Systemic and regional hemodynamic effects of isosorbide dinitrate in patients with liver cirrhosis and portal hypertension

316 HEP ooj43

Systemic and regional hemdynamic effects of isoscrkide dinitrate in patients with liver cirrhosis and portal hypertension

Pierre Mols’, Roger Hallemans’, Christian Melot’, Philippe Lejeune’ and Robert Naeije* ‘Medical lnrolrivc Carr Onir Labormory, Saint-Piene UniversiryHosprral. md ‘Deparmzenroflnremive Care, ErasmusHospiral, Bruss& (Belgiwn) (Received24 May 1988) (Accqxed 28 November1988)

In a group of 17 cirrhotic patients with portal hypertensim, administration of isosorbide dinitrate (IDN)

on central

we have investigated the effects of 5 mg sublingual

hemodynamics,on regional

namics and on blood gases. Fifteor~ mi” after drug administration, pressure from 4 + 1 to 3 f I mmHg (mean k S.E.M.,

(hepatic and renal) hemody-

we observed a decrease in the right atria1 mean

PC 0.02) and of pulrraonary arterial wedge pressure fmm 7

f 1 to 4 + 1 mmHg (P C 0.001) with decreases of the cardiac inr!ex from 4.2 f 0.2 to 3.7 5~ 0.2 Umin/m* (P 4 0.001) and the mean arterial pressure from 89 f 4 to 72 f 3 mmHg (P < 0.001) and a” increase in heart rate from 86 f 4 to 94 + 5 beats/min (P < 0.001). Arterial

Pq decreased from 73 f 2 to 66 + 2 mmHg

(P C 0.001). As a

consequence of bo.> cardiac index and arterial PO* reductions, 02 transport to the tissues was reduced from 602 + 32 to 518 f 26 mllmin.mz (P < 0.001). The hepatic venous pressure gradient decreased from 17 k 1 to 14 t 1 mmHg (P < 0.001) and hepatic vein PO2 did not change. The hepatic blood flow (HBF) remaked

unchanged. Renal blood flow (RBF)

determined in 7 patients

determined in 5 patients decreased from 0.76 ‘_ 0.11 to 0.68 t

0.11 Vmin (P < 0.001). In conclusion, isosorbide dinitrate reduces portal hypertension in patients with liver cirrhosis withom compromising hepatic perfusion. This effect, however., is associated with a decrease in 02 delivery and a slight reduction in renal perfusion.

I”lmd”Cti0”

ment in patients with

liver cirrhosis and portal hype*_

tension 131. It has atso been shown to decrease the Isosorbide dinitrate, a predominanily venous vasodilator,

has been shown to reduce portal pressere

both acutely [l-4]

and after 1 month of oral tmat-

pIeWIre

Of esophageal varices catheterized by a fme

needle under endoecopic control [S!. uld to reduce the portal venous flow assessedby an echo-Doppler

Correspondence:Dr. Pierre Mols, Garde Adultes. Hfipital Univenitaire Saint-Pierre,322 rue HPU~C,B-1W Brussels,Belgium 0168.82781W$O3.50~1989EtsevicrSciencePublishersB.V. (BiomedicalDivision)

REGIONAL HEMODYNAMICEFFECTSOF IDN IN CIRRHOSIS technique [6]. The mechanism of the nitrate-induced decrease in portal pressure is not well known. The administration of citrates to nonnovolemic patients is generally associated with a decrease in cardiac output and .hus a possible reduction in Oz delivery to the tissues inc!uding the leer. We therefore investigated central and regional hemodynamics concomitantly in the present study.

Patients and Methods P4deni.v Seventeen patients with histologically proven liver cirrhosis, 11 men and 6 women, aged 36-65 years (mean 53) were referred to our laboratory for hepatic vein catheterization to alleviate the severity of portal hyperteasion and hemodynamic alterations. Each patient was informed of the proceedings and agreed to participate in the study which was approved by the ethical committee of :he hospital. All the patients were chronic alcoholics (defined as having a daily consumption of at least 180 g of alcohol for several years). None of them had clinical evidence of intrinsic cardiac or pulmonary disease. Chest X-rays and electrocardiograms were unremarkable. Live: scans invariably showed decreased and irregular radiocolloid uptake with increased uptake by the spleen and the bone marrow. Esophageal varices were demonstrated by esophagoscopy and barium X-ray studies in all but 2 patients. On examination, 2 patients had hepatomegaly, 16 of them splenomegaly, 14 ascites, 8 wider naevi, 11 valmar ewthema and 11 ictems. Onk patient showeri slight cot&ion and 1 had a mild asterixis. Serum creatinine was 1.2 f 0.3 ma/d1 (mean + S.E.M.); blood urea, 31 + 9 mg/dl; serum albumin, 34 f 1 @I; serum bilirubin, 3.2 zk0.6 mg/dl; and serum calcium, 8.6 f 0.1 mg/dl. According to Child’s classification. there were 5 grade A patients, 11 grade B patients and 1 grade C patient. Hemodynamic Twc

and bloodgas

dekwninotiom

triple lumen thermodilution 7F Swan-Ganz catheters (Edwards Laboratories, Santa-Anna, CA)

317

were introduced into the right internal jugular vein using a Seldinger technique, and under constant pressure ~jve moniiorin8 and fluomscopic control, one was advanced into a pulmonary awry dnd cue into an hepatic vein. Correci hepa+ic vein occlusion by inflation of the bal!oon was confimted by careful injection of a sntall amount of contrast medium, to demonstrate dye retention in the occluded portion of the hepatic vein (71. A small polyethylene catheter was inserted into a radial artery for systemic presswe measurenwtts and arterial blood sampling. Press-ores were lneasured using Statham 23 P strain gauged transducers (Statham Instruments, Hato Rey, PR) and recorded on a thermal writingrecorder (Visicorder, 1858, Honeywell). The zero reference Ieve; was placed at the midchest level and pressure values were averaged for three successive respiratory cycles. Heart rate was determined from a continuowly monitored ECG lead. Cardiac output was measured in tiplicate by the tbermodilution method using an automatic pressure injector and a computer (9510-A, Edwards Laboratories, Santa-Anna, CA). Arterial, hepatic and mixed venous blood gases were measured using a Coming 175 pH/bIood gas system (Coming Medical Products, Medfield, MA). Hemoglobin concentrations and oxygen saturations were determined by a OSM2 bemoximeter (Radiometer, Copenhagen, Denmark). Standard parameters and formulae were used to calculate cardiorespiratory and hcmcdynamicvariables IS]. Hepadc hemodynamics

Hepatic blood flow (HBP) was measured by a constant infusion method with indocyanine green (KG, Hynson, Westcon and Dunning Inc., Bal:imore, MD) 191. An intravenous bolas of 15 mg KG was followed by a constant infusion of 0.33 mg KG per min and after 1 h equilibration, measurements of hepatic clearance were started. ICG absorbance was measured by spectrophotometry at 800~~1 and the absorbance of the plasma take before the infusiofi of ICG was used to correct for background turbidity. The ICG concentration was then calculated from standard curves. If the ratio of KG extraction by the liver was more

318

P. MOLS et al.

calculated by dividing the radioactivity of 1 mi” infw sio” of hippura” by the radioactivity of 1 ml of plasma, and the renal blood flow was c,kulated by dividing the hippura” clearance by (I-hematocrit). No urinary sample was required. Normal values from our laboratory from 20 healthy adult volunteers were in thcra”ge0.52-1.00 Umin.

than lo%, the clearance of KG by the liver and hep atic blood flow were calculated as follows: ICG clearance

= ,
(I.min-t)

where R = KG amount infusate in 1 mi” (mgmi”“); .4 = arterial KG concentration (mgl-‘); H = hepatic vein KG concentration (mgl-‘); hepatic blood flow = ICG clcaiance/(l-h,~matocrit) (HBF, Lmin-I). Nom~al values of HBF estimated by this method range from 815 to 2252 mUmi” [9,10]. Asscssmeni of hepatic hemodynamics included the following determinations: hepatic vein 0, tension - Torr, free hepatic vein pressure (FHP) - mmHg. occluded hcpatic vein pressuw (VviiP) - mmHg, hepatic venous prossure gradient = WXP-FHP (HVPG) - mmHg.

Right heart and hepatic vein catheterization were petfomted in the supine position after a” overnight fast and with no premcdication. No drug had been administered for at least 48 h before the test. The patients were allowed to rest after insertion of the catheters during the equilibration time of ICG and hippuran. Baseline mcasu~ements were pcrfonncd when the respiratory rate, heart rate and vascular prcswes were stable. lsosorbide dinitrate, 5 mg, was then administered subliigually. Blood gases and hemodynamic measu~cments were performed again 5 and 15 min after the tablet had been dissolved.

Renal bloodflow

Renal Mood flow (RBF, was measured by a constant infusion method with o-[‘3!I]iodohipp”rate (Hippuran, Institute of Radio-Elen:ents, Fleurus. Belgium) [ll], A” intravenous bolus of 12 &i hippure” was followed by a constant infusion of 0.2&i per min. After 1 h equilibration, measurements of the renal clearance were started. The clearance was

VALUES OF BLOOD GASES DETERMINATION SORBIDEDlNITRATESUBLtNGtJALLY

The statistics consisted of an analysis of variance for repeated meas”rcmc”ts. When the Fratios were greater than the tabulated P = 0.05 critical value, modified f-tests were performed with the Bonfcrroni

BEFORE .AND 5 AND 15 min AFTER THE DISSOLUTION

OF 5 mg ISO-

Measurements madeduring the drug regimen were comparedto baaeliue.Valuesare meansf S.E.M. _~~___ Variables

Baseline

IDN 5 niin

Hemoglobin (gll0l ml) Arterial pli Arterial P* (Ton) Arterial Oz saturation (%) Arterial PC4 (Torr) Alveolar-arterial PO gradient (Tow\ Venousadmixture (d) Mixed “enous Po2 (Torr) Mixed YF~OUS0, saturation (%) Anerio-venousO~~ntcntdiiie~nce (mUdI) 0, consumprion (mUminmz) Oz trauspat (mt/minm2) Oz extraction (W)

10.6*0.4

-

7.45+ 0.0,

7.46f 0.01

73f2 95 + 1 342 1 35 * 2 14+ 1 37 * 1 71 k 1 3.6 f 0.1 149k6 602 2 32 25*1

65f2 93f 1 34* 1 44** 18 f 2 34* 1 64fl 4.2 +_0.2 14826 4%*26 3t+ 1

IDN = isosorbide dinitrate.

P

IDN L5minp

P

N.S. < 0.001 < 0.001 N.S. < o.wt < 0.01 < 0.001 < O.Wl
7.46 f Il.01 66+2 93+* 34 + 1 42 * 2 17+2 34+1 65k.1 4.1 * 0.2 150+5 518 + 26 M+t

N.S.
__~

REGIONAL

HEMODYNAMIC

319

EFFECTS OF IDN IN CIRRHOSIS

B

to Srni”

ID 15”i”

!” 5mm

ID

B

Pamin

1D Srni”

10 15min

Kg. 1. Individual effects of 5 mg sublingual isosorbidc dinitrate (ID) on mean arterial pressure (MAP), cardiac index (CI) aad arterial PO2(PaO,) in 17 cinhoti~ patients. (B = bsseline.)

adjustment for multiple comparisons !12]. Linear correlations were also calculated.

RestlIt At basal state, blood gases showed a mild arterial hypoxemia, a respiratory akalosis and an increase of alveolar to arterial Pm gradient and of venous adTABLE

mixture (Table 1). Pulmonary and systemic vascular pressures were in the normal range and the cardiac index was at the upper limit of normal (Table 2). The bepatic venous pressure gradient ranged from 9 to 26 mmHg with a mean value of 17 mmHg. Hepatic blood flow could be determined only in 7 patients in whom the ICG extraction ratio exceeded 10% (16-571, mea” 34%) and ranged from 0.35 to 1.09 Umin (Table 3). Renal blood flow measured in the last 5

2

VALUES OF HEMGDYNAMK DETERMINATIONS ISOSORBIDE DINITRATE SUBLMGUALLY

BEFORE AND 5 AND 15 mi” AFTER THE DISSOLUTION

OF 5 mg

Measurements made during the drug regimen were compared to baseline. Valuer are means + S.E.M. Variables

Baseline

IDN 5 min

P

IDN 15 min

P

86*4 98 + 4 CO.M)l 94*s c 0.001 HR (beattimin) 3.6 f 0.2 < O.Wl 3.7 f 0.2 < O.Ml 4.2+n.z CI (l/mitvmz) < 0.w 37 * 2 4052 < O.W1 WI (mtiatsJm2) 50f2 74 + 3 < 0.001 < O.Wl g9*2 72f3 MAP (mmHg) 13f I 9+1
320

P. MOLS et al.

“%x-T-

40oJ

. B

Sam 15min

5mm 15mh

SIni” t5mm

. I. ID 10 Smin ISmin

Fig. 2. lndividuat effects of 5 mg sublingual isosorbide dinitrate (ID) on hepati venous pressure gradient (HVPG. n = 17). hepatic blood flow (HBF, I = 7). hepatic vein I’% (P hv 0,. n = 17) and renal bled flow (RBF, n = 5) in ourcirrhoticpatients. (B = baseline).

patients ranged from 0.48 to 1.G9Ilmin. lsosorbide dinitrate induced important effects after 5 min. These effects persisted 15 min later. Mean arterial pressure, mean pulmonary arterial pressure and pulmonary arterial wedge pressure decreased as did the cardiac index (Table 2, Fig. 1). Pulmonary and systemic vascular resistance indices remained unchanged and heart rate increased (Table 2). Arterial PO2fell with a concomitant increase in alveolar to arterial P,-,*.O1 transport and mixed venous Pm decreased (Table 1).

After isosorbide dinitrate administration, the hepatic venous pressure gradient decreased as a consequence of a fall in the occluded hepatic vein pressure, while the free hepatic vein pressure remained unchanged (Table 3). Hepatic vein Pm (n = 17) and hepatic blood flow (n = 7) did not change (Table 3, Fig. 2). We did not find any linear correlation between the modification of hepatic vein Pm and the variation of arterial PO2(r = +O.lO, N.S.) or of the cardiac index (I = +0.18. N.S.). The reduction cf the hepatic ve-

TABLE 3 VALUESOF HEPATIC VEIN BLOOD GASES AND HEPATIC AND RENAL HEMODYNAMK

DETERM,kAT,ONS FORE AND 5 AND 15 min AFIXR THE DISSOLUTION OF5 mg ISOSORBIDE DINITRAXE SUBLlNGUALLY

BE.

Mcarurements made during the drug regimen were compared to baseline. Values are means ? S.E.M. Variables

”

Baseline

IDN S min

P

IDN 15 min

P

FHP(mmHS) WHP(mmHS) HVPG (mmHg) Pw2 (Tot0 Sk”* W) IG extr. ratio (%) IG clear (I/min) HBF(limin) RBF (t!min)

17 17 17 I7 17 7 7 7 5

II f 1 282 1 L8+ L 34 k 6 58k4 32 ?; 5 0.41 k 0.07 0.65 i 0.11 0.76 *o.,t

to+, 24+1 t4+ 1 32k5 54*4 29+9 0.41 f 0.09 0.74 * 0.14 0.70 * 0.11

N.S. 4 O.ool F o.w, N.S. N.S. N.S. N.S. N.S. c 0.01

IOk1 24 * 1 14* 1 32 + 5 54 * 4 30 k 4 0.44? 0.08 0.71 k 0.13 063 + 0.11

N.S. < O.WI 4 o.tmt N.S. N.S. N.S. N.S. N.S. c 0.001

k% = free hepatic vein pressure; WHP = occluded hepatic vein pressure; HVPG = hepatic vein pressure grsdienr; Phvo- = bepatic vein Pal: S,,, = hepatic vein O2 saturation: IG extr. ratio = extraction ratio afindocyanine green: IO dear = clearance of indocyanine green: H B F = hcpatic blood flow: RBF = renal blood flow.

REGIONAL

now

HEMODYNAMIC

pressure

decrease

gradient

was directly

h the cardiac

correlated

ir._‘ex (r = +0.54.

to the

P c 0.001).

of isosorbide

dinitrate

dent on Child’s classifica:ion

were nor depen-

(Table

4).

Renal blood flow measured in the last 5 patients decreased in each one after isosorbide dinitrate administration (Table 3, Fig. 2). Side effects were observed in two patients isosorbide

dinitrate

induced

hypaension,

Hypotension, our

but not to the decrease in the mean arterial pressure (I = +0.43, N.S.). The effects

321

EFFECTS OF IDN IN CIRRHOSiS

bmdycychrdia

I7 patients,

have

and nausea,

also been

bradycar-

dia and nausea.

in a pre-

symptoms are px&ab:y the ronsequ~nce of an increased vagal tone which could be either secondary to hypovolemia or be due to a diwet effect of the drug [i5,16].

vious

work

[l].

These

Arrerial hypoxemia In contrast with the results

in whom

wen in 2 of

shwved

of a previous

study

[l],

literature, nitrates are shown to induce hypoxemia in normal [9 and diseased men [1@,19].The increase in the uiveolar-arterial P-gradient and the venous admixture in our pain ou: present

tients

suggests

work and in the

a deterioration

in pulmonary

gas ex-

In our patients, sublitlgual IDN administration produced a decrease in both vascular pressures and

change. As nitroglycerin.and probably IDN aIw, da not inhibitthe hypoxic pulmonaryvawconstriction in men 1171.as no drug-induced pulmonaryvasodilation was observed in OUTstudy, a reduction in mixed venous PQ secondary to a decreased cardiac output

cardiac

would

Sysfemic hemodynamics and side effects

index,

and an increase

modifications peripheral

are

venodilation

acceleration

of the heart

the hemcdynamic

in heart

consequences

rate.

These

of a drug-induced

and of a sympathetic in some

trates

is directly

the pono-systemic

pmportional shunt.

of

of our patients

might be due to the fact that the bioavailability

of ni-

to the importance

Patients

account

for the fall in arterial

Pa in

reflex

rate 1131. The amplitude

response

probably

our patients.

of

with an important

shl;nt might thus show a more significant hemodynamicresponse to IDN [14].

pono-systemic

Liverhemodynomics Until now, inman,it

hasbeendifficult

toassessthe

and hepatic circulationaccurately.Portal pressure is usually estimated by the hepatic venous pressure gradient, whereas liver blood flow and sometimes the azygos vein flow are measured directly. Although the evaluationof liver blood flow by the conportal

TABLE 4 VALUES (MEANS f S.E.M.) OF HEMODYNAMK AFTER THE DISSOLUTION OF 5 mS ISOSORBIDB PATIENTS

AND GASOMETRIC DETERMINATIONS BEFORE AND 15 min DINITRATE SUBLINGUALLY IN 5 CHILD’S A AND II CHILD’S B

_____ “adables

CI (vminJ) MAP(mmHg) FHP(mmHg) WHP (mmHg) HVPG (mmHS) %o (Tom) HBd(Umin) RBF(Vmin)

ChildS A (n = 5)

Child’s B (n = II)

baseline

IDN I5 min

Dasellne

IDN 15 min

4.5 5 0.4 85 * 2 to*2 25 + 4 15c3 33 * 5 0.70 * 0.17 0.92

4.0 +o.z* 61 I7” 10f1 22f2’ 1**2* 3223 0.70 f 0.21 0.86N.D.

4.2 i: 0.3 91k.3 II +t 29fl t8+* 36f2 0.65 * 0.18 0.72 2 0.14

3.7 * 0.3, 73 * 3” II+1 2S+t* 14+1* 344+2 0.70 + 0.17 0.63 f 0.14N.D.

Ct = cardiac index; MAP = mean arterial pressure; &AM = mean right atrial pressure; FHP = free nepatic win pressure; WHP = oe_ eluded hepatic vein Pressure; HVPG = hcpatic vein pressure gradient; P h*01= hepatic
322 stem infusion method with ICG has been used for a long time, this is now questioned. indeed, the extrection ratio of ICG is variable according to the patient, is time dependent and seems ill-correlated to the liver blood in cirrhotic patients [20,21]. In spite of there methodological limitations, our work confirms that nitrates induce a fall in portal pressures [l-4,22] and in cardiac output [1,2,4,22] and that it fails to modify the hepatic blood flow [2] in cirrho!ic patients with portal hypertension. Furthermore, our study shows that the hepatic oxygenation evaluated by the hepatic venous blood Pm dazs not deteriorate after administration sf IDN despite a fall in 02 transport. Mechanisms of decreased portal pressure after IDN are complex. In one of our studies, the hepatic venous pressure gradient was strongly correlated with the cardiac index suggesting that IDN decreases portal pressure secondary to a reduction of portal venousfiow [l]. In our present study, however, this correlation being 0.54, the fall in cardiac index cannot on its own explain the fall in portal hypertension. A splanchnic arterial vasoconstriction induced by B baroreceptor reflex after IDN administration also seems unlikely in OUTwork as we did not find a significant relation between the modification of hepatic venous pressure gradient and of mean arterial pressure. However, IDN-induced mechanisms lowering portal pressure have been more thoroughly investigated in animal models. At doses causing a slight drop in blood pressure without affecting the cardiac index, IDN and nitroglycerin cause a decrease in portal venous inflow &her with or without modification in portal pressure [23,24]. The portal venous flow reduction is caused by a mesenteric artery vasoconstriction secondary to a peripheral veuodilation [25]. At doses causing a drop in blood pressure and in cardiac output, the adaptation mechanisms fail and the reverse effect is observed: the portal venous inflow remains stable and portal pressure decreases. This suggests a vasodiiation of the mesenteric arteries and of the collateral syst:m [23]. The increased portal vascular resistance of an iso-

P. MOLS et al. lated cbnhotic rat liver perfused at a constant rate is reduced by vasodilators including sodium nitroprusside, magnesium sulfate or prostagiandin E, 1261. Since large numbers of myofibrablasts XI chren td in the cirrhotic liver [,!7.id] a .- ~:luct:o. IF gx:aZ pressure by vasodilators could be explained, in part, by the lowering of the intra-hepatic portal vascular resistance. In men with cirrhosis, IDN has been shown to reduce portal venous tlow by an echo Doppler technique [6]. The stability of the liver blood flow when portal venous flow decreases, could be explained by an increase in the arterial hepatic flow [29]. Hydrodynamic interactions between the hepatic arterial and portal venous vascular bed have been known for years [29]. A decrease in blood flow through one circuit leads to a decreased inflow resistance in the other circuit, tending to lead to a constant blood flow through the liver. Thii effect has been telmed ‘reciprocity’ between the hepatic artery and the portal vein. It is probably also relative to the increase of hepatic arterial blood flow that liver oxy genation, as assessed by hepatic vein Pm, remained canstent despite a fail of the arterial Pq in our patients. Aithougb the mean value of hepatic venous Pa aft-r administration of IDN remained constant in our study. b~dividual variations were important. As we did not observe any correlation between hepatic venoas ?, variations and either cardiac index or arterial Pm changes after IDN, a fail in hepatic venous Pm was not predictable from cardiac output or arterial blood gas monitoring. Liver oxygenation may also have remained constant in relation to Breduced spianchnic 0, consump tion after IDN. Although we did not measure splanchnit O2 consumption, this hypothesis seems unlikely because the total Oz consumption remained constant after IDN administration. Rena: hemodynandcs Renal blood flow was measured in our last 5 patients. Fifteen min after the dissolution of IDN, we observed a slight, but significant, fall in the renal blood flow. This was concomitant with the drop in

REGIONAL

HEMODYNAMIC

EFFECTS OF IDN IN CIRRHOSIS

mean arterial pressure which suggests an absence of

323

In concldon

variations in renal vascular resistance. In animals, immedia+Ay t~o&cerin, fin,

after sublingual administration renal blood flow kxxsej

It IF;C..TP 10 :ht

modifications

base!,.

of ni-

and, after 15

‘. JL!~-. i24.?0!,

.j.

of renal blood flow follow the varia-

tions of the cardiac index [24]. Therefore,

in our pa-

We confi?n

that the acufe sublingual

adminiatra-

*ion of isosorbide dinitrate is an effective treatment ol 7 hypertension which does not seem to alter the blood tlow through the liver, but which can possibly cause a reduced oxygen delivery to other organs,

tients the decrease in renal blood flow appeared to be the consequence of the decrease of the cardiac index.

1 Hallemans R, Naeije R, Mols P, et al. Treatmen, of panal hypenension wi,h isoserbide dinioate alone and in combiw&ion with vasopressin. Cri, Care Med 1983,11: 536-540. 2 Merkel C. Finucci G, Bazzerla M, e, al. Effeas of isasorbide dinhrate on portal h~rtension in alcoholic cirrhosis. 1 Hepatol19d~4: 181-189. 3 Freeman JG. Banon JR. Record CO. Effecfec( of isosorbide didrate, verapamil and labelalol on patal presswe in citrhosis. Br Med J 1985.291: X-562. 4 Blei AT, Garcia-Tsao G, Groszmann RI, Kabrilas P, Ganger D. Morre S, Fung, HL. Hemodynamic evaluation of ,so. sorbide dinitrate in alcoholic cirrhosis. Pharmacokinetio bemodynamic interactions. Gastraenterology 1987: 93: 576-F83. 5 Staritz M. Poralla, T. Meyer mm Biischenfelde KH. Intravasadar oesopbageal vaiices pressure (IOVP) assessed by cndoscoc4c fme needle puncture under basal conditions, Valsatva’s man~uvx and afrer glyceryltrinhrate application. Gw 1985; 26: 52.5-530. 6 Zoli M. Marchesini G. Rnmori A. et al. Ponal venous flow in response to acute@-blocker and vasodilatabxy treatment in patients with liver cirrhosis. Hepamlegy 1986; 6: 1248-1251. 7 Grosnnann RJ. Glickman M, Blei AT, et al. Wedged and free bepatic vene~s pressure measured with a balloon catheter. Gastroemerology ,979: 76: 253-259. 8 Mols P. Naeije R, Hallemans R, et al. Cer.tral and regional hemodynamic cflects of nhrendipiw in noonotensive patients with chronic obrlmctive tang &sease. I Cardiovasc. Phannaeol1986; 8: 77-w. 9 Caesar 3, Sheldon S, Chiandurai L, ei at. The use af mdocyanine green in the measurement of hepatic blood flow aod as a test of hepatie function. Ctin Sci 1961: 21: 43-57. 10 Lebrec D, Sicot C, Benbamou JP. Debit sanguin bepadoue. hvwnension wrtale et insuffuance cellulaire chez des &lad;; atteints d; cinhose alcoalique. Arch Fr Mel App Dig 1973; 62: 465-471. DP. Compariw of the cons,am infusion and urine collection ,eechniques for the measurememafrenal function. 1 Ctin Invest 1948; n: 710-716. 12 Wallenstein S, Zwcker CL, Fteiss JL. Some ladsdeal meshads useful in circulation research. Circ Rev 1980; 47: l-9. 13 Abram J. Current concepts. Nitroglycerin and long-acting

11Berger EY,Farber S,.Park

m,ra:es. N Engl I Med ,980: 302: 1234-1237. 14 Blri AT. O’Reilly Dl, Gasrein 1. Pormt-systemic rhuntiting ano ?.e hemodynanic effects of nitmgtyczrin in the ral. Gastroentemlogy 1984; 86: 1428-1436. 15 Khan AK, Carleton RA. Nitmglycerikinduccd hypotension and bradycardia. Arch Intern Med 1981; 141: 984. 16 Nemorowki M. Shah P. Syndrome of severe bradycardidia and hypotensionfollowing sublingualnhqlyecdn administration.Cardiology1961; 67~ 180-189. 17 Nneije R, Melo, C. Mols P, et al. Effects of vasedilatorson hyporic pulmonary va~lcanstriaioD in normal man. Chest 1982 82: 404-510. I8 Hales CA, Westphal D. Hypoxemia following the adminis~ Iration of sublingual nitroglycerin. Am I Med t97g; 65: 911-918. 19 Melo, C. Hallemans R. Naeije R. et al. Deleterious effea of nifedipinc on pulmonary gas exchange in chronic putmanary disease. Am Rev Respir Dis 1984; 130: 612-616. 2~3Clemems D, West R, ELias E. Comparison of b&w and infusion methods for estimating hepatic blood ilow c; p&ems wi,b liver disease usfng indocyanine green. .’ Hepatol 1987, 5: 282-287. 21 Buqnski FJ, Greenway CV, Sitar DS. Hepatic blood now: accuracy of ertimatica from infusion of indcqanins green in anaenhetized cats. Br 3 Pharmaml 1987; 91: 82X-659 22 Garcia-Tsao G, Gronmann RI. Partal hemodyeamier dw ing nirmglycerin administration in cirrhotic patients. I&D&logy 3987; 5: 8n5-809. 23 Blei AT, GotrsreinJ. Isosorbidedinitrate in experinrrntal portal hyf,enenrion: a study of fac,o,s thnt modvtate the hemcdy~~amic respanse. Hepa,&gy 1986; 6: tO7-Ill. 24 Vatner SF, Pagani M, Rwberford JD. et at. Effens of nitroglycerin on cardiac funcdon and re+mt blaad flow dtstribufion in conscteus dogs. Am J Pbysiol 1978: 234: H244-H252. 25 Kroeger RI, Gmszmann RI. The eKect of the mmbir&.x, of nitmglycenn and prqxanolol on splrnchnie and systemic hemodynamics in a portal hypenensive rat umdel. Hepatology 196 5: 425-430. 26 Bhathai PS, Grossmann HJ. Reduction af the increased penal vasadar resistance of the isolated perfused cirrhotic rat live, byvasodilmon. I Hepaw 1985; 1: 325-337. 27 Nakano M, Werner TM, Lieber CS. Periveaular fibrosis in alcoholic liver injury: ultms~tiure and histologic pmgres-

__.__.

324

sion. Gastroemerology 1982; 83: 717-785. 28 Bbatha, PS. Presence “f modihed flbroblaata in cirrbollc livers in man. Pathology 1972; 4: 139-144. 29 Richardson PD. Wbhrington PG. Leer blood flow. 1. Intrimic and nervous control of liver blood flow. Gastrocntcr-

P. MOLS

et a,.

OlOgy 1981; 81: 159-173. 30 Dumont L, Lamoureux C, Leforier I, ei al. fnrravenaus infusion of nitroglycerin: effects upon cardiovascular dynamks and regional blood fBw distribution in conscious dogs. CanJ Physiol Pbamlacol1982;60: 3436-1443.