- Email: [email protected]
PERCUTANEOUS TRANSHEPATIC OBLITERATION OF GASTRO-ŒSOPHAGEAL VARICES
Saturday
PERCUTANEOUS TRANSHEPATIC OBLITERATION OF GASTRO-ŒSOPHAGEAL VARICES
R. G. LONG SHEILA SHERLOCK
J.SCOTT R. DICK
Departments of Medicine and Radiology, Royal Free Hospital, London NW3 2QG
Summary
Percutaneous
transhepatic portal-vein
catheterisation was attempted to obliterate the major variceal venous supply in 13 decompensated cirrhotic patients, who continued to bleed after conservative therapy. Obliteration was achieved and bleeding stopped in 7 patients. In 5 patients obliteration was technically unsuccessful. The remaining patient had an unsuspected portal-vein block diagnosed by the transhepatic technique. 1 patient with successfully obliterated varices died after a hæmothorax and hæmoperitoneum three to
developed. Follow-up splenic venography
at
six months in the 6 successfully thrombosed patients showed that 4 had persistent obliteration and had not re-bled. 2 patients re-bled from incompletely obliterated varices. It is concluded that selective obliteration of the major variceal supply is effective in stopping acute gastro-œsophageal variceal bleeding, but that greater experience is necessary before the long-term effectiveness of the procedure can be determined. Introduction
EMERGENCY surgery for bleeding pstro-cesophageal varices in patients with severely decompensated cirrhosis
10
July 1976
has a very high mortality.’One emergency procedure is local devascularisation by injection of a thrombosing agent directly into the varices down an resophagoscope.’ * Alternatively the portal vein may be catheterised and the variceal supply selectively catheterised and obliterated. A percutaneous transhepatic method of portal venography has been described and used to treat bleeding varices.56 We used this technique to treat bleeding varices in patients with end-stage cirrhosis who would not be suitable for surgery. Materials and Methods The 13
patients with decompensated cirrhosis were actively from varices (see accompanying table). The patients bleeding had not responded to conservative therapy with blood, blood products, and intravenous vasopressin, and surgery was contraindicated because of poor hepatic function. In view of the discomfort and complications the Sengstaken oesophageal compression tube was not used in these very poor-risk patients, who could not proceed to surgery. All 13 patients had endoscopic evidence of variceal bleeding. The portal vein was localised in both anteroposterior and lateral planes and shown to be patent by splenic venography, by the venous phase of superior mesenteric arteriography, or by grey-scale ultrasound. Under local anaesthesia the liver was punctured during apncea in the mid-axillary line below the costophrenic reflection by means of a cholangiography needle with a radio-opaque catheter (Surgimed, Denmark) (length 27 cm, external diameter 1.58 mm, internal diameter 1.18 mm). The needle was advanced under image intensification in the direction of the portal vein and to within 3 cm of the vertebral column. The needle was removed and the radio-opaque catheter gradually drawn back with suction applied until blood was freely
CLINICAL AND BIOCHEMICAL DATA IN
13 CIRRHOTIC
PATIENTS
"HBaAe positive. C.A.H,=Chronk active hepatitis. .8.C.= Primary biliary cirrhosis.
7976
54
Fig.
I-A
shown.
transhepatic catheter is shown advanced into splenic hilum.
supply veins (left gastric and short gastric veins) are Gastro-epiploic veins may occasionally also feed gastro-ceso-
Major
variceal
had active chronic hepatitis with cirrhosis diagnosed in 1973 and was receiving prednisolone. She had major variceal bleeds in 1973 and 1974; and in August, 1975, she had a further major bleed and 20 1 of blood was transfused during ten days. She showed grade-n portal systemic encephalopathy and moderate ascites. Serum-bilirubin was 65 umol/1 (normal range 4-17), albumin 35 g/1 (normal range 35-50), and the prothrombin-time was prolonged seven seconds over the control after vitamin K. Surgery was believed to be contraindicated. A transhepatic portal venogram was obtained (fig. 2) and the portal pressure was 35 mm Hg. The short gastric veins and then the left gastric vein were selectively catheterised (fig. 3) and thrombosed. A final portal venogram showed incomplete obliteration of the short gastric veins (fig. 4) which were not recatheterised. The final portal pressure was 37 mm Hg. The patient had not re-bled at nine months and a splenic venogram at six months showed only a small short gastric vein.
phageal varices.
Results Thrombosis
was
achieved and
acute
bleeding
from
aspirated. A test injection of contrast medium was made to confirm the position of the catheter in the portal vein or one of its branches. The catheter was advanced over a guide wire as far as possible along the splenic vein. The portal pressure was measured and a portal venogram obtained. The major variceal supply veins (left gastric and short gastric) (fig. 1) were selectively catheterised and thrombosed. 30 ml of 500 g/1 dextrose were injected to traumatise the intima of the veins, increasing quantities of human thrombin (500-3000 units) were injected to induce complete thrombosis, and small pieces of gelatin foam were injected to stabilise the thrombus. Contrast medium was then injected to confirm successful obliteration, and if this had been achieved the catheter
carefully withdrawn. The pressure was remeasured in the portal vein and a portal venogram was obtained to ensure complete obliteration of all variceal supply veins and portal-vein patency. Finally, before complete withdrawal of the catheter from the liver, the hepatic puncture wound was plugged by an injection of gelatin foam into the subcapsular parenchyma. was
Fig.
3-Selective catheterisation of left
gastric vein.
Case-report A
twenty-nine-year-old housewife, with two
young children,
Fig. 4-Portal venogram after obliteration of left gastric vein showing incompletely obliterated short gastric vein (white arrow). Fig. 2-Portal venogram showing filling of gastro-cesophageal collaterals (white arrow) and of pancreatic venous plexus (black arrow).
Note contrast trapped in clot tric-vein radicle.
peripherally (black arrow) in left
an
gas-
55
stopped in 7 of the 13 patients. In 5 patients the procedure failed because of technical difficulty in selective catheterisation of small variceal-feeding vessels (4 patients) and because of massive ascites and a very hard liver (1patient). The remaining patient had a previously unsuspected portal-vein block so that the major portal
varices
vein could not be entered. In 3 patients the gallbladder was punctured but there was no biliary leak. Hgemothorax developed in 5 patients and aspiration was required in 2. 2 of these patients also had a haemoperitoneum. 1 recovered after transfusion. The 2nd, however, who also had a haemopehtoneuth and in whom the varices had been successfully thrombosed died. Necropsy showed no variceal bleeding but there was a 1 5 cm tear in the liver capsule. The 5 patients with bleeding complications were treated early in the study and gelatin foam had not been used to plug the exit puncture in the liver. Eight subsequent punctures using gelatin foam have been performed without bleeding. Portal-vein thrombosis did not occur, but in 3 patients strands of thrombin-induced clot were seen in the portal vein. Follow-up with splenic or hepatic venography at three to six months on the 6 successfully treated surviving patients in whom bleeding had been controlled showed that the portal vein was patent in all patients and that in 4 obliteration of varices had persisted. In 1 patient rebleeding occurred from incompletely obliterated varices ten days after the procedure; in the other patient bleeding recurred at three months from a new collateral vein filled from the left gastro-epiploic vein.
Discussion Percutaneous transhepatic portal-vein catheterisation and obliteration of the major variceal supply seems to be an effective procedure for stopping variceal bleeding. The major indication is in patients with severe decompensated cirrhosis, who have not responded to conservative treatment, and in whom, because of liver failure, surgery is contraindicated. After successful obliteration of varices - surgery may be performed electively or referred indefinitely. The long-term effectiveness of variceal obliteration has not been established. In 4 of our patients obliteration has persisted for several months; the longest follow-up period is 9 months: However, Viamonte has (personal communication) treated over 40 patients by this method and in some obliteration of varices has persisted for more than 1 year. -‘ Portal-vein catheterisation allows both fundal and oesophageal varices to be obliterated close to their origin. This is an advantage over oesophagoscopic injection where there is difficulty in obliterating all varices and especially those in the fundus of the stomach. Other approaches to the portal vein such as the translumbar routes and by umbilical-vein catheterisation9 are not as convenient as the transhepatic one. Obliteration of varices has the potential of stopping bleeding for an indefinite period by diversion of portal blood to other collateral vessels. It is well-tolerated and .
,
easily repeated. Portal systemic encephalopathy would not be a complication. In patients with severe cirrhosis, cesophageal transection may have a mortality-rate of more than 50%.’ Our results indicate that mortality from variceal obliteration is much less (1 of the 13 died). The benefit of shunt surgery is controversial since bleed-
ing may be replaced by encephalopathy and the chances of survival are probably not improved. 10 11 The major cause of failure was technical difficulty in catheterising the small variceal-supply veins. The success-rate has, however, improved partly because we have become more practised but mainly because we have used an increased range of guide wires (including moveable core and a variety of "J" shapes). Persistence of obliteration may depend on the venoocclusive agents used. Various combinations of autogenous clot, thrombin-induced clot, sclerosing agents, and gelatin foam matrix have been used.56 12 Initially we used thrombin-induced clot alone. Two of these patients re-bled after successful obliteration. We now prefer thrombin-induced clot stabilised with gelatin foam since this should give a more permanent clot. The complications of hxmothorax and haemoperitoneum (1 fatal) have not occurred since gelatin foam has been used to plug hepatic puncture wounds. Biliary puncture necessitates careful observation for biliary peritonitis. Trauma to the portal vein, pulmonary embolism, and portal-vein thrombosis are theoretical complications not observed so far, although strands of thrombin-induced clot have been seen in the portal vein. Follow-up of these patients by splenic venography showed that the portal vein was patent. Obliteration of the major variceal supply is potentially important in the management of acute variceal bleeding, particularly in patients with end-stage cirrhosis. The long-term role of the technique has yet to be established but it will probably replace both oesophageal transection and obliteration of varices by oesophagoscopy in such patients. We thank Dr Manuel Viamonte, Jr. for his assistance in developing this technique; and all members of the department of radiology who gave their technical assistance. REFERENCES
1.
George, P., Brown, C., Ridgway, G., Crofts, B., Sherlock,
S. Br. J. Surg. 1973, 60, 635. 2. Turcotte, J. G., Lambert, M. J. Surgery, 1973, 73, 810. 3. Crafoord, C., Frenckner, P. Acta Oto-Laryng. 1939, 27, 422. 4. Hunt, P. S., Johnston, G. W., Rodgers, H. W. Br. J. Surg. 1969, 56, 305. 5. Lunderquist, A., Vang, J. New Engl. J. Med. 1974, 291, 646. 6. Viamonte, M. Jr., Page, J., Lunderquist, A., Pereiras, R., Russell, E., Viamonte, M., Camacho, M. Radiology, 1975, 114, 457. 7. Kennedy, J. Nutrition, 1973, 27, 325. 8. Haverling, M., Overfors, C-O. Invest. Radiol. 1968, 3, 376. 9. Kessler, R. E., Tice, D. A., Zimmon, D. S. Surg. Gynec. Obstet. 1973, 136,
529. 10. Resnick, R. H., Iber, F. L., Ishihara, A. M., Chalmers, T. C., Zimmerman, H. Gastroenterology, 1974, 67, 843. 11. Rueff, B., Prandi, D., Degos, F., Sicot, J., Degos, J-D., Sicot, C., Maillard, J-N., Fauvert, R., Benhamou, J-P. Lancet, 1976, i, 655. 12. Reuter, S. R., Chuany, V. P., Bree, R. L., Major, M. C., Br. J. Radiol. 1975,
125, 119.
"Monopoly bargaining is a reality of the times. Nevertheless there is such a thing as being too well organised. For a profession can become so autonomous that it moves into a phase where it begins altogether to discount the experiences and evaluations of the public outside. This can mount to a point where wide community criticism threatens to reduce autonomy and thus remove some liberty. The professional symptoms which attract such criticism of an organised profession are: secret behaviour; professional colleagues protecting one another against all outside criticism; overt claims to exclusiveness or social rank; insistence on very long training without clear demonstration of its necessity; very highly conditioned loyalties and behaviour within the profession; claims to omniscience in human matters or to a monopoly of technical knowledge; and active resistance to change in society."-EARLE HACKETT, South Australian Climes, 1975, 7, 3.
PERCUTANEOUS TRANSHEPATIC OBLITERATION OF GASTRO-ŒSOPHAGEAL VARICES
R. G. LONG SHEILA SHERLOCK
J.SCOTT R. DICK
Departments of Medicine and Radiology, Royal Free Hospital, London NW3 2QG
Summary
Percutaneous
transhepatic portal-vein
catheterisation was attempted to obliterate the major variceal venous supply in 13 decompensated cirrhotic patients, who continued to bleed after conservative therapy. Obliteration was achieved and bleeding stopped in 7 patients. In 5 patients obliteration was technically unsuccessful. The remaining patient had an unsuspected portal-vein block diagnosed by the transhepatic technique. 1 patient with successfully obliterated varices died after a hæmothorax and hæmoperitoneum three to
developed. Follow-up splenic venography
at
six months in the 6 successfully thrombosed patients showed that 4 had persistent obliteration and had not re-bled. 2 patients re-bled from incompletely obliterated varices. It is concluded that selective obliteration of the major variceal supply is effective in stopping acute gastro-œsophageal variceal bleeding, but that greater experience is necessary before the long-term effectiveness of the procedure can be determined. Introduction
EMERGENCY surgery for bleeding pstro-cesophageal varices in patients with severely decompensated cirrhosis
10
July 1976
has a very high mortality.’One emergency procedure is local devascularisation by injection of a thrombosing agent directly into the varices down an resophagoscope.’ * Alternatively the portal vein may be catheterised and the variceal supply selectively catheterised and obliterated. A percutaneous transhepatic method of portal venography has been described and used to treat bleeding varices.56 We used this technique to treat bleeding varices in patients with end-stage cirrhosis who would not be suitable for surgery. Materials and Methods The 13
patients with decompensated cirrhosis were actively from varices (see accompanying table). The patients bleeding had not responded to conservative therapy with blood, blood products, and intravenous vasopressin, and surgery was contraindicated because of poor hepatic function. In view of the discomfort and complications the Sengstaken oesophageal compression tube was not used in these very poor-risk patients, who could not proceed to surgery. All 13 patients had endoscopic evidence of variceal bleeding. The portal vein was localised in both anteroposterior and lateral planes and shown to be patent by splenic venography, by the venous phase of superior mesenteric arteriography, or by grey-scale ultrasound. Under local anaesthesia the liver was punctured during apncea in the mid-axillary line below the costophrenic reflection by means of a cholangiography needle with a radio-opaque catheter (Surgimed, Denmark) (length 27 cm, external diameter 1.58 mm, internal diameter 1.18 mm). The needle was advanced under image intensification in the direction of the portal vein and to within 3 cm of the vertebral column. The needle was removed and the radio-opaque catheter gradually drawn back with suction applied until blood was freely
CLINICAL AND BIOCHEMICAL DATA IN
13 CIRRHOTIC
PATIENTS
"HBaAe positive. C.A.H,=Chronk active hepatitis. .8.C.= Primary biliary cirrhosis.
7976
54
Fig.
I-A
shown.
transhepatic catheter is shown advanced into splenic hilum.
supply veins (left gastric and short gastric veins) are Gastro-epiploic veins may occasionally also feed gastro-ceso-
Major
variceal
had active chronic hepatitis with cirrhosis diagnosed in 1973 and was receiving prednisolone. She had major variceal bleeds in 1973 and 1974; and in August, 1975, she had a further major bleed and 20 1 of blood was transfused during ten days. She showed grade-n portal systemic encephalopathy and moderate ascites. Serum-bilirubin was 65 umol/1 (normal range 4-17), albumin 35 g/1 (normal range 35-50), and the prothrombin-time was prolonged seven seconds over the control after vitamin K. Surgery was believed to be contraindicated. A transhepatic portal venogram was obtained (fig. 2) and the portal pressure was 35 mm Hg. The short gastric veins and then the left gastric vein were selectively catheterised (fig. 3) and thrombosed. A final portal venogram showed incomplete obliteration of the short gastric veins (fig. 4) which were not recatheterised. The final portal pressure was 37 mm Hg. The patient had not re-bled at nine months and a splenic venogram at six months showed only a small short gastric vein.
phageal varices.
Results Thrombosis
was
achieved and
acute
bleeding
from
aspirated. A test injection of contrast medium was made to confirm the position of the catheter in the portal vein or one of its branches. The catheter was advanced over a guide wire as far as possible along the splenic vein. The portal pressure was measured and a portal venogram obtained. The major variceal supply veins (left gastric and short gastric) (fig. 1) were selectively catheterised and thrombosed. 30 ml of 500 g/1 dextrose were injected to traumatise the intima of the veins, increasing quantities of human thrombin (500-3000 units) were injected to induce complete thrombosis, and small pieces of gelatin foam were injected to stabilise the thrombus. Contrast medium was then injected to confirm successful obliteration, and if this had been achieved the catheter
carefully withdrawn. The pressure was remeasured in the portal vein and a portal venogram was obtained to ensure complete obliteration of all variceal supply veins and portal-vein patency. Finally, before complete withdrawal of the catheter from the liver, the hepatic puncture wound was plugged by an injection of gelatin foam into the subcapsular parenchyma. was
Fig.
3-Selective catheterisation of left
gastric vein.
Case-report A
twenty-nine-year-old housewife, with two
young children,
Fig. 4-Portal venogram after obliteration of left gastric vein showing incompletely obliterated short gastric vein (white arrow). Fig. 2-Portal venogram showing filling of gastro-cesophageal collaterals (white arrow) and of pancreatic venous plexus (black arrow).
Note contrast trapped in clot tric-vein radicle.
peripherally (black arrow) in left
an
gas-
55
stopped in 7 of the 13 patients. In 5 patients the procedure failed because of technical difficulty in selective catheterisation of small variceal-feeding vessels (4 patients) and because of massive ascites and a very hard liver (1patient). The remaining patient had a previously unsuspected portal-vein block so that the major portal
varices
vein could not be entered. In 3 patients the gallbladder was punctured but there was no biliary leak. Hgemothorax developed in 5 patients and aspiration was required in 2. 2 of these patients also had a haemoperitoneum. 1 recovered after transfusion. The 2nd, however, who also had a haemopehtoneuth and in whom the varices had been successfully thrombosed died. Necropsy showed no variceal bleeding but there was a 1 5 cm tear in the liver capsule. The 5 patients with bleeding complications were treated early in the study and gelatin foam had not been used to plug the exit puncture in the liver. Eight subsequent punctures using gelatin foam have been performed without bleeding. Portal-vein thrombosis did not occur, but in 3 patients strands of thrombin-induced clot were seen in the portal vein. Follow-up with splenic or hepatic venography at three to six months on the 6 successfully treated surviving patients in whom bleeding had been controlled showed that the portal vein was patent in all patients and that in 4 obliteration of varices had persisted. In 1 patient rebleeding occurred from incompletely obliterated varices ten days after the procedure; in the other patient bleeding recurred at three months from a new collateral vein filled from the left gastro-epiploic vein.
Discussion Percutaneous transhepatic portal-vein catheterisation and obliteration of the major variceal supply seems to be an effective procedure for stopping variceal bleeding. The major indication is in patients with severe decompensated cirrhosis, who have not responded to conservative treatment, and in whom, because of liver failure, surgery is contraindicated. After successful obliteration of varices - surgery may be performed electively or referred indefinitely. The long-term effectiveness of variceal obliteration has not been established. In 4 of our patients obliteration has persisted for several months; the longest follow-up period is 9 months: However, Viamonte has (personal communication) treated over 40 patients by this method and in some obliteration of varices has persisted for more than 1 year. -‘ Portal-vein catheterisation allows both fundal and oesophageal varices to be obliterated close to their origin. This is an advantage over oesophagoscopic injection where there is difficulty in obliterating all varices and especially those in the fundus of the stomach. Other approaches to the portal vein such as the translumbar routes and by umbilical-vein catheterisation9 are not as convenient as the transhepatic one. Obliteration of varices has the potential of stopping bleeding for an indefinite period by diversion of portal blood to other collateral vessels. It is well-tolerated and .
,
easily repeated. Portal systemic encephalopathy would not be a complication. In patients with severe cirrhosis, cesophageal transection may have a mortality-rate of more than 50%.’ Our results indicate that mortality from variceal obliteration is much less (1 of the 13 died). The benefit of shunt surgery is controversial since bleed-
ing may be replaced by encephalopathy and the chances of survival are probably not improved. 10 11 The major cause of failure was technical difficulty in catheterising the small variceal-supply veins. The success-rate has, however, improved partly because we have become more practised but mainly because we have used an increased range of guide wires (including moveable core and a variety of "J" shapes). Persistence of obliteration may depend on the venoocclusive agents used. Various combinations of autogenous clot, thrombin-induced clot, sclerosing agents, and gelatin foam matrix have been used.56 12 Initially we used thrombin-induced clot alone. Two of these patients re-bled after successful obliteration. We now prefer thrombin-induced clot stabilised with gelatin foam since this should give a more permanent clot. The complications of hxmothorax and haemoperitoneum (1 fatal) have not occurred since gelatin foam has been used to plug hepatic puncture wounds. Biliary puncture necessitates careful observation for biliary peritonitis. Trauma to the portal vein, pulmonary embolism, and portal-vein thrombosis are theoretical complications not observed so far, although strands of thrombin-induced clot have been seen in the portal vein. Follow-up of these patients by splenic venography showed that the portal vein was patent. Obliteration of the major variceal supply is potentially important in the management of acute variceal bleeding, particularly in patients with end-stage cirrhosis. The long-term role of the technique has yet to be established but it will probably replace both oesophageal transection and obliteration of varices by oesophagoscopy in such patients. We thank Dr Manuel Viamonte, Jr. for his assistance in developing this technique; and all members of the department of radiology who gave their technical assistance. REFERENCES
1.
George, P., Brown, C., Ridgway, G., Crofts, B., Sherlock,
S. Br. J. Surg. 1973, 60, 635. 2. Turcotte, J. G., Lambert, M. J. Surgery, 1973, 73, 810. 3. Crafoord, C., Frenckner, P. Acta Oto-Laryng. 1939, 27, 422. 4. Hunt, P. S., Johnston, G. W., Rodgers, H. W. Br. J. Surg. 1969, 56, 305. 5. Lunderquist, A., Vang, J. New Engl. J. Med. 1974, 291, 646. 6. Viamonte, M. Jr., Page, J., Lunderquist, A., Pereiras, R., Russell, E., Viamonte, M., Camacho, M. Radiology, 1975, 114, 457. 7. Kennedy, J. Nutrition, 1973, 27, 325. 8. Haverling, M., Overfors, C-O. Invest. Radiol. 1968, 3, 376. 9. Kessler, R. E., Tice, D. A., Zimmon, D. S. Surg. Gynec. Obstet. 1973, 136,
529. 10. Resnick, R. H., Iber, F. L., Ishihara, A. M., Chalmers, T. C., Zimmerman, H. Gastroenterology, 1974, 67, 843. 11. Rueff, B., Prandi, D., Degos, F., Sicot, J., Degos, J-D., Sicot, C., Maillard, J-N., Fauvert, R., Benhamou, J-P. Lancet, 1976, i, 655. 12. Reuter, S. R., Chuany, V. P., Bree, R. L., Major, M. C., Br. J. Radiol. 1975,
125, 119.
"Monopoly bargaining is a reality of the times. Nevertheless there is such a thing as being too well organised. For a profession can become so autonomous that it moves into a phase where it begins altogether to discount the experiences and evaluations of the public outside. This can mount to a point where wide community criticism threatens to reduce autonomy and thus remove some liberty. The professional symptoms which attract such criticism of an organised profession are: secret behaviour; professional colleagues protecting one another against all outside criticism; overt claims to exclusiveness or social rank; insistence on very long training without clear demonstration of its necessity; very highly conditioned loyalties and behaviour within the profession; claims to omniscience in human matters or to a monopoly of technical knowledge; and active resistance to change in society."-EARLE HACKETT, South Australian Climes, 1975, 7, 3.