- Email: [email protected]
Portal Hypertension: The Treatment by Splenectomy and Splenorenal Anastomosis with Preservation of the Kidney
PORTAL HYPERTENSION The Treatment by Splenectomy and Splenorenal Anastomosis with Preservation of the Kidney
ROBERT R. LINTON, M.D., F.A.C.S.,# CHESTER M. JONES, M.D., F.A.C.P.f AND WADE VOLWILER, M.D.t HEMORRHAGE from esophageal varices secondary to a portal bed block is a condition which has ballied the medical profession for many years. The recent articles on portacaval and splenorenal shunts by Whipple6 and Blakemore and Lord, l to reduce the portal hypertension in these cases, have stimulated a renewed interest in the subject. They have demonstrated the feasibility of shunting the portal blood How directly into the systemic venous circulation by either a portacaval or a splenorenal anastomosis, thereby reducing the blood How through the collateral venous channels in the esophagus and at the same time lowering the portal venous pressure. It is accepted generally that the esophageal varices are secondary to a partial block in the portal venous system. They represent collateral channels whereby the portal blood returns to the systemic venous circulation. The normal venous pressure in the portal system with the patient supine is 8 to 10 cm. of water, which is considerably higher than in the systemic venous system. Due to the portal bed block in patients with esophageal varices the portal venous pressure is greatly elevated. Any reading above 11 cm. of water, as pointed out by Whipple,6 probably indicates a block in the portal system. The highest value in our series is 49 cm. and the lowest is 23; the mean value is 36.4 cm. Types of Portal Bed Block.-Portal bed block may be one of the two types, as pointed out by Whipple, 6 or a combination of both. Statistics are not yet available to estimate the relative frequency of the various types. First, and probably the most common, is intrahepatic in origin. It results from portal or Laennec's type of cirrhosis. The How of blood to the liver is obstructed secondary to the scar tissue replacement of the liver parenchyma. It probably is never completely occluded so that all the portal How is forced through the collateral From the Departments of Surgery and Medicine, Massachusetts General Hospital, Boston. # Visitiug Surgeon, Massachusetts General Hospital; Associate in Surgery, Harvard Medical School. t Physician, Massachusetts General Hospital; Clinical Professor of Medicine, Harvard Medical School. Assistant in Medicine, Massachusetts General ,Hospital.
*
1162
PORTAL HYPERTENSION
1163
channels. However, a very large portion of it may be shunted. In this type of block the blood flow of the hepatic artery is added to that of the portal vein which passes through the collateral vessels. This opinion is based on the perfusion experiments on normal and cirrhotic livers by Herrick3 and later checked by one of the authors (R. R. L.). Herrick found that if the hepatic artery of a normal liver is perfused alone with the manometer attached to the portal vein that there is a 1 mm. rise in mercury for every 40 mm. rise in the hepatic artery pressure, while in the cirrhotic liver there is a' 1 mm. rise for every 6 in the hepatic artery. He found in addition that the normal portal pressure was not affected until the arterial pressure was elevated over 100 mm., while in the cirrhotic the portal pressure was affected as low as 30 mm. The second type of portal bed block is extrahepatic. The obstruction is in the portal venous system before it enters the liver, which as a rule is normal from a morphological viewpoint. The cause of the block is an occlusion of the portal vein or its main tributaries by fibrous or scar tissue which may be (a) congenital, or (b) acquired. As Whipple6 has pointed out, the former is believed to occur from the extension into the portal system of the obliterative process that takes place after birth in the umbilical vein and ductus venosus. This appears to be a plausible explanation for the occurrence of portal bed block with esophageal varices and splenomegaly seen in young children, commonly diagnosed as Banti's disease or syndrome. The acquired form occurs probably as a result of venous thrombosis of the portal vein or one of its main tributaries. This process may develop in the splenic vein as was pointed out by Warthin 5 in 1910. The thrombosis may progress proximally to involve the portal vein itself so that a main portal bed block develops. It is also within the realm of possibility that thrombi may break off from the splenic or other tributaries of the portal vein to produce portal embolism of the veins entering the, liver. The etiology of portal venous thrombosis is less well understood than that which occurs in the lower extremities, but that it exists there is not a question of doubt. It probably is the result of inflammation, trauma, or it may be spontaneous, developing as a result of phlebosclerosis. The thrombus undergoes organization and partial canalization but a permanent obstruction to the portal blood flow is produced, which results in a persistent portal hypertension. A combination of the two main types, intrahepatic and extraheptic, may occur in the same case. Diagnosis.-A sudden massive hematemesis or severe melena is usually the first indication that a patient has a portal bed block. If roentgenological examination of the esophagus with barium shows esophageal varices the diagnosis is proven, since no other condition produces these anomalous blood vessels. Further confirmatory evidence
1164
R. R. LINTON'IC, M. JONES, W: VOLWILER
as to the diagnosis is the presence of an enlarged spleen and blood examinations which show a secondary anemia, a leukopenia and a thrombocytopenia. The site of the portal bed block can usually be determined by liver function tests. A high retention of bromsulfalein after the intravenous injection of this drug, a reversal of the albuminglobulin ratio, an elevated prothrombin time, and a positive cephalin flocculation test indicate usually an intrahepatic block, whereas· if these liver function tests are nom1al the block is most likely extrahepatic. Choice of Operation.-There is some question yet as to which shunt operation is the best. Whipple,6 Blakemore and Lordl and Blalock2 seem to favor a portacaval type, since this produces the most complete shunt of the portal blood and at the same time a maximum reduction of the portal hypertension. Our opinion on the other hand is that ·a splenectomy with an end-to-side splenorenal shunt preserving the kidney is the better procedure. This type of operation serves two purposes. First, it removes the spleen and thereby reduces the portal blood flow by as much as 40 per cent, if there is a marked splenomegaly. Second, it p~rmits an .anastomosis to be performed between the cut end of the splenic vein and the side of the renal vein without sacrificing the kidney. This shunt will reduce the hypertension by permitting some of the portal blood flow to enter directly into the systemic venous circulation but at the same time it will not cause all the blood to by-pass the liver as in the case of a portacaval shunt. It has seemed to us in view of the nutritional difficulties encountered in animals in which a complete Eck fistula has been produced that the complete shunting of the portal blood by portacaval anastomosis should be avoided if the other type will reduce satisfactorily the portal hypertension. REPORT OF FIVE PATIENTS TREATED BY SPLENECTOMY AND SPLENORENAL ANASTOMOSIS WITH PRESERVATION OF THE KIDNEY
The purpose of this paper is to report a group of five patients with extrahepatic portal bed block which have been treated by splenectomy and an end-to-side suture anastomosis between the splenic and the left renal veins with preservation of the kidney, utilizing a thoraco~ abdominal incision. The youngest patient was 6 years of age, the oldest 65 and the others, 23, 30 and 45 years. Whipple6 and· Blakemdte and Lordl reported a group of patients treated with splene.ctomY, Ii:ft nephrectomy and an end-to-end nonsuture type of sple~orenal anastomosis utilizing the Blakemore vitallium tube. The technic of the operation described in this report eliminates the necessity of sacrificing the kidney. It is believed also that the end-to-s,ide anastomosis is less likely to become blocked from thrombosis than the end-to-end
i.
!
PORTAL HYPERTENSION
1165
type, since the blood from the kidney flowing through the renal vein has a suction effect on the blood entering it from the splenic vein, thereby helping to maintain the patency of the anastomosis. Blalock2 reports experimental evidence to support this opinion. A suture anastomosis was performed in these cases instead of the nonsuture type. Technic· of Operation.-The patient is placed on the operating table lying on the right side· so that his back is at about a 60 degree angle with the table. The anesthetic is ether administered through an endotracheal tube so that positive pressure may be maintained· with an open chest and also to facilitate the administration of oxygen as well as the anesthetic. A vein in the left arm is cannulated before the intravenous administration of fluids and blood during the operation. A large bore needle such as a 16 gauge is preferable so that blood may be administered rapidly since severe and sometimes rapid hemorrhage may occur during the operative procedure because of the tremendous collateral blood supply encountered in the operative field. An additional precaution which has been found to be of great value in some of the cases is to have autotransfusion sets available in order that any large amount of blood which may escape during the operation can be saved and utilized for transfusion purposes. In one of the patients reported (No.2) twelve transfusions were given during the operation, six of these being autotransfusions. A thoracoabdominal type of incision, first used by Sweet4 in this type of case, is made over the tenth left rib from its angle posteriorly, extending anteriorly through the abdominal wall to the midline cutting across the external, internal oblique and rectus muscles. The tenth rib is resected subperiosteally and the pleural and peritoneal cavities are opened. The anterior portion of the diaphragm is divided in the line parallel to the incision to permit a direct exposure of the posterior attachments of the spleen. Once the incision has been completed the spleen if enlarged practically fills the operative wound. It is retracted forward to expose its posterior peritoneal attachments, the lienorenal and the phrenicocolic ligaments. These structures frequently contain enormous venous collateral channels so that they are best doubly clamped before dividing them. The gastrolienal ligament and the vasa brevia are divided next, leaving the spleen attached only by its pedicle. The dissection in this order is important since it gives the best access to the splenic vein, which if injured, may jeopardize this structure so that a satisfactory length of it may not be available for the venous anastomosis. The splenic artery is isolated, then divided and ligated proximal to the point it divides at the hilum. The spleen is gently compressed to evacuate as much blood from it as possible before dividing the splenic vein. This vessel is ligated and divided as close to the hilum as possible so that sufficient length of if=\Vill be available to make the
1166
R. R. LINTON, C. M. JONES, W. VOLWILER
anastomosis. The splenic flexure of the colon is retracted downward after the removal of the spleen and the tail of the pancreas is displaced ,anteriorward. The peritoneum over the kidney is incised and this organ is completely mobilized from the perirenal tissues. The renal artery is isolated proximal to the point it bifurcates so that a small bulldog arterial clamp may be placed on it to control the arterial inflow to the kidney while performing the venous anastomosis. This is important, otherwise the kidney may be damaged from the high venous back pressure while the renal vein is occluded.
Fig. 360.-Schematic drawing to show the exposure of the (1) splenic vein, (2) renal artery, (3) renal vein, (4) left kidney. The renal artery has been clamped with a bulldog arterial clamp and the renal vein with two tourniquet clamps. The ligatured end of the splenic vein is resected after applying a bulldog clamp to it, which controls the blood flow while performing the anastomosis.
Sufficient length of the renal vein must be dissected next to allow the application of two clamps so that there is enough space between them to permit the venous anastomosis to be performed. In some cases this has necessitated dividing the spermatic or ovarian and the adrenal vein; while in others it may not be necessary. The clamps for the control of the renal vein that have been used are a modification of the Bethune lung tourniquet clamp. (Fig. 360). Since the splenic vein lies slightly cephalad and anterior to the renal vein the anastomosis is performed between the end of the splenic vein and the anterosuperior surface of the renal vein. A h'ansverse incision is made in the
,
PORTAL HYPERTENSION
1167
latter; the length of it is judged so that it will approximately equal the diameter of the splenic vein. The anastomosis is performed with No. 5-0 braided silk on a No.9 needle. A continuous mattress type of suture is used, everting the cut edges of the blood vessel walls so that intima is approximated to intima. The suture line is interrupted at three equidistant points by tying it to stay sutures in order to prevent narrowing of the anastomosis lumen from a purse-string effect. The clamps on the renal vein and the one on the splenic vein are first removed after completion of the anastomosis. The hemostasis of the suture line is determined following this step. Occasionally a bleeding point may need a single suture to control it. The patency of the anastomosis may also be observed. Finally the clamp on the renal artery is removed (Fig. 361).
Fig. 361.-Schematic drawing to show the completed end-to-side splenorenal anastomosis. All the clamps have been removed. 1, Splenic vein, 2, renal artery, 3, suprarenal vein, 4, spermatic vein, 5, renal vein, 6, left kidney.
The kidney is then placed back in its fossa and the peritoneum over it resutured to maintain it in its normal position. The diaphragm and the thoraco-abdominal incision are closed in layers with interrupted nonabsorbable sutures of cotton or silk. Venous pressures are taken with a water manometer by cannulating one of the veins in the greater omentum before carrying out the splenectomy. A marked elevation above the normal was found in the cases reported in this article. The measurement is repeated after splenectomy to see the effect produced by the reduction of the portal blood inflow after ligature of the splenic artery. A final reading is also obtained after completion of the splenorenal anastomosis. Comment.-The results of this form of treatment to date in this group of patients have been very encouraging. The mortality rate has
...... ......
0:. 00
~ TABLE
1
~
PORTAL HYPER1E:\SIO:\: FIYE CASES OF EXTRAHEPATIC BU)CK TREATED BY SPLENECTOMY AND SPLENORENAL ANASTOMOSIS WITH PRESERVATION O~' THE KIDNEY
Portal Pressure in Centimeters of Water No.
Patient
2
R.M.M. J.M.
Duration of Symptom
Previous Treatment Initial
Postsplenectomy
Postanastomosis
49.0
....
21.5
33.5
....
....
__ I
;
1
Initial Symptom
Sex
Age
6 30
I
F ).f
IHenlatemesis Hematemesis ;
10 wks. 11 yrs.
None 1935-Ligation splenic artery in continuity and omentopexy
Time since Date Operation of Operation without Bleeding
6/4/46
10 mos.
"
3 4 S
F.W.B. .M. J. S.
65 . 45 23
:\1 F M
I
II H ema.,lIle8Js '" . ematj)mesis
~ematelnesis
1 day 5 mos. 6 yrs.
.... 23.0 34.5 42.0
31.0 23.0
i.i:&
26.0 24.0
~
6/8/46
6/8/46 1/11/47 2/7/47 3/13/47
~t"l
S"
~
veins
1942-Transthoracic ligation periesophageal varices 1946-Portacaval anastomosis attempted. Cholecystectomy and choledochojejunostomy None None None
p '-4
1937-Ligation left phrenic and right coronary
I
~
v~
10 mos. 3 mos. 2 mos. 4wks
I
PORTAL HYPERTENSION
1169
been zero; none of the patients has bled from the esophagus or the gastrointestinal tract since the operation except one patient (No.3) who had a hematemesis on the first and fourth postoperative days. This may have been due to an inlying nasal tube that was placed in the stomach prior to the operation by error and removed at the end of twenty-four hours. It is still too early to be sure that all dangers of bleeding have been eliminated in these patients since the longest follow-up is only ten months in two cases. The patients will have to be followed for a number of years to be sure that the portal pressure has been reduced permanently so that the load has been taken off the esophageal veins. In the cases that have had portal pressures taken before and after the splenectomy and splenorenal anastomosis there has been a gratifying drop in the level (Table 1). No deleterious effect to the kidney in any of the patients has been demonstrable as the result of interrupting its blood supply while performing the splenorenal anastomosis. The longest period of interruption was thirty-five minutes. An interesting observation has been made in this group of patients to the effect that portal hypertension per se is not the cause of ascites, as none of the patients in this group had ascites at the. time of operation. The anticoagulant heparin was Ilsed in the form of Pitkin Menstruum heparin in one case (No.1). Despite the fact that careful coagulation time determinations were made at frequent intervals, this patient developed a marked hemoperitoneum and hemothorax. It is believed that due to the magnitude of the operation and the tremendous number of blood vessels encountered in the operative procedure, anticoagulants should not be used despite the fact there would be a better chance that the anastomosis would remain patent if they were used. Studies are planned in the future to try to determine whether the anastomosis remains patent. The only evidence we have at present that they are still functioning is that none of the patients has developed bleeding since discharge from the hospital. CONCLUSIONS
1. Hemorrhage from. esophageal varices is secondary to a portal bed block and an associated portal hypertension. The· block may be either intrahepatic or extrahepatic. 2. Five patients with an extrahepatic type of block are reported in whom a;splene¢tomy and an end-to-side splenorenal anastomosis with prese\vation of the kidney have been performed. 3. The results of the surgical treatment have been encouraging since none of the patients has bled from the gastrointestinal tract since discharge from the hospital. 4. The longest follow-up is ten months in two patients, so it is too early yet to claim permanent relief.
1170
R. R. LINTON, C. M. JONES, W. VOLWILER
REFERENCES 1. Blakemore, A. H. and Lord, J. W., Jr.: The Technic of Using Vitallium Tuhcs in Establishing Portacaval Shunts for Portal Hypertension. Ann. Surg., 122:476-489 (Oct.) 1945. 2. Blalock, A.: The Use of Shunt or By-pass Operations in the Treatment of Certain Circulatory Disorders, including Portal Hypertension and PI\lmonie Stenosis. Ann. Surg., 125:129-141 (Feb.) 1947. 3. Herrick, F. c.: An Experimental Study into the Cause of the Increased Portal Pressure in Portal Cirrhosis. J. Exper. Med., 9:93-104, 1907. 4. Sweet, R. H.: Personal communication. 5. Warthin, A. S.: The Relation of Thrombophlebitis of the Portal and Splenic Veins in Splenic Anemia and Banti's Disease. Internat. Clin., 4: (20th series): 189-226, 1910. 6. Whipple, A. 0.: The Problem of Portal Hypertension in Relation to Hepatosplenopathies. Ann. Surg., 122:449-475 (Oct.) 1945.
ROBERT R. LINTON, M.D., F.A.C.S.,# CHESTER M. JONES, M.D., F.A.C.P.f AND WADE VOLWILER, M.D.t HEMORRHAGE from esophageal varices secondary to a portal bed block is a condition which has ballied the medical profession for many years. The recent articles on portacaval and splenorenal shunts by Whipple6 and Blakemore and Lord, l to reduce the portal hypertension in these cases, have stimulated a renewed interest in the subject. They have demonstrated the feasibility of shunting the portal blood How directly into the systemic venous circulation by either a portacaval or a splenorenal anastomosis, thereby reducing the blood How through the collateral venous channels in the esophagus and at the same time lowering the portal venous pressure. It is accepted generally that the esophageal varices are secondary to a partial block in the portal venous system. They represent collateral channels whereby the portal blood returns to the systemic venous circulation. The normal venous pressure in the portal system with the patient supine is 8 to 10 cm. of water, which is considerably higher than in the systemic venous system. Due to the portal bed block in patients with esophageal varices the portal venous pressure is greatly elevated. Any reading above 11 cm. of water, as pointed out by Whipple,6 probably indicates a block in the portal system. The highest value in our series is 49 cm. and the lowest is 23; the mean value is 36.4 cm. Types of Portal Bed Block.-Portal bed block may be one of the two types, as pointed out by Whipple, 6 or a combination of both. Statistics are not yet available to estimate the relative frequency of the various types. First, and probably the most common, is intrahepatic in origin. It results from portal or Laennec's type of cirrhosis. The How of blood to the liver is obstructed secondary to the scar tissue replacement of the liver parenchyma. It probably is never completely occluded so that all the portal How is forced through the collateral From the Departments of Surgery and Medicine, Massachusetts General Hospital, Boston. # Visitiug Surgeon, Massachusetts General Hospital; Associate in Surgery, Harvard Medical School. t Physician, Massachusetts General Hospital; Clinical Professor of Medicine, Harvard Medical School. Assistant in Medicine, Massachusetts General ,Hospital.
*
1162
PORTAL HYPERTENSION
1163
channels. However, a very large portion of it may be shunted. In this type of block the blood flow of the hepatic artery is added to that of the portal vein which passes through the collateral vessels. This opinion is based on the perfusion experiments on normal and cirrhotic livers by Herrick3 and later checked by one of the authors (R. R. L.). Herrick found that if the hepatic artery of a normal liver is perfused alone with the manometer attached to the portal vein that there is a 1 mm. rise in mercury for every 40 mm. rise in the hepatic artery pressure, while in the cirrhotic liver there is a' 1 mm. rise for every 6 in the hepatic artery. He found in addition that the normal portal pressure was not affected until the arterial pressure was elevated over 100 mm., while in the cirrhotic the portal pressure was affected as low as 30 mm. The second type of portal bed block is extrahepatic. The obstruction is in the portal venous system before it enters the liver, which as a rule is normal from a morphological viewpoint. The cause of the block is an occlusion of the portal vein or its main tributaries by fibrous or scar tissue which may be (a) congenital, or (b) acquired. As Whipple6 has pointed out, the former is believed to occur from the extension into the portal system of the obliterative process that takes place after birth in the umbilical vein and ductus venosus. This appears to be a plausible explanation for the occurrence of portal bed block with esophageal varices and splenomegaly seen in young children, commonly diagnosed as Banti's disease or syndrome. The acquired form occurs probably as a result of venous thrombosis of the portal vein or one of its main tributaries. This process may develop in the splenic vein as was pointed out by Warthin 5 in 1910. The thrombosis may progress proximally to involve the portal vein itself so that a main portal bed block develops. It is also within the realm of possibility that thrombi may break off from the splenic or other tributaries of the portal vein to produce portal embolism of the veins entering the, liver. The etiology of portal venous thrombosis is less well understood than that which occurs in the lower extremities, but that it exists there is not a question of doubt. It probably is the result of inflammation, trauma, or it may be spontaneous, developing as a result of phlebosclerosis. The thrombus undergoes organization and partial canalization but a permanent obstruction to the portal blood flow is produced, which results in a persistent portal hypertension. A combination of the two main types, intrahepatic and extraheptic, may occur in the same case. Diagnosis.-A sudden massive hematemesis or severe melena is usually the first indication that a patient has a portal bed block. If roentgenological examination of the esophagus with barium shows esophageal varices the diagnosis is proven, since no other condition produces these anomalous blood vessels. Further confirmatory evidence
1164
R. R. LINTON'IC, M. JONES, W: VOLWILER
as to the diagnosis is the presence of an enlarged spleen and blood examinations which show a secondary anemia, a leukopenia and a thrombocytopenia. The site of the portal bed block can usually be determined by liver function tests. A high retention of bromsulfalein after the intravenous injection of this drug, a reversal of the albuminglobulin ratio, an elevated prothrombin time, and a positive cephalin flocculation test indicate usually an intrahepatic block, whereas· if these liver function tests are nom1al the block is most likely extrahepatic. Choice of Operation.-There is some question yet as to which shunt operation is the best. Whipple,6 Blakemore and Lordl and Blalock2 seem to favor a portacaval type, since this produces the most complete shunt of the portal blood and at the same time a maximum reduction of the portal hypertension. Our opinion on the other hand is that ·a splenectomy with an end-to-side splenorenal shunt preserving the kidney is the better procedure. This type of operation serves two purposes. First, it removes the spleen and thereby reduces the portal blood flow by as much as 40 per cent, if there is a marked splenomegaly. Second, it p~rmits an .anastomosis to be performed between the cut end of the splenic vein and the side of the renal vein without sacrificing the kidney. This shunt will reduce the hypertension by permitting some of the portal blood flow to enter directly into the systemic venous circulation but at the same time it will not cause all the blood to by-pass the liver as in the case of a portacaval shunt. It has seemed to us in view of the nutritional difficulties encountered in animals in which a complete Eck fistula has been produced that the complete shunting of the portal blood by portacaval anastomosis should be avoided if the other type will reduce satisfactorily the portal hypertension. REPORT OF FIVE PATIENTS TREATED BY SPLENECTOMY AND SPLENORENAL ANASTOMOSIS WITH PRESERVATION OF THE KIDNEY
The purpose of this paper is to report a group of five patients with extrahepatic portal bed block which have been treated by splenectomy and an end-to-side suture anastomosis between the splenic and the left renal veins with preservation of the kidney, utilizing a thoraco~ abdominal incision. The youngest patient was 6 years of age, the oldest 65 and the others, 23, 30 and 45 years. Whipple6 and· Blakemdte and Lordl reported a group of patients treated with splene.ctomY, Ii:ft nephrectomy and an end-to-end nonsuture type of sple~orenal anastomosis utilizing the Blakemore vitallium tube. The technic of the operation described in this report eliminates the necessity of sacrificing the kidney. It is believed also that the end-to-s,ide anastomosis is less likely to become blocked from thrombosis than the end-to-end
i.
!
PORTAL HYPERTENSION
1165
type, since the blood from the kidney flowing through the renal vein has a suction effect on the blood entering it from the splenic vein, thereby helping to maintain the patency of the anastomosis. Blalock2 reports experimental evidence to support this opinion. A suture anastomosis was performed in these cases instead of the nonsuture type. Technic· of Operation.-The patient is placed on the operating table lying on the right side· so that his back is at about a 60 degree angle with the table. The anesthetic is ether administered through an endotracheal tube so that positive pressure may be maintained· with an open chest and also to facilitate the administration of oxygen as well as the anesthetic. A vein in the left arm is cannulated before the intravenous administration of fluids and blood during the operation. A large bore needle such as a 16 gauge is preferable so that blood may be administered rapidly since severe and sometimes rapid hemorrhage may occur during the operative procedure because of the tremendous collateral blood supply encountered in the operative field. An additional precaution which has been found to be of great value in some of the cases is to have autotransfusion sets available in order that any large amount of blood which may escape during the operation can be saved and utilized for transfusion purposes. In one of the patients reported (No.2) twelve transfusions were given during the operation, six of these being autotransfusions. A thoracoabdominal type of incision, first used by Sweet4 in this type of case, is made over the tenth left rib from its angle posteriorly, extending anteriorly through the abdominal wall to the midline cutting across the external, internal oblique and rectus muscles. The tenth rib is resected subperiosteally and the pleural and peritoneal cavities are opened. The anterior portion of the diaphragm is divided in the line parallel to the incision to permit a direct exposure of the posterior attachments of the spleen. Once the incision has been completed the spleen if enlarged practically fills the operative wound. It is retracted forward to expose its posterior peritoneal attachments, the lienorenal and the phrenicocolic ligaments. These structures frequently contain enormous venous collateral channels so that they are best doubly clamped before dividing them. The gastrolienal ligament and the vasa brevia are divided next, leaving the spleen attached only by its pedicle. The dissection in this order is important since it gives the best access to the splenic vein, which if injured, may jeopardize this structure so that a satisfactory length of it may not be available for the venous anastomosis. The splenic artery is isolated, then divided and ligated proximal to the point it divides at the hilum. The spleen is gently compressed to evacuate as much blood from it as possible before dividing the splenic vein. This vessel is ligated and divided as close to the hilum as possible so that sufficient length of if=\Vill be available to make the
1166
R. R. LINTON, C. M. JONES, W. VOLWILER
anastomosis. The splenic flexure of the colon is retracted downward after the removal of the spleen and the tail of the pancreas is displaced ,anteriorward. The peritoneum over the kidney is incised and this organ is completely mobilized from the perirenal tissues. The renal artery is isolated proximal to the point it bifurcates so that a small bulldog arterial clamp may be placed on it to control the arterial inflow to the kidney while performing the venous anastomosis. This is important, otherwise the kidney may be damaged from the high venous back pressure while the renal vein is occluded.
Fig. 360.-Schematic drawing to show the exposure of the (1) splenic vein, (2) renal artery, (3) renal vein, (4) left kidney. The renal artery has been clamped with a bulldog arterial clamp and the renal vein with two tourniquet clamps. The ligatured end of the splenic vein is resected after applying a bulldog clamp to it, which controls the blood flow while performing the anastomosis.
Sufficient length of the renal vein must be dissected next to allow the application of two clamps so that there is enough space between them to permit the venous anastomosis to be performed. In some cases this has necessitated dividing the spermatic or ovarian and the adrenal vein; while in others it may not be necessary. The clamps for the control of the renal vein that have been used are a modification of the Bethune lung tourniquet clamp. (Fig. 360). Since the splenic vein lies slightly cephalad and anterior to the renal vein the anastomosis is performed between the end of the splenic vein and the anterosuperior surface of the renal vein. A h'ansverse incision is made in the
,
PORTAL HYPERTENSION
1167
latter; the length of it is judged so that it will approximately equal the diameter of the splenic vein. The anastomosis is performed with No. 5-0 braided silk on a No.9 needle. A continuous mattress type of suture is used, everting the cut edges of the blood vessel walls so that intima is approximated to intima. The suture line is interrupted at three equidistant points by tying it to stay sutures in order to prevent narrowing of the anastomosis lumen from a purse-string effect. The clamps on the renal vein and the one on the splenic vein are first removed after completion of the anastomosis. The hemostasis of the suture line is determined following this step. Occasionally a bleeding point may need a single suture to control it. The patency of the anastomosis may also be observed. Finally the clamp on the renal artery is removed (Fig. 361).
Fig. 361.-Schematic drawing to show the completed end-to-side splenorenal anastomosis. All the clamps have been removed. 1, Splenic vein, 2, renal artery, 3, suprarenal vein, 4, spermatic vein, 5, renal vein, 6, left kidney.
The kidney is then placed back in its fossa and the peritoneum over it resutured to maintain it in its normal position. The diaphragm and the thoraco-abdominal incision are closed in layers with interrupted nonabsorbable sutures of cotton or silk. Venous pressures are taken with a water manometer by cannulating one of the veins in the greater omentum before carrying out the splenectomy. A marked elevation above the normal was found in the cases reported in this article. The measurement is repeated after splenectomy to see the effect produced by the reduction of the portal blood inflow after ligature of the splenic artery. A final reading is also obtained after completion of the splenorenal anastomosis. Comment.-The results of this form of treatment to date in this group of patients have been very encouraging. The mortality rate has
...... ......
0:. 00
~ TABLE
1
~
PORTAL HYPER1E:\SIO:\: FIYE CASES OF EXTRAHEPATIC BU)CK TREATED BY SPLENECTOMY AND SPLENORENAL ANASTOMOSIS WITH PRESERVATION O~' THE KIDNEY
Portal Pressure in Centimeters of Water No.
Patient
2
R.M.M. J.M.
Duration of Symptom
Previous Treatment Initial
Postsplenectomy
Postanastomosis
49.0
....
21.5
33.5
....
....
__ I
;
1
Initial Symptom
Sex
Age
6 30
I
F ).f
IHenlatemesis Hematemesis ;
10 wks. 11 yrs.
None 1935-Ligation splenic artery in continuity and omentopexy
Time since Date Operation of Operation without Bleeding
6/4/46
10 mos.
"
3 4 S
F.W.B. .M. J. S.
65 . 45 23
:\1 F M
I
II H ema.,lIle8Js '" . ematj)mesis
~ematelnesis
1 day 5 mos. 6 yrs.
.... 23.0 34.5 42.0
31.0 23.0
i.i:&
26.0 24.0
~
6/8/46
6/8/46 1/11/47 2/7/47 3/13/47
~t"l
S"
~
veins
1942-Transthoracic ligation periesophageal varices 1946-Portacaval anastomosis attempted. Cholecystectomy and choledochojejunostomy None None None
p '-4
1937-Ligation left phrenic and right coronary
I
~
v~
10 mos. 3 mos. 2 mos. 4wks
I
PORTAL HYPERTENSION
1169
been zero; none of the patients has bled from the esophagus or the gastrointestinal tract since the operation except one patient (No.3) who had a hematemesis on the first and fourth postoperative days. This may have been due to an inlying nasal tube that was placed in the stomach prior to the operation by error and removed at the end of twenty-four hours. It is still too early to be sure that all dangers of bleeding have been eliminated in these patients since the longest follow-up is only ten months in two cases. The patients will have to be followed for a number of years to be sure that the portal pressure has been reduced permanently so that the load has been taken off the esophageal veins. In the cases that have had portal pressures taken before and after the splenectomy and splenorenal anastomosis there has been a gratifying drop in the level (Table 1). No deleterious effect to the kidney in any of the patients has been demonstrable as the result of interrupting its blood supply while performing the splenorenal anastomosis. The longest period of interruption was thirty-five minutes. An interesting observation has been made in this group of patients to the effect that portal hypertension per se is not the cause of ascites, as none of the patients in this group had ascites at the. time of operation. The anticoagulant heparin was Ilsed in the form of Pitkin Menstruum heparin in one case (No.1). Despite the fact that careful coagulation time determinations were made at frequent intervals, this patient developed a marked hemoperitoneum and hemothorax. It is believed that due to the magnitude of the operation and the tremendous number of blood vessels encountered in the operative procedure, anticoagulants should not be used despite the fact there would be a better chance that the anastomosis would remain patent if they were used. Studies are planned in the future to try to determine whether the anastomosis remains patent. The only evidence we have at present that they are still functioning is that none of the patients has developed bleeding since discharge from the hospital. CONCLUSIONS
1. Hemorrhage from. esophageal varices is secondary to a portal bed block and an associated portal hypertension. The· block may be either intrahepatic or extrahepatic. 2. Five patients with an extrahepatic type of block are reported in whom a;splene¢tomy and an end-to-side splenorenal anastomosis with prese\vation of the kidney have been performed. 3. The results of the surgical treatment have been encouraging since none of the patients has bled from the gastrointestinal tract since discharge from the hospital. 4. The longest follow-up is ten months in two patients, so it is too early yet to claim permanent relief.
1170
R. R. LINTON, C. M. JONES, W. VOLWILER
REFERENCES 1. Blakemore, A. H. and Lord, J. W., Jr.: The Technic of Using Vitallium Tuhcs in Establishing Portacaval Shunts for Portal Hypertension. Ann. Surg., 122:476-489 (Oct.) 1945. 2. Blalock, A.: The Use of Shunt or By-pass Operations in the Treatment of Certain Circulatory Disorders, including Portal Hypertension and PI\lmonie Stenosis. Ann. Surg., 125:129-141 (Feb.) 1947. 3. Herrick, F. c.: An Experimental Study into the Cause of the Increased Portal Pressure in Portal Cirrhosis. J. Exper. Med., 9:93-104, 1907. 4. Sweet, R. H.: Personal communication. 5. Warthin, A. S.: The Relation of Thrombophlebitis of the Portal and Splenic Veins in Splenic Anemia and Banti's Disease. Internat. Clin., 4: (20th series): 189-226, 1910. 6. Whipple, A. 0.: The Problem of Portal Hypertension in Relation to Hepatosplenopathies. Ann. Surg., 122:449-475 (Oct.) 1945.