The standard technique of orthotopic liver transplantation (OLT* (1) (2-4)
Calne in 1968 and Tzakis in 1989 described the "piggy-back" (PB) technique with preservation of the recipient vena cava (PIVC) and direct anastomosis of the donor's vena cava with the recipient's hepatic veins (5, 6) (7) (8)
The basic concept of this technique was absence of total IVC occlusion, precluding the use of a VVS. Many advantages were expected for this technique: (1) reduction in complications specific to VVS (thromboembolic complications, hypothermia, hemolysis) (2, 3, 9-11) (12, 13) (10, 14-16)
The objectives of this study were: to describe the complications specifically related to OLT with preservation of the inferior vena cava (PIVC) and to their therapeutic management; to assess the impact of the different technical modalities used [type of caval anastomosis, temporary portacaval anastomosis (PCA)]; and, finally, to define the clinical and anatomical conditions limiting this PIVC technique.
PATIENTS AND METHODS
A retrospective, multicentric study was carried out based on a questionnaire sent to 17 French liver transplant centers. This questionnaire recorded the number of OLTs with PIVC performed in adults between 1991 and 1997, the technical modalities used such as the type of cavocaval anastomosis performed, PB (Fig. 1A) , end-to-side (Fig. 1B) , side-to-side (Fig. 1C) , and the use of a temporary PCA, as well as technique-related complications and mortality. Descriptive analysis of peri- and postoperative complications included indications for OLT, etiological factors, and therapeutic management. A comparative analysis of the type of caval reconstruction (PB versus side-to-side) was also completed. The χ2 test was used for statistical analysis, and the level of significance was set at P <0.05.
Figure 1: A, PB technique in standard OLT. PVCI and direct anastomosis of the donor's vena cava with the recipient's hepatic veins. B, End-to-side caval anastomosis. Anastomosis between the upper end of the donor graft's inferior vena cava and the recipient's inferior vena cava. C, Side-to-side caval anastomosis. The two ends of the donor graft's inferior vena cava are closed, and the anastomosis is performed after cavotomy of the donor's inferior vena cava and the recipient's inferior vena cava which is a side-to-side anastomosis.
RESULTS
Seventeen OLT centers replied to our research survey, allowing us to take a census of 1681 OLTs, of which 1361 (81%) were performed with PIVC.
Technical Modalities
Type of cavocaval anastomosis was side-to-side anastomosis in 50.6% of cases (n=689), PB in 42.7% (n=582), and end-to-side in 6.6% (n=90). In total, 882 temporary PCA were performed. Seven centers performed this anastomosis systematically (718 PCA; 53% of sample population OLTs), whereas for the other 10 centers PCA was performed in cases of hemodynamic intolerance (164 PCA for 643 PIVC OLT). Thus, PCA for hemodynamic intolerance at clamping was necessary in 25% of patients.
Morbidity and mortality. Fifty-five patients presented with one or more complications ascribable to the PIVC technique; i.e., overall morbidity was 4.1% (55/1361). Overall mortality was 0.7% (10/1361). Mortality rate for patients presenting with surgical complication was 18%. In total, 64 complications were noted, of which 57 (89%) were in the perioperative or immediate postoperative period (before day 1) and 7 (11%) were postoperative (day 1 to day 10). The complications were as follows: hemorrhage, n=39 (61%); occlusive venous return or Budd-Chiari syndrome (BC), n=21 (33%); hemodynamic complications, n=4 (6%).
Hemorrhagic Complications
Thirty-nine hemorrhagic complications related to PIVC were noted for 39 patients. Thirty-six were perioperative and 3 were postoperative. The incidence of these hemorrhages was 3% (39/1361). Etiological factors involved in the occurrence of these complications are presented in Table 1 . Hemostasis could be performed for 34 of 39 patients (87%). Sixteen total vascular clampings and 7 latero-caval clampings were necessary. For two patients, a further side-to-side caval anastomosis was performed after hemostasis on the original PB anastomosis. Hemorrhage caused gaseous embolism in one patient. One patient had to be retransplanted for graft ischemia after total vascular clamping for reconstruction of the cavocaval running suture.
Table 1: Etiological factors involved in the occurrence of 39 hemorrhagic complications
Mortality for these hemorrhagic complications was 13% (5/39); i.e., five patients died in a context of perioperative bleeding. For these patients, the cause of hemorrhaging was injury to the hepatic veins, n=2; injury to an IVC included in segment I, n=1; IVC agenesis, n=1; hemorrhage through portal hypertension (major epiploic vascularization), n=1.
Occlusive Venous Return and BC
The incidence for these venous complications was 1.5% (21/1361). Seventeen of the 21 patients (81%) were in the perioperative or immediate postoperative period (before day 1) and four (19%) were postoperative.
Indications for OLT in these patients are presented in Table 2 . The distribution of these vena cava complications is presented in Table 3 . Inadequate graft size is involved in the occurrence of these complications and was mentioned in six patients (29%).
Table 2: Indications for OLT for patients with vena cava complications; occlusive venous return and Budd-Chiari syndrome
Table 3: Distribution of vena cava complications
Therapeutic management of these patients with occlusive venous return and BC was among the following. (1) Therapeutic desistance, or a simple rotation of the graft, n=4 (20%). (2) Caval anastomosis reconstruction, n=5 (24%); this was with a thrombectomy for three patients presenting with caval thrombosis. (3) Placing a Blackmore catheter, n=3 (15%); this was with an anastomotic reconstruction for one patient presenting with positional BC, and was preceded by retransplantation in one other patient. (4) Omentoplasty, n=2 (10%). (5) Diaphragmatic placement of the graft in two patients (10%) with positional BC. (6) A second caval anastomosis was completed in two cases (5%) for torsion of the original caval anastomosis. (7) Endoluminal anastomotic dilation for the two patients (10%) of anastomotic stenosis. (9) Conversion to standard OLT, n=1 (5%). (10) Retransplantation, n=1 (5%)
Five patients (24%) died of these venous complications. Cause of death was caval thrombosis before retransplantation (n=1), intraoperative fulminant hepatitis (n=1), and postoperatively sepsis (n=3).
Hemodynamic Disturbances
Two patients with intolerance to lateral clamping, one of which required a cardiopulmonary bypass, and two patients with hemodynamic shock related to declamping were reported. These complications caused no deaths.
Anatomical Limits of OLT with PIVC
Inadequate graft size. Eleven patients (11%) with inadequate graft size were reported, causing six (55%) complications and two (18%) modifications of OLT technique. There were seven patients with a voluminous graft. Over and above the technical difficulties in carrying out a caval anastomosis requiring modification of standard OLT in two patients, a voluminous graft was involved in one patient with hemorrhage through release of the running suture, and in the occurrence of two patients with graft congestion. For the remaining four patients with inadequate graft size, small grafts were involved in the occurrence of patients with positional or torsion-related BC, and one caval thrombosis patient related to over-rapid postoperative regeneration of the graft.
Anatomic abnormalities of the retrohepatic venae cava. There was one patient with IVC agenesis, involved in one death from intraoperative bleeding. There was one patient with fibrous stenosis of the IVC, requiring standard OLT.
Comparative Analysis of Caval Reconstruction Techniques: PB Versus Side-to-Side.
(Table 4) Hemorrhagic complications through release of the cavocaval running suture or BC were significantly more frequent after PB anastomosis than after side-to-side anastomosis.
Table 4: Comparative analysis of caval reconstruction techniques: piggyback (PB) versus side-to-side
DISCUSSION
To date, specific complications of PIVC OLT are poorly reported in the literature. This large series of PIVC OLTs allows us to identify two types of complications: hemorrhage and occlusion of the caval venous return or BC.
These complications are rare, severe, and early. They concerned 4% of patients and caused 0.7% mortality, although mortality at 3 months is estimated at 16.8% (of 21,073 patients) in the European Liver Transplant Registry's June 1997 report. Specific mortality for these complications was 18% in our series. These complications were all observed within 10 postoperative days and in 89% of cases they occurred in the perioperative period (before day 1).
Hemorrhagic complications were most frequent, with an incidence of 3%. In standard OLT, this type of complication has been correlated to retrocaval dissection (16)
In this analysis, hemorrhage through release of the cavocaval running suture was significantly more frequent with PB anastomosis than with side-to-side anastomosis. Do these results show that carrying out a caval anastomosis in PB, with a limited amount of tissue for this anastomosis, is more difficult?
The second complication that appeared specific to this PIVC technique was caval return occlusion (1.5% of cases). This type of complication has been described after standard OLT (17) (18)
In this study, the patients with BC described were not associated with underlying thromboembolic disease and appeared to correlate with the PIVC surgical technique. Two factors seemed to be correlated to the occurrence of BC after PIVC OLT: inadequate graft size and type of caval anastomosis. Voluminous grafts can cause anastomotic occlusion through compression. Positional BCs are related to small grafts sliding into the hypochondrium, causing twisting of PB anastomoses. Furthermore, there is significantly higher incidence of caval return occlusion after PB anastomosis than after side-to-side anastomosis. Using the common trunk of the middle and left hepatic veins can limit the diameter of the anastomosis, compared with the use of cavotomy which can be wide in side-to-side anastomosis. To limit this type of complication, some authors have described enlargement of the recipient's hepatic veins for the PB anastomosis, either through horizontal incision near the right hepatic vein, or longitudinal cavotomy perpendicular to the axis of the common trunk of the hepatic veins (15, 19)
Excluding simple graft congestion (25% of patients) requiring no special additional treatment other than a final maneuvering of the graft, and anastomotic stenosis (9.5% of patients) with favorable outcome after percutaneous endoluminal dilation, these patients of BC were characterized by their severity, with 28% mortality. Therapeutic management of the 21 patients of BC was predominantly surgical, but not well codified. Treatment was more or less invasive, going from omentoplasty, diaphragmatic fixation of the graft or placement of a blackmore catheter, to anastomosis repair or conversion to standard OLT. In this study, side-to-side cavocaval anastomosis was performed in addition to the existing PB anastomosis of BC; through thrombosis and anastomotic twisting in two patients, and after hemostasis on the caval anastomosis to prevent an eventual anastomotic stenosis in two patients.
Other than the type of cavocaval anastomosis carried out, the other variable technical modality was whether a PCA was performed systematically or not.
The PCA was first described by Tsakis in children for whom a VVS could not be performed (20) (12, 20) (11)
The benefits of systematic PCA have not yet been demonstrated. For 643 PIVC OLT (47%) in this study, PCA was not the first technique considered. For this patient group, hemodynamic intolerance to laterocaval clamping was noted for 25% of cases, requiring construction of a second PCA, although only 2 of 718 patients that had had a systematic PCA presented with hemodynamic disturbances upon lateral clamping, one of whom required the use of a VVS. This appears to demonstrate the hemodynamic benefit of a PCA, although the data collected in this study are not sufficient to be able to evaluate the potential benefit of a PCA for perioperative hemorrhaging. It is not possible, based on this study alone, to determine whether a PCA should be performed systematically, or secondarily in cases of hemodynamic intolerance to laterocaval clamping.
PIVC OLT seems to be the technique of choice in France, concerning 80% of OLTs in this study. Sampling a large number of PIVC OLTs allowed us to analyze surgical complications ascribed to this technique, and to define factors that limit it.
Specific complications of PIVC OLT are rare, but serious, and our results suggest that some of them can be prevented: (1) side-to-side cavocaval anastomosis should be preferred to PB anastomosis for inadequately small donor grafts; (2) standard OLT should be preferred in cases of voluminous donor graft, or where unfavorable recipient anatomic factors exist (IVC included in segment I or anatomic abnormalities of the IVC). However, the decision not to preserve the donor's IVC can be taken perioperatively, which, for surgeons, entails a mastery of both OLT techniques.
Concerning technical modalities used in PIVC OLTs, our retrospective results define significant advantages in favor of side-to-side anastomosis in terms of vascular complications (fewer BCs, fewer releases of the cavocaval running suture).
Acknowledgments. The authors thank Dr. B. Suc, Dr. J. Carles, Dr. L. Chiche, Pr. G. Mantion, Pr. B. Descottes, Pr. J. Baulieux, Pr. C. Ducerf, Pr. P. Boissel, Pr. J. Gugenheim, Pr. P. Le Treut, Dr. B. Dousset, Pr. F. Pruvot, Pr. B. Launnois, for their participation in this study. The authors also thank Mr. Glen McCulley for preparation of the manuscript.
REFERENCES
1. Starzl TE, Groth CG, Brettschneider L, et al. Orthotopic homotransplantation of the human liver. Am Surg 1968; 168: 392.
2. Paulsen W, Whitten C, Moimsay M, Lintmalm G. Considerations for anesthetic management during veno-venous bypass in adult hepatic transplantation. Anest Analg 1989; 68: 489.
3. Paulsen N, Valek TR, Blessing W, et al. Hemodynamics during liver transplantation with veno-venous bypass. Transpl Proc 1987; 21(1): 2417.
4. Shaw BW, Martin DJ, Marquez JW, et al. Venous bypass in clinical liver transplantation. Ann Surg 1984; 200: 524.
5. Calne RY, Williams R. Liver transplantation in man: observations on technique and organization in 5 cases. BMJ 1968; 4: 535.
6. Tzakis A, Todo S, Starzl TE. Orthotopic liver transplantation with preservation of the inferior vena cava. Ann Surg 1989; 210: 649.
7. Belghiti J, Panis Y, Sauvanet A, Gayet B, Fekete F. A new technique of side to side caval anastomosis during orthotopic hepatic transplantation without inferior vena caval clamping. Surg Gynecol Obst 1992; 175: 270.
8. Belghiti J, Noun R, Sauvanet A. Temporary porto caval anastomosis with preservation of caval flow during orthotopic liver transplantation. Am J Surg 1995; 169: 277.
9. Kloury G, Martin MD, Mann M, Porot M, Abdul-Rasool I, Busitil R. Air embolism associated with veno-venous bypass during orthotopic liver transplantation. Anesthiol 1987; 67(5): 848.
10. Navalgund A, Kang Y, Sarner J, Jahr J, Gieraerts R. Massive pulmonary thromboembolism during liver transplantation. Anesth Analg 1988; 67: 400.
11. Ellis J, Lichtor J, Feinstren S, et al. Right heart disfunction, pulmonary embolism and paradoxical embolization during liver transplantation. Anesth Analg 1989; 68: 777.
12. Cherqui D, Lauzet J-Y, Rotman N, et al. Orthotopic liver transplantation with preservation of the caval and portal flows: technique and results in 62 cases. Transplantation 1994; 58(7): 793.
13. Jovine E, Mazziotti A, Grazi GL, et al. Piggyback versus conventional technique in liver transplantation: report of a randomised trial. Transpl Int 1997; 10: 109.
14. Salizzoni M, Andorno E, Bossuto E, et al. Piggyback, techniques versus classical technique in orthotopic liver transplantation: a review of 75 cases. Transpl Proc 1994; 26: 3552.
15. Lerut, JP, Molle G, Donataccio M, et al. Cavocaval liver transplantation without venovenous bypass and without temporary portocaval shunting: the ideal technique for adult liver grafting? Transpl Int 1997; 10: 171.
16. Stieber AC. One surgeon's experience with the piggyback versus the standard technique in orthotopic liver transplantation: is one better than the other? Hepato-Gastroenterology 1995; 42: 403.
17. Lerut J, Tzakis AG, Bron K, et al. Complications of venous construction in human liver transplantation. Ann Surg 1987; 205: 404.
18. Hesse UJ, Defreyne L, Pattyn P, Kerremans I, Berrevoet F, de Hemptinne B. Hepato-venous outflow complications following orthotopic liver transplantation with various techniques for hepato-venous reconstruction in adults and children. Transpl Int 1996; 9: 182.
19. Bismuth H, Castaing D, Sherlock DJ. Liver transplantation by "face-à-face" venacavaplasty. Surgery 1992; 111: 151.
20. Tzakis AG, Reyes J, Nour B, Marino IR, Todo S, Starzl TE. Temporary end-to-side portocaval shunt in orthotopic hepatic transplantation in humans. Surg Gynecol Obstet 1993; 176: 180.
* Abbreviations: OLT, orthotopic liver transplantation BC, Budd-Chiari syndrome; IVC, inferior vena cava; PB, piggyback; PCA, portacaval anastomosis; PIVC, preservation of the inferior vena cava; VVS, veno-venous shunt.
