Liver and biliary tract

              Jetra i bilijarni trakt

                 ARCH GASTROENTEROHEPATOL 2002; 21 ( No 3 – 4 ):

 

 

 

Ventilatory – perfusion disorders in cirhosis of liver 

 

Ventilatorno – perfuzioni poremećaji u cirozi jetre

 

 

( accepted September 21st, 2002 )

 

 

Djordje Ćulafić, Mirjana Perišić, Predrag Rebić

 

Institute for Digestive Diseases, Clinic of Gastroenterology and        

Hepatology, Clinical Center of Serbia, Belgrade.

 

 

Address correspondence to: Dr Djordje Culafic

                                             49 Mirijevski venac St.

                                             YU-11160 Belgrade, Serbia  

                                             Yugoslavia

                                              Tel. + 381 11 342 8020

                                              E-mail:dculafic@EUnet.yu

................................................                           .....................................................

 Ventilatory-perusion disoder in cirrhosis    Gastroenterološka sekcija SLD-

                                                                            01734,2002.

 

 

 

ABSTRACT

 

Ventilatory-perfusion disorders in chronic liver diseases are multifactorial in origin, and may be caused by impaired pulmonary vasoconstrictive response, increased airway closing volume, ascites, pleural effusion, and hepatosplenomegaly.

We examined ventilatory-perfusion indices in 50 consecutive patients with liver cirrhosis, at the Institute for Digestive System Diseases and the Institute of Pulmonary Diseases, Clinical Center of Serbia. Pulmonary function tests were used to determine  transfer factor and transfer coefficient, as indicators of alveolo-capillary diffusion rate. Restrictive ventilatory impairment was determined on the basis of spirometric parameters: vital capacity, Tiffeneau index and total lung capacity. Obstructive ventilatory disorder was determined on the basis of Tiffeneau index.

In the group of 18 non-smokers with liver cirrhosis we diagnosed 10 subjects with lower transfer factor, reduced vital capacity, decreased forced expiratory volume during the first second, and forced expiratory flow at 25% forced vital capacity. Restrictive ventilatory disorders were diagnosed in 12 (38.7%) patients with ascites. In comparison with group without ascites (c2 test), significantly higher incidence of restrictive ventilatory disorders was found in patients with ascites (p=0.03). However, ascites had no significant effect on the level of partial oxygen pressure (in comparison with the group with ascites and hypoxemia, and the group without ascites but with hypoxemia) (p=0.79, c2 test). Hypoxemia was found in 18 (36%) patients. Mean value of partial oxygen pressure was 9.68 kPa (SD=1.10). This result indicates that ventilatory-perfusion ratio causes moderate oxygenation impairment.

 

Key Words: ventilatory-perfusion disorder, liver cirrhosis.

 

 

 

SAŽETAK

 

Ventilaciono-perfuzioni poremećaji u oboljenjima jetre su multifaktorskog nastanka i prouzrokovani su: oslabljenim plućnim vazokonstriktornim odgovorom, porastom volumena zatvaranja disajnih puteva, ascitesom, pleuralnim efuzijama i hepatosplenomegaliom. U Institutu za bolesti digestivnog sistema i Institutu za plućne bolest, Kliničkog centra Srbije u Beogradu, ispitalivali smo ventilaciono-perfuzione poremećaje kod 50 pacijenata sa cirozom jetre. Funkcionim plućnim testovima odredjivali smo transfer faktor i koeficijent   transfera kao pokazatelje stanja alveolno- kapilarne difuzije. Restriktivne ventilacione smetnje definisali smo na osnovu spirometrijskih parametara:vitalnog kapaciteta, Tiffeneau indeksa i totalnog plućnog kapaciteta. Opstruktivne ventilacione smetnje određivali smo na osnovu Tiffeneau indeksa. U našem ispitivanju otkrili smo da 10 osoba u grupi od 18 nepušača sa cirozom jetre, koji su imali smanjene vrednosti transfer-faktora imaju i značajno manje vrednosti vitalnog kapaciteta, forsiranog ekspiratornog volumena u prvoj sekundi i posebno forsiranog ekspirijumskog protoka pri 25% forsiranog vitalnog kapaciteta, u odnosu na preostalih 8 ispitanika s normalnom difuzijom u plućima. Ovakav nalaz bi mogao da se objasni ranim, supstancijalnim promenama ili edemom u plućnom parenhimu. Restriktivne ventilacione smetnje smo dijagnostikovali kod 12 (38,7%) pacijenata sa ascitesom. U poređenju sa  grupom bez ascitesa (c2 test), nađena je statistički značajno veća incidenca restriktivnih ventilacionih smetnji kod pacijenata sa ascitesom (p = 0,03). Međutim, ascites nije značajno uticao na vrednost parcijalnog pritiska kiseonika (poređenje grupe sa ascitesom i hipoksemijom u odnosu na grupu bez ascitesa sa hipoksemijom) (p = 0,79, c2 test). Hipoksemiju je imalo 18 (36%) pacijenata. Srednja vrednost parcijalnog pritiska kiseonika je iznosila 9,68 kPa (SD = 1,10). Rezultat ukazuje da poremećaj ventilaciono perfuzionog odnosa obično uzrokuje blaže oksigenacione poremećaje.

 

Ključne reči:ventilaciono-perfuzioni poremećaj, ciroza jetre.

 

In patients with chronic liver diseases pulmonary functions are frequently impaired.

 

Approximatelly 30% of cases with decompensated liver cirrhosis are hypoxemic (1).

 

The genesis of arterial hypoxemia is multifactorial in origin and may be caused by

 

lower affinity of hemoglobin for oxygen, pleural effusion (hepatic hydrothorax),

 

pulmonary hypertension, parenchymal pulmonary disease, bronchitis or emphysema

 

induced by smoking, portopulmonary shunts, decreased alveolo-capillary difusion rate, 

 

ventilatory-perfusion mismatch and/or intrapulmonary arterio-venous shunts (2,3).

 

The most common functional respiratory disorder in patients with liver cirrhosis is ventilation-perfusion mismatch as well as the reduction of pulmonary diffusion phase. On the contrary, measurable disorders of air volume and air flow through respiratory pathways are rare (4).

In chronic liver diseases pulmonary ventilatory-perfusion () disorder is of  multifactorial origin, and may be caused by  impaired pulmonary vasoconstrictive response, increased closing volume, ascites, pleural effusion, and hepatosplenomegaly (4). Clinically, disorders of lung function caused by mismatch commonly manifests with moderate hypoxemia. disorder is very frequent extrahepatic manifestation of chronic liver disease (5).

 

MATERIAL AND METHODS

 

This study enrolled 50 consecutive patients with liver cirrhosis diagnosed at the Institute for Digestive System Diseases and the Institute of Pulmonary Diseases, Clinical Center of Serbia. Analyses included past medical history, physical examination, liver function tests, virological survey, laboratory tests for specific metabolic liver diseases (ceroloplasmin, alfa-1 antitripsin, transferin saturation index etc) and percutaneous liver biopsy. 

Two groups of pulmonary function tests were routinely made: blood gasses analysis and ventilation tests (spirometry, flow-volume curve and body plethysmography). Blood gas essay was carried out using Blood Gas Manager 1312. Expected values for partial oxygen pressure were calculated according to  Sorbini equation (Pa,O2=103.5 – 0.42 x years) (6).

To determine statistical, dynamic pulmonary volumes and capacities: vital capacity (VC), forced vital capacity (FVC) and forced expiratory volume during the first second (FEV1) spirometric studies were performed by open spirometric system with pneumotachograph (Pneumoscreen II spirometer). To measure total lung capacity (TLC), thoracic gas volume (TGV), residual volume (RV), and air volume resistance in the airways (RaW) body plethysmography with constant volume was used. This measurements were performed by Bodyscreen II.

Results of pulmonary function tests were used to determine transfer factor (TL,CO) and transfer coefficient (KCO= TL,CO divided by the effective alveolar volume), as indicators of alveolo-capillary diffusion state. Diffusion indices were assessed by CO measurement using the one-inspiration method, with Transferscreen II apparatus. Lower normal limit of transfer factor was determined by mathematics. The expected values of TL,CO were subtracted by SD of 1.64 (SD of 1.42 for males, SD of 1.17 for females). The expected values of TL,CO were calculated on the basis of valid standards (in males, expected TL,CO = 11.11 x height in meters – 0.066 x years – 6.03; in females, expected TL,CO = 8.18 x height in meters – 0.049 x years – 2.74) (7).

Restrictive ventilatory disoder was determined on the basis of spirometric parameters: VC, index 100 x FEV1/VC (Tiffeneau) and TLC. Lower normal limit was determined as the expected value – 1.64 SD (in males, SD for FVC = 0.61, SD for TLC = 0.70, and for Tiffeneau SD = 7.17; in females, SD for FVC = 0.43, SD for TLC = 0.60, and for Tiffeneau SD = 6.51) (8).

Obstructive ventilatory disorder was determined on the basis of the following ratio: 100 x FEV1/VC. Values below 75% were considered to be abnormal index of pulmonary ventilation. The function of minor airways was determined by MEF25 (forced expiratory flow at 25% FVC) and MEF50 (forced expiratory flow at 50% FVC).

RESULTS

 

In order to determine the effect of mismatch on partial oxygen pressure, patients without intrapulmonary shunt were analyzed. Seventeen (34%) out of 50 cases  were smokers. Hypoxemia was found in 18 (36%). Mean partial oxygen pressure was 9.68 kPa (range: 7.70 to 11.57) (SD=1.10). 

The obstructive ventilatory disorder with Tiffeneau index <75% was diagnosed  in 13 (26%) patients. We found lower transfer factor in 27 (54%), and lower transfer coefficient in 33 (66%) cases.

Ten out of 18 non-smokers with liver cirrhosis, who exibited lower transfer factor (mean value 70.9; SD=16.8), also had significantly lower VC (mean value 75.7; SD=16.1), FEV1 (mean value 78.3; SD=18.8) and especially MEF25 (mean value 65.5; SD=23.5). The remaining 8 subjects had normal pulmonary gass exchange.

In order to study the effect of ascites on partial oxygen pressure, patients without intrapulmonary shunt were analyzed. The diagnosis of ascites was made in 31 (62%) patients. In recumbent position, hypoxemia was found in 13 (65%) ascitic patients, with PaO2 mean value of 9.63 kPa (SD=1.01, min 8.23, max 11.06). In sitting position, hypoxemia was recorded in 10 (50%) patients with ascites, with PaO2 mean value 10.47 kPa (SD=1.18, min 9.30, max 12.60). Ascites had no significant effect on partial oxygen pressure (comparison of group with ascites and hypoxemia, and the group without ascites and with hypoxemia) (p=0.79, c2 test).

Restrictive ventilatory disorder was diagnosed in 12 (38.7%) patients with ascites. Mean value of restriction parameters was: VC 73.0 (SD=14.4), TLC 83.7 (SD=13.5), and Tiffeneau index 78.3 (SD=3.37). In comparison with non-ascitic group (c2 test), significantly higher incidence of restrictive ventilatory disorders demonstrated   patients with ascites (p=0.03).

Lower transfer factor with mean value 62.1 (SD=16.8) and diffusion disorder was recorded in 18 (58%) patients with ascites. In relation to non-ascitic group (TL,CO mean value 71.8; SD=13.7), no significant difference of incidence of diffusion disorders was found (p=0.59, t-test).

Hepatic hydrothorax was diagnosed in 5 (10%) cirrhotic patients. All manifested restrictive ventilation disorders and hypoxemia with mean PaO2 mean 9.36 kPa (SD=1.09, min7.50, max10.12).

DISCUSSION

 

In healthy persons, the strongest stimulus of pulmonary vasoconstriction is alveolar hypoxia (9). In liver cirrhosis vasoconstrictive response to hypoxia is impaired. Abnormal vasomotor activity and decreased vasomotor arteriolar tone cause dilatation of minor lung blood vessels. On the contrary to macroscopic arterio-venous shunts which are rarely seen in cirrhotics, pulmonary microcirculatory dilatation is common post-mortem finding (10).

In 1987, Rodriguez-Roisin et al. reported that reduction of pulmonary vascular reactivity is an important cause of hypoxemia and disorder. In their series of 15 patients with liver cirrhosis, pulmonary vascular reactivity was checked  by measuring resistance of pulmonary blood vessels when exposed to hypoxic and hyperoxic air mixture. In hypoxic condition no significant changes of distribution was noted;  expected increase of pulmonary vascular resistance was completely absent in 5 patients. Hyperoxia caused worsening of ratio and moderate drop of pulmonary vascular resistance (11). In 1990, Agusti et al. reported that beside airway diseases, vasoparalysis of pulmonary vasculature is responsible for alteration of ventilatory-perfusion ratio and accounts for paradoxical phenomenon: low pulmonary vascular resistance associated arterial hypoxemia (12).

In 1971, Ruff et al. noted for the first time an increase of closing volume in 8 out of 10 patients with liver cirrhosis using the xenon-133. During normal respiration, early airway closing (closing volume - CV) is caused by mechanical compression or interstitial lung oedema because of salt retention, hypoalbuminemia, increased capillary permeability, hormone-related water retention or reduced pulmonary lymphatic drainage. Developed microatelectases led to perfusion of non-ventilated lung regions (13).

 In patients with liver cirrhosis closing volume is frequently upregulated reflecting functional status of minor blood vessels. Closing capacity (RV+CV) has tendency to be higher than FRC in patients with liver cirrhosis and hypoxemia, suggesting premature airway closing and air trapping even during unforced breathing (14,15).

In 1980, Hara et al demonstrated lower MEF25 in 53 patients with liver cirrhosis without apparent clinical and radiological chest disease. In their study values of MEF50, VC, TLC, FRC, RV and FEV1 were within normal limits. Beside decreased MEF25, the  authors found increased closing volume (CV) indicating disordered function of minor airways. Cigarette smoking had no effect on their results (16). This observation was further verified by Furukawa and associates. Lower MEF25 was detected in the majority of his 105 patients in spite of the fact they had normal VC, FRC, RV, TLC, RV/TLC and FEV1 values (14). In 1989, Scemama-Clergeu et al. found normal values of FEV1/FVC with reduced FEF25-75 in the majority of 10 patients with liver cirrhosis (17). Marichal et al. recorded decreased flow via minor airways as only spirometric disorder in the group of 8 patients with liver cirrhosis (18).

In our study 10 out of 18 non-smokers cirrhotics with lower transfer factor demonstrated reduced VC, FEV1 and especially MEF25. This finding is explained by early, substantial changes or pulmonary parenchymal edema.

Profuse ascites and/or pleural effusion reduce lung expansion and cause restriction of pulmonary volume and capacity. All static pulmonary volumes are decreased: VC, TLC, FRC and RV. Further on, in chronic liver diseases reduction of total pulmonary capacity may be caused by inspiratory muscles weakness, due to malnutrition. Forced air flow is decreased in concordance with reduced volume. Maximal inspiratory pressure is commonly maintained. Ascites increases the diaphragmatic curvature and therefore enhances the ratio of length and contractility of this muscle (2).

We diagnosed restrictive ventilatory disorders in 12 (38.7%) patients with ascites. In comparison with non-ascitic group (c2 test), significantly higher incidence of restrictive ventilatory disorders was found in patients with ascites (p=0.03). The evacuation of ascites resulted in the improved total pulmonary capacity, functional residual capacity and airway flow. The best improvement was noted in the expiratory reserve volume (19).

In recumbent position, enlarged abdominal viscera pushes the lungs upwards, causing considerable restriction of ventilation and gas exchange. Flow reduction may be more pronounced in minor airways, leading to increase of closing volume. Reduction of FRC is particularly important. In low FRC and higher closing volume, the airway closing occurs even with calm breathing. Early airway closing makes large areas of the lower lung lobes excluded from the ventilation while continuing to be circulatory perfused what worsen ventilation-perfusion ratio. Perfusion of non-ventilated regions causes that large quantity of blood runs though lungs without being oxygenated (20).

In the upright position patients with massive peritoneal effusion usually have normal arterial blood gasses, while partial insufficiency of respiration is induced by  recumbent position change. Extremely large restrictions due to massive ascites may led to global respiratory insufficiency, which is mostly marked in supine position (21).

In order to evaluate the effect of ascites on partial oxygen pressure, patients without intrapulmonary shunts were analyzed. In recumbent position, hypoxemia was recorded in 13 (65%) patients, with mean PaO2 value 9.63 kPa (SD=1.01, min 8.23, max 11.06). In sedentary position, hypoxemia was documented in 10 (50%) ascitic patients  with mean PaO2 value 10.47 (SD=1.18, min 9.30, max 12.60). Ascites had no significant effect on partial oxygen pressure (in comparison of group with ascites and hypoxemia, and the group without ascites but with hypoxemia) (p=0.79, c2 test).

Significant effect of ascites on gass exchange has been challenged by other authors. Several group reported that there is no significant improvement of gas exchange after ample paracentesis ( assessed by PaO2 level, transfer factor and alveolo-arterial oxygen gradient) in spite of considerable improvement of ventilatory function and increased FEV1, FVC, TLC and FRC (22,23).

In 1997 Chang et al. published the results of their study in which they explored the effect of ascites on pulmonary function in two groups of patients. They compared therapeutic effects of paracentesis and diuretics. On the contrary to the patients managed by paracentesis, those treated with diuretics manifested significant improvement of gas exchange, along with PaO2 increase and large reduction of alveolo-arterial gradient. This suggests that beside mechanical effect of ascites, the interstitial pulmonary edema and fluid retention contribute additionally to gas exchange imapirement (24).Alveolo-capilary gas diffusion is impaired only in concordance with  reduced pulmonary volume (25).

We demonstrated lower transfer factor (TL,CO) and impaired diffusion in 18 (58%) patients with ascites. In relation to group without ascites, there was no significant difference of incidence of diffusion impairment (p=0.59, t-test).

Pulmonary transfer factor for CO is being improved parallely with the improvement of pulmonary volume (26).

Lower values of transfer coefficient were recorded after paracentesis in relation to values before paracentesis. This is to suggest that, after ascites evacuation, pulmonary volume is improving more than gas diffusion from alveoli to blood and, therefore, the quotient of transfer factor and effective alveolar volume is being reduced (27).

In the absence of primary lung or heart disease, patients with liver cirrhosis and pleural effusion are classified as having hepatic hydrothorax (28). The development of pleural exudatate is operated by multiple pathogenic mechanisms: hypoalbuminemia and subsequent fall of colloid osmotic pressure, anastomoses between portal and azygous systems with fluid transudation due to increased hydrostatic pressure in azygous vein, lymph transudation from ductus thoracicus, entry of fluid into pleural space via transdiaphragmatic lymphatic channels, and direct flow of peritoneal fluid through diaphragmatic openings (29).

There is some evidence that congenital diaphragmatic porus(i) and the increased intraabdominal pressure are the most important precipitating factors for the development of hepatic hydrothorax. Most commonly, these porus(i) are small diaphragmatic openings 0.03-0.12mm in diameter (30).

It is well known that pleural exudation may develop in the absence of ascites. One-way, peritoneo-pleural fluid flow occured due to valvular mechanism based on pressure difference. The fluid runs from the abdomen, where pressure is positive, to the chest, where pressure is negative accordingly to the respiratory cycle. Transdiaphragmal flow of peritoneal fluid into the pleural space is an alternative way of ascites elimination (31).

Pleural effusion in the absence of ascites represents major diagnostic problem. In 1985, Rubenstein and associates reported that intraperitoneal injection of 99m Tc-sulphur colloid may verify transdiaphragmatic passage of peritoneal fluids into pleural cavity. Instilled intraperitoneal 99mTc-sulphur colloid is not detected in pleural space in patients with effusions due to pulmonary or cardiac diseases (29).

Pleural exudation causes ventilatory restrictive disorder due to  lung compression. This is more pronounced in the upright position. Large exudates displace heart to the contralateral side thus jeopardizing heart performance and interfering with diaphragmatic function as well. The reduction of diffusion accompanies lung volume decrease what further on lowers transfer factor (31).

In our study, hepatic hydrothorax was diagnosed in 5 (10%) patients with liver cirrhosis. All patients had restrictive ventilatory disorders and hypoxemia with PaO2 mean value of 9.36 kPa.

In conclusion, we demonstrated that mismatch is result of complex of multifactorial pathogenetic mechanisms. The disorded ventilatory-perfusion ratio causes moderate oxygenation impairment. 


REFERENCES:

 

1.    Sherlock S. Diseases of the liver and biliary system. 8th ed. Oxford, Blackwell Scientific Publications 1989; 82-5.

2.    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.

3.    King P, Rumbaut R, Sanchez C. Pulmonary manifestations of chronic liver disease. Dig Dis 1996; 14: 73-82.

4.    Krowka MJ. Clinical management of hepatopulmonary syndrome. Seminars in liver disease 1993; 4:414-22.

5.    Melot C, Naeije R, Dechamps P, Hallemans R, Lejune P. Pulmonary and extrapulmonary contributors to hypoxaemia in liver cirrhosis. Am Rev Respir Dis 1989; 139: 632-40.

6.    Sorbini CA, Grassi V, Solinas E, Muiesan G. Arterial oxygen tension in relation to age  in healthy subjects. Respiration 1968; 25: 3-13.

7.    Cotes JE. Standardization of the measurement of transfer factor (diffusing capacity). The European Respiratory Journal 1993;6: 5- 40.

8.    Quanjer PH. Lung volumes and forced ventilatory flows. The European Respiratory Journal 1993; 6: 5- 40.

9.    Harris P, Heath D. The human pulmonary circulation. Churchill-Livingstone. Edinburg 1986; 545-3.

10.Daoud FS, Reeves JT, Schaefer JW. Failure of hypoxic pulmonary vasoconstriction in patients with liver cirrhosis. J Clin Invest 1972; 51: 1076-80.

11.Rodriguez-Roisin R, Roca J, Agusti AG, Mastai R, Wagner P, Bosch J. Gas exchange and pulmonary vascular reactivity in patients with liver cirrhosis. Am Rev Respir Dis 1987; 135: 1085-92.

12.Agusti AGN, Roca J, Bosch J, Rodriguez-Roisin R. The lung in patients with cirrhosis. J Hepatol 1990; 10: 251-257.

13.Ruff F, Hughes JMB, Stanley N, McCarthy D, Green R, Aronoff A. Regional lung function in patients with hepatic cirrhosis. J Clin Invest 1971; 50: 2403-13.

14.Furukawa T, Hara N, Yasumoto K, Inokuchi K. Arterial hypoxemia in patients with hepatic cirrhosis. Am J Med Sci 1984; 287:10-13.

15.Behera D, Singh M, Chawla Y, Dilawari JB. Pulmonary function abnormality in patients with portal hypertension with or without chronic liver disease. Indian J Chest Dis Allied Sci 1998; 40:33-39.

16.Hara N, Yoshida T, Furukawa T, Inokuchi K. Abnormalities in maximum flow volume curve and closing volume in patients with hepatic cirrhosis. Jpn J Surg 1980; 10: 265-269.

17.Scemama-Clergue J, Piquet J, Clerici C, Harf A, Dhumeaux D. Effects of propranolol on pulmonary  gas exchange in patients with cirrhosis. J Hepatol 1989; 9:234-239.

18.Marichal I, Dublet P, Medrano G, Hinestrosa H, Talmo C, Korchoff W, Alvarado R, Quiros E. Liver cirrhosis: pulmonary function. G E N 1991; 45:263-269.

19.Angueira CE,Kadakia SC.Effects of large - volumen paracentesis on pulmonary function in patients  with tense cirrhotic ascites. Hepatology 1994;20 : 825-828.

20.Zhu Z, Yuan Y, Li N, Ye X, He T. Pulmonary function abnormality in patients with portal hypertension with or without liver disease. Hua Hsi I Ko Ta Hsueh Hsueh Pao 1992; 23:343-346.

21.Gupta D, Lalrothuama S, Agrawal PN, Agrawal AN, Dhiman RK, Behra D, Chawla Y. Pulmonary function changes  after large volumen paracentesis.Trop Gastroentrol 2000;21:68-70.

22.Berkowitz KA, Butensky MS, Smith RL. Pulmonary function changes after large volume paracentesis. Am J Gastroenterol 1993; 88: 905-7.

23.Chao Y, Wang SS, Lee SD, Shiao GM, Chang HI, Chang SC. Effects of large volume paracentesis on pulmonary function in patients whit cirrhosis and tense ascites. J Hepatol 1994; 20: 101-5.

24.Chang SC, Chang HI, Juh F, Shiao GM, Wang SS, Lee SD. Therapeutic effects of diuretic and paracentesis on lung function in patients with non-alcoholic cirrhosis and tense ascites. J Hepatol 1997; 26: 833-8.

25.Stanley  NN,Woodgate DJ.Mottled chest radiograph and gas transfer defect in  chronic liver  disease.Thorax 1972;27:315-323.

26.Chao Y, Wang SS, Lee SD, Shiao GM, Chang HI, Chang SC. Effects of large volume paracentesis on pulmonary function in patients whit cirrhosis and tense ascites. J Hepatol 1994; 20: 101-5.

27.Chang SC,Chang HI,Chen FJ,Shiao GM,Wang  SS.Effects of ascites and body position on gas exchange in patients with cirrhosis. Proc Natl Sci Counc Repub China B  1995;19:143-150.      

28.Morrow CS, Kantor M, Armen RN. Hepatic hydrotorax. Ann Intern Med 1958; 49: 193-203.

29.Rubinstein D, Melnnes IE, Dudley FJ. Hepatic hydrotorax in the absence of clinical ascites: diagnosis and management. Gastroenterology 1985; 88: 188-91.

30.Alberts WM, Salem AJ, Solomon DA, Boyce G. Hepatic hydrothorax. Arch Int Med 1991; 151: 2383-8.

31.Chen A, Yat SH, Yen CT, Hung ST, Tung C. Diaphragmatic defect as a cause of massive hydrotorax in cirrhosis of the liver. J Clin Gastroenterol 1988; 10: 663-6.