sign in sign up
<<
Home , Gastroenterology, Hyperdynamic circulation

Gut, 1992, 33, 965-968 965 in acute failure

M Navasa, J C Garcia-Pagan, J Bosch, J R Riera, R Bafnares, A Mas, M Bruguera, J Rodes : first published as 10.1136/gut.33.7.965 on 1 July 1992. Downloaded from

Abstract hypertension, such as and hepatorenal Twenty five patients with acute syndrome.34 It has been suggested that portal were measured for hepatic venous pressure hypertension could be caused by an increased gradient as an index of portal pressure during hepatic resistance to portal blood flow, as a the course of a transjugular . consequence of sinusoidal collapse and distor- Hepatic venous pressure gradient ranged from tion of liver architecture after extensive liver cell 4 to 24.5 mm Hg with a mean of 12.8 (5.3) mm necrosis.3 5 Hg (normal values <5mm Hg). AU patients but Transjugular liver biopsy allows histological one had increased portal pressure gradient. and haemodynamic studies to be carried out Portal hypertension correlated with the degree during the course of fulminant hepatic failure.6 of architectural distortion of the liver, as This allows better characterisation of portal suggested by a direct correlation between hypertension and the examination of the hepatic venous pressure gradient and the area relationship between portal pressure, presence of reticulin collapse, evaluated by means of a of ascites and renal failure, and the extent of morphometric analysis on Sirius red stained histological damage. This study presents our liver slides (r=0.43, p<005). Hepatic venous findings in a series of 25 consecutive patients. pressure gradient was significantly higher in patients with ascites (15.1 (5) mm Hg, n=15) or renal failure (14.4 (5.3) mm Hg, n=16) than in Methods Liver Unit, those without (9.3(3.4) mm Hg and 10*1 (4) mm i Provincial, Hg, respectively; p<0.05). Portal hyperten- PATIENTS University of Barcelona was out in 25 patients with and Hospital Valle del sion was associated with systemic The study carried Nalon, Rianio-Langreo, and a hyperkinetic circulatory state, with admitted to the Liver Intensive Asturias, Spain decreased arterial pressure, and peripheral Care Unit during a three year period. Fifteen M Navasa resistance and increased . patients were male and 10 female, with an age J C Garcia-Pagan J Bosch ranging from 14 to 62 years (mean (SD) 34 (15)

J R Riera years) (Table). The aetiology of liver failure was http://gut.bmj.com/ R Bafiares Fulminant hepatic failure is a syndrome assessed by clinical history, and by serologic A Mas M Bruguera characterised by short term development of signs markers of B, hepatitis delta, and J Rodes of advanced liver failure leading to hepatic . The possibility of Wilson's Correspondence to: encephalopathy in the absence of previous liver presenting as fulminant hepatic failure was J Bosch, Hepatic by measuring plasma Hemodynamic Laboratory, disease.' Fulminant hepatic failure has a very investigated in all patients Liver Unit, Hospital Clinic i high mortality, especially when caused by acute and urine copper and plasma ceruloplasmin

Provincial, Villaroel 170, on October 1, 2021 by guest. Protected copyright. 08036 Barcelona, Spain. , which is the most common cause concentrations. diagnosis of acute liver failure was based Accepted for publication of this syndrome.'2 Over half of the patients The 21 October 1991 exhibit clinical signs associated with portal on the accepted clinical criteria: development of

Clinical data ofthe 25 patients studied Patient Age Sex Enc Creat HVPG CO SVR no (yr) MIF Aetiology FullSub (grade) Ascites (smolll) (mm Hg) (i/minl) (dynls.cm-5) Death 1 40 M D Sub IV Yes 230 20.5 4 1800 Yes 2 16 F D Ful II No 71 8 - - No 3 39 M nAnB Ful III No 177 11 - - Yes 4 27 M nAnB Sub IV Yes 265 22 7-5 938 Yes 5 41 M B Sub IV Yes 972 16 8-9 952 Yes 6 51 M nAnB Sub IV Yes 804 14 - - Yes 7 24 M nAnB Ful IV No 221 6-5 13 369 Yes 8 22 M D Sub IV No 53 12-5 - - No 9 55 F Haloth Sub II Yes 53 12 8-1 829 No 10 62 M A Ful IV Yes 946 7 8-2 829 Yes 11 27 F D Sub IV Yes 354 14 - - Yes 12 22 M B Ful IV No 35 8 - - No 13 52 F nAnB Sub IV No 318 7 6-3 977 Yes 14 51 F nAnB Sub IV Yes 557 24-5 6-4 1375 Yes 15 23 F D Sub IV Yes 97 9 - - Yes 16 62 M B Sub IV Yes 610 18 - - Yes 17 42 F nAnB Sub IV No 132 15 - - Yes 18 21 M D Sub II Yes 97 18 - - No 19 24 F B Ful II No 53 7 5 - - No 20 23 M D Sub IV Yes 919 16-5 6.5 1070 Yes 21 14 F nAnB Sub II Yes 132 12 6-2 1058 No 22 25 M nAnB Sub IV No 857 13 13 363 Yes 23 43 M B Ful IV Yes 185 9 - - Yes 24 21 M D Ful IV Yes 238 15 8-4 971 No 25 26 F B Ful IV No 132 4 5.5 1381 Yes Ful/Sub: fulminant/subfulminant; Enc: encephalopathy; Creat: creatinine; HVPG: hepatic venous pressure gradient; CO cardiac output; SVR: systemic ; M/F: male/female; D: virus delta hepatitis; nAnB: virus non-A non-B hepatitis; B: virus B hepatitis; Haloth: halothane hepatitis; A: virus A hepatitis. 966 Navasa, Garcia-Pagdn, Bosch, Riera, Baniares, Mas, Bruguera, Rodes

and fall in prothrombin and light areas. Three consecutive liver areas index below 40% within eight weeks of the onset were evaluated for each specimen by one of of symptoms in the absence of previous liver us, under blind conditions. The results were disease.' Patients were divided in two groups: expressed as the mean percentage of Sirius red (i) fulminant hepatic failure, when hepatic stained area of the section. The intra assay encephalopathy appeared in less than two weeks variability of the measurement was of 6%. Gut: first published as 10.1136/gut.33.7.965 on 1 July 1992. Downloaded from after the onset of , and (ii) subfulminant hepatic failure ifhepatic encephalopathy compli- cated liver failure two to eight weeks after the STATISTICAL ANALYSIS onset of jaundice.' All values are reported as mean (SD). The After admission, the patients were transferred Student's t test, the X' test and the coeficient of to the Liver Intensive Care Unit. Physical exami- correlation were used in the statistical analysis of nation including the presence or development of the results. Statistical significance was estab- ascites and evaluation ofneurological status were lished at p<0 05. performed at least every four hours. All patients had continuous moniforing of arterial pressure and urine output. Biochemical analysis includ- Results ing standard liver and renal function tests, as The mean time between the onset of clinical performed daily until death or recovery. Renal symptoms and the haemodynamic study was 21 failure was considered to be present when plasma days, with a range ofsix to 56 days. The course of creatinine increased above 132 [imol/l. liver failure was subfulminant in 16 patients Within 48 hours of admission, all patients (64%) and fulminant in the remaining nine underwent a haemodynamic study during trans- patients. jugular liver biopsy. Briefly, under local anaes- The aetiology of acute liver failure was viral thesia a venous introducer (USCI International hepatitis in all but one patients. Six patients Inc, Mass, USA) was placed in the right internal had acute (24%), eight acute delta jugular vein by the Seldinger technique, and the (32%), one had hepatitis A (4%), one main right hepatic vein was catheterised under had halothane hepatitis and the remaining nine fluoroscopy using a 7F balloon tipped catheter cases were considered as non-A-non-B hepatitis (Medi-Tech, Cooper Scientific Corp, Water- (36%) (Table). Ascites was detected in 15 town, MA, USA); repeated measurements of patients (60%). The ascites had a low protein wedged and free hepatic venous pressures were content (14 (6) g/l). Renal failure developed in 16 obtained by inflating and releasing the balloon. (64%). Portal pressure was estimated from the hepatic As shown in the Table, hepatic venous pressure

venous pressure gradient - that is, the difference gradient ranged from 4 to 24-5 mm Hg with http://gut.bmj.com/ between wedged hepatic venous pressure and a mean of 12-8 (5.3) mm Hg (normal values free hepatic venous pressure.7 In 13 of these <5 mm Hg) (Fig 1). All patients but one had an patients a Swan-Ganz catheter was inserted into increased hepatic venous pressure gradient. the pulmonary artery for measurements of Fifteen patients (60%) had a hepatic venous cardiopulmonary pressures and cardiac output pressure gradient equal or greater than 12 mm (Edwards Laboratory, Los Angeles, CA, USA) Hg, and were considered to have severe portal

(thermal dilution). Arterial pressure was hypertension, while nine patients had minor on October 1, 2021 by guest. Protected copyright. monitored by sphygmomanometry (Dinamap, degrees of portal hypertension (hepatic venous Critikon, Tampa, Florida, USA) and pressure gradient from 6 to 1 1 mm Hg). The 13 recorded by continuous electrocardiograph patients who had measurements of cardiac out- monitoring. Systemic vascular resistance (dyn/ put, showed a hyperdynamic circulation mani- sec.cm- 5) was calculated as (MAP-RAP)*80/ fested by a high cardiac output (7-8 (2.6) 1/min), CO, in which MAP is the mean arterial pressure, and decreased mean arterial pressure (89 (16) RAP is the right atrial pressure, and CO the mm Hg) and systemic vascular resistance (993 cardiac output. All measurements were per- (387) dyn/sec.cm - 5). formed by triplicate and permanent tracings Presence of ascites was related to the severity were obtained on a multichannel recorder of portal hypertension. The hepatic venous (Hewlett Packard 7754B). After that, an 8 F pressure gradient was significantly higher in catheter was introduced in the hepatic vein and a patients with ascites (15.1 (5) mm Hg) than in preshaped liver biopsy needle (Ingenor Medical, those without (9.3 (3.4) mm Hg), (p<001). Paris, France) was slipped into the catheter and Moreover, among the 15 patients with ascites, 12 the transjugular liver biopsy was obtained in (80%) had severe portal hypertension. On the each patient using the standard technique.6 contrary, only three of the 10 patients without There were no complications from these pro- ascites had an hepatic venous pressure gradient cedures. above 12 mm Hg (p<0 02). Liver biopsy specimens were sectioned Patients with renal failure had a higher hepatic serially and stained with haematoxylin and eosin, venous pressure gradient than those without it Masson's trichrome stain, and reticulin stain for (14-4 (5.3) mm Hg v 10-1 (4) mm Hg, respect- conventional histological analysis. In addition, ively; p<0-05). Renal failure was present in 11 of all specimens were stained with Sirius red F3BA the 15 patients with an hepatic venous pressure in order to do a quantitative morphometric gradient above 12 mm Hg. analysis of the reticulin collapse by means of an The hepatic venous pressure gradient was automatic image quantifier (Ramtek-935 1), as significantly higher in patients with subfulmi- previously described.89 The measurements were nant hepatic failure (15-3 (4.6) mm Hg) than in based on the contrast between dark (collapse) those with fulminant hepatic failure (8.4 (3 1) Portal hypertension in acute liverfailure 967

Discussion Portal hypertension usually complicates the evolution of chronic liver .`' Ascites and gastrointestinal bleeding caused by ruptured

I.- oesophageal varices are frequent consequences Gut: first published as 10.1136/gut.33.7.965 on 1 July 1992. Downloaded from of severe portal hypertension." It has been shown that the development of ascites is also common in acute liver failure,34 and has been related to the presence of severe portal hyperten- sion.35 Previous studies from our unit, however, have shown that ascites may occur in patients with acute viral hepatitis in the absence of portal hypertension. 12 These patients exhibited an 'exudative' ascites with a high protein and lymphocyte content and had a good prognosis. 12 The present study further shows that portal hypertension is common in acute liver failure. In fact, all but one of the patients studied had a hepatic venous pressure gradient above the Figure 1: Individual values ofthe hepatic venous pressure gradient (HVPG) in patients with acute liverfailure. The normal values, thus indicating intrahepatic horizontal line indicates the mean value ofthe whole series. sinusoidal portal hypertension.'° Although portal Patients withfulminant liverfailure are represented by pressure is determined by the interrelationship squares, while those with subfulminant (over two weeks flow and vascular it is course) are represented by circles. Note that the latter had a of blood resistance, greater HVPG than theformer (15 3 (4.6) v 8-4 (3 1) mm generally accepted that portal hypertension is Hg, respectively). The shadowed area represent the normal initiated by an increased resistance to portal range. blood flow.' '°1 The site of increased resistance in patients with acute liver failure has not been clearly established. There is no evidence of mm Hg, p<0 001). Among the 16 patients with sinusoidal fibrosis and nodular regeneration.' subfulminant hepatic failure, all but two (87.5%) Similarly, there is no massive infiltration of the had an hepatic venous pressure gradient above liver, as happens in portal hypertension that may 12 mm Hg. This was only seen in one of the be associated with haematologic disorders or nine patients with fulminant hepatic failure amyloidosis,'5 nor evidence of obstruction at the (p<0 001) (Table). Ascites was more common in hepatic outflow, as observed in venoocclusive

patients with subfulminant hepatic failure than disease.'6 Therefore, the most likely explanation http://gut.bmj.com/ in patients with fulminant hepatic failure (12/16 for the increased hepatic venous pressure v three of nine patients, respectively, p=0.08). gradient is the extensive necrosis and subsequent Reticulin collapse, measured by the percent- reticulin collapse, that probably causes a severe age of the total area of the section stained by distortion of the hepatic microcirculation. The Sirius red averaged 15-7 (8.4)% (mean value in direct correlation found in the present study normal liver 3-1 (0-5%).9 There was a direct between the degree of reticulin collapse and the

correlation between the degree of reticulin hepatic venous pressure gradient, which is very on October 1, 2021 by guest. Protected copyright. collapse and the hepatic venous pressure similar to that obtained by Valla et al5 in patients gradient (r=043, p<005) (Fig 2). with acute viral hepatitis, further supports this There were no significant differences in the hypothesis. The correlation between the area of degree of collapse between patients with fulmi- reticulin collapse, however, and hepatic venous nant and subfulminant hepatic failure (14.6 (90) pressure gradient, although statistically signifi- v 16-2 (8.3)%, NS). cant, was poor. This may indicate that other The whole mortality rate of the present series factors also play an important role in increasing was 68%. There were no significant differences hepatic venous pressure gradient in acute liver in hepatic venous pressure gradient between failure. In addition, it is likely that morpho- patients who survived and those who did not metric analysis in the transjugular biopsies is (11-6 (3.7) v 13.3 (5.9) mm Hg, NS). The area of subject to some degree of sampling error, which reticulin collapse, however, was significantly may obscure this relationship. Measurements of lower in survivors (9.7 (4.7) v 18-1 (8.3)%, hepatic venous pressure gradient with the p<002). balloon occlusion technique, however, estimate the sinusoidal perfusion pressure in a relatively large area of the liver.`' It is therefore possible 30 that measurements of hepatic venous pressure a) r = 0.43; p < 0-05 , 25 0 gradient provide a better estimate of the degree ~ y = 8.6 + 0.27x CO of architectural distortion in acute liver failure 20- 0

E * 0 than morphometric analysis of liver biopsies. n 0 the current of a u) D1 Nevertheless, finding statistic- E)_n *0 0 c c ally significant greater degree of reticulin 0

0 0 collapse in patients who died than in the Figure 2: Correlation CO m . * * 0. 0 between the area of reticulin CL 5- survivors clearly points out the important prog- 0) 0 collapse (as the percentage of I nostic connotation of the severity of histological the total area ofthe biopsy 0 damage in patients with acute liver failure. 8 slide stained by Sirius red) 0 5 10 15 20 25 30 35 and the hepatic venous The present study also shows a high preval- pressure gradient (HVPG). Reticulin collapse (% area) ence ofascites in patients with acute liver failure. 968 Navasa, Garcia-Pagdn, Bosch, Riera, Banlares, Mas, Bruguera, Rodes

Our data and previous studies35 indicate that the carried out in comatosed patients. Our findings presence of ascites is related to the magnitude of indicate that the combined histological haemo- the increase in portal pressure. This would dynamic evaluation provided by this procedure suggest that, as in ,'9 portal hyperten- may help to optimise the medical management of

sion plays an important role in the pathogenesis these critically ill patients. Gut: first published as 10.1136/gut.33.7.965 on 1 July 1992. Downloaded from ofascites in acute liver failure. It is important to note that patients with acute liver failure showed a hyperdynamic circulation, We are indebted to Angels Baringo for his expert assistance in a high cardiac output and decreased these studies and to Lidia Paris and Eulalia Ventura for secretarial evidenced by support. Supported in part by grants from the Fundaci6 Catalana systemic vascular resistances and mean arterial per a l'Estudi de les Malalties del Fetge. Dr J C Garcia-Pagan held pressure. This vasodilatation is thought to be the a research from Fondo de Investigaciones Sanitarias de mechanism triggering the activation of neuro- la Seguridad Social (FISS-89/1516). humoral factors contributing to sodium reten- tion, extracellular fluid volume expansion and 1 Bernuau J, Rueff B, Benhamou JP. Fulminant and sub- ascites formation.'9 In addition, splanchnic vaso- fulminant liver failure: definitions and causes. Semin Liver dilation, which results in increased portal inflow, Dis 1986; 6: 97-106. 2 Tygstrup N, Ranek L. Assessment of prognosis in fulminant may contribute to increase portal pressure, hepatic failure. Semin Liver Dis 1986; 6: 129-37. specially when the hepatic vascular resistance is 3 Lebrec D, Nouel 0, Bernuau J, RueffB, Benhamou JP. Portal hypertension in fulminant viral hepatitis. Gut 1980; 21: increased.'0 The pathogenesis of systemic vaso- 962-4. dilation remains largely conjectural. As sug- 4 Navasa M, Panes J, Teres J, Bruguera M, Rodes J. Insuficiencia hepitica aguda grave: analisis de 51 casos. gested in patients with cirrhosis, it is likely to be Gastroenterol Hepatol 1986; 9: 221-7. of multifactorial origin.'0'9 A possible explana- 5 Valla D, Flejou JF, Lebrec D, Bernuau J, Rueff B, Salzmann JL, et al. Portal hypertension and ascites in acute hepatitis: tion is the accumulation ofvasoactive substances Clinical, hemodynamic and histological correlations. of splanchnic origin in the systemic circulation 1989; 10: 482-7. 6 Lebrec D, Goldfarb G, Degott C, Rueff B, Benhamou JP. that are either metabolised by the normal liver Transvenous liver biopsy. An experience based on 1000 or abnormally released during acute liver hepatic tissue samplings with this procedure. Gastro- enterology 1982; 83: 338-40. failure.'0202' In that regard it has been reported 7 Bosch J, Mastai R, Kravetz D, Navasa M, Rodes J. Hemo- that increased levels of substance P are observed dynamic evaluation of the patient with portal hypertension. SeminLiverDis 1986; 6: 309-17. in patients with advanced liver failure and that 8 Sweat F, Fuchtler H, Rosenthal SI. Sirius red F3BA as a stain these correlate with the degree of hypotension.22 for connective tissue. Arch Pathol 1964; 78: 69-72. 9 Rubio CA, Porwit A. Quantitation offibrosis in liver biopsies. In addition, other factors such as endotoxaemia, Anal Quant Cytol Histol 1988; 10: 107-9. sepsis, and fever may contribute to further 10 Bosch J, Navasa M, Garcia-Pagan JC, DeLacy AM, Rodes J. Portal Hypertension. Med Clin North Am 1989; 73: 931-53. exaggerate these circulatory abnormalities.2' 11 Garcia-Tsao G, Groszmann RJ, Fisher RL, Conn H Renal failure was common in the present series Atterbury E, Glickman M. Portal pressure, presence of

gastroesophageal varices and variceal bleeding. Hepatology http://gut.bmj.com/ ofpatients with acute liver failure. Some ofthese 1985; 5:419-24. patients had acute tubular necrosis, which may 12 Viola C, Vineta L, Bosch J, Rodes J. Exudative ascites in the course ofacute type B hepatitis. Hepatology 1983; 3: 1013-5. be related to shock, bacteraemia hypovolaemia 13 Sikuler E, Kravetz D, Groszmann RJ. Evolution of portal and drug nephrotoxicity. Some patients, how- hypertension and mechanisms involved in its maintenance in a rat model. AmJ Physiol 1985; 248: G618-25. ever, exhibited 'functional' renal failure, with 14 Anthony PP, Oshak KG, Nayac NC. The morphology of many similarities observed in cirrhosis. As in this cirrhosis. J Clin Pathol 1978; 31: 395-414. 15 Urbano A, Navasa M, Mont LI, Bosch J, Bruguera M, Rodes condition,23 functional renal failure in acute liver J. Colestasis intrahepatica e hipertensi6n portal en un failure has been related to an imbalance between paciente con amiloidosis primaria. Gastroenterol Hepatol on October 1, 2021 by guest. Protected copyright. 1986; 9:186-8. neurohumoral factors with renal vasoconstric- 16 Kasturi TE, Manchanda SC, Tandon RK, Rajani M, Bathia tive activity and renal vasodilators.24 In this ML. Haemodynamic studies in veno-occlusive disease ofthe liver. Br Heartj 1979; 41: 594-9. regard, increased plasma renin activity and 17 Groszmann RJ, Glickman M, Blei AT, Storer E, Conn HO. decreased renal prostaglandin production have Wedged and free hepatic venous pressure measured with a balloon catheter. Gastroenterology 1979; 76: 253-8. been reported in patients with acute liver 18 Gazzard BG, Portmann B, Murray-Lyon IM, Williams R. failure.24 As already mentioned, activation of Causes ofdeath in fulminant hepatic failure and relationship to quantitative histological assessment of parenchimal neurohumoral vasoactive factors is triggered by damage. QJrMed 1975; 176: 615-26. the extreme vasodilation frequently seen in these 19 Schrier RW, Arroyo V, Bernardi M, Epstein M, Henriksen JH, Rodes J. Peripheral arterial vasodilation hypothesis: a patients.'9 Therefore, as in cirrhotics with proposal for the initiation of renal sodium and water , patients with acute liver retention in cirrhosis. Hepatology 1988; 8: 1151-7. 20 Bihari DJ, Gimson AES, Williams R. Cardiovascular, pul- and renal failure exhibit renal vasoconstriction in monary and renal complications offulninant hepatic failure. the context of marked systemic vasodilation. Semin LiverDis 1986; 6: 119-28. 21 Trewby PN, Williams R. Pathophysiology of hypotension in This indicates that there should be some areas patients with fulminant hepatic failure. Gut 1977; 18: where vasodilation is greater than others. 1021-6. 22 Hortnagl H, Lochs H, Kleinberger G, Singer EA, Lenz K. Experimental data suggest that vasodilation pre- Substance P is markedly increased in plasma ofpatients with dominates in the splanchnic territory.25 It is thus hepatic coma. Lancet 1984; i: 480-3. 23 Ring-Larsen H, Henriksen JH. Pathogenesis of ascites forma- possible that if the splanchnic vasodilation could tion and hepatorenal syndrome: Humoral and hemodynamic be overcome, there could be an amelioration in factors. Semin LtverDis 1986; 6: 341-52. 24 Guarner F, Hughes RD, Gimson AES, Williams R. Renal the systemic and renal circulation.26 This function in fulminant hepatic failure: haemodynamics and actually has been reported in patients with renal prostaglandins. Gut 1987; 28: 1643-7. 25 Fernandez-Seara J, Prieto 1, Quiroga J, Zozaya JM, Cobos cirrhosis,27 Although there has no t been a MA, Rodriquetire JL et al. Systemic and regional hemo- clinical or experimental study exploring this dynamics in patients with liver cirrhosis and ascites with and without functional renal failure. Gastroenterology 1989; 97: possibility in acute liver failure. 1304-12. Finally, it should be emphasised that we had 26 Bosch J. Splanchnic vasodilation and renal vasoconstriction: A key to the hepatorenal syndrome? Hepatology 1990; 12: no complications in obtaining transjugular biop- 1445-7. sies in patients with acute liver failure despite the 27 Lenz K, Hortnagle H, Druml W, Grimm G, Laggner A, Schneoweisz B, et al. Beneficial effect of 8-ornithin vaso- severe inherent to this condition, pressin on renal;disfunction in decompensated cirrhosis. Gut even though these studies were frequently 1989; 30:90-6.