| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Ultrasonography |
1 From the Department of Clinical and Experimental Medicine (M.B., D.S., G.B., C.M., A.G.) and the Institute of Radiology (M.C.C.), Clinica Medica 5, Policlinico Universitario, Via Giustiniani 2, 35128 Padova, Italy. Received November 3, 1999; revision requested December 7; revision received February 3, 2000; accepted February 7. Address correspondence to M.B. (e-mail: bolognes@ux1.unipd.it).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: In all patients with cirrhosis examined at one center over 4 years, the resistive (RI) and the pulsatility (PI) indexes in the right and left branches of the hepatic artery were evaluated with Doppler US. An APF was suspected when an RI decrease of at least 20% and a PI decrease of at least 30% were present in one hepatic lobe relative to values in the other lobe and portal blood flow in the lobe with the decreased values was reversed. The RI and PI in patients with an APF were compared with those in 75 patients with cirrhosis and without APFs at angiography.
RESULTS: Seven patients with an APF were identified. APFs suspected at Doppler US were always confirmed with angiography. The percent differences ± SD in the RI and the PI between the two intrahepatic branches of the hepatic artery in patients with versus in patients without an APF were as follows: RI, 35% ± 6 (range, 27%–42%) versus 5% ± 4 (range, 0%–15%) (P < .001); PI, 50% ± 5 (range, 41%–58%) versus 11% ± 7 (range, 0%–26%) (P < .001).
CONCLUSION: The intrahepatic arterial resistive and pulsatility indexes and the direction of portal blood flow should be evaluated in routine screening for APFs in patients with liver cirrhosis.
Index terms: Fistula, arteriovenous, 761.49, 952.494, 957.494 • Gastrointestinal tract, hemorrhage, 71.75, 952.717, 957.717 • Hepatic arteries, US, 952.12983 • Liver, cirrhosis, 761.794 • Liver, US, 761.12983 • Portal vein, US, 957.12983 • Ultrasound (US), Doppler studies, 761.12983, 952.12983, 957.12983
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Duplex Doppler ultrasonography (US) has already been demonstrated to be able to show the presence of APFs (3,6–8), but a screening diagnostic method to be applied during routine examinations has to our knowledge never been defined. Because the Doppler arterial resistive index and pulsatility index are decreased in patients with arteriovenous fistulas (9–13), we performed this study to determine the usefulness of the routine evaluation of the hepatic resistive index and pulsatility index and of the direction of portal blood flow for the diagnosis of APFs in patients with cirrhosis.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
All subjects were examined in the morning after an overnight fast. The arterial Doppler resistive indexes and pulsatility indexes, that is, the resistive indexes and the pulsatility indexes in the right and left branches of the hepatic artery, were measured with a model Sonolayer SSA-270A (Toshiba, Tokyo, Japan) or a model ATL-5000 (ATL, Seattle, Wash) color Doppler unit, as previously described (14,16). The resistive index and pulsatility index were calculated according to the following formulas (17): RI = (vmaxsys - venddias)/vmaxsys, and PI = (vmaxsys - venddias)/vmean, where RI is the resistive index, PI is the pulsatility index, vmaxsys is the maximum systolic velocity, venddias is the end diastolic velocity, and vmean is the mean velocity. Each result was the mean of three measurements. The presence of an APF was suspected when there was a difference greater than 20% in the resistive index and a difference greater than 30% in the pulsatility index between the two lobar hepatic arteries. The two cutoff values were chosen after examining the first patient with an APF; they were confirmed after analyzing the data from the second patient.
Moreover, the flow directions in the main trunk and in the two main branches of the portal vein were evaluated to verify the reversal of portal blood flow. Color Doppler US of the liver parenchyma was also performed to directly identify the APF, defined as an echo-free focal hepatic lesion, which contained fast and turbulent flow at Doppler and color Doppler US examinations.
All examinations were performed by equally skilled operators (M.B., D.S., G.B.), whose inter- and intraobserver reproducibilities have already been reported (14,18). When possible, evaluations of the arterial branch afferent to and of the vein draining the suspected APF were performed. Patients with APFs suspected at color Doppler US underwent an angiographic examination, both to confirm the diagnosis and, when necessary, to perform therapeutic embolization.
To analyze and quantify the decreases in the Doppler US resistive index and pulsatility index in the main arterial intrahepatic branch in the lobe with APF relative to the values in the other lobe, the differences in the Doppler resistive index and pulsatility index between the two main intrahepatic branches of the hepatic artery were compared with those in 75 patients with cirrhosis but without APFs, which was verified according to findings of angiography performed while the patients were waiting for liver transplantation. These 75 patients, who were part of the entire group of the investigated patients, were the candidates for liver transplantation consecutively examined in our center during the 4 years. Angiography was performed as part of the pretransplantation clinical evaluation.
All angiographic examinations (model DVI-II; Philips, Eindhoven, the Netherlands) were performed and all angiograms were interpreted by the same operator (M.C.C.) within 7 days of the US examination. In all cases, selective study of the proper hepatic artery was performed by injecting 15–20 mL of nonionic contrast medium (iohexol [Omnipaque; Nycomed Imaging, Oslo, Norway]) at a rate of 4–5 mL/sec; arterial and early and late parenchymal phases were evaluated after the acquisition of digitally subtracted images at a rate of one per second. APF was considered with opacification of one or more portal branches before or during the early parenchymal phase. APFs were identified as central perihilar, parenchymal, and subcapsular according to location near the porta hepatis, in the middle of the parenchyma, or in the periphery of the liver, respectively.
Results were expressed as means ± SDs. Differences among groups were evaluated by using the Student t test for paired and unpaired samples. The coefficients of variation of the measurements of the Doppler resistive index and pulsatility index were calculated in every patient with APF by dividing the SD by the mean. The null hypothesis was rejected if P values were less than .05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Patients without APFs
In patients with cirrhosis and without APF, the Doppler resistive indexes and pulsatility indexes were similar (P > .05) between the left and right lobes of the liver (Fig 1), which is in accordance with data already reported in the literature (14,15). The mean resistive index and the pulsatility index were 0.70 ± 0.07 and 1.25 ± 0.26, respectively, in the right branch and 0.71 ± 0.07 and 1.30 ± 0.29, respectively, in the left branch. Hepatic resistive indexes and pulsatility indexes were similar to those already reported (14,15,20). Differences in the Doppler arterial resistive indexes and pulsatility indexes between the two intrahepatic branches were 0.04 ± 0.03 (range, 0.00–0.11) and 0.14 ± 0.10 (range, 0.00–0.48), respectively (Fig 2). Analyzing the results as percent differences, the resistive index and pulsatility index differences between the two arterial branches were, respectively, 5% ± 4 (range, 0%–15%) and 11% ± 7 (range, 0%–26%) (Fig 3).
|
|
|
|
The direct identification of the APFs was possible with the color Doppler US examination of the liver parenchyma in all patients; however, in two patients, the presence of the APF was suspected at the beginning only because of the marked decrease in the arterial resistive index and pulsatility index in one hepatic lobe. The presence of the APF could have been overlooked in these two patients if the diagnosis had been based only on the direct visualization of the APF without the routine evaluation of the arterial resistive index and pulsatility index.
APFs were central perihilar in two patients, parenchymal in three patients, and subcapsular in the last two patients, according to their location. Resistive indexes and pulsatility indexes in the central part of the main intrahepatic branch of the artery were substantially the same independently of the location of the APF. (Resistive indexes were 0.47 and 0.48 for central perihilar APFs; 0.40, 0.40, and 0.52 for parenchymal APFs; and 0.40 and 0.46 for subcapsular APFs. Pulsatility indexes were 0.64 and 0.65 for central perihilar APFs; 0.52, 0.50, and 0.79 for parenchymal APFs; and 0.54 and 0.65 for subcapsular APFs.)
In five patients with APFs, the evaluation of the peripheral small artery afferent to the APF was possible. In these arteries, the resistive index and pulsatility index were even lower than those measured in the main lobar branch of the artery: The ranges of the resistive indexes and pulsatility indexes measured in the afferent arteries were, respectively, 0.31–0.45 and 0.37–0.59. In the main intrahepatic artery branch in the same lobe, the resistive indexes and pulsatility indexes were, respectively, 0.40–0.52 and 0.50–0.79.
The cause of the APF was identified in only five patients: One had had a previous blunt abdominal trauma, one had previously undergone catheterization of an intrahepatic bile duct, and three had undergone liver biopsy at the site where the APF was present—two for the diagnosis of the liver disease and one, the one with hepatocellular carcinoma, for the histologic diagnosis of a nodular hepatic lesion.
The three patients with the most serious upper gastrointestinal tract bleeding underwent embolization with Gianturco coils (21,22). In the first and second patients, the procedure was sufficient to occlude the APF, with the subsequent improvement (one patient) or disappearance (one patient) of esophageal varices, which before embolization kept causing relapse despite numerous sessions of sclerotherapy. In the third patient, excision of liver segments V and VI was necessary to achieve the cessation of esophageal bleeding.
Of the remaining four patients, two were on the list for hepatic transplantation; one, the one affected by hepatocellular carcinoma, did not undergo embolization because of poor clinical conditions (he died 1 month after the Doppler US evaluation); and the last patient did not undergo any particular procedure, because she had neither esophageal varices nor other serious signs of portal hypertension.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
APFs can be small and asymptomatic (1,23), but they often cause symptoms because of their consequences, which are an increase in portal pressure with subsequent esophageal varices and upper gastrointestinal tract bleeding and ascites (23,24). APFs are a reversible cause of portal hypertension; therefore, a correct diagnosis is pivotal because a definitive therapy can be proposed. However, the presence of an APF could go unnoticed in patients with liver cirrhosis, because portal hypertension is already present in these patients and worsening of the symptoms because of the onset of an APF could be misinterpreted as being due to a decrease in hepatic function that is suggested by the natural history of the disease. Moreover, patients with cirrhosis could be the patients with the highest incidence of APFs, because these patients often have transhepatic procedures, such as liver biopsy, proposed. For these reasons, in this study we chose to evaluate only patients with cirrhosis.
In patients with cirrhosis, the APF increases portal inflow, which is a relevant component in the pathophysiology of portal hypertension (25); thus, ascites and the degree of esophageal varices are worsened, with subsequent hemorrhage. Therefore, in patients with cirrhosis, the search for APFs is mandatory, because the APF can be treated, which reduces the complications of cirrhosis. Indeed, in the three patients in whom the APF was accompanied by recurrent upper gastrointestinal tract bleeding from ruptured esophageal varices resistant to endoscopic sclerotherapy, the treatment of APF was followed by the cessation of the bleeding. The presence of an APF should, therefore, be routinely looked for in all patients with cirrhosis, but a screening method has never been defined.
Color Doppler US is able to depict the presence of an APF, as a number of articles and case reports have shown (3,6–8,26–30). Therefore, color Doppler US can be proposed for the screening of APFs in patients with cirrhosis. This screening can be carried out simply, as patients with cirrhosis are usually periodically examined with Doppler US.
In particular, in every patient undergoing a Doppler US examination, the resistive index and pulsatility index in the left and right branches of the hepatic artery and the direction of blood flow in the left and right branches of the portal vein should always be evaluated. If the arterial resistive index and pulsatility index in a lobe are decreased by at least 30%–40% compared with the indexes in the other lobe and the blood flow in the intrahepatic branch of the portal vein in the lobe with the decreases is reversed while that in the other lobe is hepatopetal, the presence of an APF is highly probable.
As a consequence, an accurate color Doppler US examination of the liver parenchyma to look for the APF is mandatory. We think that this approach could unveil the presence of APFs, which may be overlooked by considering only direct APF depiction with color Doppler US. Moreover, it should be stressed that with simple real-time sonography an intraparenchymal APF could be easily misinterpreted as a small cystic or inhomogeneous area.
In this article, findings of a quantitative evaluation of the intrahepatic arterial resistive index and pulsatility index in patients with cirrhosis and an APF are reported and a systematic method for screening for APFs in patients with cirrhosis is proposed. The criteria listed in this article are easily assessable in all Doppler US laboratories. The opposite directions of blood flow in the two intrahepatic branches of the portal vein is a finding that is never present in healthy patients, but this single finding could not be relied on for the diagnosis of APF. Indeed, in patients with cirrhosis, a reversed flow in a single branch of the portal vein may be spontaneously present, even though rarely, or may be associated with a patent paraumbilical vein or with an intrahepatic partial portal vein thrombosis. Therefore, the evaluation of the intrahepatic arterial resistive index and pulsatility index is fundamental for the diagnosis of the APF.
In our patients, there was no overlap between the arterial resistive or pulsatility index percent differences in the two branches of the hepatic artery for patients with versus for patients without an APF. Indeed, the maximum percent differences in the resistive index and the pulsatility index in patients with cirrhosis and without an APF were, respectively, 15% and 26%, while the minimum differences in the resistive index and the pulsatility index in patients with cirrhosis and with an APF were, respectively, 27% and 41% (Fig 3).
Resistive index measurement in the hepatic artery has been reported to have intrinsic sampling variability (31), with an SD for a single measurement of 0.08. Therefore, small changes in these indexes should be interpreted with caution. On the other hand, if the mean of three measurements is used, as it was in our study, the variability decreases by a factor of three (31). Moreover, in the study by Paulson et al (31), resistive indexes were calculated by interrogating the proper hepatic artery near the point where it crosses the main portal vein, while in our study arterial indexes were obtained by interrogating the two main intrahepatic branches (14). At this level, resistive index measurement has no relevant interobserver variability when performed by specifically trained operators (32).
It could be hypothesized that a more central APF is likely to cause greater hemodynamic arterial changes than a more peripheral one. However, in our patients, no differences in resistive indexes and pulsatility indexes according to the location of the APF were apparent. We think that it is possible that a small APF some distance from the site of evaluation in the hepatic artery may not be detected by measuring the Doppler arterial resistive index and pulsatility index, but our data do not support this hypothesis.
Echo-Doppler US parameters seem to be accurate in the diagnosis of APF. Because the presence of an APF was always confirmed with angiography in patients who fulfilled the mentioned criteria, the method had 100% specificity, with no false-positive findings. However, the sensitivity of the method could not be assessed, because obviously angiography could not be performed in all patients. However, because no false-negative findings were present in the 75 patients who underwent angiography, the sensitivity of the method also seems to be good.
In conclusion, color Doppler US is a reliable method for diagnosing APFs in patients with cirrhosis. The evaluation of the intrahepatic arterial resistive index and pulsatility index and of the direction of portal blood flow in the liver should be used for routine screening for APFs in patients with liver cirrhosis.
|
FOOTNOTES |
|---|
Author contributions: Guarantors of integrity of entire study, M.B., A.G.; study concepts and design, M.B., D.S.; definition of intellectual content, M.B., D.S., C.M.; literature research, M.B.; clinical studies, M.B., D.S., G.B., M.C.C.; data acquisition, M.B., D.S., G.B., M.C.C.; data analysis, M.B., D.S., M.C.C., C.M.; statistical analysis, M.B., C.M.; manuscript preparation, M.B.; manuscript editing, M.B., D.S.; manuscript review, M.B., D.S., M.C.C., C.M.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  O. Catalano and A. Siani Peripheral Arterioportal Fistula: Demonstration With Contrast-Enhanced Sonography J. Ultrasound Med., July 1, 2005; 24(7): 1027 - 1030. [Full Text] [PDF]
|
|
|
|
|
|
G. Bodner, S. Peer, M. Karner, R. Perkmann, B. Neuhauser, W. Vogel, and W. Jaschke Nontumorous Vascular Malformations in the Liver: Color Doppler Ultrasonographic Findings J. Ultrasound Med., February 1, 2002; 21(2): 187 - 197. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. H. Wachsberg, P. Bahramipour, C. T. Sofocleous, and A. Barone Hepatofugal Flow in the Portal Venous System: Pathophysiology, Imaging Findings, and Diagnostic Pitfalls RadioGraphics, January 1, 2002; 22(1): 123 - 140. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Tochio, M. Kudo, S.-i. Nishiuma, and Y. Okabe Intrahepatic Spontaneous Retrograde Portal Flow in Patients with Cirrhosis of the Liver: Reversal by Food Intake Am. J. Roentgenol., November 1, 2001; 177(5): 1109 - 1112. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
) versus in patients with cirrhosis and with an APF (
). The differences are significantly higher in patients with an APF.
