Reduction of Renal Cortical Blood Flow Assessed by Dopplerin Cirrhotic Patients With Refractory Ascites

ROBERTO RIVOLTA,1 ALESSANDRA MAGGI,3 MASSIMO CAZZANIGA,3 DANIELA CASTAGNONE,1 ANNA PANZERI,3 DANIELA SOLENGHI,3

ELETTRA LORENZANO,3 FILIPPO QUARTO DI PALO,2 AND FRANCESCO SALERNO3

The usefulness in cirrhotic patients of hemodynamicmeasurements by Doppler ultrasonography (US) is still notdefined. We investigated the relationships between Dopplermeasurements and the severity of ascites. Portal blood flowvelocity and volume, and hepatic and renal arterial resis-tance indexes (RI) were measured in 57 cirrhotic patients(19 without ascites, 28 with responsive ascites, and 10 withrefractory ascites) and 15 healthy controls. The renalarterial RI were obtained for the main renal artery, interlo-bar vessels, and cortical vessels. Cirrhotic patients haddecreased portal blood flow and an increased congestionindex (CI). Only the CI was correlated to the severity ofascites, showing that it is also a reliable measure of theseverity of portal hypertension in patients with ascites. Thehepatic and renal artery RI were increased in cirrhoticpatients, and the two values were correlated (r 5 .68; P 5.00001). The RI of renal interlobar and cortical vessels werehigher in patients with refractory ascites than in patientswithout ascites (P F .02 and P F .009), and correlated withsodium excretion rate (r 5 2.45; P F .003), the renin-aldosterone system, and creatinine clearance (r 5 2.62;P F .0002). The RI decreased from the hilum of the kidneyto the outer parenchyma in healthy subjects and patientswith responsive ascites, but this difference disappeared inpatients with refractory ascites. This indicates that thedegree of renal vasoconstriction varies in different areasaccording to the severity of the ascites. Cortical vessels areinvolved mainly in patients with refractory ascites, suggest-ing that the intrarenal blood flow distribution in cirrhosistends to preserve the cortical area and that severe corticalischemia is a feature of refractory ascites. (HEPATOLOGY 1998;28:1235-1240.)

Doppler ultrasonography (US) is a noninvasive tool for theassessment of vascular patency. It has been used to measurethe hepatic arterial and venous flows of patients with portalhypertension1-4 and to document the increases in renal

resistances that occur in some cirrhotic patients.5-8 That someof these measures have prognostic value has been demon-strated.4,7-9 Although a diagnostic gray-scale US is widelyemployed in the evaluation of cirrhotic patients, Doppler israrely used. One pending problem is to establish whichDoppler measurements correlate best with the differentcomplications of portal hypertension. This could also help todetermine whether or not the Doppler is useful in monitoringthe effects of pharmacological therapies.

The aim of the present study was to correlate the Dopplermeasurements of portal blood flow and of hepatic and renalarterial resistances with the presence and severity of ascitesand renal failure in cirrhotic patients.

PATIENTS AND METHODS

Fifty-seven cirrhotic patients, 35 men and 22 women, admittedconsecutively to our hospital were enrolled in the present study.Their mean age was 57 6 9 years (range, 37-73 years), and theirmean body weight was 68 6 14 kg (range, 42-111 kg). The diagnosisof cirrhosis was based on clinical, laboratory, and, for 25 cases,histological findings. The etiology of cirrhosis was alcohol-relatedfor 15 patients, viral for 35 (31 hepatitis C virus–related and 4hepatitis B virus–related), mixed for 5, and cryptogenic for 2. Fivepatients had patent paraumbilical veins, and 6 others had completethrombosis of the portal vein. Patients with recent episodes ofgastrointestinal bleeding (n 5 2), encephalopathy (n 5 1), orbacterial infection (n 5 2) were excluded, as were those with arterialhypertension, primary cardiac (n 5 1), renal (n 5 1), and endocrinediseases (n 5 1). Nineteen patients were in Child-Pugh class A, 18were in class B, and 20 were in class C.10 The patients were dividedon the basis of presence and severity of ascites: 19 had never hadascites or edema, 28 had ascites that was responsive to therapy, and10 had refractory ascites as defined by Arroyo et al.11 Fifteen healthysubjects (normal volunteers; 8 men and 7 women; mean age, 44 611 years; range, 26-59 years) without history of personal or familialarterial hypertension or diabetes served as controls. All subjectswere studied after an overnight fast, and medications had beenstopped at least 4 days before the study. Healthy subjects andcirrhotic patients without ascites consumed free diets, and patientswith ascites were prescribed a low-sodium diet (60 meq/d). In-formed consent was obtained from each subject, and the studyprotocol conformed with the 1975 Declaration of Helsinki.

Measurement of Doppler Variables. After 1 hour of supine rest in asuitable environment (20°C-22°C), arterial pressure was recordedand mean arterial pressure calculated as diastolic pressure 1 1/3pulse pressure. Color-Doppler US examination of the liver and rightkidney was performed with an ATL Ultramark-9 HDI scanner usinga 2- to 4-MHz broadband convex probe (Advanced TechnologyLaboratories, Bothell, WA). To reduce the variability of the results,Doppler measurements were always made by the same investigator(R.R.), who was unaware of the patients’ laboratory data. Theoutcome arterial variables were resistance index (RI) measured onthe main branch of the arteria epatica propria and the main renal,

Abbreviations: US, ultrasonography; RI, resistance index; CI, congestion index;TAMV, time averaged mean velocity; PRA, plasma renin activity; PA, plasma aldosterone.

From 1Servizio di Radiologia e 2Divisione di nefrologia e dialisi, Istituto di Ricovero eCura a Carattere Scientifico (IRCCS) Policlinico, and 3Dipartimento di MedicinaInterna, Universita degli Studi di Milano, Milano, Italy.

Received October 31, 1997; accepted June 11, 1998.Alessandra Maggi and Anna Panzeri are recipients of a grant from IRCCS Policlinico

di Milano.Address reprint requests to: Francesco Salerno, M.D., Dipartimento di Medicina

Interna, Centro A. Migliavacca per lo Studio e la Cura delle Malattie di Fegato,Universita degli Studi, Via della Pace, 9, 20122 Milano, Italy. Fax: 39-2-59902598.

Copyright r 1998 by the American Association for the Study of Liver Diseases.0270-9139/98/2805-0001$3.00/0

1235

interlobar, and cortical arteries of the right kidney. RI was preferredto pulsatility index because they are equivalent,5,6,12,13 and RI is asimpler measurement. Doppler US was performed at three differentlevels of the renal arterial tree at which the RI values differ in healthysubjects.14 RI was determined from the spectral waveforms accord-ing to the following formula: [peak systolic frequency shift 2 lowestdiastolic frequency shift]/peak systolic frequency shift.15 Each RIwas calculated as the mean value obtained from 5 waveforms. The RIof kidney cortical vessels was obtained by color imaging, positioningthe exploring sample volume externally to the visualized arcuatevessels. The result was the mean of 6 waveforms (2 measured on theupper, 2 on the medium, and 2 on the lower pole of the organ).Doppler sample volume was set at 2.5 mm. The intraexaminationcoefficient of variation of RI was 3.5% 6 2.1%. All RI values arereported 3 100. Normal renal RI values previously obtained in 90normal volunteers ranged between 48 and 67 (main renal artery), 47and 64 (interlobar arteries), and 41 and 61 (cortical arteries).

The outcome venous variables were mean velocity of portal bloodflow, portal blood flow, and the congestion index (CI). A rectilinearsegment of the common tract of the portal vein was visualized in theB mode as near as possible to the bifurcation. The sample volumewas centered inside the vessel, and the beam completely insonatedthe flowing blood. The angle correction cursor was parallel to thedirection of flow, with an angle ,60° between the vessel and theultrasound beam. The time-averaged mean velocity (TAMV) (cm/sec) and portal diameters (D1 and D2) were measured during asustained shallow inspiration. Portal blood flow (mL/min) wascalculated as mean velocity 3 cross-sectional area (elliptic area 5(D1D2/4). The CI was calculated according to Moriyasu et al.16,17 ascross-sectional area (cm2)/portal blood flow mean velocity (A/TAMV).For a portal vein with diameters measuring 0.6 and 0.8 cm and aTAMV of 20 cm/s, the cross-sectional area is 3.14 3 0.6 3 0.8/4 50.3768, and the CI is 0.3768/20 5 0.0188.

Each result was the mean of five repeated measurements. Dopplerwaveforms were obtained at the lowest pulse repetition frequencypossible without aliasing. This maximized the size of the Dopplerspectrum and decreased the percentage of error of the measure-ments.

Liver and Kidney Function. On the day following the US investiga-tion, standard liver function tests were performed, and plasma andurinary electrolytes, serum creatinine, and endogenous creatinineclearance, plasma renin activity (PRA), and plasma aldosterone (PA)were measured. After an overnight fast and 2 hours of bed rest, eachsubject emptied the bladder spontaneously, and urine was discarded.Then, he/she received an oral load of tap water (10 mL/kg bodyweight) and remained supine for the following 3 hours, after whichspontaneously voided urine was again collected. The efficiency ofurine collection was confirmed by US monitoring of bladdervolumes after each voiding. At mid-point of the clearance period,blood samples were collected to measure creatinine, electrolytes,PRA, and PA. Creatinine clearance was calculated by standardformulas. PRA and PA were determined by specific radioimmunoas-says using commercial kits (Biodata, Milan, Italy; and Ares Serono,Rome, Italy) and previously described methods.18 Normal values forPRA and aldosterone in our laboratory are, respectively, 0.4 to 2.2ng/mL/h and 35 to 160 pg/mL. Within 10 days, the cirrhotic patientsalso underwent endoscopy to visualize esophageal or gastric varices.

Statistical Analysis. Data are reported as means 6 SD. To analyzedifferences between different groups of patients, a nonparametrictest (Kruskal-Wallis for independent groups) was used. When theresults of the analysis were significant, the groups were comparedwith the Wilcoxon rank sum test, with the a error correctedaccording to Bonferroni’s method. A two-way ANOVA was used toanalyze differences in renal RI according to different sites ofsampling and to different groups of patients. Spearman’s rankcorrelation coefficient was used to establish the degree of correlationbetween continuous variables. P , .05 was considered statisticallysignificant. Results were analyzed using SPSS software (SPSS Inc.,Chicago, IL, 1990).

RESULTS

Table 1 compares the whole group of cirrhotic patientswith that of healthy subjects, while Table 2 compares thethree subgroups of cirrhotic patients. Measurements obtainedon hepatic artery and renal vessels are from all 57 cirrhoticpatients, while values obtained on portal vein were for only46 patients without total portal vein thrombosis or patency ofthe paraumbilical vein. Cirrhotic patients had significantlyhigher values of hepatic artery RI, portal vein CI, and renalvessel RI, and significantly lower values of portal blood meanvelocity and flow than healthy subjects (Table 1). The hepaticartery RI was similar for patients with portal vein thrombosisand patients with patent portal vein (72 6 7 vs. 71 6 9; P 5.84) and for patients with patent paraumbilical veins andpatients without (76 6 7 vs. 72 6 7; P 5 .18), whereas the RIof the renal cortical vessels was significantly higher forpatients with portal vein thrombosis than patients withpatent portal vein (68.8 6 2.8 vs. 60.8 6 7.5; P , .001), and

TABLE 1. Values of Doppler Measurements Obtained in 15 HealthyControls and 57 Cirrhotic Patients

Outcome Healthy Cirrhotics P

Hepatic artery RI (3100) 58.4 6 2.1 72.8 6 7.8 .00001Portal blood mean velocity (cm/s)* 19.6 6 5.7 10.5 6 3.6 .00001Portal blood flow (mL/min)* 1,337 6 454 859 6 502 .0027CI* 0.06 6 0.03 0.15 6 0.09 .0016Renal artery RI (3100) 55.2 6 3.3 67.7 6 8.7 .00001Renal interlobar RI (3100) 52.6 6 4.5 65.6 6 7.0 .00001Renal cortical RI (3100) 49.1 6 3.6 62.7 6 7.9 .00001Mean arterial pressure (mm Hg) 92 6 7 89 6 10 .07

NOTE. Data are means 6 SD.*Portal hemodynamic values are only for patients without total portal vein

thrombosis and without patency of the umbilical veins (n 5 46).

TABLE 2. Values of Doppler Measurements and Related Variablesfor Cirrhotic Groups According to Severity of Ascites

Ascites

NO Responsive Refractory P*

Hepatic artery RI (3100) 69.7 6 5.8 74.6 6 8.3† 72.3 6 7.8† .03Portal blood flow mean

velocity (cm/sec)§ 11.7 6 2.9 10.8 6 4.1 8.4 6 2.3 .11Portal blood flow

(mL/min)§ 874 6 517 853 6 520 570 6 450 .34CI§ 0.11 6 0.05 0.14 6 0.06 0.23 6 0.17† .05Renal artery RI (3100) 63.3 6 8.3 69.3 6 8.1† 71.7 6 8.1† .04Interlobar RI (3100) 63.3 6 6.5 65.7 6 7.4 70.3 6 4.4† .02Cortical RI (3100) 60.8 6 7.3 61.9 6 8.1 69.3 6 4.8‡ .009Creatinine clearance

(mL/min) 105 6 27 79 6 79 54 6 22† .01Urine sodium excretion

rate (µeq/min) 102 6 56 40 6 52 27 6 52‡ .005PRA (ng/mL/h) 2.4 6 3.4 7.7 6 8.7† 30 6 18.5‡ .0002PA (ng/mL) 132 6 67 429 6 360† 1,253 6 202‡ .00001Mean arterial pressure

(mm Hg) 93 6 11 87 6 9 85 6 9 .10

NOTE. Data are means 6 SD.*Kruskal-Wallis ANOVA followed by the Mann-Whitney test with Bonfer-

roni’s correction.†P , .05 and ‡P , .01 compared with the values in the first column (no

ascites).§Portal hemodynamic values only for patients without total portal vein

thrombosis and without patency of the umbilical veins (n 5 46).

1236 RIVOLTA ET AL. HEPATOLOGY November 1998

portal blood flow was significantly higher for patients withparaumbilical veins than those without (1,515 6 461 vs.859 6 502 mL/min; P 5 .006). On comparing the threesubgroups of cirrhotic patients, it was evident that patientswith refractory ascites had the lowest values of creatinineclearance and urinary sodium excretion rate and the highestvalues of PRA and PA (Table 2). The RI measured on thehepatic artery was significantly higher for patients withascites than for those without ascites, and correlated slightlywith PRA (r 5 .40; P , .01) and PA (r 5 .45; P , .002).Portal blood flow mean velocity and portal blood flow did notdiffer among the subgroups of patients with different severi-ties of ascites, whereas CI was higher for patients withrefractory ascites than for patients without ascites (Table 2).

The main renal artery RI was significantly higher for thetwo subgroups of patients with ascites than for patientswithout ascites (Table 2) and correlated directly with thehepatic artery RI (r 5 .68; P 5 .00001) (Fig. 1). In 25% ofcirrhotic patients (95% CI: 11%-39%), however, the hepaticartery RI was increased despite a still-normal renal RI (Fig. 1).

The RI of interlobar and cortical vessels were significantlyhigher for patients with refractory ascites than for thesubgroup of patients without ascites (Table 2). When the RIobtained from different areas of the renal parenchyma waslooked at, it was evident that RI of the cortical vessels isslightly affected in patients with ascites responsive to therapy,but is markedly increased in patients with refractory ascites.The RI of renal cortical vessels correlated positively with PRA(r 5 .41; P , .05), PA levels (r 5 .72; P , .001), and serumcreatinine (r 5 .39; P , .004), and inversely with creatinineclearance (r 5 2.62; P , .0002) (Fig. 2) and urinary sodiumexcretion rate (r 5 2.45; P , .003). Similar figures were

obtained when the RI of the interlobar arteries were consid-ered (data not shown).

Of the 46 cirrhotic patients with increased renal cortical RI,20 (43% [95% CI: 29%-57%]) had normal values of creati-nine clearance (Fig. 2).

Figure 3 shows that the RI measured on different renalareas were significantly different for healthy subjects andpatients with ascites responsive to therapy (decreasing fromthe hilum toward the outer parenchyma), while such differ-ences disappeared for patients with refractory ascites.

DISCUSSION

Cirrhosis is characterized by complex changes of systemicand splanchnic hemodynamics. High cardiac output and lowperipheral resistances are the features of the so-called hyper-dynamic syndrome.19 Vasodilation occurs mainly, but notonly, in the splanchnic area, where, in concert with increasedresistance to venous outflow, it causes blood congestion.20,21

By contrast, in other areas such as the kidney, the resistance toarteriolar flow increases, and blood perfusion of the organdecreases progressively.22,23

In the present study, Doppler US showed that in cirrhoticpatients, resistances to hepatic and renal arterial blood flowsare increased. This finding is in agreement with the experimen-tal findings of Fernandez-Munoz et al.24 and Groszmann etal.25 in rats with CCl4-induced cirrhosis. The increase ofarterial resistances can be caused by local factors, such asliver fibrosis and regenerative nodules or release of localvasoconstrictors.26,27 Because peripheral resistances are deter-mined by a delicate local balance between vasoconstrictorsand vasodilators, variabilities of the regional release ofvasodilators (i.e., nitric oxide, prostacyclin) cannot be ex-cluded.28,29

The RI we measured for the hepatic artery correlated

FIG. 1. Relationship between the hepatic artery resistance index and therenal artery resistance index for 15 healthy controls (s) and 57 cirrhoticpatients (d). Dashed lines delimit the range of normal values (mean 1 2 SD).(s), Controls; (d), cirrhosis.

FIG. 2. Relationship between creatinine clearance and renal cortical RIfor 57 cirrhotic patients. Dashed lines delimit the ranges of normal values ofthe two variables.

HEPATOLOGY Vol. 28, No. 5, 1998 RIVOLTA ET AL. 1237

directly with the RI measured for the kidney artery. Thissuggests that, in addition to local factors, a general vasocon-stricting activity (i.e., activation of the sympathetic nervoussystem and renin-angiotensin system) may contribute toincrease the resistances of the two vascular beds simulta-neously. This could operate in patients with ascites, in whomincreased sympathetic nervous system activity has beendocumented.30

Increased hepatic artery RI was previously reported forcirrhotic patients by Sacerdoti et al.12 Unlike them, however,we did not find differences of the hepatic artery RI forpatients with portal vein thrombosis and those with patencyof the portal vein. However, some discrepancies of the hepaticartery RI measured in patients with total portal vein occlusionmay be simply a result of chronological differences, i.e., acuteversus chronic thrombosis.

Our study also shows that the hepatic artery RI was higherfor patients with ascites than those without ascites. This isnot surprising, because increased arterial resistance could bea physiological homeostatic response to the vascular underfill-ing occurring in ascitic patients.31,32 This can also explainwhy the RI measured on hepatic and renal arteries wererelated inversely to glomerular filtration rate and urinarysodium excretion rate, and directly to PRA and PA levels.Differences in PRA and PA levels between patients withcompensated and decompensated cirrhosis could have beenmagnified by the different amounts of sodium consumed withthe diet.

As for portal blood perfusion, we found that the CI of theportal tract was significantly higher for patients with refrac-tory ascites than for the other patients. On the contrary, portalblood flow velocity and volume, although lower for cirrhotic

patients than for healthy controls, was not related to thepresence or the severity of ascites. This confirms that the CI isthe best Doppler measurement assessing the severity of portalhypertension.8,16,33 We found portal blood flow to be higherfor patients with patent paraumbilical vein as previouslyreported by Sacerdoti et al.34 This indicates that the portalblood flow is greatly influenced by the presence of collateralcirculation and may further explain why it provides inaccu-rate estimates of the degree of portal hypertension. Thislimitation may be partly overcome by subtracting the paraum-bilical vein blood flow from the portal blood flow.

The RI measured on renal vessels was always higher forcirrhotic patients than for healthy subjects. When our pa-tients were stratified according to the presence and theseverity of ascites, we found that the RI obtained on interlo-bar and cortical small vessels were increased mainly forpatients with refractory ascites. This is in line with the resultsof other investigators5-8 who have shown that a high RImeasured on intrarenal arteries may predict the developmentof hepatorenal syndrome and a short survival. Changes of RIvalue in parenchymal vessels could be hypothetically causedby several phenomena, because RI is the result of bothupstream and downstream vascular patency. Because theexistence of stenosis of the main renal artery was alwaysexcluded by the Doppler evaluation, the increased RI wefound at interlobar and cortical levels may be explained by analtered downflow vascular tone or patency.

One of the main findings of our study was that the increaseof RI for patients with ascites is heterogeneous with respect todifferent parts of the renal vascular bed (Fig. 3). The RImeasured on the main renal artery was increased similarly inboth subgroups of patients with ascites as compared withpatients without ascites, whereas the RI of the cortical vesselswas significantly increased only in patients with refractoryascites. Therefore, the absence of a decrease of RI measuredon different areas, from the hilum toward the outer paren-chyma, was a feature of patients with refractory ascites.Because renal RI is inversely related to organ perfusion,35 thismight indicate that the intrarenal blood flow distribution incirrhotic patients preserves the flow to the cortex until theascites becomes refractory. This preservation could resultfrom intrarenal mechanisms of blood flow regulation. Whenthis regulation fails, cortical ischemia causes azotemia andoliguric renal failure,22 and, if unresolved, tubular necro-sis.22,36-38 This pattern is reminiscent of the one causingvasoconstriction of the renal cortical vessels during renaltransplant rejection.39

As depicted in Fig. 2, the RI measured on renal corticalarteries was elevated in many patients with normal values ofcreatinine clearance (43% of the whole group of patients withincreased RI). The discrepancy between Doppler measure-ments and creatinine clearances can be explained by theexistence of a glomerular filtration rate–sparing mechanismthrough increased filtration fraction in some patients and insome others by the inaccuracy of creatinine clearance as ameasure of glomerular filtration rate.40,41 However, it isknown that renal hemodynamic changes begin early in thecourse of liver disease–related functional kidney failure, evenbefore changes in creatinine clearance are detectable.37 Also,Maroto et al.7 showed that several cirrhotic patients withnormal creatinine values had increased RI, and Platt et al.8

FIG. 3. Mean values and SEM of kidney RI measured on the main renalartery, interlobar, and cortical vessels of healthy controls and cirrhoticpatients with ascites. (h), Healthy controls (n 5 15); (d), patients withascites responsive to therapy (n 5 28); (s), patients with refractory ascites(n 5 10).

1238 RIVOLTA ET AL. HEPATOLOGY November 1998

showed that cirrhotic patients with normal creatinine levelsbut with increased renal RI are at higher risk of kidney failurethan those with normal RI. Thus, Doppler investigation of thearterial kidney perfusion gives unique information on renalinvolvement in cirrhotic patients. It is easier to perform thantime-consuming clearances, and its use is also favored by itsreproducibility,5 safety, and ease of repetition as comparedwith invasive arteriography.

Another interesting finding of the present study was that RIon renal cortical vessels was markedly higher for patientswith portal vein thrombosis than for those with patent portalveins. A possible relationship between the blunting of portalvein blood flow and the increase of intrarenal resistancesmerits further investigations.

In conclusion, this study confirms that Doppler US is avalid noninvasive method to study cirrhotic patients. The CImeasured on the portal tract is a reliable variable forestimating the severity of portal hypertension. The RI mea-sured on renal vessels is related to the severity of ascites and,inversely, to renal function. RI measured on different renalvascular beds can help to directly estimate the variablepatterns of intraparenchymal hemodynamic derangementoccurring in the kidney of cirrhotic patients. Whether or notsuch Doppler measurements would also be useful for monitor-ing the effects of pharmacological therapy for renal dysfunc-tion should also be investigated further.

REFERENCES

1. de Vries PJ, Hoekstra JB, de Hooge P, van Hattum J. Portal venous flowand follow-up in patients with liver disease and healthy subjects.Assessments with duplex Doppler. Scand J Gastroenterol 1994;29:172-177.

2. Gaiani S, Bolondi L, LiBassi S, Santi V, Zironi G, Barbara L. Effect of mealon portal hemodynamics in healthy humans and in patients with chronicliver disease. HEPATOLOGY 1989;9:815-819.

3. Lomas DJ, Britton PD, Summerton CB, Seymour CA. Duplex Dopplermeasurements of the portal vein in portal hypertension. Clin Radiol1993;48:311-315.

4. Nishihara K, Sakata K, Yagyu T, Nakashima K, Suzuki T. Relationshipbetween peripheral portal blood flow and liver function in patients withliver cirrhosis. Pulsed Doppler ultrasonographic study. Scand J Gastro-enterol 1994;29:859-864.

5. Sacerdoti D, Bolognesi M, Merkel C, Angeli P, Gatta A. Renal vasoconstric-tion in cirrhosis evaluated by duplex Doppler ultrasonography. HEPATOL-OGY 1993;17:219-224.

6. Colli A, Cocciolo M, Riva C, Martinez E. Abnormal renovascularimpedance in patients with hepatic cirrhosis: detection with duplex US.Radiology 1993;187:561-563.

7. Maroto A, Gines A, Salo J, Claria J, Gines P, Anibarro L, Jimenez W, et al.Diagnosis of functional kidney failure in cirrhosis with Doppler sonogra-phy: prognostic value of resistive index. HEPATOLOGY 1994;20:839-844.

8. Platt JF, Ellis JH, Rubin JM, Merion RM, Lucey MR. Renal duplexDoppler ultrasonography: a noninvasive predictor of kidney dysfunctionand hepatorenal failure in liver disease. HEPATOLOGY 1994;20:362-369.

9. Zoli M, Iervese T, Merkel C, Bianchi G, Magalotti D, Marchesini G, GattaA, et al. Prognostic significance of portal hemodynamics in patients withcompensated cirrhosis. J Hepatol 1993;17:56-61.

10. Pugh RNH, Murray-Lyon IM, Dawson JL, Pietroni MC, Williams R.Transection of the esophagus for bleeding esophageal varices. Br J Surg1973;60:646-664.

11. Arroyo V, Gines P, Gerbes AL, Dudley FJ, Gentilini P, Laffi G, ReynoldsTB, et al. Definition and diagnostic criteria of refractory ascites andhepatorenal syndrome in cirrhosis. HEPATOLOGY 1996;23:164-176.

12. Sacerdoti D, Merkel C, Bolognesi M, Amodio P, Angeli P, Gatta A.Hepatic arterial resistance in cirrhosis with and without portal veinthrombosis: relationships with portal hemodynamics. Gastroenterology1995;108:1152-1158.

13. Petersen LJ, Petersen JR, Ladefoged SD, Mehlsen J, Jensen HA. Thepulsatility index and the resistive index in renal arteries in patients with

hypertension and chronic renal failure. Nephrol Dial Transplant 1995;10:2060-2064.

14. Rivolta R, Mascagni B, Berruti V, Quarto di Palo F, Elli A, Scorza R,Castagnone D. Renal vascular damage in systemic sclerosis patientswithout clinical evidence of nephropathy. Arthritis Rheum 1996;39:1030-1034.

15. Taylor KJW, Holland S. Doppler US. Part I. Basic principles, instrumenta-tion and pitfalls. Radiology 1990;174:297-307.

16. Moriyasu F, Nishida O, Ban N, Nakamura T, Miura K, Sakai M, Miyake T,et al. Measurement of portal vascular resistance in patients with portalhypertension. Gastroenterology 1986;90:710-717.

17. Moriyasu F, Nishida O, Ban N, Nakamura T, Sakai M, Miyake T, UchinoH. ‘‘Congestion Index’’ of the portal vein. Am J Roentgenol 1986;143:735-739.

18. Salerno F, Badalamenti S, Incerti PL, Tempini S, Restelli B, Bruno S,Bellati G, et al. Repeated paracentesis and i.v. albumin infusion to treattense ascites in cirrhotic patients: a safe alternative therapy. J Hepatol1987;5:102-108.

19. Groszmann RJ. Hyperdynamic circulation of liver disease 40 years later:pathophysiology and clinical consequences. HEPATOLOGY 1994;20:1359-1363.

20. Sato S, Ohnishi K, Sugita S, Okuda K. Splenic artery and superiormesenteric artery blood flow: nonsurgical Doppler US measurement inhealthy subjects and patients with chronic liver disease. Radiology1987;164:347-352.

21. Sabba G, Ferraioli G, Genecin P, Colombato LA, Buonamico P, Lerner E,Taylor KJW, et al. Evaluation of postprandial hyperemia in superiormesenteric artery and portal vein in healthy and cirrhotic humans in anoperator-blind echo-Doppler study. HEPATOLOGY 1991;13:714-718.

22. Epstein M, Schneider N, Befeler B. Relationship of systemic andintrarenal hemodynamics in cirrhosis. J Clin Lab Med 1977;89:1175-1187.

23. Fernandez-Seara J, Prieto J, Quiroga J, Zozaya JM, Cobos MA, Rodriguez-Eire JL, Garcia-Plaza A, et al. Systemic and regional hemodynamics inpatients with liver cirrhosis and ascites with and without functionalrenal failure. Gastroenterology 1989;97:1304-1312.

24. Fernandez-Munoz D, Caramelo C, Santos JC, Blanchart A, Hernando L,Lopez-Novoa JM. Systemic and splanchnic hemodynamic disturbancesin conscious rats with experimental liver cirrhosis without ascites. Am JPhysiol 1985;249:G316-G320.

25. Groszmann RJ, Vorobioff J, Riley E. A combined method for theevaluation of the splanchnic circulation in rats. In: Lautt WW, ed.Hepatic Circulation in Health and Disease. New York: Raven, 1981:15-23.

26. Bhathal PS, Grossman HJ. Active contractile and mechanical factors inthe pathogenesis of portal hypertension—a study utilizing three experi-mental models. In: Sarin SK, Nayac NC, eds. Animal Models of PortalHypertension. New Delhi: Kuni Publishing, 1988;111-122.

27. Rockey DC, Weisiger RA. Endothelin induced contractility of stellatecells from normal and cirrhotic rat liver: implications for regulation ofportal pressure and resistance. HEPATOLOGY 1996;24:233-240.

28. Graves J, Poston L. b-Adrenoceptor agonist mediated relaxation ofisolated resistance arteries: a role for the endothelium and nitric oxide.Br J Pharmacol 1993;108:631-637.

29. Mittal M, Gupta T, Lee F, Seiber C, Groszmann R. Nitric oxide modulateshepatic vascular tone in normal rat liver. Am J Physiol 1994;267:G416-G422.

30. Floras JS, Legault L, Morali GA, Hara K, Blendis LM. Increasedsympathetic outflow in cirrhosis and ascites: direct evidence fromintraneural recordings. Ann Intern Med 1991;114:373-380.

31. Abuelo GJ. Diagnosing vascular causes of renal failure. Ann Intern Med1995;123:601-614.

32. Epstein M. Hepatorenal syndrome: emerging perspectives of pathophysi-ology and therapy. J Am Soc Nephrol 1994;4:1735-1753.

33. Zironi G, Gaiani S, Fenyves D, Rigamonti A, Bolondi L, Barbara L. Valueof measurement of mean portal flow velocity by Doppler flowmetry inthe diagnosis of portal hypertension. J Hepatol 1992;16:298-303.

34. Sacerdoti D, Bolognesi M, Bombonato G, Gatta A. Paraumbilical veinpatency in cirrhosis: effects on hepatic hemodynamics evaluated byDoppler sonography. HEPATOLOGY 1995;22:1689-1694.

35. Norris CS, Barnes RW. Renal artery flow velocity analysis: a sensitivemeasure of experimental and clinical renovascular resistance. J Surg Res1984;36:230-236.

36. Schroeder ET, Shear L, Sanceta SM, Gabuzda GJ. Renal failure in patients

HEPATOLOGY Vol. 28, No. 5, 1998 RIVOLTA ET AL. 1239

with cirrhosis of the liver. III. Evaluation of intrarenal blood flow byparaamino-hippurate extraction and response to angiotensin. Am J Med1967;43:887-896.

37. Epstein M, Berk DP, Hollenberg NK, Adams DF, Chalmers TC, AbramsHC, Merril JP. Renal failure in patients with cirrhosis: the role of activevasoconstriction. Am J Med 1970;49:175-185.

38. Platt JF, Rubin JM, Ellis JH. Acute renal failure: possible role of duplexDoppler US in distinction between acute prerenal failure and acutetubular necrosis. Radiology 1991;179:419-423.

39. Quarto di Palo F, Rivolta R, Elli A, Castagnone D. Relevance of resistiveindex ultrasonographic measurement in renal transplantation. Nephron1996;73:195-200.

40. Papadakis MA, Arieff AJ. Unreliability of clinical evaluation of renalfunction in cirrhosis: a prospective study. Am J Med 1987;82:945-952.

41. Caregaro L, Menon F, Angeli P, Amodio P, Merkel C, Bortoluzzi A,Alberino F, et al. Limitations of serum creatinine level and creatinineclearance as filtration markers of cirrhosis. Arch Intern Med 1994;154:201-205.

1240 RIVOLTA ET AL. HEPATOLOGY November 1998