Long-term renal sodium handling in patients with cirrhosis treated with transjugular intrahepatic portosystemic shunts for refractory ascites

$Page 1: Long-term renal sodium handling in patients with cirrhosis treated with transjugular intrahepatic portosystemic shunts for refractory ascites$
Long-term Renal Sodium Handling in Patientswith Cirrhosis Treated with Transjugular

Intrahepatic Portosystemic Shunts for RefractoryAscites

Winnie Wong, MD, Peter Liu, MD, Laurence Blendis, MD, Florence Wong, MD

PURPOSE: The long-term effects of transjugular intrahepaticportosystemic shunts on renal sodium excretion are not known.We sought to determine these long-term effects, as well as tomeasure the effects of a sodium load in patients who are free ofascites.SUBJECTS AND METHODS: Ten patients with cirrhosis whohad been successfully treated with transjugular intrahepaticportosystemic stent shunt for refractory ascites were studiedbefore the shunt and again at 6 and 14 months after the shuntwhile on a 22 mmol sodium/day diet. At 14 months they werealso studied on a 200 mmol sodium/day diet for 7 days withoutdiuretics. Renal sodium handling, central blood volume, neu-rohumoral factors, and hepatic function were measured.RESULTS: Sodium balance was negative at 6 months (urinarysodium excretion [mean 6 SD] 51 6 11 mmol/day versus 7 6 2mmol/day pre-shunt; P ,0.05), was maintained at 14 months(22 6 4 mmol/day; P ,0.05 versus pre-shunt), and was asso-

ciated with normalization of renin activity and aldosterone lev-els, but not norepinephrine levels, as well as significantly im-proved renal hemodynamic measurements. Sodium loadingwith 200 mmol/day resulted in weight gain associated with in-creased central blood volume and appropriate renal sodiumhandling in most but not all patients (urinary sodium excretion188 6 14 mmol/day), despite persistent nonsuppressibility ofsympathetic hyperactivity.CONCLUSIONS: In cirrhotic patients with refractory ascitestreated with a transjugular intrahepatic portosystemic stentshunt, long-term renal sodium handling is improved. Adequateintravascular filling in ascites-free cirrhotic patients with nor-mal portal pressure permits an improved but not normalizedrenal response to a sodium load, possibly due to persistentlyelevated sympathetic activity. Therefore, these patients shouldincrease their sodium intake cautiously. Am J Med. 1999;106:315–322. q1999 by Excerpta Medica, Inc.

The pathogenesis of renal sodium retention in pa-tients with cirrhosis has been explained by the “pe-ripheral arterial vasodilatation hypothesis,” which

proposes that systemic arterial vasodilatation leads to de-creased effective arterial blood volume. This in turncauses compensatory increases in the activities of thesympathetic nervous system and the renin-angiotensin-aldosterone system, thereby promoting renal sodium andwater retention (1). In ascitic patients who have beentreated with a peritoneovenous shunt, the reconstitutionof the intravascular space with ascitic fluid is associatedwith suppression of the renin-angiotensin-aldosteronesystem and of the sympathetic nervous system, leading tonatriuresis (2,3). However, when patients are ascites-free

and challenged with a high sodium intake, these patientshave sodium retention despite continued suppression ofthe renin-angiotensin-aldosterone system (4). This sug-gests that other mechanisms, in addition to arterial un-derfilling, are involved in the pathogenesis of sodium re-tention in cirrhosis.

The role of sinusoidal portal hypertension in causingsodium retention is suggested by the effectiveness oftransjugular intrahepatic portosystemic shunts in reliev-ing ascites (5–7) as well as by studies that have shown thatascites does not occur if the sinusoidal portal pressure is,8 mm Hg (8,9). Transjugular intrahepatic portosys-temic shunt insertion is followed by systemic vasodilata-tion (10). In the absence of diuretics, natriuresis is de-layed for 2 weeks after the procedure, at which time adiuresis begins despite an initial deterioration in hepaticfunction (11) and continues until most patients eventu-ally become free of ascites. However, the long-term ben-efits of the transjugular intrahepatic portosystemic shuntare unproven (12), and it is unclear whether renal sodiumhandling returns to normal when patients become as-cites-free.

The aims of this study were to assess long-term renalsodium handling in cirrhotic patients following success-ful treatment of refractory ascites with a transjugular in-trahepatic portosystemic shunt and the ability of these

From the Departments of Medicine (WW, LB, FW) and Nuclear Car-diology (PL), The Toronto Hospital, University of Toronto, Toronto,Ontario, Canada; and the Institute of Gastroenterology (LB), Tel AvivMedical Center, Ichilov Hospital, Tel Aviv, Israel.

Florence Wong is the recipient of a Fellowship from the MedicalResearch Council of Canada. Part of this work was performed whileholding a Fellowship in Hepatology from Schering Canada. Peter Liu isa Career Investigator of the Heart and Stroke Foundation of Ontario,Canada.

Requests for reprints should be addressed to Florence Wong, 9EN/220, The Toronto Hospital, 200 Elizabeth Street, Toronto M5G 2C4,Ontario, Canada.

Manuscript submitted May 19, 1998, and accepted in revised formOctober 29, 1998.

q1999 by Excerpta Medica, Inc. 0002-9343/99/$–see front matter 315All rights reserved. PII S0002-9343(99)00029-7

patients to handle a sodium load when they had become freeof ascites. The associated pathophysiological changes andhormonal status were also measured to gain insight into thepathogenesis of renal sodium retention in cirrhosis.

METHODS

Approval for the study was granted by the Human Exper-imentation Committee of the University of Toronto, andall patients gave informed consent.

PatientsFrom November 1993 to November 1995, 20 patientswith biopsy proven cirrhosis and refractory ascites re-ceived the shunt as a treatment for their ascites. In thepost-shunt period, 3 patients died during the early post-shunt period, 4 patients received a liver transplant, 1 pa-tient was lost to follow-up, and 2 patients did not consentto participate in the study. The remaining 10 patients whohad survived for more than 1 year after undergoing asuccessful shunt insertion were studied. The initial resultsof some of these patients pre- and up to 1 month post-shunt have been previously reported (5,11). Refractoryascites was defined as being unresponsive to 400 mg spi-ronolactone or 30 mg amiloride plus 160 mg furosemidedaily for 2 weeks. All patients fulfilled the selection crite-ria for shunt insertion of having no renal or cardiovascu-lar disease, no evidence of sepsis, and having a patentportal vein (13). The study group consisted of 7 men and3 women, aged 35 to 76 years (mean 6 SD, 58 6 12). Theetiology of the cirrhosis was alcohol-related in 7 patients,hepatitis B viral infection in 2 patients, and alcohol to-gether with hepatitis C viral infection in 1 patient. Child-Pugh class was determined (14). There were 9 Child-Pugh class B patients (score 7–9) and 1 Child-Pugh classC patient (score $10). All patients had refractory ascitesfor more than 6 months. Patients with alcoholic cirrhosishad abstained from alcohol for at least 6 months beforeshunt insertion.

Study ProtocolPatients were studied on three occasions: before, at 6months, and at a mean of 14 6 2 months after shuntinsertion (when they had remained free of ascites for atleast 6 months). All studies were performed as inpatientsin the Clinical Investigation Unit of the Toronto Hospi-tal. On each occasion, patients were admitted for stabili-zation for 7 days before clinical studies and maintainedon a diet of 22 mmol sodium and 1.5 L fluid per day,without diuretic therapy. They were monitored with dailyweight, serum electrolyte level, and 24-hour urinary so-dium excretion measurements. At 14 months post-shunt,after completion of the studies on the 22 mmol sodium,1.5 L fluid per day diet, dietary sodium intake was in-creased to 200 mmol per day with a 2 L fluid per day

intake for a further 7 days, and all studies were repeated.Doppler ultrasound was performed to assess shunt pa-tency and flow velocity during the post-shunt admissionsand was repeated at 14 months to exclude the reappear-ance of ascites. A chest radiograph was performed duringeach admission to detect pleural effusions. Smoking wasdiscouraged and caffeine-containing drinks and fooditems were not allowed.

On each occasion, on the eighth day of the diet, after atleast 2 hours of bedrest, the patients had blood drawnfrom a forearm venous catheter inserted the previousnight to measure plasma norepinephrine and aldosteroneconcentrations, and plasma renin activity. Renal hemo-dynamic values, sodium homeostasis, and liver functiontests, including indocyanine green clearance, were as-sessed in the morning. In the afternoon, systemic hemo-dynamic values and central blood volume were measuredby radionuclide angiography (15). Forearm circulationwas assessed at 14 months post-shunt, before and aftersodium loading.

Glomerular filtration rate was measured by inulinclearance, and renal plasma flow was measured by para-amino-hippurate clearance. Renal vascular resistance wasestimated (5,18). Lithium clearance was used as an indi-cator of proximal renal tubular reabsorption of sodium(16,17). Forearm blood flow was measured by venousocclusion plethysmography (18) and forearm vascularresistance calculated from the mean arterial pressure andforearm blood flow. Cardiac output and systemic vascu-lar resistance were estimated from mean arterial pressure;heart rate and stroke volume measurements were ob-tained during radionuclide angiography (15). Centralblood volume was corrected for body surface area usingthe subject’s height and weight. The cardiac output wascorrected for body surface area to yield the cardiac index.Cardiac and central vascular blood volume were calcu-lated from total central blood volume minus right and leftpulmonary vascular volumes. Glomerular filtration rateand renal plasma flow were corrected for body surfacearea and expressed per 1.73 m2. Renal vascular resistancewas defined as mean arterial pressure divided by renalblood flow, where renal blood flow was equal to renalplasma flow divided by (1 2 hematocrit). Proximal tubu-lar reabsorption of sodium was calculated as glomerularfiltration rate minus lithium clearance divided by the glo-merular filtration rate (19). Indocyanine green (Cardio-green, Becton Dickinson and Co., Cockeysville, Mary-land) percent retention at 20 minutes was calculated fromthe indocyanine green clearance curve, which has beenvalidated as a quantitative measurement of liver functionand a predictor of survival in patients with cirrhosis (20).

Serum and urinary sodium and lithium concentra-tions were determined by standard flame photometrytechniques. Measurements of plasma renin activity,

Long-term Effects of Transjugular Intrahepatic Portosystemic Shunt/Wong et al

316 March 1999 THE AMERICAN JOURNAL OF MEDICINEt Volume 106

plasma aldosterone level, and plasma norepinephrinelevel were performed as previously described (5).

Statistical AnalysisAll results are expressed as mean 6 standard deviation(SD). Differences by time for the studies performed whileon a 22 mmol sodium diet were assessed by repeated mea-sures of analysis of variance. Dunnett’s test was used todetermine the statistical significance of differences be-tween times. Student’s t tests for independent and pairedsamples were used to analyze the effect of sodium load-ing. A P value of less than 0.05 was considered statisticallysignificant.

RESULTS

Shunt insertion was successful in all patients. The cor-rected sinusoidal pressure fell from 23 6 3 mm Hg to 7 61 mm Hg (P ,0.001) immediately after shunt insertion.Deterioration of liver function occurred immediately af-ter shunt insertion, as indicated by significant increases inthe prothrombin time (16 6 2 sec), the serum bilirubinlevel (44 6 15 mmol/L), and the Pugh score (10 6 2),and a significant decrease in the serum albumin level(28 6 4 g/L). The Pugh score improved during the fol-low-up period, when measured at 6 and 14 months ascompared with baseline, primarily due to the eliminationof ascites (Table 1). Serum albumin level and prothrom-bin time were unchanged at 6 and 14 months post-shuntcompared with baseline. Serum bilirubin level increasedat 6 months, mainly from unconjugated bilirubin due toshunt hemolysis. There was a corresponding fall in hemo-globin concentration that recovered by 14 months (Table1). Hepatic function, as measured by the half-life of in-docyanine green and plasma retention at 20 minutes im-proved at 6 and 14 months; the improvement in plasmaretention at 20 minutes was significant at 14 months (Ta-ble 1).

During the initial 3-month follow-up period, 4 pa-tients with shunt stenosis required additional shunts tomaintain adequate flow and normal sinusoidal pressure.

However, throughout the post-shunt study period, all 10patients had functioning shunts, with a shunt flow veloc-ity of 142 6 47 cm/sec. Ascites had disappeared in allpatients at 6 months, while right-sided pleural effusionsin 2 patients had decreased substantially in size at 14months. However, 4 patients developed lower limbedema after elimination of ascites, 3 of whom requireddiuretics intermittently. (These were discontinued beforeadmission for studies.) After the first 3 months, enceph-alopathy was easily controlled in all 10 patients with orallactulose. By 14 months post-shunt, all 10 patients hadreturned to independent living in the community, withthe 2 youngest patients returning to work full time.

Long-term Follow-up StudiesSystemic and renal hemodynamic measurements. Fol-lowing shunt insertion, systemic hemodynamic measure-ments were largely unchanged, although there was atrend toward an increase in the cardiac index and a de-crease in the systemic vascular resistance (Table 2). A sig-nificant fall in renal vascular resistance (6,259 6 2,925dyne.s.cm25 versus 13,453 6 7,009 dyne.s.cm25 at base-line, P ,0.01), associated with an increase in renal bloodflow (555 6 320 mL/min versus 398 and 232 mL/min atbaseline, P ,0.05) and in glomerular filtration rate(103 6 69 mL/min versus 66 6 45 mL/min at baseline, P,0.05), was also observed at 14 months post-shunt.

Renal sodium and water handling. The 24-hour uri-nary sodium excretion at both 6 months (51 6 36 mmol/day) and 14 months (22 6 10 mmol/day) were signifi-cantly (P ,0.05) increased compared with that at base-line (7 6 5 mmol/day) while on the same low-sodium (22mmol) diet for 7 days. At 14 months, the 24-hour urinarysodium excretion on the day of admission was 115 6 78mmol/day, reflecting ambulatory sodium intake. Afterbeing maintained on a diet of 22 mmol sodium per daywithout diuretics, the patients lost 1.8 6 0.9 kg bodyweight by day 7, and their 24-hour sodium excretion fellto 22 6 10 mmol/day, confirming sodium balance. So-dium excretion measured during the 3-hour renal studyperiod also increased throughout the follow-up period,

Table 1. Mean (6 SD) Clinical, Hematological Measurements and Liver Function in 10 Ascitic Patients before, and 6 and 14 Monthsafter, Transjugular Intrahepatic Portosystemic Stent Shunt (TIPS) Insertion while on a 22 mmol Sodium per Day Diet

Measurement (Normal Range) Pre-TIPS 6 Months 14 Months

Serum albumin (35–55 g/L) 33 6 6 34 6 5 32 6 8Bilirubin (#20 mmol/L) 25 6 13 37 6 24 36 6 17Prothrombin time (10 6 1 sec) 13.7 6 1.9 14.9 6 2.3 14.1 6 1.4Hemoglobin (160 6 20 g/L) 118 6 8 108 6 20 122 6 23Pugh score 8.2 6 1.0 6.3 6 1.3† 6.9 6 1.5†

Indocyanine-green retention at 20 minutes (,4%) 47 6 20 34 6 15 32 6 19*Half-life of indocyanine-green (minutes) (,10.4 minutes) 21 6 9 14 6 8 15 6 6

* P ,0.05 compared with baseline (pre-TIPS).† P ,0.01 compared with baseline.


March 1999 THE AMERICAN JOURNAL OF MEDICINEt Volume 106 317

and the increase was significant at both 6 months and at14 months compared with baseline (Figure 1). This wasdue to a reduction in proximal renal tubular reabsorptionof sodium leading to a significant increase in the frac-tional excretion of sodium (Figure 1).

Central blood volume. Total central blood volume andcardiac and central vascular volumes increased frombaseline to 14 months post-shunt, with the increase incentral blood volume almost reaching statistical signifi-cance at 6 and 14 months (P 5 0.06). In contrast, pul-

Table 2. Mean (6 SD) Systemic Hemodynamic Measurements, Total Central Blood Volume, Pulmonary Vascular Volume, Cardiacand Central Vascular Volume in 10 Ascitic Patients before and at 6 and 14 Months after Transjugular Intrahepatic PortosystemicStent Shunt (TIPS) Insertion while on a 22 mmol Sodium per Day Diet


Mean arterial pressure (70–105 mm Hg) 78 6 7 82 6 5 84 6 7Cardiac index (2.4–3.8 L/min/m2) 2.8 6 1.0 3.8 6 0.4 3.5 6 0.1Systemic vascular resistance (700–1600 dyne.s.cm25) 1459 6 541 1154 6 401 1058 6 339Central blood volume (1280 6 320 mL/m2) 1204 6 313 1463 6 370 1518 6 464Right pulmonary vascular volume (220 6 65 mL/m2) 158 6 70 186 6 69 174 6 54Left pulmonary vascular volume (185 6 58 mL/m2) 154 6 57 141 6 46 133 6 27Cardiac and central vascular volume (880 6 260 mL/m2) 991 6 248 1173 6 355* 1177 6 356*

* P ,0.05 compared with baseline (pre-TIPS).

Figure 1. Daily urinary sodium excretion (a), proximal tubular reabsorption of sodium (b), and fractional excretion of sodium (c)during the 3-hour renal study period in 10 ascites-free cirrhotic patients before and 6 and 14 months after transjugular intrahepaticportosystemic stent shunt (TIPS) insertion while on a 22 mmol sodium per day diet. *P ,0.05 versus baseline; †P ,0.01 versusbaseline.



monary vascular volumes did not change significantlyduring the follow-up period (Table 2).

Hormonal profile. At baseline, plasma renin activities,plasma aldosterone levels, and plasma norepinephrinelevels were greater than normal. Both plasma renin activ-ity and plasma aldosterone levels were normal at 6 and 14months (Table 3). In contrast, plasma norepinephrinelevels remained unchanged and elevated throughout thepost-shunt follow-up period (Table 3).

Effects of Sodium Loading at 14 MonthsSodium loading resulted in a weight gain of 2.1 6 1.3 kgon day 7 of the 200 mmol sodium per day diet. One pa-tient with ischemic heart disease developed signs of heartfailure at the end of this period, but made an unremark-able recovery. Two other patients had a slight increase inpedal edema. No patient developed clinically appreciableascites. Repeat ultrasound performed on day 7 showedreaccumulation of a small amount of ascites in 1 patientonly. No reaccumulation of pleural effusion occurred inthe 2 patients who had pre-shunt effusions.

Sodium loading was not associated with any change insystemic, renal, or forearm hemodynamics (Table 4), butwas associated with a significant increase in central bloodvolume that was distributed in its component compart-ments (Table 5). Sodium loading promptly resulted in anincrease in daily urinary volume from 1,236 6 768 mL to1,810 6 649 mL (P ,0.01), and sodium excretion in-creased from 22 6 10 mmol/day to 188 6 32 mmol/dayon day 7 of the high sodium diet (P ,0.01). Sodiumexcretion during the 3-hour study period also increased

significantly when patients were changed from the low-to high-sodium diet (Figure 2). This was due to a reduc-tion in both the proximal and distal renal tubular reab-sorption of sodium, leading to a significant increase in thefractional excretion of sodium (Figure 2).

Increased sodium intake led to a further physiologicalsuppression of the renin-aldosterone system. Plasma re-nin activity decreased from 0.46 6 0.46 ng/L/sec during22 mmol sodium intake to 0.16 6 0.10 ng/L/sec withsodium loading (P ,0.01), while plasma aldosteronelevels fell from 532 6 342 pmol/L with 22 mmol sodiumintake to 172 6 81 pmol/L with sodium loading (P,0.01). In contrast, plasma norepinephrine levels re-mained elevated (1.6 6 0.9 nmol/L with 22 mmol sodiumper day diet versus 1.7 6 1.1 nmol/L with sodium load-ing).

DISCUSSION

Transjugular intrahepatic portosystemic shunting is ef-fective in managing refractory ascites in cirrhosis in suit-ably selected patients (5–7,21). We have demonstratedthat sodium balance is maintained in these patients whileon a low sodium diet 6 months after shunt placement,with reduced sinusoidal portal pressure and a normal re-nin-aldosterone system. Furthermore, sodium handlingis improved in the post-shunt patients who are ascites-free and challenged with a sodium load of 200 mmol perday. However, sodium balance was not maintained in allpost-shunt patients, as evidenced by the development ofminimal ascites or pedal edema.

Table 4. Mean (6 SD) Systemic and Renal Hemodynamic Measurements and Status of the Forearm Circulation in 10 Ascites-freeCirrhotic Patients before and after Sodium Loading with 200 mmol per Day, 14 Months after Transjugular Intrahepatic Portosys-temic Stent Shunt (TIPS) Insertion for Refractory Ascites

Measurement (Normal Range)22 mmol

Sodium Diet200 mmol

Sodium Diet

Cardiac index (2.4–3.8 mL/min per 1.73 m2) 3.5 6 0.4 3.8 6 0.7Systemic vascular resistance (700–1600 dyne.s.cm25) 1058 6 339 1054 6 389Glomerular filtration rate (124 6 26 mL/min per 1.73 m2) 103 6 69 106 6 57Renal plasma flow (654 6 163 mL/min per 1.73 m2) 555 6 320 545 6 253Renal vascular resistance (6760 6 1730 dyne.s.cm25) 6257 6 2925 6813 6 3735Forearm blood flow (2.3 6 0.8 mL/min per 100 mL tissue) 4.2 6 2.4 4.1 6 1.5Forearm vascular resistance (50 6 10 mm Hg/mL/min per 100 mL tissue) 26 6 15 23 6 9

Table 3. Mean (6 SD) Hormonal Measurements in 10 Ascitic Cirrhotic Patients before and at 6 and 14 Months after TransjugularIntrahepatic Portosystemic Stent Shunt (TIPS) Insertion while on a 22 mmol Sodium per Day Diet


Plasma renin activity (1.0 6 0.9 ng/L/sec) 2.5 6 1.2 0.6 6 0.4* 0.5 6 0.5*Plasma aldosterone level (650 6 280 pmol/L) 1950 6 639 698 6 447* 532 6 342*Plasma norepinephrine level (,1.2 nmol/L) 2.3 6 1.2 1.8 6 1.0 1.6 6 0.9

* P ,0.01 compared with baseline (pre-TIPS).



The negative sodium balance during the first 6 monthswas associated with elimination of ascites and peripheraledema, together with normalization of plasma renin ac-tivity and serum aldosterone concentrations, indicatingadequate effective arterial blood volume. This confirmsthat, in the decompensated patient with cirrhosis and as-cites, effective arterial underfilling is at least partly re-sponsible for renal sodium retention. The importance ofthe renin-angiotensin-aldosterone system as a cause ofsodium retention was first suggested by the observationthat many, but not all, of these patients have increasedplasma renin activity and hyperaldosteronism (22,23),and that those who respond to diuretics appear to be ex-quisitely sensitive to the proximal sodium retaining ef-fects of angiotensin II (24).

We have previously reported that plasma renin activityand plasma aldosterone levels were normal in patientswithin the first month after shunt placement, at a timewhen sodium excretion was significantly lower (5). Sig-nificant worsening of hepatic function can occur in theearly post-shunt period (5,11,25) and was also observedin this group of patients. This could have delayed theonset of natriuresis after shunt insertion. A subsequentreturn of liver function to baseline levels at 6 months, as

indicated by the improved indocyanine green clearance,along with an adequately filled intravascular volume anda normal sinusoidal pressure coincided with further na-triuresis.

Once ascites was eliminated, renal sodium excretionwas in balance with the 22 mmol sodium daily intake.This was maintained at 14 months post-shunt, and wasassociated with normalization of the glomerular filtrationrate. Intravascular volume was adequate, as judged by therenin-aldosterone activity and central blood volume. Themean arterial pressure was maintained by the combina-tion of increased sympathetic activity and improved sys-temic hemodynamic measurements and intravascularfilling.

Sodium loading was performed for two reasons.Firstly, post-shunt patients without portal hypertensionmay handle a sodium load differently. Second, a low-sodium diet is not a very attractive diet: It is unpalatableand requires a great effort in food selection. For patientswho no longer have ascites, the incentive to reduce so-dium intake is no longer present. Therefore, it would beimportant to be able to advise patients whether to in-crease their sodium intake based on solid scientific evi-dence.

Figure 2. Daily urinary sodium excretion (a) and fractional excretion of sodium (b) during the 3-hour renal study period in 10ascites-free cirrhotic patients before and after sodium loading with 200 mmol per day 14 months after transjugular intrahepaticportosystemic stent shunt (TIPS) insertion.

Table 5. Total Central Blood Volume, Pulmonary Vascular Volume, Cardiac and Central Vascular Volume Measurements in 10Ascites-free Cirrhotic Patients before and after Sodium Loading with 200 mmol per Day 14 Months after Transjugular IntrahepaticPortosystemic Stent Shunt (TIPS) Insertion for Refractory Ascites

Measurement (Normal Range) 22 mmol Sodium Diet 200 mmol Sodium Diet

Central blood volume (1280 6 320 mL/m2) 1518 6 464 1913 6 773*Right pulmonary vascular volume (220 6 65 mL/m2) 174 6 54 215 6 88Left pulmonary vascular volume (185 6 58 mL/m2) 133 6 27 183 6 62†

Cardiac and central vascular volume (880 6 260 mL/m2) 1177 6 356 1431 6 651

* P ,0.05 compared with baseline.† P ,0.01 compared with baseline.



Sodium loading in the post-shunt patients at 14months resulted in replenishment of the intravascularvolume with a mean weight gain of 2.1 kg, similar to thatobserved in normal control subjects (25). However, in 4patients, the increase in weight and volume resulted inpedal edema or cardiac failure. However, on average, thehandling of a sodium load in these post-shunt patientswas better than in patients with cirrhosis who have not yetdeveloped ascites; such patients uniformly retain sodium(25). Excess volume in preascitic patients with portal hy-pertension is redistributed to the splanchnic area, ratherthan to the periphery as in the post-shunt patients withnormal portal pressure. Sodium handling following a so-dium load in these post-shunt patients was also betterthan that in patients following peritoneovenous shunts,who excreted a mean of 56 mmol of sodium while on a100 mmol sodium diet (4). Both shunted groups had nor-mal renin-aldosterone systems that were suppressed nor-mally when patients were challenged with a high-sodiumdiet. The main difference between these two groups ofshunted patients is that following transjugular intrahe-patic portosystemic shunting, sinusoidal portal hyper-tension has been eliminated, confirming the importanceof portal hypertension in the pathogenesis of sodium re-tention in cirrhosis.

Suppression of the renin-angiotensin-aldosterone sys-tem by volume expansion associated with sodium load-ing was mainly responsible for the reduction in tubularreabsorption of sodium, as there were not any furtherchanges in renal hemodynamic measurements. The ab-sence of portal hypertension in these post-shunt patientsmeant that the excess volume associated with sodium re-tention was not preferentially localized to the splanchniccirculation but remained in the systemic circulation, asindicated by an increase in central blood volume. How-ever, sodium loading did not result in any further in-crease in volume in either the renal or forearm circula-tions, highlighting the heterogeneity of the response ofvascular beds to physiological stimuli, particularly in thepresence of increased sympathetic activity. This has pre-viously been demonstrated in preascitic cirrhotic pa-tients, in whom deactivation of baroreceptors led to fore-arm vasoconstriction but not to any changes in the renalcirculation (18). Persistent activation of the sympatheticnervous system may be a central effect, at least in an ani-mal model of cirrhosis (26). Alternatively, increased sym-pathetic nervous system activation may be a consequenceof hepato-renal dysfunction, associated with hormonal(27) or neurogenic hepato-renal interactions (28,29). Al-though persistent sodium retention apparently resultedin clinical evidence of fluid retention in 4 patients, pedaledema may have been aggravated by local capillary fac-tors associated with increased systemic arterial vasodila-tation.

In summary, following transjugular portosystemic

shunt, many patients with cirrhosis who are free of ascitescan often tolerate a daily intake of 200 mmol of sodium.This is associated with a reduction in sinusoidal portalpressure, and adequate filling of the intravascular vol-ume. However, persistently elevated sympathetic activitymay predispose some of these patients to a sodium-re-taining tendency. Therefore, caution should still be exer-cised when advising these patients to increase their so-dium intake.

ACKNOWLEDGEMENTSSpecial thanks are extended to Yasmin Allidina of the NuclearCardiology Department for the image analysis and data entry.The authors also gratefully acknowledge the assistance of Ms.Sue Chung, dietetic assistant; Mrs. Nancy Law, nurse teacher;and the dietetic and nursing staff of the Clinical InvestigationUnit at the Toronto Hospital.

REFERENCES1. Schrier RW, Arroyo V, Bernardi M, et al. Peripheral arterial vaso-

dilatation hypothesis: a proposal for the initiation of renal sodiumand water retention in cirrhosis. Hepatology. 1988;8:1151–1157.

2. Blendis LM, Greig PD, Langer B, et al. The renal and hemodynamiceffects of the peritoneovenous shunt for intractable hepatic ascites.Gastroenterology. 1979;77:250 –257.

3. Blendis LM, Sole MJ, Campbell P, et al. The effect of peritone-ovenous shunting on catecholamine metabolism in patients withhepatic ascites. Hepatology. 1987;7:143–148.

4. Greig PD, Blendis LM, Langer BR, et al. The renal and hemody-namic effects of the peritoneovenous shunt. Long term effect. Gas-troenterology. 1981;80:119 –125.

5. Wong F, Sniderman K, Liu P, et al. Transjugular intrahepatic por-tosystemic stent shunt: effects on hemodynamics and sodium ho-meostasis in cirrhosis and refractory ascites. Ann Intern Med. 1995;122:816 – 822.

6. Somberg KA, Lake JR, Tomlanovich SJ, et al. Transjugular intrahe-patic portosystemic shunts for refractory ascites: assessment of clin-ical and hormonal response and renal function. Hepatology. 1995;21:709 –716.

7. Quiroga J, Sangro B, Nunez M, et al. Transjugular intrahepaticportosystemic shunt in the treatment of refractory ascites: effect onclinical, renal, humoral and hemodynamic parameters. Hepatology.1995;21:986 –994.

8. Morali G, Sniderman K, Deitel KM, et al. Is sinusoidal portal hy-pertension a necessary factor for the development of hepatic ascites.J Hepatology. 1992;16:249 –250.

9. Rector WG Jr. Portal hypertension: a permissive factor only in thedevelopment of ascites and variceal bleeding. Liver. 1986;6:221–226.

10. Azoulay D, Castaing D, Dennison A, et al. Transjugular intrahe-patic portosystemic shunt worsens the hyperdynamic circulatorystate of the cirrhotic patient: preliminary report of a prospectivestudy. Hepatology. 1994;19:129 –132.

11. Wong F, Sniderman K, Liu P, Blendis LM. The mechanism of theinitial natriuresis post-TIPS in cirrhotic patients with refractoryascites. Gastroenterology. 1997;112:899 –907.

12. Lebrec D, Giuily N, Hadengue A, et al. Transjugular intrahepaticportosystemic shunts: comparison with paracentesis in patients



with cirrhosis and refractory ascites. A randomized trial. J Hepatol-ogy. 1996;25:135–144.

13. Wong F, Blendis LM. Transjugular intrahepatic portasystemicshunt-Tipping the sodium balance. Hepatology. 1995;22:358 –364.

14. Pugh RNH, Murray-Lyon IM, Dawson JL, et al. Transection of theoesophagus for bleeding oesophageal varices. Br J Surg. 1973;60:646 – 649.

15. Wong F, Liu P, Tobe S, et al. Central blood volume in cirrhosis:measurement by radionuclide angiography. Hepatology. 1994;19:312–321.

16. Wong F, Massie D, Hsu P, Dudley F. The renal response to a salineload in well compensated alcoholic cirrhosis. Hepatology. 1994;20:873– 881.

17. Thomsen K, Schou M, Steiness I, Hansen HE. Lithium as an indi-cator of proximal sodium reabsorption. Pflugers Arch. 1969;308:180 –184.

18. Wong F, Logan AG, Blendis LM. Systemic hemodynamic, forearmvascular, renal and humoral response to sustained cardiopulmo-nary baroreceptor deactivation in cirrhosis. Hepatology. 1995;21:717–724.

19. Boer WH, Koomans HA, Dorhout Mees EJ. Renal hemodynamicsand sodium handling after hypertonic albumin infusion in sodium-restricted normal man. Eur J Clin Invest. 1987;17:442– 447.

20. Merkel C, Gatta A, Zoli M, et al. Prognostic value of galactose elim-ination capacity, aminopyrine breath test, and ICG clearance inpatients with cirrhosis. Comparison with Pugh score. Dig Dis Sci.1991;36:1197–1203.

21. Ochs A, Rossle M, Haag K, et al. The transjugular intrahepatic por-

tosystemic stent shunt procedure for refractory ascites. NEJM.1995;332:1192–1197.

22. Wong F, Legault L, Tobe S, et al. Refractory ascites in cirrhosis: theroles of volume expansion and plasma atrial natriuretic factor ele-vation. Hepatology. 1993;18:519 –528.

23. Salo J, Gines A, Anibarro L, et al. Effect of upright posture andphysical exercise on endogenous neurohormonal systems in cir-rhotic patients with sodium retention and normal plasma renin,aldosterone and norepinephrine. Hepatology. 1995;22:479 – 487.

24. Tobe S, Blendis LM, Morali GA, et al. Angiotensin II modulatesANF induced natriuresis in cirrhosis with ascites. Am J Kidney Dis.1993;21:472– 479.

25. Wong F, Liu P, Allidina Y, Blendis LM. Pattern of sodium handlingand its consequences in pre-ascitic cirrhosis. Gastroenterology.1995;108:1820 –1827.

26. Breitman DR, Lee SS. Blunted responsiveness of neuronal activa-tion marker FOS in brainstem cardiovascular nuclei of cirrhoticrats. Gastroenterology. 1997;112(suppl):A1231. Abstract.

27. Wong F, Logan AG, Blendis LM. Effects of insulin on renal hemo-dynamics, sodium homeostasis, forearm blood flow and sympa-thetic nervous activity in pre-ascitic cirrhosis. Hepatology. 1996;23:414 – 422.

28. Lang F, Tschernko E, Schulze E, et al. Hepatorenal reflex regulatingkidney function. Hepatology. 1991;14:590 –594.

29. Koepke J, Jones S, DiBona GF. Renal nerves mediate blunted natri-uresis to atrial natriuretic peptide in cirrhotic rats. Am J Physiol.1987;252:R1019 –R1023.



Documents

Long-term renal sodium handling in patients with cirrhosis treated with transjugular intrahepatic portosystemic shunts for refractory ascites