A 64-year-old woman had a long history of alcoholic cirrhosis and hypoxemia. She was treated with oxgyen. The terminal event was acute massive intestinal hemorrhage. At autopsy, the lungs appeared normal. One was perfused via the pulmonary artery with a barium-gelatin mixture.
Figure 1. Chest Radiograph (from another patient)
Frontal chest radiograph shows minimal linear and reticular abnormalities that are difficult to appreciate. This radiograph was prospectively interpreted as negative for acute cardiopulmonary disease.
Figure 2. Section of non-perfused lung--Elastic van Gieson stain
Compare the size and number of small vessels with the H&E-stained section of normal lung, figure 3, and the high-power view of this lung in figure 4.
Find the small muscular artery.
Find the small vein. What abnormality does it show?
Figure 3. Normal Lung at Same Magnification as Figure 2--H&E Stain
Figure 4. High-power View of Lung in Figure 2
The capillary diameters are much greater than those in figure 3, providing space for several RBCs rather than just one or two.
What is the differential diagnosis?
Scroll down after answering the questions.
Hepatopulmonary Syndrome (HPS)
HPS may be defined as hepatic dysfunction (usually cirrhosis) with hypoxemia (PaO2< 70 mmHg) or alveolar-arterial gradient (AaPO2) >20 mmHg and intrapulmonary vascular dilations (IPVD) demonstrated by positive contrast-enhanced transthoracic echocardiography (CEE) (bubble study) or the 99mtechnetium-labeled macroaggregated albumin scan (99mTcMAA) (>6% radioactivity detected over the brain). The syndrome has also been described in patients with acute and chronic hepatitis , extrahepatic portal hypertension , including the Budd-Chiari syndrome , or fulminant hepatic failure . In one study the syndrome had an incidence between 13-15% in endstage liver disease . In another study the incidence of HPS was 11% in cirrhotics, but only 2% in patients with prehepatic portal hypertension . Mortality of 50% at 41 mo after diagnosis has been described . Rarely, patients develop pulmonary hypertension as a sequel.
Most patients present with liver disease (82%) and the rest present with dyspnea with or without platypnea (dyspnea worse on standing). Dyspnea occurs an average of 4.8 y before HPS is diagnosed. Pulmonary disease can worsen independent of the liver disease, and death is often due to a non-pulmonary cause. Cyanosis, clubbing, and cutaneous spider nevi are frequent .
The hypoxemia of HPS is related, in part, to a diffusion-perfusion abnormality (correctable by breathing 100% oxygen) in enlarged gas-exchanging vessels, and to true arterio-venous (A-V) shunts, parenchymal or pleural, in non-gas-exchanging regions (not correctable by breathing 100% oxygen). The hyperdynamic circulation (increased cardiac index and low systemic pressure) contributes to the diffusion-perfusion abnormality .
The lungs of patients with HPS have a normal appearance on gross and routine histologic examination. The lesions became apparent when postmortem lungs were perfused with radiopaque media that distended vessels. Radiographs of these perfused lungs showed multiple, small, parenchymal A-V fistulas . These would be expected to act as true shunts, as the vessels were precapillary in size. In two studies, including 13 patients with cirrhosis (not defined as having HPS) and 12 patients with fulminant liver failure and hypoxemia, injection of the pulmonary artery with micropaque-gelatin showed two other types of lesions. In the patients with cirrhosis, arteriolar vessels were found to be dilated compared to normal (see images above) , and in fulminant liver failure, both arterioles and venules, as well as alveolar wall vessels were dilated, and arterioles had thinned walls . The lung architecture was otherwise normal, and the number of vessels was not increased, suggesting dilation of preexisting vessels and not proliferation . All types of involved vessels showed some patchy intimal fibrosis. In addition, prominent collections of dilated pleural vessels, termed spider nevi, were filled with the perfusate in 14 of the 25 cases [3,6]. The "spiders" predominated in the lower lobes. Subsequently, the term type I lesion was applied to the dilated parenchymal arterioles/capillaries/venules and the term type II lesion to the A-V shunts in either the parenchyma or pleura [4,7]. The type I lesions are thought to account for the diffusion-perfusion abnormality, correctable by breathing 100% oxygen, and the type II lesions represent true shunts.
Blood gases: Hypoxemia is indicated by a PaO2<70 mmHg. To demonstrate orthodeoxia, arterial blood gas analysis is performed both supine and standing. A PaO2 more than 3 mmHg lower when standing compared to that found supine indicates orthodeoxia . Alternatively, hypoxemia can be demonstrated by an AaPO2 gradient greater than 20 mmHg . Orthodeoxia and platypnea are explained by gravitational pooling of blood in the basal vascular dilations, which increases the shunt . If PaO2 on breathing 100% oxygen is less than 300 mmHg, it signifies a true shunt (cardiac or discrete A-V shunts (type II lesions)) in patients without underlying pulmonary disease. Angiography can be used to demonstrate the presence of focal A-V shunts, which may be treated by embolization. If the PaO2 is greater than 300 mmHg on 100% oxygen (type I lesions), true shunts are probably absent .
Chest radiographic findings: Chest radiographic findings in patients with HPS usually consist of bilateral, basilar predominant, reticular or nodular opacities, usually with normal lung volumes. The chest radiographic findings of HPS can be subtle, and chest radiographs in patients with HPS are often interpreted prospectively as normal.
Thoracic CT in patients with HPS may show peripherally enlarged pulmonary arteries. In normal patients, pulmonary arteries are too small to be discretely visualized at or very near the pleural surface, but in patients with HPS, intrapulmonary vascular dilations are often easily identified in the immediate subpleural regions of lung. These dilated vessels account for the opacities seen on chest radiography. True AVMs are not usually seen on thoracic CT in patients with HPS.
While the peripheral pulmonary arteries are commonly dilated in patients with HPS, even compared to patients with normoxemic cirrhosis, the central pulmonary arteries are usually normal in caliber . The vascular dilation that occurs in patients with HPS and that may be seen on thoracic CT may become appreciable at the distal segmental arterial level, where there is enough spatial separation to compare the arterial and bronchial sizes. In patients with cirrhosis and hypoxemia, the diameter of the distal segmental arteries (normally the same size as the adjacent airway) may be as much as twice that of the accompanying bronchus when vessels are measured approximately 2 cm from the pleural surface . Contrast-enhanced CT and MRI are useful to exclude other types of lung disease.
Nuclear medicine study: The pulmonary vascular dilations may be quantified by a 99mTcMAA study, which gives the fraction of labeled particles that are not trapped in the lung. While characteristic of right-to-left shunts, positive studies also occur with cardiac or extracardiac shunts.
Echocardiography: Intracardiac and intrapulmonary shunts may be distinguished by contrast-enhanced echocardiography using agitated saline. The immediate appearance of echo contrast material in the left heart after appearance in the right heart suggests an intracardiac shunt, whereas delayed appearance of the material in the left heart after visualization in the right heart (typically 3-6 cardiac cycles) suggests an intrapulmonary shunt. This test is very sensitive, being positive in a large number of cirrhotics without the complete syndrome .
Pulmonary angiography performed in patients with HPS may show dilation of peripheral pulmonary arteries, occasionally accompanied by early venous filling. As with thoracic CT, true arteriovenous malformations are not usually seen. Angiography is usually performed in patients with marked hypoxemia unresponsive to breathing 100% O2 (<300 mmHg maximum PaO2) . Large A-V shunts found in this way can be treated by embolization.
Pharmacologic therapy based on hypotheses about the molecular abnormality is still elusive. The main therapy other than supplemental oxygen is orthotopic liver transplant (OLT), which can reverse IPVD and improve oxygenation, although sometimes slowly. In a small study of OLT candidates who were examined for HPS, presence of cirrhosis, IPVD by CEE, and an AaPO2 gradient >20 mmHg were found to give the best delineation of patients with HPS . Using the same criteria for diagnosis of HPS, Krowka, et al. found that 8/40 candidates for OLT with HPS were denied transplant, but reasons were nonpulmonary in all. (Nevertheless the PaO2 was 47 mmHg in those denied transplant vs 52 mmHg in the others.) Hospital mortality was high in HPS patients receiving transplants (5/32 (16%)). In those undergoing transplant, non-survivors in hospital had a lower pretransplant PaO2 (37 mmHg) than did survivors (55 mmHg) . Survival with benefit or cure of HPS was 71% at 1 year among 24 patients undergoing OLT. Predictors of death during this time were PaO2 <50 mmHg and 99mMAA shunt fraction >/= to 20% . Thus, OLT seems to be indicated in those with progressive, but not severe, hypoxemia .
Currently, increased levels of nitric oxide (NO) are believed to be responsible for the pulmonary vasodilation of HPS. It has been shown that patients with cirrhosis exhale increased amounts of NO compared to control subjects. After OLT, the concentration of exhaled NO in these patients fell in proportion to the AaPO2 gradient, especially in the 5 patients with HPS . An experimental model of HPS in the rat has given some clues about the development of the pulmonary vascular dilations. Ligation of the common bile duct (CBD) produces biliary cirrhosis that is accompanied by pulmonary abnormalities seen in human HPS. The ligation is followed in 2 weeks by increased hepatic production, and increased plasma levels, of endothelin-1. At the same time, pulmonary vascular endothelial endothelin-B (ETB) receptor protein and endothelial nitric oxide synthase (NOS) levels are increased and can be noted immunohistochemically in the pulmonary microvascular endothelium. Stimulation of the ETB receptor by endothelin-1 increases NO production. The AaPO2 gradient increases and an intrapulmonary shunt develops. Increased numbers of intravascular mononuclear cells producing inducible nitric oxide synthase accompany these changes . These macrophages may develop as a result of Gram-negative translocation from the gut and can be decreased by treatment with antibiotics .
In the CBD-ligation model, HPS could be diminished by continuous postoperative endothelial ETB receptor blockade, which partly prevented the increase in AaPO2 gradient, intrapulmonary shunt, and the increase in endothelial NOS and ETB receptor protein levels . TNF-alpha inhibition by continuous postoperative pentoxifylline also prevented the development of HPS, in part by decreasing the number of intravascular macrophages producing NO . Comparison of the findings in CBD-ligation animals with animals undergoing partial portal vein ligation or toxin-induced cirrhosis, which are not accompanied by HPS, suggests that TNF-alfa and endothelin-1 have a synergistic role in producing HPS .
Summary of Imaging Studies:
Summary of Pathologic Findings:
References: To return to reference section after viewing abstract, click here before clicking on "abstract".
1. Hoeper M, Krowka M, Strassburg C. Portopulmonary hypertension and hepatopulmonary syndrome. Lancet 2004; 363:1461-1468. Abstract
2. Gupta D, Vijaya D, Gupta R, Dhiman R, Bhargava M, Verma J, Chawla Y. Prevalence of hepatopulmonary syndrome in cirrhosis and extahepatic portal venous obstruction. Am J Gastroenterol 2001; 96:3395-3399. Abstract
3. Williams A, Trewby P, Williams R, Reid L. Structural alterations to the pulmonary circulation in fulminant hepatic failure. Thorax 1979: 34:447-453. Abstract
4. Castro M, Krowka M. Hepatopulmonary syndrome. A pulmonary vascular complication of liver disease. Clin Chest Med 1996; 17:35-48.
5. Krowka M, Mandell M, Ramsay M, Kawut S, Fallon M, Manzarbeitia C, Pardo Jr M, et al. Hepatopulmonary syndrome and portopulmonary hypertension: a report of the multicenter liver transplant database. Liver Transpl 2004:10:174-182. Abstract
6. Berthelot P, Walker J, Sherlock S, Reid L. Arterial changes in the lungs in cirrhosis of the liver--lung spider nevi. N Engl J Med 1966: 274:291-298.
7. Scott V, Dodson S, Kang Y. The hepatopulmonary syndrome. Surg Clin N Am 1999: 79:23-41.
8. Krowka M, Wiseman G, Burnett O, Spivey J, Therneau T, Porayko M, Wiesner R. Hepatopulmonary syndrome. A prospective study of relationships between severity of liver disease, PaO2 response to 100% oxygen, and brain uptake after 99mTc MAA lung scanning. Chest 2000; 118:615-624. Abstract
9. Lee K, Lee H, Shin W, Webb W. Hypoxemia and liver cirrhosis (hepatopulmonary syndrome) in eight patients: comparison of the central and peripheral pulmonary vasculature. Radiology 1999; 211:549-553. Abstract
10. Lima B, Franca A, Pazin-Filho A, Araujo W, Martinez J, Maciel B, Simoes M, et al. Frequency, clinical characteristics, and respiratory parameters of hepatopulmonary syndrome. Mayo Clin Proc 2004; 79:42-48. Abstract
11. Arguedas M, Abrams G, Krowka M, Fallon M. Prospective evaluation of outcomes and predictors of mortality in patients with hepatopulmonary syndrome undergoing liver transplantation. Hepatology 2003; 37:192-197. Abstract
12. Rolla G, Brussino L, Colagrande P, Scappaticci E, Morello M, Bergerone S, Ottobrelli A, et al. Exhaled nitric oxide and impaired oxygenation in cirrhotic patients before and after liver transplantation. Ann Intern Med 1998; 129:375-378. Abstract
13. Ling Y, Zhang J, Luo B, Song D, Liu L, Tang L, Stockard C, et al. The role of endothelin-1 and the endothelin B receptor in the pathogenesis of hepatopulmonary syndrome in the rat. Hepatology 2004; 39:1593-1602. Abstract
14. Rabiller A, Nunes H, Lebrec D, Tazi K, Wartski M, Dulmet E, Libert J-M, et al. Prevention of Gram-negative translocation reduces the severity of hepatopulmonary syndrome. Am J Respir Crit Care Med 2002; 166:514-517. Abstract
15. Sztrymf B, Rabiller A, Nunes H, Savale L, Lebrec D, Le Pape A, de Montpreville V, et al. Prevention of hepatopulmonary syndrome and hyperdynamic state by pentoxifylline in cirrhotic rats. Eur Respir J 2004; 23:752-758. Abstract. See also: Editorial, Dinh-Xuan A, Naeije R. The hepatopulmonary syndrome: NO way out? ibid 661-662.
16. Luo B, Liu L, Tang L, Zhang J, Ling Y, Fallon M. ET-1 and TNF-alfa in HPS: analysis in prehepatic portal hypertension and biliary and nonbiliary cirrhosis in rats. Am J Physiol Gastrointest Liver Physiol 2004; 286:G294-G303. Abstract
Rad/Path Home Page
Last modified 10/26/04
Small muscular artery
Small vein with intimal fibrosis partly narrowing the lumen
Histologic differential diagnosis: 1. Pulmonary veno-occlusive disease. The venous intimal fibrosis and capillary congestion suggest this diagnosis, but neither the vein wall nor the adventitia is thickened. Hemosiderin-laden alveolar macrophages are absent. The pulmonary artery pressure was normal.
2. Chronic passive congestion. The absence of hemosiderin-laden alveolar macrophages and lack of history of heart failure exclude this diagnosis.
3. Lung post-cavopulmonary anastomosis or hepatic vein diversion to left atrium for congenital heart disease. The histologic appearance is consistent with this diagnosis, but the patient is an adult without a history of congenital heart disease. See: Vettukattil J. Pathogenesis of pulmonary arteriovenous malformations: role of hepatopulmonary interactions. Heart 2002:88:561-563. Article
4. Hepatopulmonary syndrome (HPS). The normal lung architecture with subtle dilation of vessels smaller than muscular arteries is consistent with HPS in a patient with cirrhosis. A 99mtechnetium-labeled macroaggregated albumin shunt fraction was 15%. The final diagnosis was HPS.
It should be noted that diagnosis is not based on this routine histologic appearance (see Discussion).