|Home | About | Journals | Submit | Contact Us | Français|
Two general kinds of liver transplantation have been attempted clinically. With one approach, the host liver is removed and replaced with a homograft (orthotopic liver transplantation); alternatively, an extra liver is inserted at an ectopic site (auxiliary homotransplantation). This review concerns only orthotopic liver transplantation.
The first effort to replace a human liver was made at the University of Colorado on March 1, 1963. That patient died as did four others during the next 7 months (1, 2) (Table 1). In September, 1963 and January, 1964, other unsuccessful attempts at liver replacement were made in Boston (3) and Paris (4) (Table 1). The first clinical trials were not frivolously undertaken. Members of the Boston and Denver teams had developed techniques for liver replacement in dogs in the late 1950s and, in both laboratories, research on liver transplantation had been continuously performed for more than 4 years.
Nevertheless, the consecutive failures in three institutions halted all clinical trials until our sixth and seventh equally unsuccessful attempts in October, 1966 and May, 1967. Finally, on July 23, 1967, the first extended survival of a human recipient was achieved (5). The patient, a 1½-year-old girl, lived for more than 13 months before dying of metastases from the hepatocellular carcinoma for which she had been treated. From then until the first week of May, 1982, we have treated 230 more patients, 163 at the University of Colorado and 67 at the University Health Center of Pittsburgh for a total of 237. The yearly frequency of transplantation throughout our experience is shown in Figure 1. The highest number of 30 was reached in 1981; the number of new cases in 1982 is projected at between 60 and 80.
A signal event in the development of orthotopic liver transplantation occurred on May 2, 1968 when Calne and Williams of the University Hospital at Cambridge and the King’s College in London, respectively, treated the first patient in their program (6, 7) which had since generated more than 125 well-studied cases. Calne’s contributions in experimental renal transplantation has been a major factor in the early development (8) of chemical immunosuppression, without which transplantation of any organ was not realistic until 2 decades ago. The fact that both men had the personal qualities to be able to accept defeat or victory with equal grace was fortunate since failure was the dominant theme with all such efforts until recent times.
In this review, emphasis will be placed on the more than 350 cases in these two series. However, between 1968 and 1978, other single attempts or small series were reported from Boston (9, 10), Los Angeles (11), Montreal (12), Bonn (13, 14), Sao Paulo (15), Calgary (16), New York City (17), Richmond (18), Minneapolis (19), Manchester (20), and Oslo (21). It is probable that these documented cases were a minority of those attempted in that decade, exclusive of the English and Colorado series. Since 1978, programs have been reopened or started new from which important information can be expected. In the United States, clinical liver transplantation programs are active at the University of Minnesota (J. S. Najarian, unpublished observations) and the University of Tennessee (J. W. Williams, unpublished observations). Series of six or more cases each have been reported or are in preparation from Holland (22), East Germany (23), West Germany (24, 25), France (26), and the Republic of China (H. Q. Hong, unpublished observations). The number of cases in the West German series (25) has passed 50, but the results have not yet been published in detail.
The first known efforts at experimental orthotopic transplantation of the liver were made by Dr. Jack Cannon of Los Angeles (27). This report was so brief that it lacked a title, description of methods, and notation of the animal species. The animals did not survive operation. In June, 1958, a program of orthotopic transplantation of the canine liver was initiated at the Peter Bent Brigham Hospital under the direction of Dr. Frances D. Moore (28), and in August, our first experiments in the same species were begun at Northwestern University in Chicago (29,30).
The technical problems of liver replacement and the features of rejection in untreated canine recipients were delineated (28–31). Eventually, using immunosuppression with azathioprine (32–34) and antilymphocyte serum or its antilymphocyte globulin (ALG) derivative (32, 35–38), chronic survival was achieved in mongrel dogs of which one lived for almost 12 years (39).
In 1965, Garnier of Paris (40) made the important observation that rejection of pig liver homografts was mild in comparison to that in dogs. Several porcine recipients lived for long times without immunosuppression. Workers in Bristol (41), Cambridge (42), and Denver (32) promptly confirmed Garnier’s work. The value of the pig for transplantation research has been demonstrated frequently in the investigations of Calne and his associates.
Hundreds of significant experimental studies in various species have since been published. Review of this work will not be attempted although some of it has influenced the clinical trials as will be noted later.
A training period in the animal laboratories is an important preparatory step for teams planning clinical programs; however, not all of the experimental techniques are identical in humans. Our methods of orthotopic liver transplantation (1, 5, 32, 43–52) and the modifications introduced by Calne et al. (6, 47, 53) have been described.
The operation is simple in principle (Figure 2), but its execution has been exceptionally difficult because of the almost invariable debilitation of the recipients and the profoundly abnormal vascularization patterns caused by portal hypertension in endstage liver disease. Defects in clotting have been present in most cases (1, 32, 54, 55), and adhesions or other alterations secondary to previous operations are often complicating factors. With such a background, it is not surprising that the postoperative care of many patients has been an exercise in resuscitation (32). The most common difficulties have been pulmonary insufficiency (requiring mechanical ventilation for several months in some cases), renal failure with massive fluid shifts, and persistent clotting abnormalities. These problems are managed with conventional methods of intensive care with emphasis on biochemical and hemodynamic monitoring. Recovery can be expected from encephalopathy and the hepatorenal syndrome (50).
The ability to survive this critical period depends upon what has transpired in the operating room. Thus, in the following remarks, we will touch upon details of surgical technique which require reemphasis or points of view which have changed from those expressed in the past.
Until 1976, techniques for preserving the liver either severely limited the acceptable time of cold ischemia or were too complicated for use in outlying hospitals (32). In 1976, simple methods that permitted reasonably long storage were developed, and clinical trials were started. We have employed an electrolyte (Collins) solution with a composition similar to that found intracellularly (57). The Cambridge-King’s College team uses a plasma solution for similar cold infusion of the homograft (58). In dogs, the two methods yielded similar results (57) and allowed safe preservation for up to 12 hr. These techniques permit shipment of livers from city to city. More than two-thirds of the livers used in the Pittsburgh program have been obtained outside of the normal procurement area for this region; the longest transit being from Phoenix, Ariz. Efficient air travel arrangements are necessary; to meet this objective, several Pittsburgh corporations have donated their private jet airplanes.
Liver procurement outside of the local area has depended on cooperation with other procurement centers whose main function was previously to provide cadaveric kidneys. Fear that kidney grafts would be jeopardized by giving too high a priority to the liver has been allayed by standardization of techniques which protect all organs equally and which can be adapted to the local surgeons’ wishes (59).
A midline incision is made from the pubis to the neck, and the sternum is split. The structures entering and leaving the liver are skeletonized, and the necessary preliminary steps for kidney removal are taken in the presence of an intact circulation. Then a cannula is placed through the splenic vein into the superior mesenteric vein. Rapid infusion of cold lactated Ringers’ solution is used to start cooling of the liver. After 1 or 2 liters have been infused, cold Collins solution is infused through the terminal aorta at the same time as the distal thoracic aorta is cross-clamped and the cadaveric donor is exsanguinated from a cannula previously placed in the distal inferior vena cava. After removal, the cold liver is flushed with Collins solution and protected by a plastic bag which is placed in slushed ice. The performance of kidneys removed from liver donors with this technique was as good or better than has been achieved in most centers with kidney removal alone (59). With minor modifications, it has also been possible to remove the heart in addition to kidneys and liver from seven donors.
During 1981 when 30 liver homografts were used, there were 176 offers of organs. The inability to use the other 146 livers was usually because there was a disparity in size between the donor and recipient, the donor blood group was incompatible with the recipient, or Pittsburgh facilities for liver recipient care were saturated. An actual shortage of pediatric donors has existed in the 1- to 5-year age group.
The most demanding aspect of liver transplantation is removal of the diseased native organ. The technical difficulties are usually determined by the underlying disease. The easiest situations are in patients with primary hepatic malignancies or primary biliary cirrhosis. The most difficult are in recipients with the shrunken livers of macronodular cirrhosis or in patients with multiple previous operations. On the average, hepatectomy is easier in infants and children than in adolescents and adults.
The preexisting pathologic changes often necessitate deviations from a standard plan; however, the first step is to find the hilum and dearterialize the liver which expedites hilar dissection and slows hemorrhage from the liver surface. During hilar dissection, the bile duct is transected as high as possible so that the option of duct-to-duct anastomosis is retained. The portal vein is left intact until later, in order not to aggravate the portal hypertension.
The inferior vena cava below the liver is encircled with minimal dissection. The left triangular and falciform ligaments are incised until the suprahepatic vena cava is identified. The suprahepatic vena cava is encircled to allow placement of a cross-clamp.
If all of these maneuvers are successfully executed, the liver can be isolated from the circulation by cross-clamping vessels which have been encircled but are intact. Cuffs of the suprahepatic and infrahepatic vena cava are fashioned from these vessels as the liver is removed. Sufficient infrahepatic vena cava is not difficult to obtain, but development of an adequate suprahepatic cuff may require tailoring of the vena cava which is mobilized from within the liver (46). The technique of isolating the liver and peeling it out in a bloodless state permits all residual tissue connections of the right triangular ligament and the bare areas (including the right adrenal vein) to be ligated under direct vision. The penalty with this approach is an increase (usually about 30 min) in the time of portal and vena caval cross-clamping, compared to cross-clamp time with previously described techniques (1, 32, 43).
Such cross-clamping is usually tolerated in patients with chronic disease in spite of major declines in cardiac output and variable hypotension (60): the same thing has been demonstrated in dogs subjected to chronic bile duct obstruction (61). Because of this, venous bypasses which were used in our first cases were discontinued (1, 2).
Some patients are jeopardized by the venous cross-clamping. If severe hypotension occurs after cross-clamping, Calne et al. (52, 53) recommend femoral vein-to-femoral artery bypass with an intervening oxygenator. About 10% of the English patients are so treated. One death in our last 67 patients (OT 233, Table 14), as well as a cardiac arrest which was successfully treated, may have been avoided by this precaution.
The fact that most patients recover from portal and vena inferior caval cross-clamping may have created a false impression about the safety of this practice. Usually, there is gross swelling of the intestine during the period of occlusion. Subsequently, many patients suffer from third-space fluid sequestration and postoperative renal failure. The extent to which these complex physiologic events contribute to the high perioperative mortality has not yet been delineated. For this reason, we returned in recent cases (not reported in this communication) to the practice of venous bypass which had been abandoned. Cannulas are placed into the inferior vena cava through an iliac or femoral vein and into the portal system through the open end of the transected portal vein. During the anhepatic phase, the blood is returned to a reservoir and pumped to a large vein in the neck or arm. Although this kind of bypass requires total body heparinization, bleeding has not been excessive since the “heparin effect” can be effectively reversed even in patients with severe liver disease.
Difficulties with biliary tract reconstruction were frequently lethal in our experience (43, 62, 63) and in that of the English workers (7, 49–51) until the mid-1970s. Anatomic studies by Terblanche et al. (64) suggest that deficient blood supply of the homograft duct system may be a contributory factor; however, our principal problems were due to the frequent (and inappropriate) use of cholecystoduodenostomy and to failure to diagnose the complications. The latter deficiency was resolved with the frequent use of postoperative cholangiography (Figure 3) and reoperation, if necessary. The incidence of complications has been reduced with better primary reconstruction.
We now consider duct-to-duct anastomosis to be the procedure of choice (Figure 2A). When this is not feasible, the duct is anastomosed to a Roux limb of jejunum (Figure 2B). With either technique, the homograft gall-bladder is removed. The results of biliary tract reconstruction in the last 67 cases are shown in Table 2. Using choledochocholedochostomy or choledochojejunostomy, an eventually satisfactory result was obtained in more than 95% of cases, although the use of an internal stent (instead of a T-tube stent) for duct-to-duct anastomoses led to a high rate of reoperation. Only two deaths resulted directly from biliary tract reconstruction.
Since 1976, Calne et al. (48) have used a technique in which the homograft common duct and gallbladder are fashioned into a common channel and anastomosed to recipient common duct or a Roux limb. Waddell and Grover (65) had described such a common channel procedure for difficult biliary tract problems.
The spleen was removed in most of our early patients, in part to achieve immunodepression, but mainly to relieve hypersplenism and leukopenia which prevented the effective use of azathioprine or cyclophosphamide (32). With the advent of cyclosporin A, splenectomy was discontinued.
Technical complications that have occurred after liver transplantation are legion. All have been potentially treatable except for irreversible ischemic injury of the graft or early loss of its blood supply. Retransplantation is the only hope for a patient who has been given an irreversibly damaged or devascularized organ. In one of our patients, portal vein stenosis at the anastomosis was diagnosed 8 months after transplantation, and, at reoperation, the stenosis was successfully resected in spite of the fact that thrombosis had occurred requiring thrombectomy (Figure 4).
The time constraints of liver preservation and urgent recipient need usually preclude systematic efforts at tissue matching. With the random donor-recipient pairing, good matches at the A, B, and DR loci have never been obtained.
Hepatic transplantation has been performed (45, 50, 51, 66) against the recipient anti-donor T-warm antibodies which cause hyperacute rejection of kidney homografts. To our knowledge, hyperacute rejection of the liver has never been seen.
However, much more experience will be required before concluding that acceptance of “positive cross-matches” against T -warm recipient antibodies is without jeopardy. Data from the first 53 Pittsburgh cases is summarized in Table 3. Patients for whom cross-matches could not be performed or whose sera cross-matched negative with recipient cells had better results than those with cross-match positive donors. Although hyperacute rejection was not observed in the latter recipients, the postoperative courses were stormy. At least two of the livers developed delayed massive necrosis.
From experimental studies, it is known that the liver is resistant to hyperacute rejection. However, in animal xenograft models in which the recipient has performed heterospecific cytotoxins, humoral antibody rejection of the liver is merely slower than that of the kidney; the mechanisms of destruction are the same. The extent (if any) to which the outlook after clinical liver transplantation is depreciated by preformed antibodies is unknown; many liver candidates have widely reacting T-warm cytotoxic antibodies which reflect sensitization by previous blood transfusions. For the time being, we continue to treat such highly sensitized patients.
If donor-recipient ABO blood group incompatibility exists, renal grafts can be destroyed by isoagglutinins (67). Liver grafts are resistant to this kind of hyperacute rejection (45). In two of our Pittsburgh cases, two incompatible donors were used (Table 4). The course of one of these patients is shown in Figure 6. We have tried to avoid this practice since blood group compatible donors can usually be found.
It was important to demonstrate in animals that chronic survival is possible after liver transplantation under immunosuppression. However, such laboratory investigations contributed relatively little to the immunosuppressive regimens which have been used clinically. Liver transplantation, either in animals or man, was too complex to be used as a model to evaluate drugs or drug combinations. Instead, all methods to prevent or reverse rejection of whole organs have depended upon observations after renal transplantation.
The immunosuppressive protocols that have been developed for human renal transplantation are summarized in Table 5, exclusive of the historically important trials with total body irradiation (68). Because the first genuinely promising drug, azathioprine (8), proved to be effective only rarely when given alone (69), the “modern” era was not entered until it was realized that azathioprine and prednisone have an additive (or possibly synergistic) effect (67, 70–73). At the outset, our policy was to begin therapy after renal transplantation with azathioprine and to add high doses of prednisone with the first signs of rejection (70). Because it was rare to escape rejection even after transplantation from closely related donors, our recommendation (67) was to begin treatment with both drugs immediately after transplantation with a gradual reduction in prednisone. Such “double-drug therapy” has been the most commonly used immunosuppression for almost 20 years.
With transplantation from consanguineous donors under double-drug treatment, chronic renal graft function was achieved almost immediately in more than two-thirds of cases. However, during the first year after cadaveric renal transplantation, the graft loss rate in multicenter compilations remains at about 50% (74, 75). Liver recipients for whom cadaveric donors were obligatory, and who did not have the option of fall-back maintenance on an artificial organ therapy analogous to renal dialysis in the event of rejection, were confronted with a bleak outlook.
Between 1963 and 1979, several alternative therapeutic programs were introduced for renal transplantation (Table 5); all were modifications of or additions to the original double-drug therapy. A promising approach involved lymphoid depletion with ALG (36) which was given i.m. or i.v. as an adjunct to azathioprine and prednisone during the first few weeks or months when the risk of rejection is the greatest. “Triple-drug therapy” has been the second most commonly used technique of immunosuppression. A conceptually important but pragmatically inconsequential detail was that cyclophosphamide could be freely substituted for azathioprine (76). The results of 1-year graft survival after cadaveric renal transplantation under triple-drug therapy were improved in most centers. After the discontinuance of ALG, there was an unacceptable rate of delayed rejection which, not surprisingly, also occurred after liver transplantation (32). The alternative of temporary lymphoid depletion with thoracic duct drainage (TDD) (77) in preparation of patients for cadaveric renal transplantation (78) had the same disadvantage (79). Efforts to use preoperative TDD in liver recipients usually created insurmountable problems because of the prodigious quantities (as much as 2 liters per hr) of thoracic duct lymph which patients with hepatic insufficiency produced (80). Lymphoid depletion by total lymphoid irradiation for conditioning before grafting (81, 82) has not been tried in liver recipients.
There was widespread discontent with all techniques of immunosuppression from 1963 to 1978. Many kidney transplant surgeons attempted to escape the consequences of this therapeutic cul de sac by exploiting developments in tissue typing and matching, or by systematically conditioning prospective renal recipients with preoperative blood transfusions. The former efforts yielded disappointing results after cadaveric kidney transplantation; the latter practice of conditioning by transfusion allowed an increased success rate in patients not accidentally sensitized during their preparation. In any event, liver transplantation candidates usually were too ill to wait for a well-matched liver or to undergo stages of preoperative preparation. For future trials of liver transplantation, it was necessary to hope for better immunosuppressive drugs. This did not seem realistic until the advent of cyclosporin A.
Cyclosporin A is an extract from the fungi Cylindrocarpon lucidum and Trichoderma polysporum. It was discovered and characterized biochemically by scientists at the Sandoz Corp., Basel, Switzerland. Cyclosporin A was shown to be immunosuppressive by Borel et al. (83, 84) in mice, rats, and guinea pigs. The drug depressed humoral and cellular immunity with a preferential and quickly reversible action against T-lymphocytes. These effects were not accompanied by bone marrow depression which frequently limits the doses of azathioprine and cyclophosphamide. The unusual effectiveness of cyclosporin A in preventing or delaying rejection of mouse skin homografts was demonstrated by Borel et al. (83, 84). Analogous observations in which heart, kidney, liver, and pancreatic grafts were protected in rats, rabbits, dogs, and pigs were reported by Kostakis (85), Calne (86–88), and Green (89) and their associates.
When cyclosporin A was first used in patients by Calne and coworkers (90,91), it was hoped that no other drug would be routinely required. Our dissenting opinion is that cyclosporin A should be combined with steroid therapy from the outset (92, 93). The extent to which steroids are required with cyclosporin A remains to be clarified, but it is clear that kidney survival of greater than 80% can be expected 1 year after primary cadaveric transplantation (93,94). Long-term follow-up of our original recipients and those of Calne has not shown a tendency for patients under cyclosporin A to have “catchup” graft losses or unexpected delayed morbidity from other causes. We and Calne have not had the disillusionment reported by Carpenter et al. (95) and Sweny et al. (96) in their first trials with cyclosporin A for cadaveric renal transplantation.
As new teams begin using cyclosporin A, it will be important to avoid unrealistic expectations about early convalescence that could be engendered by the high success rates achieved after cadaveric renal transplantations. In a recent analysis of 42 consecutive cadaveric renal recipients (97), only one-third had a completely uneventful recovery. Of the remainder, most developed rejection which was usually reversed with augmented steroid therapy. In every case, the major differential diagnosis was rejection vs. nephrotoxicity from cyclosporin A.
Nephrotoxicity of cyclosporin A was first noted by Calne (90, 91) and Powles (98), and has been confirmed elsewhere (92, 99). To sharpen the interrelationship between therapeutic effect and toxicity, pharmacologic (cyclosporin A blood and/or plasma levels) and immunologic monitoring were advocated by Keown (100), Rynasiewicz (101), and Kahan (102). Our techniques of management have not depended upon these monitoring techniques. Fortunately, nephrotoxicity usually has promptly reversed with reduction of cyclosporin A doses. As a last resort, a change from cyclosporin A to azathioprine has been made but at an increased risk of rejection (99).
Most other side effects of cyclosporin A (90–94) are not serious and include gingival hyperplasia, tremor, regional flushing or vague abdominal discomfort just after drug ingestion, and development of breast fibroadenomas in women. Although hepatotoxicity occurs in about one-fifth of cases (103), it is rarely serious enough to necessitate a change to azathioprine.
The most publicized question about cyclosporin A concerns its potential oncogenicity. It has been known for 15 years that conventional immunosuppression results in an increased incidence of de novo tumors, of which approximately one-third are lymphomas (104). Early reports by Calne (91) of lymphoma development in patients treated with cyclosporin A were not surprising, although the incidence of three lymphomas in 34 recipients was sobering. Calne attributed this high incidence to the concomitant use of other cytotoxic drugs and possibly steroids. In our own experience with cyclosporin A and steroid therapy in almost 200 cadaveric renal recipients, there have been two lymphomas. One was an incidental finding at autopsy following a fatal infection (93). The other was successfully treated by intestinal resection after it had caused a perforation (97). To our knowledge, no de novo epithelial tumors have been seen in renal recipients. As experience with cyclosporin A accumulated worldwide, the spectre of this drug being a spectacular tumor producer has receded. None of the liver recipients treated with cyclosporin A and steroids has developed new malignancies.
Two patients are known to have been given orthotopic liver grafts without immunosuppression or with steroid therapy only. The first patient was in the Cambridge series (7); the other was treated in Oslo (21). One factor in these decisions may have been the demonstration in dogs (32, 33) and pigs (32, 40–42) that rejection of liver grafts was less severe than that after renal transplantation. In addition, the English recipient had hepatitis which it was feared would be reactivated by immunosuppression. Both organs promptly failed with early death of the patients.
All other patients were given some variant of the double- or triple-drug treatment summarized in Table 5. Our first five recipients and occasional ones later were treated with azathioprine and prednisone. The same treatment was used for almost all patients in the Cambridge series from 1968 through 1979.
Triple-drug treatment was used in the majority of recipients from 1966 through 1979. The most common regimen was azathioprine, prednisone, and a variable course of i.m. ALG which was begun on the day of operation. The duration of ALG was usually limited to a few weeks because of sensitization of the recipients to horse, rabbit, or goat globulin; however, treatment with ALG was continued in some cases for 6 to 12 months.
In a modification of triple-drug therapy (Table 5), cyclophosphamide instead of azathioprine was given to 16 patients (OT 42–57) from March, 1971 to August, 1972 (105). Six (37.5%) recipients lived for at least 1 year, and four (OT 42,46, 53, 56) are still alive more than 10 years later. From a few months to several years after transplantation, all surviving patients were switched to azathioprine. Because the results during this period were not markedly different than with the original triple-drug management, cyclophosphamide was not further used as a first line drug.
In 1978 and 1979, TDD was used as an adjunct to therapy with azathioprine and prednisone in 21 patients (80). TDD was started 10 to 18 days before transplantation in 2 patients, on the day of operation in 17 patients, and 2 and 4 weeks after transplantation in the other two. Six (31.6%) of 19 recipients who had TDD started prior to or on the day of transplantation lived for at least 1 year and five are alive after 3⅔ to 4⅓ years. The management dilemma was that in the kidney transplantation model, TDD was ineffective unless applied at least 3 weeks in advance of transplantation (78), but potential liver recipients could not tolerate the chronic, high-volume thoracic lymph drainage associated with hepatic disease. If TDD is to be tried again in liver transplantation, a closed system will be required in which lymphocytes can be removed in transit without the necessity for lymph removal and later reinfusion.
Calne et al. (91) were the first to use cyclosporin A for liver transplantation. In their first two cases, cyclosporin was used alone. However, most of their experience has been with delayed administration of the drug (53, 106). Azathioprine (1.5 mg per kg per day) and prednisolone (0.4 mg per kg per day) were used until renal and hepatic functions were adequate. Then, cyclosporin A (10 mg per kg per day) was begun, and the steroid dose was slowly reduced to zero. The supervention of acute rejection during treatment with azathioprine and prednisone was troublesome and, in the last review by Calne et al. (53), they recommend shortening this period.
Our practice (107–109) has been to start cyclosporin A a few hours preoperatively with an p.o. dose of 17.5 mg per kg (Figures 5 to to7).7). Cyclosporin A is continued daily, but with reduced i.m. or i.v. quantities (Figure 7) until p.o. diet is resumed. Subsequently an p.o. dose of 17.5 mg per kg per day is given, usually with half the daily dose every 12 hr. The quantities are reduced subsequently if toxic manifestations develop, of which nephrotoxicity has been the most important (Figures 5 and and6).6). Usually, steroids are also started on the day of operation. For adult patients who leave the operating room in relatively good condition, a 5-day burst of prednisolone is given, starting at 200 mg and stopping with a maintenance dose of 20 mg per day (Figure 5). Further reductions of cyclosporin A and steroid doses are made on an individualized basis. Initial and maintenance therapy with steroids are reduced in infants and children (Figure 6).
If the patient was in poor postoperative condition, the initial burst of high-dose steroid therapy was omitted for a few days or greatly reduced (Figure 7). A few patients suspected of having nephrotoxicity from cyclosporin A were switched temporarily to azathioprine with resumption of cyclosporin A treatment as renal function improved (Figure 6). With less severe renal impairment (Figures 5 and and7),7), the dose of cyclosporin A was reduced. No patient has been changed to azathioprine permanently. If rejection occurred in spite of this therapy, the principal responses have been to administer intermittently large i.v. doses of hydrocortisone (or prednisolone) (Figures 5 and and7),7), repeat the original 5-day burst of steroids (Figure 7), and settle at a higher maintenance level of steroids. Although cyclosporin A does not permit much dose maneuverability, it has sometimes been possible to increase the amounts given despite the risk of nephrotoxicity.
To interpret much of the statistical information to be presented later in this review, it will be necessary to describe the features of homograft rejection as these have been perceived by the pathologist and surgeon.
Despite treatment with any of the immunosuppressive regimens so far used, rejection of human hepatic homografts has been observed. A clear picture can be pieced together from studies of orthotopic liver transplantation in rats, dogs, pigs, baboons, and humans of the sequence of pathologic events which occur when a modified or unmodified recipient rejects a liver graft (110–113).
In untreated members of all species (the principal observations have been in dogs), there is a quiescent phase of at least 2 or 3 days during which only rare small lymphocytes are found in the tissue spaces of the liver which remain normal except for nonspecific changes. However, during this period, large pyroninophilic cells start proliferating in the paracortical zones of the host lymph nodes. About 3 days after transplantation, lymphoid cells begin to leave portal vein tributaries randomly throughout the graft. The venous endothelium is separated from the basement membrane, and fibrin collects in the subendothelial space. After passing through the vessel wall, lymphocytes accumulate in the portal tracts. Smaller numbers of lymphoid cells migrate through the walls of the central vein and the endothelial lining of sinusoids. These cells invade the space of Disse and some enter between hepatocytes. Immunoglobulins are rare in the cytoplasm of the infiltrating cells at this time.
Associated with cellular infiltration, the cells of many sinusoids disintegrate, blood flow through the liver begins to decrease, and some centrilobular cells die. As the centrilobular necrosis progresses to midzonal necrosis, liver function becomes affected. Inspissated bile appears in surviving bile canaliculi and lipid droplets accumulate in the hepatocytes around the portal tract. Shortly before the death of the untreated recipient, foci of fibrinoid necrosis sometimes occur in the walls of small branches of the hepatic artery associated with deposition of immunoglobulin and complement in the intima and media.
When rejection is mild, as in pigs (115), or is modified by immunosuppressive agents, as in dogs and humans (110, 112–114, 116), destruction of hepatocytes ceases, cellular infiltration diminishes and may disappear, but the central part of the lobular reticulin framework often collapses. Accumulation of bile in surviving centrilobular hepatocytes and in bile canaliculi occurs; the cause of this severe cholestasis is not known and may be secondary to widespread loss or distortion of canalicular microvilli. An alternative possibility has been put forward by Myburgh et al. (111) who drew attention to the progressive hypertrophy and dilatation of smooth endoplasmic reticulum in the centrilobular hepatocytes and suggested that these intracellular changes might be caused by humoral antibody and result in disrupted cholesterol and bile salt metabolism with production of excess lithocholate. In some patients, the larger interlobular bile ducts disappear as modified rejection continues. This phenomenon is characterized by rapid and relentless rise in serum bilirubin (49–51, 116).
As rejection progresses, connecting bands of reticulin are often laid down between the central areas, subdividing the lobules. What triggers progression to hepatic fibrosis in some grafts is unknown; excess lithocholate has been implicated (111). In some patients, cirrhosis is produced. Another characteristic feature of chronic rejection is progressive thickening of the intima of the branches of the hepatic artery in the homograft. The intima contains fat-laden smooth muscle cells and macrophages, and the lumen is narrowed or occluded. These arterial changes occur in many long-surviving liver homografts. The accumulation of immunoglobulins and complement in the altered vessel wall has raised the possibility that this damage is a late manifestation of rejection brought about by circulating antibody. Deposition of immunoglobulins is less striking in hepatic grafts than in transplant kidneys (117), prompting speculation that hepatic rejection was more a phenomenon of cell-mediated immunity (as opposed to damage by circulating antibodies) than has been thought to be the case with rejecting kidney grafts.
Some histopathologic changes in liver homografts may be caused by hepatotoxic drugs, viral hepatitis, or other factors. However, a comparison of changes in human specimens with those in animal homografts emphasizes that the major alterations are immunologic in etiology.
Rejection as defined by us (32) and Williams and Calne (7) in patients treated with conventional immunosuppression also occurs during treatment with cyclosporin A and steroids. Many patients lose their appetites and become depressed. Fever, vague upper abdominal pain, and ascites are variable. By palpation, grafts are frequently swollen, hard, and mildly tender. Radioisotopes used for liver scanning are poorly concentrated, whether these depend upon parenchymal or reticuloendothelial function. Elevations may occur in serum bilirubin, alkaline phosphatase, and transaminases. Failure of synthetic function is most readily detected by measurements of prothrombin time.
The various manifestations of rejection occur in different combinations, to variable degrees, and at unpredictable times. The resulting patterns have been categorized as “anicteric,” “indolent,” and “crisis” (32). With a rejection crisis, jaundice can develop with astonishing rapidity, usually just after or accompanied by major rises in transaminases. The insidious indolent rejections have been the most difficult to reverse.
All patterns of graft deterioration are nonspecific. Proven alternative etiologies include ischemic injury, biliary obstruction, cholangitis, hepatitis (B virus, cytomegalovirus, adenovirus, herpes), and drug toxicity. Consequently, diagnostic procedures, often including cholangiography and needle biopsy, must be considered if the postoperative evolution is not satisfactory. In the interim, steroid dosage is temporarily increased and returned to baseline if a diagnosis other than rejection is established. Clinical management is particularly difficult and fraught with error if good initial graft function was not achieved.
A devastating complication of rejection, termed “septic hepatic gangrene,” occurred in patients who were treated with azathioprine, ALG, and relatively low doses of prednisone. After days or weeks of slightly abnormal hepatic function, these patients developed massive hepatic necrosis with extremely elevated serum transaminase activities and deterioration of other measures of hepatic function. Raging fevers, bacteremia, and disappearance on liver scans of large portions of hepatic parenchyma signaled regional infarctions within the transplants. Kinking of lobar or segmental hepatic arteries was originally postulated to be responsible for this complication (5), but the most important factor is poorly controlled rejection (32). With diminution in total hepatic blood flow, as was documented in canine experiments by Groth et al. (31), invasion of the ischemic homograft by microorganisms from the adjacent intestinal tract is not surprising.
Study of patients with septic hepatic gangrene and confirmatory observations by Brettschneider et al. (118) after canine and porcine liver transplantations clarified the interaction between rejection and homograft bacterial colonization. In animal studies, the normally low incidence of positive cultures from liver tissue of dogs and pigs increased after sham operations. When an hepatic ischemic injury was added by performance of simulated autotransplantation, all livers became contaminated, primarily with the same organisms concomitantly present in the upper intestine. Bacterial counts were somewhat lower if the common bile duct was left intact than when the duct was ligated and bile drainage restored by cholecystoenterostomy. The bacterial changes were more pronounced in liver homografts transplanted to unmodified or immunosuppressed animal recipients.
Presumably, any ischemic or necrotic area can become a septic focus, particularly if the host is given immunosuppressive therapy and cannot respond normally to invading microorganisms. The spectrum of resulting infections under conventional immunosuppression has been well studied (32, 119–121). Paradoxically, one of the most important ways to prevent nontreatable liver infection is to protect the graft with potent immunosuppression, especially during the early postoperative period. The second obvious step is to provide systematically designed antibiotic therapy intraoperatively and for several days thereafter.
Acute cellular rejection in grafts sampled by biopsy or at autopsy has been encountered many months or years after transplantation. Some patients were known to have discontinued their medications but others had been given unwise advice about lowering maintenance medications. Increased steroid therapy was given under such circumstances.
The diagnosis of chronic rejection was restricted to patients whose grafts had arterial intimal thickening, hepatic fibrosis, and other findings described previously. These findings are not necessarily time-related since they often develop within the first few months. Clinical manifestations of chronic rejection were not much different from those of chronic liver failure from endstage disease of differing etiologies. Treatment with increased immunosuppression was ineffective.
The introduction of cyclosporin A and steroids has had such a major influence upon results that patients have been divided into those treated before and after this event. By so doing, it is possible to reexamine several factors whose significance was previously unclear.
Between 1963 and the end of 1979, 170 patients had liver replacement. In previous publications, the identification of individual recipients by orthotopic transplantation (OT) numbers made it possible for the interested reader to follow their progress from report to report. The same code numbers will be used throughout this review.
In past reports (45, 80, 107), the first 170 recipients were divided into the three successive series summarized in Table 6 and Figure 8. In the first, second, and third of these consecutive groups, 1-year survival was 28.8, 50, and 34.5%.
Of the 56 recipients who survived the first postoperative year, 23 died at the times shown in Table 6. Although 13 of 23 late deaths were in the second postoperative year, deaths occurred as late as 6 years. Of the original 170 patients, 33 (19.4%) are alive after follow-ups of 2½ to 12½ years. Twenty have survived for more than 5 years and four are into the second postoperative decade.
There was an almost equal division in the total period of 1963 to 1979 between adult (≥19 years) and pediatric (≤ 18 years) recipients. From the sixth month onward, the younger patients had about a 10% survival advantage (Figure 9).
Table 7 summarizes the results in the Cambridge-King’s College trials from 1968 through early 1980 (53, 122). In this series, 22 (23.7%) of the first 93 recipients lived for at least 1 year, with 11 subsequent deaths during the second to sixth years; the 11 survivors had been followed for 1 to 6 years. The better 1-year survival in the American compared to the English trial (33 vs. 24%) was partly illusory since Calne and Williams accepted for surgery few pediatric patients with whom (Figure 9) our best results were obtained in those years.
The longest follow-ups for our patients treated with cyclosporin A and steroids are only 2⅓ years. During the first 9 months of 1980, 14 patients entered into this trial at the University of Colorado (108); 2 died during the operation, and 1 died after 19 days. The other 11 patients survived for more than 1 year. Eight are alive after 21 to 28 months. The 1-year survival of 78.6% could have represented a sampling accident. However, at the University Health Center of Pittsburgh, 26 patients were treated in 1981. Five died in the first postoperative month, and additional deaths occurred in the second, third, and fourth months. With follow-ups of 6 months to more than 1 year, the remaining 18 (69.2%) recipients are alive and at home; none has poor hepatic function.
The pattern of predominantly early mortality seen in 1980 and 1981 has continued into 1982. Of the first 27 recipients treated in 1982, 8 died. Six, one, and one of these deaths were in the first, second, and third postoperative months, respectively.
It is too early to assess the rate at which late deaths will occur, since only 18 patients treated with cyclosporin A have reached or passed the 1-year mark. Of these, three died in their thirteenth, sixteenth, and twentieth postoperative months for reasons that will be considered in the next section.
The actuarial survival calculated from the cases in the cyclosporin era studied to date is shown in Figure 10. In comparison to our previous experience in the precyclosporin era and in comparison to the Cambridge-King’s College compilation, survival has more than doubled. The results calculated for the first 12 postoperative months have not been different in adults and children (Figure 11). A breakdown of results according to original disease in the pediatric and adult cases is provided in Tables 10 and and1111.
The influence of cyclosporin A upon survival in the Cambridge-King’s College trials has not been clearly defined, because the drug has not been regularly used and because it was started late in most cases after an initial course of azathioprine and steroids. Nevertheless, improved results have been attributed by Calne et al. (53) to better immunosuppression.
The appalling early mortality after liver transplantation has prompted exhaustive clinical-pathologic analyses of our failed cases. Using the OT code numbers of the patients, the results have been reported in such a way that individual assessment of almost every early death in the first 170 cases can be made by the interested reader (43, 45, 80). Mortality figures included the use of grafts damaged by ischemia, massive operative hemorrhage, thrombosis of the reconstituted homograft blood supply, intraoperative cerebral air embolism (44), unsuspected recipient abnormalities (such as prior thrombosis of the portal vein), hopeless anatomical situations created by multiple previous operations, irreversible preexisting debilitation, and (above all) defective biliary tract reconstruction.
With or without such factors, overwhelming infection was frequently a terminal event. At autopsy, histopathologic findings of acute rejection were found in 10 to 15% of cases, prompting speculation that over immunosuppression, especially with prednisone, may have been responsible for unnecessary deaths (43).
When serial biopsies were obtained in later cases (45, 80), this simplistic view had to be revised. Many biopsies contained unmistakable findings of rejection for which the appropriate response had been more steroids. After death caused by infection, the findings of rejection were absent. The conclusion was reached that even after a perfect operation, the unacceptable acute mortality would remain until improved immunosuppression became available. Both nonimmunologic and immunologic complications have continued to cause early deaths in the cyclosporin era albeit at a reduced rate.
Assessment of the reasons for late death in older cases may help to predict the spectrum of problems which can be expected in future patients. The causes of mortality after 1 year in patients treated with conventional immunosuppression are listed in Table 12. Recurrent liver failure was responsible for death in ¾ of 23 patients, if the four who died after attempted retransplantation are included. In three patients, the main mortality factor was recurrence of malignancy. One of the late deaths was caused by chicken pox hepatitis during an epidemic on the transplantation ward. The patient (OT 112) whose death was classified under self-abuse was an alcoholic, drug abuser, and derelict who resumed the same life style after transplantation. Fifty-six months after transplantation, he was found unconscious in a ditch in Florida and died of pneumonitis.
The dominant pathologic diagnoses of the 24 first or second grafts which functioned chronically in these 23 patients are listed in Table 13. Chronic rejection was the most common final diagnosis, followed by biliary obstruction and recurrent cancer. There were two examples each of chronic hepatitis and portal vein thrombosis.
These findings differ from those reported by Calne et al. (53) in 11 patients who died after 1 year; recurrent carcinoma was the main homograft abnormality in five patients. In the other six grafts, there was biliary sludge and cholangitis. Chronic rejection was not mentioned. Our findings suggest that ongoing problems with immunologic control will continue to take a gradual toll long after successful transplantation, whereas interpretation of the pathologic findings in the English recipients is different. Clarification of this divergence of observations will be important.
Twenty-two of the 67 patients treated in 1980 to 1982 died. Three deaths were after 1 year, and the other 19 were early.
Fourteen of 19 early deaths, including two on the operating table, occurred in the first postoperative month. In the second, third, and fourth postoperative months, there were 1, 1, and 2 more death(s), respectively (Table 14).
Eight deaths were directly attributable to preexisting anatomic conditions including multiple previous operations (OT 178), earlier portacaval shunt (OT 180), and right-to-left pulmonary shunts secondary to the liver disease (systemic arterial pO2 was 30 mm Hg) which did not subsequently close (OT 203). However, the most important abnormalities were in liver blood supply or the vena cava (OT 217, 220, 228, 232, 233) which had not been diagnosed preoperatively. At operation, it was not possible to vascularize adequately homografts in 4 of these latter 5 recipients. In the fourth (OT 233), the superior vena cava was discovered at autopsy to have been replaced by two innominate veins which descended into the abdomen and emptied into the inferior vena cava below the renal veins. During the vena caval cross-clamping of the anhepatic phase of transplantation, the child developed an acute superior vena caval syndrome with irreversible brain injury.
Eight early deaths were technical and thus avoidable, including the use of inadequately preserved grafts (OT 185 to 188), hepatic artery thrombosis (OT 183, 225), and complications of biliary tract reconstruction (OT 201, 208). Problems in preservation were encountered in the first four cases at a new institution.
Patients died late (Table 14) of recurrent Budd-Chiari syndrome (OT 174), recurrent duct cell carcinoma (OT 176), and after retransplantation after a primary graft was chronically rejected (OT 181).
In assessing ways of reducing patient mortality, it was obvious almost from the beginning that aggressive attempts at retransplantation offered the only chance of survival for many patients whose first grafts failed either early or late. Such efforts, which have been made in 27 patients (2 of the 27 also were given third grafts) since 1968 (Figure 1), usually have born bitter fruit. The few successes that have been achieved have served as an important stimulus for further trials.
The attempts at retransplantation in 27 patients are summarized in Table 15. Eighteen of 27 second transplantations were within the first 3 months, 3 were between 3 and 12 months, and 6 were after 12½ to 29 months.
Extended subsequent survival occasionally was achieved with early and later retransplantation. The fate of six patients whose lives were significantly prolonged is summarized in Table 16. Much of life for the first four recipients was a nightmare of morbidity because of the combination of high steroid needs and slowly failing graft function. However, the two patients who had successful retransplantation in the cyclosporin A era, 1 and 3 weeks after primary grafting, have had perfect results and are at home 6 and 12 months later on daily prednisone doses of 5 and 15 mg per day, respectively. In the first patient, the primary graft had developed a huge fungus abscess; in the second patient, the first graft had been rejected.
The performance of retransplantation has sometimes been surprisingly easy. The procedure has been greatly simplified by retaining cuffs from the suprahepatic and infrahepatic vena cava and from the portal vein of the first graft. Usually, it has been necessary to perform the arterial anastomosis proximal to the previous site.
Evaluation of the influence on survival of preexisting hepatic disease is complicated by the fact that many patients have more than one diagnosis (Table 17). Ten (14.9%) of our last 67 patients had two coexisting hepatic diseases; however, in most of our analyses (Table 8 to to11),11), only the most important diagnosis was tabulated.
No disease for which transplantation has been used can be categorically excluded for further trials. This can be appreciated by examination of results with different kinds of disease in pediatric and adult patients before and after the introduction of cyclosporin A (Tables 8 to to1111).
Nevertheless, special problems can be expected in treating some hepatic diseases. In the past, some patients with biliary atresia died because of unexpected anomalies which jeopardized performance of a technically satisfactory transplantation (32); specific examples are listed in Table 14.
Noncompliance may be a problem in patients treated for Laennec’s cirrhosis. Of our first nine patients with alcoholic cirrhosis, eight died too soon to evaluate this potential problem (123). There were several subsequent successes (Table 9); only one recipient returned to toxic drinking.
The possibility that immunosuppression may accelerate metastatic tumor growth has been recognized (32). Evaluations of transplantation in treating hepatic malignancies were made in 1981 by Iwatsuki et al. (124) and by Calne (122). Although recurrent disease exceeded 50% in both series, the results did not allow definitive recommendations about continuation of these efforts.
Our case material has been divided into three groups. In the first category were three children whose livers contained malignant tumors (2 hepatocellular cancers, 1 hepatoblastoma) that had not been suspected preoperatively (Table 18). The two recipients who survived operation have no evidence of recurrence after 4¼ to 12½ years. These observations suggest that malignancies can be cured by liver replacement.
In the second category were eight patients, all treated early in our experience, who died less than 1 month after liver replacement for hepatic or duct cell cancer (Table 19). From this case collection, it was possible to determine by autopsy studies the frequency with which extra-hepatic tumor spread had been missed in preoperative evaluation. Only 1 of the 8 recipients had metastases.
Twenty-two additional patients, including one (OT 176) whose neoplastic lesion was missed at the initial pathologic examination, lived long enough to evaluate the influence of transplantation upon the malignancy (Table 20). The first 12 were treated in the precyclosporin era; nine recipients developed metastases. A tenth patient with an unclassified sarcoma had extrahepatic metastases at transplantation, and she is well 5½ years later with no clinical evidence of advancing disease. Five patients survived for longer than 1 year; even for those who eventually died of metastases, the extension of useful life seemed to be worthwhile.
Ten more patients have been treated in the cyclosporin era (Table 20). Three had duct cell carcinomas (one with sclerosing cholangitis); one of whom (OT 185) died early without evidence of residual cancer. The second patient (OT 176) died of metastatic cholangiocarcinoma after 1 year. The third (OT 200) is alive in the ninth postoperative month and has metastases. All patients with hepatocellular carcinoma are alive. The tumors were enormous in three cases. In the other four, the lesions were smaller but could not be resected with conventional techniques because of coexisting cirrhosis. It seems likely that selected patients with hepatic or possibly biliary duct malignancies can be effectively treated with transplantation; however, no patient with duct cell carcinoma has ever been cured (53, 122, 124). The prospects may be more favorable for young patients whose hepatocellular cancers could be treated with conventional partial hepatectomy were it not for coexisting cirrhosis.
Heroic efforts may be justifiable for patients with the recently described “fibrolamellar hepatoma” which is characterized by indolent primary growth and late metastases (125, 126). Three of our last 10 patients with hepatic malignancy have had this diagnosis. In all three, the tumors were massive. In one patient who was treated more than 2 years ago (OT 172), a large tumor thrombus originating in a hepatic vein was extracted at operation from the vena cava and right atrium. He is tumor-free. A second patient, who also is tumor-free after almost 1 year (OT 194), had complete obstruction of the portal vein by tumor. A third patient treated 2 months ago (OT 231) developed recurrence in the residual lateral segment after a right trisegmentectomy 4½ years earlier. The tumor-laden residual segment was replaced with a new liver with a satisfactory result thus far.
At the moment, the prospects for cure seem bleak for patients with duct cell carcinomas, and scarcely better for those with unresectable conventional hepatocellular carcinomas. Patients with smaller malignancies in livers with other diseases, or those with fibrolamellar hepatomas may be more susceptible to treatment.
In pediatric recipients (Tables 8 and and10),10), recurrence of nonneoplastic hepatic disease has not been observed. A special feature of transplantation in the younger age group has been the metabolic “cure” of at least five and possibly six so-called inborn errors (Tables 8 and and10)10) (126–131). With cyclosporin A, the prospects of using transplantation to treat children with a variety of diseases has been heightened because chronic high-dose steroid therapy can be avoided (132).
That the original disease can be recapitulated in homografts was demonstrated in adult recipients. Two patients with Australia antigenemia and chronic aggressive hepatitis redeveloped their original disease and died (133). Other patients with recurrent or newly developing Australian antigenemia have lived for as long as 8 years with the carrier state. It has been our policy to treat HBsAG positive transplant recipients with hyperimmune globulin postoperatively. Antigenemia has returned in every case, sometimes after becoming undetectable for months. With such treatment, Johnson et al. reported permanent antigen clearing in a patient (134).
Recurrent primary biliary cirrhosis (135) as described in the English series was not seen in five grafts studied at autopsy after a few days to more than 6 months, and five patients still living have had no evidence of recurrence in spite of the reappearance of antimiochondrial antibodies in the longest survivors (2¼ and almost 4 years).
We have treated three patients for the Budd-Chiari syndrome. One (OT 174) developed the same disease in the graft and died after 15 months. The terminal course of this patient was triggered by unwise discontinuance of anticoagulant therapy in preparation for a closed liver biopsy. Calne et al. (53) reported a similar occurrence.
The technical problems engendered by prior surgery have been so great that the Cambridge-King’s College Team consider multiple earlier operations as a relative contraindication to transplantation. However, the majority of candidates evaluated by us and the English workers have had previous operations. The influence of this factor was evaluated in the first 40 patients treated with cyclosporin A and prednisone (Table 21) in whom the results were more analyzable than in our earlier cases.
Fourteen of 40 recipients had major procedures on portal triad structures (Table 21) including six portal-systemic shunts and five biliary duct reconstructions. The risk of death in the first postoperative month was double that in patients with lesser or no earlier operations. There were two operative deaths (OT 176 and 180) (Table 21) including one that occurred while trying to take down a portacaval shunt. The portal vein after portacaval shunt (especially side to side) may have suboptimal length and can be so sclerotic that suturing is difficult or impossible. For patients who survive the first postoperative month, the background of previous major operations was no longer a factor (Table 21). These results suggest that a prudent decision may be against transplantation in patients with a complex surgical history, but that a “clean abdomen” is not a criterion of selection.
The alternative to orthotopic liver transplantation is transplantation of an extra liver (auxiliary transplantation) without removal of the diseased native organ. Clinical trials have been discouraging, as summarized by Fortner et al. (136) from the compiled world experience. Of nearly 50 well-documented auxiliary transplantations, only one was an unequivocal success. Subsequently a report from Paris described a second success (137).
Our opinion has been that auxiliary liver transplantation should be restricted to patients with potentially reversible liver disease. In such a situation, the extra liver could be construed as a temporary support organ which can be removed later. However, we have encountered increasing numbers of patients whose portal vein has clotted in the hepatic hilum, making it technically impossible to consider liver replacement. Other candidates are those with extensive previous surgery in the right upper quadrant. Such patients can theoretically be helped by an auxiliary liver transplantation, particularly when the superior mesenteric vein or other distal tributories to the main portal circulation are still open. The optimal conditions for vascularization of an auxiliary liver graft require input from the portal circulation (2, 32, 138, 139), largely because of its high concentrations of endogenous hormones.
Through 1980 in the United States, almost all liver transplantations were performed in the Clinical Research Centers (CRC) supported by the National Institutes of Health. The fraction of the per case cost born by this government agency shrank from year to year because of the increasing willingness of many third party insurance carriers to pay for part or all of the service. Third party payments were collected by the institution and remanded back to the National CRC headquarters which included such collections as part of the total grant funding. In the last years of the program at the University of Colorado, approximately 85% of CRC expenditures for liver transplantation were paid for in this way. By having CRC support for the other 15%, it was not necessary to screen candidates for their ability to pay. This creative practice of federal and private cost sharing was conceived and made practical by Dr. William DeCaesare, Director of the Clinical Research Center Division, Bethesda, Md. Thanks to the administrative leadership that flowed from DeCaesare’s office for more than 2 decades, almost all modern-day techniques of immunosuppression and transplantation of all organs, beginning with the kidney, were developed on CRC units.
However, it is not reasonable to look indefinitely to the CRC for support. So far, none of the liver recipients at the University Health Center of Pittsburgh has been treated on the pediatric or adult CRC. The principal financing for their care has been from health insurance programs or, less commonly, from private sources including community fund raising. The collection rate for bills has been greater than in the Colorado experience. With the exception of Blue Cross/Blue Shield and some health maintenance organization programs, the major health insurance companies have willingly accepted their responsibility for liver transplant candidates providing they were forewarned. The patterns of support from Blue Cross/Blue Shield have been irregular, probably because of their close association with MediCare.
It is ironic that government decisions or opinions have impeded the movement of liver transplantation to the private sector of medicine. A pronouncement by a state or federal official of MediCare to the effect that liver transplantation is “experimental” and not fundable has often been the basis for a similar decision by regional officials of Blue Cross and/or Blue Shield or by a cost-conscious health maintenance organization. Even so, a rapidly growing number of state (or Blue Cross) agencies (including those in Pennsylvania, New York, and New Jersey) have classified liver transplantation as a service.
In spite of the advantage of a preexisting federally funded organ procurement network in the United States as part of the End Stage Renal Disease Program, the financing of liver procurement is not on a solid base. There are no formal guidelines about how to proportion the extra costs of removing extrarenal organs from a multiple organ donor, or how to ensure against the potential malpractice and other liability that could be incurred. Etna Life Insurance Company, the carrier for the National Kidney Procurement Program, recently drew attention to these policy gaps in a document that had a chilling effect on some transplant coordinators.
The approximate average cost of a liver transplantation in Pittsburgh has been $55,000 (range $23,000 to $150,000). The procedure offers hope of genuine rehabilitation. Dying of endstage liver disease with no hope of real recovery may be even more expensive. O’Donnell et al. (140) reported from Boston that the average cost of nonsurgical treatment of patient for variceal hemorrhage was $35,000. The use of any operative procedure increased the total to $53,000. In many of our patients, the expenses incurred during repeated hospitalizations before transplantation dwarfed those incurred by transplantation itself.
Five to 10 years from now, we believe that every major center for the treatment of liver disease will have either transplantation capabilities or direct access to this kind of service. The surgical techniques are within the grasp of many practicing surgeons. The frequency with which liver transplantation can be used will be great although this has not been properly assessed. Our estimate is that at least 20 centers will be required in the United States.
Knowledge that the provision of new liver tissue is a realistic objective at the end of the line will influence decisions about treatment. It will be increasingly important to avoid major and often futile surgical operations that jeopardize ultimate candidacy for transplantation. Fortunately, there are alternative approaches. Sclerosing therapy for the control of variceal hemorrhage instead of portal diversion has become increasingly accepted. “Interventional radiologists” often have been able to ameliorate duct strictures in sclerosing cholangitis and other diseases as effectively as can surgeons at open operation. When procedures such as porticoenterostomy (Kasai) are performed in infants with biliary atresia, it will be worthwhile to avoid deviations from the standard Roux-Y technique and multiple reoperations which make transplantation difficult or impossible.
The presence of regional units undoubtedly will move the timing of transplantation forward in the course of the disease. The fact that there has been a very high preoperative mortality of patients accepted as candidates for new livers is an indication of the lateness of referrals. Throughout the years, many others have reached the operating room in such appalling condition that there was little hope of survival.
An avalanche of new scientific information should become available to hepatologists and surgeons as the result of progress in transplantation. Much has already been learned about the synthesis of proteins whose origin was not previously clear (2, 9, 141–143). Further improvements in surgical techniques and immunosuppression will increase the harvest. The history of medicine is that what was inconceivable yesterday and barely achievable today often becomes routine tomorrow.
This study was supported by grants from the National Institutes of Health (AM-17260, AM-07772, AM-29961, and AM-30183) and from the General Clinical Research Centers Program of the Division of Research Resources, National Institutes of Health (RR-00051, RR-00069, and RR-0084).