- Email: [email protected]
Emergency portacaval shunt
Emergency
Portacaval
WILLIAM P. MIKKELSEN, M.D., Los Angeles, Pasadena,
From the Department of Surgery, University oJ Soulbern CaliJornia, School of Medicine, Los Angeles, California.
T esophagogastric
HE management
of acute bleeding from varices in the patient with hepatic cirrhosis continues to be a most chalIenging problem. When current conservative measures are used approximateIy one-haIf of these patients succumb. This has Ied severa surgeons to explore the feasibility of operative control of bleeding. It has been recognized, however, that any surgica1 procedure of magnitude will not be tolerated by the cirrhotic patient in whom severe hepatrc decompensation deveIops. Such a patient must be managed by conservative means, regardless of the persistence of uncontrohed bIeeding. These are the patients who are mainIy responsible for the continued high mortality from acute bleeding. In contrast to the foregoing situation there is the patient whose hepatic reserve is good, whose bleeding ceases spontaneousIy or is quickly controlled by conservative means. Here an elective operative procedure may be carried out at a convenient time. This is the group of patients who, in any reported series, survive their bIeeding episode. Between the two extremes just described there lies a third group of patients for whom an emergency operative procedure may prove Iifesaving. These patients initiahy remain in reasonably good condition, but if bleeding continues or recurs, they may sink into Iethal hepatic decompensation or cssanguinate. This group has not been clearly delineated but, in all likelihood, it is a considerablJ- smaller group than either of the other t \vo. CURRENT
EMERGENCY
PROCEDURES
Experience with operative intervention during active varical bIeeding has been limited but has been sufficient to indicate that certain pro-
Calijornia,
Shunt
AND ARTHUR C. PATTISON, M.D.,
Calijornia
cedures cannot be reIied upon to terminate bIeeding. Into this category faI1 such operations as spIenectomy, coronary vein ligation, esophageal devascularization or ligation of one or al1 of the branches of the celiac axis. WhiIe eIective portacava1 shunt is now generally accepted as the best means of eliminating recurrent bIeeding from varices, its effectiveness in controIIing acute bIeeding has remained undetermined. Furthermore, this operation has been considered to be too formidable for these acuteIy bIeeding patients. Consequently, there have evolved less formidabIe procedures designed to control bleeding temporarily and to be followed later by a more extensive definitive procedure for permanent contro1. Direct transesophageal suture of the bleeding varix, usuatly together with suture of the remainder of the visibIe varices in the lower esophagus and cardia, has received the greatest attention. Linton and Ellis [I] reported on twenty patients treated in this manner hy a transthoracic approach. Three patients died immediately postoperatively and one died live weeks later following partial gastrectomv for recurrent varicaI bleeding. Another patient died fifteen months Iater from hepatic failure. The remaining fifteen patients had portacaval shunts of some type constructed later. The subsequent course of these latter patients was not described. Reference is not made as to the number of patients who re-bIed before a definitive procedure couId be carried out, although in a previous communication [2] five of nine survivers did re-bIeed. Linton now rccommends immediate tramesophageal vat-ix ligation in any patient who bleeds severely enough to require balloon tamponade. He admits, however, that he refuses to operate upon any patient who exhibits signs of “impending liver faiIure.” Hamihon’s [j] results with transthoracic
Mikkelsen varix Iigation have been much more discouraging. Six of eight patients so treated died in their immediate postoperative period. Cohn [4], utiIizing a combined thoracoabdomina1 incision, added spIenectomy to transesophagea1 varix Iigation. Three of his tweIve cirrhotic patients died postoperatively, and six of the nine who survived bIed again. WeIch’s [T] unsatisfactory experience with three transthoracic varix Iigations Ied him to advocate an abdominal approach. He reported on five patients so managed, three of whom died in the immediate postoperative period. AIso utihzing an abdominal approach, Hsii et al. [6] Iigated the bIeeding varix in five cirrhotic patients, in four of whom the cause was schistosomiasis. There was one postoperative death and one of the survivors bled shortIy after surgery. Other authors recommending Iigation of the varix but reporting no data are Julian [7], Cameron and Hughes [8] and Sanger [9]. Upper gastric transection, or bisection, first suggested by Tanner has been used by Allen [IO] in acute bIeeding. Of three patients in whom this procedure aIone was used, two died postoperativeIy. In two additional patients spIenorena1 shunt was added to gastric transection. Both survived but one died ten months later in hepatic failure. None of the five patients had recurrent bleeding. AIIen used an abdomina1 approach for the first two patients, but found this technicaIIy so difficult that in the Iast three patients he utiIized a thoracic approach. SegmentaI esophagogastrectomy has been used a few times during acute bIeeding. Shumacker and King [I I ] and NachIas [ 121 each report one cirrhotic patient handIed in this manner. Both patients died postoperativeIy. Rousselot et a1. [13] refer to five patients treated by esophagogastrectomy, but do not give detaiIs other than to state that the expected mortaIity wiI1 be 40 to 50 per cent. Lyons and Patton [14] aIso recommend this procedure when conservative measures fai1, but do not report any cases. From the foregoing it appears that these temporizing procedures have faiIed to accompIish what was intended. AI1 have proved to be formidabIe, poorIy torerated by the cirrhotic patient. Further, Iigation of the varix has often been followed by recurrent bIeeding, even though initia1 contro1 apparently had been achieved.
and
Pattison Despite its magnitude, portacava1 shunt has been empIoyed as an emergency measure in a few cirrhotic patients. ChiId [IS] described eight such patients. Three died postoperativeIy and three who survived died from six to thirtynine months Iater. O’Sullivan and Payne [16] added one patient to Child’s series. This patient survived but his foIIow-up was short. One of the patients who succumbed in the postoperative period continued to bIeed. Necropsy disclosed a thrombosed anastomosis. None of the remaining eight patients bled after surgery. In aI1 nine of these cases baIIoon tamponade was continued while the patient was taken to the operating room and had been effective in controIIing overt bleeding for one to eight days prior to surgery. PostoperativeIy, baIIoon tamponade was continued for one day in two patients and for Ionger periods up to eight days in the remainder. Criticism might be expressed in considering the operation as an emergency procedure in a11 these cases since bleeding had been controlled by tamponade for severa days in some. The question of whether or not porta decompression wouId immediateIy contro1 bleeding remained unanswered, since tamponade was continued postoperativeIy. NevertheIess, credit must be acknowIedged to these authors since many of their patients were desperateIy iI and by recognized criteria were poor shunt candidates. HufnageI [r7] refers to portacaval shunts performed during acute bIeeding without mortaIity but does not record any detaiIs. HamiIton [3] records one emergency portacava1 shunt with recovery. Recognizing certain Iimitations of baIIoon tamponade, this however continues to be our treatment of choice and surgical intervention is avoided if possibIe. When properIy used, tamponade is a very effective means of controIIing bIeeding. Effective use requires that the gastric baIIoon be inffated with at Ieast 300 to 400 cc. of air, then withdrawn with sufficient tension that the upper end of the gastric baIIoon is puIIed up in pear-shaped fashion into the Iower esophagus. Maintenance of this tension is most easiIy accompIished with a we11 fitting face mask to which the tube is anchored. Tamponade is maintained for twenty-four to forty-eight hours and then reIeased. In most instances bIeeding does not recur and an eIective portacava1 shunt can be considered Iater. If bIeeding recurs or is not controIIed by tam-
184
Emergency ponade, a decision must operative intervention.
be made
Portacaval
concerning
MATERIAL
Dissatisfied with the magnitude and undependabihty of palhative or temporizing surgical procedures, we have elected to utihze portacavaI shunt as an emergency procedure in eIcven seIected cirrhotic patients during acute varical bIeeding. The rationaIe in seIecting this operation lay in the beIief that this definitive procedure could be accomplished with apprcciabIy no greater stress to the patient than that associated with temporizing procedures. Further, it seemed IikeIy that porta decompression couId be reIied upon to terminate \,aricaI bIeeding immediateIy and permanentIy. In this group of eIeven cases being reported the indications for emergency surgery were: in one case, ineffectiveness of tamponade in controlling bleeding; in five cases, immediate recurrence of bIeeding on reIease of tamponade; in four cases, earIy recurrence of bIeeding whiIe recovering from previous hemorrhage; and in one case, mistaken diagnosis of bleeding peptic ulcer. ,411patients were either activeIy bIeeding at the time or had bIed activeIy within tweIve hours of surgery. None were in coma and onIy two had exhibited encephalopathy in the days prior to surgery. The Iaboratory data quoted in the case reports were the most current avaiIabIe and in most cases represented the values within two days or less of surgery. In most instances vitamin K therapy had preceded the quoted Human serum aIbumin prothrombin vaIue. had been used in some. In five patients the cirrhosis was secondary to aIcohoIism; in five it was secondary to hepatitis and in one it was juvenile. Liver biopsy obtained during surgery confirmed the diagnosis of cirrhosis in all cases. AI1 patients had end-to-side portacava1 shunt performed through a right thoracoabdomina1 Portal pressures were measured by incision. means of a standard saIine manometer with the base IeveI at the anatomica Ievel of the porta vein. In every instance baIIoon tamponade was disconnected during or immediateIy after completion of the operation. Four aIcoholic cirrhotic patients have been excIuded from this study because overt bIeeding had either not recurred or had been controlled by baIIoon tamponade for a period of five days in two and for six and seven days in the others. Two of these patients died post-
Shunt
operatively and that too long a time of bIeeding representative of
two survived. It is belie\,ed period had eIapsed from the to consider their surgery as the designation emergent>..
CASE
REPORTS
CASE I. A sixty-seven year old whitr woman with a duodenal ulcer and cirrhosis secondary to hepatitis was hospitalized on December I I, I()jh, with her first episode of bleeding. Hepatomegaly and splenomegaly were present. There was no ascites. Laboratory data consisted of brornsulphalein” retention of 25 per c’cnt, srrum albumin 3.5 gm. per cent, prothrombin IOO per crnt, t hgmol turbidity 2.6 units, ccphalin lloculation test I plus and tota serum bilirubin 1.8 mg. per Cent. Bleeding ceased spontaneously shortly after admission but recurred ten days later. BalIoon tamponade immediately controlled the bIeeding and it \vas released the following day, December 22nd. On the night of December 3oth bleeding recurred. Effcctive balloon tamponade was again instituted and the patient was operated upon eight hours latcar on December 3rst, with the tamponade tube in place. Blood replacement prior to surgery amounted to 4,000 cc. The operation consisted of splcnoportography, portacaval shunt, vagotomy and pyloroplasty. Because tamponade had effectively contralIed the bleeding and no bleeding was noted in the duodenal ulcer, it was concluded that the bleeding had been varical. Portal pressure was reduced from 41 to 15 cm. of saline. Operating time was not accurately recorded. During the postoperative course icterus became marked but quickly subsided to normal. The patient was discharged on the sixteenth postoperative day. Her subscquc~rrt course was excellent unti1 December 22, 1956, when she became confused for the lirst time. Hepatic coma developed and she died on December 31, 1956, two years to the day after surgery. There had been no bleeding following surgery. CASE II. A thirty-four year old white woman with cirrhosis secondary to hepatitis was hospitalized on November IO, 1955, with her third cpisodc of bleeding. Hepatomegaly, splcnomegaly and mild ascites were present. Laboratory data consisted of bromsuIphaIein retention of IO per crnt, serum albumin 3.3 gm. per cent, thymol turbidity 5 units, prothrombin IOO per cent and tota serum bilirubin 1.0 mg. per cent. BalIoon tamponade did not control bleeding well. Nevertheless, bleeding slo\sly ceased and in forty-eight hours the tube w-as removed. Severe bIeedinp recurred on the cvoning of November 18th. In view of the lack of succcs~~ with tamponade previously, the patient was operated upon immediately without insertion of a tamponade tube. BIood replacement prior to surgery was 8,500 cc. Portacaval shunt reduced the
MikkeIsen
and Pattison CASE IV. A nine year oId white girl with juveniIe cirrhosis was hospitalized on December 12, 1955, with her first episode of bleeding. BIeeding continued and on December 15th successfu1 baIIoon tamponade was instituted. The foIIowing day tamponade was released. On this day the patient lapsed into coma which persisted for two days. Jaundice steadily increased. Frequent right thoracenteses were required over the next few days. Hepatomegaly, spIenomegaIy and ascites were present. Laboratory data consisted of serum albumin 2.1 gm. per cent, prothrombin 40 per cent and tota serum bilirubin 24.6 mg. per cent. On December 28th balloon tamponade was again required. This was reIeased the folIowing day. On December 3Ist bleeding recurred and tamponade was once again instituted and the foIIowing day removed. On the night of January 5, 1956, bleeding recurred once again. Tamponade was once again instituted and the patient operated upon the folIowing morning. Prior to surgery blood replacement had been 17,500 cc. PortacavaI shunt reduced the porta pressure from 46 to 22 cm. of saline. Operating time was two hours, fifteen minutes. Following surgery there was no further bleeding, but within twelve hours the patient Iapsed into coma and died twenty-four hours after surgery. Autopsy revealed a patent anastomosis.
porta pressure from 50 to 17 cm. of saline. During the time that the porta vein was clamped the stomach could be seen to fil1 with blood from recurrence of acute varicaI bIeeding. Operating time was two hours, fifty-five minutes. The postoperative course was smooth and the patient was discharged on the fourteenth postoperative day. Her subsequent course has been excellent with ho further bleeding. At this writing her follow-up is twenty-seven months. CASE III. A sixty-seven year oId white woman with cirrhosis secondary to hepatitis was hospitaIized on October 24, 1955, with her first episode of bIeeding. The Iiver and spIeen were not paIpabIe. There was no ascites. Because of continued massive bleeding she was operated upon the folIowing morning and a wedge resection of the stomach for a gastric poIyp was performed. Ten days Iater, November 4th, massive bIeeding recurred and she was immediately reoperated upon. At this operation an empiric subtota1 gastrectomy was performed. A Iiver biopsy was reported as showing cirrhosis. OmentaI vein pressure was recorded as 30 cm. of saIine. BIeeding recurred the foIIowing day and continued intermittentIy for the next five days, at which time baIIoon tamponade was instituted which effectively controIIed the bleeding. Laboratory data consisted of bromsuIphaIein retention of 32 per cent, serum albumin 2.8 gm. per cent, prothrombin IOO per cent, thymo1 turbidity 1.0 units, cephaIin ffoccuIation 2 plus and tota seruin biIirubin 1.1 mg. per cent. During the ensuitig sixteen days tamponade was reIeased and reinstituted eight times. The Iongest period bf freedom from bleeding without tamponade during this time was about tweIve hours. On two occasions she became irrationa1 and disoriented, and a flapping tremor developed. At no time, however, did she Iapse into coma. On November 24th one of tis was caIIed in consuItation. The foIIowing morning tamponade was reIeased for an x-ray study for varices. During this examination the patient bIed again. Tamponade was once again instituted a&I she was operated upon that afternoon. BIood repIacement prior to surgery was 19,000 cc. Portacava1 shunt reduced the portal pressure from 33 to I7 cm. of saline. Operating time was not accurateIy recorded. Four hours after surgery the patient vomited about 200 cc. of bIood. Tamponade wab once again instituted and was removed IO hours later. Further bleeding did not occur. Her postoperative recovery in spite of a11 that she had been through was remarkabIy smooth and she was discharged on the thirteenth postoperative day. Her subsequent course during twenty-six months since surgery has been characterized by occasiona episodes of miId disorientation. Reduction in protein intake has improved this probIem. During the foIIow-up there has been no further bIeeding.
CASE v. A thirty-nine year oId white woman with cirrhosis secondary to hepatitis was hospitalized on June 9, 1956, with her fifth episode of bleeding. HepatomegaIy and spIenomegaIy were present. There was no ascites. Laboratory data consisted of bromsuIphaIein retention of IO per cent, serum aIbumin 3.3 gm. per cent, prothrombin 84 per cent, cephalin AoccuIation 4 pIus, aIkaIine phosphatase 26 Bodansky units and tota serum biIirubin I. I mg. per cent. An unsuccessful attempt to insert the baIloon tube was made. However, the patient stopped bIeeding spontaneousIy in a few hours. On June 14th bIeeding recurred and she was immediately operated upon without utiIizing tamponade because of the previous unsuccessfu1 attempt. BIood replacement prior to surgery was 3,500 cc. Portacaval shunt reduced the porta pressure from 46 to 24 cm. of saline. Operating time was two hours, forty minutes. The postoperative course was uneventfu1 and the patient was discharged on the tenth postoperative day. Three months later rehospitalization for two weeks was required for sudden development of jaundice that was thought to represent homoIogous serum hepatitis. SubsequentIy her course has been exceIIent without bleeding during a twenty-month foIIow-up. CASE VI. A lifty-seven year oId white man with cirrhosis secondary to aIcohoIism was hospitaIized on September 6, 1956, with his first episode of 186
Emergency
Portacaval
Shunt
consisted of serum albumin 3.0 gm. per rent. prothrombin 65 per cent, blood urea nitrogen 76 mg. per cent and total serum bilirubin 5.8 mg. per cent. Balloon tamponade was instituted which immc>diately controlled bleeding. During the ensuing eleven days bleeding recurred shortly after release of tampnnade on each occasion. On Fehruarg z2t h shortly after the lifth recurrence of bleeding, the patient was operated upon with balloon tamponadc in place. Blood replacement prior to surgery was 13,000 cc. Portacaval shunt reduced the portal pressure from 37 to 17 cm. of saline. Incidental cholecystectomy was carried out hccausc of cholvlithiasis. Operating time was one hour, fifty-t\vo minutes. Because of cxcessivc trachcobronchial secretions a tracheostomy was performed thr following day. The patient’s coursr was progressively downhiIi and he died in coma on the seventh postoperative day. Bleeding had not recurred.
bleeding. Hepatomegaly and ascites were present. The spleen was not palpable. Laboratory studies consisted of bromsulphalein retention of 15 per cent, serum albumin 3.0 gm. per cent and total serum bilirubin I 1.0 mg. per cent. Effective balloon tamponade was iminediately instituted. During the ensuing nine days tamponade was released on four occasions. Each time bleeding recurred and shock dcvclopcd within twenty-four hours. After the fourth episode on September 15th the patient was operated upon with balloon tamponade in place. Blood replacement prior to surgery was 8,000 cc. Portacaval shunt reduced the porta pressure from 4j to 22 cm. of saline. Operating time was t\(c) hours, forty minutes. Five hours after surgery the palirnt died in irreversible shock. Autopsy rcvcalcd a patent anastomosis. C*ASI-VII. A forty-eight year old white man with cirrhosis secondary to alcoholism was hospitalized on February 8, 1957, with his second bleeding episode. He was markedly obese with hepatomegaly and moderate ascites. The spleen was not palpabIe. I.aboratory data consisted of bromsulphalein rrtenticm of 46 per cent, serum albumin 3.4 gm. per cent, thymol turbidity 2.6 units, prothrombin -0 per cent and total serum bilirubin 1.6 mg. per cent. Bleeding ceased spontaneously but recurred massively four days later, on February 12th. Balloon tamponade was instituted and the patient taken immediately to the operating room. Blood rcplacrment prior to surgery was 3,000 cc. Portac*e\,al shunt reduced the porta pressure from 50 to 40 cm. of saline. Fifty per cent of the lumen of the portal vein was occluded by an organized thrombus. Thrombectomy was performed, but it was thought that not all of the proxima1 portion of the thrombus uas removed. Operating time was three hours. The postoperative course was smooth and the patient was discharged on the fifteenth postoperative day. 1Iis subsequent course was good for two months and he returned to part time employment. At this time jaundice developed and steadily progressed over the next month. He died in coma three months after surgery. Autopsy reveaIed a patent. shunt with no evidence of persisting portal vein thrombus. Comparison of liver biopsy specimen with the surgical specimen three months before suggested acute hepatitis, although this was equivocal and was read in different fashions by different pathologists. It, however, is believed that the patient succumbed to homoIogous serum hepatitis superimposed on cirrhosis. Bleeding had not rccurrcd.
CASE IS. A forty year old white woman \vith cirrhosis secondary to alcoholism was hospitalized on October I 3, 1957, with her first episode of bleccling. The bleeding was manifested by melena onI> and had been continuing for several days. Hcmoglobin on entry was 2.5 gm. per cent. Hepatomegaly was detected but splenomegalv and ascitrs were absent. Laboratory data consisted of brnmsulphalein retention of 25 per cent, serum albumin 3.3 gm. per cent and prothrombin 62 per cent. Melena continued and the day after admission the patient was operated upon with a prcoperati\.c* diagnosis of bleeding peptic ulcer. Blood replacement prior to surgery was 6,000 cc. Complete abdominal exploration, including pyloroduodcnotomy and gastrotomy, failed to reveal the presence of an ulcer. The stomach at the time of exploration contained no blood. The liver was obviously cirrhotic, the spleen enlarged and omental vein prrssure increased. Inspection of the lower esophagus through a sigmoidoscope revealed varices but therr was no bleeding at this time. The decision was made that in all probability the source of the bleeding had been varices. Therefore a portacaval shunt was performed which reduced the portal pressure from 38 to 17 cm. of saline. Gastrostomy was pcrformcd. Operating time was four hours. Mild icterus dcvcloped postoperatively and right-sided pleural lluid accumulated requiring several thoracentrses. The patient was discharged on the twentieth postopcrative day. Bleeding has not recurred in a fourmonth follow-up. CASE x. A fifty-seven year old white woman with cirrhosis secondary to alcoholism was hospitalized with her second bleeding episode on October 26, 1957. A left thoracoplasty for pulmonary tubercuIosis had been performed five years before. Hepatomegaly and splenomegaIy without ascitcs \vere present. Laboratory data consisted of 1,rom-
year old white man CASE VIII. A sixty-three with cirrhosis secondary to alcoholism was hospitaIitcd on February I I, 1957, with his third episode of bleeding. Hepatomegaly and spIenomegaIy were Ijrescnt. Ascites was absent. Laboratory data 18;
MikkeIsen suIphalein retention of IO per cent, serum aIbumin 4.2 gm. per cent, prothrombin 90 per cent, alkaline phosphatase 2.7 Bodansky units and tota serum biIirubin 1.4 mg. per cent. Balloon tamponade was instituted and it immediately controlled the bleeding. Tamponade was released forty-eight hours Iater. On the evening of October 29th severe bleeding recurred, baIIoon tamponade was once again instituted and the patient was operated upon the folIowing morning. BIood repIacement prior to surgery was 8,500 cc. Portacaval shunt reduced the porta pressure from 37 to 13 cm. of saIine. Gastrostomy and tracheostomy were aIso performed. Operating time was two hours, thirty-two minutes. Significant jaundice developed postoperativeIy and was slow to resoIve. The patient was discharged on the twenty-sixth postoperative day free of icterus and having been active about the ward. She was rehospitaIized three days Iater in light coma which rapidIy progressed to deep coma, and she died seven days Iater (thirty-six days after surgery). Necropsy was refused. BIeeding had not recurred. A nineteen year oId girl with cirrhosis CASE XI. ptobabIy secondary to hepatitis (perhaps better cIassified as juvenile cirrhosis) was hospitalized on November I, 1957, with her first episode of bleeding. The liver was not paIpable. There was mild ascites and marked spIenomegaIy. Laboratory data consisted of bromsuIphaIein retention of 18 per cent, serum aIbumin 4.3 gm. per cent, prothrombin 50 per cent, and total serum biIirubin 1.7 mg. per cent. Urinary gonadotropins were positive and it was estimated that gestation was of about two and a half months. Bleeding quickIy subsided without the use of baIIoon tamponade. EIective surgery was scheduIed for a convenient time approximately two weeks after admission. The day before the scheduIed eIective time, bIeeding recurred. The patient was immediateIy operated upon without institution of balloon tamponade. BIood repIacement prior to surgery was 5,000 cc. PortacavaI shunt reduced the porta pressure from 50 to 14 cm. of saIine. Operating time was two hours. The postoperative course was smooth; abortion did not occur and the patient was discharged on the eIeventh postoperative day. She has retained her pregnancy and at this writing continues we11 without bleeding during the two and a half months since surgery. RESULTS
AND
COMMENTS
Four of these eIeven patients died in their earIy postoperative period. Of these four, one patient died five hours after surgery in irreversibIe shock. Another patient died twentyfour hours Iater in coma. The remaining two died
and
Pattison in coma on the seventh and thirty-sixth postoperative day, respectively. Death occurring in the fourth patient is puzzling. Her “Iiver profiIe” indicated that she was a better operative candidate than any patient in the entire series. Her postoperative course was compIicated by rather persistent jaundice which, however, cIeared and she was very active about the ward when she was discharged on her twenty-sixth day. Three days Iater she was readmitted in coma, which although somewhat atypica1 of hepatic faiIure persisted unti1 her death seven days later. Necropsy was not permitted. Late death occurred in two of the seven survivors. Both died in hepatic coma. One patient was we11 and asymptomatic unti1 a few days before her death, two years to the day after surgery. The other died three months after surgery due to what is beIieved to have been homologous serum hepatitis superimposed upon aIcohoIic cirrhosis. Five patients are alive two and a haIf, four, twenty, twenty-six and twenty-seven months folIowing surgery. With one exception, varica1 bIeeding ceased abruptIy with compIetion of surgery in a11 patients even though baIIoon tamponade was not empIoyed postoperativeIy in any. In this one exception (Case III) the patient had been bIeeding aImost continuousIy for thirty-two days and baIIoon tamponade had been used intermittentIy for sixteen days preceding the operation. Four hours after surgery hematemesis of 200 cc. necessitated the reinstitution of tamponade for ten hours. FoIIowing remova of the tamponade tube at this time no further bIeeding was encountered. The concIusion that portal decompression as accomplished by endto-side portacava1 shunt can be relied upon to terminate varical bIeeding immediately appears to be warranted. An anaIysis of the postoperative deaths reveaIs two further saIient features: (I) Three of the four deaths occurred in patients who were significantly jaundiced at the time of surgery. In none of the survivors was the tota serum biIirubin over 1.8 mg. per cent. (2) Three of the five patients with aIcohoIic cirrhosis died, whiIe none of those with cirrhosis secondary to hepatitis died. Child’s suspicion that aIcohoIic cirrhotic patients fare more poorIy after surgery than those with cirrhosis due to hepatitis wouId seem to be supported by these resuIts. CriticaI evaIuation further reveaIs that the
Emergency
PortacavaI
“Iiver proMe” in the seven who survived categorized them, in generaI, as satisfactory candidates for eIective portacava1 shunt. With one exception such was not true of those who died. It wouId appear from this series that a further justifiable concfusion is that patients whose hepatic function indicates they wouId be satisfactorv elective shunt candidates, will tolerate this procedure in the presence of acute bIecding. Comparison of our resuIts with the other series of emergency procedures is exceedingly difficuIt. So much depends on the preoperative condition of the patient and this information is avaiIahIe onIy in the series reported by WeIch [5], ChiId [r5] and O’SuIIivan and Payne [r6]. WeIch’s two survivors of transabdomina1 varix ligation could both be considered, on the basis of “Iiver profiIe,” as shunt candidates. His three patients who died couId not be so categorized. The nine patients of ChiId and O’SuIIivan and Payne are not as clear-cut. Three of the six who survived could be cfassified as good shunt candidates, while one was borderline. The two remaining survivors had significant jaundice but both had primary biIiary cirrhosis, and in such patients jaundice does not carry the usuaI ominous significance. Of the three patients who died, one was borderline but two were definitely poor shunt candidates. Fifty cases of Iigation of the varix have been published. Seventeen of these patients died immediately postoperativeIy, representing an operative mortality rate of 34 per cent. Two deaths among the five patients subjected to upper gastric transection resuIts in a 40 per cent mortaIity rate. The mortaIity rate for esophagogastrectomy was IOO per cent for the two recorded cases. Three deaths among ten recorded patients subjected to portacava1 shunt represents a 30 per cent mortality rate. Four deaths among the eIeven patients in this series having portacava1 shunt represents a mortality rate of 36 per cent. It is hard11 justifiabIe to apply percentage mortaIity figures for these small groups, hut it appears that there is little to choose betaecn them in this regard. ‘The average operating time for portacaval shunt in the nine patients in which this information u-as recorded was two hours, thirty-nine This includes the few patients in minutes. whom ;tnciIIary procedures were performed. If the length of the operation is a measure of the stress suffered by the patient, it wouId seem
Shunt
that portacava1 shunt, in this series, would offer no more than the temporizing procedures. The real purpose of any emergency operative procedure is to salvage a Iife that would otherwise be Iost. Thus seven patients in this series might be considered to have been salvaged. A more critical evaluation of the present series, however, might conclude that only one patient was truly salvaged. This patient \vas Case III. Her bIeeding was unreIenting and surely would have caused her death if operative intervention had not been used. Perhaps Case II should also be so cIassilied. Her bIeeding \vas torrenti: and balloon tamponade had previously been unsuccessful. Ejolvever, Meeding on three previous occasions had terminated spontaadmittedly none \j.as as neousI,v, aIthough as the episode Ieadlng to surgery. severe Antithetically, Case x might have survived surgery if more time had been permitted for her liver function to improve. The other three patients who died had been certainly given an adequate tria1 by conservative means. Criticism cannot be directed to an operative attempt to control unreIenting b1eeding. Nevertheless, none of these three patients survived. In the remaining five patients who survived it cannot be categoricaIIy assumed that hIeeding cvould not have ceased quickly and permitted an eIective procedure at a later time. Many cirrhotic patients who hieed, quickly manifest evidence of hepatic encephalopathy. Few if any of these patients will tolerate an> type of emergency surgery. From our expertence a further contraindication to emergcncl surgery is the presence of significant jaundice even though other evidences of hepatic failure are absent. An exception to this hypothesis, is jaundice in primary biliary cirperhaps, rhosis. In this disease jaundice does not carry the same ominous significance that it does in aIcohoIic cirrhosis. The opinion has been expressed by others that perhaps a11 cirrhotic patients with acute bIeeding should be subjected to some type of operative control immediately after hospitalization. WC c;tnnot concur in this opinion. Laboratar?, in addition to clinical cvaIuwtion i5 required in their management, and in many of these patients this would not be immediately avaiIable. Further, some patients who initialI>. hepatic stabiIizec1 deteriorntc into appear coma within tlvelve to twenty,-four hours cvcn though blcaeding has been controlled II\
MikkeIsen
and
6. Hsii, CHI-HO, YANG TOA-HUA and LI SHEN. Surgicat treatment of acute massive hemorrhage from esophagea1 and gastric varices associated with portal hypertension. Chinese M. J., 74: 359, 1956. 7. JULIAN, 0. C. Choice of operation in portaI hypertension with varices. Surg., Gynec. @ Obst., IOO:
tamponade. These patients wouId not toIerate emergency surgery but might be salvaged by conservative means for future elective surgical management. It therefore seems doubtfuI to us that any emergency surgical procedure as practiced today wiII significantIy aIter tota mortality rates from acutely bleeding varices in cirrhotic patients. AdmittedIy, an occasiona patient may be saIvaged. In such a patient it seems reasonabIe to believe that a portacava1 shunt wil1 be toIerated as we11 as any other of the emergency procedures proved to be effective in controIIing bIeeding.
753, 1955. 8. CAMERON, A. and HUGHES, J. F. Massive gastrointestinal hemorrhage due to esophagea1 varices. Delaware M. J., 28: 161, 1956. g. SANGER, P. W. Discussion. Ann. h-g., 141: 647, ‘955. IO. ALLEN, J. G. and HEAD, L. R. The diagnosis of portal hypertension with notes on treatment. S. Clin. Nortb America, p. I rg, 1956. I I. SHUMACKER,H. B. and KING, H. SpIenic studies. II. PortaI hypertension in chiIdren associated with gastroesophageal hemorrhage. Arch. Surg., 65:
SUMMARY
Eleven cirrhotic patients had portacava1 shunt performed during acute bIeeding from esophagogastric varices. Four died in the postoperative period and two died at a Iater date. Five remain alive and have not bIed again. OnIy one patient bIed after surgery in spite of the fact that baIIoon tamponade was not empIoyed postoperativeIy in any. The bIeeding in this one patient was minor in degree. The concIusion that porta decompression as accompIished by end-to-side portacava1 shunt, wiI1 in aImost a11 cases immediately terminate bIeeding seems warranted. Preoperative jaundice was an ominous sign with regard to survival. Comparison of our resuIts with those utilizing emergency temporizing procedures suggests that the operative mortaIity is IittIe different and that effective contro1 of bIeeding is more IikeIy to be obtained by portacava1 shunt. It stiI1 remains doubtfu1 that any of the emergency operations currently in use will significantIy improve the total mortaIity for these cirrhotic patients with severe varica1 bIeeding. REFERENCES I. LINTON, R. R. and ELLIS, D. S. Emergency
2.
3.
4. 5.
Pattison
and definitive treatment of bleeding esophageal varices. J. A. M. A., 160: 1017, 1956. LINTON, R. R. and WARREN. R. The emergency treatment of massive bIeeding from esophageal varices by transesophagea1 suture of these vesseIs at the time of acute hemorrhage. Surgery, 33: 243, ‘953. HAMILTON, J. E. The management of bleeding esophagea1 varices associated with cirrhosis of the liver. Ann. Surg., 141: 637, 1955. COHN, R. Surgical treatment of bleeding esophageal varices. C&Jo&a Med., 83: 348, 1955. WELCH, C. S. Ligation of esophageal varices by the transabdominal route. New England 3. Med., 255: 677, 1956.
499. 1952. 12. NACHLAS. M. M. Treatment of bIeedine esoohaaeaI varices by resection of the Iower esophagus. Arch. Surg., 72: 634, 1956. 13. LIEBOWITZ, H. R., RUZICKA, F. F., ROUSSELOT, L. M. and GREEN, C. H. Panel discussion on esophageal varices-porta hypertension. Am. J. Gastroenterol., 27: 325, 1957. 14. LYONS, C. and PATTON, T. B. BIeeding esophagea1 varices. Surgery, 39: 540, 1956. 15. CHILD. C. G. The Shattuck Iecture: the oorta1 circuIation. New England J. Med., 252: 837: 1955. 16. O’SULLIVAN, W. D. and PAYNE, M. A. The emergency portacaval shunt. Surg., Gynec. ti Obst., 102: 668, 1956. 17. HUFNAGEL, C. A. SurgicaI treatment of portal hypertension. Am. J. Gastroenterol., 23: 522, 1955. v
I
Y
DISCUSSION RAY COHN (San Francisco, CaIif.): It is fair to say that the general consensus of surgeons is that portacaval anastomosis is the best surgica1 method to Iower the porta venous pressure. It is also fair to say that it is the genera1 consensus of physiologists that diversion of the porta bIood ffow is a most damaging procedure with respect to function of the Iiver. The enzyme activities of the liver are the latest in the long Iist of functions of the liver which have been shown to be damaged by shunting procedures [r]. In fact, the most amazing feature of this disease to both cIinician and physioIogist is that a patient shouId be as cIinicaIIy we11 as some cirrhotic patients appear to be, with nothing but a shrunken noduIar liver about one-fifth norma size when seen at operation. In the earIy days of this operation many patients who were poor risks did not survive. Because of this fact, criteria such as that of Linton were established with a consequent fall in the immediate mortality rate. It then became the fashion to compare the results after operation on this highly
1MANNING and DELP. Management of hepatocerebra1 intoxication. New EngLand J. Med., 258: 62, 1958. 190
Emergency
PortacavaI
selected group of cirrhntic patients with the dire consequences of massive hemorrhage on the unselrcted series of those with cirrhosis. This fallacy has been recognized by surgeons reporting series of portacaval anastomosis, the latest of whom has bcrn Hallenbeck. From collected data of the San Francisco Hospital it can be noted that 86 per cent of the patients with gastrointestinal bleeding who died, did so on the first bleeding episode. Now of the patients who survived and who satisfied the usual laboratory and clinical criteria for portacaval anastomosis, eleven or roughly 30 per cent survived five years or more with medical therapy. The authors have recorded eleven cases, the longest follow-up being twenty-seven months. Only five patients survived, less than half. Obviously the authors’ eleven cases were extremely poclr risks and no one can state with assurance what their fate would have been had operation not been performed. I would agree with the authors that only in Case 111 does it really appear that operation was life saving and in this patient cirrhosis sc,emcd to be minimal. It would appear to mc that the number of patients for whom portacaval anastomosis is indicated is extremely smaI1, and that when the operation is applied to the desperately ill, acute bleeder the selection of the proper cases would appear even more dificuIt, probably impossible. .IACK M. FAHRIS (Los Angeles, Calif.): Dr. Smith and I have been interested in this problem for some time and have followed the work of Drs. Mikkelsen and Pattison with a great deaI of interest. Although \Vc’ ha\c had less experience with portacaval anastomosis, we have operated upon three such patients as emergcncics where the balIoon had been in place, for a number of days and attempts to release it had been followed by recurrence of the hemorrhage requiring further transfusion. Because of the conviction that other operations do not approach the fundamental problem, we carried out emergency operations. I would like to present one case briefly, if 1 may, to support the views of Drs. lZlikkelsen and Pattison that survival may at least be enhanced by emergency portacaval anastomosis in the exceedingly poor-risk patient.* This man, in his forties, had typical alcoholic cirrhosis and was operated upon at another hospital with a mistaken diagnosis of bleeding ulcer. At that opcraticm the diagnosis of cirrhosis was established by biopsy and measurements of portal pressures. Following operation conservative treatment was instituted and during the ensuing five days an attempt was made to release the balloon on three dift’crent occasions; each attempt was followed
a A slide was employed by Dr. Farris illustrating the stable pulse rate between 68 and 78 throughout portacuval anastomosis at temperature of 31°c.
Shunt
by a recurrrnce of brisk hemorrhage requiring transfusion. At this point the patient was transferred to us via ambulance, a distance of about $0 miles, having received eighteen blood transfusions during this phase of his postoperative course. Actuallv, he was getting blood while being transported via ambulance and during this his wound disrupted. He arrived, semi-stuporous, with copious amounts of ascitic fluid draining from his abdominal wound. There was also clinical jaundice. Although hc presented what in the past would be considered absolute contraindications for portacaval anastomosis, it was apparent that he could not tolerate the balloon much longer. We were aware of cases of perforation of the esophagus from prolonged presence of the tamponade. Therefore we decided to operate upon him as an emergency, and with the conviction that portacaval anastomosis was tht operation of choice we determined to carry it out under hypothermia. His preoperative pulse was between IX) and 130 and after one hour of cooling in ice it was full and strong at 68. During the operation of about two and a half hours’ duration the pulse remained between 68 and 80 with a lo\% temperature of 31”~. We were able to revise his previous wound disruption and perform an end-toside portacaval anastomosis; and although there was some increase in his jaundice postoperatively, he made a surprisingly good convalescence and was able to be clischarged from the hospital on his twenty-first postoperative day. He has now been followed up for two years and is pursuing his usual occupation. There has been no recurrent bleeding. Although observations such as this arc not subject to scientific control, we believe that recovery in this patient would have been unlikely without portacaval anastomosis. Furthermore, we believe that hypothermia has a definite place in the management of this type ol case. The cooling with resultant lowering of mrtabolic requirements provides a logical protection to the liver both from the standpoint of lowering its metabolism and also protecting it from the trauma of the operation and toxic effects of anrsthetic agents. The most important benefit, howcvcr, to IX derived from the cooling technic is the prrdictablc and consistent lowering of the pulse rate. To undertake a major and prolonged surgical operation in a patient with a preoperative pulse in excess ot’ I 20 in the past has been fraught with hazard; the pulse invariably increases during the operation with a diminished stroke volume and cardiac output progressing frequently to cardiogcnic shock. 1t appears that cooling nullifies the corticosteroid and autonomic response which in the usual patic*nt arc normal and acceptable defense mechanisms. \\‘hcn fever and tachycardia, however, exceed the bounds of homecjstatic control, survival is endangered.
MikkeIsen
and Pattison
We wouId Iike to comphment the authors on the exceIlence of their presentation. HORACE J. MCCORKLE (San Francisco, CaIif.): I enjoyed Dr. Mikkelsen’s presentation and Iearned a great dea1 from it, as I have also from his recent pubIications. Last year I was abIe to rescue a patient with persistent bleeding from esophageal varices by an operation that he apparentIy does not think too much of at this time, namely, the Tanner procedure. This young woman was at term and she began to bIeed from esophagea1 varices immediately after the dehvery of her chiId. She was treated with the Sengstaken tube over a period of about three weeks, and every time the tube was removed or Ioosened she wouId almost bleed to death. Finally we performed a Iaparotomy and were unable to find either porta or spIenic vein. Her spIeen had been previousIy removed and there was no possibility of performing a porta vein decompression procedure. Therefore al1 the numerous, very Iarge veins on the undersurface of the patient’s diaphragm were Iigated and the Tanner procedure was performed. This was about eight months ago and she has had no subsequent hemorrhage. WILLIAM P. MIKKELSEN (cIosing): I shouId like to refer to Dr. McCorkIe’s suggestion of the vaIue of a Tanner procedure (gastric transection). It is true that about two years ago we recommended this operation for patients with portal hypertension in whom major venous thrombosis prevented the construction of a venous shunt. Eight patients have been so treated. Two of these have since exsanguinated from recurrent bleeding and five others have bled severeIy one or more times since surgery. The remaining patient has not re-bIed but persists in demonstrating Iarge varices. This aIarmingIy poor experience has Ied us now to abandon this operation. Dr. Farris’ suggestion that hypothermia wouId
be of vaIue is we11 taken. CertainIy it has theoretical advantages. Hypothermia was empIoyed with two patients in this series. Both succumbed, but admittedIy, both were in the poor risk category. Dr. Cohn for years has insisted that portacaval shunt does not proIong the Iife of these patients and that it does not result in Iasting reduction of porta pressure. Nevertheless, it remains our delinite impression that Iife is proIonged and certainIy that the threat of recurrent bleeding is aImost eIiminated. Persistence of reduction in porta pressure by end-to-side portacava1 shunt is demonstrated in reviewing pressure studies in nine patients before and up to ninety weeks after eIective surgery.* Catheterization in these cases was performed by Drs. ReynoIds and Redeker of the Department of Medicine at the University of Southern Cahfornia. Preoperative wedged hepatic vein and operative portat pressures were compared with postoperative wedged hepatic vein ‘pressures and portal pressures obtained by passage of the catheter through the anastomosis and out into the porta vein. It wiI1 be noted that folIowing surgery wedged hepatic vein pressure was reduced by about one-third but stil1 remained at a moderateIy hypertensive Ievel, whereas the direct porta pressure in a11 instances was reduced to normal. ,It shouId be pointed out that after division of the portal vein, wedged hepatic vein pressure no Ionger reflects portal pressure and that its minimal reduction after a shunt reflects directly the reduced hepatic bIood ffow ordy. ClinicaIIy, nearly a11 our shunt patients, now numbering over 120, have demonstrated disappearance of varices by roentgenography. Further, recurrent gastrointestinal bIeeding has been a rather rare occurrence. Thus our confidence in end-to-side portacava1 shunt continues. * Dr. MikkeIsen illustrated his discussion with a slide.
192
Portacaval
WILLIAM P. MIKKELSEN, M.D., Los Angeles, Pasadena,
From the Department of Surgery, University oJ Soulbern CaliJornia, School of Medicine, Los Angeles, California.
T esophagogastric
HE management
of acute bleeding from varices in the patient with hepatic cirrhosis continues to be a most chalIenging problem. When current conservative measures are used approximateIy one-haIf of these patients succumb. This has Ied severa surgeons to explore the feasibility of operative control of bleeding. It has been recognized, however, that any surgica1 procedure of magnitude will not be tolerated by the cirrhotic patient in whom severe hepatrc decompensation deveIops. Such a patient must be managed by conservative means, regardless of the persistence of uncontrohed bIeeding. These are the patients who are mainIy responsible for the continued high mortality from acute bleeding. In contrast to the foregoing situation there is the patient whose hepatic reserve is good, whose bleeding ceases spontaneousIy or is quickly controlled by conservative means. Here an elective operative procedure may be carried out at a convenient time. This is the group of patients who, in any reported series, survive their bIeeding episode. Between the two extremes just described there lies a third group of patients for whom an emergency operative procedure may prove Iifesaving. These patients initiahy remain in reasonably good condition, but if bleeding continues or recurs, they may sink into Iethal hepatic decompensation or cssanguinate. This group has not been clearly delineated but, in all likelihood, it is a considerablJ- smaller group than either of the other t \vo. CURRENT
EMERGENCY
PROCEDURES
Experience with operative intervention during active varical bIeeding has been limited but has been sufficient to indicate that certain pro-
Calijornia,
Shunt
AND ARTHUR C. PATTISON, M.D.,
Calijornia
cedures cannot be reIied upon to terminate bIeeding. Into this category faI1 such operations as spIenectomy, coronary vein ligation, esophageal devascularization or ligation of one or al1 of the branches of the celiac axis. WhiIe eIective portacava1 shunt is now generally accepted as the best means of eliminating recurrent bIeeding from varices, its effectiveness in controIIing acute bIeeding has remained undetermined. Furthermore, this operation has been considered to be too formidable for these acuteIy bIeeding patients. Consequently, there have evolved less formidabIe procedures designed to control bleeding temporarily and to be followed later by a more extensive definitive procedure for permanent contro1. Direct transesophageal suture of the bleeding varix, usuatly together with suture of the remainder of the visibIe varices in the lower esophagus and cardia, has received the greatest attention. Linton and Ellis [I] reported on twenty patients treated in this manner hy a transthoracic approach. Three patients died immediately postoperatively and one died live weeks later following partial gastrectomv for recurrent varicaI bleeding. Another patient died fifteen months Iater from hepatic failure. The remaining fifteen patients had portacaval shunts of some type constructed later. The subsequent course of these latter patients was not described. Reference is not made as to the number of patients who re-bIed before a definitive procedure couId be carried out, although in a previous communication [2] five of nine survivers did re-bIeed. Linton now rccommends immediate tramesophageal vat-ix ligation in any patient who bleeds severely enough to require balloon tamponade. He admits, however, that he refuses to operate upon any patient who exhibits signs of “impending liver faiIure.” Hamihon’s [j] results with transthoracic
Mikkelsen varix Iigation have been much more discouraging. Six of eight patients so treated died in their immediate postoperative period. Cohn [4], utiIizing a combined thoracoabdomina1 incision, added spIenectomy to transesophagea1 varix Iigation. Three of his tweIve cirrhotic patients died postoperatively, and six of the nine who survived bIed again. WeIch’s [T] unsatisfactory experience with three transthoracic varix Iigations Ied him to advocate an abdominal approach. He reported on five patients so managed, three of whom died in the immediate postoperative period. AIso utihzing an abdominal approach, Hsii et al. [6] Iigated the bIeeding varix in five cirrhotic patients, in four of whom the cause was schistosomiasis. There was one postoperative death and one of the survivors bled shortIy after surgery. Other authors recommending Iigation of the varix but reporting no data are Julian [7], Cameron and Hughes [8] and Sanger [9]. Upper gastric transection, or bisection, first suggested by Tanner has been used by Allen [IO] in acute bIeeding. Of three patients in whom this procedure aIone was used, two died postoperativeIy. In two additional patients spIenorena1 shunt was added to gastric transection. Both survived but one died ten months later in hepatic failure. None of the five patients had recurrent bleeding. AIIen used an abdomina1 approach for the first two patients, but found this technicaIIy so difficult that in the Iast three patients he utiIized a thoracic approach. SegmentaI esophagogastrectomy has been used a few times during acute bIeeding. Shumacker and King [I I ] and NachIas [ 121 each report one cirrhotic patient handIed in this manner. Both patients died postoperativeIy. Rousselot et a1. [13] refer to five patients treated by esophagogastrectomy, but do not give detaiIs other than to state that the expected mortaIity wiI1 be 40 to 50 per cent. Lyons and Patton [14] aIso recommend this procedure when conservative measures fai1, but do not report any cases. From the foregoing it appears that these temporizing procedures have faiIed to accompIish what was intended. AI1 have proved to be formidabIe, poorIy torerated by the cirrhotic patient. Further, Iigation of the varix has often been followed by recurrent bIeeding, even though initia1 contro1 apparently had been achieved.
and
Pattison Despite its magnitude, portacava1 shunt has been empIoyed as an emergency measure in a few cirrhotic patients. ChiId [IS] described eight such patients. Three died postoperativeIy and three who survived died from six to thirtynine months Iater. O’Sullivan and Payne [16] added one patient to Child’s series. This patient survived but his foIIow-up was short. One of the patients who succumbed in the postoperative period continued to bIeed. Necropsy disclosed a thrombosed anastomosis. None of the remaining eight patients bled after surgery. In aI1 nine of these cases baIIoon tamponade was continued while the patient was taken to the operating room and had been effective in controIIing overt bleeding for one to eight days prior to surgery. PostoperativeIy, baIIoon tamponade was continued for one day in two patients and for Ionger periods up to eight days in the remainder. Criticism might be expressed in considering the operation as an emergency procedure in a11 these cases since bleeding had been controlled by tamponade for severa days in some. The question of whether or not porta decompression wouId immediateIy contro1 bleeding remained unanswered, since tamponade was continued postoperativeIy. NevertheIess, credit must be acknowIedged to these authors since many of their patients were desperateIy iI and by recognized criteria were poor shunt candidates. HufnageI [r7] refers to portacaval shunts performed during acute bIeeding without mortaIity but does not record any detaiIs. HamiIton [3] records one emergency portacava1 shunt with recovery. Recognizing certain Iimitations of baIIoon tamponade, this however continues to be our treatment of choice and surgical intervention is avoided if possibIe. When properIy used, tamponade is a very effective means of controIIing bIeeding. Effective use requires that the gastric baIIoon be inffated with at Ieast 300 to 400 cc. of air, then withdrawn with sufficient tension that the upper end of the gastric baIIoon is puIIed up in pear-shaped fashion into the Iower esophagus. Maintenance of this tension is most easiIy accompIished with a we11 fitting face mask to which the tube is anchored. Tamponade is maintained for twenty-four to forty-eight hours and then reIeased. In most instances bIeeding does not recur and an eIective portacava1 shunt can be considered Iater. If bIeeding recurs or is not controIIed by tam-
184
Emergency ponade, a decision must operative intervention.
be made
Portacaval
concerning
MATERIAL
Dissatisfied with the magnitude and undependabihty of palhative or temporizing surgical procedures, we have elected to utihze portacavaI shunt as an emergency procedure in eIcven seIected cirrhotic patients during acute varical bIeeding. The rationaIe in seIecting this operation lay in the beIief that this definitive procedure could be accomplished with apprcciabIy no greater stress to the patient than that associated with temporizing procedures. Further, it seemed IikeIy that porta decompression couId be reIied upon to terminate \,aricaI bIeeding immediateIy and permanentIy. In this group of eIeven cases being reported the indications for emergency surgery were: in one case, ineffectiveness of tamponade in controlling bleeding; in five cases, immediate recurrence of bIeeding on reIease of tamponade; in four cases, earIy recurrence of bIeeding whiIe recovering from previous hemorrhage; and in one case, mistaken diagnosis of bleeding peptic ulcer. ,411patients were either activeIy bIeeding at the time or had bIed activeIy within tweIve hours of surgery. None were in coma and onIy two had exhibited encephalopathy in the days prior to surgery. The Iaboratory data quoted in the case reports were the most current avaiIabIe and in most cases represented the values within two days or less of surgery. In most instances vitamin K therapy had preceded the quoted Human serum aIbumin prothrombin vaIue. had been used in some. In five patients the cirrhosis was secondary to aIcohoIism; in five it was secondary to hepatitis and in one it was juvenile. Liver biopsy obtained during surgery confirmed the diagnosis of cirrhosis in all cases. AI1 patients had end-to-side portacava1 shunt performed through a right thoracoabdomina1 Portal pressures were measured by incision. means of a standard saIine manometer with the base IeveI at the anatomica Ievel of the porta vein. In every instance baIIoon tamponade was disconnected during or immediateIy after completion of the operation. Four aIcoholic cirrhotic patients have been excIuded from this study because overt bIeeding had either not recurred or had been controlled by baIIoon tamponade for a period of five days in two and for six and seven days in the others. Two of these patients died post-
Shunt
operatively and that too long a time of bIeeding representative of
two survived. It is belie\,ed period had eIapsed from the to consider their surgery as the designation emergent>..
CASE
REPORTS
CASE I. A sixty-seven year old whitr woman with a duodenal ulcer and cirrhosis secondary to hepatitis was hospitalized on December I I, I()jh, with her first episode of bleeding. Hepatomegaly and splenomegaly were present. There was no ascites. Laboratory data consisted of brornsulphalein” retention of 25 per c’cnt, srrum albumin 3.5 gm. per cent, prothrombin IOO per crnt, t hgmol turbidity 2.6 units, ccphalin lloculation test I plus and tota serum bilirubin 1.8 mg. per Cent. Bleeding ceased spontaneously shortly after admission but recurred ten days later. BalIoon tamponade immediately controlled the bIeeding and it \vas released the following day, December 22nd. On the night of December 3oth bleeding recurred. Effcctive balloon tamponade was again instituted and the patient was operated upon eight hours latcar on December 3rst, with the tamponade tube in place. Blood replacement prior to surgery amounted to 4,000 cc. The operation consisted of splcnoportography, portacaval shunt, vagotomy and pyloroplasty. Because tamponade had effectively contralIed the bleeding and no bleeding was noted in the duodenal ulcer, it was concluded that the bleeding had been varical. Portal pressure was reduced from 41 to 15 cm. of saline. Operating time was not accurately recorded. During the postoperative course icterus became marked but quickly subsided to normal. The patient was discharged on the sixteenth postoperative day. Her subscquc~rrt course was excellent unti1 December 22, 1956, when she became confused for the lirst time. Hepatic coma developed and she died on December 31, 1956, two years to the day after surgery. There had been no bleeding following surgery. CASE II. A thirty-four year old white woman with cirrhosis secondary to hepatitis was hospitalized on November IO, 1955, with her third cpisodc of bleeding. Hepatomegaly, splcnomegaly and mild ascites were present. Laboratory data consisted of bromsuIphaIein retention of IO per crnt, serum albumin 3.3 gm. per cent, thymol turbidity 5 units, prothrombin IOO per cent and tota serum bilirubin 1.0 mg. per cent. BalIoon tamponade did not control bleeding well. Nevertheless, bleeding slo\sly ceased and in forty-eight hours the tube w-as removed. Severe bIeedinp recurred on the cvoning of November 18th. In view of the lack of succcs~~ with tamponade previously, the patient was operated upon immediately without insertion of a tamponade tube. BIood replacement prior to surgery was 8,500 cc. Portacaval shunt reduced the
MikkeIsen
and Pattison CASE IV. A nine year oId white girl with juveniIe cirrhosis was hospitalized on December 12, 1955, with her first episode of bleeding. BIeeding continued and on December 15th successfu1 baIIoon tamponade was instituted. The foIIowing day tamponade was released. On this day the patient lapsed into coma which persisted for two days. Jaundice steadily increased. Frequent right thoracenteses were required over the next few days. Hepatomegaly, spIenomegaIy and ascites were present. Laboratory data consisted of serum albumin 2.1 gm. per cent, prothrombin 40 per cent and tota serum bilirubin 24.6 mg. per cent. On December 28th balloon tamponade was again required. This was reIeased the folIowing day. On December 3Ist bleeding recurred and tamponade was once again instituted and the foIIowing day removed. On the night of January 5, 1956, bleeding recurred once again. Tamponade was once again instituted and the patient operated upon the folIowing morning. Prior to surgery blood replacement had been 17,500 cc. PortacavaI shunt reduced the porta pressure from 46 to 22 cm. of saline. Operating time was two hours, fifteen minutes. Following surgery there was no further bleeding, but within twelve hours the patient Iapsed into coma and died twenty-four hours after surgery. Autopsy revealed a patent anastomosis.
porta pressure from 50 to 17 cm. of saline. During the time that the porta vein was clamped the stomach could be seen to fil1 with blood from recurrence of acute varicaI bIeeding. Operating time was two hours, fifty-five minutes. The postoperative course was smooth and the patient was discharged on the fourteenth postoperative day. Her subsequent course has been excellent with ho further bleeding. At this writing her follow-up is twenty-seven months. CASE III. A sixty-seven year oId white woman with cirrhosis secondary to hepatitis was hospitaIized on October 24, 1955, with her first episode of bIeeding. The Iiver and spIeen were not paIpabIe. There was no ascites. Because of continued massive bleeding she was operated upon the folIowing morning and a wedge resection of the stomach for a gastric poIyp was performed. Ten days Iater, November 4th, massive bIeeding recurred and she was immediately reoperated upon. At this operation an empiric subtota1 gastrectomy was performed. A Iiver biopsy was reported as showing cirrhosis. OmentaI vein pressure was recorded as 30 cm. of saIine. BIeeding recurred the foIIowing day and continued intermittentIy for the next five days, at which time baIIoon tamponade was instituted which effectively controIIed the bleeding. Laboratory data consisted of bromsuIphaIein retention of 32 per cent, serum albumin 2.8 gm. per cent, prothrombin IOO per cent, thymo1 turbidity 1.0 units, cephaIin ffoccuIation 2 plus and tota seruin biIirubin 1.1 mg. per cent. During the ensuitig sixteen days tamponade was reIeased and reinstituted eight times. The Iongest period bf freedom from bleeding without tamponade during this time was about tweIve hours. On two occasions she became irrationa1 and disoriented, and a flapping tremor developed. At no time, however, did she Iapse into coma. On November 24th one of tis was caIIed in consuItation. The foIIowing morning tamponade was reIeased for an x-ray study for varices. During this examination the patient bIed again. Tamponade was once again instituted a&I she was operated upon that afternoon. BIood repIacement prior to surgery was 19,000 cc. Portacava1 shunt reduced the portal pressure from 33 to I7 cm. of saline. Operating time was not accurateIy recorded. Four hours after surgery the patient vomited about 200 cc. of bIood. Tamponade wab once again instituted and was removed IO hours later. Further bleeding did not occur. Her postoperative recovery in spite of a11 that she had been through was remarkabIy smooth and she was discharged on the thirteenth postoperative day. Her subsequent course during twenty-six months since surgery has been characterized by occasiona episodes of miId disorientation. Reduction in protein intake has improved this probIem. During the foIIow-up there has been no further bIeeding.
CASE v. A thirty-nine year oId white woman with cirrhosis secondary to hepatitis was hospitalized on June 9, 1956, with her fifth episode of bleeding. HepatomegaIy and spIenomegaIy were present. There was no ascites. Laboratory data consisted of bromsuIphaIein retention of IO per cent, serum aIbumin 3.3 gm. per cent, prothrombin 84 per cent, cephalin AoccuIation 4 pIus, aIkaIine phosphatase 26 Bodansky units and tota serum biIirubin I. I mg. per cent. An unsuccessful attempt to insert the baIloon tube was made. However, the patient stopped bIeeding spontaneousIy in a few hours. On June 14th bIeeding recurred and she was immediately operated upon without utiIizing tamponade because of the previous unsuccessfu1 attempt. BIood replacement prior to surgery was 3,500 cc. Portacaval shunt reduced the porta pressure from 46 to 24 cm. of saline. Operating time was two hours, forty minutes. The postoperative course was uneventfu1 and the patient was discharged on the tenth postoperative day. Three months later rehospitalization for two weeks was required for sudden development of jaundice that was thought to represent homoIogous serum hepatitis. SubsequentIy her course has been exceIIent without bleeding during a twenty-month foIIow-up. CASE VI. A lifty-seven year oId white man with cirrhosis secondary to aIcohoIism was hospitaIized on September 6, 1956, with his first episode of 186
Emergency
Portacaval
Shunt
consisted of serum albumin 3.0 gm. per rent. prothrombin 65 per cent, blood urea nitrogen 76 mg. per cent and total serum bilirubin 5.8 mg. per cent. Balloon tamponade was instituted which immc>diately controlled bleeding. During the ensuing eleven days bleeding recurred shortly after release of tampnnade on each occasion. On Fehruarg z2t h shortly after the lifth recurrence of bleeding, the patient was operated upon with balloon tamponadc in place. Blood replacement prior to surgery was 13,000 cc. Portacaval shunt reduced the portal pressure from 37 to 17 cm. of saline. Incidental cholecystectomy was carried out hccausc of cholvlithiasis. Operating time was one hour, fifty-t\vo minutes. Because of cxcessivc trachcobronchial secretions a tracheostomy was performed thr following day. The patient’s coursr was progressively downhiIi and he died in coma on the seventh postoperative day. Bleeding had not recurred.
bleeding. Hepatomegaly and ascites were present. The spleen was not palpable. Laboratory studies consisted of bromsulphalein retention of 15 per cent, serum albumin 3.0 gm. per cent and total serum bilirubin I 1.0 mg. per cent. Effective balloon tamponade was iminediately instituted. During the ensuing nine days tamponade was released on four occasions. Each time bleeding recurred and shock dcvclopcd within twenty-four hours. After the fourth episode on September 15th the patient was operated upon with balloon tamponade in place. Blood replacement prior to surgery was 8,000 cc. Portacaval shunt reduced the porta pressure from 4j to 22 cm. of saline. Operating time was t\(c) hours, forty minutes. Five hours after surgery the palirnt died in irreversible shock. Autopsy rcvcalcd a patent anastomosis. C*ASI-VII. A forty-eight year old white man with cirrhosis secondary to alcoholism was hospitalized on February 8, 1957, with his second bleeding episode. He was markedly obese with hepatomegaly and moderate ascites. The spleen was not palpabIe. I.aboratory data consisted of bromsulphalein rrtenticm of 46 per cent, serum albumin 3.4 gm. per cent, thymol turbidity 2.6 units, prothrombin -0 per cent and total serum bilirubin 1.6 mg. per cent. Bleeding ceased spontaneously but recurred massively four days later, on February 12th. Balloon tamponade was instituted and the patient taken immediately to the operating room. Blood rcplacrment prior to surgery was 3,000 cc. Portac*e\,al shunt reduced the porta pressure from 50 to 40 cm. of saline. Fifty per cent of the lumen of the portal vein was occluded by an organized thrombus. Thrombectomy was performed, but it was thought that not all of the proxima1 portion of the thrombus uas removed. Operating time was three hours. The postoperative course was smooth and the patient was discharged on the fifteenth postoperative day. 1Iis subsequent course was good for two months and he returned to part time employment. At this time jaundice developed and steadily progressed over the next month. He died in coma three months after surgery. Autopsy reveaIed a patent. shunt with no evidence of persisting portal vein thrombus. Comparison of liver biopsy specimen with the surgical specimen three months before suggested acute hepatitis, although this was equivocal and was read in different fashions by different pathologists. It, however, is believed that the patient succumbed to homoIogous serum hepatitis superimposed on cirrhosis. Bleeding had not rccurrcd.
CASE IS. A forty year old white woman \vith cirrhosis secondary to alcoholism was hospitalized on October I 3, 1957, with her first episode of bleccling. The bleeding was manifested by melena onI> and had been continuing for several days. Hcmoglobin on entry was 2.5 gm. per cent. Hepatomegaly was detected but splenomegalv and ascitrs were absent. Laboratory data consisted of brnmsulphalein retention of 25 per cent, serum albumin 3.3 gm. per cent and prothrombin 62 per cent. Melena continued and the day after admission the patient was operated upon with a prcoperati\.c* diagnosis of bleeding peptic ulcer. Blood replacement prior to surgery was 6,000 cc. Complete abdominal exploration, including pyloroduodcnotomy and gastrotomy, failed to reveal the presence of an ulcer. The stomach at the time of exploration contained no blood. The liver was obviously cirrhotic, the spleen enlarged and omental vein prrssure increased. Inspection of the lower esophagus through a sigmoidoscope revealed varices but therr was no bleeding at this time. The decision was made that in all probability the source of the bleeding had been varices. Therefore a portacaval shunt was performed which reduced the portal pressure from 38 to 17 cm. of saline. Gastrostomy was pcrformcd. Operating time was four hours. Mild icterus dcvcloped postoperatively and right-sided pleural lluid accumulated requiring several thoracentrses. The patient was discharged on the twentieth postopcrative day. Bleeding has not recurred in a fourmonth follow-up. CASE x. A fifty-seven year old white woman with cirrhosis secondary to alcoholism was hospitalized with her second bleeding episode on October 26, 1957. A left thoracoplasty for pulmonary tubercuIosis had been performed five years before. Hepatomegaly and splenomegaIy without ascitcs \vere present. Laboratory data consisted of 1,rom-
year old white man CASE VIII. A sixty-three with cirrhosis secondary to alcoholism was hospitaIitcd on February I I, 1957, with his third episode of bleeding. Hepatomegaly and spIenomegaIy were Ijrescnt. Ascites was absent. Laboratory data 18;
MikkeIsen suIphalein retention of IO per cent, serum aIbumin 4.2 gm. per cent, prothrombin 90 per cent, alkaline phosphatase 2.7 Bodansky units and tota serum biIirubin 1.4 mg. per cent. Balloon tamponade was instituted and it immediately controlled the bleeding. Tamponade was released forty-eight hours Iater. On the evening of October 29th severe bleeding recurred, baIIoon tamponade was once again instituted and the patient was operated upon the folIowing morning. BIood repIacement prior to surgery was 8,500 cc. Portacaval shunt reduced the porta pressure from 37 to 13 cm. of saIine. Gastrostomy and tracheostomy were aIso performed. Operating time was two hours, thirty-two minutes. Significant jaundice developed postoperativeIy and was slow to resoIve. The patient was discharged on the twenty-sixth postoperative day free of icterus and having been active about the ward. She was rehospitaIized three days Iater in light coma which rapidIy progressed to deep coma, and she died seven days Iater (thirty-six days after surgery). Necropsy was refused. BIeeding had not recurred. A nineteen year oId girl with cirrhosis CASE XI. ptobabIy secondary to hepatitis (perhaps better cIassified as juvenile cirrhosis) was hospitalized on November I, 1957, with her first episode of bleeding. The liver was not paIpable. There was mild ascites and marked spIenomegaIy. Laboratory data consisted of bromsuIphaIein retention of 18 per cent, serum aIbumin 4.3 gm. per cent, prothrombin 50 per cent, and total serum biIirubin 1.7 mg. per cent. Urinary gonadotropins were positive and it was estimated that gestation was of about two and a half months. Bleeding quickIy subsided without the use of baIIoon tamponade. EIective surgery was scheduIed for a convenient time approximately two weeks after admission. The day before the scheduIed eIective time, bIeeding recurred. The patient was immediateIy operated upon without institution of balloon tamponade. BIood repIacement prior to surgery was 5,000 cc. PortacavaI shunt reduced the porta pressure from 50 to 14 cm. of saIine. Operating time was two hours. The postoperative course was smooth; abortion did not occur and the patient was discharged on the eIeventh postoperative day. She has retained her pregnancy and at this writing continues we11 without bleeding during the two and a half months since surgery. RESULTS
AND
COMMENTS
Four of these eIeven patients died in their earIy postoperative period. Of these four, one patient died five hours after surgery in irreversibIe shock. Another patient died twentyfour hours Iater in coma. The remaining two died
and
Pattison in coma on the seventh and thirty-sixth postoperative day, respectively. Death occurring in the fourth patient is puzzling. Her “Iiver profiIe” indicated that she was a better operative candidate than any patient in the entire series. Her postoperative course was compIicated by rather persistent jaundice which, however, cIeared and she was very active about the ward when she was discharged on her twenty-sixth day. Three days Iater she was readmitted in coma, which although somewhat atypica1 of hepatic faiIure persisted unti1 her death seven days later. Necropsy was not permitted. Late death occurred in two of the seven survivors. Both died in hepatic coma. One patient was we11 and asymptomatic unti1 a few days before her death, two years to the day after surgery. The other died three months after surgery due to what is beIieved to have been homologous serum hepatitis superimposed upon aIcohoIic cirrhosis. Five patients are alive two and a haIf, four, twenty, twenty-six and twenty-seven months folIowing surgery. With one exception, varica1 bIeeding ceased abruptIy with compIetion of surgery in a11 patients even though baIIoon tamponade was not empIoyed postoperativeIy in any. In this one exception (Case III) the patient had been bIeeding aImost continuousIy for thirty-two days and baIIoon tamponade had been used intermittentIy for sixteen days preceding the operation. Four hours after surgery hematemesis of 200 cc. necessitated the reinstitution of tamponade for ten hours. FoIIowing remova of the tamponade tube at this time no further bIeeding was encountered. The concIusion that portal decompression as accomplished by endto-side portacava1 shunt can be relied upon to terminate varical bIeeding immediately appears to be warranted. An anaIysis of the postoperative deaths reveaIs two further saIient features: (I) Three of the four deaths occurred in patients who were significantly jaundiced at the time of surgery. In none of the survivors was the tota serum biIirubin over 1.8 mg. per cent. (2) Three of the five patients with aIcohoIic cirrhosis died, whiIe none of those with cirrhosis secondary to hepatitis died. Child’s suspicion that aIcohoIic cirrhotic patients fare more poorIy after surgery than those with cirrhosis due to hepatitis wouId seem to be supported by these resuIts. CriticaI evaIuation further reveaIs that the
Emergency
PortacavaI
“Iiver proMe” in the seven who survived categorized them, in generaI, as satisfactory candidates for eIective portacava1 shunt. With one exception such was not true of those who died. It wouId appear from this series that a further justifiable concfusion is that patients whose hepatic function indicates they wouId be satisfactorv elective shunt candidates, will tolerate this procedure in the presence of acute bIecding. Comparison of our resuIts with the other series of emergency procedures is exceedingly difficuIt. So much depends on the preoperative condition of the patient and this information is avaiIahIe onIy in the series reported by WeIch [5], ChiId [r5] and O’SuIIivan and Payne [r6]. WeIch’s two survivors of transabdomina1 varix ligation could both be considered, on the basis of “Iiver profiIe,” as shunt candidates. His three patients who died couId not be so categorized. The nine patients of ChiId and O’SuIIivan and Payne are not as clear-cut. Three of the six who survived could be cfassified as good shunt candidates, while one was borderline. The two remaining survivors had significant jaundice but both had primary biIiary cirrhosis, and in such patients jaundice does not carry the usuaI ominous significance. Of the three patients who died, one was borderline but two were definitely poor shunt candidates. Fifty cases of Iigation of the varix have been published. Seventeen of these patients died immediately postoperativeIy, representing an operative mortality rate of 34 per cent. Two deaths among the five patients subjected to upper gastric transection resuIts in a 40 per cent mortaIity rate. The mortaIity rate for esophagogastrectomy was IOO per cent for the two recorded cases. Three deaths among ten recorded patients subjected to portacava1 shunt represents a 30 per cent mortality rate. Four deaths among the eIeven patients in this series having portacava1 shunt represents a mortality rate of 36 per cent. It is hard11 justifiabIe to apply percentage mortaIity figures for these small groups, hut it appears that there is little to choose betaecn them in this regard. ‘The average operating time for portacaval shunt in the nine patients in which this information u-as recorded was two hours, thirty-nine This includes the few patients in minutes. whom ;tnciIIary procedures were performed. If the length of the operation is a measure of the stress suffered by the patient, it wouId seem
Shunt
that portacava1 shunt, in this series, would offer no more than the temporizing procedures. The real purpose of any emergency operative procedure is to salvage a Iife that would otherwise be Iost. Thus seven patients in this series might be considered to have been salvaged. A more critical evaluation of the present series, however, might conclude that only one patient was truly salvaged. This patient \vas Case III. Her bIeeding was unreIenting and surely would have caused her death if operative intervention had not been used. Perhaps Case II should also be so cIassilied. Her bIeeding \vas torrenti: and balloon tamponade had previously been unsuccessful. Ejolvever, Meeding on three previous occasions had terminated spontaadmittedly none \j.as as neousI,v, aIthough as the episode Ieadlng to surgery. severe Antithetically, Case x might have survived surgery if more time had been permitted for her liver function to improve. The other three patients who died had been certainly given an adequate tria1 by conservative means. Criticism cannot be directed to an operative attempt to control unreIenting b1eeding. Nevertheless, none of these three patients survived. In the remaining five patients who survived it cannot be categoricaIIy assumed that hIeeding cvould not have ceased quickly and permitted an eIective procedure at a later time. Many cirrhotic patients who hieed, quickly manifest evidence of hepatic encephalopathy. Few if any of these patients will tolerate an> type of emergency surgery. From our expertence a further contraindication to emergcncl surgery is the presence of significant jaundice even though other evidences of hepatic failure are absent. An exception to this hypothesis, is jaundice in primary biliary cirperhaps, rhosis. In this disease jaundice does not carry the same ominous significance that it does in aIcohoIic cirrhosis. The opinion has been expressed by others that perhaps a11 cirrhotic patients with acute bIeeding should be subjected to some type of operative control immediately after hospitalization. WC c;tnnot concur in this opinion. Laboratar?, in addition to clinical cvaIuwtion i5 required in their management, and in many of these patients this would not be immediately avaiIable. Further, some patients who initialI>. hepatic stabiIizec1 deteriorntc into appear coma within tlvelve to twenty,-four hours cvcn though blcaeding has been controlled II\
MikkeIsen
and
6. Hsii, CHI-HO, YANG TOA-HUA and LI SHEN. Surgicat treatment of acute massive hemorrhage from esophagea1 and gastric varices associated with portal hypertension. Chinese M. J., 74: 359, 1956. 7. JULIAN, 0. C. Choice of operation in portaI hypertension with varices. Surg., Gynec. @ Obst., IOO:
tamponade. These patients wouId not toIerate emergency surgery but might be salvaged by conservative means for future elective surgical management. It therefore seems doubtfuI to us that any emergency surgical procedure as practiced today wiII significantIy aIter tota mortality rates from acutely bleeding varices in cirrhotic patients. AdmittedIy, an occasiona patient may be saIvaged. In such a patient it seems reasonabIe to believe that a portacava1 shunt wil1 be toIerated as we11 as any other of the emergency procedures proved to be effective in controIIing bIeeding.
753, 1955. 8. CAMERON, A. and HUGHES, J. F. Massive gastrointestinal hemorrhage due to esophagea1 varices. Delaware M. J., 28: 161, 1956. g. SANGER, P. W. Discussion. Ann. h-g., 141: 647, ‘955. IO. ALLEN, J. G. and HEAD, L. R. The diagnosis of portal hypertension with notes on treatment. S. Clin. Nortb America, p. I rg, 1956. I I. SHUMACKER,H. B. and KING, H. SpIenic studies. II. PortaI hypertension in chiIdren associated with gastroesophageal hemorrhage. Arch. Surg., 65:
SUMMARY
Eleven cirrhotic patients had portacava1 shunt performed during acute bIeeding from esophagogastric varices. Four died in the postoperative period and two died at a Iater date. Five remain alive and have not bIed again. OnIy one patient bIed after surgery in spite of the fact that baIIoon tamponade was not empIoyed postoperativeIy in any. The bIeeding in this one patient was minor in degree. The concIusion that porta decompression as accompIished by end-to-side portacava1 shunt, wiI1 in aImost a11 cases immediately terminate bIeeding seems warranted. Preoperative jaundice was an ominous sign with regard to survival. Comparison of our resuIts with those utilizing emergency temporizing procedures suggests that the operative mortaIity is IittIe different and that effective contro1 of bIeeding is more IikeIy to be obtained by portacava1 shunt. It stiI1 remains doubtfu1 that any of the emergency operations currently in use will significantIy improve the total mortaIity for these cirrhotic patients with severe varica1 bIeeding. REFERENCES I. LINTON, R. R. and ELLIS, D. S. Emergency
2.
3.
4. 5.
Pattison
and definitive treatment of bleeding esophageal varices. J. A. M. A., 160: 1017, 1956. LINTON, R. R. and WARREN. R. The emergency treatment of massive bIeeding from esophageal varices by transesophagea1 suture of these vesseIs at the time of acute hemorrhage. Surgery, 33: 243, ‘953. HAMILTON, J. E. The management of bleeding esophagea1 varices associated with cirrhosis of the liver. Ann. Surg., 141: 637, 1955. COHN, R. Surgical treatment of bleeding esophageal varices. C&Jo&a Med., 83: 348, 1955. WELCH, C. S. Ligation of esophageal varices by the transabdominal route. New England 3. Med., 255: 677, 1956.
499. 1952. 12. NACHLAS. M. M. Treatment of bIeedine esoohaaeaI varices by resection of the Iower esophagus. Arch. Surg., 72: 634, 1956. 13. LIEBOWITZ, H. R., RUZICKA, F. F., ROUSSELOT, L. M. and GREEN, C. H. Panel discussion on esophageal varices-porta hypertension. Am. J. Gastroenterol., 27: 325, 1957. 14. LYONS, C. and PATTON, T. B. BIeeding esophagea1 varices. Surgery, 39: 540, 1956. 15. CHILD. C. G. The Shattuck Iecture: the oorta1 circuIation. New England J. Med., 252: 837: 1955. 16. O’SULLIVAN, W. D. and PAYNE, M. A. The emergency portacaval shunt. Surg., Gynec. ti Obst., 102: 668, 1956. 17. HUFNAGEL, C. A. SurgicaI treatment of portal hypertension. Am. J. Gastroenterol., 23: 522, 1955. v
I
Y
DISCUSSION RAY COHN (San Francisco, CaIif.): It is fair to say that the general consensus of surgeons is that portacaval anastomosis is the best surgica1 method to Iower the porta venous pressure. It is also fair to say that it is the genera1 consensus of physiologists that diversion of the porta bIood ffow is a most damaging procedure with respect to function of the Iiver. The enzyme activities of the liver are the latest in the long Iist of functions of the liver which have been shown to be damaged by shunting procedures [r]. In fact, the most amazing feature of this disease to both cIinician and physioIogist is that a patient shouId be as cIinicaIIy we11 as some cirrhotic patients appear to be, with nothing but a shrunken noduIar liver about one-fifth norma size when seen at operation. In the earIy days of this operation many patients who were poor risks did not survive. Because of this fact, criteria such as that of Linton were established with a consequent fall in the immediate mortality rate. It then became the fashion to compare the results after operation on this highly
1MANNING and DELP. Management of hepatocerebra1 intoxication. New EngLand J. Med., 258: 62, 1958. 190
Emergency
PortacavaI
selected group of cirrhntic patients with the dire consequences of massive hemorrhage on the unselrcted series of those with cirrhosis. This fallacy has been recognized by surgeons reporting series of portacaval anastomosis, the latest of whom has bcrn Hallenbeck. From collected data of the San Francisco Hospital it can be noted that 86 per cent of the patients with gastrointestinal bleeding who died, did so on the first bleeding episode. Now of the patients who survived and who satisfied the usual laboratory and clinical criteria for portacaval anastomosis, eleven or roughly 30 per cent survived five years or more with medical therapy. The authors have recorded eleven cases, the longest follow-up being twenty-seven months. Only five patients survived, less than half. Obviously the authors’ eleven cases were extremely poclr risks and no one can state with assurance what their fate would have been had operation not been performed. I would agree with the authors that only in Case 111 does it really appear that operation was life saving and in this patient cirrhosis sc,emcd to be minimal. It would appear to mc that the number of patients for whom portacaval anastomosis is indicated is extremely smaI1, and that when the operation is applied to the desperately ill, acute bleeder the selection of the proper cases would appear even more dificuIt, probably impossible. .IACK M. FAHRIS (Los Angeles, Calif.): Dr. Smith and I have been interested in this problem for some time and have followed the work of Drs. Mikkelsen and Pattison with a great deaI of interest. Although \Vc’ ha\c had less experience with portacaval anastomosis, we have operated upon three such patients as emergcncics where the balIoon had been in place, for a number of days and attempts to release it had been followed by recurrence of the hemorrhage requiring further transfusion. Because of the conviction that other operations do not approach the fundamental problem, we carried out emergency operations. I would like to present one case briefly, if 1 may, to support the views of Drs. lZlikkelsen and Pattison that survival may at least be enhanced by emergency portacaval anastomosis in the exceedingly poor-risk patient.* This man, in his forties, had typical alcoholic cirrhosis and was operated upon at another hospital with a mistaken diagnosis of bleeding ulcer. At that opcraticm the diagnosis of cirrhosis was established by biopsy and measurements of portal pressures. Following operation conservative treatment was instituted and during the ensuing five days an attempt was made to release the balloon on three dift’crent occasions; each attempt was followed
a A slide was employed by Dr. Farris illustrating the stable pulse rate between 68 and 78 throughout portacuval anastomosis at temperature of 31°c.
Shunt
by a recurrrnce of brisk hemorrhage requiring transfusion. At this point the patient was transferred to us via ambulance, a distance of about $0 miles, having received eighteen blood transfusions during this phase of his postoperative course. Actuallv, he was getting blood while being transported via ambulance and during this his wound disrupted. He arrived, semi-stuporous, with copious amounts of ascitic fluid draining from his abdominal wound. There was also clinical jaundice. Although hc presented what in the past would be considered absolute contraindications for portacaval anastomosis, it was apparent that he could not tolerate the balloon much longer. We were aware of cases of perforation of the esophagus from prolonged presence of the tamponade. Therefore we decided to operate upon him as an emergency, and with the conviction that portacaval anastomosis was tht operation of choice we determined to carry it out under hypothermia. His preoperative pulse was between IX) and 130 and after one hour of cooling in ice it was full and strong at 68. During the operation of about two and a half hours’ duration the pulse remained between 68 and 80 with a lo\% temperature of 31”~. We were able to revise his previous wound disruption and perform an end-toside portacaval anastomosis; and although there was some increase in his jaundice postoperatively, he made a surprisingly good convalescence and was able to be clischarged from the hospital on his twenty-first postoperative day. He has now been followed up for two years and is pursuing his usual occupation. There has been no recurrent bleeding. Although observations such as this arc not subject to scientific control, we believe that recovery in this patient would have been unlikely without portacaval anastomosis. Furthermore, we believe that hypothermia has a definite place in the management of this type ol case. The cooling with resultant lowering of mrtabolic requirements provides a logical protection to the liver both from the standpoint of lowering its metabolism and also protecting it from the trauma of the operation and toxic effects of anrsthetic agents. The most important benefit, howcvcr, to IX derived from the cooling technic is the prrdictablc and consistent lowering of the pulse rate. To undertake a major and prolonged surgical operation in a patient with a preoperative pulse in excess ot’ I 20 in the past has been fraught with hazard; the pulse invariably increases during the operation with a diminished stroke volume and cardiac output progressing frequently to cardiogcnic shock. 1t appears that cooling nullifies the corticosteroid and autonomic response which in the usual patic*nt arc normal and acceptable defense mechanisms. \\‘hcn fever and tachycardia, however, exceed the bounds of homecjstatic control, survival is endangered.
MikkeIsen
and Pattison
We wouId Iike to comphment the authors on the exceIlence of their presentation. HORACE J. MCCORKLE (San Francisco, CaIif.): I enjoyed Dr. Mikkelsen’s presentation and Iearned a great dea1 from it, as I have also from his recent pubIications. Last year I was abIe to rescue a patient with persistent bleeding from esophageal varices by an operation that he apparentIy does not think too much of at this time, namely, the Tanner procedure. This young woman was at term and she began to bIeed from esophagea1 varices immediately after the dehvery of her chiId. She was treated with the Sengstaken tube over a period of about three weeks, and every time the tube was removed or Ioosened she wouId almost bleed to death. Finally we performed a Iaparotomy and were unable to find either porta or spIenic vein. Her spIeen had been previousIy removed and there was no possibility of performing a porta vein decompression procedure. Therefore al1 the numerous, very Iarge veins on the undersurface of the patient’s diaphragm were Iigated and the Tanner procedure was performed. This was about eight months ago and she has had no subsequent hemorrhage. WILLIAM P. MIKKELSEN (cIosing): I shouId like to refer to Dr. McCorkIe’s suggestion of the vaIue of a Tanner procedure (gastric transection). It is true that about two years ago we recommended this operation for patients with portal hypertension in whom major venous thrombosis prevented the construction of a venous shunt. Eight patients have been so treated. Two of these have since exsanguinated from recurrent bleeding and five others have bled severeIy one or more times since surgery. The remaining patient has not re-bIed but persists in demonstrating Iarge varices. This aIarmingIy poor experience has Ied us now to abandon this operation. Dr. Farris’ suggestion that hypothermia wouId
be of vaIue is we11 taken. CertainIy it has theoretical advantages. Hypothermia was empIoyed with two patients in this series. Both succumbed, but admittedIy, both were in the poor risk category. Dr. Cohn for years has insisted that portacaval shunt does not proIong the Iife of these patients and that it does not result in Iasting reduction of porta pressure. Nevertheless, it remains our delinite impression that Iife is proIonged and certainIy that the threat of recurrent bleeding is aImost eIiminated. Persistence of reduction in porta pressure by end-to-side portacava1 shunt is demonstrated in reviewing pressure studies in nine patients before and up to ninety weeks after eIective surgery.* Catheterization in these cases was performed by Drs. ReynoIds and Redeker of the Department of Medicine at the University of Southern Cahfornia. Preoperative wedged hepatic vein and operative portat pressures were compared with postoperative wedged hepatic vein ‘pressures and portal pressures obtained by passage of the catheter through the anastomosis and out into the porta vein. It wiI1 be noted that folIowing surgery wedged hepatic vein pressure was reduced by about one-third but stil1 remained at a moderateIy hypertensive Ievel, whereas the direct porta pressure in a11 instances was reduced to normal. ,It shouId be pointed out that after division of the portal vein, wedged hepatic vein pressure no Ionger reflects portal pressure and that its minimal reduction after a shunt reflects directly the reduced hepatic bIood ffow ordy. ClinicaIIy, nearly a11 our shunt patients, now numbering over 120, have demonstrated disappearance of varices by roentgenography. Further, recurrent gastrointestinal bIeeding has been a rather rare occurrence. Thus our confidence in end-to-side portacava1 shunt continues. * Dr. MikkeIsen illustrated his discussion with a slide.
192