Incidence of Renal and Liver Rejection and Patient Survival Rate Following Combined Liver and Kidney Transplantation


  • Caroline Creput,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a Antoine Durrbach,

    Corresponding author
    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a,d , Didier Samuel,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • b Pascal Eschwege,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a Mounia Amor,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a Faycal Kriaa,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a Henri Kreis,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • c Gérard Benoit,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • a Henri Bismuth,

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author
  • and b Bernard Charpentier a,d

    1. a Services d'Urologie et de Néphrologie, Le Kremlin- Bicêtre, Paris, FrancebCentre Hépato-Biliaire, Hôpital Paul Brousse, Villejuif, Paris, France cService de Transplantation, Hôpital Necker, Paris, France dINSERM U542 Villejuif, France
    Search for more papers by this author

Corresponding author: Antoine Durrbach,


Multiple organ transplantations are used to treat chronic multiple organ failure. However, long-term mortality and graft tolerance remain to be evaluated. We carried out a retrospective and comparative analysis of 45 patients who underwent a combined liver and kidney (LK) transplantation (LKT) from the same donor. They were compared to 86 matched patients who underwent kidney (K) transplantation (KT). All patients had an organic renal failure associated with cirrhosis (n = 35) or with inherited disease (n = 10). Nineteen (42.9%) had been transplanted previously. The patients' survival rate was 85% at 1 year and 82% at 3 years. Seven patients died within the first 3 months, due to severe polymicrobial infection. Two patients in the LK population (4.2%) developed acute rejection of the kidney graft compared to 24 of the 86 matched renal transplanted patients (32.6%). In parallel, acute liver rejection was observed in 14 cases (31.1%) in the LK population. The occurrence of acute rejection was not associated with panel-reactive lymphocytotoxic antibodies (n = 16), nor with positive cross-matches (n = 3). Four of the 45 patients (8.8%) subsequently developed chronic renal allograft rejection, and 16 cases of chronic hepatic dysfunction were noted (42.2%). In conclusion, the overall survival rate following combined liver kidney transplantation is acceptable, and LKT can be proposed to patients with kidney failure associated with liver dysfunction, primary oxaluria or amyloid neuropathy. The main cause of mortality in this population was severe infectious complications. The frequency of acute kidney rejection was lower than in single transplantation.


The number of cases of chronic renal insufficiency associated with chronic liver failure has increased during the last decade. This is mainly due to glomerulonephritis associated with chronic alcoholic hepatitis and/or viral hepatitis (1). In many cases, both the renal and kidney diseases progress in parallel, although an increasing number of patients with a renal or liver failure transplantation develop secondary hepatic or renal failure, respectively. Secondary hepatic failure is often caused by chronic viral hepatitis, whereas secondary renal failure is often due to toxicity of anticalcineurin treatment. In patients with both kidney and liver failure, dialysis is associated with a high frequency of mortality. Combined liver and kidney transplantation (LKT) remains the only solution (1,2), although this was frequently considered to be contraindicated due to the high mortality rate during the postoperative period. The causes of death and predisposing factors remain to be determined. However, recent studies have reported encouraging results for LKT, even in cases of chronic viral hepatitis (3–8), suggesting that LKT can be proposed to patients with hepatic and renal failure.

Recent findings on inherited diseases, such as primary hyperoxaluria and familial amyloid polyneuropathy, have opened up new possibilities for LKT (9,10). It has been shown that the production of mutated proteins (alanine glyoxylate transferase and transthyretin) by the liver leads directly or indirectly to end-stage renal disease. Although major advances have been made in the management of these diseases, their prognosis remains extremely poor and liver transplantation is the only therapy that allows the replacement of the mutated protein by the correct one, thus limiting the progression of the disease (1,11). Following renal failure, attempts have been made to associate liver and kidney transplantations (1,9–13).

Based on the study of small groups of patients who have undergone LKT, some authors have suggested that LKT is associated with a reduced incidence of acute renal rejection compared to renal transplantation alone (5,14). In contrast, Katznelson & Cecka compared the frequency of acute renal rejection following LKT with that in patients transplanted with the control lateral kidney. They found a similar elevated frequency of acute graft rejection in both groups (15). However, they did not take the immunological status of the recipient into account, even though this dramatically affects the outcome of the graft. An increasing number of patients who have previously been transplanted or who harbor panel-reactive anti-HLA antibodies are eligible for LKT. This population is at higher risk of acute graft rejection, which might affect the survival of the graft. For these reasons, we decided to compare a cohort of 45 patients who underwent combined liver and kidney transplantations between 1986 and 1999 to 86 matched patients with a single kidney transplantation (KT). The goal of this study was to evaluate whether the frequency of renal acute rejection is reduced during LKT and to compare it to that occurring during KT. The second objective was to identify factors of death in the LK population.

Patients and Methods

Combined liver and kidney transplantation population

Forty-five LKT were performed between 1986 and 1999 to treat an organic nephropathy associated with cirrhosis (n = 35) (patients with hepatorenal syndrome were excluded) or associated with an inherited disorder (n = 10): two had a familial amyloid neuropathy due to a valine-to-methionine mutation at position 30 of transthyretin and eight had primary hyperoxaluria confirmed by the reduced alanine glyoxylate transferase activity in the liver (Table 1). For the chronic liver disease, patients were graded according to the Child-Pugh classification (A: n = 21, B: n = 9 and C: n = 5).

Table 1.  Characteristics of combined transplantation population (LKT). CS (corticosteroid), CyA (cyclosporin A), AZA (azathioprine), no (no immunosuppressive treatment before the transplantation), HCV (hepatitis C virus), HBV (hepatitis B virus), HDV (hepatitis D virus).
NoSexAgeKidney diseaseDialysisCockcroft clearanceLiver diseasePrevious transplantationImmunosuppressive treatment before the second graft
1M33Interstitial nephritisYes HCVKidneyCS
2M35GlomerulonephritisYes HCVKidneyCS/CyA
4M43UnknownYes HCV  
6M40GlomerulonephritisNo HBV/HDV  
7M62Interstitial nephritisNo28.39HBV/HCV  
8F38GlomerulonephritisYes HBV/HDVKidneyno
10F43Primary hyperoxaluriaYes Primary hyperoxaluriaKidneyno
11M24UnknownYes Unknown  
14F37Primary hyperoxaluriaYes Primary hyperoxaluria  
15F57Primary hyperoxaluriaYes Primary hyperoxaluria  
16M39GlomerulonephritisYes Alcohol  
17F21UnknownNo24.75Auto-immune disease  
18M51Primary hyperoxaluriaYes Primary hyperoxaluria  
20M53UnknownYes Unknown  
21F50UnknownYes Unknown  
22M57Chronic PyelonephritisYes HCV/ alcohol  
23M27UnknownYes HBV/HCV/HDV  
24F30GlomerulonephritisYes Budd-Chiari syndromeKidneyno
25M49GlomerulonephritisYes HBV/HCVKidneyno
26F44GlomerulonephritisYes HBV/HCVKidneyno
27M28GlomerulonephritisYes HBV/HCVKidneyno
28M57NephroangiosclerosisYes Alcohol  
29M26Primary hyperoxaluriaYes primary hyperoxaluria  
30M49GlomerulonephritisYes HBVKidneyno
31F51UnknownYes HBV/HCVKidneyCS
32M53GlomerulonephritisYes HBVKidneyCS
33F33Primary hyperoxaluriaNo10.11Primary hyperoxaluria  
35F53Interstitial nephritisYes HBVKidneyCS
36F52Chronic PyelonephritisYes HCVKidneyCS
37M49GlomerulonephritisYes HBV/HCVKidneyno
38M56UnknownYes HBVKidneyCS
39M24UnknownYes HBVKidneyno
40M40GlomerulonephritisYes HCVKidneyno
42M44Primary hyperoxaluriaYes Primary hyperoxaluria  
43M40Primary hyperoxaluriaYes Primary hyperoxaluria  
44F49GlomerulonephritisYes Alcohol  

Thirty-two patients were treated by hemodialysis. The others had a mean creatinine clearance of 29 ± 11.3 mL/min (Cockroft clearance) (Table 1).

Nineteen patients had previously been transplanted (18 kidneys and one liver). In these cases, the immunosuppression therapy was continued for 47.4% of the patients, and 52.6% of them received no immunosuppression before the second transplantation (Table 1).

The combined transplantation was done from a single donor matched according to the ABO group as previously described (16). Patients were not matched for HLA. Cross-matches were retrospectively analyzed.

Sequential immunosuppression was given using antithymocyte globulins (ATG) for 14 days and high doses of methyl prednisolone, followed by prednisolone (decreasing doses) and azathioprine (maximum 2 mg/kg/day). Three patients received mycophenolate mofetil (2 g daily) instead of azathioprine. A calcineurin inhibitor [cyclosporin (n = 31) or FK 506 (n = 12)] was given as soon as the renal function tended to return to normal value or 12 days post transplantation. Anti-calcineurin doses were adapted to their blood concentrations (ELISA).

Patients with persistent renal failure during the postoperative period (primary dysfunction) were treated with hemodialysis. Patients with primary hyperoxaluria underwent hemodialysis systematically for the first 10 days after the transplantation.

Kidney transplant control population

To compare the frequency of acute renal rejection, 86 matched renal transplanted patients out of our 2005 transplanted patients were analyzed retrospectively (Table 2). They were matched for age (±5 years), date of transplantation (± 6 months), presence of PRA and the number of previous transplantations. Because HLA matching was done before KT but not LKT, due to the short delay required for LKT, HLA matching was better in the KT than the LKT group (Table 2).

Table 2. : Comparison of LKT and matched KT. Data are median (range) or number of patients. M: Male; F: Female, LKT: Liver-Kidney transplantation; KT: Kidney transplantation; HLA: Human leukocytes antigen
Number of patients4586
Median age42.942.3
Gender M/F29 / 1644 / 42
Number of second transplantations19 (42.2%)36 (41,9%)
Panel reactive antibodies12 (26.6%)22 (25.6%)
HLA matching
<133 (73.3%)4 (4.7%)
=29 (20%)15 (17.4%)
>23 (6.7%)67 (77.9%)
Renal diseases
Glomerulonephritis15 (33.3%)42 (49%)
Interstitial nephritis6 (13.3%)10 (11.8%)
Vascular nephritis1 (2.2%)2 (2.3%)
Inherited disease10 (22.2%)15 (17.4%)
Unknown13 (28.8%)17 (19.6%)
Immunosuppressive treatment
Antibody induction4586
Cyclosporin A33 (74%)68 (79%)
Tacrolimus12 (26%)18
Acute renal rejection2 (4.4%)28 (32.5%)
Grade 1222
Grade 203
Grade 303

Cross-match determination and panel-reactive antibodies

Lymphocytotoxic cross-matches were performed on isolated T and B donor cells.

Panel-reactive antibodies (PRA) were assessed by testing recipient pretransplant serum. PRA were positive if it they reacted with more than 10% of the panel cells.

Follow-up and rejection

Acute and chronic kidney rejections were suspected on clinical and biological criteria (increase in serum creatinine > 25%) and were confirmed by a graft biopsy. Acute and chronic hepatic rejections were suspected if liver transaminases or alkaline phosphatase increased, and were confirmed by a transjugular biopsy. Hepatic and renal rejections were classified according the Banff 99 and 97 criteria, respectively. Proven rejections were treated with pulses of methylprednisolone and with ATG for 10 days for corticoresistant rejection.

Statistical analysis

Survival rates of patients, kidneys and livers were determined by Kaplan-Meier analysis and compared using the Log-Rank method. Comparisons were made using Student's t-test, Chi-square tests or a model of Cox regression, depending on the type of data.


Patient survival

Forty-five patients received combined liver and kidney grafts, with a median follow-up period of 5 years (range 14 months−14 years) (Table 1). Forty-two per cent had previously been transplanted, and one patient had already been transplanted twice. The patient survival rate was 85% at 1 year and 82% at 2 years (Figure 1). Eight patients died. Most of them (n = 7) died within the first 2 months due to severe sepsis (n = 6) or arrhythmia (n = 1). One died 17 months after transplantation, due to neurovascular complications of hyperoxaluria.

Figure 1.

Kaplan–Meier curve of LKT patient survival.

Although sepsis was frequent among the patients who died during the first 2 months (n = 6; 85%), it was also frequent among those who survived (n = 27; 73%) (ns) (Table 3). The frequency of septicemia was similar for patients who died (n = 3; 42.8%) and for those who survived (n = 14; 36.8%) (ns) (Figure 2), and thus death cannot be directly attributed to severe sepsis or septicemia. Bacterial analysis indicated that methicillin-resistant Staphylococcus aureus (n = 7 vs. n = 8, p < 0.001) and fungal infections (n = 6 vs. n = 3; p < 0.001) were more frequent in patients who died within the first 2 months (early death) than in patients who survived. Candida albicans was identified in four cases and Torulopsis glabrata in two cases. These fungi occurred frequently as multifocal infection involving the urine, blood, lung and graft bed. They were frequently associated at the diagnostic time with others pyogens, such as Staphylococcus aureus, suggesting that they did not result from the use of broad-spectrum antibiotics for the treatment of a primary infection.

Table 3.  : Factors influencing patient survival after LKT. A: Univariate analysis of predictive factors for death. B: Multivariate analysis (Cox. Model). BMI:Body mass index;ns nonsignificant Thumbnail image of
Figure 2.

Infections during LKT. * p < 0.05.

Analysis of risk factors for death

Several predisposing factors may influence the patients' survival and the occurrence of severe infections. These factors include malnutrition, the previous use of immunosuppressive drugs or/and hepatic failure with splenoportal anastomosis. Most of the patients had a low body mass index (BMI), which was slightly but not significantly lower for the patients who died (Table 3A). Retrospective analysis indicated that age was similar and the frequency of a previous transplantation was not significantly different for patients who died within the first 2 months and those who survived. In addition, the survival rate was not associated with the severity of the cirrhosis according to the Child-Pugh system. The only factor associated with a higher frequency of death was being female: 31.3% of females died compared to 6.9% of males (p = 0.05) (Table 3A). However, having normal liver function and an inherited disease tended to be associated with a higher frequency of death (Table 3A). Indeed, 30% of the patients with inherited diseases (n = 10) died, compared to 11.4% of the patients with cirrhosis (n = 35) (p = 0.07). Moreover, among the eight patients with hyperoxalemia, three died. Multivariate analysis (Cox model) indicated that only having an inherited disease and the age of the recipient were independently associated with a higher risk of death (Table 3B).

Early recovery of graft functions

After transplantation, the recovery of liver function assessed by normalization of transferase activities and proconverting time occurred in all patients except one. This patient had a primary nonfunctional hepatic graft and was retransplanted on day 10. He died 2 months later due to septic shock.

Seventeen patients were dialyzed after LKT: nine for primary renal dysfunction and eight because of hyperoxaluria independent of their renal function. Most of the surviving patients recovered satisfactory renal function within 4 weeks and had a mean serum creatinine value of 130 μmol/L after 6 months (Figure 3). When patients who systematically underwent hemodialysis for the treatment of hyperoxaluria were excluded, the need for hemodialysis was associated with the age of the donor, but not with the time of cold ischemia, diuresis in the donor, serum creatinine concentration or use of inotropic drugs before transplantation (Table 4). The mean age of the donor was 43.6 ± 12.8 years (n = 9) for dialyzed patients (except patients with hyperoxaluria) and 32 ± 11.9 years (n = 28) for nondialyzed patients (p < 0.03).

Figure 3.

Mean serum creatinine concentration in LKT patients. Error bars indicate the standard error.

Table 4.  : Dialysis requirement (hyperoxalemia excluded): characteristics of donors
Evolution of recipientsDialyzedNot dialyzed
  1. Results are expressed as mean value (range). None of these results are significantly different between the two groups.

Age (year)43.5 (26-62)32 (14-50)
Renal ischemia time (hours)15.9 (13.5-18)13.6 (4.8-19)
Diuresis (mL)(50-2000)(120-1000)
Creatinemia (μmol/L)69.1 (51-89)89.8 (49-132)
Need of catecholamine (n)27

Acute graft rejections

Acute graft rejection was suspected in 15 patients with elevated concentrations of hepatic enzymes (n = 15) or serum creatinine (n = 8). Fourteen of them had grade 1 acute liver rejection revealed by biopsy (31.6%). Rejection occurred within the first 3 months post-LKT. It was not associated with a particular type of immunosuppressive therapy. Eleven of them were efficiently treated with steroids, and one also required ATG therapy. Acute hepatic rejection was not associated with previous transplantation or the presence of PRA (Table 5). Interestingly, three of the transplanted patients had a positive cross-match. One of them developed acute rejection. In all cases, a second cross-match carried out after the hepatic transplantation and before the renal engraftment was negative.

Table 5.  : Immunological status of patients presenting an acute liver graft rejection
 Acute rejection 
Number of events1431 
Positive cross matches12ns
Panel reactive antibodies
T lymphocytes7 (50%)9 (29%)ns
B lymphocytes2 (14.3%)4 (12.9%)ns
IgG5 (35.7%)6 (19.%)ns
IgM6 (42.8%)7 (22.6%)ns

During the same period, renal biopsies from patients with suspected acute renal rejection (n = 8) revealed only two cases of acute rejection (one grade 1 and one borderline) (4.2%). These two patients had concomitant acute hepatic and renal rejections. One was efficiently treated with steroid boluses and the other received ATG therapy. Both recovered satisfactory renal function. This low frequency of renal acute rejection was compared to that observed for single renal transplanted patients. As 42% of the study population had previously been transplanted or had PRA before the LKT, and because acute rejection was more frequent for immunized patients, we studied the incidence of acute rejection in 86 K matched patients (Table 2). Acute renal rejection occurred in 32.5% of the K patients. About one quarter (25.3%) had grade 1 and 7% had grade 2 or 3 acute graft rejection. Acute graft rejection was more frequent in the KT group (32.5%) than in the LKT group (4.2%) (p < 0.001).

Analysis of chronic liver and renal dysfunction

During the follow-up period, seven patients (15.5%) developed chronic renal dysfunction, identified by a sustained increase of serum creatinine. None of these patients was dialyzed or retransplanted. Analysis of graft biopsies revealed three cases of anticalcineurin drug nephrotoxicity, four cases of chronic graft rejections (8.9%) and two cases of recurrent nephrocalcinosis (type 1 hyperoxaluria) (4.4%), which occurred during episodes of dehydration, although the amounts of hyperoxaluria and the concentration of oxalemia decreased but were not normalized. For the six other patients who had had hyperoxalemia, the mean creatinemia was 137 μmol/L 6 months post transplantation. The rate of chronic dysfunction was similar for the matched renal transplanted patients (n = 13/86; 15.1%) (Table 6). Eight chronic graft rejections (9.3%) and two acute graft rejections (2.3%) were observed due to the interruption of immunosuppressive treatment for these patients. The rate of rejection was not associated with a positive cross-match or with the presence of pretransplantation PRA. Finally, the survival rate of the grafts was similar to that observed in the matched controls (Figure 4). The renal graft survival was 85%, 85%, 82% at 6 months, 1 year and 2 years, respectively, for LKT patients. It was 92%, 87% and 82% at the same times for KT patients (p = 0.62). The type of transplantation (KT vs. LKT) did not affect graft survival [p = 0.5, OR = 1.142 IC (0.775–1.682)], although an acute rejection appeared as an independent factor [p = 0.043, OR = 0.626, (0.393–0.996) (Cox model analysis)].

Table 6.  : Analysis of chronic graft dysfunction in LKT and matched KT patients
 Combined transplantationMatched renal transplantation
Total number of patients454586
Chronic dysfunction16 (35.5%)8 (17.8%)13 (15.1%)
Drug toxicity3 (6.6%)3 (6.6%)3 (3.4%)
Chronic rejection2 (4.4%)4 (8.9%)8 (9.3%)
Recurrences8 (17.8%)2 (4.4%)0
B hepatitisn=5  
C hepatitisn=3  
De novo diseases3 (6.6%)02 (2.3%)
Others005 (5.8%)
Figure 4.

Kaplan–Meier curve of LK and K graft survival.

Nineteen cases of chronic liver dysfunction were observed (42.2%). Azathioprine treatment had to be discontinued in two patients with peliosis and one with nodular regenerative hypertrophia (Table 6), and two patients exhibited chronic graft rejection. Two patients developed de novo hepatitis C, one had an alcohol-related hepatic disease and eight had recurrent viral hepatitis. In most cases the viral disease recurred during the first 6 months. Five of the 15 patients (30%) with hepatitis B had recurrence of the disease on the graft, whereas none of them had detectable viral B DNA in blood before transplantation. Three of the 14 patients (21.5%) with viral hepatitis C had recurrence. None of them was positive for hepatitis C DNA after the transplantation, but three of patients who were positive for the virus before transplantation did not have recurrence of the disease. In conclusion, there was no correlation between positive PCR and recurrence of the disease.


We report here the long-term follow-up of 45 LKT patients. We have shown that the overall survival was 85% at 1 year and 82% at 2 years. Whereas previous studies have reported a high mortality rate, recent studies by Gonwa et al. and Katznelson & Cecka reported similar results with an overall survival rate of 76–79% at 3 years (5,15). These results are partly due to the precise evaluation and selection of patients, excluding those suffering from severe malnutrition, advanced cardiovascular disease or by limiting advanced cirrhosis.

In our study, death was most common during the first 3 months post transplantation. Death was often associated with severe infections. However, high frequencies of infection, severe sepsis or septicemia were also observed in patients who survived, and thus these factors cannot be direct causes of death. However, a detailed analysis of infection showed that disseminated yeast infections or Staphylococcus aureus infections were significantly more frequent in the patients who died. Disseminated fungal infection was concomitant to bacterial infection, suggesting that it was not selected by the broad-spectrum antibiotics used to treat bacterial infections. This indicates that fungal infections should be systematically detected and treated as soon as they are identified. Interestingly, we observed a high incidence of methicillin-resistant Staphylococcus aureus, as observed for patients with a chronic preterminal or terminal renal failure. Therefore, it is important to determine whether patients are colonized with this pathogen before transplantation. This suggests that the initial empirical treatment for sepsis should be anti-staphylococcal drugs such as vancomycin or teicoplanin, rather than methicillin.

Several predisposing factors may be associated with death. These include nutritional status, the previous use of immunosuppressants and the age of the patient. None of these factors was found to be associated with the mortality in our study. Surprisingly, only gender was found to be associated with death. Females have a higher risk of death than males according to the monovariate analysis. This was not confirmed by the multivariate analysis, possibly because the patients with inherited disease were mostly women (see below). Unexpectedly, patients with inherited diseases and with normal hepatic functions had a higher risk of death. Two of them died within the first 2 months, and one developed cerebral vascular thrombosis due oxalate deposit at 17 months post transplantation. This may be due to the high frequency of vascular lesions in patients with primary hyperoxaluria and heart amyloidosis. Indeed, most of them had been on dialysis for several years and had tophi, bone lesions and vascular deposits.

It has been suggested that LKT leads to a reduction in acute renal, but not hepatic rejection (5,6,8). In this study, 26% of LK patients displayed acute liver rejection. This rate was lower than that reported for single liver transplantation (17), but was similar to that observed in our center for transplanted patients with alcoholic cirrhosis or hepatitis B cirrhosis treated by long-term anti-hepatitis B surface antigen immunoglobulins (17). However, we did not observe any differences in the frequency of hepatic acute rejection for these two subgroups of patients compared to other LK patients with cirrhosis due to hepatitis C or inherited disease. This may be due to the immunosuppressive protocol used, which was based on the sequential association of ATG, steroids, azathioprine and calcineurin inhibitors, whereas the immunosuppressive protocols usually used for liver transplantation do not include ATG (6,8). Alternatively, LKT might reduce acute liver and renal rejection. Our results for the kidney were consistent with those of previous studies with smaller numbers of patients (3–5,7,8,14), indicating that the frequency of acute renal rejection was low in LKT patients (4.2%). In contrast, the study by Katznelson & Cecka, which was based on the American register data-base, reported that the frequency of kidney graft rejection for LKT was equally as high as in patients transplanted with the contralateral kidney (15). However, no information was available on the immunosuppressive drugs used, the way in which these drugs were monitored or the immunological status of the patients. As graft acceptance is mainly dependent on the immunological status of the recipient, and as up to 40% of our patients had previously been transplanted and/or had panel-reactive lymphocytotoxic antibodies (26.6%), we compared the frequency of acute renal rejection in LK patients with that in matched renal transplanted patients. We observed that the frequency of renal acute graft rejection was significantly lower in the study group than in the group of matched patients (4.2% vs. 32.5%), although they received similar immunosuppressive treatment. Thus, we decided to reduce the dosage of calcineurin inhibitors for LK patients since 2 years in order to obtain basal tacrolimus levels between 5 and 7 ng/mL without rejection. This suggests that LKT can prevent acute renal rejection, as reported for experimental combined transplantation of the pancreas-intestine or the pancreas with the liver. This occurs via a mechanism of tolerance which remains to be determined. Several mechanisms have been proposed, including the production of soluble HLA class I antigens by the liver allograft, which may neutralize both pre-existing allo-antibodies and cytotoxic T lymphocytes (18), the secretion by the liver of some immunomodulatory factors or cytokines that will favor a Th2 immune response (19), and the possibility of the release of cells such as donor leukocytes (19), stem cells or immature dendritic cells from the allogeneic liver graft that will educate the immune system (20,21) or regulatory T cells, such as the CD4+CD25+ population (22), that will modulate the immune response.

Another hypothesis is that alterations in liver function may be a more sensitive marker of acute rejection than a rise in serum creatinine. However, this is unlikely. Indeed, we observed that eight patients had simultaneous renal and liver biopsies for liver and renal dysfunction. Only two cases of kidney rejection were associated with liver rejection, suggesting that the occurrence of liver dysfunction does not mask acute renal rejection. However, we cannot exclude the possibility that the graft rejection began in the liver and was not initially detectable in the kidney. This may be due to the important antigenic mass of the liver and/or due to its cellular composition, which includes antigen-presenting cells required to induce an allograft rejection.

As for liver transplantation, LKT was performed without cross-matching, but the ABO blood-group compatibility system was respected. Cross-matches were carried out secondarily and three of our patients had a positive cross-match. As previously reported (14,23–25), none of them developed hyperacute or acute rejection. In all cases, a second cross-match was carried out after the liver transplantation but before the renal transplantation and had become negative, as reported by Morrissey et al. (14). This might explain the absence of renal humoral acute rejection and suggests that most of the antibodies were trapped within the liver without affecting its function. Recent reports have shown that positive cross-matches and panel-reactive lymphocytotoxic antibodies favor acute rejection in liver-transplanted patients via an antibody-driven mechanism (26–28). In addition, in our patients, we did not find any relationship between positive cross-matches or panel-reactive lymphocytotoxic antibodies and hyperacute, acute or chronic rejection. This is consistent with the results reported by Morrissey et al. for LK patients (14).

In this report we have also shown that graft survival was excellent. Apart from the patients who died soon after the transplantation and one patient who died due to a vascular complication related to chronic hyperoxaluria, all of the patients had functional grafts. None of them was retransplanted or dialyzed, within the mean follow-up period of 5 years. However, 15.5% of our patients had chronic renal dysfunction, mostly due to chronic renal graft rejection (n = 4) or renal toxicity of calcineurin inhibitors (n = 3). These results are similar to those observed for the matched renal transplanted population, suggesting that LKT does not prevent chronic rejection. In two of the patients transplanted for primary hyperoxaluria, who had massive oxalic acid bone deposits, chronic graft dysfunction was due to recurrence of the renal lithiasis. The prevention of recurrence by maintaining an important diuresis should be sustained for a long time since hyperoxaluria tends to reach normal values only 2 years post-transplantation. The survival of the liver was also good. Only one patient was retransplanted during the early post-transplantation stage. However, 35.5% of patients developed chronic liver dysfunction, mostly due to the recurrence of viral disease (n = 8; 17.8%) or azathioprine toxicity (n = 3), but rarely due to chronic rejection. It is noteworthy that three out of the 15 patients transplanted for chronic hepatitis B, but who had no detectable circulating hepatitis B DNA, had recurrent disease. Three of the patients with viral hepatitis C were positive for viral DNA before the transplantation. One had recurrence of the disease, whereas two patients who were initially negative developed the disease. This suggests that the recurrence of the disease is not related to the weak replication of the virus, as reported for simple hepatic transplantation. This also indicates that patients with combined hepatic and renal failure due to viral hepatitis can be transplanted even if they have weak viral replication.

Altogether, these results indicate that the survival rate of grafts in LK patients is satisfactory, with few cases of acute renal rejection compared to K patients, even in retransplanted patients or patients with panel-reactive lymphocytic antibodies or positive cross-matches.