Massive Pancreatic Pleural Effusion: Pathogenesis of Pancreatic Duct Disruption DAISUKE SEMBA, YOSHIYUKI WADA, YUKIO ISHIHARA, TOSHIO KAJI, AKIRA KURODA, and YASUHIKO MORIOKA First Surgical Department, The University of Tokyo, Tokyo, Japan
Massive pancreatic pleural effusion is believed to be caused by pancreatic duct disruption. To elucidate the cause of the duct disruption, pancreas specimens resected from three cases of massive pancreatic pleural effusion were investigated histopathologitally. There was no evidence of either chronic or acute pancreatitis except focal fibrosis along a single dilated duct in the resected pancreas. From our study, the pathogenesis of pancreatic duct disruption is suggested as follows: (a) alcohol ingestion can induce focal acute inflammation on a single branch of the duct system and elicit protein plug formation; and (b) focal stenotic change occurs on this branch at a point near the main pancreatic duct, and transient obstruction can occur by means of the protein plugs. The “upstream” extent of this branch will then dilate until it ruptures. However, on the frequency of this atypical pancreatitis, we have no clues to mention.
R
eactive pleural effusion is often observed during the course of acute pancreatitis. Its volume is usually small and self-limited, and it resolves spontaneously as the pancreatic inflammation subsides (1,2). In contrast, massive pancreatic pleural effusion is rarely encountered. Its fluid is usually bloody and is characterized by its volume, high protein and amylase content, dominant pancreatic type amylase isoenzyme, and close relationship to alcohol. The etiology of this type of pleural effusion has been controversial (3-5). In 1973, Anderson et al. (6) reported that it may be caused by pancreatic duct disruption with leakage of pancreatic juice into the pleural cavity. However, it has not yet been explained why and how pancreatic duct disruption occurs without remarkable abdominal signs and symptoms. In this study, we investigate histopathologically tbree cases of resected pancreas of
massive pancreatic pleural effusion and discuss the pathogenesis of the pancreatic duct disruption. Case Reports Case 1 A 46-year-old man was admitted to our surgical unit with a complaint of dyspnea. He had a history of alcoholinduced acute pancreatitis 2 months before this admission. A plain chest radiograph at the time of admission showed a massive pleural effusion on the left side. ‘I’horacentesis yielded approximately 1000 mL of bloody fluid with high amylase content (198,000 IU/L). Isoamylase analysis proved
that the pancreatic type was dominant. Exocrine pancreatic function was normal by N-benzoyl-ltyrosol-p-aminobenzoic acid (BT-PABA) testing. Endoscopic retrograde pancreatography (ERP) disclosed an internal fistulous tract, with contrast medium extravasated upward from the tail to the esophageal hiatus (Figure 1A). At laparotomy, a retroperitoneal fistulous tract was shown originating from the anterosuperior aspect of the tail and terminating in the esophageal hiatus. Distal pancreatectomy was performed. Histological findings. The specimen was divided frontally along the main pancreatic duct, locating the disruption site in the anterior half. Serial sagittal sections of 5 mm width on this half were studied. Dense fibrosis with focal ductular proliferation were seen around the fistular origin of the pancreas (Figure 2A). In the neighboring section, a dilated cavity with atrophic ductal epithelium was observed [Figure 2B); this was hence regarded as a dilated duct that derived directly from the main pancreatic duct in the next section (Figure 2C). Fibrous change existed around the wall of this dilated duct, which seemed stenotic at its proximal
Abbreviations used in this paper: BT-PABA, N-benzoyl-l-tyrosylp-aminobenzoic acid; ERP, endoscopic retrograde pancreatograPhy. 0 1990by the Aamican Gastroenterological Association 0016-5065/90/$3.00
August 1990
MASSIVE PANCREATIC PLEURAL EFFUSION
529
Figure I. Endoscopic retrograde pancreatography. A. The ERP of case 1 shows extrapan-
creatic fistular tract from the tail. The pancreatic duct appears normal. B. The ERP of case 2 shows a wide fistular tract from the tail to the pleural cavity. C. The ERP of case 3 shows extrapancreatic pooling of contrast medium. site, with protein plugs in it. However, acinar tissue except this periductal area was normal (Figure 2B and C).
Case 2 The patient was a %-year-old alcoholic man. Clinical findings, including chest radiographs, ERP (Figure lB),
and laboratory findings, were almost the same with those of case 1. Laparotomy showed a wide fistulous tract originating from the tail of the pancreas and terminating in the esophageal hiatus. Distal pancreatectomy was performed. Histological findings were similar to those of case 1. Marked fibrosis that made the disrupted duct stenotic in its proximal portion to the main pancreatic duct was identified;
530 SEMBA ET AL.
GASTROENTEROLOGY
Vol. 99, No. 2
the only lesion in this specimen (Figure 3Aj. In the area adjacent to this, however, preservation of acinar tissue and structures was observed (Figure 3BJ.
it was
Case 3 A %-year-old man, also alcoholic, was admitted to our unit because of dyspnea caused by left massive pleural effusion. Abdominal radiographs showed no calcification, but a mass was palpable on the left upper quadrant. Endoscopic retrograde pancreatography showed extra pancreatic pooling of contrast medium (Figure lC), giving a diagnosis of a pancreatic pseudocyst. Cystography after drainage showed that this cavity extended upward to the left diaphragm and downward to the left iliac crest in the retroperitoneal space. The pleural effusion was hence suggested to have been caused by the rupture of the retroperitoneal fluid accumulation into the pleural cavity. Laparotomy proved that the cyst originated from the tail of the pancreas and extended to the retroperitoneal cavity. The cyst and the tail were removed. Microscopically, except for mild periductal fibrosis, we found no changes compatible with those of chronic and/or acute pancreatitis in this specimen (Figure 4). Discussion Regarding the etiology of massive pancreatic pleural effusion, it is believed that this type of effusion is the result of leakage of pancreatic juice from the duct system into the pleural space, either by a sinus tract or communicating pseudocyst (6). However, the cause of the duct disruption is not yet explained. From our study on the resected specimens, we identified some similar characteristics. Initially, we could find no evidence compatible with either acute or chronic pancreatitis, such as fat necrosis, cellular infiltration, and diffuse interlobular fibrosis, in our specimens. The site of pancreatic duct disruption was identified in two cases. In the section of this site, fibrosis and ductular proliferation with focal acinar destruction were observed in case 1 (Figure 2A). Furthermore, we could observe marked periductal fibrosis of the dilated duct (Figure 2B) and fibrous change surrounding the proximal portion of the disrupted duct (Figures 2C and 3A), which were the primary branches derived directly from the main pancreatic duct. It was most impressive that in both cases, these inflammatory changes were localized only around a single duct extended to the disruption site that there was not any such lesion apart from this focal area, and that most of the acinar tissue and its structure were preserved well in each specimen (Figures 2B and 3B). Therefore, we considered these findings as evidence for a sequence of events leading to duct disruption, beginning with focal inflammation of a single primary branch of the duct system. Histolog-
Figure2. Case 1.A. Around the site of disruption (arrow], fibrosis and ductular proliferations are seen but there are no cellular infiltrations. Main pancreatic duct (arrowhead] is divided (H&E; original magnification x 10). B. Section adjacent to that of A. The intrapancreatic dilated cavity is surrounded by fibrosis. Because the inner surface of this cavity is lined with single columnar epithelium, the cavity is proved to be a dilated duct. Arrowhead indicates the main duct (H&E; original magnification x6). C. Section adjacent to that of B. The dilated duct derives directly from the main pancreatic duct (arrowhead). The lumen of this duct appears to be narrowing at its proximal portion. Protein plugs are seen in the duct. (H&E; original magnification x6).
MASSIVE PANCREATIC
August 1990
Figure 3. Case 2. A. Marked fibrosis disrupted duct stenotic which is filled heads) injected postoperatively. The disrupted duct and the white curved (H&E; original magnification x25).
PLEURAL EFFUSION
531
near the main duct made the with contrast medium (arrowblack curved arrow indicates arrow indicates the main duct
B. Area adjacent to A. Preservation of acinar tissue and structures are observed. Black and white curved orrows indicate the same as in A (H&E; original magnification x 12.5).
ically, it is not difficult to conceive that these lesions were neither acute nor chronic pancreatitis; clinically, there was no calcification on abdominal radiographs, no persistent abdominal pain, and no abdominal trauma in all. But it is also apparent that inflammatory sequence occurred in this limited area. We concluded that these pathological conditions were the result of an atypical type of pancreatitis, namely, focal acute pancreatitis. Generally, an obstructive element is essential for the duct disruption, and we presume that it was in this limited area, the proximal portion of the primary branch to the main duct, that focal inflammation
Figure 4. Case 3.Only pmiductal fibrosis of mild degree are seen (H&JQoriginal magnification x 10j
associated with protein plugs led to duct disruption. We believe that the protein plugs are expelled and dissolved after disruption, leaving behind a stenotic segment of the duct through which most of the pancreatic secretions escape. Once the pancreatic juice escapes into the retroperitoneal space, it will usually move upward because of the transdiaphragmatic pressure gradient between abdominal and pleural cavity (7). It will then enter the posterior mediastinum, either through the aotic or esophageal hiatus (8-10) or, less commonly, directly through the dome of the diaphragm (11,12). In most cases, mediastinal accumulation will rupture into the pleural cavity through the hilum of the lung. Because proteolytic enzymes are in an inactive form and the disruption site of the pancreas resembles a pin hole covered by the peritoneum, few abdominal signs and symptoms appear. In one of our cases, only an extrapancreatic pseudocyst was found, which was drained to the retroperitoneal space. Also in this case, we believe that the above sequence of duct disruption occurred and that the evidence of focal lesion was repaired in the course of time. Therefore, in this specimen we found no changes suggestive of chronic or acute pancreatitis (Figure4). We did find several reports of this type of retroperitoneal extension of the pseudocyst, e.g., extending up to the mediastinum (11) and to the pelvic (13) or inguinal region (14). Finally, only focal lesions were observed on the resected specimens in our cases, but we cannot explain the association with alcohol ingestion. However,
GASTROENTEROLOGY
532 SEMBA ET AL.
Sarles (15) emphasized that the focal nature of pancreatitis is characteristic of alcohol-induced pancreatitis. We concur with this formulation on the basis of our own studies, but on the frequency of this atypical inflammation, we have no clues to discuss. In most published reports, massive pancreatic pleural effusion is reported to be associated with chronic pancreatitis. However, we believe that the direct cause may be focal acute pancreatitis, as in our cases, although there is a paucity of information in the literature about this pathological entity.
6.
9.
10. 11. 12.
References 1. Goldman M, Goldman G, Fleischner
2. 3. 4. 5.
6.
7.
FG. Pleural fluid amylase in acute pancreatitis. N Engl J Med 1962;226:715-716. Walton TP, Quinity RJ. Supradiaphragmatic aspect of benign pancreatic disease. Am Surg 1975:41:32-36. Moreaux J, Clot J-P, Mercadier M, Hepp J. Pancreatites et Epanchements. Ann Chir 1966:20:413-423. Kaye MD. Pleuropulmonary complications of pancreatitis. Thorax 1966;23:297-306. Geffroy Y, Bourreille J. Morere P, Sageot M-C, Colin R. Les epanchements pleuraux des pancreatopathies chroniques. Semin Hop Paris 1966;26:1356-1368. Anderson WJ, Skinner DB, Zuidema GD, Cameron JL. Chronic pancreatic pleural effusions. Surg Gynecol Obstet 1973;137:627830. Faling LJ, Gerzof SG, Daly BDT, PuGatch RD, Snider GL.
13.
14.
15.
Vol. 99, No. 2
Treatment of chronic pancreatic pleural effusion by percutaneous catheter drainage of abdominal pseudocyst. Am J Med 1964;76329-333. Rignault D, Tardat M, Berthet A, Pailler JL. Communication directe au tours d’une pleuresie enzymatique. mis en evidence par wirsungographie. Nouv Presse Med 1976;13:2536-2536. Marceau P, Trolliet P, Bourque R-A, Piche P, Lou W. Epanchement pleural revelateur de pancreatite. Can J Surg 1980;23:8789. Dewan NA, Kinney WW, O’Donohue WJ. Chronic massive pancreatic pleural effusion. Chest 1984;85:497-501. Jaffe BM, Ferguson TB, Holtz S, Shields JB. Mediastinal pancreatic pseudocysts. Am J Surg 1972;124:600-606. Tombroff M, Loicq A, DeKoster J-P, Engleholm L, Govaerts J-P. Pleural effusion with pancreaticopleural fistula. Br Med J 1973;1:330-331. Farman J, Kutcher R, Dallemand S, Lane FC, Becker JA. Unusual pelvic complications of a pancreatic pseudocyst. Gastrointest Radio1 1978;3:43-45. Lombrozo R, Wolloch Y, Dintsman M. Retroperitoneal dissection of a pancreatic pseudocyst to the left inguinal region. Isr J Med Sci 1977;13:309-312. Sarles H. Alcoholism and pancreatitis. Stand J Gastroenterol 197X6:193-198.
Received June 12.1989. Accepted February 22.1990. Address requests for reprints to: Daisuke Semba, M.D., First Surgical Department, The University of Tokyo, 7-3-l. Hongo, Bunkyo-ku, Tokyo, 113, Japan.
Massive pancreatic pleural effusion is believed to be caused by pancreatic duct disruption. To elucidate the cause of the duct disruption, pancreas specimens resected from three cases of massive pancreatic pleural effusion were investigated histopathologitally. There was no evidence of either chronic or acute pancreatitis except focal fibrosis along a single dilated duct in the resected pancreas. From our study, the pathogenesis of pancreatic duct disruption is suggested as follows: (a) alcohol ingestion can induce focal acute inflammation on a single branch of the duct system and elicit protein plug formation; and (b) focal stenotic change occurs on this branch at a point near the main pancreatic duct, and transient obstruction can occur by means of the protein plugs. The “upstream” extent of this branch will then dilate until it ruptures. However, on the frequency of this atypical pancreatitis, we have no clues to mention.
R
eactive pleural effusion is often observed during the course of acute pancreatitis. Its volume is usually small and self-limited, and it resolves spontaneously as the pancreatic inflammation subsides (1,2). In contrast, massive pancreatic pleural effusion is rarely encountered. Its fluid is usually bloody and is characterized by its volume, high protein and amylase content, dominant pancreatic type amylase isoenzyme, and close relationship to alcohol. The etiology of this type of pleural effusion has been controversial (3-5). In 1973, Anderson et al. (6) reported that it may be caused by pancreatic duct disruption with leakage of pancreatic juice into the pleural cavity. However, it has not yet been explained why and how pancreatic duct disruption occurs without remarkable abdominal signs and symptoms. In this study, we investigate histopathologically tbree cases of resected pancreas of
massive pancreatic pleural effusion and discuss the pathogenesis of the pancreatic duct disruption. Case Reports Case 1 A 46-year-old man was admitted to our surgical unit with a complaint of dyspnea. He had a history of alcoholinduced acute pancreatitis 2 months before this admission. A plain chest radiograph at the time of admission showed a massive pleural effusion on the left side. ‘I’horacentesis yielded approximately 1000 mL of bloody fluid with high amylase content (198,000 IU/L). Isoamylase analysis proved
that the pancreatic type was dominant. Exocrine pancreatic function was normal by N-benzoyl-ltyrosol-p-aminobenzoic acid (BT-PABA) testing. Endoscopic retrograde pancreatography (ERP) disclosed an internal fistulous tract, with contrast medium extravasated upward from the tail to the esophageal hiatus (Figure 1A). At laparotomy, a retroperitoneal fistulous tract was shown originating from the anterosuperior aspect of the tail and terminating in the esophageal hiatus. Distal pancreatectomy was performed. Histological findings. The specimen was divided frontally along the main pancreatic duct, locating the disruption site in the anterior half. Serial sagittal sections of 5 mm width on this half were studied. Dense fibrosis with focal ductular proliferation were seen around the fistular origin of the pancreas (Figure 2A). In the neighboring section, a dilated cavity with atrophic ductal epithelium was observed [Figure 2B); this was hence regarded as a dilated duct that derived directly from the main pancreatic duct in the next section (Figure 2C). Fibrous change existed around the wall of this dilated duct, which seemed stenotic at its proximal
Abbreviations used in this paper: BT-PABA, N-benzoyl-l-tyrosylp-aminobenzoic acid; ERP, endoscopic retrograde pancreatograPhy. 0 1990by the Aamican Gastroenterological Association 0016-5065/90/$3.00
August 1990
MASSIVE PANCREATIC PLEURAL EFFUSION
529
Figure I. Endoscopic retrograde pancreatography. A. The ERP of case 1 shows extrapan-
creatic fistular tract from the tail. The pancreatic duct appears normal. B. The ERP of case 2 shows a wide fistular tract from the tail to the pleural cavity. C. The ERP of case 3 shows extrapancreatic pooling of contrast medium. site, with protein plugs in it. However, acinar tissue except this periductal area was normal (Figure 2B and C).
Case 2 The patient was a %-year-old alcoholic man. Clinical findings, including chest radiographs, ERP (Figure lB),
and laboratory findings, were almost the same with those of case 1. Laparotomy showed a wide fistulous tract originating from the tail of the pancreas and terminating in the esophageal hiatus. Distal pancreatectomy was performed. Histological findings were similar to those of case 1. Marked fibrosis that made the disrupted duct stenotic in its proximal portion to the main pancreatic duct was identified;
530 SEMBA ET AL.
GASTROENTEROLOGY
Vol. 99, No. 2
the only lesion in this specimen (Figure 3Aj. In the area adjacent to this, however, preservation of acinar tissue and structures was observed (Figure 3BJ.
it was
Case 3 A %-year-old man, also alcoholic, was admitted to our unit because of dyspnea caused by left massive pleural effusion. Abdominal radiographs showed no calcification, but a mass was palpable on the left upper quadrant. Endoscopic retrograde pancreatography showed extra pancreatic pooling of contrast medium (Figure lC), giving a diagnosis of a pancreatic pseudocyst. Cystography after drainage showed that this cavity extended upward to the left diaphragm and downward to the left iliac crest in the retroperitoneal space. The pleural effusion was hence suggested to have been caused by the rupture of the retroperitoneal fluid accumulation into the pleural cavity. Laparotomy proved that the cyst originated from the tail of the pancreas and extended to the retroperitoneal cavity. The cyst and the tail were removed. Microscopically, except for mild periductal fibrosis, we found no changes compatible with those of chronic and/or acute pancreatitis in this specimen (Figure 4). Discussion Regarding the etiology of massive pancreatic pleural effusion, it is believed that this type of effusion is the result of leakage of pancreatic juice from the duct system into the pleural space, either by a sinus tract or communicating pseudocyst (6). However, the cause of the duct disruption is not yet explained. From our study on the resected specimens, we identified some similar characteristics. Initially, we could find no evidence compatible with either acute or chronic pancreatitis, such as fat necrosis, cellular infiltration, and diffuse interlobular fibrosis, in our specimens. The site of pancreatic duct disruption was identified in two cases. In the section of this site, fibrosis and ductular proliferation with focal acinar destruction were observed in case 1 (Figure 2A). Furthermore, we could observe marked periductal fibrosis of the dilated duct (Figure 2B) and fibrous change surrounding the proximal portion of the disrupted duct (Figures 2C and 3A), which were the primary branches derived directly from the main pancreatic duct. It was most impressive that in both cases, these inflammatory changes were localized only around a single duct extended to the disruption site that there was not any such lesion apart from this focal area, and that most of the acinar tissue and its structure were preserved well in each specimen (Figures 2B and 3B). Therefore, we considered these findings as evidence for a sequence of events leading to duct disruption, beginning with focal inflammation of a single primary branch of the duct system. Histolog-
Figure2. Case 1.A. Around the site of disruption (arrow], fibrosis and ductular proliferations are seen but there are no cellular infiltrations. Main pancreatic duct (arrowhead] is divided (H&E; original magnification x 10). B. Section adjacent to that of A. The intrapancreatic dilated cavity is surrounded by fibrosis. Because the inner surface of this cavity is lined with single columnar epithelium, the cavity is proved to be a dilated duct. Arrowhead indicates the main duct (H&E; original magnification x6). C. Section adjacent to that of B. The dilated duct derives directly from the main pancreatic duct (arrowhead). The lumen of this duct appears to be narrowing at its proximal portion. Protein plugs are seen in the duct. (H&E; original magnification x6).
MASSIVE PANCREATIC
August 1990
Figure 3. Case 2. A. Marked fibrosis disrupted duct stenotic which is filled heads) injected postoperatively. The disrupted duct and the white curved (H&E; original magnification x25).
PLEURAL EFFUSION
531
near the main duct made the with contrast medium (arrowblack curved arrow indicates arrow indicates the main duct
B. Area adjacent to A. Preservation of acinar tissue and structures are observed. Black and white curved orrows indicate the same as in A (H&E; original magnification x 12.5).
ically, it is not difficult to conceive that these lesions were neither acute nor chronic pancreatitis; clinically, there was no calcification on abdominal radiographs, no persistent abdominal pain, and no abdominal trauma in all. But it is also apparent that inflammatory sequence occurred in this limited area. We concluded that these pathological conditions were the result of an atypical type of pancreatitis, namely, focal acute pancreatitis. Generally, an obstructive element is essential for the duct disruption, and we presume that it was in this limited area, the proximal portion of the primary branch to the main duct, that focal inflammation
Figure 4. Case 3.Only pmiductal fibrosis of mild degree are seen (H&JQoriginal magnification x 10j
associated with protein plugs led to duct disruption. We believe that the protein plugs are expelled and dissolved after disruption, leaving behind a stenotic segment of the duct through which most of the pancreatic secretions escape. Once the pancreatic juice escapes into the retroperitoneal space, it will usually move upward because of the transdiaphragmatic pressure gradient between abdominal and pleural cavity (7). It will then enter the posterior mediastinum, either through the aotic or esophageal hiatus (8-10) or, less commonly, directly through the dome of the diaphragm (11,12). In most cases, mediastinal accumulation will rupture into the pleural cavity through the hilum of the lung. Because proteolytic enzymes are in an inactive form and the disruption site of the pancreas resembles a pin hole covered by the peritoneum, few abdominal signs and symptoms appear. In one of our cases, only an extrapancreatic pseudocyst was found, which was drained to the retroperitoneal space. Also in this case, we believe that the above sequence of duct disruption occurred and that the evidence of focal lesion was repaired in the course of time. Therefore, in this specimen we found no changes suggestive of chronic or acute pancreatitis (Figure4). We did find several reports of this type of retroperitoneal extension of the pseudocyst, e.g., extending up to the mediastinum (11) and to the pelvic (13) or inguinal region (14). Finally, only focal lesions were observed on the resected specimens in our cases, but we cannot explain the association with alcohol ingestion. However,
GASTROENTEROLOGY
532 SEMBA ET AL.
Sarles (15) emphasized that the focal nature of pancreatitis is characteristic of alcohol-induced pancreatitis. We concur with this formulation on the basis of our own studies, but on the frequency of this atypical inflammation, we have no clues to discuss. In most published reports, massive pancreatic pleural effusion is reported to be associated with chronic pancreatitis. However, we believe that the direct cause may be focal acute pancreatitis, as in our cases, although there is a paucity of information in the literature about this pathological entity.
6.
9.
10. 11. 12.
References 1. Goldman M, Goldman G, Fleischner
2. 3. 4. 5.
6.
7.
FG. Pleural fluid amylase in acute pancreatitis. N Engl J Med 1962;226:715-716. Walton TP, Quinity RJ. Supradiaphragmatic aspect of benign pancreatic disease. Am Surg 1975:41:32-36. Moreaux J, Clot J-P, Mercadier M, Hepp J. Pancreatites et Epanchements. Ann Chir 1966:20:413-423. Kaye MD. Pleuropulmonary complications of pancreatitis. Thorax 1966;23:297-306. Geffroy Y, Bourreille J. Morere P, Sageot M-C, Colin R. Les epanchements pleuraux des pancreatopathies chroniques. Semin Hop Paris 1966;26:1356-1368. Anderson WJ, Skinner DB, Zuidema GD, Cameron JL. Chronic pancreatic pleural effusions. Surg Gynecol Obstet 1973;137:627830. Faling LJ, Gerzof SG, Daly BDT, PuGatch RD, Snider GL.
13.
14.
15.
Vol. 99, No. 2
Treatment of chronic pancreatic pleural effusion by percutaneous catheter drainage of abdominal pseudocyst. Am J Med 1964;76329-333. Rignault D, Tardat M, Berthet A, Pailler JL. Communication directe au tours d’une pleuresie enzymatique. mis en evidence par wirsungographie. Nouv Presse Med 1976;13:2536-2536. Marceau P, Trolliet P, Bourque R-A, Piche P, Lou W. Epanchement pleural revelateur de pancreatite. Can J Surg 1980;23:8789. Dewan NA, Kinney WW, O’Donohue WJ. Chronic massive pancreatic pleural effusion. Chest 1984;85:497-501. Jaffe BM, Ferguson TB, Holtz S, Shields JB. Mediastinal pancreatic pseudocysts. Am J Surg 1972;124:600-606. Tombroff M, Loicq A, DeKoster J-P, Engleholm L, Govaerts J-P. Pleural effusion with pancreaticopleural fistula. Br Med J 1973;1:330-331. Farman J, Kutcher R, Dallemand S, Lane FC, Becker JA. Unusual pelvic complications of a pancreatic pseudocyst. Gastrointest Radio1 1978;3:43-45. Lombrozo R, Wolloch Y, Dintsman M. Retroperitoneal dissection of a pancreatic pseudocyst to the left inguinal region. Isr J Med Sci 1977;13:309-312. Sarles H. Alcoholism and pancreatitis. Stand J Gastroenterol 197X6:193-198.
Received June 12.1989. Accepted February 22.1990. Address requests for reprints to: Daisuke Semba, M.D., First Surgical Department, The University of Tokyo, 7-3-l. Hongo, Bunkyo-ku, Tokyo, 113, Japan.
