liver dysfunction.

In the past, hepatic artery caliber assessed angiographically has been studied as an index for liver function in chronic liver disease [17]. Severe alcoholic hepatitis is associated with major changes in portal flow defined as reversal or alternating flow in the portal trunk and/or in intrahepatic portal branches and is independently associated with poor survival. Resolution of AH can contribute to both the restoration of portal blood flow toward normal and the improvement in liver function [11]. In a similar manner, our study found hepatic artery dilatation to correlate with the severity of alcoholic hepatitis and it is reasonable to suggest that resolution of SAAH can result in regression of hepatic artery dilatation. In SAAH as portal flow decreases, HABR compensatory response tries to maintain hemodynamics, adequate liver function and metabolic homeostasis [18]. The increase in hepatic arterial blood flow however is capable of buffering up to 25-60% of the decreased portal flow [12,18].

Hepatic artery dilatation could also be due to high levels of endotoxinemia that has been documented among patients who have alcoholic hepatitis; probably because of increased intestinal permeability [19]. In models of continuous intravenous infusion of Escherichia coli in rats, portal venous flow was reduced, and increased hepatic arterial flow resulted in unchanged total hepatic blood flow. The increased hepatic arterial flow could be a result of an active HABR [20,21]. Sato et al showed the effect of portal pressure (shear stress) to the triggering of regeneration and suggested that there was an upper limit to a beneficial effect of elevated portal pressure. Furthermore, it has been demonstrated that prevention of shear stress after partial hepetectomy blocked the activation of regeneration cascade [22]. Hepatic artery dilatation could also be attributed to increase in oxygen consumption and attempted liver regeneration although altered immune response, presence of portocaval shunts and abnormal intrahepatic hemodynamics remain a major cause of reduced hepatic regenerative capacity. These hypotheses could explain why even in the absence of treatment, the 1-month spontaneous survival of patients with a MDF ≥32 fluctuates between 50% and 65% [4,5].

We did not study portal vein hemodynamics due to procedural time constraint and discomfort to SAAH study group. A hepatic artery Doppler study in a larger cohort of SAAH with follow-up scan after steroid treatment could be useful in assessing the prognostic value of this non-invasive tool.

In conclusion, duplex Doppler ultrasound parameters of the hepatic artery may be a useful tool in the diagnosis of SAAH in appropriate clinical setting and HAD may have a prognostic role in these patients.