Liver and biliary
tract
Jetra i bilijarni
trakt
ARCH
GASTROENTEROHEPATOL 2002; 21 ( No 1 – 2 ):
Hepatopulmonary
syndrome: Experience from one centre
Hepatopulmonalni sindrom: Iskustvo jednog
centra
( accepted
January 30th, 2002 )
Đorđe
Ćulafić
Clinic
for Gastroenterology and Hepatology, Institute of Digestive Diseases,
Clinical
Centre of Serbia, Belgrade.
Address correpondence to:
Dr Đorđe Ćulafić
49 Mirjevski Venac Str,
YU – 11160 Belgrade,
Serbia, Yugoslavia
…………….
………………………………
Hepatopulmonary
syndrome
Gastroenteroloska sekcija SLD-
01725, 2002.
ABSTRACT
Hepatopulmonary syndrome is defined by hypoxemia, an increase of alveolo-arterial gradient during inhalation of room air, and intrapulmonary vascular dilatation in the presence of hepatocelullar insufficiency. Pathoanatomic substrate of hepatopulmonary syndrome is intrapulmonary vascular dilatation due to dilated precapillaries, direct arterio-venous communications, and engorged pleural blood vessels. During 1998, 50 consecutive patients with liver cirrhosis enrolled this prospective study. After hepatological examinations, the morphological and pulmonary function tests were performed. Arterial blood gas analyses were performed in both supine and sitting positions while inhaling room air, and after 15 minutes of hyperoxic mixture exposure. Hypoxemia was documented in 25 (50%) patients. Further functional and morphological studies established the diagnosis of hepatopulmonary syndrome in 9 (18%) patients. Orthodeoxia was confirmed in all patients with hepatopulmonary syndrome and in 3 patients with subclinical intrapulmonary arterio-venous shunts. Radioisotope marker 99mTc-MAA skipped intrapulmonary circulation, being accumulated in the brain and kidneys in 9 patients with hepatopulmonary syndrome, and in 3 patients with subclinical intrapulmonary shunts, what correlated with functional findings. Precapillary dilatation, subclinical intrapulmonary shunts and true anatomic shunts indicate that changes of pulmonary blood vessels may evolve, defining the degree of hypoxemia. After liver transplantation, the reversion of vascular abnormalities and the improvement of oxygenation are possible in patients with subclinical intrapulmonary shunts and mild precapillary dilatation. Regression of vascular abnormalities is not expected in patients with true anatomic shunts and marked praecapillary dilatation documented by incomplete oxygenation response to 100% oxygen exposure.
Key words: hepatopulmonary syndrome.
SAŽETAK
Hepatopulmonalni sindrom je definisan hepatocelularnim ostecenjem jetre, hipoksemijom, povećanjem alveolno-arterijskog gradijenta tokom udisanja atmosferskog vazduha i intrapulmonalnom vaskularnom dilatacijom. Patoanatomski supstrat intrapulmonalne vaskularne dilatacije čine: dilatirani prekapilari, direktne arterio-venske komunikacije i dilatirani pleuralni krvni sudovi. U prospektivnoj studiji sprovedenoj tokom 1998. godine ispitali smo 50 pacijenta sa cirozom jetre. Nakon hepatoloških, izvršena su morfološka i funkcionalna ispitivanja pluća. Arterijske gasne analize smo uzimali u ležećem i sedećem položaju, pri udisanju sobnog vazduha i posle 15 minuta udisanja hiperoksične smese. Hipoksemija je imalo 25 (50%) pacijenata, medjutim, morfološkim i funkcionalnim ispitivanjem pluća hepatopulmonalni sindrom je dijagnostikovan kod 9 (18%). Ortodeoksija je potvrđena kod svih pacijenata sa hepatopulmonalnm sindromom, ali i kod 3 pacijanata sa sa subkliničkim intrapulmonalnim šantom. Perfuzionom scintigrafijom sa 99mTc-MAA, detaktovana je akumulacija radiofarmaka u mozgu i bubrezima kod svih pacijenata sa hepatopulmonalnm sindromom i subkliničkim intrapulmonalnim šantovima, što je korelisalo sa nalazom funkcionalnih testova. Prekapilarna dilatacija, subklinički intrapulmonalni šantovi i pravi anatomski šantovi ukazuju da promene u plućnim krvnim sudovima evoluiraju, što određuje stepen hipoksemije. Posle transplantacije jetre, reverzija vaskularnih abnormalnosti i poboljšanje oksigenacije je moguća kod pacijenata sa subkliničkim intrapulmonalnim šantovima i blagom prekapilarnom dilatacijom. Regresija vaskularnih poremećaja se ne očekuje kod pacijenata sa pravim anatomskim šantom i izraženom prekapilarnom dilatacijom, koje karakteriše nepotpun oksigenacioni odgovor na udisanje 100 % kiseonika.
Kljucne reci: hepatopulonalni sindrom.
INTRODUCTION
Kennedy and Knudson in 1977 introduced the term hepatopulmonary syndrome (HPS) describing an association between hepatocellular insufficiency with severe hypoxemia due to “dilatation of pulmonary blood vessels” (intrapulmonary vascular dilatation) and intrapulmonary arterio-venous shunts (1). By definition, this syndrome characterised advanced liver disease accompanied with hypoxemia, an increase of alveolo-arterial gradient during inhalation of room air, and intrapulmonary vascular dilatation (IPVD). Pathoanatomic substrate of intrapulmonary vascular dilatation consists of: dilated praecapillaries, direct arterio-venous communications, and dilated pleural blood vessels. Dilated pleural blood vessels resemble the cutaneous (spider-like) nevi, and therefore, they are denoted as pleural spiders. Hypoxemia with partial oxygen pressure ( PaO2 ) less than 8 kPa in patients with liver cirrhosis in the absence of other cardiorespiratory diseases suggests the presence of intrapulmonary vascular dilatation. Severe hypoxemia with PaO2 below 6.6 kPa indicates the diagnosis of HPS (2-4).
The most common pulmonary symptom in HPS is dyspnea, affecting almost 40% of patients with HPS. Another characteristic manifestation of intrapulmonary vascular dilatation in cirrhosis is platypnea, specifying that dyspnea is more pronounced in the upright and sitting position than in a recumbent one (5). Orthodeoxia is another significant clinical sing in HPS, representing the decrease of partial oxygen pressure more than 10% when patient is changing the position from a recumbent to a sitting one. It is presumed that the partial oxygen pressure decrease, occurring during the recumbent-to-sitting position change, is caused an increase of blood flow via dilated basal pulmonary blood vessels by gravitation. Platypnea and orthodeoxia conclusively indicate the presence of intrapulmonary vascular dilatation (6).
MATERIAL AND METHODS
During 1998, we prospectively studied 50 consecutive patients with liver cirrhosis, diagnosed at the Institute of Digestive Diseases and Institute of Pulmonary Diseases, Clinical Center of Serbia. The initially performed hepatological examinations were based on: medical history, physical examination, liver function tests, specific test for aetiological diagnosis such as serum ceruloplasmin, afla-1 antitiripsin , serum feritin, transferin saturation, viral markers, and percutaneous liver biopsy. The degree of liver insufficiency was made according to the Child classification.
After throughout hepatologic investigations, morphological and pulmonary function tests were performed. Two groups of pulmonary function tests were made: arterial blood gases measurement and ventilation tests (spirometry, flow-volume curve and body plethysmography). Gas analyses were carried out using Blood Gas Manager 1312 equipment, while predicted values for the partial oxygen pressure were calculated using Sorbini equation (PaO2 =103,5 – 0,42 x years) (7). Arterial blood gases analysis was performed in both supine and sitting positions while exposed to the room air and after 15 minutes of breathing of hyperoxic mixture using a direct method, which includes arterial blood sampling with a heparinized syringe. Spirometry was used for the determination of the statical, dynamic pulmonary volumes and capacities: vital capacity (VC), forced vital capacity (FVC) and forced expiratory volume during the first second (FEV1). These tests were performed using Pneumoscreen II spirometer. Body plethysmography was used to measure of total lung capacity (TLC), thoracic gas volume (TGV), residual volume (RV), and airflow resistance in the airways (RaW). The measurements were performed using Bodyscreen II.
In our study, the alveolo-arterial gradient of oxygen pressure (P(A-a)O2=PAO2-PaO2) was calculated in the way usually used in practice. The oxygen pressure in the alveolar gas is calculated on the basis of the measured atmospheric pressure and CO2 in the exhaled air, while PaO2 is measured directly in the arterial blood (8).
Given to the variations of respiration changes in cirrhotics while assuming different body positions, arterial blood changes occurred; the change of P(A-a),O2 for at least 0.66 kPa is a sign of significant alteration of alveolo- capilary exchange with the assumption of different body positions (9).
In our study standard chest radiography and perfusion pulmonary scintigraphy was routinely performed. With the patient in a supine position, albumin macroaggregates labelled with radioactive technetium (99mTc-MAA) were injected. Following the contrast medium injection, the patient was placed in a sitting position with his back turned toward the gamma camera (Siemens Gammasonics ZLC 3700) connected with the PC (Microdelta computer RT-11, Clinic Menu V 87A). The visualization was performed in six standard planes – gamma camera positions: anterior, posterior, left lateral, right lateral, left lateral oblique and right lateral oblique until 500 000 impulses per projection was accumulated in the PC matrix 64 x 64. The scanning of extra thoracic organs was used for the measurement of the total number of the brain and kidney impulses in order to confirm intrapulmonary arterio-venous shunting.
RESULTS
The patients were classified according to Child's score system. Class A included 16 (32%), class B 20 (40%) and class C 14 (28%) patients.
Hypoxemia was documented in 25 (50%) patients; however, based on functional and morphological investigations the diagnosis of HPS was made in 9 (18%) patients with liver cirrhosis. The majority of the HPS patients (6) belonged to the Child C group, while 3 patients were categorised as Child B.
Hypoxemia caused by ventilation-perfusion mismatching (Va/Q) in supine position was evidenced in 18 (36%) patients. Mean PaO2 was 9.68 kPa (MD=1.10), while hypoxemia in the sitting position was evidenced in the additional 12 (24%) patients due to ascites. The mean PaO2 value was 10.37 kPa (MD=1.09). No statistically significant difference between a supine and sitting position was found (p=0.102, t-test). Hypoxemia caused by HPS detectable in a supine position was evidenced in 7 (14%) patients. The mean PaO2 value was 8.72 kPa (MD=1.02). Hypoxemia caused by HPS detectable in a sitting position was evidenced in 9 (18%). The mean PaO2 value was 7.41 kPa (MD = 1.81).
When compared to
the HPS-free group, the patients with HPS had more severe hypoxemia in a
sitting position. When measured in a sitting position, the difference in PaO2
values in patients with and without HPS was highly significant
(p=0.001, t- test ).
Orthodeoxia was recorded in all patients with HPS, and in 3 patients with subclinical intrapulmonary arterio-venous shunts. When exposed to the room air, the average PaO2 value in a supine position was 10.49 kPa (MD=2.14), while in a sitting position this was 8.34 kPa (MD=2.31). When exposed to 100% oxygen, in supine position the average PaO2 value was 32.64 kPa (MD=13.64), while in the sitting position this was 19.19 kPa (MD=9.94).
The perfusion lung scintigraphy in normal conditions, after intravenous injection 99mTc-MAA, usually demonstrated that more than 95 % of injected material is trapped within pulmonary capillaries, whose diameter is 7 mm. When intrapulmonary arterio-venous shunts are present,
Radionucleide labelled particles of 20-100 mm in diameter pass into a systemic circulation and are detected in the brain and the kidneys. In our study, the radioisotope marker 99mTc-MAA skipped intrapulmonary circulation, being accumulated in the brain and kidneys in 9 patients with HPS and in 3 patients with subclinical intrapulmonary arterio-venous shunts, what absolutely correlated with functional findings.
DISCUSSION
Disorders of oxygenation with consequent arterial hypoxemia is frequently seen in patients with impaired liver function. Between 12% and 30% patients with cirrhosis demonstrate moderate hypoxemia, small percentage of them have gross hypoxia (10). Approximately 50% of patients planned for liver transplantation have some degree of arterial oxygenation impairment, while 13% to 47% of these cases may have eventually HPS (11). There is evidence that in advanced liver diseases ventilation-perfusion (Va/Q) defect is an important cause of hypoxemia (12,13).
When liver disease, s associated mild hypoxemia, HPS (intrapulmonary arterio-venous shunting) is absent, in such cases Va/Q mismatch is primarily responsible for impaired oxygenation. The genesis of Va/Q disorder in liver disease is multifactorial and may be caused by: ascites, pleural effusion, reduced pulmonary vasoconstrictive response, and an increase of the closing volume. During normal breathing. the premature airway closure is result of the mechanical airway compression or interstitial pulmonary oedema caused by salt retention, low serum albumin, capillary permeability increase, hormone dependent water retention, and reduced pulmonary lymph drainage. Microatelectases may lead to perfusion of the unventilated pulmonary area (14). In patients with liver cirrhosis and prominent hypoxemia, intrapulmonary right-to-left shunts represent the major pathogenetic mechanism in the development of severe respiratory disorders, in spite of accompanied Va/Q disorder (15-17).
In our series, 18 (36%) patients had hypoxemia, but without intrapulmonary shunts. However, all patients had only mild oxygenation disorders, with the average PaO2 value of 9.68 kPa. Hypoxemia caused by intrapulmonary arterio-venous shunts was documented in 9 (18 %) patients with the average PaO2 value of 7.41 kPa. Severe hypoxemia was observed in patients with shunting in a sitting position, and this was statistically significant when compared to the group without shunts.
An important functional index of intrapulmonary vascular shunting is the fall of partial oxygen pressure when body position is changed to supine one. In liver cirrhosis, deoxygenation phenomenon associated with body position change, called orthodeoxia, is highly sensitive and specific (18). In this study we confirmed the sensitivity of orthodeoxia phenomenon, which was diagnosed in all patients with HPS, and in 3 patients with subclinical intrapulmonary arterio-venous shunts.
The determination of alveolo-arterial gradient of oxygen pressure, P(A-a)O2 = PAO2-PaO2, is the simplest method to assess the ratio of ventilation and perfusion of the lungs. It is postulated that P(A-a)O2 ³ 2 kPa with or without hypoxemia indicates the intrapulmonary vascular dilatation (19). Frequently, the hyperventilation of patients with liver cirrhosis may increase P(A-a)O2 values even in the case of normoxemia. Therefore, alveolo-arterial gradient is sensitive parameter of pulmonary gas exchange disorder. High diagnostic significance of these tests has been corroborated by finding of the unchanged basal partial oxygen pressure in the presence of subclinical intrapulmonary shunts, in spite of the increase of alveolo-arterial gradient of oxygen pressure (20).
Subclinical intrapulmonary shunts, as specific entity, and present the transitory phase in the development of HPS. In spite of intrapulmonary vascular dilatation and the increased alveolo-arterial gradient, the partial arterial oxygen pressure may stay unchanged for a long period of time. This is caused by alveolar hyperventilation, hyperdynamic circulation, and an increase of cardiac output characteristic for cirrhotics (21,22).
Subclinical lung vasculature vasodilatation, with concomitant normoxemia may be found in the incipient liver cirrhosis. With the progression of pulmonary changes, the hypoxemia becomes increasingly manifested. During the course of the disease, the oxygenation may be intensified even without the aggravation of liver function (23).
On the basis of response to exposssure to 100%
oxygen, two types of
vascular abnormalities in patients with hepatocellular insufficiency have been
described. HPS type-1 is characterized by diffuse vascular dilatation at
praecapillary level, while HPS type-2 is defined by true anatomic pulmonary
arterio-venous shunting. The application of 100% oxygen leads to significant increase of
partial oxygen pressure in the HPS of type-1, while the effect is minimal in
the type-2 (4).
In cirrhosis various types of pulmonary vascular abnormalities have been
described by angiography: minimal spider-like dilatations, diffuse spongiotic
dilatations, and discrete arterio-venous communications. It has been observed
that the patients with significant hypoxemia and excellent response to
inhalation of 100% oxygen (PaO2 > 53.3 kPa) usually have normal
pulmonary angiogram, because their vascular abnormalities is microscopic. The
patients with severe hypoxemia (PaO2 < 8 kPa) and poor response
to 100% oxygen exposure, usually have fixed arterio-venous communications or
advanced diffuse spongiotic dilatation (24).
In order to confirm the diagnosis of pulmonary vascular changes in liver
cirrhosis the visualisation of the intrapulmonary vascular dilatation is
necessary. In his series of 135 paediatric cases with chronic liver diseases of
various aetiology and candidates for liver transplantation Grimmon recorded
significant hypoxemia in 26 patients. In his study perfusion pulmonary
scintigraphy was sensitive and specific method to detect intrapulmonary shunts,
particularly when hypoxemia was insignificant (25). Great diagnostic
significance of perfusion pulmonary scintigraphy using 99mTc-MAA was
also confirmed by the fact that when subclinical intrapulmonary arterio-venous
shunts were present, partial oxygen pressure remained unchanged in spite of the
increase of alveolar-arterial oxygen gradient (20).
In our study, perfusion pulmonary scintigraphy enabled us to diagnose
pulmonary arterio-venous shunts in 9 (18%) patients with HPS and in 3 patients
with subclinical intrapulmonary arterio-venous shunts. This method
significantly correlated with pulmonary function tests, primarily based on the
phenomenon of orthodeoxia.
In 1998, Abrams et al. reported that to diagnose intrapulmonary arterio-venous shunts, the contrast two-dimensional echocardiography was far more sensitive than perfusion scintigraphy with 99mTc-MAA (26). In 1999, Aller et al tested the specificity of transesophageal contrast echocardiography in diagnosing intrapulmonary vascular abnormalities in 71 patients with liver cirrhosis. They found that transesophageal contrast echocardiography was the most sensitive diagnostic method, which, as pulmonary angiography, should be included in the liver pretransplant assessment programme (27).
In the pretransplantation assessment, it is important to diagnose and differentiate the vascular pulmonary abnormalities including search for probable intracardiac shunt, pulmonary embolism, and pulmonary hypertension. The patients with PaO2 < 9.3 kPa require a special preoperative staging, in order to assess whether adequate oxygenation may be provided during their operative and postoperative periods (28). In 1994, Hobeika et al. concluded that partial oxygen pressure below 8 kPa makes liver transplantation hazardous. In their series of 9 transplanted patients with HPS, there was 4 death. All had PaO2 < 8 kPa, while among 5 cases with PaO2 > 8 kPa, only 1 patient died (29).
Praecapillary dilatation, subclinical intrapulmonary shunts evolving to the true anatomic shunts indicate that changes of pulmonary blood vessels may undergo evolution, thus worsening the degree of hypoxemia. At the onset of the disease, mild hypoxemia correlates with the decrease of Va/Q ratio, while severe hypoxemia is the sequel of marked precapillary dilatation and anatomic intrapulmonary microvascular right-to-left shunts. The progression of vascular pulmonary dilatation leads to deterioration of arterial oxygenation, indicating fairly poor transplant prognosis (4,30). Therefore HPS with good oxygenation response to inhalation of 100% oxygen is a new indication for liver transplantation.
The reversion of vascular abnormalities and the improvement of oxygenation after liver transplantation is possible in patients with subclinical intrapulmonary shunts and mild praecapillary dilatation characterised by complete oxygenation response to 100% oxygen as well as angiographic findings of minimal spider-like dilatation. Regression of vascular abnormalities does not occur in patients with true anatomic shunts and marked. praecapillary dilatation characterised by incomplete oxygenation response to 100% oxygen and angiographic finding of diffuse spongiotic dilatation and/or discrete arterio-venous communications (31).
1. Kennedy TC, Knudson RJ. Exercise-aggravated hypoxaemia and orthodeoxia in cirrhosis. Chest 1977; 72: 305-9.
2. Krowka MJ, Cortese DA. Pulmonary aspects of chronic liver disease and liver transplantation. Mayo Clin Proc 1985; 60: 407-18.
3. Rodriguez-Roisin R. The hepatopulmonary syndrome: new name, old complexities. Thorax 1992; 47: 897-902.
4. Krowka MJ, Cortese DA. Hepatopulmonary syndrome: Current concepts in diagnostic and therapeutic considerations. Chest 1994; 105: 1528-37.
5. Robin ED, Laman D, Horn BR, Theodore J. Platypnea related to ortodeoxia caused by true vascular lung shunts. N Engl J Med 1976; 294: 941-3.
6. Krowka MJ, Cortese DA. Severe hypoxaemia associated with liver disease: Mayo Clinic experience and the experimental use of almitrine bismesylate. Mayo Clinic Proc 1987; 62: 164-73.
7. Sorbini CA, Grassi V, Solinas E, Muiesan G. Arterial oxygen tension in relation to age in healthy subjects. Respiration 1968; 25: 3-13.
8. Rebic P. Merenje odnosa ventilacije i perfuzije u plucima. U: Sekulic S, ed. Plucne bolesti. Elit Medica. Beograd 2000; 79-81.
9. Hourani JM, Bellamy PE, Tashkin DP, Batra P, Simmons MS. Pulmonary dysfunction in advanced liver disease: frequent occurrence of an abnormal diffusing capacity. Am J Med 1991; 90 : 693-700.
10. Sherlock S. Vasodilatation associated with hepatocellular disease: relation to functional organ failure. Gut 1990; 31: 365- 71.
11. Hopkins WE, Waggoner AD, Barazilai B. Frequency and significance of intrapulmonary right-to-left shunting in end-stage hepatic disease. Am J Cardiol 1992; 70: 516-9.
12. Melot C, Naeije R, Dechamps P, Hallemans R, Lejeune P. Pulmonary and extrapulmonary contributors to hypoxemia in liver cirrhosis. Am Rev Respir Dis 1989; 139: 632-40.
13. Agusti AG, Roca J, Bosch J, Rodriguez-Roisin R. The lung in patients with cirrhosis. J Hepatol 1990; 10: 251-7.
14. Ruff F, Hughes JM, Stanley N, et al. Regional lung function in patients with hepatic cirrhosis. J Clin Invest 1971 ; 50: 2403-13.
15. Edell ES, Cortese DA, Krowka MJ, Rehder K. Severe hypoxemia and liver disease. Am Rev Respir Dis 1989 ; 140: 1631-5.
16. Andrivet P, Cadranel J, Housset B, Herigault R, Harf A, Adnot S. Mechanisms of impaired arterial oxygenation in patients with liver cirrhosis and severe respiratory insufficiency. Effects of indomethacin. Chest 1993; 103 : 500-7.
17. Robert V, Chabot F, Vial B, Guito P, Poussel JF, Polu JM. Hepatopulmonary syndrome: physiopathology of impaired gas exchange. Rev Mal Respir 1999; 16:769- 79.
18. Thorens JB, Junod AF. Hypoxemia and liver cirrhosis: a new argument in favor of a "diffusion-perfusion" defect. Eur Respir J 1992 ; 5 : 754-6.
19. Rodriguez-Roisin R. The hepatopulmonary syndrome: new name, old complexities. Thorax 1992; 47: 897-902.
20. Lange PA, Stoller JK. The hepatopulmonary syndrome. Ann Int Med 1995 ; 22 : 521-9.
21. Agusti AG, Roca J, Rodriguez-Roisin R. Mechanisms of gas exchange impairment in patients with liver cirrhosis. Clin Chest Med 1996; 7: 49-66.
22. Miki K, Shinohara T, Ogushi F, Sone S, Yamada H, Oishi Y, Wakatsuki T, Ito S, Yogita S, Tashiro S. Hepatopulmonary syndrome-discussion of cardiopulmonary parameters. J Med Invest 2000; 47:164-9.
23. Mimidis KP, Karatza C, Spiropoulos KV, Toulgaridis T, Charokopos NA, Thomopoulos KC, Margaritis VG. Prevalence of intrapulmonary vascular dilatations in normoxemic patients with early liver cirrhosis. Scand J Gastroentrol 1998; 33: 988-92.
24. King P, Rumbaut R, Sanchez C. Pulmonary manifestations of chronic liver disease. Dig Dis 1996; 14: 73-82.
25. Grimon G, Andre L, Bernard O, Raffestin B, Desgrez A. Early radionuclide detection of intrapulmonary shunts in children with liver disease. J Nucl Med 1994 ; 35: 1328-32.
26. Abrams GA, Nanda NC, Duvovsky EV, Krowka MJ, Fallon MB. Use of macro aggregated albumin lung perfusion scans to diagnose hepatopulmonary syndrome: a new approach. Gastroenterology 1998; 114: 305-10.
27. Aller R, Moya J, Moreira V, Boixeda D, Picher J, Garcia- Rull S, de Luis D. Etiology and frequency of gas exchange abnormalities in cirrhosis. Rev Esp Enferm Dig 1999; 91: 568-74.
28. Schwarzenberg SJ, Freese DK, Regelmann WE, Gores PF, Boudreaz RJ, Payne WD. Resolution of severe intrapulmonary shunting after liver transplantation. Chest 1993; 103: 1271-3.
29. Hobeika J, Houssin D, Bernard O, Devictor D, Grimon G, Chapuis Y. Orthotopic liver transplantation in children with chronic liver disease and severe hypoxaemia. Transplantation 1994; 57: 224-8.
30. Van Obbergh L, Carlier M, Clety SC, Sokal E, Rennotte MT, Veycxemans F. Liver transplantation and pulmonary gas exchanges in hypoxemic children. Am Rev Respir Dis 1993; 148: 1408-10.
31. Glattki GP, Maurer C, Satake N, Lange R, Metz K, Wettstein M, Costable U. Hepatopulmonary syndrome. Med Klin 1999 ; 94: 505 - 12.